MUSCLE PHYSIOLOGYAND BIOCHEMISTRY

Muscle Physiology and Biochemistry

Edited by

SHOICHIIMAI MAKOTOENDO

Department of Pharmacology Saitama Medical School

Niigata University School of Medicine Moroyamamachi
No. 757, Asahimachi-dori 1 Saitama 350-04
Niigata 951, Japan Japan

IWAO OHTSUKI

Faculty ofMedicine
Kyushu University
Fukuoka 812-82
Japan

Reprinted from Molecular and Cellular Biochemistry, Volume 190 (1999)

Springer-Science+Business Media, B.V.

Library of Congress Cataloging-in-Publication Data

Musc1e physiology and biochemistry/edited by Shoichi Imai, Makoto

Endo, Iwao Ohtsuki.
p. cm. -- (Developments in molecular and cellular
biochemistry)
ISBN 978-1-4613-7534-0 ISBN 978-1-4615-5543-8 (eBook)
DOI 10.1007/978-1-4615-5543-8
1. Musc1es--Physiology. 2. Musc1e contract ion. 3. Musc1es-.
-Molecular aspects. 4. Musc1es--Metabolism. 1. Imai, Shoichi,
1931- . II. Endo, Makato. 1933- . III. Ohtsuki, Iwao.
IV. Series.
QP321.M8917 1998
616.7' 4--dc21 98-36698

ISBN 978-1-4613-7534-0

Printed an acid-free paper

Ali rights reserved

© 1999 Springer Science+Business Media Dordrecht

Originally published by K1uwer Academic Publishers in 1999
Softcover reprint of the hardcover 1st edition 1999

No part ofthe material protected by this copyright notice may be reproduced or

utilized in any farm or by any means, electronic or mechanical,
inc1uding photocopying, recarding ar by any information storage and
retrieval system, without written permis sion from the copyright owner
Molecular and Cellular BiochelDistry:
An International Journal for Chemical Biology in Health and Disease
CONTENTS VOLUME 190, Nos. 1 & 2, January (I) 1999
MUSCLE PHYSIOLOGY AND BIOCHEMISTRY
Shoichi Imai, Makoto Endo and Iwao Ohtsuki

Preface I
M. Endo: Dedication 3-4
J. Gergely: Professor Ebashi' s impact on the study ofthe regulation of striated muscle contraction 5-S
S.Y. Perry: Troponin I: Inhibitor or facilitator 9-32
I. Ohtsuki: Calcium ion regulation ofmuscle contraction: The regulatory role oftroponin T 33-38
K. Yamada: Thermodynamic analyses ofcalcium binding to troponin C, calmodulin and parvalbumins by using microcalorimetry 39-45
M. Yazawa, K.-i. Nakashima and K. Yagi: A strange calmodulin of yeast 47-54
A.G. Szent-Gyorgyi, Y.N. Kalabokis and C.L. Perreault-Micale: Regulation by molluscan myosins 55-62
Y. Yazawa and M. Kamidochi: The properties and function of invertebrate new muscle protein 63-66
A. Weber: Actin binding proteins that change extent and rate of actin monomer-polymer distribution by different mechanisms 67-74
M. Tanokura and Y Suzuki: A phosphorus-31 nuclear magnetic resonance study on the complex ofchicken gizzard myosin subfragment
I with adenosine diphosphate 75-78
DJ. Hartshorne and K. Hirano: Interactions of protein phosphatase type 1, with a focus on myosin phosphatase 79-84
K. Fujita, L.-HYe, M. Sato, T. Okagaki, Y. Nagamachi and K. Kohama: Myosin light chain kinase from skeletal muscle regulates
anATP-dependent interaction between actin and myosin by binding to actin 85-90
T. Murahashi,A. Fujita and T. Kitazawa: Ca 2+-induced Ca 2+ desensitization ofmyosin light chain phosphorylation and contraction in
phasic smooth muscle 91-98
T. Masuda, K. Ohmi, H. Yamaguchi, K. Hasegawa, T. Sugiyama, Y. Matsuda, M. lino and Y. Nonomura: Growing and differentiating
characterization ofaortic smooth muscle cell line, p53LMACO 1 obtained from p53 knock out mice 99-104
K. Sobue, K. Hayashi and W. Nishida: Expressional regulation of smooth muscle cell-specific genes in association with phenotypic
modulation 105-118
I. Niki and H. Hidaka: Roles of intracellular Ca2+ receptors in the pancreatic ~-cell in insulin secretion 119-124
Y. Soeno, H. Yajima, Y. Kawamura, S. Kimura, K. Maruyama and T. Obinata: Organization ofconnectinltitin filaments in sarcomeres
of differentiating chicken skeletal muscle cells 125-131
K.-i. Kusano, H. Abe and T. Obinata: Detection of a sequence involved in actin-binding and phosphoinositide-binding in the N-
terminal side ofcofilin 133-141
E. Ozawa, Y. Hagiwara and M. Yoshida: Creatine kinase, cell membrane and Duchenne muscular dystrophy 143-151
T. Masaki, H. Ninomiya,A. Sakamoto and Y. Okamoto: Structural basis of the function of endothelin receptor 153-156
Y. Yoshida,A. Toyosato, M.O. Islam, T. Koga, S. Fujita and S. Imai: Stimulation of plasma membrane Ca 2+-pump ATPase ofvascular
smooth muscle by cGMP-dependent protein kinase: Functional reconstitution with purified proteins 157-167
H. Yamamoto and M. Kawakita: Chemical modification ofan arginine residue in theATP-binding site ofCa2+-transportingATPase
of sarcoplasmic reticulum by phenylglyoxal 169-177
M. Hirata, M. Yoshida, T. Kanematsu and H. Takeuchi: Intrinsic inhibitor of inositol I ,4,5-trisphosphate binding 179-184
M. lino: Dynamic regulation of intracellular calcium signals through calcium release channels 185-190
Y. Ogawa, T. Murayama and N. Kurebayashi: Comparison of properties ofCaH release channels between rabbit and frog skeletal
muscles 191-201

Index to Volume 190 203-204

Molecular and Cellular Biochemistry 190: 1, 1999.

Preface

The papers in this issue were contributed by close friends, the publication of this issue possible. Owing to some un-
coworkers and pupils of Professor Setsuro Ebashi. They are expected troubles of one ofthe editors (M. E.) the publication
dedicated to him to commemorate his great and pioneering of this issue has been greatly delayed, for which he sincerely
contribution to the advancement of muscle physiology and apologizes to all the contributors and other editors. We
biochemistry, which in course of time exerted a great in- believe that this issue reveals the present state of research on
fluence on the whole field of life science. We would like to muscle and/or calcium that had been opened up by Professor
express our cordial thanks to an the contributors who made Ebashi.

Shoichi Imai, Niigata, Japan

Makoto Endo, Saitama, Japan
Iwao Ohtsuki, Fakuoka, Japan
Molecular and Cellular Biochemistry 190: 3-4, 1999.

Dedication

Setsuro Ebashi was born in Tokyo on 31 st August, 1922.

There is a Japanese saying that 'Sandalwood is fragrant even
in seed leaf.' Genius displays itself even in childhood.
Finishing the six-year course ofprimary school in five years
and the five-year course of middle school in four years, he
entered the First High School, the most prestigious high
school in Japan, at the age of only 15.
In July 1942, when he was an undergraduate student of
Faculty of Medicine, Tokyo Imperial University (now called
University ofTokyo), he by chance visited the laboratory of
Dr. Hiroshi Kumagai, at that time Lecturer in Pharmacology,
to have a practical training during the summer vacation. This
was the beginning of an admirable relationship of love and
kindness between a pupil and a teacher as well as the start of
Dr. Ebashi' s muscle research. However, the War severed the
relationship: the teacher went to Indonesia to teach in Jakarta
Medical School, and the pupil received his M.D. degree in
1944 and served in the war as a naval surgeon. When Dr.
Ebashi was demobilized in 1946, he went straight to Dr.
Kumagai's laboratory again.
Dr. Ebashi's research was at first electrophysiology of research was right. Having inquired further into the relaxing
smooth muscle in which Dr. Kumagai had a deep interest. factor, he demonstrated in 1955 that the essential component
However, in 1950 Dr. Ebashi was deeply impressed with a J. of the relaxing factor was in the particulate fraction, against
Physiol. paper by Hodgkin and Katz (1949) which completely the general beliefat that time that it may beATP-regenerating
elucidated the mechanism of excitation as he felt. At about soluble enzyme(s).
the same time he was also deeply inspired by a book As for the mechanism of relaxation by the relaxing factor,
'Chemistry of Muscular Contraction' by A. Szent-Gyorgyi once again against the general belief at that time that the
(1949). These readings led him to change the subject of his relaxing factor might produce some (organic) substance
research to the contractile mechanisms. which in tum acts on the actomyosin system to cause relaxa-
He raised the following question. Although Szent-Gyorgyi tion, Dr. Ebashi showed in the early 60s that removal ofCa 2+
demonstrated thatATP added to the actin-myosin system such ion from the medium by the relaxing factor is the cause of
as actomyosin thread or glycerinated muscle induces con- relaxation. His evidence consisted oftwo important discoveries
traction, removal ofATP does not cause relaxation, which is that the particulate relaxing factor strongly accumulates Ca2+
quite different from, for example, acetylcholine-induced ion from the medium in the presence ofATP, and that a minute
contraction of living muscle, where the removal of acetyl- amount of Ca 2+ ion is necessary for the contractile reaction
choline causes relaxation. His idea was that there must be of well-washed Ca2+-free natural actomyosin system. Al-
something in living muscle to cause relaxation, which was though physiologists had recognized the contraction-
lost and absent in the actomyosin systems. He started to inducing action of Ca2+ ion, it had not been recognized by
search for the relaxing factor in homogenized muscle and muscle biochemists before Dr. Ebashi, because all the
soon he found the factor and reported to a meeting of a biochemical experiments were done in the presence of
Japanese muscle physiology group in 1952. Sometime after sufficient amount of Ca2+ion contaminated from reagents or
this Dr. Kumagai found a paper by Marsh in Nature (1951) exuded from glasswares. Dr. Ebashi further demonstrated
that had already reported the same factor. However, this was electronmicroscopically that the relaxing factor has a vesi-
not a disappointment for young Dr. Ebashi but rather an cular structure, indicating that it is the fragment ofthe sarco-
encouragement because it proved that his direction of plasmic reticulum (SR). Since relaxation is the reverse of
Molecular and Cellular Biochemistry 190: 5-8, 1999.
© 1999 Kluwer Academic Publishers.

Professor Ebashi's impact on the study of the

regulation of striated muscle contraction

John Gergely
Muscle Research Group, Boston Biomedical Research Institute; Department ofNeurology, Massachusetts General Hospital;
Department ofBiological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, MA, USA

Abstract
The field of striated muscle regulation has changed tremendously over the last forty years. Many of the problems solved by Dr.
Ebashi and by those stimulated by him offer new challenges for future generations of scientists. Many questions remain to be
solved, and it should give particular pleasure to Dr. Ebashi to see how the seeds sown by him and his col1eagues have now
grown into a beautiful tree that bears rich fruit at present and will continue to do so for a long time in the future. (Mol Cel1
Biochem 190: 5-8, 1999)

Key words: troponin, tropomyosin, thin filament regulation, Ca 2+

Introduction interaction of actin and myosin in the presence of ATP Ca2+

sensitive. Native tropomyosin was soon separated into two
I am indeed greatly pleased and honored to be able to join in components: tropomyosin and a new entity that became
the celebration of the remarkably productive and influential known as troponin. Troponin was identified as the receptor
contribution of Professor Ebashi's life in science. This gives for Ca 2+, whose role in actomyosin activation had earlier
me particular pleasure since so much of what my colleagues been established [4, 5] with four Ca-binding sites in troponin
and I were able to do in the last four decades has depended [6]. There were indications that troponin is a multi-component
on his contributions. system [7]. The work initiated in the Ebashi laboratory
Dr. Ebashi became known very early in his career by started a new era in muscle research involving many
establishing the nature ofthe so-called relaxing factor. This was laboratories all over the world. It became clear that there
followed by the identification of calcium ions as key mes- was a calcium binding component and another one that
sengers in the activation process of muscle contraction and inhibited the Mg 2+ stimulatedATPase ofpurified actomyosin
the discovery of the troponin complex which, together with [8,9]. By 1972 a general consensus was reached that troponin
tropomyosin, was identified as the actin-bound regulatory consists of three subunits [10-13] whose names indicate their
system of striated muscle (See [I]). In this brief review (For roles: viz. troponin C (TnC), troponin I (Tnl), and troponin T
detailed reviews see: [1,52-57]) I should like to trace some (TnT) for calcium binding, inhibition, and troponin binding,
of the developments concerning the regulation of striated respectively.
muscle that were sparked by the ground-breaking work of Dr. The availability of purified components oftroponin made
Ebashi and his colleagues and to point to some questions that it possible to build on the findings that emerged in the earlier
currently await answers. stages. Calcium binding studies [14], utilizing the calcium
buffer system originally developed by Dr. Ebashi and his
colleagues [5], located four calcium binding sites in TnC and
The discovery of troponin showed that there are two classes each containing two
calcium binding sites. Two sites bind calcium with high
The new component discovered in Dr. Ebashi's laboratory affinity, as well as Mg2+ although with lower affinity, while
was first known as native tropomyosin [2,3]. It rendered the the other two sites ofTnC are essentially specific for calcium.

Address/or offprints: J. Gergely, Muscle Research Group, Boston Biomedical Research Institute, 220 Staniford Street, Boston, MA 02114-2500, USA
6

Troponin C - The Ca2+ receptor ATPase activity. Distance determinations by resonance

energy transfer between probes on appropriately placed
While research on troponin began to flourish studies on a engineered Cys residues showed a Ca 2+-induced change
related protein led to important results. Parvalbumin, which corresponding to the expectations based on the model [24].
had first been found in fish muscle as a cytoplasmic rather Finally, solution ofthe high resolution NMR structure ofTnC
than a myofibrillar protein, was characterized both in terms with four Ca2+ bound [25] brought definitive proof for the
of primary structure and crystal structure as having two Ca 2+ postulated structure. The opening of the N terminal domain
binding sites [15]. This work led to the concept of the of TnC may be considered as the molecular switch in TnC.
so-called EF hands suggested by a bent middle finger, the The NMR structure revealed some differences between helix
thumb and index finger as depicting a calcium binding loop B in the N-terminal domain and the corresponding helix in
flanked by two a helices as the model of calcium binding sites the C terminal domain even when both sites in each domain
originally found in parvalbumin and by now known to occur were occupied by Ca2+. It also pointed to some flexibility of
in large super families of Ca 2+binding proteins. the central helix in solution. A recent comparison ofthe high
When the amino acid sequence ofTnC became known [16] resolution NMR structures of cardiac and skeletal TnC in the
a high degree of homology with parvalbumin was recognized 4-Ca2+ state shows that the extent of opening of the hydro-
and the Ca2+binding sites were identified. A variety ofstudies phobic surface is much less in the case of cardiac muscle, a
have shown that sites I and II in the N-terminal domain are finding whose full implications are yet to be explored [26].
Ca-specific sites; sites III and IV are the high affinity Ca-Mg The opening of the N terminal domain of TnC may be
sites in the C domain. The former are recognized as the func- considered as the molecular switch in TnC.
tionally important triggering sites (see [17]) and references
therein).
The similarities between the TnC and parvalbumin struc- The molecular switch in troponin I
tures led to speculations about how two parvalbumin-like
halves could be fitted into the structure ofTnC [18]. When the The next question that has received some partial answers over
structure ofTnC was solved by x-ray crystallography [19, 20], the years is the status of the molecular switch in Tn!. In the
it showed two domains - each a parvalbumin-like structure - absence ofhigh resolution structures for TnI and its complexes
connected, however, by a single a helix instead of a compact such answers must remain tentative. There is evidence that
molecule essentially containing two paravalbumin-like portions of Tnl move under the influence of activation from
structures. TnC to actin, and under conditions corresponding to relaxation
they return to TnC. One of the sites that has been used is
cysteine 133 [27] and current studies are further exploring
The molecular switch in troponin C movements in Tnl by labeling cysteine residues introduced by
genetic engineering as has been done in the case ofTnC. There
An important step toward our current understanding of the are some not fully answered questions concerning the relation
chain of events initiated by calcium binding to the triggering of the region that comes into close contact with actin and the
sites in TnC came from insights ofHerzberg et al. [21] gained so-called inhibitory region that emerged in earlier studies and
in comparing the structure of the N- and C-terminal homo- contains the stretch of residues 96-116 in Tnl [28]. Evidence
logous domains in TnC. Owing to the conditions of crystal- is accumulating that this inhibitory region is indeed inter-
lization the former contained no bound calcium, while the acting with both domains of TnC [29-31] but its mode of
latter had two sites occupied by Ca 2+. Thus the difference interaction with actin needs further elucidation. During recent
between the two domains would give a clue to the con- years a reasonable consensus has emerged concerning the
formational changes brought about by calcium when it overall arrangement of the polypeptides in TnC- and Tnl
becomes bound to the N terminal sites. This led to the relative to each other. Both crosslinking [32, 33] and fragment
suggestion that the connector between helices Band C binding [34] studies suggest that the two chains run in
together with the link between them moves away from helix opposite directions; that is, the N terminus of TnC interacts
D which is part ofthe long helix connecting the two domains, mainly with the C terminus of TnI and vice versa. However,
exposing a hydrophobic area which was presumed to become evidence is also at hand indicating that within Tnl certain
an interacting site with Tn!. Soon thereafter various pieces stretches may run locally in opposite directions while the
of evidence emerged for this view. Site directed mutagenesis overall trend is preserved. Recent work on troponin with Tnl
of charged residues [22] or disulfide formation between containing only Cys 133 and Cys 48 thiols for placement of
genetically engineered Cys residues [23], in segments whose probes for resonance energy transfer studies has shown metal
separation was expected to change upon Ca 2+ binding dependent conformational changes in Tnl modulated by the
according to the model, led to changes in Ca 2+-binding and interaction oftroponin with actin [35].
 7

Research on certain aspects of TnI/TnC interaction has the question of how changes in solution are related to those
been stimulated by studies on calmodulin, which is an in the actin filament itself have been intriguing (see [45] and
activator of a large number of enzymes. In light of the close references therein). Recent x-ray diffraction studies on
similarities between TnI and calmodulin with respect to their reconstituted actomyosin gels and on muscle fibers have
chemical and crystallographic structure the question arises thrown new light on tropomyosin movement associated with
whether the structural changes occurring when TnI binds to Ca 2+ activation [46-48). It appears that in the regulated thin
TnC are similar to those taking place on the interaction of filament Ca 2+ binding to troponin is accompanied by a
calmodulin with one of its target proteins. In the case of movement about 30° azimuthally towards the central groove
calmodulin, both x-ray diffraction [36] and multidimensional from a position where it would block strong myosin binding,
NMR [37] studies showed that at least with the M 13 peptide according to the current model of myosin-actin interactions
derived from myosin light chain kinase - a well known target (see [49]). Binding of myosin, which would take place to a
of calmodulin playing a role in the activation of smooth site partially unblocked by calcium, causes a small but
muscle contraction - is accompanied by a large structural significant further change in the position of tropomyosin,
change in calmodulin bringing the two globular domains consistent with a cooperative role of myosin - first pointed
homologous to those in TnC close together. As far as the to by A. Weber and her colleagues [50] - in the full activation
TnI'TnC complex is concerned, evidence points in the of the thin filament. This two step model of activation seems
opposite direction indicating an essentially extended to be in harmony with the three state model based on kinetic
structure for TnC based both on resonance energy transfer studies in solution, the third state being the Ca 2+ free,
distance determinations [38] and low angle x-ray and 'blocked' state [51].
neutron diffraction studies [39,40]. The latter studies also
point to the existence of masses derived from TnI beyond
the Nand C terminal domains of TnC, a picture whose References
details remain to be filled in in terms of the course of the
polypeptide chain in Tn!. Students of this field are eagerly I. Ebashi S, Endo M: Calcium ion and muscle contraction. [Review].
awaiting a more definitive x-ray crystallographic study of Prog Biophys Mol Bioi 18: 123-183,1968
the intact troponin complex. Very recently crystallization of 2. Ebashi S: Third component participating in the precipitation of 'natural
actomyosin'. Nature 200: 1010, 1963
a complex between an N-terminal fragment of TnI and TnC
3. Ebashi S, Ebashi F: A new protein component participating in the
has been achieved and the high resolution structure derived superprecipitation of myosin. J Biochem 55: 604-613,1964
from x-ray diffraction reported [41]. This work provides some 4. Weber A: On the role of calcium in the activity of adenosine-5'-
interesting interim results pending the availability of crystals triphosphate hydrolysis by actomyosin. J Bioi Chern 234: 2764-2769,
of the full complex. 1959
5. Ebashi S: Calcium binding and relaxation in the actomyosin system. J
Biochem 48: 150---151, 1960
6. Ebashi S, Ebashi F, KodamaA: Troponin as the Ca 2+-receptive protein
Troponin T in the contractile system. J Biochem 62: 137-138, 1967
7. Ebashi S, Wakabayashi T, Ebashi E: Troponin and its components. J
Biochem 69: 441-445, 1971
Although the shape of TnT has been long established starting
8. Hartshorne OL, Perry SV, Schaub MC: A protein factor inhibiting the
with the immuno-electron microscopic demonstration that the magnesium-activated adenosine triphosphatase of desensitized
globular C-terminal portion and the highly a-helical N- actomyosin. Biochem J 104: 907-913,1967
terminal portion of TnT occupy distinguishable sites along 9. Hartshorne OJ, Mueller H: Fractionation oftroponin into two distinct
tropomyosin [42], little is known about the role of TnT in the proteins. Biochem Biophys Res Comm 31: 647--653, 1968
10. Greaser ML, Gergely J: Reconstitution oftroponin activity from three
mechanism of regulation. While early work has considered
protein components. J Bioi Chern 246: 4226-4233, 1971
TnT mainly as an anchor tying the rest of the complex to II. Greaser ML, Gergely J: Purification and properties of the components
tropomyosin, hence the suffix T to troponin, recent evidence from troponin. J Bioi Chern 248: 2125-2133,1973
assigns a more active role to TnT serving as a signal transmitter 12. Ebashi S: Separation oftroponin into its three components. J Biochem
between TnC and TnI [43] as well as modulating the effect 72: 787-90, 1972
13. Perry SV, Cole HA, Head JF, Wilson FL: Localization and mode of
of myosin heads on the activity ofthe troponin complex [44).
action of the inhibitory protein component of the troponin complex.
Cold Spring Harbor. Symp. Quant Bioi 37: 251-262, 1972
14. Potter JO, Gergely J: The calcium and magnesium binding sites on
Filament regulation and cooperativity troponin and their role in the regulation of myofibrillar adenosine
triphosphatase. J BioI Chern 250: 4628-4633, 1975
15. Kretsinger RH, Nockolds CE: Carp muscle calcium binding protein.
From the earliest days of the identification of the participants II. Structure determination and general description. J Bioi Chern 248:
in the regulatory machinery, viz. troponin and tropomyosin, 3313-3326, 1973
8

16. Collins lH, Potter lD, Hom Ml, Wilshire G, lackman N: The amino 36. Meador WE, Means AR, Quiocho FA: Target enzyme recognition by
acid sequence of rabbit skeletal muscle troponin C: Gene replication calmodulin: 2.4 A structure of a calmodulin-peptide complex. Science
and homology with calcium-binding proteins from carp and hake 257: 1251-1255,1992
muscle. FEBS Lett 36: 268-272, 1973 37. Ikura M, Clore GM, Gronenborn AM, Zhu G, Klee CB, Bax A:
17. Grabarek Z, Drabikowski W, Leavis PC, Rosenfeld SS, Gergely L: Solution structure of a calmodulin-target peptide complex by multi-
Proteolytic fragments oftroponin C. Interactions with the other troponin dimensional NMR. Science 256: 632-638, 1992
subunits and biological activity. 1 Bioi Chern 256: 13121-13127, 1981 38. Gong B-1, Wang Z, Tao T, Gergely J: Troponin C remains extended in
18. Kretsinger RH, Barry CD: The predicted structure ofthe calcium-binding the ternary trroponin complex. Biophys 1 66: A346, 1994
component oftroponin. Biochim Biophys Acta 405: 40-52,1975 39. Olah GA, Rokop SE, Wang C-LA, Blechner SL, Trewhella J: Troponin
19. Sundaralingam M, Bergstrom R, Strasburg G, Rao ST, Roychowdhury I encompasses an extended troponin C in the Ca'+-bound complex - a
P, el al.: Molecular structure oftroponin C from chicken skeletal muscle small-angle X-ray and neutron scattering study. Biochemistry 33:
at 3-angstrom resolution. Science 227: 945-948, 1985 8233-8239, 1994
20. Herzberg 0, lames MN: Structure of the calcium regulatory muscle 40. Olah GA, Trewhella J: A model structure ofthe muscle protein complex
protein troponin-C at 2.8 A resolution. Nature 313: 653-659, 1985 4-Ca'+-troponin C-troponin I derived from small-angle scattering data
21. Herzberg 0, Moult 1, lames MN: A model for the Ca'+-induced - implications for regulation. Biochemistry 33: 12800-12806, 1994
conformational transition of troponin C. A trigger for muscle 41. Vassylyev DG, Takeda S, Wakatsuki S, Maeda K, Maeda, Y: Crystal
contraction. 1 Bioi Chern 261: 2638-2644, 1986 structure of troponin C in complex with troponin I fragment at 2.3A
22. Fujimori K, Sorenson M, Herzberg 0, Moult 1, Reinach FC: Probing resolution. Proc Natl Acad Sci (USA) 95: 4747--4852, 1998
the calcium-induced conformational transition of troponin C with 42. Ohtsuki I: Molecular arrangement oftroponin-T in the thin filament. J
site-directed mutants. Nature 345: 182-184, 1990 Biochem 86: 491--497,1979
23. Grabarek Z, Tan RY, Wang 1, Tao T, Gergely J: Inhibition of mutant 43. Potter JD, Sheng Z, Pan BS, Zhao J: A direct regulatory role for
troponin C activity by an intra-domain disulphide bond. Nature 345: troponin T and a dual role for troponin C in the Ca'+ regulation of
132-135,1990 muscle contraction. 1 BioI Chern 270: 2557-2562, 1995
24. Wang Z, Gergely 1, Tao T: Characterization of the Ca'+-triggered 44. Schaertl S, Lehrer SS, Geeves MA: Separation and characterization
conformational transition in troponin C. Proc Natl Acad Sci USA 89: of the two functional regions of troponin involved in muscle thin
11814-11817, 1992 filament regulation. Biochemistry 34: 15890-15894, 1995
25. Siupsky CM, Sykes BD: NMR solution structure of calcium-saturated 45. Potter JD, Gergely 1: Troponin tropomyosin and actin interactions in
skeletal muscle troponin C. Biochemistry 34: 15953-15964, 1995 the Ca2+ regulation of muscle contraction. Biochemistry 13:2697-2703,
26. Gagne SM, Li MX, McKay RT, Sia SK, Spyracopoulos L, el al.: The 1974
calcium induced structural change in that triggers skeletal and cardiac 46. Lorenz M, Poole KJV, Popp D, Rosenbaum G, Holmes KC:An atomic
muscle contraction. Biophys 1 72: A332, 1997 model of the unregulated thin filament obtained by X-ray fiber
27. Tao T, Gong BJ, Leavis PC: Calcium-induced movement oftroponin-l diffraction on oriented actin-tropomyosin gels. J Mol Bioi 246: 108-
relative to actin in skeletal muscle thin filaments. Science 247: 1339- 119,1995
1341,1990 47. Poole KJV, Evans G, Rosenbaum G, Lorenz M, Holmes KC: Thc effect
28. Syska H, Wilkinson 1M, Grand RJ, Perry SV: The relationship between of crossbridges on the calcium sensitivity of the structural change in
biological activity and primary structure of troponin I from white the regulated thin filament. Biophys J 68: A365, 1995
skeletal muscle of the rabbit. Biochem J 153: 375-387, 1976 48. Holmes KC: The actomyosin interaction and its control by tropo-
29. Leszyk 1, Grabarek Z, Gergely J, Collins JH: Characterization of myosin. Biophys 168: S2-S7, 1995
zero-length cross-links between rabbit skeletal muscle troponin C and 49. Rayment 1, Holden HM, Whittaker M, Yohn CB, Lorenz M et al.:
troponin I: Evidence for direct interaction between the inhibitory region Structure of the actin-myosin complex and its implications for muscle
oftroponin I and the NH,- terminal, regulatory domain oftroponin C. contraction. Science 261: 58-65, 1993
Biochemistry 29: 299-304, 1990 50. Bremel RD, Murray JM, Weber A: Manifestations of cooperative
30. Pearlstone lR, Smillie LB: Evidence for two-site binding oftroponin behavior in the reglated actin filament during actin activated ATP
I inhibitory peptides to the Nand C domains of troponin C. Bio- hydrolysis in the presence of calcium. Cold Spring Harbor. Symp Quant
chemistry 34: 6932-6940, 1995 BioI 37: 267-275,1972
31. Pearlstone JR, Sykes BD, Smillie LB: Interactions of structural C- 51. McKillop D, Geeves MA: Regulation of the interaction between actin
domain and regulatory N-domains of troponon C with repeated and myosin subfragment-I - evidence for 3 states of the thin filament.
sequence motifs in troponin I. Biophys J 72: A331, 1997 Biophys J 65: 693-701,1993
32. Kobayashi T, Tao T, Gergely J, Collins 1: Structure of the troponin 52. Leavis PC, Gergely 1: Thin filament proteins and thin filament-linked
complex. Implications of photocross-linking oftroponin I to troponin regulation of vertebrate muscle contraction. CRC Crit Rev Biochem
C thiol mutants. 1 Bioi Chern 269: 5725-5729, 1994 16: 235-305,1984
33. lha PK, Mao C, Sarkar S: Photo-cross-linking of rabbit skeletal 53. Ohtsuki 1, Maruyama K, Ebashi S: Regulatory and cytoskeletal proteins
troponin I deletion mutants with troponin C and its thiol mutants: The of vertebrate skeletal muscle. Adv Prot Chern 38: 1-67, 1986
inhibitory region enhances binding oftroponin I fragments to troponin 54. Zot AS, Potter lD: Structural aspects of troponin-tropomyosin
C. Biochemistry 35: 11026-11035, 1996 regulation of skeletal muscle contraction. Ann Rev Biophys Biophys
34. Farah CS, Miyamoto CA, Ramos C, Dasilva A, Quaggio RB, el al.: Chern 16: 535-559, 1987
Structural and regulatory functions of the NH,- and COOH-terminal 55. Chalovich 1M: Actin mediated regulataion of muscle contraction.
regions of skeletal muscle troponin I. J Bioi Chern 269: 5230-5240, Pharmac Ther 55: 95-148,1992
1994 56. Farah CS, Reinach FC: The troponin complex and regulation of muscle
35. Luo Y, Wu l-L, Gergely 1, Tao T: Troponin T and Ca2+ dependence of contraction. [Review]. FASEB 19: 755-767, 1995
the distance between Cys48 and Cys 133 of troponin I in the ternary 57. Tobacman LS: Thin Filament-mediated Regulation of Cardiac
troponin complex and reconstituted thin filaments. Biochemistry 36: Contraction [Review]. Ann Rev Physiol58: 447--481,1996
11027-11035,1997
Molecular and Cellular Biochemistry 190: 9-32, 1999.
© 1999 Kluwer Academic Publishers.

Troponin I: Inhibitor or facilitator

S.Y: Perry
Department ofPhysiology, Medical School, University ofBirmingham, Birmingham, UK

Abstract
TN-I occurs as a homologous group of proteins which form part of the regulatory system of vertebrate and invertebrate striated
muscle. These proteins are present in vertebrate muscle as isoforms, M, 21000-24000, that are specific for the muscle type and
under individual genetic control. TN-I occupies a central position in the chain of events starting with the binding of calcium to
troponin C and ending with activation of the Ca2+ stimulated MgATPase ofthe actomyosin filament in muscle. The ability of
TN-I to inhibit the MgATPase of actomyosin in a manner that is accentuated by tropomyosin is fundamental to its role but the
molecular mechanism involved is not yet completely understood. For the actomyosin ATPase to be regulated the interaction of
TN-I with actin, TN-C and TN-T must undergo changes as the calcium concentration in the muscle cell rises, which result in
the loss of its inhibitory activity. A variety of techniques have enabled the sites of interaction to be defined in terms of regions
ofthe polypeptide chain that must be intact to preserve the biological properties ofTN-I. There is also evidence for conformational
changes that occur when the complex with TN-C binds calcium. Nevertheless a detailed high resolution structure of the troponin
complex and its relation to actin/tropomyosin is not yet available. TN-I induces changes in those proteins with which it interacts,
that are essential for their function. In the special case of cardiac TN-I its effect on the calcium binding properties ofTN-C is
modulated by phosphorylation. It has yet to be determined whether TN-I acts directly as an inhibitor or indirectly by interacting
with associated proteins to facilitate their role in the regulatory system. (Mol Cell Biochem 190: 9-32, 1999)

Key words: troponin I, troponin C, troponin T, troponin, tropomyosin, actin, actomyosin, calcium activated MgATPase, calcium
sensitivity, skeletal, cardiac muscle, muscle regulation, protein kinase A, protein kinase C, phosphorylation, phosphorylation
site, inhibitory peptide, actin binding site, binding site

Introduction about the same time we had shown [2, 3] that the MgATPase
activity of crude preparations of actomyosin, euphemistically
I was probably one of the first muscle scientists from the West called 'natural actomyosin' was sensitive to low concentra-
to meet Dr Setsuro Ebashi in Japan after World War II. The tions of EDTA and glycolcomplexon, the name then used for
occasion was not a scientific conference but during a tour of EGTA, and which we had obtained from Schwarzenbach
Japan in 1953 by the Cambridge University rugby team, for before it was commercially available. Thus it was clear to us
which at the time I was acting as manager. I met Dr Ebashi at that time that a trace of calcium was required for the
and his mentor, Professor Kumagai, during a diversion from MgATPase of 'natural actomyosin'. On the other hand we
my sporting responsibilities when I gave a lecture at the reported that the MgATPase of actomyosin prepared from the
University ofTokyo about our work on the muscle proteins. purified proteins was insensitive to calcium chelators.
From that time we have maintained a friendship and my Unfortunately we were not smart enough to show why these
respect for him has grown steadily with his many achieve- two preparations differed in behaviour. It was left to Ebashi
ments. It is a very special pleasure for me to contribute to this in 1963 [4] to report that 'natural tropomyosin' (tropomyosin
volume dedicated to my long standing friend. + troponin) was responsible for the effect and thus open up
Our research careers have been dominated by similar the whole field of calcium regulation in striated muscle. As
interests. In 1955 he reported [1] the association of relaxing a consequence ofhis discovery oftroponin my colleagues and
factor activity with the particulate fraction from muscle. At I were able later to identify a component of the complex,

Address for offprints: S. V. Perry, Department of Physiology, Medical School, University of Birmingham, Birmingham, B15 2TT, UK
10

troponin I. Over the years troponin I has become one of my skeletal form [17]. It is of interest in this respect that the foetal
favourite proteins and this review of its properties, the heart contains the slow skeletal isoform which is slowly
development of ideas on its function and thoughts on its mode replaced during immediate post natal development by the
of action represents my contribution to this commemorative cardiac isoform [18, 19]. This process is under the control
volume. ofthe transcription factor GATA-4 [20]. Replacement of the
endogenous cardiac isoform in myocytes by adenovirus-
mediated skeletal slow TN-I transfer increased the Ca 2+
Discovery and early studies on troponin I sensitivity ofthe tension development ofpermeabilized single
myocytes [21]. Although mammalian skeletal TN-I appears
During the development of' desensitized actomyosin' for the to be under the control oftwo genes a recent report provides
assay of troponin (or EGTA sensitising activity, as it was evidence for the expression ofthree genes in whole myotomal
called at that time) it was noted that occasionally, and muscle ofsalmon fry [22]. During the later foetal stages both
especially after ageing, the troponin extracts developed fast and slow skeletal isoforms are present in mammalian
inhibitory activity that was not calcium sensitive [5]. The fact skeletal muscle muscle cells [23].
that the inhibitory factor was specific for the MgATPase of The genes responsible for encoding the isoforms of TNI
actomyosin and inhibited its superprecipitation in the absence and TNT are organised in pairs. Those for the fast skeletal
ofEGTA suggested that it might be derived from the troponin isoform of TNI (TNNI2) and the fast skeletal form of TNT
complex or was a modified form of it [6, 7]. This hypothesis (TNNT3) are both located on chromosome llp15.5 [23a].
was confirmed when it was later shown that troponin could The genes for the slow skeletal muscle TNI and cardiac TNT
be fractionated into inhibitory (troponin B) and calcium are on chromosome lq32 whereas those for cardiac TNI
sensitising (troponin A) factors [8, 9]. Later it became clear (TNNI3) and slow TNT are located on chromosome 19q13.4
that some inhibitory protein fractions also contained another [23b]. This organisation contrasts with that ofother sarcomeric
basic protein of higher molecular weight, the '37000 protein genes and could be explained if the troponin genes
component', later called troponin T, [10], which was shown were derived by triplication of an ancestral TNI/TNT gene
to form a viscous complex with tropomyosin. In 1972 the pair [23a].
nomenclature proposed by Greaser et al.[ll] for the com- With further postnatal development the expression of the
ponents of the troponin complex, troponin C, I and Twas gene not appropriate for the cell type is suppressed with the
adopted. Before the nomenclature was rationalised the result that the mature adult skeletal muscle cell usually
various research groups working on the fractionation of contains only the isoform characteristic for its function [24].
troponin used their own names to distinguish their fractions The detection of TN-I isoforms by specific antibody staining
and the inhibitory factor was variously known as troponin B, provides a convenient and reliable method of muscle cell
troponin 2, troponin II and component II (see [12] for details). typing [25] and the presence of TN-I in serum can be used for
the detection ofdisease in skeletal and cardiac muscle [26]. The
immunological detection ofserum cardiac TN-I is widely used
in cardiology as an index ofmyocardial damage [27]. Slow and
Isoforms of troponin I and their fast isoforms are present in the same skeletal muscle cell as a
distribution consequence of cross innervation [28], hormone intervention
[29], and in pathological conditions [30].
Muscle is the only tissue that has been shown to contain Invertebrate troponins have not been as widely studied as
significant amounts of TN-I. In vertebrates it is restricted to their vertebrate counterparts but it is clear that they do not
striated muscle and thin filament regulation in smooth muscle represent such a homogeneous group as the latter. Attention
involves another actin-binding protein, caldesmon, which has been directed particularly to striated arthropod muscle
possesses some properties similar to those of TN-I. Isoforms where there were early reports of troponin-like systems in
of TN-I have been reported to be present in invertebrate insect flight muscles [31], crayfish [32], lobster [33], and
smooth muscle in a few instances. These include the adult horseshoe crab [34].
body wall of the ascidian, Halocynthia roretzi [13], adductor Components similar to TN-I, C and T of vertebrate muscle
muscle of the scallop [14] and the oviduct myoepithelial have been identified in many invertebrates but they frequently
sheath of Caenorhabitis elegans [15]. differ in molecular mass from their vertebrate counterparts.
Three isoforms, fast skeletal, slow skeletal and cardiac TN- Often the component identified as troponin I has a higher
I , each the product of a separate gene, are present in mam- molecular mass than the vertebrate form due to the addition
malian striated muscle [16]. As yet there is no evidence of a of residues at the N-terminus. It is of interest that TN-I from
distinct foetal form and the earliest isoform detected in the crayfish, Astacus leptodactylus, possesses an additional
myocytes growing in culture and early embryos is the slow N-terminal sequence ofabout 30 residues, similar to vertebrate
 11

cardiac TN-I. This region, however, only exhibits a sequence identified as HL I which corresponds to one of the comple-
identity score of22% with the vertebrate cardiac isoform and mentation groups that constitute the haplolethal region of the
does not possess an equivalent phosphorylation site. Com- Shaker gene complex of Drosophila. This region encodes a
parison of the sequence of the whole molecule with rabbit family of TN-I proteins that are expressed in a develop-
cardiac troponin I reveals a sequence identity of 26% [35]. mentally regulated manner. As judged from the gene sequence
Sequence analysis does not reveal any polymorphism in this the protein sequence of Drosophila isoform p6a 10 has 65%
species of crayfish whereas tail muscle of another crayfish, identity with the well-characterised TN-I of crayfish [35]. It
Procambarus c1arkii, is reported to contain two isoforms of is of particular interest that Barbas et at. [42] also report the
troponin I with molecular masses of 25 and 23 kDa [36]. the existence of mutant phenotypes in which the expression
The 52kDa subunit of the troponin system of the striated of certain TN-I isoforms is impaired or prevented. In addition
adductor of the Akazara scallop, Chlamys nipponensis to defects confined to specific muscles in these mutants,
akazara, that has been provisionally identified as a TN-I, can aberrant neurogenesis is observed. This is the first time that
be cleaved into two major fragments [37]. The C-terminal mutations in a TN-I gene have been reported to produce an
fragment of 17 kDa exhibits 39% sequence homology with effect in nervous tissue. These findings suggest that TN-I may
crayfish troponin I [38]. On the basis of this and the increased have a role in nerve cell development in addition to its well
inhibition of actomyosin ATPase obtained with scallop defined function in I filament regulation of striated muscle.
tropomyosin alone it has been concluded that the 52 kDa
protein is indeed a troponin I. It is reported in this study that
scallop tropomyosin alone inhibits the actomyosin ATPase Structure of troponin I
by 88% which is increased to 95% in the presence of TN-I.
The N-terminal 35 kDa fragment does not have inhibitory The isoforms of TN-I represent a homologous group of
activity. It has a unique amino acid composition with glutamic proteins with the molecular ratios of those of vertebrate origin
acid, arginine and alanine accounting for approximately 75% lying in the range of about 20000-24000. All consist of a
of the total. In the ascidian, Halocynthia roretzi, the troponin single polypeptide chain and rabbit fast skeletal muscle TN-
I isoform present in the striated cardiac and the smooth body 1, which the most widely studied, contains 181 amino acid
wall muscles is similar in polypeptide length to vertebrate residues and possesses a M, of about 21073 (Fig. I). The
striated muscle TN-I isoforms. The ascidian larvae express amino acid sequences of fast skeletal, slow skeletal and
two isoforms with truncation of about 30 amino acid residues cardiac TN-I ofthe rabbit are 60% identical [43]. The identity
at the C-terminus indicating that at least three genes are of sequence is even greater between a given isoform type in
responsible for encoding TN-I in this organism [39]. different species, for example the amino acid residues in fast
The original report of the presence of troponin in Letho- skeletal muscle isoforms of rabbit and chicken are 85%
cerus flight muscle [31] suggested that the system consisted identical. As is the case with TN-T the cardiac forms of TN-
of components similar to those of the vertebrate system. I have a slightly higher molecular ratio, about 24000, than
Nevertheless reinvestigation, taking precautions to minimise the skeletal isoforms. In the case of TN-I this is due to an
proteolysis [40], has so far failed to identify a protein that is additional, strongly conserved, N-terminal peptide of about
strictly analogous to vertebrate TN-I. It is concluded that the 30 residues in which is located an important phosphorylation
troponin system in flight muscle consists of TN-C, a TN-T site (see section on phosphorylation). TN-I from invertebrate
ofhigher molecular mass, 53 kDa, than its vertebrate counter- muscle is more variable in size than the vertebrate protein. It
part, and troponin H, molecular mass 80 kDa. Unlike TN-I, ranges from the smallest reported to date isolated from
TN-H did not inhibit actomyosin MgATPase alone but did ascidian larval muscle consisting of 142 residues [39], to very
in the presence ofTN-T and tropomyosin. The inhibition was much larger molecules of molecular ratio greater than 50000.
relieved in a calcium-sensitive manner in the presence ofTN- In some insect flight muscles itmay occur as a fusion protein
C. Lethocerus TN-H can be proteolytically digested to give with a tropomyosin-like sequence (see above).
a peptide with sequence homologies to vertebrate TN-I (B. Two regions corresponding to residues 17-23 and 97-121
Bullard, personal communication). It is immunologically in the rabbit fast skeletal muscle protein that are considered
similar to a protein of comparable molecular mass present in to be of functional importance are particularly strongly
Drosophila muscle that is a fusion protein of tropomyosin and conserved in all isoforms. The region represented by residues
a hydrophobic sequence rich in proline [41]. In view of the 97-121 includes the smallest sequence that possesses in-
fact that locus of the TN-I gene family has been identified in hibitory activity and is known to interact with actin. The
Drosophila it is surprising that a similar TN-I isoform has not region of residues 137-144 has some common features of
been isolated from Lethocerus. Drosophila also expresses sequence with the inhibitory peptide region [44]. In extension
TN-H, a protein which apparently replaces TN-I in Letho- of these observations it has been pointed out that the C-
cerus muscles. Barbas et at. [42] have studied a region terminal portion ofthe polypeptide chain exhibits a conserved
12

10
Met Gly ASp Glu Glu Lys Arg Asn Arg Ala Ile Tb.r Ala Arg Arg GIn

20 30
His Leu 11'5 Ser Val Met Leu GIn Ile Ala Ala Tb.r Glu Leu Glu Lys

.to
Glu Glu Gly Arg Arg Glu Ala Glu Lys GIn Asn Tyr Leu Ala Asp Ser

SO 60
Cys Pro Pro Leu Ser Leu Pro GIn Ser Met Ala Glu Val GIn Glu Leu

70 80
Cys Lys GIn Leu His Ala Lys I Ie ASp A~a Ala GI u Glu GI u Lys Tb.r

90
Asp Met Glu Ile Lys Val GIn Lys Ser Ser Lys Glu Leu Glu ASp Met

100 _-------------110-----
Asn GIn Lys Leu Pb.e Asp Leu Arg Gly Lys Phe Lys Arg Pro Pro Leu

120
_A_r_g_ _A_r_g_ _v_a_I_Ar
__ g ,~ Se r Al a ASp Ala Me~ Leu Lys Ala Leu Leu Gly

130 140
Ser Lys His Lys Val Cys Met ASp Leu Arg Ala Asn Leu Lys GIn Val

150 160
Lys Lys Glu Asp T~r Glu Lys Glu Arg Asp Leu Arg Asp Val Gly Asp

170
'::."? Arg Lys Asn Ile GI u Glu Lys Ser Gl7 Met Gl u G17 A.rg Lys r.ys

180
Met Phe Glu Ser GI uSer

Fig. I. Sequence of rabbit fast skeletal TN-I derived from a cDNA clone expressed in E. coli [97]. This corresponds to a protein of 182 amino acid residues
with a molecular weight of 21162 kDa. The sequence corresponding to the inhibitory peptide [60] is underlined and that corresponding to the minimum
inhibitory peptide [62] is enclosed in a box. Note: This sequence is slightly different from that originally proposed by Wilkinson and Grand [43] from amino
acid analysis of the isolated protein in that a methionine residue replaces the N-terminal acetyl group and argl53, asp 154 and leu 155 are inserted. The
evidence would suggest that the N-terminal of rabbit fast skeletal TN-I is acylated as are other myofibrillar proteins, M, 21073. In most of the work
published the original numbering of residues [43] has been used and has been retained (up to glul52) in this review on the assumption that the naturally
expressed protein in the fast skeletal muscle of the rabbit is acylated and consists of 181 residues. Authors should make clear, particularly in mutation
studies, which sequence is used. When results are quoted in this review the residue numbers used by the authors are retained.

repeat motif in three positions ofthe sequences ofmammalian TN-I is a basic protein with a high isoelelectric point and
and avian TN-I sequences currently available [45]. Using the in the isolated form tends to aggregate under physiological
residue numbering ofthe rabbit fast skeletal isoform these are pH values and ionic strength. For this reason it has been
in the region of residues 101-114 (designated ex), residues difficult to study the isolated protein in solution and, as it has
121-132 (p) and residues 135-146 (y). not so far been crystallised, a high resolution structure is not
 13

available. Despite its tendency to aggregate the isolated rabbit myosin, but not the calcium activated ATPase (CaATP as
fast skeletal isoform is readily degraded by proteolysis and substrate) ofactomyosin or myosin [6]. This clearly indicates
phosphorylated at residues thrll and serll7 by kinases that in some way TN-I blocks the interaction of actin with
indicating that most of the molecule is available to solvent. myosin that is responsible for activation of the MgATPase.
Evidence that suggests the polypeptide chain in the isolated With in vitro systems using purified TN-I inhibition of the
molecule is in an extended and flexible form. Structural MgATPase of actomyosin can be obtained with a molar ratio
predictions based on the amino sequence of the region of actin monomer to TN-I of I: I [12]. In the presence of
represented by residues 96-148 indicate that it consists ofa tropomyosin the inhibitory action is much enhanced and
series of alternating coil and helix [45]. values approaching 90-95%, depending on the conditions,
TN-I is designed to function as a complex in association are obtained. With molar ratios of actin monomer to TN-I of
with the other myofibrillar proteins and an intrinsic ability 3-4: I and higher, inhibition is obtained in in vitro conditions.
of its polypeptide chain to adapt to interacting proteins is vital The results of immunochemicallocalisation studies and the
for its function in the troponin complex. As the interaction known protein composition ofthe myofibril indicate that TN-
with TN-C is central to its function, the conformation it adopts I is located at every seventh actin monomer. It is considered
in the TN-I1TN-C complex is of particular interest. Despite that all the actin monomers of the thin filament, excluding
the availability of the high resolution structure of TN-C, those that may possibly be blocked even in stimulated muscle
similar detailed information about the structure of TN-I in the by the troponin complex itself, have the capability of inter-
free form or complexed with other proteins is not yet avail- acting with myosin. This implies that one TN-I molecule exerts
able. Some progress has been made towards solving this its inhibitory action over seven actin monomers occupying a
problem by the determination of the conformation of the distance of 38.5 nm along the I filament. The mechanism of
peptides corresponding to residues 104-115 [46,47] by NMR this remarkable cooperative property is still uncertain.
methods and the N-terminal region by X-ray crystallography
[48a]. In the latter study crystals of the complex oftroponin
C with the N-terminal fragment ofTN-I consisting of residues Tropomyosin and troponin I function
1-47 were examined. From neutron and low angle X-ray
scattering data the so-called 'dumbell structure' model has The current dogma of the field is that tropomyosin is the
been proposed for the TN-I1TN-C complex (see section on component that blocks the sites on actin involved in inter-
complex below). The more recent neutron scattering patterns action with myosin in resting muscle. It is postulated that on
obtained with deuterated TN-I reconstituted in the skeletal stimulation the tropomyosin moves, leaving actin free to
trononin complex suggest that the TN-I in the complex is interact with myosin, the MgATPase is stimulated and the
elongated and consists of two subdomains [49]. The bulk of muscle contracts, the so-called 'steric hypothesis'. Tropo-
the molecular mass (65%) exists as highly oblate ellipsoid of myosin lies in the groove of the actin double helix and is
revolution with the remainder of the molecule present as a responsible for the intensification of the 2nd and 3rd layer
highly prolate ellipsoid of revolution. The best fit of the data line reflections of actin in the muscle X-ray diffraction
was obtained when the axes of the ellipsoids were sharply pattern. This hypothesis was proposed by Haselgrove [50]
inclined to each other. Unlike the TN-I, in these studies the and Huxley [51] to explain changes in the layer line re-
TN-C was not observed to undergo a measureable global flections on contraction of frog and toad muscles. These
change in shape on addition of calcium. It also appeared to changes were interpreted as arising from movement of the
be elongated as in crystals of the isolated protein. tropomyosin molecule in the filament groove from the
blocking position to that which permits the MgATPase of
myosin to be activated by interaction with actin. Due to an
Functions of troponin I incorrect assumption about the polarity of the actin filament
in the original study, the position of the tropomyosin in the
TN-I can interact with all the major proteins of the I filament, groove has been changed [52]. Nevertheless the basic tenets
actin, tropomyosin, TN-C and TN-T, properties that clearly of the hypothesis still stand. The hypothesis is supported by
indicate its central role in the regulatory process in striated electron microscope image reconstruction studies that show
muscle. A great deal of information, some of which is at times tropomyosin to move when the thin filament is activated by
confusing, exists about the amino acids and peptide regions calcium [53] and the fact that the X-ray diffraction studies
involved in these interactions. There is a great need for a high indicate tropomyosin movement can be detected before
resolution three dimensional structure of TN-I to which this tension development in intact muscle.
data can be related. The most striking property of TN-I is its The problem is to demonstrate clearly that in resting muscle
ability to inhibit the magnesium activated ATPase of acto- tropomyosin blocks the site(s) on actin, interaction of which
14

with an as yet undefined site(s) on myosin leads to activation From the results of the enzymic studies carried out in
ofthe MgATPase. It is difficult to postulate with any precision solution it has been concluded that tropomyosin inhibition
on this matter whilst there is uncertainty about the detailed is correlated with complex formation between it and actin [54].
nature of the interaction of actin with myosin that is respon- This correlation is far from complete, for smooth muscle
sible for the contractile process. In recent years the concept tropomyosin shows maximal potentiation of skeletal acto-
of strong and weak binding states corresponding to the myosinATPase at high ionic strength when actin-tropomyosin
relaxed and contracting states respectively has been proposed. interaction is considered to be at a maximum [56]. To explain
Ifthis actually is the case the steric hypothesis would require these inconsistencies yet another hypothesis has been in-
that tropomyosin is involved in the change from weak to the troduced. This postulates that the actin- tropomyosin complex
strong binding state of the actomyosin interaction. In its exists in strong and weak myosin-binding states, the pro-
current form the hypothesis demands that tropomyosin can portion of which determining the ATPase activity of the
interact with actin in a manner that inhibits the actomyosin system under any given conditions.
MgATPase. The maximal inhibition obtained with tropomyosin in
There is experimental evidence from enzymic studies systems using myosin fragments at higher ionic strengths,
suggesting that this can occur, but the effect very much under which conditions the intrinsic MgATPase activity is
depends on the conditions of assay [54]. For this reason it is low, about 60%. The much higher levels of inhibition
very difficult to relate the results to events occurring in the obtained in the presence of TN-I are more comparable to
myofibril which is essentially a protein gel system. Both those that exist in resting muscle. This could be explained by
inhibitory and potentiating effects of tropomyosin have been the steric hypothesis on the assumption that TN-I increases
reported with systems reconstituted from the component the binding constant of actin for tropomyosin. It is difficult
proteins [55] and often with those containing myosin frag- to see how this achieves greater inhibition according to the
ments to produce soluble enzyme systems [54, 56, 57]. In hypothesis ofWilliams et al. [56] unless it is considered that
addition to the ionic conditions and the relative concen- TN-I converts all of the actin-tropomyosin complex into the
trations of the protein components being factors in de- weak myosin binding state.
termining the effect of tropomyosin on the actomyosin Even if it could be clearly shown that in resting muscle
MgATPase, the physical state of the actomyosin system may tropomyosin blocks the interaction site on actin there is no
also be important. This was first indicated by Katz [55] who direct information about the nature of the process that causes
reported in his study of the superprecitation of actomyosin tropomyosin to move when muscle is stimulated. One
that tropomyosin inhibited the clearing (solution) phase but suggestion has been that the conformational changes occurring
enhanced the MgATPase during superprecipitation. A signifi- when TN-C binds calcium are transmitted to tropomyosin
cant observation suggesting that tropomyosin may have a through TN-lor TN-T, both of which can be demonstrated
direct effect on the myosin itself was that treatment of the to interact with TN-C and tropomyosin. Nevertheless it is
myosin with a sulphydryl reagent destroyed the inhibitory difficult to visualize how interaction with the troponin
effect on superprecipitation, but left the enhancement of the complex restricted to one region of the tropomyosin, would
ATPase unchanged. cause sideways movement of the molecule along the whole
Some what different findings have been reported when the of its length .
effects of tropomyosin are studied on the enzymic activity of The fact that tropomyosin in the absence of TN-C and
actomyosin extracted directly from myofibrils. This prepara- TN-T can extend the inhibitory activity of one molecule of
tion, 'desensitized actomyosin', from which endogenous TN-I to actin monomers with which it is not in contact,
troponin and tropomyosin has been removed may corre- suggests that tropomyosin has an important role in controlling
spond more closely to the in vivo situation. At low ionic the relationship between the actin monomers in the filament.
strength, under which conditions desensitized actomyosin The intact molecule of TN-I is not essential for the inhibitory
has MgATPase activity comparable to that of the intact activity in the presence of tropomyosin to be extended over
myofibrils, no inhibitory activity could be detected with more than one actin monomer. The effect is obtained with the
tropomyosin [58, 59]. Some activation of the MgATPase was inhibitory peptide representing residues 96-116 of the rabbit
evident but this was reduced or eliminated ifthe tropomyosin fast skeletal isoform [60]. Inhibition can also be obtained with
was further purified. On the other hand the CaATPase of residues 101-115 [61] and a synthetic peptide corresponding
'desensitised actomyosin' was inhibited under these condi- to residues 105-114, which is the minimum length required to
tions suggesting that when CaATP was the substrate, tropo- produce this effect [62]. The evidence from affinity chroma-
myosin modifies the enzymic process, but not with MgATP tography [63] and from fluorescence studies on pyrene-labelled
as the substrate. This complements the evidence of Katz [55] tropomyosin [64] is that any direct interaction between TN-I
indicating that tropomyosin may have some direct effect on and tropomyosin is weak. In the presence of TN-T and its
the enzymic activity of myosin. tropomyosin binding fragments, TN-I is bound more strongly
 15

in the ternary complex [64]. This presumably reflects the own [66]. Nevertheless the fact that there is evidence that
ability of TN-T to link TN-I to tropomyosin rather than any conformational changes occur in actin indicates that these
change in affinity of the latter protein for tropomyosin. take place without disrupting the periodicity of the double
The association of inhibitory activity with small peptides helical structure. Presumably this is stabilised by the tropo-
in the presence of tropomyosin suggests that the interaction myosin molecules in the large grooves ofthe filament. Proton
of TN-I at a relatively small region of the actin molecule is NMR studies indicate that the binding of TN-I to the N-
responsible for the cooperative effect. There are a number of terminal sites of actin produce changes at the C-terminus
possible explanations, for example where it is not bound. As a result of these changes the alkali
(I) Binding at the site induces a conformational change in light chain (ALC I) of myosin no longer binds to the C-
the actin monomer which is transmitted to neighbouring terminal region of actin [67]. From the analysis of three
monomers with which TN-I is not bound. The conformational dimensional images reconstructed from cryo-electron micro-
change might be responsible for the following graphs Ishikawa and Wakabayashi [68] have reported changes
(a) The affinity of tropomyosin for the site on actin in the structure of reconstituted actin filaments in the presence
that interacts with myosin and which is responsible for the of calcium ions. Although the results of this study are
activation of the MgATPase, is increased. The result would suggestive of conformational changes in the actin filament
be that all the actins in the thin filament are unable to interact the authors were not able to conclude whether they were due
with myosin. There are steric difficulties with this explana- to tropomyosin movement or conformational change in the
tion, particularly in the relation oftropomyosin to TN-I at the actin itself.
actin monomer(s) where the troponin is bound. All the
evidence suggests that this is at or close to the activation site
on the actin. There is also the in vitro enzymic evidence in The interaction of troponin I with actin
which correlation between tropomyosin binding and in-
hibition is not as close as might be expected. This model gives The fact that the inhibitory peptide, residues 96-116, is the
tropomyosin an active role. only fragment ofTN-I present in the cyanogen bromide digest
(b) The conformational change on binding TN-I with inhibitory activity emphasises the special significance
transmitted to neighbouring actins with which TN-I is not of this region for interaction with actin. Its properties
directly associated, is such to render them unable to interact resemble those of the intact molecule in that in addition to
with myosin. The tropomyosin in this model does not play its inhibitory activity being enhanced by tropomyosin, its
an active role but acts as a kind of template that supports the effect is neutralised by troponin C. Indeed cardiac fibre
actin filament in a manner that permits the conformational bundles reconstituted with the cardiac inhibitory peptide are
changes induced by TN-I binding to spread to adjacent able to undergo sequential contraction- relaxation cycles [69].
monomers. It is not unreasonable to expect that if con- On a molar basis it, and the shorter synthetic duodecapeptide,
formational change occurs in one monomer in the actin residues 104-115, is 45-70% as effective as TN-I [60,62].
filament the neighbouring monomers must undergo change The shorter sequence represents less than half of the highly
to maintain the symmetry of the I filament structure. In this conserved region of rabbit fast skeletal muscle TN-I, which
model the tropomyosin does not have a blocking role but its extends from residues 97-121 (Fig. I). The corresponding
movement is purely an adjustment on the surface of the I minimum length inhibitory peptide from rabbit cardiac TN-
filament in response to the conformational changes occurring I is slightly less effective as an inhibitor of the actomyosin
in the actin monomers. ATPase than the skeletal peptide. It differs from the latter in
(2) In this model it is the binding ofthe myosin head to actin, that pro II 0 is replaced by threonine and arg 113 by leucine.
once the sites on actin are rendered available for myosin Studies with hybrid peptides indicate that substitution of
interaction, that induces cooperative activity between actin residue 110 had little effect on activity and the substitution
monomers in the filaments. There is already some evidence for of arg 113 was probably largely responsible for the difference
this type of cooperative behaviour [65]. Tropomyosin could in activity of the two forms [70]. The importance of the
play active or passive roles as outlined above in such a model. arginine residues, which are principally located in the
Conformational changes that occur in the actin monomers C-terminal moiety of the inhibitory peptide, for the inter-
during the contraction-relaxation cycle do not lead to any action with actin is indicated by proton NMR studies of the
marked changes in the helical parameters of the actin fila- binding of the inhibitory peptide to defined cleavage frag-
ments and the subunit repeat remains constant [51]. From a ments of actin [71]. The inhibitory region represented by
recent reanalysis of the low angle X-ray data it has been residues 96-116 in the rabbit fast skeletal isoform is strongly
concluded that in addition to the tropomyosin movement conserved in vertebrates with an identity score of85%. In the
there are small but plausible actin subdomain movements. corresponding segment of crayfish TN-I only 57% of the
The data cannot be explained by a tropomyosin shift on its residues are identical or functionally conserved [35].
16

10 THR 11 CYS12 (ml

(1021
20

,... 30

40
BINDS
TROPONIN 50
T
223-227
60
I
I
I 70
I
I
I
L- 80

90

BINDS TO ~--------'00 83
I
ACTIN
1-7 I
P ~--
M
134
(921
23-i7 ~ 110
----------
120

130
J

BINDS [--------140
ACTIN
-------150

MET 155
160 (8n

170 TROPONIN
(
180
(

Fig. 2. Scheme summarising information on Ihe sites ofinleraclion of rabbit fast skeletal TN-I with TN-C and other proteins of the myofbril. I P, inhibitory
peptide region, residues 96-116, indicated by hatching. M I P, minimum inhibitory peplide region, residues 104-115 indicated by cross halching. Numbers
indicate the residue positions in the primary sequences of the proteins, mutant forms indicated by (m). Interactions are indicated on arrows logether with
numbers of the TN-I residues in the regions involved. The TN-T binding site on TN-I is not well defined but cysteines 48 and 64 are inaccessible to
iodoacetamide labelling in the presence ofTN-T [99]. Reference numbers are given in brackets.

Two N-terminal regions of actin, residues 1-7 and 23-27, TN-I binding site on actin has also been obtained by cross-
were identified as interaction sites with the inhibitory peptide linking studies using the 'zero-length' carbodiimide reagent
[72] . Similar regions have been identified for interaction of that is specific for lysine-carboxylate contacts [73]. By
the inhibitory peptide from cardiac TN-l with actin (J.P. comparison of peptide patterns after proteolysis in the
Trayer, personal communication). Evidence for aN-terminal absence and presence of the cross linker it was concluded that
 17

the region represented by residues 1-12 was cross linked to

TN-I. Positions 98, 105 and 107 are occupied by lysine
residues in the inhibitory peptide. In view of the fact that the
minimum peptide required for interaction with actin corre-
sponds to residues 105-114 and in the light of the NMR
results [72], it is likely that lysines 105 or 107 are involved
in the cross linking (Fig. 2).

t
Thus the evidence suggests that the regions represented by
residues 104-115 of the TN-I molecule and the N-terminus
of actin are involved in the interaction between the two 96 115 131 140 148
proteins that is functionally significant. This interaction is ~~ COOH

~/////)/
essential for inhibition and in initiating events that lead to
amplification of the effect by tropomyosin. In the myofibril
at the moment of stimulation induced by the rise in calcium
concentration some movement of the TN-I must take place Actin-Tm
to enable the actin to activate the actomyosin MgATPase.
Resonance energy transfer measurements on addition ofCa 2+
to reconstituted thin filaments indicate that both the N- Fig. 3. Scheme indicating the interactions of TN-I with TN-C and actin/
tropomyosin as proposed by Tripet et al. [81]. Regions involved in
terminal (cys48 [m]) and the C-terminal (cysI33) regions of interactions are indicated by arrows. The shaded regions of TN-I are those
TN-I move away from the C-terminus of actin (cys374) essential for full inhibitory activity. Hatched regions are those presumed to
towards TN-C. The results suggest that the movement of the interact with TN-C only. Figures indicate residue numbers in the TN-I
C-terminal of TN-I, about 15 A, is more extensive than that polypeptide chain.
of the N-terminal region [74, 75]. Similar movement is
observed when the reconstituted filament system is activated
by binding myosin S I to the filament in the absence of Ca 2+. TN-I with the inhibitory region, residues 96-116, deleted is
Somewhat less movement is observed in the presence ofCa 2+ completely inactive as an inhibitor [79]. Another mutant with
[76]. In association with these changes, when the low affinity residues 105-115 deleted is partially active as an inhibitor
calcium binding sites of TN-C are occupied, cys98 of the suggesting that the whole ofthe inhibitory peptide region and
latter protein moves closer to the inhibitory peptide [77]. not simply the minimal inhibitory peptide region, residues
From binding and ATPase studies on the effects of TN-l 104-115, is required for function [79]. The minimum sequence
on the actomyosin - S I ATPase in the presence of the to obtain inhibitory activity comparable to the intact rabbit
regulatory proteins it has been confirmed that the TN-I fast skeletal TN-I molecule is residues 98-148 [80, 81].
inhibitory activity is independent of the presence ofTN-C or Removal of residues 140-148 reduces the inhibitory action
TN-T [78]. Although Ca 2+ dissociates intrinsic TN-l from an to that ofthe inhibitory peptide. This observation, and the fact
actin-tropomyosin site it remains bound to the complete I that a synthetic peptide corresponding to residues 128-148
filament system suggesting that another region in addition to bound to the actin-tropomyosin filament and induced a weak
the inhibitory peptide region is involved in actin binding. The inhibitory activity, has lead to the suggestion that the region
inhibitory peptide, residues 98-116, is able to maximally consisting of residues 140-148 is a second actin binding site
inhibit to the same extent as the intact molecule but is about (Fig. 3) [81].
50% as effective as equimolar amounts of intact TN-I in Similar findings about the role of the C-terminal part of
producing 50% inhibition ofthe actomyosin MgATPase [60]. the molecule in the inhibitory activity of TN-I have been
This implies that another region of the molecule that may obtained with the cardiac isoform [82]. Using mutants of
interact with actin is required for full inhibitory activity. mouse cardiac TN-I (211 residues) it has been concluded that
Proton NMR studies have so far been largely restricted to the residues 152-199 are required to obtain inhibition equal to
interaction of the inhibitory peptide with the N-terminus of that of the wild type protein. These correspond to residues
actin and less is known about other regions of the molecule 120-167 of rabbit fast skeletal TN-I and it is suggested that
that may be involved. their results can be interpreted to imply that there are two actin
Because the sequence of residues 121-146 of rabbit fast binding sites in the corresponding region of cardiac TN-I.
TN-I has some common features with residues 108--115 of Until the nature of the interaction in this region, which is very
the inhibitory region, it has been suspected that regions similar in amino acid sequence in the skeletal and cardiac
C-terminal to the inhibitory peptide may be involved in actin isoforms, can be defined precisely in terms of the amino acid
interaction. Additional interaction in this region would be residues involved, e.g. by NMR studies, judgement on this
expected to have a modulating role, for a mutant of skeletal interpretation must be reserved.
18

Interaction of troponin I with troponin C Two dimensional NMR study of the interaction of the
C-terminal part of the inhibitory peptide using a synthetic
The properties and functions of TN-I and TN-C are relatively peptide, N-acetyl TN-I (104-115) amide has enabled the
well defined compared with those of the third component of conformational changes that occur when it complexes with
the troponin complex, TN-T. The fact that TN-I has an troponin C to be described [46]. The sequence of this part of
inhibitory property that is neutralised by TN-C implies that the inhibitory peptide is somewhat unusual in that it consists
they interact together and this interaction has a central role of six basic residues alternating with hydrophobic residues,
in the regulatory process. The stability of this complex in the interrupted in the centre by two adjacent proline residues. It
presence ofcalcium is clearly illustrated by the demonstration is concluded from the data that when bound the peptide forms
that when the two proteins are present in equimolar amounts an helical amphiphilic structure bent round the central proline
it migrates as a single band on electrophoresis in 6-8 M urea residues so that the hydrophobic residues are brought closer
[12,83]. In the presence of EGTA the complex dissociates together to form a hydrophobic face. Campbell and Sykes
and the TN-C migrates as a single band of different mobility. [46] suggest this interacts with an exposed hydrophobic
The stability of this complex and its calcium dependence region of troponin C. The interaction of the complete in-
enabled the development ofa simple method for the isolation hibitory peptide with TN-C must be more extensive than this
oftroponin-I in one stage from whole muscle homogenates for earlier NMR studies [71] indicated that its N-terminal
using a TN-C affinity column [16]. On application of this residues were also perturbed in the presence ofTN-C. These
technique to the cyanogen bromide digest of rabbit fast residues are presumably less important in actin binding for
skeletal muscle TN-I only peptides consisting of residues 1- the shorter peptide acts as an inhibitor of the actomyosin
21,1-47 and 96-116 are bound to theTN-C affinity column ATPase that is neutralised byTN-C. Nevertheless in view of
[60]. The implication that these two regions are involved in the studies of Zang et at [79] with mutants they are of some
TN-C binding and are important for the function of the significance for inhibitory function.
complex is supported by the fact that they contain the two The inhibitory peptide region in isolated TN-I must be
regions of troponin I that are strongly conserved in all exposed to solvent for it is readily susceptible to limited
vertebrate isoforms. Subsequent investigations by a number proteolysis by chymotrypsin, cleavage occurring at asp I0 I
ofworkers have supported these findings (for reviews see [44, or Iys I07 [90]. Despite the evidence for the involvement of
84,85]). this region in interaction with TN-C it is equally susceptible
to chymotryptic attack in the TN-I/TN-C complex.
Comparison between the skeletal and cardiac systems has
Inhibitory peptide region been made using the 104-115 residue peptide in which
proline 110 is replaced with glycine as an analogue of the
The fact that TN-C neutralises inhibition by residues 105- cardiac inhibitory peptide [47]. In rabbit cardiac TN-I prollO
114 also indicates that it will interact with this region to is replaced by threonine. It is concluded that on interaction
displace the actin from TN-I so that it can activate the myosin with bovine cardiac TN-C the peptide analogue undergoes
MgATPase. TN-C interacts with lysine, leucine and phenyl- similar conformational changes to those occurring in the
alanine, residues that are more abundant in the N-terminal skeletal system, but the cardiac peptide appears more flexible
region of the inhibitory peptide [71]. The NMR evidence about the glycine residue. It is a little difficult to be certain
indicates that TN-C interacts with both N- and C-terminal whether this is a real difference between the fast skeletal and
parts of the inhibitory peptide region of TN-I in the presence cardiac systems or simply due to the fact that comparison has
of calcium. Cross linking studies confirm that this is the case not been made with strictly homologous peptides.
for they demonstrate that in the TN-I/TN-C complex the The interaction of TN-I or the inhibitory peptide with actin
central helical region and both the Nand C-terminal domains is independent of calcium whereas that with TN-C is much
ofTN-C all interact with the inhibitory region of TN-I (Fig. strengthened in the presence of this cation. With isolated
2). The interaction appears to be antiparallel in so far as TN-C proteins TN-C can neutralise the inhibitory activity of TN-I
cys98 cross links with residues 103-111 [86] and cys57 in in the presence of EGTA [91] indicating that calcium is not
TN-C cross links to residues 113-121 in TN-I [87]. Cross essential for the interaction. In other words the interaction
linking experiments with mutants in which single cysteine between the two proteins is sufficiently strong in the absence
residues are inserted at positions 6, 48, 89, 104, 133 or 179 of calcium to neutralise inhibitory activity. Nevertheless
are also consistent with an antiparallel arrangement of the proton NMR investigations indicate that TN-C interaction
polypeptide chains of the two proteins [88]. Gly I04 of the with the inhibitory peptide is less strong than with the intact
inhibitory peptide also cross links with metl55 in a region protein for it is modulated by calcium binding at con-
consisting of residues 154-159 ofTN-C, at the end of helix centrations that would suggest the high affinity sites ofTN-C
H ([89], Fig. 2). are filled [71].
 19

N-terminal region oftroponin I of human cardiac TN-I have confirmed that TN-C interacts
with the homologous N-terminal site on this protein [95]. The
In the original studies on the digests of fast skeletal TN-I strength ofTN-C binding to the N-terminus ofTN-I has been
obtained by a variety of specific cleavage methods the confirmed by Ngai and Hodges [96] who have reported that
terminal peptide representing residues 1-47 was the most a synthetic peptide consisting of residues 1-40 of TN-I can
strongly bound to the TN-C affinity columns. A shorter dispiace TN-I from a preformed TN-I1TN-C complex. It also
peptide obtained by cyanogen bromide digestion consisting prevents TN-C from neutralising the inhibitory activity of the
of residues 1-21 was also bound, but less strongly [60]. These inhibitory peptide on the MgATPase of actomyosin. In the
observations indicate that in addition to the centrally located light of these results it has been suggested that the N-terminal
inhibitory peptide sequence, the N-terminus of TN-I interacts region of TN-I plays an important role in modulating the
strongly with TN-C. Further substantiation to this view was calcium sensitive control ofthe actomyosin MgATPase. Ifthe
provided by the finding that two cyanogen bromide fragments results obtained with the fragments and synthetic peptides can
representing different regions of rabbit fast TN-C form be applied to the intact proteins, binding of TN-C to the
calcium dependent complexes with TN-I [92]. One of these N-terminus of TN-I would be expected to occur later as the
consisting of residues 83-134 was shown to neutralise the calcium concentration rises on stimulation. Ngai and Hodges
inhibitory activity and inhibit the phosphorylation of ser117 results imply that this should lead to detachment of TN-C
ofTN-I much more effectively than thrll. Clearly it interacts from the inhibitory region and inhibition of the ATPase, ie
with the inhibitory region. Leavis et al. [93] employing a wide speed up relaxation. Nevertheless the N-terminus does not
range of specific cleavage fragments also came to the appear to be essential for function, at least in in vitro, systems.
conclusion that there were two sites of attachment between A mutant in which the first 57 N-terminal residues are deleted
TN-C and TN-I. [97] has inhibitory properties similar to wild type TN-I and
Recent photocrosslinking studies with single cysteine retains calcium dependent interaction with TN-C.
mutants have identified a specific cross link between TN- From an extensive study with deletion mutants Farah et al.
C 158 and met21 of TN- I and a range of crosslinks between [98] conclude that the N-terminal region does not have a
TN-C21 and residues 9frl31 of TN-I [93a]. Proton NMR regulatory role but ascribe such a function to the C-terminal
studies [71] of the interaction of peptides representing region of the molecule. They consider that residues 1-102
residues 1-21 and 9frl16 of TN-I with TN-C indicate they of TN-I playa structural role in the troponin complex, being
do not compete on binding implying that separate sites are responsible for binding to the C-terminal domain of TN-C
involved and are close together. By applying a spin label to and for stabilising the incorporation of TN-T into the ternary
cysteine 98 ofTN-C it was concluded that the sites were both troponin complex. The mutant TN-I10J_1R2 despite containing
within 15 A of cys98. In both cases the interactions were only 44% of the residues of the intact molecule and only part
calcium sensitive although dissociation at the N-terminal site of the TN-C binding region of the inhibitory peptide region,
of TN-I occurred at higher calcium concentrations than were inhibits the actomyosin ATPase as well as the intact molecule.
required for dissociation at the inhibitory peptide site. It interacts with the N- and C-terminal domains ofTN-C in a
Although Katayama and Nozaki [94] agree on the basis of calcium dependent manner and can be reconstituted into a
electrophoresis studies that the binding of TN-I fragment functional calcium sensitive filament.
consisting of residues 1-21 to TN-C is calcium dependent
they conclude that their inhibitory fragment, residues lOl-
llS, is not. This is surprising for the NMR data indicates that C-terminal region oftroponin 1.
the slightly longer intact inhibitory peptide, consisting of
residues 9fr116, does require a trace of calcium, sufficient No evidence was obtained by affinity chromatography for
to saturate the high affinity sites, to bind to TN-C [71]. It is TN-C binding to specific cleavage fragments of TN-I pro-
reported that the mutant of mouse cardiac TN-I with 53 N- duced from the region C-terminal to the inhibitory peptide,
terminal residues deleted does not bind to cardiac TN-C or one ofwhich corresponded to the C-terminal46 residues [60].
restore calcium activation to the myofibrillar ATPase. Nevertheless the sequence in regions of the C-terminal
Somewhat surprisingly the mutant with additional N-terminal portion of TN-I is strongly conserved between isoforms,
residues deleted, TN-I sQ-211' binds weakly to TN-C and suggesting this region is of functional importance. The use
partially restores calcium activation [94a]. of deletion mutants of TN-I has thrown new light on the
Although the function ofthe inhibitory region of TN-I has possible role of the C-terminal domain [98]. TN-I I2 Q-1S2' as
been clear for some time that of the other calcium -dependent would be expected, had no significant inhibitory or activating
binding site at the N-terminus has received less attention. The effect on actomyosinATPase whereas the inhibition obtained
evidence with peptide fragments indicates that its presence with TN-I 10J-1S2 was 80% relieved by TN-C. Mutants in which
is not essential for inhibition. Recent studies on mutant forms parts ofthe C-terminal region are removed possess inhibitory
20

activity which is neutralised when complexed with TN-C in The interaction ofTN-C with TN-I is clearly complex and at
the presence or absence of calcium, i.e. the system is not the heart of the regulatory process. Even in the absence of
calcium sensitive. This suggests that the C-terminal regions calcium the interaction is strong enough in the isolated protein
have a role in determining the calcium sensitivity of the TN- system for the inhibitory activity of TN-I to be neutralised.
I/TN-C complex. These results must be interpreted in the light Values for the equilibrium binding constants vary [104-107]
ofthe observation that neutralisation of the inhibitory activity but generally accepted values would be approximately 106
of isolated rabbit fast skeletal TN-I with its homologous M- I and 109 M- I in the absence and presence of calcium
TN-C is also virtually complete at equimolar ratios whether respectively [84]. This approximately 1000 fold increase in
calcium is present or not [91]. Farah et al. [98] conclude that affinity on binding calcium is responsible for initiating
the COOH domain of TN-I has a regulatory role and interacts changes that transform the muscle from the resting to the
with the N-terminal domain ofTN-C. contracted state. The binding of TN-I also increases the
The region in the vicinity of cysl33 of TN-I appears to be affinity for calcium of both the high affinity and low affinity
of significance for the function of the TN-I/TN-C complex. domains ofTN-C [106-108].
This residue is located at about 16 residues on the COOH side The N-terminal and inhibitory regions of skeletal TN-I are
of the strongly conserved inhibitory region occupied by actin fairly well defined as sites of interaction with TN-C but the
and/or TN-C. The phosphorylation of serl17 in rabbit fast nature of the interaction and location of the interaction
skeletal TN-I is blocked by TN-C. This fits in well with the C-terminal to the inhibitory peptide region is less well
proposal for a 'second TN-C binding site' in the regions of defined. Recently NMR spectroscopy studies with the peptide
residues 115-131 [81]. Nevertheless the adjacent cys 133 representing residues 115-131 of skeletal TN-I indicate that
appears to be exposed on the surface of TN-I when it is this region binds at the N-terminal hydrophobic pocket ofTN-
complexed with other components of the troponin complex C [I 08a]. This conclusion is supported by studies with
[99] which suggests that the binding ofTN-C in the region of cysteine mutuants of the skeletal protein demonstrating
serl17 does not extend up to cys 133. The conclusion that there crosslinking of met 121 with the hydrophobic region [1 08b].
is a binding site in this region is supported by fluorescent The corresponding region of human cardiac TN-I [residues
energy transfer experiments [100, 101] indicating that when 148-164] also binds to the regulatory domain of cardiac TN-
TN-I binds to TN-C the distance between cys98 ofTN-C and C but there are differences from the skeletal system in the
cys 133 oftroponin I decreases. In the complex the N-terminal conformational changes induced and the affinity of the
of TN-C is also close to this region for a mutant in which interaction [1 08c, 108d]. Much effort has gone into defining
residue 12 ofTN-C is replaced by cysteine a cross link between the complementary sites on TN-C (for reviews see [44, 84,
this residue occurs at or near metl 34 ofTN-1 [102]. With aTN- 98]). The interaction interface of TN-C would appear to be
C mutant in which residue 98 is replaced by leucine and residue extensive as judged by the regions in the primary sequence
89 by cysteine, cross-linking occurred with residues in the that have been reported to be involved in interacting with TN-
region of 108-113 of troponin I. These results are further evi- I (Fig. 2). This may be more apparent than real for despite
dence of an antiparallel arrangement of the peptide chains of the availability ofthe crystal structure ofTN-C it is uncertain
the two proteins in the region of the inhibitory peptide (Fig. 2). if the molecule in the troponin complex is as extended as the
From the results of a number of investigations involving crystallographic evidence would suggest. Some information
binding, enzymic and crosslinking studies it is emerging that on this point has been provided by small angle X-ray and
the region ofTN-I represented by residues 96-131 is involved neutron scattering. This data can be interpreted by an
in binding to the Nand C terminal domains of TN-C ([45, extended 'dumbbell like' structure for the skeletal TN-I/TN-C
81,89, 93a, 98] Fig. 3). An additional site(s) for binding actin complex. In this structure the TN-I winds round the extended
has been proposed on the C-terminal side of this site (see TN-C and makes contact with the hydrophobic patches
section on actin binding). considered to be present in the Nand C domains of the TN-C
The C-terminal region of cardiac TN-I is essential for the [109, 110]. This model is supported by the results of floures-
full development of Ca 2+-sensitive force in reconstituted cence lifetime, acrylamide quenching and photocrosslinking
skinned cardiac fibre bundles. Replacement of wild type TN- studies [111]. It might be expected with such a model that the
I with the mutant TN-I 1-151 reduced force development by two distance between the Nand C terminal domains of TN-I,
thirds [103]. about 40 A, would change when it complexed with TN-C.
This has been shown not to be the case using fluorescent
probes on cys48 and cys133 of TN-I. In the presence of TN-T
Troponin I-troponin C complex and in the reconstituted thin filament this distance increased
to about 50 A. Removal of calcium from the thin filament
There is much to be learnt about the nature of the calcium system caused a further increase to about 60 A [112]. This
regulation of the inhibition by TN-I in the troponin system. fits in well with the observations of Stone et al. [49] that TN-
 21

I in the troponin complex is less elongated when calcium is fast skeletal isoform, is ofparticular importance for its function.
bound to TN-C. The interactions ofthis region with other proteins of the myo-
In contrast to the above evidence suggesting that the TN- fibril are illustrated in the scheme ofTripetet ai. (Fig. 3). This
C molecule is in an extended form in the TN-C/TN-I com- region must be intact for full inhibitory activity and is essential
plex, a recent crystal structure ofTN-C (two Ca2+ bound state) for Ca 2+ sensitivity of the actomyosin MgATPase [80, 82].
complexed with the 47 residue N-terminal fragment of TN- Although the region of the molecule N-terminal to residue 96
I indicates that the TN-C is in a compact globular form [48a]. is not essential for inhibitory activity it is required for maximum
The amphiphilic C terminal end of this a-helical peptide is ATPase activity in the regulated actomyosin sytem [80].
bound in the hydrophobic pocket of the N-terminal regulatory
domain ofTN-C. It should be noted that there is evidence with
isolated peptides that a region C-terminal to the inhibitory Interaction of troponin I with troponin T
peptide of TN-I also binds to this hydrophobic pocket [81,
108a, 108c]. It remains to be demonstrated whether both It has long been known that TN-T is an essential component
regions of TN-I can interact with this pocket when the of the troponin complex for the inhibitory action of TN-I to
complex between the intact proteins is formed. It is possible be regulated by calcium, but the precise relationship between
that some isolated peptides from TN-I can interact with this the two proteins is not as well defined as that between TN-I
hydrophobic pocket on TN-C in a non-specific manner. and TN-C (see Perry [12Ia] for review). The fact that in the
There are two fairly well defined sites on TN-C involved early studies on TN-I preparations often contained TN-T [10]
in interaction with the inhibitory region of TN-I. The first is suggested that there might be some interaction between the
the region of the E helix ofTN-C (residues 93-103), which two proteins. The affinity ofTN-I for TN-T, however, would
has been shown to be implicated by NMR [71] and cross appear to be rather low for interaction has not been demon-
linking studies involving residues 89 and 98 [102, 113, 114]. strated under the conditions of electrophoresis as is the case
The second involves the N-terminal calcium binding domains. with the complexes of TN-C with TN-I and TN-T. The
Cross links have been demonstrated with the C helix in tendency to copurify may also reflect the similar physical
calcium binding site II ofTN-C [115] and with the region of properties of the two proteins, as both, unlike TN-C, are
alanine 57 ofTN-C [87, 116, 117]. This latter interaction fits strongly basic with high isoelectric points.
in well with that suggested by Herzberg et ai. [118] in their Complexing TN-T with TN-I produces changes in the
model for the mechanism of action of TN-C. It is proposed reactivities of TN-I Iysines in the region of residues 40-98
that on binding calcium at sites I and II the B/C pair of helices [122]. These changes were similar to those observed when
move away from the AID pair to expose a patch of hydro- the reactivities of the Iysines in isolated TN-I were compared
phobic residues that becomes a binding site for TN-I. Hydro- with those of TN-I in the troponin complex. It was concluded
phobic side chains are known to be involved in the interaction that this region ofTN-I is involved in interaction with TN- T.
ofthe inhibitory peptide with TN-C [71, 119] as are the acidic The binding of calcium to the complex produced changes in
side chains of glu84 and asp85 of TN-C [120]. Despite the the reactivities of some Iysines in this region suggesting that
evidence for a range of groups being involved in the TN-I the TN-I1TN-T interaction is modified when TN-C binds
interaction, only a sub-region ofthe TN-C N-terminal domain calcium. Further evidence of TN-T interaction at this region
would appear to be sensitive to bulky covalent adducts. is that cysteine residues 48 and 64, which are accessible to
Covalently linked peptide or biotin at residues 45, 81, 84, and acetamide labelling in TN-I and the TN-I/TN-C complex, are
85 of cardiac C only had a major effect on the transmission of not accessible in the TN-I/TN-T complex or in whole
the calcium signal, as measured by the ATPase level of the troponin [99]. The C-terminal peptide ofTN-T, consisting of
activated myofibril, in the case of residue 81 [121]. residues 159-259, has been shown to be the region involved
The use of mutant forms, and synthetic peptides corre- in interaction with TN-I,23_,26' In the light of the fact that the
sponding to fragments of TN-I, in reconstituted regulated reactivity of Iysines 223 and 226 of TN-T were reduced in
actomyosin and skinned muscle fibre systems has enable the the TN-I/TN-T complex it has been suggested that the region
assignment of distinct functions to different regions of the of residues 223-227 of TN- T is directly involved in the
molecule. By combining the biochemical and biophysical interaction with TN-I. The report that the deletion mutant TN-
evidence of the sites of interaction with the results of studies T I _201 does not bind TN-I supports this conclusion [127].
on the properties of deletion mutants of TN-I Farah et ai. [98] Earlier conclusions about the involvement of the region
have proposed a scheme for the interaction ofthe two proteins. represented by residues 40-98 of TN-I in interaction with
In this scheme the three domains of each protein are aligned TN-T is compatible with more recent studies with deletion
in an antiparallel manner with direct interaction between the mutants of the former protein [98, 127]. From these it is
Nand C-terminals of both proteins. It is clear that the central concluded that the N-terminal consisting of residues 1-98 is
region of the TN-I molecule, residues 96-148 in the rabbit necessary for the incorporation of TN-T in the complex but
22

is not necessary for the calcium regulation of the inhibitory number of workers [131-133]. Both phosphorylase and
activity ofTN-I in a reconstituted filament. This observation protein A kinases were used as phosphorylating agents and
with the skeletal isoform correlates with recent studies on the as the composition of the troponin complex was not well
cardiac protein which has about 30 additional residues at the defined at the time there was some confusion as to which
N-terminus. Whereas mutant mouse cardiac TN-I with residues protein component was the major target. In due course it
I-53 deleted binds to cardiacTN-T, with residues 1-79 deleted became clear that both enzymes phosphorylated TN-I and
it does not [94a]. The TN-T binding region of skeletal TN-I only phosphorylase kinase phosphorylated TN-Tat a signi-
probably does not extend as far as residue 98 for a deletion ficant rate [132-134]. With the isolated protein thrll of rabbit
mutant missing residues I-57 is unable to bind TN-T [128]. fast skeletal TN-I is the major site for phosphorylase kinase,
Analysis of the sequences of TN-I and TN-T from widely with serll7 phosphorylated much more slowly. In contrast
different species confirms the conclusions about the inter- phosphorylation by protein kinase A is largely restricted to
action sites on the two proteins made by more direct studies. serll7 [135, 136]. These sites would appear to be partially
It has been proposed that the evolutionary conserved heptad phosphorylated in vivo for when TN-I is isolated from rabbit
repeat motif with the potential for a-helical coiled coil fast skeletal muscle under conditions designed to minimise
formation that is observed in the amino acid sequences ofTN- endogenous enzymic activity it usually contains about 0.5
I and TN-T represents regions where the proteins interact mole phosphate per mole of protein. This phosphate does
[128a]. In the case of human fast skeletal muscle similarities not appear to be in rapid equilibrium with the intracellular
in sequence exist between residues 20-110 of TN-I and pools for Ribulow et al. [137] were unable demonstrate
residues 160-240 ofTN-T. Both regions contain heptad repeats incorporation of3 2P in TN-I as a consequence of contractile
that could be involved in a-helical coiled coil formation. activity in stimulated frog muscle. The functional signi-
The study of the properties ofthe isolated proteins suggests ficance of phosphorylation at these sites is not clear but
that an important role ofTN-T is to change the TN-C/TN-I serine or threonine residues are present in homologous
complex from a calcium insensitive to a calcium sensitive positions in the sequences of all the isoforms of TN-I. The
form in the MgATPase system. It could do this by reducing N-terminal site is close to a site of interaction with TN-C
the binding constant for the TN-I/TN-C complex in the and the other is adjacent to the inhibitory peptide region
absence of calcium. The results obtained with mutant forms where both actin and TN-C are known to interact. In view of
of TN-I, however, exclude TN-T having this effect by the presence of charged residues in the interaction sites of
interaction in the region of TN-I represented by residues 1- both proteins it is likely that electrostatic forces playa part
98. This mutant exhibits normal calcium sensitivity when in these interactions. The introduction of negatively charged
incorporated into a reconstituted thin filament [98]. Other phosphate residues at or close to the interaction sites would
mutant studies with the N-terminal region of TN-I deleted be expected to affect the binding constants. In this respect it
also emphasise the importance of this region and its inter- is of interest that if ser43/ser45 of mouse cardiac TN-I are
action with TN-T for the calcium sensitivity of the troponin mutated to alanine the calcium sensitivity and Ca2+ activated
regulated MgATPase of actomyosin [121 a, 129]. MgATPase are both reduced [138]. In the presence ofTN-C
Malnic et al. [129a] have confirmed the interaction of the the phosphorylation of fast skeletal TN-I is markedly
C-terminal region ofTN-T (residues 216-263) with the N- inhibited suggesting that on complex formation the phos-
terminal region of TN-I (residue 1-98). In their revised model phorylation sites are blocked. Hydroxy amino acids that are
for the calcium switch of the troponin complex in the thin potential targets for protein kinases are present in homo-
filament they postulate that TN-T interacts with the N-terminal logous positions in skeletal and cardiac TN-I isoforms. In
region of TN-I and not directly with TN-C. It is difficult to the case of the mouse cardiac TN-I ser43/ser45 and thrl44
reconcile this model with the convincing evidence that TN-C are in homologous positions to thrll and serl16 of the rabbit
interacts with the C-terminal region of TN-T (for a review see fast skeletal isoform respectively. The mouse cardiac sites
[12Ia]) and studies that have shown that TN-I with the first have been shown be phosphorylated by protein kinase C
N-terminal 57 residues deleted can function in the troponin [138].
complex with the calcium regulation unchanged [128] The TN-C peptide consisting of residues 83-134 also
inhibits phosphorylation of serine 117 [92]. These observa-
tions fit in well with current views on the regions of TN-I
Phosphorylation of troponin I considered to be involved in interaction with TN-C (see
earlier section and Fig. 2). The lack of incorporation of 32p
Skeletal muscle into TN-I of intact contracting skeletal muscle indicates that
in the myofibril the sites of phosphorylation are not available
The original report [130] that the troponin complex could be to the endogenous kinases presumably because they are
phosphorylated by protein kinase A was followed up by a blocked by other components of the troponin complex.
 23

Cardiac muscle I to exert its effect on the calcium sensitivity of force

development.
N-terminai site It is unlikely that phosphorylation of TN-I is responsible
Following the report [139] that cardiac troponin was phos- for the rise in contractile force that results from adrenaline
phorylated by endogenous protein kinase A it was shown that intervention but represents a negative feed-back mechanism
rabbit cardiac troponin was phosphorylated 5-10 times more in response to the rise in the calcium transients induced by
rapidly than the skeletal complex and that virtually all the intervention with adrenaline. By reducing the sensitivity of
phosphate was incorporated in TN-I [134, 140]. Rabbit the myofibrillarATPase to calcium it has been suggested that
cardiac TN-I differs from its skeletal isoform in that when phosphorylation of cardiac TN-I contributes to the speeding
isolated by affinity chromatography under conditions that up of relaxation that is associated with adrenaline inter-
inhibit endogenous enzymes, its phosphate content is much vention. There is no direct experimental evidence for this,
higher. The amount depends on the conditions of slaughter indeed recent studies employing laser flash photolysis in
and varies from about 1-2 moles per mole of protein. skinned guinea pig trabeculae suggest no change in relaxation
The rate of phosphorylation of the isolated protein by rate occurs on phosphorylation of the TN-I [148]. Experi-
protein kinase A is about 30 times faster than that of the ments with phospholamban knock-out mice also indicate that
skeletal isoform [141]. Also it is not significantly inhibited phosphorylation of cardiac TN-I plays a minor role in PKA
by TN-C. On the other hand phosphorylation of cardiac TN- mediated accelerated relaxation [148a].
I by phosphorylase kinase is strongly inhibited by TN-C. This The major part of the amino acid sequence of cardiac TN-
suggests that the site homologous with thrll in fast skeletal I shows strong homology with the skeletal isoforms but
muscle, which is the preferred site for phosphorylase kinase, differs from them in that the N-terminus is extended by about
is located in a TN-C binding site on cardiac troponin I. Thus 30 residues [43]. This additional N-terminal sequence is
it would appear that unlike the sites close to the TN-C strongly conserved between species and contains the site that
interaction sites, the preferred phosphorylation site on cardiac is further phosphorylated on adrenaline intervention in vivo.
TN-I for protein kinase A, is freely available to enzymes even When rabbit cardiac TN-I is phosphorylated in vitro with
when incorporated in the troponin complex. protein kinase A the majority of phosphate is incorporated
The fact that the phosphate content depends on the con- into the N-terminal site with a few percent ofthe total at serine
ditions of isolation, being lowest when the rabbit was 146 which is homologous with the protein kinase A specific
slaughtered under anaesthesia [141], suggested that the site of rabbit fast skeletal TN-I [149].
phosphorylation state might be related to the physiological The N-terminal site was originally considered to correspond
state of the heart. Further evidence for such an association to serine 20 from the sequence studies that were available at
was provided by the correlation between the contractile the time [43]. It is now known from the work of Mittman et
force developed in the perfused rat heart on application of ai. [150] that there were errors in the original sequence and that
adrenaline and the extent of TN-I phosphorylation [142]. the phosphorylation site consists oftwo residues, serine 22 and
Similar observations were made with the perfused rabbit 23 in the rabbit, both of which can be phosphorylated. This
heart in which the phosphorylation site was found to be report explained the inability of Moir et ai. [151] to identify a
located in the N-terminal cyanogen bromide fragment second site ofphosphorylation in the perfused rabbit heart after
comprising residues 1-48 [143]. In contrast to an earlier adrenaline intervention. In these studies despite the fact that
claim that phosphorylation increased the sensitivity of the virtually all the phosphate was located in the N-terminal peptide
myofibrillar ATPase to calcium [144] it was later shown to and associated with 'serine 20' the value rose to approximately
cause a decrease in calcium sensitivity i.e. the calcium 2 mole per mole after adrenaline intervention.
concentration required for 50% activation was increased It has been long known that interaction of TN-I with TN-C
[143, 145]. These findings have been confirmed and ex- markedly increases the affinity of the latter protein for
tended by many workers (for review see [85]). The change calcium. It is therefore not surprising that covalent modifi-
in calcium sensitivity is independent of the TN-C isoform cation of TN-I, particularly with highly charged phosphate
for it can be obtained in skinned cardiotrabeculae with groups should modify the effects of the interaction. In this
cardiac or skeletal TN-C [146]. Surprisingly, in the light of respect it may be significant that cardiac TN-C differs from
these findings, phosphorylation of recombinant human the fast skeletal isoform in possessing only one calcium
cardiac TN-I introduced into rat and rabbit skinned soleus specific binding site for regulating the contractile response.
fibres did not result in a change in the calcium sensitivity TN-I isolated under conditions to restrict endogenous enzyme
of force production [147]. If both of these findings are activity from the normal beating heart in the anaesthetised
confirmed this would suggest another cardiac specific rabbit usually contains about one mole ofphosphate per mole
myofibrillar protein(s) other than TN-C, possibly TN-T, is of protein, mainly associated with the ser22/ser23 site [151].
essential for the N-terminal phosphorylation of cardiac TN- The decrease in calcium sensitivity of the actomyosin ATPase
24

to calcium accompanies a rise to approximately two moles the special needs of cardiac muscle regulation. The arginine
ofphosphate per mole ofprotein. This would suggest that the residue immediately N-terminal to the serine residue is of
phosphorylation ofboth serine residues is required to produce special significance. Replacement by alanine or methionine
the increase in calcium sensitivity. Recent studies with mutant leads to a reversal of the rate of phosphorylation of the two
forms of cardiac TN-I introduced into skinned fibres have serines [158]. A feature ofthe sequence is that there is another
indicated that this is the case [152]. serine, N-terminal to the phosphorylation site, that is not
The state of the TN-I in the normal unstimulated rabbit phosphorylated.
heart is uncertain but four species could be present, namely If indeed the basal form is mainly ser23serP24, in the
the ser22ser23, serP22ser23, ser22serP23 and serP22serP23 human heart for example, it is questionable whether protein
forms, with the total amount of phosphate not exceeding kinase A is sufficiently active after adrenaline intervention
about one mole per mole of protein. To produce a significant to complete phosphorylation of the site or whether another
change in calcium sensitivity after adrenaline the diphos- enzyme may be involved. In the case of phospholamban
phorylated form must be absent or present as a very small which also possesses an N-terminal dual phosphorylation site,
proportion ofthe total. In that event the TN-I would be present two different enzymes are involved. After adrenaline inter-
as the serP22 and serP23 forms, the proportion depending on vention the TN-I phosphate reverts to the basal level implying
the relative susceptibility of the two serine residues to the that ordered dephosphorylation has taken place. There is
endogenous protein kinases. It is not yet clear as to the precise much yet to be learnt about the enzymic mechanisms involved
phosphorylation state of the normal beating heart in the in the phosphorylation and dephosphorylation of TN-I in the
absence of adrenaline, although there are suggestions that functioning heart.
both monophosphorylated forms are present [150, 153, 154]. The N-terminal peptide characteristic of cardiac troponin
In a recent study of the distribution of the phosphorylated I would appear to be a somewhat flexible structure compared
forms of TN-I in different regions of rabbit and bovine hearts with the rest of the molecule. NMR studies suggest that
the diphosphate was the main form present [154a]. In contrast, regions of cardiac TN-I not interacting with cardiac TN-C are
the monophosphate was the predominant species present in the highly flexible [158a]. The phosphorylation site is available
human heart samples analysed. The physiological states ofthe to kinases and phosphatases in the functioning heart and any
hearts used in these studies were not defined. interactions that occur with other components of the troponin
The fact that ordered phosphorylation of the two serine complex are such as to enable the site to be available to these
residues occurs [155, 156] suggests that if protein kinase A is enzymes. In view of the marked changes in properties of TN-
the only enzyme involved in TN-I phosphorylation, serP23 will Ion conversion from the mono to the diphosphorylated form
be the major form present in the unstimulated rabbit heart. This it is important to distinguish the structural effects associated
conclusion follows from enzymic studies on synthetic peptides with these two forms. Most of the results so far reported have
corresponding to the N-terminal region ofhuman cardiac TN- been obtained with the fully phosphorylated TN-I.
I. In this cardiac isoform, with the phosphorylation site at ser23/ The nature ofthe interaction of cardiac TN-I and TN-C has
24, ser24 is almost completely phosphorylated by protein been less intensively studied than the skeletal system. As
kinase A before traces ofthe diphosphorylated peptide appear might be expected in view of the sequence homology of the
[156]. In this system little evidence ofphosphorylation ofser23 two systems, NMR studies [159] indicate that the con-
was obtained before ser24 was phosphorylated. This implies formation of the two complexes are similar. In both cases
that ser23, whether in the non or monophosphorylated forms complex formation with TN-I decreases interdomain flex-
ofthe N-terminal peptide, is phosphorylated much more slowly ibility and maintains TN-C in an extended conformation.
than ser24. This would appear to be a general property of Evidence has been obtained that the two proteins bind in an
cardiac TN-I phosphorylation by protein kinase A. Enzymic antiparallel fashion and that binding occurs between the
studies on mouse cardiac mutants in which each ofthe serines inhibitory region and the region of the C-terminus oftroponin
of the phosphorylation site (in this species ser22 and ser23) C represented by residues metI20 and met157 [160,161]. A
were replaced in tum by alanine, suggested that in the wild hint of the role of the cardiac N-terminal peptide is provided
type molecule ser23 was phosphorylated by protein kinase A by indications that by binding to the C-terminal ofTN-C it
more rapidly than ser22 [152, 157]. These results strongly can modulate the flexibility of the interdomain linker [159].
suggest that before adrenaline intervention the C-terminal Phosphorylation leads to conformational changes in the
serine of the phosphorylation site is the residue that is phos- TN-I itself[156, 158, 161-163], in the interaction with TN-C
phorylated. This might be expected in view of its position in [164-167] and in the binding of calcium to TN-C in the
the peptide sequence and the known specificity ofthe enzyme. complex [168]. Effects that might be expected in view of the
The sequence RRRSS, which is the N-terminal phosphoryla- change in calcium sensitivity of the troponin system that
tion site of all vertebrate cardiac TN-I so far sequenced, is occurs as a consequence ofphosphorylation. By studying the
rather unusual and has presumably evolved to accommodate changes in distance between trp 192 and cysteine residues
 25

introduced at different positions in mutant cardiac TN-I skeletal complex is stable. Unlike the skeletal system the
molecules it has been concluded from flourescent resonance binding of cardiac TN-I to actin/tropomyosin (skeletal) shows
energy transfer measurements that the structural changes marked cooperativity. This is lost on phosphorylation [165].
induced by PKA phosphorylation are confined to the N- NMR studies on the phosphorylation by protein kinase A
terminal half of the molecule [168a]. The binding constant of a synthetic peptide corresponding to the region of the
was shown to be little changed by removing the N-terminal N-terminal phosphorylation site of human cardiac TN-I have
32 residues of mouse cardiac TN-I [163]. The results of Al- thrown some light on the conformational changes that occur
Hillawi et al. [165] indicate that calcium strengthens the [158] . Phosphorylation of ser24 leads to interaction between
binding of the cardiac isoforms of TN-I to TN-C much less ser24 and arg 22 and some decrease in flexibility in this
than is the case with the skeletal isoforms. Surprisingly region. The major change occurs on phosphorylating ser23
phosphorylation desensitised the interaction to calcium to the which leads to a marked strengthening of the interaction
extent that the cardiac complex no longer migrated as a between the phosphate groups and arginine residues 20, 21
complex on electrophoresis under conditions in which the and 22. The result is that, unlike the monophosphate, the

Fig. 4. Conformation about the phosphorylation site of the N-terminal region of human cardiac TN-I deduced from NOE data obtained from a synthetic
peptide corresponding to residues 17-30 [158].
26

diphosphate is unable to bind a paramagnetic anionic probe to alanine the calcium sensitivity and the Ca 2+ activated
due to screening of the arginine side chains. Phosphorylation MgATPase is reduced. Unlike the situation with the wild type
also abolishes the binding ofthe peptide to TN-C presumably isoform phosphorylation of ser23/ser24 of the mutant does
because the arginine side chains are no longer available for not further decrease the calcium sensitivity [138, 174). It has
electrostatic interactions. It should be pointed out that any been suggested that phosphorylation at these sites might be
interactions that occur in vivo between the N-terminal region responsible for the negative inotropic effect of phorbol esters
of cardiac TN-I and TN-C cannot involve tight binding as is on various cardiac preparations [173]. Evidence that these
the case with other regions of TN-I for the N-terminal site sites have been shown to be phosphorylated in cardiac
must be freely available to kinase and phosphatase. The myocytes has been taken to support such a role. It should be
structure of the peptide chain in the region of the phos- noted that phosphorylation of such sites is not unique to
phorylation site as computed from NOE spectrum data is cardiac muscle. The phosphorylation of ser146 of rabbit
illustrated in Fig. 4. The looped constrained structure is cardiac TN-I by protein kinase A in not markedly inhibited
compatible with the conclusions made from fluorescence by TN-C, unlike phosphorylation of other sites by phos-
studies on a cys5 mutant of mouse cardiac TN-I labelled with phorylase kinase [141, 175].
the sulphydryl reagent IAANS [162). The protein kinases are not completely specific for the
potential phosphorylation sites and under some conditions,
Phosphorylation at sites other than N-terminal for example, protein kinase C will phosphorylate ser23/ser24
Cardiac TN-I possesses potential phosphorylation sites in of mouse cardiac TN-I [138,176). It is clear that the modula-
homologous positions to thr11 and ser117 in the rabbit skeletal tion of calcium sensitivity by phosphorylation of the N-
muscle isoform. These sites are partially phosphorylated in terminal extension is unique to cardiac muscle and it is very
skeletal muscle but as the phosphate does not rapidly equili- probable that in the response in vivo to adrenaline, protein
brate with the intracellular phosphate pools they are probably kinase A is the enzyme involved. It is more likely that
involved in long term modulation ofthe TN-IffN-C interaction phosphorylation of the homologous potential phosphoryla-
[169]. The role of phosphorylase and protein A kinases in tion sites in TN-I isoforms, if indeed they have a modulatory
phosphorylating homologous sites in cardiac troponin has not function, has a role that is common to the contractile process
been investigated in any detail. Serl46 of rabbit cardiac TN-I in striated muscle in general, rather one than specific for
has been shown to be phosphorylated by protein kinase A in cardiac tissue. Much has to be learnt about the detailed role
vitro although at a much slower rate than ser22/23 [149, 151). of protein kinase C in regulating cardiac contractility in vivo
Cardiac TN-I contains a number of hydroxy amino acids for it will also phosphorylate TN- T. In the latter case the
that are potential sites for phosphorylation of the isolated pattern of phosphorylation depends on the isoforms ofprotein
protein. In the rabbit there are 10 serines and 11 threonine kinase C present [177] and presumably the same will apply
residues and all the vertebrate cardiac TN-Is so far se- to TN-I. The protein kinase C gene family is complex and
quenced (Swissprot) have a similar number with the hy- the pattern of isoform expression is tissue specific. Most of
droxy amino acids in homologous positions. Whether these the in vitro phosphorylation studies on cardiac TN-I have
residues are available for rapid phosphorylation and de- been carried out with porcine brain protein kinase C which
phosphorylation, as is required for a dynamic modulatory consists principally of the ex isoform. Confirmation with
role, will depend on how exposed the site is in the troponin isoforms specific for cardiac muscle would indicate whether
complex in the living muscle. Sites available for phos- phosphorylation of TN-I by protein kinase C has a unique
phorylation in all isoforms will be limited as much of the modulatory role in that tissue. Some progress in this direction
surface must be blocked by the three proteins, TN-C, TN-T has been made by the indication that there is decreased
and actin, with which it must interact to function. This is myofilament responsiveness to calcium in hearts from
amply born out by the inhibition by TN-C of the phos- transgenic mice in which the protein kinase C ~2 isoform is
phorylation by phosphorylase and protein A kinases of rabbit over expressed [1 77a).
fast skeletal TN-I [134]. Significantly the major regulatory
site is located on the flexible exposed extension of the
polypeptide chain at the N-terminus of the cardiac isoform Conclusions
and is not blocked by TN-C.
Protein kinase C can phosphorylate mouse cardiac TN-I Although detailed understanding of the mode of action of the
at ser43/ser45 [138, 170-172). Phosphorylation at these sites troponin complex is still not yet available it is clear that TN-
does not alter the calcium sensitivity but results in a decrease I occupies a central position in the molecular mechanisms
in the Ca 2+ activated MgATPase of actomyosin [138, 174]. involved in the control of contractile activity. It acts as the
Apart from their possible role as phosphorylation sites these major link between the troponin complex and the contractile
residues have functional significance. If both are mutated proteins. As such it conveys the conformational changes
 27

resulting from the binding ofcalcium on TN-C to the enzymic 2. Perry SV, Grey TC: A study of the effects of substrate concentration
site on the myosin head where the MgATP is hydrolysed. This and certain relaxing factors on the magnesium activated myofibrillar
ATPase. Biochem J 64: 184-192, 1956
unique role requires that it must interact with each of the 3. Perry SV, Grey TC: Ethylenediaminetetracetate and the ATPase
proteins involved in this process, directly with actin, TN- activities of actomyosin systems. Biochem J 64: 5, 1956
C, TN-T and possibly indirectly with tropomyosin and 4. Ebashi S: Third component participating in the superprecipitation of
myosin. Its extended, apparently flexible, structure must be natural actomyosin. Nature 200: 1010,1963
important for this role. Considerable progress has been 5. Perry SV, Davies V, Hayter D: 'Natural actomyosin' and the factor
sensitising actomyosin adenosine triphosphatase to ethylene bis
made in identifying regions of the TN-I molecule concerned (ethyleneamine) tetraacetate. Biochem J 99: Ic, 1966
with its function. With the present state of knowledge there 6. Hartshorne DJ, Perry SV, Davies V: A factor inhibiting the adenosine
may, however, be dangers in being too positive about the triphosphatase activity and superprecipitation of actomyosin. Nature
assignment of TN-I properties to precisely defined regions 209: 1352, 1966
of the polypeptide chain. It is likely that some of the pro- 7. Hartshorne DJ, Perry SV, Schaub MC: A protein factor inhibiting the
magnesium activated adenosine triphosphatase of desensitized
perties ofthe wild type protein depend on the entire molecule actomyosin. Biochem J 104: 907-913, 1967
being intact. Mutation or chemical modification at one 8. Hartshorne DJ, Mueller H: Fractionation of troponin into two
position could cause changes elsewhere in the molecule that components. Biochem Biophys Res Com 31: 647---653, 1968
are important for the property under investigation. This may 9. Schaub MC, Perry SV: The relaxing factor system of striated muscle.
explain some of the discrepancies that exist in assigning Resolution of the troponin complex into inhibitory and calcium
ion-sensitizing factors and their relationship to tropomyosin. Biochem
properties to certain regions e.g. the role of N-terminal J 115: 993-1004,1969
regions in regulating the Ca2+sensitization of the actomyosin 10. Wilkinson JM, Perry SV, Cole HA, Trayer IP: The regulatory proteins
MgATPase [cf 80, 97, 98, 129, 138]. of the myofibril. Characterisation and biological activity of the
TN-I has derived its name from the property that led to its components of inhibitory factor. Biochem J 127: 215-224, 1972
discovery but there must be some doubt whether its role in the II. Greaser ML, Yamaguchi M, Brekke C, Potter J, Gergely J: Troponin
subunits and their interactions. Cold Spring Harbour Sym Quant BioI
myofibril is to act directly as an inhibitor. According to the 37: 235-244,1972
steric hypothesis tropomyosin has this role. Perhaps its most 12. Perry SV, Cole HA, Head JF, Wilson FJ: Localization and mode of
significant property is its ability to induce different con- action of the inhibitory protein component of the troponin complex.
formational changes in each of the proteins with which it Cold Spring Harbour Sym Quant Bioi 37: 251-262, 1972
interacts in order to facilitate their function. With TN-C it 13. Endo T, Obi nata T: Troponin and its components from ascidian muscle.
J Biochem 89: 1599-1608, 1981
increases the affinity for calcium so that this cation effectively 14. Ojima T, Nishita K: Isolation oftroponins from striated and smooth
triggers contraction. In cardiac muscle, by virtue of an addi- adductor muscle of Akazara scallop. J Biochem 100: 821-824, 1986
tional N-terminal phosphorylation site, it is able to modulate 15. Myers CD, Goh P- Y, Allen TS, Bucher EA, Bogaert T: Develop-
the process by changing the affinity ofTN-C for calcium in a mental genetic analysis of troponin T mutations in striated and
dynamic way in response to hormonal influences. One mole- non-striated muscle cells of Caenorhabitis elegans. J Cell Bioi 132:
1061-1077,1996
cule of TN-I also interacts with actin and apparently induces 16. Syska H, Perry SV, Trayer IP: A new method for the isolation of
cooperativity of function over a stretch of seven actin troponin I (inhibitory protein) using affinity chromatography. Evidence
monomers. Whatever the explanation of this effect, whether for three different forms oftroponin 1 in striated muscle. FEBS Letts
tropomyosin has an active or passive role, the crucial event is 41: 253-257,1974
the binding ofTN-I to actin. A more apt description ofthe role 17. Sunderland CJ, Esser KA, Eisum KA, Gordon MC, Hardeman EC:
Identification of a program of contractile protein gene expression
of TN-I might be to call it a facilitator in that by interaction initiated upon skeletal muscle differentiation. Dev Dyn 196: 25-36,
with associated proteins it endows them with properties of 1993
functional importance that they do not possess in its absence. 18. Sabry MA, Dhoot GK: Identification and pattern of expression of a
developmental isoforrn oftroponin I in chicken and rat cardiac muscle.
J Muscle Res Cell Mot 10: 85-91,1989
19. Saggin L, Gorza L, Ausoni S, Schiaffino S: Troponin I switching in
Acknowledgements developing heart. J BioI Chern 264: 16299-16302, 1989
20. Murphy AM, Thompson WR, Peng FL, Jone " L: Regulation of the
cardiac troponin I gene by the transcription factor GATA-4. Biochem
I am grateful to Val Patchell and Barry Levine for their helpful J 322: 393-401, 1997
comments on this review and their help in its preparation. 21. Westfall MV, Rust EM, Metzger JM: Slow skeletal troponin I gene
transfer, expression and myofilament incorporation enhances adult
cardiac myocyte contractile function. Proc Nat! Acad Sci 94: 5444-
5449, 1997
References 22. Jackson DM, Heeley DH: Isoforrns oftroponin 1in salmonid myotomal
muscle. Biophys J 372: A58, 1997
I. Kumagi H, Ebashi S, Takeda F: Essential relaxing factor in muscle other 23. Dhoot GK, Perry SV: The components of the troponin complex and
than myokinase and creatine phosphokinase. Nature 176: 166, 1955 development in skeletal muscle. Expt! Cell Res 127: 75-87, 1980
28
23a. Barton PJR, Townsend PJ, Brand NJ, Yacoub MH: Localization of 46. Campbell AP, Sykes BD: Interaction oftroponin I and troponin C. Use
the fast skeletal muscle troponin I gene (TNNT2) to IlpI5.5: Genes oftwo-dimensional nuclear magnetic resonance transferred nuclear Over-
for troponin I and troponin T are organised in pairs. Hum Genet 61: hauser effect to determine the structure ofthe inhibitory troponin I peptide
519-523, 1997 when bound to skeletal troponin C. J Mol Bioi 222: 405-421,1991
23b. Mogensen J, KruseTA, BorglumAD: Assignment ofthe human cardiac 47. Campbell AP,Van Eyk JE, Hodges RS, Sykes BD: Interaction of
troponin I gene (TNNI3) to chromosome 19q13.4 by radiation hybrid troponin I and troponin C; use of the two dimensional transferred
mapping. Cytogenet Cell Genet 79: 272-273, 1997 nuclear Overhauser effect to determine the structure of a gly-II 0
24. Dhoot GK, Perry SV: Distribution of the polymorphic forms oftroponin inhibitory troponin I peptide analogue when bound to cardiac troponin
components and tropomyosin in skeletal muscle. Nature 278: 714- C. Biochem Biophys Acta 1160: 35--54, 1992
718,1979 48. Saijo Y, Takeda S, Scherer A, Kobayashi T, Maeda Y, Taniguchi H,
25. Dhoot GK, Gell PGH, Perry SV: Localisation of the different forms Yao M, Wakatsuki S: Production, crystallisation and preliminary X-ray
of troponin I in skeletal and cardiac muscle. Exp Cell Res 117: 357- anaysis of rabbit skeletal muscle troponin C and fragment (1-47) of
370, 1978 troponin I. Protein Sci 6: 916--918,1997
26. Cummins P, Perry SV: Troponin I from human skeletal and cardiac 48a. Vassylyev DG, Tokeda S, Wakatsuki S, Maeda K, Maeda Y: Crystal
muscles. BiochemJ 171: 251-259,1978 structure of troponin C in complex with troponin I fragment 2.3A
27. Mair J: Cardiac troponin I and troponin T: Are enzymes still relevant resolution. Proc Natl Acad Sci USA 95: 4847-4852, 1998
as markers? Clin Chern Acta 257: 99-115, 1997 49. Stone DB, Timmins PA, Schneider DK, Krylova I, Ramos CHI,
28. Dhoot GK, Perry SV, Vrbova G: Changes in the distribution of the Reinach FC, Mendelsohn RA: The effect of regulatory Ca'· on the in
components of the troponin complex in muscle fibres after cross situ structures oftroponin C and troponin I: A neutron scattering study.
innervation. Exp Neur 72: 513-530, 1981 J Mol BioI 281: 689-704, 1998
29. Dhoot GK, Perry SV: The effect of thyroidectomy on the distribution 50. Haselgrove JC: X-ray evidence for conformational change in the actin-
of slow forms of troponin I in rat soleus muscle. FEBS Letts 133: containing filaments of vertebrate striated muscle. Cold Spring Harbor
225--229, 1981 Symp Quant Bioi 37: 341-352, 1972
30. Dhoot GK, Perry SV: The use of antibodies to the regulatory proteins 51. Huxley HE: Structural changes in actin- and myosin-containing
in the study of phenotype expresion in normal and diseased muscle. filaments during contraction. Cold Spring Harbor Symp Quant Bioi
In: S. Ebashi (ed). Muscular Dystrophy. University of Tokyo Press, 37: 361-378,1972
Tokyo,1982,pp.467-470 52. Taylor K,Amos LA: A new model for the geometry of binding of myosin
31. Bullard B, Dabrowska R, Winkelman L: The contractile and regulatory crossbridges to muscle thin filaments. J Mol Bioi 147: 297-324,1981
proteins of insect flight muscle. Biochem J 135: 277-286, 1973 53. Lehman W, Craig R, Vibert P: Cal+ induced tropomyosin movement
32. Benzonana G, Kohler L, Stein EA: Regulatory proteins of crayfish in Limulus thin filament revealed by three dimensional reconstruction.
tail. Biochem Biophys Acta 368: 247-258, 1974 Nature 368: 65-67, 1994
33. Regenstein JM, Szent-Gyorgyi AG: Regulatory proteins of lobster 54. Eaton BL, Kominz DR, Eisenberg E: Correlation between the inhibition
striated muscle. Biochemistry 14: 917-925,1975 of the acto-heavy meromyosin ATPase and the binding of tropomyosin
34. Lehman W, Regenstein JM, Ransom AL: The stoichiometry of the to F-actin; effects of Mg'., KCI, troponin I and troponin C. Bio-
components of arthropod thin filaments. Biochem Biophys Acta 434: chemistry 14: 2718-2725,1975
215-222,1976 55. Katz AM: Influence of tropomyosin upon the reactions of actomyosin
35. Kobayashi T, Takagi T, Konishi K, Cox JA: Amino acid sequence of at low ionic strength. J BioI Chern 239: 3304-3311, 1964
crayfish troponin I. J BioI Chern 264: 1551-1557, 1989 56. Williams Jr DL, Greene LE, Eisenberg E: Comparison of effects of
36. Shinoda Y, Yamada A, Yagi, K: Identification oftroponin I of crayfish smooth and skeletal muscle tropomyosins on interactions of actin and
myofibrils. J Biochem 103: 636-Q40, 1988 myosin subfragment I. Biochemistry 23: 4150-4155, 1984
37. Ojima T, Tanaka H, Nishita K: Cyanogen bromide fragments of 57. Heeley DH, Watson MH, Mak AS, Dubord P, Smillie LB: Effect of
Akazara scallop M, 52000 troponin I. J Biochem 108: 519-521, 1990 phosphorylation on the interaction and functional properties of rabbit
38. Ojima T, Tanaka H, Nishita K: Amino acid sequence ofC-terminal 17 skeletal muscle aa-tropomyosin. J BioI Chern 264: 2424-2430, 1989
kDa CNBR fragment of Akazara scallop troponin I. J Biochem 117: 58. Schaub MC, Hartshorne DJ, Perry SV: The adenosine triphosphatase
158-162,1995 activity of desensitised actomyosin. Biochem J 104: 263-269, 1967
39. Yuasa HJ, Sato S, Yamamoto H, Takagi T: Primary structure oftroponin 59. Schaub MC, Ermini M: Effect of bivalent cations on the adenosine
I isoforms from Halocynthia roretzi. J Biochem 123: 374-380, 1997 triphosphatase activity of actomyosin and its modification by
40. Bullard B, Leonard K, Larkins A, Butcher G, Karlik C, Fyrberg E: tropomyosin and troponin. Biochem Jill: 777-783, 1969
Troponin of asynchronous flight muscle. J Mol Bioi 204: 621-637, 60. Syska H, Wilkinson JM, Perry SV: The relationship between biological
1988 activity and primary structure oftroponin I from white skeletal muscle
41. Karlik CC, Mahaffey JW, Coutu MD, Fyrberg EA: Organisation of of the rabbit. Biochem J 153: 375-387,1976
contractile genes within the 88F subdivision of the D rnelanogaster 61. Nozaki S, Kobayashi K, Katayama E, Matsumura I: Synthetic studies
third chromosome. Cell 37: 469-481, 1984 on troponin I active site. Preparation of a pentadecapeptide with
42. Barbas JA, Galceran J, Krah-Jentgens I, de la Pompa JL, Canal I, inhibitory activity towards actomyosin adenosine triphosphatase. Chern
Pongs 0, FerrusA: Troponin I is encoded in the haplolethal region of Letts 3: 345--348, 1980
the Shaker gene complex of Drosophila. Genes Dev 5: 132-140,1991 62. Talbott JA, Hodges RS: Synthetic studies on the inhibitory region of
43. Wilkinson JM, Grand RJA: Comparison of amino acid sequence of rabbit skeletal troponin I. J Bioi Chern 256: 2798-2802, 1981
troponin I from different striated muscles. Nature 271: 31-35, 1978 63. Pearlstone JR, Smillie LB: Effects oftroponin I plus -C on the binding
44. Farah CS, Reinach FC: The troponin complex and regulation of muscle oftroponin T and its fragments to a-tropomyosin. J Bioi Chern 258:
contraction. FASEB J: 755-767,1995 2534-2542, 1983
45. Pearlstone JR, Sykes BD, Smillie LB: Interactions of structural C and 64. Zhou X, Morris EP, Lehrer SS: Troponin I and troponin-C binding to
N regulatory domains oftroponin C with repeat motifs in troponin 1. actin tropomyosin and dissociation by myosin S I. Biophys J 68: A167,
Biochemistry 36: 7601-7606, 1997 1995
 29
65. Bremel RD, Weber A: Cooperative behaviour within the functional 85. Solaro RJ, Van Eyk J: Altered interactions among thin filament proteins
unit of the actin filament in vertebrate striated muscle. Nature 238: modulate cardiac function. J Mol Cell Cardiol 28: 217-330, 1996
97-101,1977 86. Leszyk J, Collins JH, Leavis PC, Tao T: Cross linking ofrabbit skeletal
66. AI-Khayat HA, Vagi N, Squire JM: Structural changes in actin- troponin I with the photoactive reagent 4-maleimidobenzophenone:
tropomyosin during muscle regulation: computor modelling of low Identification of residues in troponin I that are close to cysteine 98
angle X-ray diffraction data. J Mol Bioi 252: 611-{j32, 1995 oftroponin C. Biochemistry 26: 7042-7047, 1987
67. Grand RJA, Henry DG, Moir AJG, Perry SV, Trayer IP, Dalgarno, 87. Kobayashi T, Tao T, Grabarek Z, Gergely J, Collins JH: Cross-linking
DC, Levine BA, Parker SB: Modulation by troponin C of the troponin of residue 57 in the regulatory domain of mutant rabbit skeletal muscle
I inhibition of skeletal actomyosin interactions. A PMR study, In: B. troponin C to the inhibitory region of troponin I. J BioI Chern 266:
de Barnard, G.L. Sottocasa, G. Sandri, E. Carafoli, A.N. Taylor,T.C. 13746-13751,1991
Vanaman and RJ.P. Williams (eds).Calcium Binding Proteins. Elsevier/ 88. Luo J, Qian J, Wu J-L, Tao T, Gergely J: Inter-subunit crosslinking
North Holland, Amsterdam, 1983, pp. 379-380 of single cys mutant troponin I in binary and ternary complexes.
68. Ishikawa T, Wakabayashi T: Calcium induced change in the three Biophys J 72: A59, 1997
dimensional structure of thin filaments of rabbit skeletal muscles as 89. Ngai SM, Sonnichensen FD, Hodges RS: Photochemical cross-linking
revealed by cryoelectron microscopy. Biochem Biophys Res Com 203: between native rabbit skeletal troponin C and benzylbenzoyl inhibitory
951-958,1994 peptide, residuesI04-115. J Bioi Chern 269: 2165-2172, 1994
69. Van Eyk JE, Strauss JD, Hodges RS, Ruegg JC: A synthetic peptide 90. Tao T, Gong B-J, Gergely J: The conformation oftroponin I studied
mimics troponin I function in the calcium regulation of muscle by chymotryptic digestion. Biophys J 72: A59, 1997
contraction. FEBS Lett 323: 223-228, 1993 91. Amphlett GW, Vanaman TC, Perry SV: Effect of the troponin C-like
70. Talbot JA, Hodges RS: Comparative studies on the inhibitory region of protein from bovine brain (brain modulator protein) on the Mg 2._
selective species oftroponin I. J Bioi Chern 256: 12374-12378,1981 stimulated ATPase of skeletal muscle actomyosin. FEBS Letts 72:
71. Dalgarno DC, Grand RJA, Levine BA, Moir MG, Scott GMM, Perry 162-168, 1976
SV: Interaction between troponin I and troponin C. Definition of 92. Weeks RA, Perry SV: Characterisation of a region of the primary
topography by proton magnetic resonance spectroscopy. FEBS Letts sequence oftroponin C involved in calcium ion dependent interaction
150: 54-58, 1982 with troponin I. Biochem J 173: 449--457, 1978
72. Levine BA, Moir MG, Perry SV: The interaction of troponin I with 93. Leavis PC, Rosenfeld S, Gergely J, Grabarek Z, Drabikowski W:
the N-terminal region of actin. Eur J Biochem 172: 389-397, 1988 Proteolytic fragments of troponin C. Localisation of high and low
73. Grabarek Z, Gergely J: Troponin I binds to the N-terminal twelve affinity Ca 2• binding sites and interactions with troponin I and
residue segment of actin. Biophys J 51: 331 a, 1987 troponin T. J BioI Chern 253: 5452-5459, 1978
74. Jiang J, Wu J-L, Gergely J, Tao T: Ca'· induced movement of the N- 93a. Leszyk J, Tao T, Nuwaysir LM, Gergely J: Identification of
terminal region of troponin I relative to troponin C and actin. Biophys photocrosslinking sites in troponin I with 4-maleimidobenzophenone
J 68: A165, 1995 labelled mutant troponin-Cs having single cysteines at positions 21
75. Tao T, Gong BJ, Leavis PC: Calcium induced movement oftroponin I and 158. J Muse Res Cel Motil 19: 479--490, 1998
relative to actin in skeletal muscle thin filaments. Science 247: 1339- 94. Katayama E, Nozaki S: Ca2• dependent binding of synthetic peptides
1441,1990 corresponding to some regions oftroponin I to troponin C. J Biochem
76. Li HC, Palm T, Sale K, Fajer PG: Crossbridges-induced movement of 91: 1449-1452,1982
troponin I relative to actin in skeletal muscle thin filaments. Biophys 94a. Rarick HM, Tang H-P, Guo X-D, Tu X-H, Martin AF, Solaro RJ:
J 72: A248, 1997 Interations of the N-terminal region of cardiac TNI with the structural
77. Li Z, Gergely J, Tao T: Resonant energy transfer and photocrosslinking cTNC site and cTNT are involved in the Ca 2• dependent activation
studies of the interaction between troponin C and the inhibitory region of MgATPase activity. Biophys J 74: A52, 1998
of a monocysteine mutant oftroponin I. Biophys J 68: A 165, 1995 95. Bingham RP, Trayer IP: Characterisation ofa troponin C binding site
78. Lehrer SS, Chai M, Geeves MA: Effects of troponin I on actin S I on the N-terminal oftroponin I. J Muscle Res Cell Mot 18: 260, 1997
ATPase and S 1 binding kinetics in the absence and presence of rabbit 96. Ngai SM, Hodges RS: Biologically important interactions between
skeletal tropomyosin and troponin. Biophys J 72: A60, 1997 synthetic peptides of the N-terminal region oftroponin I and troponin
79. Zang R, Sheng Z, Jones M, Parsons B, Potter JD: Effect of mutations C. J BioI Chern 267: 15715-15720, 1992
in the inhibitory region of skeletal muscle troponin I on its biological 97. Sheng Z, Pan B-S, Miller TE, Potter JD: Isolation, expression and
activity. Biophys J 70: A 165, 1996 mutation of a rabbit muscle eDNA clone for troponin I. J BioI Chern
80. Van Eyk JE, Thomas LT, Tripet, B, Wiesner RJ, Pearlstone JR, Farah 267: 25407-25413, 1992
CS, Reinach FC, Hodges RS: Distinct regions of troponin I regulate 98. Farah CS, Miyamoto CA, Ramos CH, da Silva ACR, Quaggio RB,
Ca'·-dependent activation and Ca'· sensitivity of the acto-SI-TM- Fujimori K, Smillie LB, Reinach FC: Structural and regulatory
ATPase activity ofthin filaments. J Bioi Chern 272: 10529-10537,1997 functions of the NH 2 and COOH-terminal regions of skeletal muscle
81. Tripet B, Van Eyk JE, Hodges RS: Mapping of a second actin- troponin I. J Bioi Chern 269: 5230-5240, 1994
tropomyosin and a second troponin C binding site within the C terminus 99. Chong PSC, Hodges RS: Proximity of sulphydryl groups to the sites
oftroponin I and their importance in the Ca'·-dependent regulation of of interaction between components of the troponin complex from
muscle contraction. J Mol Bioi 271 : 728-750, 1997 rabbit skeletal muscle. J Bioi Chern 257: 2549-2555,1982
82. Rarick HM, Tu X-H, Solaro RJ, Martin A: The C terminus of cardiac 100. Wang CK, Cheung HC: Proximity relationship in binary complex formed
troponin 1 is essential for full inhibitory activity and Ca'· sensitivity between troponin I and troponin C. J Mol Bioi 190: 509-521, 1984
of rat myofibrils. J BioI Chern 272: 26887-26892, 1997 101. Tao T, Gowell E, Strasbourg GM, Gergely J, Leavis PC: Ca2• de-
83. Head JP, Perry SV: The interaction of the calcium binding protein pendence of the distance between cys98 oftroponin C and cysl33 of
(troponin C) with bivalent cations and the inhibitory protein (troponin troponin I in the ternary complex. Biochemistry 28: 5902-5908, 1989
I). Biochem J 137: 145-154,1974 102. Kobayashi T, Tao T, Gergely J, Collins JH: Structure of the troponin
84. Grabarek Z, Tao T, Gergely J: Molecular mechanism of troponin C complex. Implications of photo crosslinking oftroponin I to troponin
function. J Muscle Res Cell Mot 13: 383-393, 1992 C thiol mutants. J BioI Chern 269: 5725-5729, 1994
30

103. Keller RS, Rarick HM, Martin AF, Solaro RJ: Effect of truncated 117. Wang Z, Wang J, Sarkar S, Gergely J, Tao T: The Band C helices in
and a C-terminal mutant of cardiac TN-Ion myofilament force. the N-terminal domain oftroponin C interact with troponin I. Biophys
Biophys J 72: A379, 1997 J 57: 154a, 1990
104. Ingraham RH, Swenson CA: Binary interactions oftroponin subunits. 118. Hertzberg 0, Moult J, James MNG: A model for the Ca2 +-induced
J Bioi Chern 259: 9544-9548,1984 conformational transition oftroponin C. J Bioi Chern 261: 2638-26,
105. Leavis PC, Gowell E, Tao T: Flourescence lifetimes and acrylamide 1986
quenching studies of the interaction between troponin subunits. 119. Grand RJA, Levine BA, Perry SV: Proton magnetic resonance studies
Biochemistry 23: 4156-4161,1984 on the interaction of rabbit skeletal muscle troponin I with troponin
106. Wang CK, Cheung HC: Energetics of the binding of calcium and C and actin. Biochem J 203: 61-68,1982
troponin I to troponin C. Biophys J 48: 727-739, 1985 120. Kobayashi T, Zhao X, Wade R, Collins JH: Effect of the acidic residues
107. Cheung HC, Wang CK, Malik NA: Interactions oftroponin subunits: of the N-terminal domain of troponin C on the calcium dependent
Free energy of binary and tertiary complexes. Biochemistry 26: 5904- interaction with the inhibitory peptide. Biophys J 70: A38, 1996
5907, 1987 121. Lin X, Dotson DG, Putkey JANA: Covalent binding of peptides to
108. Potter JD, Gergely J: The calcium and magnesium binding sites on the N terminal hydrophobic region of cardiac troponin C has limited
troponin and their role in the regulation of myofibrillar adenosine effects on function. J Bioi Chern 271: 244-249, 1996
triphosphatase. J BioI Chern 250: 4628-4633, 1975 12 Ia. Perry SV: Troponin T: Genetics, properties and function. J Muse Res
108a. McKay RT, Tripet BP, Hodges RS, Sykes BD: Interaction of the Cell Motil 19: 575-602, 1998
second binding region of troponin I with the regulatory domain of 122. Hitchcock-Gregori SE: Study of the structure of troponin I by
troponin C as determined by NMR. Biophys J 74: A298, 1998 measuring the relative reactivities oflysines with acetic anhydride. J
I08b. Luo Y, Qian Y, Lesyk J, Gergley G, Tao T, Residues 48 and 82 at the BioI Chern 257: 7372-7380, 1982
N-terminal hydrophobic patch of troponin-C photocrosslink to 123. Ohtsuki I: Molecular arrangement oftroponin T in the thin filament.
methionine 121 oftroponin-I. Biophys J 74: A143, 1998 J Biochem 86: 491-497,1979
I08c. Li MX, Spyrocoupolos L, Sykes BD: Interaction of human cardiac 124. Katayama E: Interaction of troponin I with troponin T and its
Troponin-I (residues 148-161) wi th the regulatory domain of human fragment. J Biochem 85: 1379-1381,1979
cardiac Troponin C. Biophys J 74: A51, 1998 125. Pearlstone JC, Smillie LB: The binding sites oftroponin I on troponin
I08d. Davis B, Liu W, Putkey JA: Effect of troponin I association of the T. Can J Biochem 58: 649, 1980
calcium-bound conformation of cardiac troponin C. Biophys J 74: 126. Tanokura M, Tawada Y, OnoA, Ohtsuki I: Chymotryptic subfragments
A143,1998 oftroponin T from rabbit skeletal muscle. Interaction with tropomyosin,
109. Olah GA, Trewewhella J: A model structure of the muscle protein troponin I and troponin C. J Biochem 93: 331-337, 1983
complex 4Ca H troponin C. troponin I derived from small angle 127. Jha Pk, Leavis PC, Sarkar S: Interaction of deletion mutants of
scattering data: Implications for regulation. Biochemistry 33: 12800- troponins I and T: COOH-terminal truncation oftroponin T abolishes
12806,1994 troponin I binding and reduces Ca 2+ sensitivity of the reconstituted
110. Olah GA, Rokop SE, Wang CLA, Blechner SL, Trewhella J: Troponin regulatory system. Biochemistry 35: 16573-16580,1996
I encompasses an extended troponin C in the Ca 2' bound complex: A 128. Potter JD, Sheng Z, Pan B-S, Zhao: A direct regulatory role for
small angle X-ray and neutron scattering study. Biochemistry 33: troponin T and a dual role for troponin C in the Ca 2+ regulation of
8233-8239, 1994 muscle contraction. J Bioi Chern 270: 2557-2562, 1995
Ill. Tao T, Qian Y, Boldogh I, Gergely J: Flourescent lifetime, acrylamide 128a. Stefancsik R, Jha PK, Sarkar P: Identification and mutagenesis of a
quenching and photo crosslinking studies between troponin I and highly conserved domain in troponin T responsible for troponin I
troponin T and monocysteine mutants of troponin C. Biophys J 68: binding: Potential for coiled coil interaction. Proc Nat! Acad Sci USA
A166,1995 95: 957-962, 1998
112. Luo Y, Wu J-L, Gergely J, Tao T: Troponin T and Ca 2' dependence 129. Thomas L, Van Eyk JE, Farah CS, Reinack FC, Hodges RS, Smillie
on the distance between cys48 and cys 133 in the ternary troponin LB: The role oftroponin I and troponin T in the calcium dependent
complex and reconstituted thin filaments. Biochemistry 36: 11027- regulation of reconstituted thin filaments. Biophys J 70: A 166, 1996
11035,1997 129a. Malnic B, Farah CS, Reinach FC: Regulatory properties of the NH 2-
113. Leszyk J, Collins JH, Leavis PC, Tao T: Cross-linking of rabbit and COOH-terminal domains oftroponin T. J BioI Chern 273: 10594-
skeletal muscle troponin subunits: Labelling of cys 98 of troponin C 10601,1998
with 4-maleimidobenzophenone and analysis of products formed in 130. Bailey C, Villar-Palasi C: Cyclic AMP dependent phosphorylation
the binary complex with troponin T and the ternary complex with oftroponin. Feb Proc 30: 1147, 1971
troponins I and T. Biochemistry 27: 6983-6987,1988 131. Stull JT, Brostrom CO, Webs EG: Phosphorylation of the inhibitor
114. Leszyk J, Grabarek Z, Gergely J, Collins JH: Characterisation of zero- component of troponin by phosphorylase kinase. J BioI Chern 247:
length cross links between rabbit skeletal muscle troponin C and 5272-5274, 1972
troponin I: Evidence for a direct interaction between the inhibitory 132. Pratje E, Heilmeyer LMG: Phosphorylation ofrabbit muscle troponin
region of troponin I and the NH 2-terminal, regulatory domain of and actin by 3'-5'cAMP dependent kinase FEBS Lett 27: 89-73,1972
troponin C. Biochemistry 29: 299-304, 1990 133. Perry SV, Cole HA: Phosphorylation of the '37000 component' of
115. Leszyk J, Grabarek Z, Gergeiy J, Collins J: Characterisation of zero the troponin complex (troponin T). Biochem J 131: 425-428,1973
cross links between rabbit skeletal muscle troponin C and troponin I: 134. Perry SV, Cole HA: Phosphorylation of troponin and the effects of
Evidence for a direct interaction between the inhibitory region of interactions between the components. Biochem J 141: 733-744, 1974
troponin I and the NH 2-terminal regulatory domain of troponin C. 135. Moir AJG, Wilkinson JM, Perry SV: The phosphorylation sites of
Biochemistry 29: 299-304,1990 troponin I from white skeletal muscle of the rabbit. FEBS Lett 42:
116. Wang Z, Sarkar S, Gergely J, Tao T: Ca 2+ dependent interactions 253-256,1974
between the C helix oftroponin C and troponin I. Photo crosslinking 136. HuangTS, Bylund DB, Stuil JT, Krebs EJ: The amino acid sequences
and fluorescent studies using a recombinant troponin C. J Bioi Chern of the phosphorylated sites in troponin I from skeletal muscle. FEBS
265:4953-4957,1990 Lett 42: 249-252, 1974
 31

137. Ribulow H, Barany K, Steinschneider A, Barany M: Lack of 154a. Ardell P, Dorka P, Jaquet K, Heilmeyer LMG Jr, Kortke H, Korer R,
phosphate incorporation in troponin I in living frog muscle. Arch Notohamiprodjo G: Microanalysis and distribution ofcardiac troponin
Biochem Biophys 179: 81-88, 1977 I phosphate species in heart areas. Bioi Chern 379: 341-347, 1998
138. Noland Jr TA, Guo X, Raynor RL, Jideama NM, Averyhart-Fullard 155. Mittman K, Jaquet K, Heilmeyer Jr LMG: Ordered phosphorylation
V, Solaro RJ, Kuo JF: Cardiac I mutants. Phosphorylation by protein of a minimal recognition motif for cAMP dependent protein kinase
kinases C and A and regulation of the Ca 2·-stimulated MgATPase of present in cardiac troponin I. FEBS Lett 302: 133-137,1992
reconstituted actomyosin-S-I. J Bioi Chern 270: 25445-25454, 1997 156. Quirk PG, Patchell VB, Gao Y, Levine BA, Perry SV: Sequential
139. Reddy YS, Ballard D, Giri NY, Schwartz A: Phosphorylation of phosphorylation ofadjacent serine residues on the N-terminal region
cardiac native tropomyosin and troponin, inhibitory effect of of cardiac troponin I: Structure activity implications of ordered
actomyosin and possible presence ofendogenous myofibrillar located phosphorylation. FEBS Lett 370: 175-178, 1995
cyclic AMP dependent protein kinase. J Mol Cell Cardiol 5: 461- 157. Malhotra A, Nakouzi A, Bowman J, Buttrick PM: Role of protein
471,1973 kinase A (PKA) and protein kinase C (PKC) phosphorylation of
140. Reddy YS, Schwartz A: Phosphorylation of cardiac contractile proteins troponin I in cardiac muscle. Biophys J 70: A52, 1996
by cyclic AMP dependent protein kinase. Fed Proc 33: 1294, 1974 158. Keane NE, Quirk PG, Gao Y, Patchell VB, Perry SV, Levine BA:
141. Cole HA, Perry SV: The phosphorylation oftroponin I from cardiac The ordered phosphorylation of cardiac troponin I by the cAMP-
muscle. Biochem J 149: 525-533, 1975 dependent protein kinase. Structural consequences and functional
142. England PJ: Correlation between contraction and phosphorylation implication. Eur J Biochem 248: 329-337, 1997
of the inhibitory subunit oftroponin in perfused rat heart. FEBS Lett 158a. Gasmi-Seabrook G, Howarth JW, Finley N, Abbott MB, Britto RM,
50: 57--60, 1975 Rosevear PR: Structure ofthe C-terminal domain ofcardiac troponin
143. Solaro J, Moir AJG, Perry SV: Phosphorylation oftroponin I and the C complexed with the N-terminal of cardiac troponin I. Biophys J
ionotropic effect of adrenaline in the perfused rabbit heart. Nature 74: A299, 1998
262:615--617,1976 159. Kleerekoper 0, Howarth J, Guo X, Solaro Rj, Rosevear PR: Cardiac
144. Rubio R Bailey C, Villar-Palasi C: Effects of cyclic AMP dependent troponin I induced conformational changes in cardiac troponin C as
protein kinase on cardiac actomyosin: increase in Ca 2• sensitivity monitored by NMR using site-directed spin and isotope labelling.
and possible phosphorylation oftroponin I. J Cyclic Nucleotide Res Biochemistry 34: 13343-13352, 1995
1: 143-150,1975 160. Krudy GA, Kleerekoper 0, Guo X, Howarth JW, Solaro RJ,
145. Ray KP, England PJ: Phosphorylation of the inhibitory subunit of Rosevear PR: NMR studies delineating space relationships within
troponin and its effect on the calcium dependence of cardiac myofibril the cardiac troponin I troponin C complex. J Bioi Chern 269: 23731-
adenosine triphosphatase. FEBS Lett 70: 11-16, 1976 23735, 1994
146. Ding X-L, Sonnenblick EH, Gulat J: Pathways of communication 161. Howarth JW, Krudy GA, Putkey JA, Rosevear PR: An NMR and
between cardiac TN-I and troponin C during acidosis and TN-I spin label study of the effects of binding calcium and troponin I
phosphorylation by protein kinase. Biophys J 72: A59, 1997 inhibitory peptide to cardiac troponin C. Protein Sci 4: 671--680,
147. Kogler H, Plathow C, AI-Hillawi E, Trayer IP, Ruegg JC: Effects of 1995
reconstitution with wild type and mutant cardiac troponin I in skinned 162. Dong W-J, Chandra M, Xing J, Solaro JD, Cheung HC: Conformation
soleus fibres. Eur J Physiol433: SS P263, 1997 of the N-terminal segment of a monocysteine mutant oftroponin 1of
148. Johns EC, Simmet SJ, Muffigan IP, Ashley CC: Troponin phos- cardiac muscle: Biochemistry 36: 6745--6753, 1997a
phorylation does not increase the rate of relaxation following laser 163. Dong W-J, Chandra M, Xing J, She M, Solaro JD, Cheung HC:
flash photolysis of diazo-2 in guinea pig skinned trabeculae. Eur J Phosphorylation induced distance change in a cardiac muscle troponin
Physiol 433: 842-844, 1997 I mutant. Biochemistry 36: 6754--6761, 1997b
148a. Li L, Chu X, Kranias EG, Bers DM: Phosphorylation of phos- 164. Liao R, Wang CK, Cheung HC: Effect of phosphorylation of cardiac
pholamban is more crucial than troponin I phosphorylation in PKA troponin I on the fluorescence properties of its single tryptophan as
mediated acceleration of relaxation in cardiac myocytes. Biophys J determined by pico second spectroscopy. SPIE Proceedings 204: 699-
74: A248, 1998 705, 1990
149. Moir AJG, Perry SV: The sites of phosphorylation of rabbit cardiac 165. AI-Hillawi E, Bhandari DG, Trayer HR, Trayer IP: The effect of
troponin I by adenosine 3'5'-cyclic monophosphate dependent protein phosphorylation of cardiac troponin I on its interaction with actin
kinase. Effect of interaction with troponin C. Biochem J 167: 333- and cardiac troponin C. Eur J Biochem 228: 962-970, 1995
343,1977 166. Dong M, Xing J, She M, Chandra M, Solaro RJ, Cheung HC:
150. Mittman K, Jaquet K, Heilmeyer Jr LMG: A common motif of two Phosphorylation increased distance in cardiac troponin I. Biophys J
adjacent phosphoserines in bovine, rabbit and human cardiac troponin 70: A38, 1996
I. FEBS Lett 273: 41-45,1990 167. Chandra M, Pan B-S, Solaro JS: Effect of PKA dependent phos-
151. Moir AJG, Solaro J, Perry SV: The site of phosphorylation oftroponin phorylation of cardiac troponin I with Nand C-terminal domains of
I in the perfused heart: the effect of adrenaline. Biochem J 185: 505- cardiac troponin C. Biophys J 70: A53, 1996
513,1980 168. Robertson SP, Johnson JD, Hoiroyde MJ, Kranias EG, Potter JD:
152. Zhang R, Zhao J, Potter JD: Phosphorylation of both serine residues The effect of troponin I phosphorylation on the Ca 2·-binding
in cardiac troponin I is required to decrease the Ca 2• affinity of cardiac properties of the Ca 2• regulatory site of bovine troponin. J Bioi Chern
troponin C. J Bioi Chern 270: 30773-30780, 1995 257: 260-263, 1982
153. Jaquet K, Thieleczek R, Heilmeyer Jr LMG: Pattern formation on 168a. Dong W-J, Chandra M, Xing J, Solaro RJ: Phosphorylation induced
cardiac troponin I by consecutive phosphorylation and dephos- conformational change in single cysteine mutants of cardiac troponin
phorylation. Eur J Biochem 231: 486-490, 1995 I. Biophys J 74: A51, 1998
154. Reiffert SU, Jaquet K, Heilmeyer Jr LMG. Ritchie M, Geeves MA: 169. Perry SV: Phosphorylation and the modulation of the calcium
Bisphosphorylation of cardiac troponin I modulates the Ca 2• regulation of contraction. In: K. Maruyama, Y. Nonomura, K.
dependent binding of myosin subfragment S I to reconstitued thin Kohama (eds). Calcium as Cell Signal. Igaku-Shion, Tokyo, 1995,
filaments. FEBS Lett 384: 43-47,1996 pp. 49-59
32

170. Katoh N, Wise BC, Kuo JF: Phosphorylation of cardiac troponin 175. Moir AJG, Perry SV: Phosphorylation of rabbit cardiac-muscle
inhibitory subunit (troponin I) and tropomyosin binding unit (troponin troponin I by phosphorylase kinase. The effect of adrenaline. Biochem
T) by cardiac lipid-sensitive Ca2+ dependent protein kinase. Biochem J 191: 547-554, 1980
J 209: 189-195,1983 176. Venema RC, Kuo JF: Protein kinase C mediated phosphorylation of
171. Mazzei GF, Kuo JF: Phosphorylation of skeletal muscle troponin I troponin I and C protein in isolated cardiac cells is associated with
and troponin T by phospholipid-sensitve Ca2+ dependent protein inhibition of myofibrillar actomyosin ATPase. J Bioi Chern 268:
kinase. Biochem J 218: 361-369,198 2705-2711, 1993
172. Noland Jr TA, Raynor RL, Kuo JF: Identification of sites phos- 177. Jideama NM, NolandTA, Raynor RL, Blobe GC, Fabbro D, Kazanietz
phorylated in bovine cardiac troponin I and troponin T by protein kinase MG, Blumberg PM, Hannun YA, Kuo JF: Phosphorylation speci-
C and comparative substrate activity of synthetic peptides containing ficities of protein kinase isozymes for bovine cardiac troponin I and
phosphorylation sites. J BioI Chern 264: 20778-20785, 1989 troponin T and sites within these proteins and regulation of myo-
173. Noland TA, Kuo JF: Protein kinase C phosphorylation of cardiac filament properties. J Bioi Chern 271: 23277-23283, 1996
troponin I or troponin T inhibits Ca'+-stimulated actomyosin 177a. Takeishi Y, Chu G, Kirkpatrick DM, Li Z, Wakasaki H, Kranias EG,
MgATPase activity. J Bioi Chern 266: 4974-4978, 1991 King GL, Walsh RA: In vitro phosphorylation of cardiac troponin I by
174. Noland JrTA, Guo X, Raynor RL, Jideama, NM,Averyhardt-Fullard protein kinase C132 decreases cardiomyocyte calcium responsiveness
V, Solaro J, Kuo JF: Cardiac troponin I mutants. J Bioi Chern 270: and contractility in transgenic mouse hearts. J Clin Invest 102: 72-78,
25445-25454, 1995 1998
Molecular and Cellular Biochemistry 190: 33-38, 1999.
© 1999 Kluwer Academic Publishers.

Calcium ion regulation of muscle contraction: The

regulatory role of troponin T

Iwao Ohtsuki
Department ofPharmacology, Faculty ofMedicine, Kyushu University, Fukuoka, Japan

Abstract
The relaxation and contraction in vertebrate skeletal muscle are regulated by Ca 2+through troponin and tropomyosin, which
are located in the thin filament. Troponin is composed of three components, troponins C, I and T. In this review article, the
Ca2+-regulatory mechanism is discussed with particular reference to the regulatory properties oftroponin T. (Mol Cell Biochem
190: 33-38, 1999)

Key words: calcium ion, troponin, troponin T, tropomyosin, skeletal muscle

Introduction [13]. The essential roles of the troponin components in the

Ca 2+-regulation are as follows. Troponin I alone inhibits the
It was in the early summer of 1962 that Professor Setsuro contractile interaction of actin-tropomyosin with myosin and
Ebashi first isolated a protein factor from minced muscle that this inhibition by troponin I is then removed by the addition
sensitized actomyosin to Ca2+[1-3]. This finding is the real oftroponin C, regardless of the Ca 2+-concentration. The Ca2+-
dawn of the molecular biology ofliving muscle contraction. sensitivity is conferred on the contractile interaction only in
The Ca 2+-sensitizing factor called native tropomyosin was the concomitant presence of troponins I, C and T. In this
thereafter separated into two proteins. One was tropomyosin, respect, troponin I and troponin C are the inhibitory and
a fibrous protein which had been found, while the other was de-inhibitory (activating) components on the contraction,
a new globular protein, troponin [4, 5]. respectively, whereas troponin T is the regulatory component
Troponin is the sole Ca 2+-receptive protein in the con- of the troponin complex.
traction of vertebrate striated muscles [5] and is regularly In this article, the Ca2+-regulatory mechanisms of skeletal
distributed along the entire length of the thin filament [6, 7]. muscle contraction are discussed with special reference to the
The contractile interaction of myosin and actin is regulated regulatory role of troponin T.
by troponin and tropomyosin through Ca 2+. In the absence of
Ca2+, troponin, in collaboration with tropomyosin, depresses
the contractile interaction of actin with myosin. The depres- Troponin T subfragments
sion by troponin-tropomyosin is removed through the action
of Ca2+on troponin and the contraction is then activated [8] Troponin T is a tropomyosin-binding component and its
(Fig. la). presence is essential for the Ca2+-regulation of contraction
Troponin consists of three different components, namely; through troponin-tropomyosin, even though troponin T itself
Ca2+-binding component (troponin C), inhibitory component does not significantly affect the contractile interaction of
(troponin I) and tropomyosin-binding component (troponin myosin-actin-tropomyosin or the inhibitory action oftroponin
T) [9-12]. The presence of all three components oftroponin I [13]. Troponin T binds to tropomyosin, troponin I and
is required for Ca2+-regulation of the contractile interaction troponin C. The interaction with tropomyosin is highly
of myosin-actin-tropomyosin, which itself stays in the sensitive to ionic strength, while the interaction with troponins
activated state irrespective ofthe Ca2+-concentration (Fig. 1b) C and I is relatively stable in relation to the ionic strength.

Address for offprints: 1. Ohstuki, Department of Pharmacology, Faculty of Medicine, Kyushu University, Fukoka 812 82, Japan
34

a) 101 residues (Fig. 2) . Both subfragments are highly soluble,

although troponin T is insoluble at low ionic strengths. The
Contraction ~------------:;---""""'== properties of these subfragments can thus be examined under
physiological conditions (13,18,19].
Although both troponin T subfragments bind to tropo-
myosin, the binding stability oftroponin T 1 is much higher
than that oftroponin T z. This indicates that troponin T binds
to tropomyosin mainly through the troponin T I region. The
binding region within the troponin T, is in the a-helical region
of residues 71-150 (CB2 fragment) which can thus form the
stable triple stranded binding structure with the coiled-coil
Relaxation 1..=======-- _ tropomyosin [20, 21] Regarding the binding of troponin Tz
to tropomyosin, studies using troponin T z fragments indicate
[Ca 2 +] that the main binding region to tropomyosin is located at the
small C-terminal region of 17 residues. This binding to
b) tropomyosin though the troponin T z region is necessary for
TN·I+TN·C the Caz+-regulating action of troponin T.
Contraction ~--------''....'..!.~'....'..!.~---::--'-'''==
Troponin T z also binds to troponin I and troponin C, while
troponin T, shows no affinity to these troponin components.
It has been demonstrated by the use oftroponin T z fragments
that the small region in troponin T z including residues 222-
227, which are highly conserved among several kinds of
troponin T [22], is critical for the binding to troponin I,
although the broad region oftroponin T z also weakly interacts
with troponin I. Troponin C interacts with the broad region
oftroponin T z, in which the N-terminal side region (troponin
Relaxation ~~=~===~T~N~'I~====­ Tz~ III) is mainly related to the Ca z+- dependent interaction
[Ca 2 +] and the remaining C-terminal region is involved in the
Caz+-independent interaction. It is, however, uncertain as to
whether or not the troponin C-troponin T z interaction is
Fig. I. Ca'+-regulatory mechanism of troponin and tropomyosin. (A)
actually present and operates in the ternary troponin complex.
Ca'+-regulation of the contractile interaction of myosin-actin in the absence
and presence oftroponin (TN) and tropomyosin (TM). (8) Ca2+-regulation A water soluble fragment of troponin T called the 26 K
of the contractile interaction of myosin-actin-tropomyosin in the absence fragment is produced by endogenous protease in the muscular
and presence oftroponin components. Cited from Ohtsuki et al. [13]. TN'C, tissue. This fragment is devoid of the N-terminal45 residues
I and T - troponins C, I and T. of troponin T and has the same properties as the intact
troponin T with regard to both the Caz+-regulating action and
tropomyosin-binding [23].
Troponin T from the rabbit skeletal muscle is a single peptide
of 259 amino acid residues, which contains 130 charged
residues, and its N-terminal region is highly rich in acidic Ca2+-regulating action of troponin T
residues, while the C-terminal region is rich in basic residues
[14]. The most significant finding obtained from the studies on the
Troponin T is split into two subfragments, troponin T, and chymotryptic subfragments of troponin T is that the Caz+-
troponin T z, by mild treatment with chymotrypsin [15]1. regulating action of troponin T is almost fully retained in
Troponin T 1 is an acidic fragment of the N-terminal 158 troponin T z. But the Caz+-regulating activity of the troponin
residues and troponin T z is a basic fragment ofthe C-terminal T z subfragment disappears by the removal of the C-terminal
17 residues with a concomitant decrease in the affinity to
'Two domain structure of troponin T was first indicated by the finding that tropomyosin, while the affinity to troponin C and troponin I
antitroponin T antibody formed a pair of striations in each troponin period remains mostly unchanged [24, 25]. Tropomyosin-binding
along the thin filament bundle [16, 17]. Study of chymotryptic troponin T through the C-terminal region oftroponin T z therefore plays
subfragments actually demonstrated that the striation by antitroponin T on
the Z-band side was formed by the N-terminal(T,) region of troponin T
a critical role in the Ca2+-regulating action oftroponin T (Fig.
and the other striation on the filament-top side by the C-terminal (T,) region 3). The strong binding of the range of residues 71-150 of
[15]. troponin T I to tropomyosin, at the same time, stably fixes the
 35

Troponin C binding
.- .
Tropomyosin
Tropomyosin binding
.... ~!!!~!!".~..
Troponin I binding

AS Ser Ser

(~_e_r ..... r-- _'_~6_" "_"' 'l' "i

-"'' Il ',.'_4_9 -,--,...,.,,,r----.,i'T,....- - _.......,...,--I1

T, TZ(lI)

Tza'
TzQI

222
I

Fig 2. Distribution of the interacting properties of rabbit skeletal troponin T along its amino acid sequence. Cited from Ohtsuki et al. [13].

position of the troponin CI·T complex to the filamentous specific isoforms of t.oponin T (troponin T 2f) as well as
tropomyosin-actin and hence supports the regulatory inter- tropomyosin (aa-tropomyosin) has been indicated to correlate
action oftroponin T2 with tropomyosin and troponin CI. with the higher cooperativity ofthe Ca2+-activated contraction
The N-terminal region oftroponin T, which has no affinity of the skinned fibers [29]. Based on this finding, it has been
to tropomyosin, is not related to the Ca2+- regulating activity proposed that the N-terminal region oftroponin T affects the
of troponin T. In fact, the 26 K fragment devoid of the cooperative property of the Ca 2+-activation of contraction
N-terminal residues from troponin T shows exactly the same through the interaction with the C-terminal region of tropo-
Ca2+-activation profiles of contraction as those seen in intact myosin including the end-to-end connecting portion, for the
troponin T [23, 26]. This conclusion has also been further end-to-end interaction oftropomyosin is known to be involved
confirmed under more physiological conditions as follows. in the cooperativity of the Ca 2+-activation [30]. Detailed
There are several isoforms oftroponin T in the muscle fibers, studies using the replacement technique specific for troponin
in which the variety is restricted in the N-terminal region [27, within the myofibrillar lattice in situ [31,32], however, have
28]. Regarding the significance of the variable N-terminal revealed that the Ca2+-activation profiles, in terms ofboth the
region of troponin T isoforms, the relative content of the Ca 2+-sensitivity and its cooperativity, do not change even after

Fig. 3. Arrangement oftroponin T, and troponin T, in the troponin-tropomyosin complex. Cited from Nagano and Ohtsuki [39]. Two domains oftroponin
T (TI' T,) bind to tropomyosin (TM) antiparallelly in relation to the amino acid sequence; the N-terminal troponin T, region occupies the portion oftroponin
T at the C-terminal side of tropomyosin and the C-terminal troponin T, region at the N-terminal side of tropomyosin [7, 15, 17]. The shape and size of this
model approximately coincide with those observed in the fresh troponin preparations [40].
36

the exchange of the endogenous troponin T 2f with another

isofonn (troponin T 1f), and vice versa, in the myofibrils [33]
and also after the exchange ofthe whole endogenous troponin
T with its fragments devoid of the N-tenninal residues in the
myofibrils [32] and the skinned fibers [34]. The N-tenninal
region oftroponin T is therefore by no means involved in the
Ca2+-activation profiles ofcontraction including cooperativity.
The troponin T , subfragment itself shows the strong
inhibitory action on actomyosin ATPase activity in the
presence of tropomyosin-troponin I·C regardless of Ca 2+- ACTIN ACTIN
concentrations [25]. But this depression by troponin T I is not (dlbited)

affected by the addition of troponin T 2 and would thus be

caused through the interaction with tropomyosin, which does
not exist in the thin filament. Actually troponin T I sub-
fragment interacts with the C-tenninal region oftropomyosin,
-B
Ca2+ (+)

If)
which shows no affinity to the troponin complex [13, 35].
This strongly suggests that the C-tenninal region of tropo-
myosin is out of the range for the troponin T-binding under
physiological conditions.

__- ....\ TM
Ca 2+-regulation in the thin filament
In the absence of Ca2+, actin in the thin filament is inhibited ACTIN ACTIN
from interacting with myosin for contraction. This relaxed
state is caused by the inhibitory action of troponin I on
actin-tropomyosin and then is released for contraction on
Ca 2+-binding to troponin C. The inhibitory activity is localized Fig. 4. Ca2'-reguJation of the interaction of the troponin components and
actin-tropomyosin. (A) The thin filament (actin-tropomyosin-troponin C·IT
in the relatively small region (inhibitory region; residues 96-
filament). Actin-tropomyosin is inhibited by troponin I in the absence of
116) oftroponin I of 178 residues from rabbit skeletal muscle Ca 2' and the inhibition by troponin I is removed in the presence of Ca 2',
[36]. This inhibitory region fragment also binds to troponin (B) The actin-tropomyosin-troponin C·I filament (without troponin T). Actin
C and its affinity is potentiated by the Ca 2+-binding to is not inhibited by troponin I in the absence and presence of Ca 2', though
troponin C. The inhibitory interaction of the inhibitory region troponin I binds to actin-tropomyosin only in the absence of Ca 2'.
Abbreviations: TM - tropomyosin; C, I and T - troponin C, I and T.
in troponin I with actin-tropomyosin predominates in the
absence of Ca 2+, while, in the presence of Ca 2+, the affinity
of the inhibitory region of troponin I to troponin C prevails
over its affinity to actin-tropomyosin and hence actin mole- by weakening the interaction of troponin I with troponin C
cules are released from the inhibited state for the contractile in the absence ofCa2+ [13].
interaction with myosin (Fig. 4).
The Ca2+-dependent competition oftroponin C and actin-
tropomyosin to the inhibitory region of troponin I thus is Conclusion
considered to be the essential mechanism of the Ca 2+-
regulatory pathway in the thin filament. These competitive The essential features ofCa2+-regulation in the thin filament
processes, however, do not operate when troponin T is are the depression by the inhibitory action troponin I and the
removed from the thin filament. The binary complex of Ca2+-dependent removal of this depression by the activating
troponin C·I (without troponin T) does not confer Ca 2+- (de-inhibitory) action oftroponin C. However, the contraction
sensitivity on actin-tropomyosin, and the contractile inter- is activated regardless of the Ca 2+-concentration in the
action is activated (de-inhibited) even in the absence ofCa 2+ absence of troponin T (Fig. Ib) . Troponin T is the in-
in this situation, though the binary troponin Col complex binds dispensable component for the inhibition of contraction by
to actin-tropomyosin in the absence ofCa 2+and is dissociated troponin complex-tropomyosin in the absence ofCa 2+. The
from actin-tropomyosin in the presence ofCa2+[36, 37] (Fig. role of troponin T is thus thought to integrate the two
4) . Troponin T thus makes the inhibitory region oftroponin Ca2+-independent actions oftroponin I and C into the Ca2+-
I fully available for the interaction with actin-tropomyosin, dependent function.
 37

Studies on chymotryptic troponin T subfragments have 3. Ebashi S: Calcium ion and muscle contraction. In: K. Maruyama et.
clarified that the Ca2+-regulating activity is localized in the ai. (eds) Calcium as cell signal. Igaku-shoin Tokyo-New York, 1995,
pp I-II
troponin T2 region, which contains the C-terminus. Although 4. Ebashi S, Kodama A: A new protein factor promoting aggregation of
the involvement of the N-terminal region has also been tropomyosin. J Biochem 58: 107-108, 1963
postulated in relation to the cooperative property of the Ca2+- 5. Ebashi S, Kodama A, Ebashi F: Troponin I. Preparation and physio-
activated contraction, this possibility was later ruled out logical function. J Biochem 64: 465-477, 1968
based on more detailed examinations under physiological 6. Ohtsuki I, Masaki T, Nonomura Y, Ebashi S: Periodic distribution of
troponin along the thin filament. J Biochem 61: 817-819, 1967
conditions. The tight coupling ofthe Ca 2+-regulating action 7. Ohtsuki I: Localization oftroponin in thin filament and tropomyosin
of troponin T with its binding to tropomyosin through the paracrystal. J Biochem 75: 753-765, 1974
troponin T2 region also suggests that the steric arrangement 8. Ebashi S, Endo M, Ohtsuki I: Control of muscle contraction. Q Rev
of troponin components and tropomyosin in the thin fila- Biophys 2: 351-384, 1969
ment is delicately regulated through troponin T, in which 9. Hartshorne OJ, Mueller H: Fractionation oftroponin into two distinct
proteins. Biochem Biophys Res Comm 31: 647--653, 1968
the C-terminal T2 region is situated at the pivotal position 10. Schaub MC, Perry SV: The relaxing protein system of striated muscle:
and thus plays a crucial role in the Ca 2+-regulation of Resolution of the troponin complex into inhibitory and calcium-ion
contraction. sensitizing factors and their relationship to tropomyosin. Biochem J
Although the contraction of most striated muscles is 115: 993-1004,1969
regulated by troponin-tropomyosin through Ca 2+, the Ca2+- II. Greaser ML, Gergely J: Reconstitution oftroponin activity from three
protein components. J BioI Chern 246: 4226--4233, 1971
regulation of contraction in some striated muscles of in- 12. Ebashi S: Separation oftroponin into its three components. J Biochem
vertebrates such as molluscs is myosin-linked, where the 72: 787-790, 1972
action of Ca 2+ on the essential light chain (Ca 2+-binding 13. Ohtsuki I, Maruyama K, Ebashi S: Regulatory and cytoskeletal proteins
subunits) is mediated to the active site of heavy chain within of vertebrate skeletal muscle. Adv Prot Chern 38: 1--67,1986
the myosin molecule [37]. The contraction is activated 14. Pearlstone JR, Carpenter MR, Johnson P, Smillie LB: Amino acid
sequence of tropomyosin-binding component of rabbit skeletal muscle
regardless of Ca2+concentration when the regulatory light troponin. Proc Natl Acad Sci 73: 1902-1906,1976
chain (RLC) is removed from myosin, indicating that RLC 15. Ohtsuki I: Molecular arrangement oftroponin T in the thin filament. J
is the inhibitory subunit. The myosin-linked Ca 2+-regulatory Biochem 86: 491-497,1979
system, however, does not contain a subunit, of which 16. Ebashi S, Ohtsuki I, Mihashi K: Regulatory proteins of muscle with
function corresponds to that of troponin T. No protein special reference to troponin. Cold Spring Harbor Symp Quant Bioi
37: 215-224, 1973
factors with high affinity to tropomyosin are also apparently 17. Ohtsuki I: Distribution of troponin components in the thin filament
involved in the Ca 2+-regulation of smooth muscle con- studied by immunoelectron microscopy. J Biochem 77: 633--639, 1975
traction. Troponin T is the truely regulatory component 18. Tanokura M, Tawada Y, Ohtsuki I: Chymotryptic subfragments of
specific for the troponin-linked regulation. The detailed troponin T from rabbit skeletal muscle. I. Determination of the primary
examination of the regulatory properties oftroponin Twill structure. J Biochem 91: 1257-1265,1982
19. Tanokura M, Tawada Y, OnoA, Ohtsuki I: Chymotryptic subfragments
certainly lead to the clarification of the characteristic oftroponin T from rabbit skeletal muscle. Interaction with tropomyosin,
mechanisms of the Ca 2+-regulation by troponin-tropomyosin troponin I and troponin C. J Biochem 93: 331-337, 1983
under physiological conditions. 20. Jackson P, Amphlett GW, Perry SV: The primary structure oftroponin
T and the interaction with tropomyosin. Biochem J 151: 85-97, 1975
21. Nagano K, Miyamoto S, Matsumura M, Ohtsuki I: Possible formation
of a triple-stranded coiled-coil region in tropomyosin-troponin T
Acknowledgement binding complex. J Mol Bioi 141: 217-222, 1981
22. Endo T, Matsumoto K, Hama T, Ohtsuka Y, Katura G, Obinata T:
It is my great pleasure to dedicate this article to Professor S. Distinct troponin T genes are expressed in embryonic larval tail striated
Ebashi, who discovered troponin and inaugurated a new era muscle and adult body wall smooth muscle of ascidian. J Bioi Chern
of the molecular biology of the Ca 2+-regulation of muscle 271: 27855-27862,1996
23. Ohtsuki I, Shiraishi F, Suenaga N, Miyata T, Tanokura M: A 26K
contraction. I also express my sincere gratitude to Professor fragment of troponin T from rabbit skeletal muscle. J Biochem 95:
S. Ebashi for his thoughtful advice and encouragement 1337-1342,1984
throughout my scientific life. 24. Ohtsuki I, Yamamoto K, Hashimoto K: Effect of two C-terminal side
chymotryptic troponin T subfragments on the Ca 2+-sensitivity of
superprecipitation and ATPase activities of actomyosin. J Biochem
90: 259-261, 1981
References 25. OnoyamaY, Ohtsuki I: Effect of chymotryptic troponin T subfragments
on the calcium ion-sensitivity of ATPase and superprecipitation of
I. Ebashi S: Third component participating in the superprccipitation of actomyosin. J Biochem 100: 517-519, 1986
'natural actomyosin'. Nature 200: 1010,1963 26. Pan B-S, Gordon AM, Potter JD: Deletion of the first 45 NH 2 -terminal
2. Ebashi S, Ebashi F: A new protein component participating in the residues of rabbit skeletal troponin T strengthens binding of troponin
superprecipitation of myosin B. J Biochem 55: 604-613, 1964 to immobilized tropomyosin. J Bioi Chern 266: 12432-12438, 1991
38

27. Wilkinson JM, Moir AJ, Waterfield MD: The expression of multiple 95-100, 1996
forms of troponin T in chicken fast skeletal muscle may result from 34. Morimoto S, Yanaga F: Troponin T and its fragment devoid of the
differential splicing ofa single gene. Eur J Biochem 143: 47-56, 1984 N-terminal 51 residues from chicken skeletal muscle show the same
28. Briggs MM, Lin J J-C, Schachat FH: The extent of amino-terminal Ca'+-regulating action in skinned fiber contraction. In preparation.
heterogeneity in rabbit fast skeletal muscle troponin T. J Muscle Res 35. Suenaga N: Effect oftroponin and Troponin T , on the iodination of
Cell Mot 8: 1-12,1987 tyrosyl residues of a-tropomyosin. Fukuoka Acta Medica 79: 493-
29. Schachat FH, Diamond M, Brandt PW: Effect of different troponin 496, 1988
T-tropomyosin combinations thin filament activation. J Mol Bioi 198: 36. Syska H, Wilkinson JM, Grand RJA, Perry SV: The relationship
551-554,1987 between biological activity and primary structure oftroponin I from
30. Tawada Y, Ohara H, Ooi T, Tawada K: Non polymerizable tropomyosin white skeletal muscle of rabbit. Biochem J 153: 375-387,1976
and control of the superprecipitation of actomyosin. J Biochem 78: 37. Potter JD, Gergely J: Troponin, tropomyosin, and actin interactions in
65-72, 1975 the Ca'+-regulation of muscle contraction. Biochemistry 13: 2697-
31. Hatakenaka M, Ohtsuki 1: Effect of removal and reconstitution of tro- 2703, 1974
ponins C and I on the Ca2+-activated tension development of single 38. Szent-Gyorgyi AG, Chantler PD: Control of contraction by calcium
glycerinated rabbit skeletal muscle fibers. Eur J Biochem 205: 985- binding to myosin. In: A.G. Engel, C. Franzini-Armstrong (eds)
993,1992 Myology. McGraw-Hill, New York, 1994, pp 506-528
32. Shiraishi F, Kambara M, Ohtsuki I: Replacement of troponin com- 39. Nagano K, Ohtsuki 1: Prediction of approximate quaternary structure
ponents in myofibrils. J Biochem III: 61-65, 1992 oftroponin complex. Proc Japan Acad 58: 73-77, 1982
33. Hatakenaka M: The Ca'+-activation profile of rabbit fast skeletal 40. Ohtsuki I, Onoyama Y, Shiraishi F: Electron microscopic study of
myofibrils is not affected by troponin T isoforms. Biomed Res 17: troponin. J Biochem 103: 913-919, 1988
Molecular and Cellular Biochemistry 190: 39-45, 1999.
© 1999 Kluwer Academic Publishers.

Thermodynamic analyses of calcium binding to

troponin C, calmodulin and parvalbumins by using
microcalorimetry

Kazuhiro Yamada
Department ofPhysiology, Gita Medical University, Gila, Japan

Abstract
Results ofmicrocalorimetric titrations ofcalcium-binding proteins with calcium or magnesium have been reviewed and evaluated.
Results were analyzed mostly in terms of heat capacity changes, which is most closely related to the structural changes of the
molecule on metal binding. Two high-affinity sites of rabbit skeletal troponin C are distinguishable in terms of their affinity to
calcium and associated enthalpy changes. Heat capacity changes on calcium binding to one of the two high-affinity sites is
negative and is in the range ascribed to the ligand binding. In contrast, that to the other of the high-affinity sites is large and
positive, indicating that a substantial area of hydrophobic groups become exposed to the solvent. In frog skeletal troponin C,
the anomalous positive heat capacity changes occur in one ofthe low-affinity calcium-specific sites, so that this may be involved
in the regulation of contraction. Unlike skeletal troponin C, both ofthe two high-affinity sites of cardiac troponin C show negative
heat capacity changes. In calmodulin, heat capacity changes are positive but small, indicating that calcium binding may induce
clustering of the hydrophobic residues on the surface of the molecule. In parvalbumins, heat capacity changes are negative,
characteristic of most ligand binding. (Mol Cell Biochem 190: 39-45, 1999)

Key words: microcalorimetry, calcium, troponin C, calmodulin, parvalbumin

Calcium binding proteins and structure of parvalbumins have been extensively studied
because their primary structures are homologous to those of
Calcium in its ionic form is essential to many cellular troponin C and calmodulin [II].
functions and these are mediated through calcium-binding The crystal structure oftroponin C revealed a dumbbell-
proteins. Among all, Ebashi and Endo [I] have established shaped protein with two globular domains connected by a
that calcium is the final activator of the contractile system, long central helix [12, 13]. Each troponin C domain is made
and troponin is the protein primarily affected by calcium. up of two EF-hand calcium-binding sites. The EF-hand motif
Troponin C, the calcium-binding component of troponin, consists of two approximately perpendicular a-helices
binds four calcium ions [2] and the signal of calcium binding separated by a 12-amino acid loop, which provides the
to troponin C is transmitted to the thin filament of the calcium-binding ligands. The four troponin C calcium-
contractile network of muscle [3]. binding sites are numbered I to IV, according to their order
Calmodulin, also one of the calcium-binding proteins and in the primary structure. Sites I and II in the NH 2-terminal
among the members ofthe EF-hand superfamily like troponin domain bind calcium specifically with a relatively low affinity
C, is a ubiquitous mediator ofthe intracellular calcium signal of approximately 105 M- ' . Sites III and IV in the COOH-
[4-8]. The EF-hand superfamily includes parvalbumins, terminal domain bind calcium with a high affinity (K. =10 7
which are found in fast-twitch muscle fibers especially of M- 1) and bind magnesium ion with a binding constant of =
lower vertebrates [9, 10] and are thought to function as a 10 3 M-1, so that they are occupied by magnesium in relaxed
cyotosolic calcium buffer. The physicochemical properties muscle [2,14,15].

Addressfor offprints: K. Yamada, Department of Physiology, Oita Medical University, Oita 879-55, Japan
40

Nature of the structural changes of Troponin C from rabbit skeletal muscle

proteins on ligand binding related to
When calcium was added successively to metal free troponin
their function C prepared from rabbit skeletal muscle, we noticed at least
three transitions [20]. When added calcium was in the range
Proteins are amphiphilic molecules. They contain amino to saturate one of the high-affinity Ca-Mg site (sites III or IV),
acids with hydrocarbon and other hydrophobic side chains, quite substantial amount of heat was produced (Fig. 1).
and also amino acids with hydrophilic (ionic and uncharged When calcium was added further to saturate one of the
polar) side chains. In general, in water-soluble proteins about other high-affinity Ca-Mg site, heat production became very
25-30% of the amino acid side chains are hydrophobic, and small. The distinction between the two high-affinity sites
45-50% are hydrophilic. In the native conformation of such almost disappeared in 1 mM Mg 2+. The stepwise binding to
a protein, a substantial fraction ofthe hydrophobic side chains the two high-affinity Ca-Mg sites of skeletal troponin C has
are buried in the interior ofthe molecule (hydrophobic effect). never been noticed by any method other than calorimetric
However, possible stable structures are limited because ionic titration. Tsalkova and Privalov [32] have also shown that
amino acid side chains must remain at the surface of the sites III and IV are different in their changes in the structure
globular structure and these cannot be separated from on binding of divalent cations by thermodynamic studies
hydrophobic amino acids in the sequence. This phenomenon, using scanning microcalorimetry with proteolytic fragments
the hydrophobic effect, was described by Tanford [16] as of rabbit skeletal muscle.
'perhaps the most important single factor in the organization Further additions of calcium to saturate the two low-
of the constituent molecules of living matter into complex affinity Ca sites (sites I and II) produced substantial amount
structural entities'. of heat again and these two sites were indistinguishable in
Therefore, the native conformation of a protein molecule the case of rabbit skeletal troponin C (see below and compare
may possess only marginal stability because it is highly with frog skeletal troponin C ).
constrained as described above. Other conformational states, Metal binding to troponin C induces changes in the
in which a much larger fraction of the hydrophobic side structure of the molecule, hence the interactions between its
chains is exposed to the solvent than in the native state, are components; the action of the calcium-sensitive molecular
thus readily accessible [16]. Similar readjustments of surface switch. These structural changes oftroponin C may accom-
groups but in lesser scale are likely to occur whenever a pany changes in the structure of the surrounding water
complex ligand binds to a site on a protein molecule [16]. molecules and may give rise to large heat capacity changes
Refined methods including microcalorimetry for examining [16]. Heat capacity changes (~Cp) can be obtained by the
protein structure may be applicable to detect these structural enthalpy (~H) changes at different temperatures (Kirchhoff's
changes that are related to functions. formula, Eq. 1). Therefore, the above described calorimetric
titrations with calcium were performed at different tempera-
tures of 5, 15, and 25°C (Fig. 2).
Microcalorimetric titration studies
~C =a~H/(n
p
[1]
For some years we have applied microcalorimetric titration
of calcium or magnesium to calcium-binding proteins. Some When calcium titrations were performed in rabbit skeletal
of these studies revealed unexpected results. Here I would troponin C in magnesium-free condition the heat capacity
review results ofthese microcalorimetric studies with special change induced by binding of calcium to one of the high-
attention to the hydrophobic effect first and then discuss these affinity sites was large and negative (-1660--1900 J K-I
in terms of results of other more recent structural studies. mol-I). In contrast, that to the other high-affinity site was also
Yamada et al. [17] have first shown that binding ofcalcium large but positive (970-1440 J K~I mol-I). These char-
produces heat in rabbit skeletal troponin. Calcium binding acteristic changes in heat capacity mostly disappears in 1 mM
produces heat also in rabbit skeletal troponin C [18-20], in Mg 2+. Magnesium itself does not induce characteristic
bullfrog skeletal troponin C [21-23], in bovine cardiac changes caused by calcium. These enthalpy titration results
troponin C [24], and in bullfrog skeletal parvalbumins [25- may indicate that the calcium-loaded and magnesium-loaded
27] and in toad skeletal parvalbumins [28, 29], while binding structures of the COOH-terminal domain oftroponin Care
of calcium to bovine brain calmodulin absorbs heat [30, different. This is in agreement with the notion that magnesium
31]. In these studies, we introduced calcium to apo forms of binding to the Ca-Mg sites causes different structural changes
calcium-binding proteins in a stepwise manner and measured than that of calcium [33]. Calcium binding to low-affinity
the heat produced or absorbed by using a microcalorimeter sites caused moderately negative heat capacity changes; these
specially modified for this purpose (see [20]). are not much affected by magnesium.
 41

Mg - free K-' mol-I which are accounted for by a decrease in exposed

60
0.2 mM Mg surface, possibly due to tightening of the structure The
heat capacity change associated with the calcium binding
50
to one of the two high-affinity sites oftroponin C described
(5 40 1 mM Mg
above (the first stage seen in the microcalorimetric calcium
E titration) is negative and is in the range ascribed to the
-,
.><: 30 ligand binding In contrast, that to the other of the high-
affinity sites (the second stage) is large and positive,
«i
CIl
J: 20 although less than that seen with unfolding. This may
Mg indicate that during this stage a substantial area of hydro-
10 phobic groups become exposed to the solvent. Sturtevant
[35] defined different causes of changes in heat capacity
0 during reactions of proteins, which may be separately
0 2 3 4 5 6 7 8 9
evaluated. The hydrophobic contribution has been shown
Ca / TnC or Mg / TnC to be much greater than other contributions (vibrational and
conformational; see [20]). Tsalkova and Privalov [32] have
Fig. 1. Total heat produced per mol troponin C in relation to the shown that the site III is the least stable and forms an ordered
concentration ratio of Ca'· of Mg2+ added and troponin C, at pH 7 and 150 structure only in the presence of calcium ions. This result
C. Calcium titrations were also performed in 0.2 and I mM Mg H Adapted might indicate that the site which exhibits an anomalous
from Yamada and Kometani [20). positive heat capacity change on calcium binding is the site
IV. Slupsky and Sykes [36] and Gagne et ai. [37] suggested
from NMR solution structure that when troponin C becomes
Typical values for the heat capacity change (L\C p ) for
fully saturated with calcium a major structural change
reactions of proteins show that unfolding of proteins results
occurs in the NH 2-terminal domain resulting in the opening
in large positive change (6500-13000 J K- I mol-I) , which
of a hydrophobic pocket to present itself to its target protein
are ascribed to the exposure of hydrophobic groups to the
troponin I.
aqueous environment [34]. In contrast, most ligand binding
Troponin C at low concentration remains as single molecules
processes result in negative values for L\C p (-350--1700 J
irrespective of the level of calcium bound [19] whereas
troponin aggregates when the level of bound calcium is
reduced [17]. Slupsky et ai. [38] showed that troponin C
10 undergoes a calcium-induced dimerization with a disso-
Mg - free 1 mM Mg ciation constant for the dimerization of 0.4 mM. The
0 2 concentration of troponin C in the microcalorimetric titra-
2 tions were 30-90 ~M.

~
-10 Farah et ai. [39] have proposed, based on the crystal
'7 structure [12, 13] and other evidence, that the complex
~
'0
E -20 formed between troponin C and troponin I is arranged in an
-;
..>:: 1 antiparallel fashion; the NH 2-terminal regions oftroponin C
'-
:r: -30 interact with the COOH-terminal regions oftroponin I, and
'J the COOH terminal regions of troponin C interact with the
-40 NH 2 terminal regions oftroponin I, especially in the presence
of calcium. The fact that a group of hydrophobic residues
-50 become exposed to the solvent by binding calcium to one of
the high-affinity calcium binding sites (possibly the site IV)
-60 as evidenced by the calorimetric calcium titration studies
5 15 25 5 15 25 supports the above idea because when exposed to the solvent
the hydrophobic residues may easily interact with the
Temperature ( °c ) hydrophobic NH 2-terminal regions of troponin I. Sorenson
et ai. [40] have shown that skeletal troponin C mutants ofsites
Fig. 2. Temperature dependance of enthalpy changes associated with III and IV are unable to remain bound to thin filaments.
calcium binding to Mg-free troponin C and troponin C in the presence of I
mM MgH Numbers represent stages of calcium additions so that I and 2
Szczensna et ai. [41] have shown from mutation studies that
corresponds to the two high-affinity Ca-Mg sites, and 3 and 4 the two both of the high affinity calcium-binding sites are important
low-affinity Ca specific sites. Adapted from Yamada and Kometani [20). for maintaining the structural stability of troponin C in the
42

thin filament. The similar processes have been ascribed to C (see above). The amino acid composition of bullfrog
the interaction between the NH 2-terminal regions oftroponin troponin C is different from that of rabbit troponin C.
C and the COOH-terminal regions oftroponin I, allowing Imaizumi and Tanokura [23] also performed the enthalpy
a structural change in the thin filament that activates titrations at different temperatures so that the heat capacity
contractile events [33]. changes associated with the calcium binding can be deduced.
In the crystal structure [12, 13] the two COOH-terminal As can be expected the results indicated that the site with
Ca-Mg sites of troponin C are occupied by calcium; no anomalous enthalpy characteristics (one of the two low-
troponin C structure with bound Mg has been solved [39]. affinity sites) shows positive heat capacity change, indicating
However, because magnesium is displaced from the Ca-Mg that exposure of hydrophobic groups to the solvent phase
sites on calcium binding, the COOH-terminal region should occurs. Taking into account the fact that these thermodynamic
assume the same structure when it binds calcium in the characteristics are not affected by Mg 2+, the exposure of
presence of magnesium. The fact that the characteristic hydrophobic groups on the surface of the molecule may be
anomalous heat capacity changes are not seen in the presence involved in the regulation of contraction.
of magnesium may indicate that similar structural changes Li et al. [44] have also shown from NMR spectroscopy
associated with the calcium binding occur also on magnesium that the calcium binding to the regulatory N-domain of
binding. This agrees with the fact that troponin C binds skeletal troponin C from chicken muscle occurs in a stepwise
troponin I with a high affinity in the presence of magnesium manner. The avian muscle used in this study had more than
(in the absence of calcium) [33 and references therein]. 90% sequence identity with the frog muscle that we used.
Results of our microcalorimetric titrations [20] might also
indicate that in the presence of magnesium the characteristic
anomalous heat capacity changes occur in the calcium- Cardiac troponin C
specific sites (Fig. 1). This may mean that magnesium binding
alters the structure of troponin C so that the anomalous heat Microcalorimetric titrations have been used to study the
capacity changes become induced on calcium binding to the binding of calcium to cardiac troponin C [24]. Cardiac
calcium-specific sites. This is of great interest because the troponin C exhibited a decrease in enthalpy and an increase
same feature is seen in frog skeletal troponin C in the presence in entropy associated with calcium binding. Enthalpy changes
of magnesium (see below). increased linearly with temperature, indicating that the
Under physiological conditions, the rate of calcium calcium binding causes negative changes in the heat capacity
binding to the high-affinity COOH-terminal domain sites of troponin C. Moreover, only two identical calcium sites
is limited by magnesium-calcium exchange, which is too were detected in the calorimetry, the affinity corresponding
slow to account for the rate of calcium-induced force to the high-affinity sites. If cardiac troponin C has one
development and relaxation in muscle fibers [42]. On the low-affinity site [45, 46], calcium binding to this site could
other hand, the rate of calcium binding to the low-affinity be thermally neutral.
NH 2-terminal domain sites is fast, suggesting that calcium Above results are in agreement with the results ofmutation
binding to sites I and II is the initial event in the activation studies of cardiac troponin C; either site III or IV of the
of the actomyosin interaction. However, there is an indica- COOH-terminal domain is required for tight association of
tion that site III might be involved in the regulation of cardiac troponin C with the troponin complex [47].
contraction because thrombin fragments of troponin C Cardiac troponin C differs from skeletal troponin C in many
containing site III can restore force when reconstituted into respects. As described above, in skeletal troponin C in the
skinned muscle fibers [43]. absence of magnesium, two high-affinity calcium sites differ
from each other in their affinity for calcium, whereas the two
high-affinity sites in cardiac troponin C are identical both in
Troponin C from frog's skeletal muscle the absence and in the presence of magnesium. In skeletal
troponin C calcium binding causes characteristic positive heat
Microcalorimetric titrations of bullfrog skeletal troponin C capacity changes associated with the calcium binding to one
with calcium have been carried out [21-23]. The enthalpy of the high-affinity sites in the COOH-terminal domain, while
titration curve could be divided into three stages. The binding heat capacity changes are negative in both of the two high-
of calcium to two high-affinity sites were indistinguishable affinity sites in cardiac troponin C. These results can be
with each other and were extremely exothermic. Whereas that interpreted to show that in cardiac troponin C the calcium
to two low-affinity sites were distinguishable and occurred binding might cause hydrophobic groups to become clustered
stepwise; one with higher affinity was endothermic and the at the interface with the aqueous environment as is the case for
other moderately exothermic. These characteristic enthalpy calmodulin (see below). This is in agreement with the results
titration curve is different from that ofrabbit skeletal troponin of NMR structural studies on mutated cardiac troponin C,
 43

showing that the calcium binding to the high-affinity sites of most ligand binding processes but unlike other calcium-
results in an ordering ofthe aromatic hydrophobic cluster [48]. binding proteins as described above.

Calmodulin Conclusions
Enthalpy titrations of calmodulin with calcium (or mag- Microcalorimetric enthalpy titration studies of calcium
nesium) in the presence and the absence of magnesium have binding proteins (troponin C, calmodulin and parvalbumins)
been carried out by microcalorimetry in bovine brain [30, were reviewed. From heat capacity changes on calcium
31] and wheat germ calmodulin [49]. Quite unexpectedly, binding to these proteins, derived from enthalpy titrations at
the binding of both calcium and magnesium to calmodulin different temperatures, the following results may be deduced.
are endothermic, and thus is driven solely by the large Exposure of large hydrophobic area to the solvent phase
entropy changes. All the four binding sites are Ca-Mg sites, occurs on calcium binding to one of the high-affinity sites
calcium displacing magnesium at each of the four binding (possibly site III) in rabbit skeletal troponin C. This may
sites. These results are markedly different from the findings enable COOH-terminal domain of troponin C to interact
for troponin C (see above). The enthalpy titrations were with the NH 2 -terminal domain of troponin I, one of. the
performed at different temperatures. When calcium was structural basis for the regulation of muscle contractiOn.
added to metal-free calmodulin in the absence of mag\- However, the situation may be quite different in frog and
nesium, heat capacity changes are negative but much avian skeletal troponin C, in cardiac troponin C, and also
smaller than those of skeletal and cardiac troponin C. Both in calmodulin. In frog skeletal troponin C, exposure of the
of these results were interpreted by that calcium binding hydrophobic groups occur in the low-affinity calcium-
causes hydrophobic groups to become clustered at the specific site on calcium binding, so that this may be related
interface with the aqueous environment. Thus when cal- to the regulation of contraction. Parvalbumins are simply
modulin binds calcium, the nonpolar groups are assembled ligand-binding proteins.
and form a region which is large enough to be able to interact
with calmodulin inhibitors, which mostly contain large
nonpolar regions.
In recent years crystal and solution structures of cal-
Acknowledgements
modulin in complex with its target, a myosin light chain
kinase (MLCK) peptide [50, 51] have been reported. The I express my sincere gratitude to Professor Setsuro Ebashi
results showed that the NH 2- and COOH-domains of ~al­ for his advice and encouragement during the course of these
modulin make contacts with the COOH- and NH 2- termmal microcalorimetric studies on calcium-binding proteins. It is
halves of the MLCK peptide, respectively. Thus the anti- my greatest pleasure to dedicate Professor Ebashi this article.
pararell nature of the interaction is as proposed for the I also thank all my colleagues who have carried out experi-
troponin I-troponin C complex [33]. ments in collaboration with myself: Dr. Kaoru Kometani
In the presence of trifluoperazine (TFP), one of the (Professor, Kyushu Institute of Technology), Dr. Masaru
inhibitors or the targets, at the TFP-calmodulin ratio of more Tanokura (Professor, University ofTokyo), and Dr. Masamoto
than 7, calcium binding to the calmodulin-TFP complex Imaizumi (Department of Ophthalmology, Oita Medical
produces heat [52], while heat is absorbed in the absence of University) . I also thank Dr. Tomoko Nawata for her help in
TFP (see above). preparing the figures.

Parvalbumins References
I. Ebashi S, Endo M: Calcium and muscle contraction. Prog Biophys
The possible structural changes of the two major isotypes of
Mol Bioi 18: 123-183, 1968
parvalbumins from bullfrog skeletal muscle caused by 2. Potter JD, Gergely 1: The calcium and magnesium binding sites on
calcium binding have been analyzed by microcalorimetric troponin and their role in the regulation of myofibrilar ATPase. J BioI
titrations [25, 27]. Titrations of the parvalbumins with Chern 250: 4628--4633, 1975
calcium have been made in both the absence and the presence 3. Ohtsuki I, Maruyama K, Ebashi S: Regulatory and cytoskeletal proteins
of vertebrate skeletal muscle Adv. Protein Chern 38: 1--67,1986
ofmagnesium at pH 7.0 and at 5, 15, and 25°C. The reactions
4. Kakiuchi S, Yamazaki R: Calcium dependent phosphodiesterase
of the parvalbumins with calcium are exothermic in both the activity and its activating factor (PAF) from brain. III. Studies on cyclic
presence and absence of magnesium at every temperature. 3', 5'nucleotide phosphodiesterase. Biochem Biophys Res Commun
The heat capacity changes are always negative, characteristic 41: 1104--1110,1970
44

5. Cheung WY: Cyclic 3',5'- nucleotide phosphodiesterase. Demonstration 27. Tanokura M, Yamada K: Heat capacity and entropy changes of the
of an activator. Biochem Biophys Res Commun 38: 533-538, 1970 two major isotypes of bullfrog (Rana catesbelana) parvalbumins
6. Teo TS, Wang JH: Mechanism of activation of a cyclic adenosine-3', induced by calcium binding. Biochemistry 26: 7668-7674, 1987
5'-monophosphate phosphodiesterase from bovine heart by calcium 28. Tanokura M, Imaizumi M, Yamada K: A calorimetric study of Ca'+
ions. J Bioi Chern 248: 5950--5955, 1973 binding by the parvalbumin of the toad (Bufo): distinguishable binding
7. Cheung WY: Calmodulin plays a pivotal role in cellular regulation. sites in the molecule FEBS Lett 209: 77-82, 1986
Science 207: 19-27,1980 29. Tanokura M, Goto K, Toyomori Y, Yamada K: Preparation and
8. Klee CB, Crouch TH, Richman PG: Calmodulin. Annu Rev Biochem characterization of the major isotype of parvalbumin from skeletal
49:489-515,1980 muscle of the toad ( Bufo bufo japonicus). J Biochem 102: 1133-
9. Pechere J-F, Demaille J, Canopy J-P: Muscular parvalbumins: 1139,1987
Preparative and analytical methods of general applicability. Biochim 30. Tanokura M, Yamada K: A calorimetric study of Ca'+- and Mg'+-
Biophys Acta 236: 391--408,1971 binding by calmodulin. J Biochem 94: 607-609, 1983
10. Cox JA, Winge DR, Wnuk W, Stein EA: Soluble calcium-binding 31. Tanokura M, Yamada K: Heat capacity and entropy changes of
proteins from invertebrate muscle. Fed Proc Fed Am Soc Exp BioI calmodulin induced by calcium binding. J Biochem 95: 643-649, 1984
35: 1363, 1976 32. Tsalkova TN, Privalov PL: Thermodynamic study of domain
II. Dedman JR, Jackson RL, Schreiber WE, Means AR: Sequence organization in troponin C and calmodulin. J Mol BioI 181: 533-
homology of the Ca'+-dependent regulator of cyclic nucleotide 544,1985
phosphodiesterase from rat testis with other Ca'+-binding proteins. J 33. Farah CS, Reinach FC: The troponin complex and regulation of muscle
BioI Chern 253: 343-346, 1978 contraction. FASEB J 9: 755-767, 1995
12. Herzberg 0, James MNG: Refined crystal structure oftroponin C from 34. Edsall JT, Gutfreund H: Biothermodynamics. John Wiley & Sons,
turkey skeletal muscle at 2.0 A resolution. J Mol Bioi 203: 761-779, Chichester, 1983, pp 215-217
1988 35. Sturtevant JM: Heat capacity and entropy changes in processes
13. Satyshur KA, Rao SST, Pysalska D, Drendel W, Greaser M, involving proteins. Proc Natl Acad Sci USA 74: 2236-2240, 1977
Sundaraligam M: Refined structure ofchicken skeletal muscle troponin 36. Siupsky CM, Sykes BD: NMR solution structure of calcium-saturated
C in the two-calcium state at 2-A resolution. J BioI Chern 263: 1628- skeletal muscle troponin C. Biochemistry 34: 15953-15964, 1995
1647,1988 37. Gagne SM, Tsuda S, Li MX, Smillie LB, Sykes BD: Structures of the
14. Zot HG, Potter JD: A structural role forCa'+-Mg2+ sites on troponin C troponin C regulatory domains in the apo and calcium-saturated states.
in the regulation of muscle contraction: preparation and properties of Nature Struct BioI 2: 784--789, 1995
troponin C depreted myofibrils. J BioI Chern 257: 7678-7683, 1982 38. Siupsky CM, Kay CM, Reinach FC, Smillie LB, Sykes BD: Calcium-
15. Leavis PC, Rosenfeld SS, Gergely J, Grabarek Z, Drabikowski W: induced dimerization of troponin C: Mode of interaction and use of
Proteolytic fragments of troponin C: Localization of high and low trifluoroethanol as a denaturant of quaternary structure. Biochemistry
affinity Ca'+ binding sites and interactions with troponin I and troponin 34: 7365-7375, 1995
T. J BioI Chcm 253: 5452-5459, 1978 39. Farah CS, Miyamoto CA, Ramos CHI, da Silva ACR, Quaggio RB,
16. Tanford C: The Hydrophobic Effect. John Wiley & Sons, New York, Fujimori K, Smillie LB, Reinach FC: Structural and regulatory
1973, pp. 120--125 functions of the NH,- and COOH-terminal regions of skeletal muscle
17. Yamada K, Mashima H, Ebashi S: The enthalpy change accompanying troponin I. J BioI Chern 269: 5230--5240, 1994
the binding of calcium to troponin relating to the activation heat 40. Sorenson MM, da Silva ACR, Gouveia CS, Sousa VP, Oshima W,
production of muscle. Proc Japan Acad 52: 252-255, 1976 Ferro JA, Reinach FC: Concerted action of the high affinity calcium
18. Potter JD, Hsu F-J, Pownall HJ: Thermodynamics ofCa'+ binding to binding sites in skeletal muscle troponin C. J Bioi Chern 270: 9770--
troponin-C. J BioI Chern 252: 2452-2454, 1977 9777,1995
19. Yamada K: The enthalpy titration oftroponin C with calcium. Biochim 41. Szczesna D, Guzman G, Miller T, Zhao J, Farokhi K, Ellemberger H,
Biophys Acta 535: 342-347, 1978 Potter JD: The role of the four Ca'+ binding sites oftroponin C in the
20. Yamada K, Kometani K: The changes in heat capacity and entropy of regulation of skeletal muscle contraction. J Bioi Chern 271: 8381-
troponin C induced by calcium binding. J Biochem 92: 1505-1517, 8386, 1996
1982 42. Robertson SP, Johnson JD, Potter JD: The timecourse ofCa'+ exchange
21. Imaizumi M, Tanokura M, Yamada K: A calorimetric study on calcium with calmodulin, troponin, parvalbumin, and myosin in response to
binding by troponin C from bullfrog skeletal muscle. J BioI Chern transient increase in Ca'+. Biophys J 34: 559-569, 1981
262: 7963-7966, 1987 43. Francois J-M, Sheng Z, Szczesna D, Potter JD: The functional role of
22. Imaizumi M, Tanokura M, Yamada K: Calorimetric studies on calcium the domains of troponin-C investigated with thrombin fragments of
and magnesium binding by troponin C from bullfrog skeletal muscle. troponin-C reconstituted into skinned muscle fibers. J BioI Chern 270:
J Biochem 107: 127-132, 1990 19287-19293,1995
23. Imaizumi M, Tanokura M: Heat capacity and entropy changes of 44. Li MX, Gagne SM, Tsuda S, Kay CM, Smillie LB, Sykes BD: Calcium
troponin C from bullfrog skeletal muscle induced by calcium binding. binding to the regulatory N-domain of skeletal muscle troponin C
Eur J Biochem 192: 275-281, 1990 occurs in a stepwise manner. Biochemistry 34: 8330--8340, 1995
24. Kometani K, Yamada K: Enthalpy, entropy and heat capacity changes 45. Van Eerd JP, Takahashi K: The amino acid sequence of bovine cardiac
induced by binding of calcium ions to cardiac troponin C. Biochem troponin-C. Comparison with rabbit skeletal troponin-C. Biochem
Biophys Res Commun 114: 162-167, 1983 Biophys Res Commun 64: 122-127,1975
25. Tanokura M, Yamada K: A calorimetric study of Ca'+ binding to two 46. Dotson DG, Putkey JA: Differential recovery ofCa'+ binding activity
major isotypes of bullfrog parvalbumin FEBS Lett 185: 165-169, 1985 in mutated EF-hands of cardiac troponin C. J BioI Chern 268: 24067-
26. Tanokura M, Aramaki H, Goto K, Hashimoto Toyomori Y, Yamada 24073, 1993
K: Preparation and characterization of two major isotypes of parv- 47. Negele JC, Dotson DG, Liu W, Sweeney HL, Putkey JA: Mutation of
albumins from skeletal muscle of bullfrog (Rana catesbelana). J the high affinity calcium binding sites in cardiac troponin C. J BioI
Biochem 99: 1211-1218, 1986 Chern 267: 825-831, 1992
 45

48. Brito RM, Krudy GA, Negele JC, Putkey JA, Rosevear PR: Calcium by multi-dimensional NMR Science 256: 632--638, 1992
plays distinctive roles in the N- and C-terminal domains of cardiac 51. Meador WE, Means AR, Quiocho FA: Target enzyme recognition by
troponin C. J BioI Chern 268: 20966-20973,1993 calmodulin; 2.4 A structure ofa calmodulin-peptide complex. Science
49. Tanokura M, Yamada K: A calorimetric study ofCa2+ binding by wheat 257: 1251-1255, 1992
germ calmodulin; regulatory steps driven by entropy. J Bioi Chern 52. Tanokura M, Yamada K: Effects of trifluoperazine on calcium binding
268: 7090-7092,1993 by calmodulin; a microcalorimetric study. J Bioi Chern 260: 8680-
50. Ikura M et al.: Solution structure ofa calmodulintarget peptide complex 8682, 1985
Molecular and Cellular Biochemistry 190: 47-54, 1999.
© 1999 Kluwer Academic Publishers.

A strange calmodulin of yeast

Michio Yazawa,l Ken-ichi Nakashima l and Koichi Yagil

'Division of Chemistry, Graduate School ofScience, Hokkaido University, Sapporo; 2Agroscience Research Laboratory,
Hokkai Sankyo Co. Ltd., Kitahiroshima, Japan

Abstract
Calmodulin of Saccharomyces cerevisiae has different Ca2+ binding properties from other calmodulins. We previously
reported that the maximum number of Ca2+ binding was 3 mol/mol and the fourth binding site was defective, which was
different from 4 mol/mol for others. Their macroscopic dissociation constants suggested the cooperative three Ca 2+bindings
rather than a pair of cooperative two Ca2+bindings of ordinary calmodulin. Here we present evidence for yeast calmodulin
showing the intramolecular close interaction between the N-terminal half domain and the C-terminal half domain, while the
two domains of ordinary calmodulin are independent of each other. We will discuss the relationship of the shape and the
shape change caused by the Ca2+binding to the enzyme activation in yeast. The functional feature of calmodulin in yeast
will also be considered, which might be different from the one of vertebrate calmodulin. (Mol Cell Biochem 190: 47-54,
1999)

Key words: calmodulin, yeast calmodulin, Ca2+binding, Ca2+binding protein, Saccharomyces cerevisiae, interdomain interaction

Abbreviations: YCMO - wild type yeast calmodulin expressed in Escherichia coli; YCMLl- a recombinant yeast calmodulin
in which 17 residues from the C terminus were deleted; Y 12 - a recombinant N-terminal half of yeast calmodulin (Ser l-Lys77);
Y34 - a recombinant C-terminal half of yeast calmodulin (Thr70-Lys 146); PDE - cyclic-nucleotide phosphodiesterase; MLCK
- myosin light-chain kinase.

Introduction question. Kakiuchi applied this procedure to investigate the

involvement of Ca2+in calmodulin [1].
An activator protein ofphosphodiesterase was independently In 1977 we found a Ca2+binding subunit of myosin light
discovered by S. Kakiuchi [1] and w.y Cheung [2] in brain. chain kinase from rabbit skeletal muscle. The kinase alone
Now we call this protein calmodulin. Kakiuchi [1] revealed was inactive in the presence ofCa 2+, and it became active in
the Ca2+-dependent activation ofphosphodiesterase and J.H. association with the subunit, which was calmodulin [6]. This
Wang [3] showed that the calmodulin was a Ca2+ binding result led to surprising evidence that the cAMP elimination
protein. in brain and myosin kinase activation in muscles are con-
In 1965, five years before the findings of calmodulin, S. trolled by the same Ca2+-binding protein, calmodulin [7]. At
Ebashi [4] had discovered a Ca2+ binding protein in the that time essential regulatory function of Ca 2+ through
striated muscle cell and designated it troponin. Troponin troponin system had been established already by Ebashi [8],
functions as the regulator of muscle contraction by receiving and the Ca2+regulation ofthe skeletal muscle contraction was
intracellular Ca2+. Ebashi developed a genius method. He found to be pleiotypic through different Ca2+binding proteins
used EGTA to eliminate Ca2+ from the reaction medium troponin and calmodulin [7]. Calmodulin was identified
containing K+, Na+, and Mg2+ by the selective binding of a further in skeletal muscle as the activator of phosphorylase
minute amount of Ca2+[5]. By this method we can decide kinase [9-11] and protein phosphatate 2B [12] which promote
whether Ca2+is indispensable to the particular reaction in breakdown of glycogen to support the muscle contraction.

Address/or offprints: K. Vagi, Agroscience Research Laboratory, Hokkai Sankyo Co. Ltd., Kitahiroshima 061-11, Japan
48

It became clear that calmodulin is a ubiquitous protein in Among many calmodulins, the yeast calmodulin showed
all eucaryotic cells which modulates more than 30 different unusually low similarity (only 60%) to the primary sequence
enzymes [13]. Its primary structure consisting of four Ca 2+ of the vertebrate calmodulin [31, 37]. It can activate very
binding sites of the EF-hand motif [14] is highly conserved poorly the target enzymes of vertebrate origin [30, 33]. Since
among eucaryotes [13, 15]. the calmodulin is essential for living yeast [37, 38], the
Recent structural analyses by NMR and the X-ray crystal- specific structure and function of yeast CaM would be
lography offered a structural basis for the Ca2+-dependent essential for the interaction with targets proteins in yeast cells.
activation process of target proteins by calmodulin [16--18]. Recent studies using chimeric calmodulin revealed a new
Apo CaM, similarly to the Ca2+-saturated form, has the N- information on the function ofthe defective Ca2+binding site
and C-terminal half molecular domain each containing two (site IV) at the C-terminal of yeast calmodulin. The result
Ca 2+binding sites and flickers in an independent manner [19, may give a new approach to the study of function of yeast
20]. In the Ca 2+-saturated form, each domain exposes a calmodulin and also those of recoverin [39] and calcineurin B
hydrophobic core with many Met residues on the molecular [40]. We will further discuss a different way of Ca2+regulation
surface [16], while they are buried in the molecular inside in from the one proposed to the vertebrate calmodulin.
the apo form [21, 22]. In the process of target recognition,
these sites playa critical role for binding to targets (reviewed
in [23]). Earlier biochemical studies such as the limited tryptic Yeast calmodulin binds only 3 Ca 2+
digestion of CaM (24) showed these features of CaM con-
sisting ofthe two independent halfmolecular domain [25, 26] Earlier results of Ca 2+ binding measurements on yeast
each containing a Ca2+-dependent hydrophobic core [27,28]. calmodulin were shown in Fig. 1 [30,31]. Yeast calmodulin
These results were helpful for speculating the activation binds only 3 mol of Ca2+while ordinary calmodulin binds 4
mechanism based on the 3D-structural analyses. mol. Careful inspection of its primary structure suspected a
Calmodulin had been isolated from eucaryotes, but attempts functional defect in the fourth Ca 2+binding site (site IV) at
to find calmodulin in procaryotes were unsuccessful. In the the most C-terminal region [30, 31,37]. This was confirmed
course of the survey experiments, we came across unusual by preparing a recombinant calmodulin, YCMt1 I, in which
properties of calmodulin of Saccharomyces cerevisiae [29, a part of the C-terminal sequence of yeast calmodulin
30]. Although the conformational similarity of yeast cal- spanning the Ca 2+ binding loop (residues 130 to the C
modulin to the vertebrate calmodulins has been suggested [29- terminus) was deleted [31]. YCMt1 bound 3 mol Ca 2+ with
32], remarkable differences have been proposed from the CD higher affinity than that of the native yeast calmodulin (Fig.
spectral change [31, 33], small angle X-ray scattering [34], 1). Results of quantitative analyses by the curve fitting to the
and cooperativity among three Ca 2+bindings [30, 35, 36]. Adair equation and estimated macroscopic dissociation
constants were summarized in Table 1 [36] (K. Nakashima
and M. Yazawa, manuscript in preparation). Yeast cal-
modulin or YCMO 1 (a recombinant yeast calmodulin

-
'0
E
4

•
•
expressed in Escherichia coli) binds 3 Ca 2+ in a highly
cooperative manner, and a deletion of the defective site IV
sequence resulted in dissociation of the feature and a
::::= 3

-:s
0 cooperative binding of the high-affinity 2 Ca 2+was separate-
E ly observed from the weaker 1 Ca 2+ binding (Figure 1 and
Q
c 2 Table 1). Since Ca2+binding sites in the N-terminal half have
been assigned as the high affinity sites [35] (S. Ohki, K.
c Hikichi, and M. Yazawa, unpublished results), the deleted
iii 1
+ residues in the site IV sequence may be involved in the
N
ftI interdomain interaction leading to the cooperative binding
U
of 3 Ca2+observed in yeast calmodulin.
~0-7 10-4

Yeast calmodulin is a poor activator for

Fig. 1. Ca'· binding to yeast calmodulin and fragment YCMD.. Ca binding
H
target enzymes of vertebrate origin
was measured by the flow-dialysis method using "CaCI, in a medium of
0.1 M NaCI and 20 mM MOPS-NaOH, pH 7.0, at 25°C. • , YCMO; .... ,
YCMD.. Solid lines are the best-fit curves to the Adair's equation. Dotted Since the primary structure ofyeast calmodulin is considerably
line shows a typical Ca'· binding curve of vertebrate calmodulin. divergent compared with that of ordinary calmodulin (about
 49

Table I. Ca 2+binding to yeast calmodulin and fragments' these poor efficiency, yeast calmodulin showed ordinary
Ca2+-dependent regulatory function based on the Ca 2+ binding
Macroscopic Dissociation Constant fuM)
CaM K2 K3
property satisfying the range ofCa 2+ concentration observed
K,
in the intracellular Ca 2+ regulation (Fig. 3) [30]. Similar
YCMO 2.75 3.63 2.18 3 \.4 results have been reported by Ohyaet al. [33] with use ofPDE
YI2 12.8 1.21 26.6
and plant NAD kinase. As shown in Fig. 1, YCM~ bound 3
Y34 5.23 -16.5
YCMti 2.18 0.299 2.20 55.6 Ca 2+ in the similar range of Ca 2+ concentration, but it hardly
activated these target enzymes (Fig. 2) [31].
• Each set of macroscopic dissociation constants K" K" .... , K, were
Davis et al. [37] and Ohya et al. [41] indicated that
estimated by the curve-fitting to the Adair equation,
calmodulin was essential for S. cerevisiae, and over expres-
y = (x/K, + 2 x'/K,K, + + nx'/K,K 2 . . . • . K,)I sion of vertebrate calmodulin or even each one of the
(I + x/K, + x'/K,K, + + x'/K,K, K,) + jx
half-molecular fragment of yeast calmodulin could com-
where y is the amount ofCa2+ bound to calmodulin or its fragment in molar plement the function of yeast calmodulin [37, 38,41].
ratio at the free Ca 2+concentration x.

60% identical), its Ca2+ - dependent regulatory properties were Interdomain interaction in yeast
examined using target enzymes of vertebrate calmodulin,
cyclic-nucleotide phosphodiesterase (PDE) I and myosin calmodulin
light chain kinase (MLCK) '. As shown in Fig, 2, yeast
In ordinary calmodulin, each of the half-molecular domains
calmodulin is a poor activator for these vertebrate enzymes [30,
containing two Ca 2+ binding sites is independent of the other
31]. More than I,OOO-fold molar excess of yeast calmodulin
and Ca 2+ binding properties of calmodulin fit to the simple
was required for the maximum activation ofPDE and MLCK.
sum ofthe Ca2+ bindings of each half molecular fragment [25,
The maximum activity ofMLCK expressed by vertebrate cal-
26]. Similar experiments were performed using a pair of
modulin could not be attained by yeast calmodulin. Under
recombinant half-molecular fragments, Y 12 ' (the N-terminal
half) and Y34 1 (the C-terminal half), of yeast calmodulin
(Table I). A simple sum of the results for each fragment
produced a different curve from the one of yeast calmodulin
(data not shown), which is consistent to the results of Fig. I
~

~
UJ suggesting some interdomain interaction in yeast calmodulin.
0
a. The interdomain interaction was studied by the measure-
'0 ment of CD spectrum. Ca 2+ binding to ordinary calmodulin
c 50 decreases the molar ellipticity (6) around 222 nm which may
.~

- correspond to the apparent increase in a helix content (Fig.

"lO
>
v 4A). On the other hand, as shown by Ohya et al. [33], Ca2+
« binding to yeast calmodulin caused the inverse effect (Fig. 4B).

~100 Chicken 15
:.... Chicken CaM
:.c: ~

u
..... "v
~
a.
lI\

-
'0 10
c >-

-
0 50 .;
~
> v
ti
Yeast «
« :.c: 5
~ u
.....
• ~ YCMI1 ~
Yeast CaM

.-.,_ _--4e..-e-._ _e_

0 -e-~ --6-6--
o
10 9 8 7 6
-LOG[Calmodulin] (M)
5 4
0
.- .........
7 6 5 4
pCa
Fig. 2. Activation of pig brain POE and rabbit skeletal MLCK. 0, Yeast
calmodulin; e, YCMO; .A., YCMti. Activation profiles by vertebrate Fig. 3. Ca'+ dependence of MLCK activity. 0, Vertebrate (chicken)
(chicken) calmodulin were shown by solid lines without data points. calmodulin; e, yeast calmodulin.
50

As fragments of yeast calmodulin, YCML1, YI2 and Y34, o r----..,...--.....,~--..,...--.....,~~-...,

A
showed Ca2+ -dependent decrease in e (Fig. 4C for YCML1)
[31], interdomain interaction through the C-terminal sequence
may be responsible for the inverse CD change observed in
yeast calmodulin. Ca2+ dependence of the global structural
change in yeast calmodulin detected by the CD spectra was
quantitatively analyzed (K. Nakashima and M. Yazawa,
manuscript in preparation). Initially, the relative distribution,
2
PCaCaM' PCa2CaM and PCa3CaM' of intermediate species of yeast
calmodulin, that is, the single-site- (Ca-CaM), two-sites-
(Ca 2-CaM) and three-sites-saturated calmodulin (Ca 3-CaM),
respectively, were estimated from concentrations of total
protein and Ca 2+ using macroscopic dissociation constants
(Table I). Then the relative contribution of each intermediate
was estimated according to the following equation,

Y = aP CaCaM + pAp Ca2CaM + PCa3CaM

where y is the relative spectral change, and a and Pare the

relative contribution factors of Ca-CaM and Ca 2-CaM,
respectively, to the spectral change ofthe final Ca 2+-saturated
conformation (Ca 3 -CaM). A pair of values a and p which
gave the best-fit curve (solid line in Fig. 5) to the relative
spectral changes (open circle in Fig. 5) of yeast calmodulin
was -0.39 and 1.0, respectively. Therefore, the initial 2 Ca2+
bindings to the N-terminal half (the high-affinity sites in yeast
calmodulin) were sufficient to induce the final global structure
of whole molecule. Since the spectral change in the ordinary
calmodulin increased linearly to the molar ratio of Ca 2+/
calmodulin of 4, there must be some information transfer in
yeast calmodulin through the C-terminal half.
Similar effect was observed by the I H NMR measurement
(K. Nakashima, S. Ohki, K. Hikichi and M. Yazawa, manu- 2
script in preparation). The chemical shift ofC2H of His 107
in the C-terminal half showed the three-state change in a slow
exchange manner, corresponding to the apo-, intermediate
and the Ca 2+-saturated state. Ca 2+ dependence of the signal
200 210 220 230 240 250
intensity of the intermediate species fitted to the distribution
of the sum ofCa-CaM and Ca2-CaM. Ca 2+ binding to the N- Wavelength (nm)
terminal half of yeast calmodulin induced the structural
change in the C-terminal half around His I07. On the other Fig. 4. Effect of Ca'· on the far UV CD spectra (shown as the molar
ellipticity) ofCaMs and fragments. CD spectra were measured in a medium
hand, the C2H signal of His I07 in vertebrate calmodulin of 0.1 M NaCI and 20 mM MOPS-NaOH, pH 7.0, at 25°C in the presence
showed the two-state change depending on Ca2+ saturation of 0.2 mM CaCl, (spectrum c in each panel) or O. I mM EGTA (spectrum
of the C-terminal half (the high-affinity Ca 2+ sites), and the e in each panel). (A) Invertebrate (scallop) calmodulin, (B) YCMO, (C)
similar 3-state change was observed only in the presence of YCM~.

peptide corresponding to the calmodulin binding domain of

the target enzyme of which binding induced the interdomain Possible role of the defective site IV
interaction in calmodulin [42].
These results suggested some interdomain interaction of sequence
yeast calmodulin, and the amino acid sequence around the
defective site IV may playa role for this interaction. Involve- Figure 6 summarizes interdomain interactions ofcalmodulins
ment ofthe site III sequence to the interaction is also probable, which can explain the experimental results. Ordinary cal-
since YCML1 also showed a similar effect (Fig. I, Table I). modulin is shown as the dumbbell shape (Ca-CaM), in which
 51

YCMd affected the Ca2+ binding (Table 1), a part of the site
1 III sequence may also be involved in the interdomain inter-
action (Ca-YCMd). The models are consistent to the results
N
N
N
ofthe solution X-ray scattering measurements, in which yeast
CD calmodulin was shown to form an asymmetric globular-like
.~ shape on Ca 2+ binding [34]. The structure ofyeast calmodulin
~O suggests its specific mode of target recognition which is

~ different from vertebrate calmodulin.

Recent progress in the 3D-structural analyses accumulated
a lot of structure of Ca 2+ binding proteins of EF-hand family
[43], and several modes of interaction with target proteins [39,
, ~.:.::::-.. :.~ :".~.~ .. 40, 44, 45] were presented other than the typical mode of
calmodulin-peptide complex shown in Fig. 6. Among them
3 6 crystal structure of amphioxus sarcoplasmic Ca 2+ binding
[Cal I [Calmodulin] (M/M)
protein [46] and the one of recoverin [39] are interesting to
understand the mechanism of target recognition of yeast
Fig. 5. Effect of total Ca2• concentration on the CD spectral changes of calmodulin. Both of them consisting of 4 EF-hand motifs
yeast calmodulin (YCMO). Relative changes in the molar ellipticity at 222
adopt the folded structure with interacting half molecular
nm (0) were calculated and they were compared with the statistical
distributions ofPC.-C.M (dotted line), PC.2-C.M (dotted-broken line) and PC.)-C.M
domains, and in their C-terminal half domains site IV does
(broken line). These lines were calculated using the macroscopic disso- not bind Ca2+. In spite of the defect in site IV the two EF-hand
ciation constants given in Table I and the total concentration of YCMO. motifs in the C-terminal half maintain the apparent two-fold
Solid line is the best-fit curve (details in the Text). rotation symmetry observed in the typical EF-hand pairs [39,
46]. The observed high affinity for Ca2+ of Y34 may be a
result of this symmetry maintained in the C-terminal half of
each half-molecular domain is independent of the other and yeast calmodulin. Then the interdomain interaction within the
is connected by the flexible linker. The interdomain inter- folded structure of these proteins is helpful to understand the
action is stable only in the ternary complex formed with the interdomain interaction of yeast calmodulin. In recoverin,
target peptide (Ca-CaM-Peptide), and this is an example of part of the site III sequence contacts with the site II sequence
the typical modes of target recognition and a step in the and the C-terminal segment of the site IV sequence folds up
mechanism of target activation. On the other hand, yeast to interact with the N-terminal half molecular domain which
calmodulin binds only 3 Ca 2+ and the defective site IV form a hydrophobic core on the inverse face of the four Ca 2+
sequence is involved in the interdomain interaction which binding loops [39]. As predicted by the authors, this might
causes the cooperative 3 Ca 2+ binding (Ca-YCMO). Since form a different type of interaction with targets. Interaction

Ca·CaM· Pep tide Ca·CaM Ca·YCMO Ca·YCMll

Fig. 6. Schematic structure ofCa 2• -saturated calmodulins. In each structure larger circles are the globular half-molecular domains, and small circles indicate
bound Ca 2•• Thick solid lines indicate the linker between the terminal half-molecular domains. In YCMO and YCML'i, the C-terminal half-molecular domain
binds I Ca2., and the defective site IV segments are shown by open tails.
52

ofcalcineurin B with the helical segment of the A subunit is II. Cohen P, Burchell A, Foulkes JG, Cohen PTW, Vanaman TC, Nairn
likely to be the case. Calcineurin B interacts with subunit A AC: Identification of the Ca'· -dependent modulator protein as the
fourth subunit of rabbit skeletal muscle phosphorylase kinase. FEBS
on the inverse face of the four Ca 2+ binding loops where the Lett 92: 287-293,1978
long C-terminal segments and the linker of the two half 12. Stewart AA, Ingebritsen TS, Manalan A, Klee CB, Cohen P: Discovery
molecular domains make contact to the helical structure of ofa Cal+_ and calmodulin-dependent protein phosphatase. FEBS Lett
the A subunit [40]. If this can be applied to the structure of 137: 80-84, 1982
yeast calmodulin, the hydrophobic C-terminal sequence may 13. Klee CB, VanamanTC: Calmodulin.Adv Protein Chern 35: 213-321,
1982
be indispensable not only to maintain the stable structure of 14. Kretsinger RH, Nockolds CE: Carp muscle calcium-binding protein.
the EF-hand pair in the C-terminal half, but also to contact to II. Structure determination and general description. J Bioi Chern 248:
the N-terminal half to induce the cooperative three Ca 2+ 3313-3326, 1973
binding. In yeast calmodulin, the C-terminal sequence may 15. Toda H, Yazawa M, Sakiyama F, Vagi K: Amino acid sequence of
be directly involved in recognizing target proteins as found calmodulin from wheat germ. J Biochem 98: 833-842, 1985
16. Babu YS, Bugg CE, Cook WJ: Structure of calmodulin refined at 2.2
in the structure ofcalcineurin. The difference in the mechanism it.. resolution. J Mol Bioi 204: 191-204, 1988
to recognize target proteins may be one of the major reason to 17. lkura M, Clore GM, Gronenborn AM, Zhu G, Klee CB, Bax A:
the observed poor activation of enzymes of vertebrate origin. Solution structure of a calmodulin-target peptide complex by multi-
Ifthis is the case, isolated target proteins in yeast may be more dimensional NMR. Science 256: 632--638, 1992
efficiently activated by yeast calmodulin than by vertebrate 18. Meador WE, Means AR, Quiocho FA: Target enzyme recognition by
calmodulin: 2.4 it.. structure ofa calmodulin-peptide complex. Science
calmodulin. The reason why the novel mode of interaction 257: 1251-1255, 1992
is predominant in yeast may be another interesting point. We 19. Barbato G, Ikura M, Kay LE, Pastor RW, Bax A: Backbone dynamics
are involved in the 3-D structural analysis ofyeast calmodulin of calmodulin studied by I5N relaxation using inverse detected
and the isolation of target proteins in yeast is in progress. two-dimensional NMR spectroscopy: The central helix is flexible.
Biochemistry 31: 5269-5278, 1992
20. Tjandra N, Kuboniwa H, Ren H, Bax A: Rotational dynamics of
calcium-free calmodulin studied by I5N-NMR relaxation measure-
Acknowledgements ments. Eur J Biochem 230: 1014-1024, 1995
21. Zhang M, Tanaka T, Ikura M: Calcium-induced conformational
transition revealed by the solution structure ofapo calmodulin. Nature
This work was supported in part by a Grant-in-Aid for
Struct Bioi 2: 758-767, 1995
Scientific Research 07309013 from the Ministry ofEducation, 22. Kuboniwa H, Tjandra N, Grzesiek S, Ren H, Mee CB, Bax A: Solution
Science and Culture of Japan. structure ofcalcium-free calmodulin. Nature Struct Bioi 2: 768-776, 1995
23. lkura M: Calcium binding and conformational response in EF-hand
proteins. Trends Biochem Sci 21: 14-17, 1996
24. Walsh M, Stevens FC, Kuznicki J, Drabikowski W: Characterization of
References tryptic fragments obtained from bovine brain protein modulator ofcyclic
nucleotide phosphodiesterase. J Bioi Chern 252: 7440-7443, 1977
I. Kakiuchi S, Yamazaki R, Nakajima H: Properties of a heat-stable 25. Drabikowski W, Brzeska H, Venyaminov SY: Tryptic fragments of
phosphodiesterase activating factor isolated from brain extract. Proc calmodulin. Cal+- and Mg'·-induced conformational changes. J Bioi
Japan Acad 46: 587-592, 1970 Chern 257: 11584-11590, 1982
2. CheungWY: Cyclic 3',5'-nucleotide phosphodiesterase. Demonstration 26. Minowa 0, Vagi K: Calcium binding to tryptic fragments ofcalmodulin.
of an activator. Biochem Biophys Res Commun 38: 533-538, 1970 J Biochem 96: 1175-1182, 1984
3. Teo TS, Wang JH: Mechanism of activation of a cyclic adenosine 3',5'- 27. Vogel HJ, Lindahl L, Thulin E: Calcium-dependent hydrophobic
monophosphate phosphodiesterase from bovine heart by calcium ions. interaction chromatography of calmodulin, troponin C and their
Identification of the protein activator as a Ca'· binding protein. J Bioi proteolytic fragments. FEBS Lett 157: 241-246, 1983
Chern 248: 5950-5955,1973 28. Brzeska H, Szynkiewicz J, Drabikowski W: Localization of hydro-
4. Ebashi S, Kodama A: A new protein factor promoting aggregation of phobic sites in calmodulin and skeletal muscle troponin C studied using
tropomyosin. J Biochem 58: 107-108, 1965 tryptic fragments. A simple method of their preparation. Biochem
5. Ebashi S: Calcium binding and relaxation in the actomyosin system. J Biophys Res Commun 115: 87-93, 1983
Biochem 48: 150-151, 1960 29. Nakamura T, Fujita K, Eguchi Y, Yazawa M: Properties of calcium-
6. Yazawa M, Vagi K: A calcium-binding subunit of myosin light chain dependent regulatory proteins from fungi and yeast. J Biochem 95:
kinase. J Biochem 82: 287-289, 1977 1551-1557,1984
7. Vagi K, Yazawa M, Kakiuchi S, Ohshima M, Uenishi K: Identification 30. Luan Y, Matsuura I, Yazawa M, Nakamura T, Vagi K: Yeast calmodulin:
of an activator protein for myosin light chain kinase as the Cal+- Structural and functional differences compared with vertebrate
dependent modulator protein. J Bioi Chern 253: 1338-1340, 1978 calmodulin. J Biochem 102: 1531-1537,1987
8. Ebashi S, Endo M: Calcium ion and muscle contraction. Prog Biophys 31. Matsuura I, Ishihara K, Nakai Y, Yazawa M, Toda H, Vagi K: A
Mol Bioi 18: 123-183, 1968 site-directed mutagenesis study of yeast calmodulin. J Biochem 109:
9. Ozawa E: Activation of muscular phosphorylase b kinase by a minute 190-197,1991
amount ofCa ion. J Biochem 71: 321-331,1972 32. Matsuura I, Kimura E, Tai K, Yazawa M: Mutagenesis of the fourth
10. Brostrom CO, Hunkeler FL, Krebs EG: The regulation ofskeletal muscle calcium binding domain of yeast calmodulin. J BioI Chern 268: 13267-
phosphorylase kinase by Cal+. J Bioi Chern 246: 1961-1967, 1971 13273, 1993
 53
33. Ohya Y, Uno I, Ishikawa T, Anraku Y: Purification and biochemical 40. Griffith JP, Kim JL, Kim EE, Sintchak MD, Thomson JA, Fitzgibbon
properties of calmodulin from Saccharomyces cerevisiae. Eur J MJ, Fleming MA, Caron PR, Hsiao K, Navia MA: X-ray structure of
Biochem 168: 13-19, 1987 calcineurin inhibited by the immunophilin-immunosuppressant
34. Yoshino H, Izumi Y, Sakai K, Takezawa H, Matsuura I, Maekawa H, FKBPI2-FK506 complex. Cell 82: 507-522, 1995
Yazawa M: Solution X-ray scattering data show structural differences 41. Ohya Y, Anraku Y: Functional expression of chicken calmodulin in
between yeast and vertebrate calmodulin: Implications for structure/ yeast. Biochem Biophys Res Commun 158: 541-547,1989
function. Biochemistry 35: 2388-2393, 1996 42. Yazawa M, Ikura M, Hikichi K, Luan Y, Yagi K: Communication
35. Starovasnik MA, Davis TN, Kievit RE: Similarities and differences between two globular domains of calmodulin in the presence of masto-
between yeast and vertebrate calmodulin: An examination of the paran or caldesmon fragment. J Bioi Chern 262: 10951-10954, 1987
calcium-binding and structural properties of calmodulin from the yeast 43. Strynadka NW, James MNG: Crystal structures of the helix-loop-helix
Saccharomyces cerevisiae. Biochemistry 32: 3261-3270, 1993 calcium-binding proteins. Annu Rev Biochem 58: 951-998, 1989
36. Nakashima K, Maekawa H, Yazawa M: Chimeras of yeast and chicken 44. Rayment I, Rypniewski WR, Schmidt-Base K, Smith R, Tomchick
calmodulin demonstrate differences in activation mechanisms of target DR, Benning MM, Winkelmann DA, Wesenberg G, Holden HM:
enzymes. Biochemistry 35: 5602-5610, 1996 Three-dimensional structure of myosin subfragment-I: A molecular
37. Davis TN, Urdea MS, Masiarz FR, Thorner J: Isolation of the yeast motor. Science 261: 50---58, 1993
calmodulin gene: Calmodulin is an essential protein. Cell 47: 423- 45. HoudusseA, Cohen C: Target sequence recognition by the calmodulin
431,1986 superfamily: Implications from light chain binding to the regulatory
38. Sun GH, Ohya Y, Anraku Y: Half-calmodulin is sufficient for cell domain of scallop myosin. Proc Nat! Acad Sci USA 92: 10644-10647,
proliferation. Expression ofN- and C-terminal halves of calmodulin in 1995
the yeast Saccharomyces cerevisiae. J Bioi Chern 266: 7008-7015, 1991 46. Cook WJ, Jeffrey LC, Cox JA, Vijay-Kumar S: Structure of sarco-
39. Flaherty KM, Zozulya S, Stryer L, McKay DB: Three-dimensional struc- plasmic calcium-binding protein from amphioxus refined at 2.4 A
ture of recoverin, a calcium sensor in vision. Cell 75: 709-716,1993 resolution. J Mol Bioi 229: 461--471,1993
Molecular and Cellular Biochemistry 190: 55-62, 1999.
© 1999 Kluwer Academic Publishers.

Regulation by molluscan myosins*

Andrew G. Szent-Gyorgyi, Vassilios N. Kalabokis and Cynthia L.

Perreault-Micale
Department ofBiology, Brandeis University, Waltham, Massachusetts, USA

Abstract
Molluscan myosins are regulated molecules that control muscle contraction by the selective binding of calcium. The essential
and the regulatory light chains are regulatory subunits. Scallop myosin is the favorite material for studying the interactions of
the light chains with the myosin heavy chain since the regulatory light chains can be reversibly removed from it and its essential
light chains can be exchanged. Mutational and structural studies show that the essential light chain binds calcium provided that
the Ca-binding loop is stabilized by specific interactions with the regulatory light chain and the heavy chain. The regulatory
light chains are inhibitory subunits. Regulation requires the presence of both myosin heads and an intact headrod junction.
Heavy meromyosin is regulated and shows cooperative features of activation while subfragment-I is non-cooperative. The
myosin heavy chains of the functionally different phasic striated and the smooth catch muscle myosins are products of a single
gene, the isoforms arise from alternative splicing. The differences between residues of the isoforms are clustered at surface
loop-I of the heavy chain and account for the different ATPase activity of the two muscle types. Catch muscles contain two
regulatory light chain isoforms, one phosphorylatable by gizzard myosin light chain kinase. Phosphorylation of the light chain
does not alter ATPase activity. We could not find evidence that light chain phosphorylation is responsible for the catch state.
(Mol Cell Biochem 190: 55---Q2, 1999)

Key words: myosin light chain, myosin isoforms, heavy meromyosin

Introduction chain (RLC) and the essential light chain (ELC). The ELC
contains the Ca 2+-binding site, which is stabilized by the
In this paper we will review the outstanding features of interaction with the RLC. The presence of the RLC is also
regulation by molluscan myosins, and then discuss in some required for the inhibition ofATPase activity, tension genera-
detail both the structural features ofthe molecule that give rise tion and motility in the absence of Ca2+. Most of the results
to its regulation and cooperative activations and the modulation on myosin-linked regulation were obtained from studies on
ofATPase activity by slight sequence variations. We will also both scallop muscle and its myosin because of the ease of
examine the effects oflight chain phosphorylation. reversibly removing the RLCs, and the ability to exchange the
Contraction of molluscan muscles is elicited by the binding ELCs. Foreign RLCs can readily hybridize with myosin
of Ca 2+to myosin (cf[ 1D. Molluscan myosins are unique in deprived of its own RLCs. The hybrids are functional if the
that they bind Ca2+selectively; this results in a conformational source ofthe RLCs is a regulated myosin. Hybrids using rabbit
change which acts as a Ca2+-dependent 'on' and 'off' switch. or chicken skeletal RLCs do not restore specific Ca2+-binding,
Molluscan myosins are regulated molecules consisting oftwo and theirATPase activity is not stimulated by Ca2+. The require-
heavy chains and two pairs oflight chains, the regulatory light ments for the ELC are even more restrictive, and Ca2+-binding
is restored only by foreign ELCs derived from Ca2+-binding
*This paper is dedicated to Professor Setsuro Ebashi whose experimental
myosins. Hybridization with foreign ELC is frequently limited,
and conceptual insight taught us how to think about relaxation of muscle, since the exchange depends on relative affinities. A significant
and how to explore its mechanism at the molecular level. step in the studies on the role of the ELC was the demon-

Address for offprints: A.G. Szent-Gyorgyi, Department of Biology, Brandeis University, Waltham, Massachusetts 02254-9110, USA
56

stration that Ca2+-binding is retained by the regulatory domain activity of single-headed myosins 3-4 fold compared to the
(RD) which consists ofa RLC, an ELC and - 10 kDa heavy 10-20 fold activation of intact myosin [7]. More recently,
chain fragment in stoichiometric amounts. The RD complex methods have been developed to obtain pure single-headed
dissociates in the presence of high concentrations of urea myosin preparations without loss ofATPase activity per head
and can be reconstituted into a Ca 2+-binding complex. The in the presence ofCa2+. The Ca2+-sensitivity ofsingle headed
role of the heavy chain fragment is less specific. Heavy «
myosin I - ATPase EGT/ ATPase ca2 J x 100) is about 70%,
chain fragments from chicken gizzard or rabbit skeletal compared to the 90-95% sensitivity of intact myosins [8]. The
myosin can also support Ca2+-binding by the ELC, although limited sensitivity of single-headed myosin preparation
with diminished affinity [2]. The RD crystallizes, and its could be due to the presence of unregulated and fully
atomic structure has been determined [3,4]. Therefore, the regulated molecules, or that the single-headed myosin
system is well suited to relate structure to function. Both preparations may consist ofa single population of molecules
light chains consist of four EF hands. However, only one of that are not fully regulated. Two lines of experiments
the four EF hands retains the ability to bind divalent cations;
this ability is lost in the other EF hands due to mutations,
deletions and substitutions.
Although the ELC contains the Ca2+-binding site, Ca 2+-
binding requires the presence of all three subunits of scallop
~0.6
'Ul
A
myosin. The isolated light chains do not bind Ca 2+ in the Gl
Ul
presence of excess magnesium. Chimeras between the «l

scallop and the non-functional skeletal myosin light chains ~ 0.4

show that the regions of scallop light chains necessary for i..,
Ca 2+-binding and Ca 2+-sensitivity are domain- I of the ELC .....,~
and domain-3 of the RLC [5, 6]. These findings complement ~ 0.2
the structural studies that show that interactions between the el
RLC and ELC are restricted to these domains of the two 0
light chains. The atomic structure also shows that the key
residues establishing a bond between the light chains are
glycine-23 of the ELC and glycine-117 of the RLC [3]. The
1.00
B
0
1
interaction between these residues allows for the stabi- ] 1
... 0.75 1
.a
lization of the unusual Ca 2+-binding loop of the ELC by the I
RLC. Replacement of glycine-117 of the RLC, even by ...
alanine, abolishes Ca 2+-binding and Ca 2+-sensitivity of ) 0.50
ATPase. Skeletal RLCs lack glycine in the corresponding
position. A single site mutation of cysteine to glycine
a
~ 1
~
0.25
converts the non-functional rabbit RLC into one that can
support Ca 2+-binding and calcium sensitivity [6].
8 7 6 5

Cooperative effects
Fig. I. Comparison ofCa" -binding and Ca" -activation of ATPase activity
of myosin and single-headed myosin. Panel A. The Ca" -activated ATPase
Subfragment-I (S 1) represents a single myosin head. It is
activity of myosin (e) and single headed myosin (shM) (0) is plotted versus
not a regulated molecule, although it binds Ca 2+and contains the concentration of free Ca". The ATPase activity was measured in 40
both the RLC and ELC. S 1 is fully active even in the mM KCI, 20 mM MOPS, 10 mM MgCI" 0.5 mM DTT, 0.1 mM CaCI" and
absence of Ca 2+ ions. In contrast, the soluble two headed various concentrations of EGTA, at pH 7.0, and 22°C. Myosin was
heavy meromyosin (HMM) is regulated similarly to myosin activated cooperatively by Ca". Half maximal activation occurred at 0.8
11M Ca". Activation of shM by Ca" was hyperbolic with half-maximal
filaments. S 1 also lacks the S2 region of the rod that
activation at 0.09 11M. Note that only the Ca" -activated portion of the
connects the myosin head with the light meromyosin portion curves is shown. The ATPase activities at pCa 8.7,0.06 sec-I for myosin
ofthe rod. Analysis ofsingle-headed myosin (scallop myosin and 0.26 sec"" for shM, were subtracted from all values. Ca" increased
from which a head was removed by limited papain digestion) the ATPase activities of these myosin and shM preparations 11- and
may help to determine whether regulation depends on the 2.8-fold, respectively. Panel B. Ca"-binding was measured in an identical
solution as used for the ATPase studies except for the absence of ATP.
presence of both myosin heads and/or an intact head rod The dissociation constants of the complexes of myosin (e) and shM (0)
junction. Early experiments on mixtures ofsingle-headed and with Ca2+ were 0.2 and 0.09 11M respectively. (Kalabokis et al., J Bioi
intact myosins indicated that Ca 2+ activated the ATPase Chern, 271: 26779-26782, 1996).
 57

indicate that the population is uniform. Myosin is activated cannot be explained by head-to-head interactions. The com-
cooperatively by Ca2+, but the activation of single headed munication between the Ca2+-binding site formed by the two
myosin is non-cooperative (Fig. I). Single turnover studies light chains and the heavy chain in the regulatory domain,
also demonstrate that the single-headed myosin preparation and the nucleotide binding site of the motor domain is
consists ofa single population of molecules. Intact myosin probably affected by the head-rod junction, since scallop S I
has a very slow turnover rate in the absence of Ca 2+with a is unregulated. The interaction of the two heads of myosin,
half life time of greater than three minutes (Fig. 2) [9]. In possibly between the RLCs, may account for the cooperative
contrast, the turnover rate of single-headed myosin is less activation of scallop myosin and its greater degree of regula-
than the 30 sec which is the time resolution of the method tion than that of single-headed myosin. It appears that both
employed (Fig. 2). The absence of a slow phase of Pi release the double-headed nature of myosin and the attachment of
suggests that the 3 fold activation of the ATPase activity of the coiled-coil rod contribute to the 'off state of scallop
single-headed myosin by Ca 2+is not due to the presence of myosin [8].
two populations of molecules, i.e. one regulated with In general, both myosin and HMM show similar differ-
properties similar to those of myosin and one unregulated ences from S I. A communication between the ATP-binding
that has high ATPase activity in the presence ofEGTA. The site of the motor domain and the Ca 2+-binding site of the
'off state is the unique state that distinguishes regulated regulatory domain is demonstrated by the Ca2+activation of
scallop myosin from the unregulated skeletal myosin. the ATPase activity. This communication exists in myosin and
Evidently, to reach this state, an interaction between the two HMM but not in S I. Ca2+is bound in a cooperative manner
heads is required; the interaction is disrupted by Ca 2+. to HMM and to myosin in the presence of ADP or non-
However, the partial regulation of single-headed myosin hydrolyzable ATP analogues. Ca 2+-binding to S I is non-

40000

30000

20000

10000

A A
x A x·
A A
XA x A A

2 4 6 8 10 12 14
Time (min)

Fig. 2. Single turnovers of myosin and single-headed myosin with [y_3 2P]ATP. Ten-fold (0; 16 11M myosin heads, 160 11M ATP) and 1.2-fold (0; 16 11M
myosin heads, 19.2 11M ATP) molar excess of [y_32P]ATP to myosin heads in the presence of EOTA. The rates of Pi release were 0.22 and 0.23 min-I with a
10- and 1.2-fold molar excess of ATP to myosin heads, respectively. Ten-fold (x) molar excess of [y_32P]ATP to myosin heads in the presence ofCa'+. _, the
amount of radioactivity remaining on the filter when myosin, treated with a 10-fold molar excess of [y-3'P]ATP to myosin heads in the presence of EOTA,
was filtered and washed with buffer containing 0.1 mM CaCI, instead of I mM EOTA. Single-headed myosin (8.5 11M) reacted with a IO-fold molar excess
of [y-32P]ATP (85 11M) in the presence of Ca'+ (.A.) or in the presence of EOTA (6). The high turnover rate of shM in the presence of EOTA (typical values
were 0.2--0.3 sec-') did not allow for the measurement of the rate of Pi release with this method, which has a resolution time of -30 sec. (Kalabokis et al., J
BioI Chern, 271: 26779-26782, 1996).
58

cooperative under the same conditions. There is antagonism preparations obtained from the closely related species,
between the binding of nucleotide analogues and Ca 2+ in Argopecten irradians.
myosin and HMM which is not observed in Sl [23). The myosin light chains are not responsible for the dif-
ferences in ATPase activities. Both the ELCs and the RLCs
of the three myosins have been cloned and sequenced. The
Surface loop-l can modulate ATPase activity nucleotide sequences of the ELCs fromPlacopecten striated
and catch muscle myosins are identical and show 95%
The actin-activated ATPase activities of the striated muscle identity and 98% similarity to the Argopecten ELC, in-
myosin and S I ofthe sea scallop (Placopecten magellanict!s) dicating that the ELCs do not account for the tissue and
is about 2-3 fold higher than the activity ofthe smooth, catch species specific differences in ATPase activities. Three RLC
muscle myosin and S1. The activity of the Placopecten isoforms have been identified from Placopecten, one from
striated myosin and S I is also 2-3 fold higher than the striated and two from catch muscles as described previously

Regulatory light chain phosphorylation

-terminal sequence surrounding the phosphorylatable erine:

Placo smoA MADKERAQRATSNVFARLPQKLMQEMKEAFTMID

Placo smoH ---- ••••• -A-G-LTK----QI-----------
Placo str ---- ••••• -A-G-LTK----QI-------S-M-
Argo str ---- ••••• -A-G-LTK----QI-------S---
Merc trans -DSDKK-KA---S-LTKFT-NQI-----------

Urea gel of RLC phosphorylation:

Species: Placopecten Argopecten Mercenar;f1

Muscle: smoA striated smoB striated tran
ATP: + + + + +

Fig. 3. Upper panel. A comparison of RLC N-terminal sequences that surround the phosphorylatable serine. Plaeopecten smoA isoform has a five amino
acid insertion compared to the Plaeopecten smoB and striated and the Argopeeten striated sequences. This insertion provides the consensus sequence for
phosphorylation by the smooth muscle myosin light chain kinase, which has a requirement for two arginine residues preceding the phosphorylatable serine
[20]. Although the Mereenaria RLC has an extended N-tenninus which also contains additional basic residues, it cannot be phosphorylated (see lower
panel). Sequences are compared to the Placopeeten smoA isoform. Placo = Plaeopecten, Argo = Argopeeten, mere = Mereenaria, str = striated, trans =
translucent, P = phosphorylatable serine. Lower panel. 12.5% urea polyacrylamide gel of phosphorylation of several molluscan RLCs. -ATP and +ATP
refers to the absence or presence of ATP. Only the smoA isoform was phosphorylated. The unphosphorylated smoA isoform migrates differently on urea
gels compared to the other molluscan RLCs due to the additional charges of arginine present in its five amino acid insertion.
 59

for the RLCs of Patinopecten yessoensis [10] (Fig. 3). the regulation ofthe catch state, since about 40% ofthe catch
Sequence differences are restricted to the nucleotides en- muscle myosin in Patinopecten yessoensis contains a phos-
coding some of the N-terminal 52 amino acids. Analysis of phorylatable RLC isoform (smoA RLC). These authors
the genomic DNA indicated that these RLC isoforms are the suggest that RLC phosphorylation shuts offATPase activity
products of a single gene that is alternatively spliced in the [17]. RLC phosphorylation modulates the behavior of a
5' region only. The ATPase activities of Placopecten and number ofdifferent muscles. It triggers contraction ofsmooth
Argopecten myosins are not altered when hybridized with the muscles [18] and alters Ca 2+-sensitivity of vertebrate skeletal
recombinant RLC isoforms ofPlacopecten or with the RLC muscles [19].
of Argopecten, indicating that it is the heavy chains that are Cloning and expression ofPlacopecten smoA RLC makes
responsible for the differences in enzymatic activities [11]. it possible to test the effect of RLC phosphorylation on
To identify the regions ofthe heavy chain sequence that may ATPase activity without altering the heavy chain or para-
be responsible for the differences in enzymatic activity, the myosin by phosphorylation. Serine-II of the smoA RLC can
cDNA encoding the myosin heavy chains ofPlacopecten catch be specifically phosphorylated by gizzard myosin light chain
and striated muscle [12] have been cloned and sequenced. The kinase (MLCK). 77% phosphorylation has been achieved
deduced protein sequences indicate that the myosins of the without phosphorylating threonine-l 0 (Fig. 4). The Placo-
two muscles are 97% identical. The differences between the pecten smoA RLC isoform contains an N-terminal five amino
two isoforms arise by differential splicing offive alternative acid insertion that includes the two arginine residues neces-
exons from a single heavy chain gene. Three of the alter- sary to encode the putative consensus sequence, RxxRxxS,
natively spliced exons are present in the heavy chain of S1. for phosphorylation by smooth muscle MLCK [20]. This
Exon-5 encodes the phosphate binding loop and contains sequence is not present in Placopecten smoR RLC and
the flexible surface loop-l at the 25-50 kDa junction that striated RLC isoforms or in Argopecten striated RLC.
is sensitive to proteolysis. Exon-6 contributes to the ATP Although the RLC of Mercenaria translucent adductor has
binding site. Exon- 13 contributes to both the actin-binding an extended Nterminus which also contains basic residues but
andATP binding sites and is near surface loop-2 at the 50- no arginine, it cannot be phosphorylated by smooth MLCK.
20 kDa junction. Sequence diffierences in the S 1 heavy Recombinant Placopecten smoA isoform can be hybridized
chain are clustered in exon 5 (Table I, cf [12-15]). There both in the phosphorylated and unphosphorylated form with
are also differences to a lesser degree in the deduced amino desensitized Placopecten catch myofibrils.ATPase activities
acid sequences of exon 6 and exon 13 However, exons 6 and Ca2+ sensitivity of the hybrids are not altered by phos-
and 13 are identical between Argopecten and Placopecten phorylation (Table 2). Similar results have been obtained on
striated myosin heavy chains in spite of the differences in striated Placopecten and Argopecten myofibrils hybridized
ATPase activity of these myosins. Therefore, it appears the with unphosphorylated and phosphorylated smoA RLCs.
surface loop-l has a major influence on ATPase turnover rates Therefore, phosphorylation of the RLC cannot activate
and is responsible for the different enzymatic activities of ATPase activity and is unlikely to serve as an 'on' switch.
molluscan myosins. These results, however, do not necessarily negate the pos-
sibility that phosphorylation may induce the catch state in the
absence ofCa2+, since both the resting state and the catch state
Light chain phosphorylation and ATPase activity would have a very low ATPase activity. However, if regu-
lation of the catch state requires RLC phosphorylation, then
There are two muscle types in molluscan adductor muscles: all catch muscle myosins should contain a phyosphorylatable
a relatively fast contracting phasic muscle and the slower, RLC. Comparison of the RLCs of catch and phasic muscles
highly specialized tonic catch muscles. The catch muscles by urea gel electrophoresis can help to obtain evidence for
have the ability to retain the tension produced by the con- the presence of phosphorylatable RLCs. The smoA RLCs of
tractile system for an extended period at the expense ofa very Patinopecten andPlacopecten have a slower mobility on urea
low metabolic turnover rate. In the catch state the animal can gels than the other isoforms due to the arginine residues in
keep its shells closed by compressing the elastic ligament at the consensus sequence. In contrast, the RLCs of the phasic
the hinge. The catch state is most likely due to a very slow and catch muscle myosins in Ostrea and Mercenaria have
cycling of the cross-bridge, although a role for paramyosin, identical mobilities. Experiments to detect phosphorylation
present in these muscles in high concentrations, has also been by endogenous light chain kinases have also been negative
suggested [16]. The differences between tetanic contraction [21]. As of now, there is no direct evidence to associate the
and the catch state indicate that the control of the catch state catch state with RLC phosphorylation by MLCK Phos-
requires a regulatory mechanism in addition to the 'on' and phorylation of myosin heavy chain and/or paramyosin,
'off' switch controlling the normal contraction cycle. Phos- possibly by an endogenous cAMP dependent kinase, remains
phorylation of the RLC has been proposed to be involved in a possibility [22].
60
Table 1. Deduced protein sequences spanning residues 1-508 of the myosin heavy chain of Placopecten striated and catch muscle

Placo Str MNIDFNDPDFQYLAVDRKKMMKEQTAPFDGKKNCWVPDPKEGFASAEIQSSKGEEITVKI

Placo Cth
Argo Str -----S-------------L------A-----------E--------------D------
Argo Cth -----S-------------L------A-----------E--------------D------

ATP-binding
61. VSDNSTRTVKKDDIQQMNPPKFEKLEDMANMTYLNEASVLNNLRGRYTAGLIYTYSGLFCIAVN

-A-S-----------S------------------------Y---S---S---------------
-A-S-----------S------------------------Y---S---S---------------

exon 5 s,c
125. PYRRLPIYTDSVIAKYRGKRKTEIPPHLFSVADNAYQNMVTDRENQSCLITGESGAGKTENTKK
------------------------------------------------------------S---
------------------------------------------------------------S---
ATP-binding (25/50 KDa) exon 6 S,C ATP-binding
189. YXMYLAXV'ACAV'KKKTSEEEEADQUeGSLEDOXXOABPYLBAYGNAX'1'TRNNNSSRFGKFIRI
----F-R--ANLY-OKE-PVPNLRAeeSN------E-------------Y--------------
---------------DE-ASDKKEe •• -------------------------------------
----F----ANLY-OKO--PTTTHARASN------E-------F-----Y--------------

253. HFGPTGKIAGADIETYLLEKSRVTYQQSAERNYHIFYQICSNAIPELNEVMLITPDSGLYSFIN

------------------------------------------------D---V-----------
------------------------------------------------D---V-----------

ATP-Binding
317. QGCLTVDNI DDVEEFKLCDEAFDI LGFTKEEKTSMFKCTAS ILHMGEMKFKQRPREEQAESDGT

--------------------------------Q-------------------------------
--------------------------------Q-------------------------------
actin-binding
381. AEAEKVAFLCGI NAGDLLKALLKPKYKYGTEMVTKGQNLQQVINSVGALSKSLYDRMFNWLVKR

--------------------------------------MN--V------A------------R-
--------------------------------------MN--V------A------------R-

445. VNRTLDTKAKRNYYIGVLDIAGFEIFDFNSFEQLCINYTNERLQQFFNHHMFVLEQEEYKKEGI

--K-------------------------------------------------I-----------
--K-------------------------------------------------I-----------
Regions involved in actin andATP-binding (underlined) were derived from the SI crystal structure [13]. 's' = exon spliced into striated (Str) and 'c' = exon
spliced into catch (Cth). These regions that are differentially spliced in Placopecten striated and catch muscles are shown in bold. Proteolysis of the SI
region of the MHC results in cleavage at two flexible surface loops to yield fragments of25, 50 and 20 kDa. The junctions at which cleavage occurs are the
flexible surface loops indicated in the figure as 25/50 kDa. This region was not visible in the crystal structure. Placopecten MHC sequences are compared
to MHC sequences from Argopecten striated [14] and Argopecten catch [15] muscles, third and fourth sequences from top, respectively. '.' indicates
differences in spacing. The number of amino acids is 1941, 1950, 1938 and 1951 for Placopecten striated, Placopecten catch, Argopecten striated and
Argopecten catch muscle MHCs, respectively. Reproduced with permission (Perreault-Micale et al., J Muscle Res Cell Motil 17: 543-553, 1996).
 61

Phosphorylation of RLC smoA by chicken gizzard LC kinase:

no 5 I 2 3 4
ATP min hr hrs hrs hrs

unP
P

Phosphorylation (%) 0 61 66 75 73 77

Fig. 4. Time course of phosphorylation of Placopecten smoA RLC isoform by chicken gizzard myosin light chain kinase as shown on a 12.5% urea
polyacrylamide gel. The extent of phosphorylation was determined by densitometry. 'No ATP' refers to the absence of ATP, but this sample contained all
other reagents and was incubated for the entire 4 h. The kinase concentration was 12 I!g/ml. unP = unphosphorylated, P = phosphorylated, min = minutes, hr
= hours.

Acknowledgments References
We would like to thank Elizabeth O'Neall-Hennessey for I. Szent-Gyorgyi AG, Chantler PD: Control of contraction by calcium
useful comments on the manuscript and to Rita Purcell for her binding to myosin. In: A.G. Engel, C. Franzini-Annstrong, C. (eds).
help with the preparation of this manuscript. This work was Myology, McGraw-Hill, New York, 1994, pp 506-528
2. KalabokisVN, O'Neall-Hennessey E, Szent-GyorgyiAG: Regulatory
supported by NIAMSINIH Grants ARl5963 and AR41808 domains ofmyosins: Influence of the heavy chain on calcium binding.
and a Muscular Dystrophy Association grant to A.G.S.-G. J Muscle Res Cell Motil 15: 547-553,1994
3. Xie X, Harrison D, Schlichting I, Sweet RM, KalabokisVN, Szent-
Gyorgyi AG, Cohen C: Structure of the regulatory domain of scallop
Table 2. Actin-activated ATPase activities of RLC hybrids with myosin at 2.8 A resolution. Nature 368: 306-318, 1994
unphosphorylated and phosphorylated smoA RLC 4. Houdusse A, Cohen C: Structure of the regulatory domain of scallop
myosin at2 A resolution: Implications for regulation. Structure 4: 21-
-Calcium +Calcium Sensitivity 32, 1996
(I!mol/min/mg) (I!mol/min/mg) (%) 5. Fromherz S, Szent-GyorgyiAG: Role of essential light chains EF hand
Placo catch 0.005 ± 0.003 0.18 ± 0.02 96 + 2 domains in calcium binding and regulation of scallop myosin. Proc
Desensitized 0.057 ± 0.0 I 8 0.058 ± 0.02 -I ± 8 Natl Acad Sci USA 92: 7652-7656, 1995
6. Jancso A, Szent-Gyorgyi AG: Regulation of scallop myosin by the
Hybrids resensitized with RLC from: regulatory light chain depends on a single glycine residue. Proc Natl
Argopecten 0.016 ± 0.005 0.17 ± 0.01 91 ± 2 Acad Sci USA 91: 8762-8766, 1994
Unphosphorylated smoA 0.018 ± 0.007 0.14 ± 0.03 87 ± 7 7. Stafford WF, Szentkiralyi EM, Szent-Gyorgyi AG: Regulatory
Phosphorylated smoA 0.013 ± 0.006 0.16 ±0.04 91 ±4 properties ofsingle-headed fragments of scallop myosin. Biochemistry
18: 5273-5280, 1979
These hybrids were made using desensitized Placopecten catch myofibrils. 8. Kalabokis VN, Vibert P, York ML, Szent-Gyorgyi AG: Single-
We also tested the unphosphorylated and phosphorylated smoA RLC on headed scallop myosin and regulation. J Bioi Chern 27 I: 26779-
hybrids made withPlacopecten striated and Argopecten striated myofibrils 26782, 1996
and obtained similar results. Activities were measured using a pH stat by 9. Wells C, Bagshaw CR: The calcium sensitivity of the actin-activated
following the rate of proton liberation at pH=7.6 in 10 ml of2 mM MgATP, ATPase ofscallop heavy meromyosin. FEBS Lett 168: 260-264, 1984
20 mM NaCI, I mM MgCI, and O. I mM EGTA. After recording the ATPase 10. Morita F, Kondo S: Regulatory light chain contents and molecular
activity in the presence of EGTA, 0.2 mM Ca2+ was added. Assays were species of myosin in catch muscle of scallop. J Biochem (Tokyo) 92:
done using 2 mg myofibrils and I mg actin at 23°C. Ca2+ sensitivity was 977-983, 1982
calculated from (I -ATPase EGTA/ATPaseC,";,m) x 100. Values represent the II. Perreault-Micale CL, Jancso A, Szent-Gyorgyi AG: Isoforms of the
mean and standard deviations of three independent experiments, each with essential and regulatory light chains of Placopecten striated and catch
duplicate measurements. muscle myosins. J Muscle Res Cell Motil 17: 533-542, 1996
62

12. Perreault-Micale CL, Kalabokis VN, Nyitray Land Szent-Gy6rgyi phosphorylation. A molecular mechanism for catch contraction. J
AG: Sequence variations in the surface loop near the nucleotide binding Biochem (Tokyo) 104: 102-107, 1988
site modulate the ATP turnover rates of molluscan myosins. J Muscle 18. Sobieszek A, Small JV: Regulation of the actin-myosin interaction in
Res Cell Motil 17: 543-553, 1996 vertebrate smooth muscle: activation via a myosin light-chain kinase
13. Rayment I, Rypniewski WR, Schmidt-Base KS, Smith RS, Tomchick and the effect of tropomyosin. J Mol Bioi 112: 559-576, 1977
DR, Benning MM, Winkelmann DA, Wesenberg G, Holden HM: 19. PersechiniA, Stull H, Cooke R: The effect of myosin phosphorylation
Three-dimensional structure of myosin subfragment-I: A molecular on the contractile properties of skinned rabbit skeletal muscle fibers. J
motor. Science 261: 50-58, 1993 Bioi Chern 260: 7951-7954,1985
14. Nyitray L, Goodwin EB, Szent-Gyorgyi AG: Complete primary 20. Kemp BE, Pearson RB: Spatial requirements for location of basic
structure ofa scallop striated muscle myosin heavy chain. J Bioi Chern residues in peptide substrates for smooth muscle myosin light chain
266: 18469-18476, 1991 kinase. J Bioi Chern 260: 3355-3359, 1985
15. Nyitray L, Jancso A, Ochiai Y, Graf L, Szent-Gy6rgyi AG: Scallop 21. Castellani L, Cohen C: Myosin rod phosphorylation and the catch state
striated and smooth muscle myosin heavy chain isoforms are produced of molluscan muscles. Science 235: 334-337, 1987
by alternative RNA splicing from a single gene. Proc Natl Acad Sci 22. Sohma H, Inoue K, Morita F: A cAMP-Dependent regulatory protein
USA 91:12686-12690,1994 for RLC-a myosin kinase catalyzing the phosphorylation of scallop
16. Riiegg JC: Smooth muscle tone. Physiol Rev 51: 201-248, 1971 smooth muscle myosin light chain. J Biochem 103: 431--435, 1988
17. Takahashi M, Sohma H, Morita F: The steady state intermediate of 23. Kalabokis VN, Szent-GyorgyiAG: Cooperativity and regulation ofscallop
scallop smooth muscle myosin ATPase and effect of light chain myosin and myosin fragments. Biochemistry 36: 15834-15840, 1997
Molecular and Cellular Biochemistry 190: 63-66, 1999.
© 1999 Kluwer Academic Publishers.

The properties and function of invertebrate new

muscle protein

Yoichi Yazawa and Mika Kamidochi

Department ofNutritional Physiology, Hokkaido University ofEducation at Asahikawa, Hokkaido, Japan

Abstract
We found out a new protein from natural actomyosin prepared from adductor muscle of Hokki clam, bivalve shell. We isolated
this protein and determined some properties. It has a large molecular weight (230 kDa) and the star diagram of amino acid
composition was very similar to that of paramyosin (110 kDa). When this protein was added to Hokki clam myosin, the
Mg 2+ -ATPase activity was more activated in the presence of 10-7 M Ca 2+ and further inhibited in the presence of 10-4 M Ca2 + as
compared with those of myosin. From these results, we suggest that Hokki clam adductor muscle contains another myosin-I inked
regulatory protein, myonin, which is different from the myosin-linked system, the myosin light chain-linked system. We named
this protein 'myonin'. (Mol Cell Biochem 190: 63---66, 1999)

Key words: Hokki clam, adductor muscle, paramyosin, myonin, regulatory protein

Introduction those results, we suggested that this protein might be a new

myosin-linked regulatory protein.
In vertebrate striated muscles, the actin-myosin interaction
is regulated by troponin-tropomyosin system dependent on
the Ca 2+ concentration. These regulatory proteins, troponin Materials and methods
and tropomyosin, are located on the thin filaments. [1, 2] In
contrast, we reported molluscan actomyosin ATPase is Adductor muscles were obtained from live, Hokki clam (surf
regulated by the dual Ca 2+ regulation system [3,4], the clam = Supisula Pseudocardium sachalinensis) caught in the
myosin-linked system [5, 6] and an actin-linked system [3, Sea of Okhotsk. Hokki clam natural actomyosin and myosin
7-9], where the myosin-linked system has been regarded as were prepared from washed adductor muscle according to the
essential (Y. Yazawa et ai, to be published). procedure described previously by us.[ 10] Mg2+ -ATPase acti-
Although molluscan smooth muscle consists of thick and vities were measured in solutions containing 50 mM KCI, 10
thin filaments, certain features of the filament structure, mM MOPS (pH 6.8), 2 mM MgCI 2 and 2 mM ATP with
protein composition and physiological behaviour of mollus- various Ca 2+ concentrations. Ca 2+ sensitivities (%) were
can smooth muscles are distinct from those of striated measured by comparing the ATPase activities in the presence
muscle. of 0.1 ~M Ca 2+ and 0.1 mM Ca 2+ by using the following
Molluscan smooth muscles contain large amounts of an relationship:
unique protein, paramyosin, which gives rise to the period-
icities observed in the thick filaments. ATPase (I x I Q-4 MCa 2.) - ATPase (I x 10-' MCa'·)
Ca'· sensitivity = x 100
The location of myosin in these muscles has not been ATPase (I x 10-4 MCa2+)
established, although it has been suggested that myosin may
be present at the surface of the thick filaments. The liberated Pi was determined according to Fiske and
In this paper, we isolated a new protein from Hokki clam SubbaRow method with a slight modification.[11] The free
adductor muscle and characterized some properties. From Ca 2+ concentration was calculated using the binding constant

Address for offprints: Y. Yazawa, Department of Nutritional Physiology, Hokkaido University of Education at Asahikawa, Hokkaido 070, Japan
64

ofCa2+to EGTAreported by Ogawa [12]. SDS-PAGE (Poly- contained in myosin fractionated with 32% saturated ammo-
acrylamide gel electrophoresis) was performed according to nium sulfate but actin and tropomyosin slightly contaminated
the method ofLaemmli.[ 13] Gels were stained with Coomassie (Fig. I, lane 2)
brilliant blue (CBB). The density of stained bands was The precipitate fractionated with 32% saturated ammonium
measured at 600 nm for CBB using a Shimadzu CS-900 sulfate was collected and further purified by three repetition
dual-wave-length TLC scanner. of the dilution and precipitation procedure. The precipitate
Amino acid analysis was carried out essentially by the was collected and suspended by adding water. After the
method of Moore and Stein[14] with a Hitachi amino acid suspend was dialyzed in 6 M Urea, I mM EDTA, 0.1 mM
analyzer (Model KLA-3B). Samples of the enzyme were 2-mercaptothanol and 10 mM Tris-HCl (pH 7.6), it was
hydrolyzed in constant boiling HCl (6N) in evacuated, applied to DEAE-Toyopearl column and eluated by using
sealed tubes for 24 h. After evaporation ofHCl hydrolyzates NaCllinear gradient. As shown in Fig. 2, the elution pattern
were dissolved in pH 2.2 solution and analyzed. Total showed two major peaks. As shown in Fig. 3, the peak eluted
half-cystine was determined by measuring cysteic acid at 0.1 M NaCl contained myonin and minor contaminants
produced by performic acid oxidation. [15] Tryptophan (paramyosin and actin). Actin and tropmyosin could be
content was determined from the absorption spectrum of the removed by Sepharose 6 B gel column in 5 M Guanidine
enzyme dissolved in 6M guanidine-HCl, pH 6.5, as de- hydrochloride (data not shown).
scribed by Edelhoch.[16] The protein concentration was Table I shows amino acid compositions of Hokki clam
determined by biuret reaction and also by the absorbance myonin estimated by us and compared with that of para-
at 280 nm. myosin reported by other group. [17] Myonin consisted of
2,0 I0 amino acids and Asx, Glx, Ala, Leu and Arg contents
were higher than those of other amino acids and Pro was
Results and discussion scarecely contained. The amino acid composition of myonin
was very similar to that of paramyosin, though the marked
Figure I shows the pattern of SDS-PAGE of Hokki clam differences were found in their molecular weights and total
NAM and myosin. NAM was consisted of well-known amino acids numbers.
several proteins (myosin heavy chains, paramyosin, actin Effects of myonin on Mg2+-ATPase activities of Hokki
and tropomyosin etc) and unknown new protein. The new
component could be observed in large amounts upper the
band of myosin heavy chain and its molecular weight was
estimated to be 230 kDa, The protein content was estimat-
ed to be 25% from the densitometer tracings at 600 nm. We 2.0
called this new protein, 'myonin', because that binds to
myosin under the physiological conditions. Myonin wasn't

1 2

-
~

o
01.0
co
N ..- Q3 i
<t:
MN. . . .
HC- u
o
Q1 z
PM...
40

A-
TM- Fig. 2. Chromatography of the fraction precipitated with 32% saturated
ammonium sulfate of Hokki clam NAM. The precipitate of NAM
fractionated with (NH.) ,SO. (154 mg) was collected and dialyzed against 6
Fig. 1. SOS-PAGE patterns of Hokki clam natural actomyosin (NAM) M urea, I mM EOTA, 1.5 mM 2-mercaptoethanol, \ 0 mM Tris-HCI buffer
and myosin. Electrophorsis was carried out using 5% polyacrylamide gel (pH 7.6) and applied to a OEAE-Toyopearl column (2.\ x 8 cm) and eluated
containing 0.1 % SOS. Lane I, natural actomyosin (NAM). Lane 2, myosin. by using a linear gradient formed for the dialysis solution and 0.3M NaCI
MN, HC, PA, A, and TM represent myosin heavy chain, paramyosin, actin, containing the same solution. NaCl concentration is shown as a dotted line.
and tropomyosin, respectively. Each fraction is 4.27 ml.
 65

1 2 3 4 5 6 7 8 9101112
M~ -MN c

-He E 0.3
HC- en
E
"-

PM- ------ a..

(5
0.2
E
:::x....
>-

->
u
0
0.1
C1I
Vl
0
a..
J-
<tI
Fig. 3. SOS-PAGE patterns (6% polyacrylamide gel) of Hokki clam myonin 0-
:2:
fractions. Lane I, the fraction of Hokki clam natural actomyosin precipi-
tated between 32-55% saturated ammonium sulfate. Lane 2, Hokki clam o 7 6 5 4
natural actomyosin(NAM). Lane 3, the fraction of Hokki clam NAM pea
precipitated with 32% saturated ammonium sulfate. Lanes 4-12 were tube
numbers, 12, 15,21,23,25,26,28,34 and 38 eluated by OEAE-Toyopearl Fig. 4. Effects of Hokki clam myonin on Mg2+-ATPase activities ofHokki
column chromatography, respectively. clam myosin at varions Ca 2+concentrations. Reaction mixture contained 2
mM MgCI 2, 10 mM MOPS buffer (pH 6.8), and 2 mM ATP at 25°C. 0,
0.2 mg/ml ofHokki clam adductor muscle myosin;~, 0.2 mg/ml ofHokki
clam myosin were investigated (Fig. 4). Mg2+-ATPase clam myosin and 0.03 mg/ml of Hokki clam myonin.
activities of Hokki clam myosin were 0.042 and 0.282 f..lmol
Pi/mg-min in the presence of 1.1 x 10-7 M and 1.0 x 10-4 M
Ca2+, respectively, and the Ca 2+sensitivity was 55.5%. The
Ca 2+, sensitivity of natural actomyosin (NAM) was 93.5%
Table I. Amino acid composition of myonin and paramyosin
(data not shown).
mol/mol protein*' Residues/l 00 Residues
When Hokki clam myonin was added to myosin, Mg2+_
Amino Acid Hokki clam Venus clam Hokki clam Venus clam ATPase activity at 1.1 x 10-7 Ca2+ was 0.081 f..lmol Pil
myonin paramyosin myonin paramyosin mg.min and that at 1.0 x 10-4 M Ca 2+ was 0.182 f..lmol Pil
mg.min. The Ca 2+sensitivity was 55.5%, which was lower
Asx 258 114 13.0 13.9
Thr 91 36 4.5 4.4 than that of myosin. This low Ca 2+ sensitivity of myosin
Ser 105 39 5.2 4.11 added myonin is probably due to the myosin-linked regula-
Glx 420 169 21.0 20.7 tion by myosin.
Gly 42 15 2.1 1.8 Szent-Gyorgyi et al.[ 18] showed that scallop myosin has two
Ala 235 108 l1.7 13.2
kinds oflight chains, -SH light chain and regulatory light chain
Val 69 28 3.4 3.4
Met 27 II 1.3 1.4
which is a Ca2+receptor to switch contraction ofscallop muscle.
lie 83 22 4.1 2.7 They also showed that EDTA treatment brought reversible
Leu 235 106 11.7 13.0 removal of regulatory chain with accompanying loss of Ca2+
Tyr 39 18 1.9 2.2 sensitivity. Previously we reported the presence ofactin-linked
Phe 19 6 0.9 0.7
system in scallop striated muscle in addition to myosin light
Lys 145 59 7.2 7.2
His 9 4 0.4 0.5
chain-linked system. Our results in this paper confirm the
Arg 191 81 9.5 9.9 presence ofanother myosinlinked regulatory system in addition
Pro -{) 1.5 -0 0.2 to myosin light chain-linked regulatory system reported by
CYS*2 39 NO 1.9 NO Szent-Gyorgyi et al.[18] We could restore the Call sensitivity
Trp*J 3 NO 0.2 NO of myofibril or natural actomyosin by reconstituting the
Total 2,010 817.5 100.0 100.0
actomyosin system of myosin and thin filaments in scallop
*'Values are mol of amino acid/mol of myonin or paramyosin. The striated muscles. When myonin was added to Hokki clam
molecular weight of myonin was estimated to be 230 kOa according to myosin, the Call sensitivity was lower than those ofNAM and
the pattern of SOS-PAGE and the values for Venus clam (Mercenaria myosin (Fig. 4).
mercenaria) were derived Komintz et al. [17]. The molecular weight of
It is well-known that many kinds of invertebrate muscles
paramyosin was estimated to be 110 kOa. *20etermined as cysteic acid
after performic acid oxidation and hydrolysis in 6N HCI for 18 h [15]. contain paramyosin in large amounts but the physiological
*JOetermined by the spectrophotometric method of Edelhoch[l 6]. role is still unkown. We consider that myonin might consist
66

of two domains of paramyosin type and the junction region 6. Lehman W, Szent-Gyiirgyi AG: Regulation of muscular contraction.
might be sensitive to any protease contained in adductor Distribution ofactin control and myosin control in the animal kingdom.
J Gen Physiol66: 1-30,1975
muscle because myonin and paramyosin were very similar 7. Lehman W, Head JF, Grant PW: The stoichiometry and location of
in amino acid composition (Table 1) in spite of the large troponin-I and troponin C-like proteins in the myofibril of the bay
difference of the molecular weights. scallop, Aequi-Pecten irradiens. Biochem J 187: 447-456, 1980
Our results have shown the existence of a new myosin- 8. Lehinan W: Thin filament-linked reguration in mollascan muscles.
linked system in addition to the myosin light chain-linked Biochim Biophy Acta 668: 349-356, 1981
9. Ojima T, Nishita K: Troponin from Akazara scallop striated adductor
system and actin-linked system in Hokki clam adductor muscles. J Bioi Chern 261: 16749-16752, 1986
muscle. This strongly suggests that the contraction of Hokki 10. Asakawa T, Yazawa Y, Azuma N: Light chains of Abalone myosin.
clam adductor muscle has triple-regulation systems, two UV absorption difference spectrum and resensitization of desensitized
kinds of myosin-linked system and one actin-linked system. scallop myosin. J Biochein 89: 1805-1814,1981
This seems applicable also to other molluscan muscles con- II. Fiske CH, SubbaRow Y: The colourimetric determination of phos-
phorus. J Bioi Chern 66: 375-400, 1925
taining myonin. It is therefore likely that the triple regulated 12. Ogawa Y: The approved binding constant of glycoletherdiamine
system is common among invertebrates and concerned in tetraacetic acid for calcium at neutral PH. J Biochem 64: 255-257,1968
catch contraction. [19] Further studies are now in progress. 13. Laemmli UK: Cleavage of structural proteins during the assembly of
the head of bacteriophage T4. Nature 277: 680-685,1970
14. Moore S, Stein WH: Chromatographic determination of amino acids
by the use of automatic recording equipment. Meth Enzymol 6: 819-
References 831,1963
15. Moore S: On the determination of cystine as cysteic acid. J Bioi Chern
I. Ebashi S, Ebashi F: A new protein component particpating in the super- 283: 235-237,1963
precipitation of myosin B. J Biochem 55: 604-613, 1964 16. Edelhoch H: Spectroscopic determination of tryptophane and tyrosine
2. Ebashi S, Endo M: Calcium ion and muscle contraction. Prog Biophys in proteins. Biochemistry 6: 1948--1954, 1967
Mol BioI 18: 123-183,1968 17. Kobo S, Kitagawa S, Tonoinura Y: Molecular biology of muscle
3. Yazawa Y: Actin-linked regulation in scallop striated muscle. Proc proteins. In: M. Kotani et al. (cds). Molecular Biology. Asakura Shoten,
Japan Acad 61 Ser B: 497-500, 1985 Tokyo,1963,pp.253-294
4. Lehman W: The distribution oftroponin-like proteins on thin filaments 18. Szent-Gyiirgyi AG, Cohen C, Kendrick-Jones J: Paramyosin and
of the bay scallop, Aeqipecten irradiens. J Muscle Res Cell Motil4: filaments of molluscan 'Catch' muscle II. Native filaments: Isolation
379-389, 1983 and characterization. J Mol Bioi 56: 239-258, 1971
5. Lehinan W, Kendrick-Jones J, Szent-Gyiirgyi AG: Myosin-linked 19. Ruegg JC: On the tropomyosin-paramyosin system in relation to the
Regulation Systems; Contraction Studies. Cold Spring Harbor Symp visco-ustone of lamellibranch 'Catch' muscle. Proc Roy Soc London
Quant BioI 37: 319-330,1972 B154: 224-249, 1961
Molecular and Cellular Biochemistry 190: 67-74, 1999.
© 1999 Kluwer Academic Publishers.

Actin binding proteins that change extent and rate

of actin monomer-polymer distribution by
different mechanisms

Annemarie Weber
Department ofBiochemistry and Biophysics, University ofPennsylvania, Philadelphia, PA, USA

Abstract
Actin binding proteins control actin assembly and disassembly by altering the critical concentration and by changing the kinetics
ofpolymerization. All ofthese control mechanisms in some way or the other make use ofthe energy of hydrolysis ofactin-bound
ATP. Capping of barbed filament ends increases the critical concentration as long as ATP hydrolysis maintains a difference in
the actin monomer binding constants of the two ends. A further increase in the critical concentration on adding a second cap,
tropomodulin, to the other, pointed filament end also requiresATP hydrolysis as described by the model presented here. Changes
in the critical concentration are amplified into much larger changes ofthe monomer pool by actin sequestering proteins, provided
their actin binding equilibrium constants fall within a relatively narrow range around the values for the two critical concentrations
of actin. Cofilin greatly speeds up treadmilling, which requires ATP hydroysis, by increasing the rate constant of
depolymerization. Profilin increases the rate of elongation at the barbed filament end, coupled to a lowering of the critical
concentration, only if ATP hydrolysis makes profilin binding to the barbed end independent of its binding constant for actin
monomers. (Mol Cell Biochem 190: 67-74, 1999)

Key words: actin, actin depolymerizing factor, actin monomer pool, actin monomer sequestration, ATP hydrolysis, capping
proteins, cofilin, critical concentration changes, coordinated calcium regulation, depolymerization rate constant, elongation
rates, Ebashi, lamellipodia, Listeria monocytogenes, leucocytes, myosin heads, platelets, profilin, thymosin-~4' Thyone acrosome,
treadmilling, troponin, tropomodulin

Introduction is called negative control: during relaxation inhibition of actin

- myosin interaction by the 'soluble relaxing factor', its action
Ebashi, among his many great contributions, also ushered in overridden by calcium at the start of contraction. Ironically,
the era of signalling. For this achievememnt he never received this basic concept was absolutely correct. What was wrong
the recognition he deserves. Everybody knows that Ebashi was the role assigned to the only 'relaxing factor' then
provided crucial information to the understanding of the known. This misassignment was caused by ignorance of the
regulation ofmuscle contraction. Everybody knows that Ebashi fact that all reagents, all muscle extracts, troponin, always
discovered troponin. Very few know that he showed that contained calcium, the ubiquitous contaminant of all solu-
calcium is the signal which coordinates contraction with tions and tissue extracts. Only the soluble relaxing factor
metabolism. At this point, I would like to digress for a moment responded to calcium addition. Consisting, as recognized
to consider briefly the status of the regulation of actomyosin later, of the vesicles ofthe sarcoplasmic reticulum, its calcium
contraction at the time of Ebashi 's entry into the field. It is storage capacity well exceeded the usual amounts of con-
an amusing example of how ignorance of one small fact can taminating calcium.
confuse the interpretation of the experimental evidence. At Ebashi's first step towards elucidating the mechanism of
the time, contraction was thought to be under what nowadays regulation consisted of showing that the soluble relaxing

Address for offprints: A. Weber, Department of Biochemistry and Biophysics, University of Pennsylvania, Philadelphia, PA 19104, USA
68

factor was the Meyerhof Kielley ATPase, a membrane actin tails depend on the dynamics of actin monomer-polymer
fraction. His next step was to show that this membrane interactions.
fraction bound calcium in an ATP-dependent manner. This Actin filaments in the first group should be stable. Gen-
was a new concept and aroused general excitement. Later on, erally, these actin filaments are crosslinked and sometimes
advances in biochemical techniques, among them the newly capped on both ends. Capping proteins protect against
emerging calcium chelating agents (initially commercially elongation and depolymerization at the capped filament ends.
available only from Japan), calcium indicators and the By contrast, actin functions depending on the dynamic
introduction of radioactive calcium led to the recognition of interactions between monomers and polymers are very
the ubiquitous calcium contamination and thus helped in sensitive to the free monomer concentration. Actin filaments
preparing the way towards finding troponin, the true actin- assemble or disassemble with rising or failing monomer
bound relaxing factor. concentration. In the following, I shall survey various
The emergence of a complete regulatory mechanism was mechanisms by which actin binding proteins alter the free
very exciting at a time when the scientific focus began to shift actin monomer concentration, the actin distribution between
from exclusive attention to energy and work - energy coupling monomers and polymers, and the rates of actin filament
towards basic concepts of regulation. Allostery and co- assembly and disassembly.
operativity had emerged as major topics of discussion.
However, the concept of using one signal to coordinate
distinctly different functions, like contraction and meta- Definition of the steady state free
bolism, did not yet exist. Rather one thought of first
switching on actomyosin contraction followed by scaling
monomer concentration as the critical
up ATP resynthesis in response to the accumulaton of ATP concentration
degradation products. Thus phosphorylase was thought to
be activated in vivo by the allosteric action of increasing Monomer-polymer equilibrium
concentrations of AMP and glycolysis by increasing ADP
combined with decreasing levels ofATP. Ebashi recognized When a monomer - polymer system consists of identical
the advantage of using the same signal to simultaneously molecules, the polymer is in equilibrium with a single
activate metabolism and contraction. Thus, he investigated monomer concentration, the critical concentration ([ I],
the calcium response of the phosphorylase-phosphorylase review). This concentration is independent of the number of
kinase system and found that calcium activated glucose polymers according to the expression: [monomer][end] /[end]
release from glycogen in the same concentration that turned = Kd; [monomer] = Kd. The critical concentration reflects
on contraction by binding to troponin. What makes this the binding constant of the monomer for the polymer. For
finding special is that Ebashi did not happen to discover the polymers that react with monomers only at their ends, this
calcium activation of glycogenolysis by chance but was led binding constant is the same for both ends, even if the ends
to it by his intuition for the big biological pattern. differ from each other because the monomer domains exposed
In addition to troponin Ebashi also discovered other actin at the filament ends are different
binding proteins, the topic of this review.

Monomer-polymer steady state associated with AT?

Actin binding proteins that alter the hydrolysis

monomer-polymer distribution of actin In the presence of ATP, actin molecules are no longer
identical. This is the result ofATP hydrolysis after monomer
Actin functions fall into two groups: functions that are carried incorporation into filaments. The free actin monomers consist
out by stable actin filaments and functions depending on of ATP-actin and the interior of the filaments consists of
monomer-polymer transitions. Stable actin filaments are ADP-actin. The terminal actin molecules at the filament ends
partners in force generation in the sliding filaments ofmuscle, may be at intermediate stages ofATP hydrolysis, containing
stress fibers or during organelle transport (such as mito- boundATP or ADP.P or ADP ([2], review). The distribution
chondria transport into the yeast bud). Stable crosslinked between ATP-, ADP.P- or ADP-actin depends on how
actin meshworks support cell membranes and, when tightly rapidly new ATP-actin monomers are added to the filament
crosslinked into bundles in the stereocilia, they are the end. For instance, if new monomers are added to the ends
receptors for sound. very fast, before the ATP of the previously incorporated
On the other hand, cell motility associated with lamellipodia actin molecules had time to be even hydrolyzed, the ends
formation or the intracellular motility of bacteria driven by will containATP-actin. However, the rate ofATP hydrolysis
 69

under physiological salt conditions is very fast [2] and Regulation of the critical concentration
ATP-actin ends are not likely unless the concentration of
unpolymerized actin is high, much higher than the steady
by actin binding proteins
state monomer concentration. In contrast to the hydrolytic
step, the release of phosphate from theADP.P intermediate Binding proteins that increase the critical concentration
of ATP hydroysis is quite slow. Thus, one may find either
ADP.P in the terminal actin molecules at intermediate Proteins that cap the barbed filament ends
elongation rates or ADP when the rate of monomer addition The steady state G-actin concentration ofa treadmilling actin
is low. filament (coo = (k-b + k-p) / (k+ b + k+ p)) is close to the critical
When the actin molecules are no longer identical but concentration for the barbed end (The barbed end pre-
contain different nucleotides, the critical concentration is dominates because it has higher rate constants than the
defined by the ratio of the rate constant of actin dissociation pointed end). Capping the barbed filament end restricts
from and the rate constant ofactin binding to the filament end: monomer-polymer exchange entirely to the pointed end and
k_ / k+ rather than by an equilibrium constant. therefore significantly increases the critical concentration
After polymerization is complete, at steady state, the (from the range between 0.05 and 0.1 !!M to the range
barbed filament ends consist mostly ofADP.P-actin and the between 0.5-0.7 !!M).
pointed ends largely of ADP-actin. The reason for the Capping an increasing fraction of barbed filament ends
difference is that the rate constans for ATP-actin monomer results in an abrupt increase in the critical concentration when
addition (on-rate constant, kJ differ about IO-fold between the extent of capping exceeds 90% [5]. At a lower extent of
the two ends ([3], review). (This difference in the absolute capping the critical concentration increases relatively little
value of the rate constants is intrinsic to the filament ends: (Because their on-rate constant is so high, a relatively small
even when all actin molecules are identical, containing fraction offree barbed ends can keep the steady state G-actin
nothing butADP, the on-rate constant is lower at the pointed concentration at a low level). Conversely, uncapping a small
- than at the barbed filament end. However, when actin fraction of barbed filaments results in a large drop in the
consists of identical ADP-actin molecules the ratio of the rate critical concentration.
constants, k_ / k+, is the same at the two ends). Because of A large number of barbed end capping proteins exist in
the difference in rate constants,ATP-G-actin binds to pointed cells. Most important among them appears to be Cap Z [6]
filament ends frequently after phosphate release and to which is present in all cells, from yeast to muscle.
barbed ends usually before phosphate release.
The ratio of the rate constants, k_ / k+, which defines the
critical concentration, is determined by the actin species at Tropomodulin
the filament end, it is lower for ADPP-actin than for ADP-
actin. Thus, the critical concentration is lower for the barbed Under certain conditions, the critical concentration of
filament ends than for the pointed ends. barbed-end capped actin is increased further when the pointed
With a different critical concentration for each filament end filament ends are also, partially, capped by tropomodulin [7].
[4], actin filaments undergo net depolymerization at the The necessary conditions are first, the presence of ATP (as
critical concentration for the barbed ends and net poly- for the effect of barbed end capping) and second, the absence
merization at the critical concentration for the pointed ends. of tropomyosin (Tropomyosin does not interfere with the
Steady state, i.e. no change in G-actin and F-actin over time, increase in critical concentration caused by barbed-end
is reached at the G-actin concentration where the rate of net capping.). The tropomodulin-induced increase in the critical
elongation at the barbed ends is equal to the rate of net concentration can be explained by its inhibitory effect on
depolymerization at the pointed ends. This steady state ATP-actin addition to the pointed filament ends, as will be
G-actin concentration is higher than the barbed end - and discussed next.
lower than the pointed end critical concentration. At this Tropomodulin, recently discovered by Fowler ([8, 9],
intermediate G-actin concentration, the critical concentration review), is the only known intracellular pointed end capping
for uncapped actin, the actin filaments undergo a process protein. It binds exclusively to actin molecules at the pointed
called treadmilling. Treadmilling has the interesting feature filament ends and does not bind actin monomers or actin
that the whole filament is constantly moving in the direction molecules inside the filaments. Tropomodulin is found in
of its barbed end although each single actin molecule in the muscle and in some non-muscle tissues such as red cells, brain,
filament stays in the same spot: New actin molecules are lens. Tropomodulin caps the pointed ends of tropomyosin-
added to the barbed end while the pointed end is dismantled containing actin filaments with high affinity (Kd<nM), while
at the same rate [4]. its affinity for tropomyosin-free filaments is relatively low
for a capping protein (Kd about 0.2 !!M). It raises the critical
70

concentration of tropomyosin-free filaments at the pointed contain at any moment in time 50% tropomodulin-capped
ends (gelsolin-capped actin filaments) by about a factor of actin filaments and 50% free actin filaments. This, however,
two [7]. does not imply that over a period of time the same filaments
Since the purpose of capping proteins is thought to be are always capped and the same filaments are always free.
blocking the interaction between monomers and filaments Capped filaments would stay capped for extended periods only
one needs to consider the concept of a critical concentration if the capping protein has a very low off-rate constant, like
in conjunction with the capping of both filament ends. Ifboth gelsolin: with a halftime of dissociation somewhere between
filament ends are tightly capped and all interactions between 2 hand 200 h, gelsolin is unable to rapidly move from one
F-actin and G-actin blocked, the concept of a critical con- filament to another, and the gelsolin cap will stay on the same
centration becomes meaningless. Under the conditions of our filament for a long time.
experiments where tropomodulin raised the critical con- By contrast, tropomodulin binds rapidly to tropomyosin-
centration of gelsolin-capped actin filaments, saturation with free actin and with a low binding constant. A half time for
tropomodulin was incomplete and pointed-end monomer- dissociation as short as 0.4 sec cannot be ruled out, i.e. in 4
polymer interaction was not fully blocked. Capping at the sec nearly all of the tropomodulin molecules could have
highest tropomodulin concentrations used was about 80- moved from one pointed filament end to another. This is fast
85%. By contrast, the extent of gelsolin saturation at the compared to the rate of ATP-actin addition to the pointed
barbed filament end was very high. (The nucleating filaments filament ends, with a halftime around 14 sec during steady
were capped initially at gelsolin concentrations of 10 7 X Kd state (in our actin preparations). Thus, from the perspective
(M = 10-12 M- 10- 13 M). of actin monomers, one can look at 50% saturation in terms
Because of the limited extent of tropomodulin saturation of the fraction oftime each filament is capped (unaccessible):
in our experiments, we do not know what would happen to At 50% tropomodulin saturation each filament end is capped
the critical concentration on complete tropomodulin satura- by tropomodulin 50% of the time. This means the rate of
tion. If tropomodulin should completely block monomer- ATP-actin binding to each filament end is reduced by a factor
polymer interaction, fully tropomodulin-saturated actin of two at 50% saturation, by a factor of 5 at 80% saturation
would not have a critical concentration. By contrast, monomer- and by a factor of 10 at 90% saturation. Thus, with increasing
polymer interaction would continue after complete tropo- tropomodulin saturation, the fraction of filament ends con-
modulin saturation if the tropomodulin cap were leaky. verted to ADP-actin is increased, and the critical concentration
Leaky capping means that the on- and off-rate constants are goes up. The critical concentration comes fairly close to the
reduced to a fraction of the control value, but not to zero. value for ADP-actin observed with our preparations (about 1.1
Since monomer-polymer interaction persists in a leaky cap, J.lM). (The ADP-actin critical concentration refers to ADP-G-
although at a very low rate, tropomodulin-saturated actin actin in equilibrium with ADP-actin filament ends, whereas
would have a critical concentration. Because of the very low the tropomodulin - induced critical concentration refers to
rates of monomer-polymer exchange, phosphate would be ATP-G-actin at steady state with ADP-actin filament ends).
released from the terminal actin molecules of virtually all It is not clear whether the tropomodulin-induced increase
filament ends before the addition of a new ATP-actin mole- in the critical concentration of barbed-end capped actin
cule. Therefore, all filament ends would contain only ADP- occurs under physiological conditions since it is restricted to
actin, and the critical concentration would reach its maximal tropomyosin-free actin filaments.
value. Unfortunately, we are not certain whether or not the
tropomodulin cap is leaky in tropomyosin-free actin filaments.
As mentioned above, tropomodulin forming a blocking cap Binding proteins that decrease the critical concentration
(in contrast to a leaky cap) would completely prevent
monomer-polymer interaction, and fully tropomodulin-satu Any binding protein that binds stoichiometrically to actin and
rated actin filaments would not be able to establish a critical prefers F-actin over G-actin will shift actin from monomer
concentration. Nevertheless, at partial saturation tropo- to polymer and lower the critical concentration. This shift
modulin could increase the critical concentration of the does not require ATP hydrolysis. For instance, at low salt,
mixture of capped and uncapped actin filaments by exactly where the critical concentration (at 20 D C) is about 20 J.lM, a
the same mechanism as described for the leaky-cap-forming stoichiometric amount of myosin heads, S-I, will, in the
tropomodulin, i.e. increased conversion of ADP.P-actin to absence of ATP, convert G-actin into F-actin, lowering the
ADP-actin. free monomer concentration to very low values [10]. The
To understand how this is possible one needs to define reason is that S-l binds very tightly (Kd about nM) to actin
partial tropomodulin saturation in terms of time rather than molecules inside the filaments, but S-l does not bind
in terms of fractions of capped and uncapped filaments. measurably to monomeric actin. The free energy of S-1
Partially tropomodulin-saturated actin filaments, e.g. by 50%, binding to actin filaments drives the polymerization until
 71

nearly all G-actin is removed. Since in the presence of ATP sequestering protein, e.g. vitamin D binding protein with a
myosin heads lose their tight binding to actin filaments Kd of about 1 nM, is saturated with actin at 0.5 flM G-actin
myosin does not physiologically affect the critical con- (the critical concentration for capped actin), i.e. it is capable
centration. However, there might be other actin binding of forming a large pool of sequestered actin. Unfortunately,
proteins which have the same effect in the presence of ATP. it is still saturated at 0.1 flM, the critical concentration for
uncapped actin, so that practically none of the actin se-
questered by vitamin D binding protein can be used to
Relationship between critical increase the concentration of polymerized actin. This may be
the reason why vitamin D binding protein is not found inside
concentration and the actin monomer cells. At the other extreme, a hypothetical low affinity protein
pool available for polymerization with a Kd of 5 flM for actin monomers would be only 10%
saturated at 0.5 flM G-actin, i.e. it would form only a very
Cells which need to respond to changes in their environment small monomer pool. Therefore, a low affinity protein also
with an increase in their polymerized actin, such as platelets would not be able to supply enough actin for polymerization,
or leukocytes, require a pool of monomeric actin that can be even though it would release its bound actin in proportion to
used for this purpose. In response to stimulation, leukocytes the decrease in free G-actin.
increase their F-actin content by about 50 flM and platelets, To provide a usable monomer reservoir for polymerization,
probably the most extreme cell in this regard, by nearly 200 the Kd of the sequestering protein for G-actin must be in the
flM. This is 400 x more than the critical concentration of range of the two critical concentrations (for capped and
barbed-end-capped actin (about 0.5 flM). These amounts are uncapped actin filaments). With a Kd of about 0.5 flM, the
drawn from a pool of actin monomers sequestered by saturation of thymosin-~4 with actin is about 50%, and 2/3
monomer-binding proteins. The size of this pool depends on of the complexed G-actin is released when the critical
the value for the critical concentration because the se- concentration changes from 0.5---Q.l flM. A similar amount
questered actin is at equilibrium with free G-actin. of actin for polymerization would be provided by a protein
Platelets and leukocytes contain thymosin-~4 (~4) which with a Kd of 0.1 flM, where a larger amount of sequestered
binds actin monomers (G) according to the equilibrium actin would compensate for a smaller fraction that can be
[G][~4] 1 [~4-G] = Kd (or [~4-G] 1 [~4] = [G] 1 Kd). In a released.
resting platelet, about 112 of the thymosin-~ 4 is complexed
with monomers (~4-G) creating a sequestered monomer pool
ofabout 300 flM thymosin-~4-complexed G-actin [II]. (This Cofilin, strategy to speed up
estimate is based on the assumption that almost all of the treadmilling by increasing the rate
measured amount of monomeric actin is bound to thymosin-
~4 and that in the resting platelet profilin is mostly bound to constant of depolymerization at the
the membrane phospholipid PIP 2 (12]. It can be calculated pointed filament ends
(using the measured Kd of about 0.5 flM (0.4-0.7 flM) and
the measured ratio of total actin and actin-complexed to Cofilin binds stoichiometrically to G-actin and has been
thymosin-~4) that this sequestered monomer pool is in shown to shift actin from the polymeric towards the monomeric
equilibrium with a critical concentration of about 0.5 flM. state. This is of course what every monomer sequestering
When is the monomeric actin released from its complex binding protein does, as for instance thymosin-~4' The action
with thymosin-~4 to increase the amount of polymerized of cofilin, however, is much more complex.
actin? It is obvious from the above discussion that a decrease Cofilins/destrins/actin depolymerizing proteins (ADF) are
in the G - actin concentration would result in the dissociation members of a family of small molecular weight actin binding
of the complex until a new equilibrium has been reached. proteins (about 21 k)([13], review) which first became known
Uncapped actin filaments appearing anywhere in the cell as depolymerizing factors for actin filaments. This was
would take up monomers and lower the G-actin concentration attributed to their ability to both sever actin filaments and
resulting in the gradual dissociation of the thymosin-~4-actin sequester actin monomers. Later, it was found that these
complex. This would continue until the critical concentration proteins also are able to stoichiometrically bind to poly-
for uncapped filaments has been reached (about 0.1 flM). merized actin. Their preference for monomeric or polymeric
How much can the concentration ofF-actin be increased? actin was shown to change with pH. Members of this family
This depends first, on the total concentration of the sequest- are essential proteins, present in all cells [14].
ering protein, and second on its Kd for actin monomers.The Carlier and Pantalone [15] suggested very recently that
latter is best illustrated by considering two different extreme cofilins rapidly depolymerize actin by a mechanism other
ranges of Kd values. First, a high-affinity actin monomer than a combination of monomer sequestration and severing.
72

Unlike monomer sequestering proteins, cofilin in increasing In keeping with a crucial role in lamellipodia movement,
concentrations did not depolymerize increasing amounts of the cofilins are phosphorylationcontrolled [20] and cofilin
F-actin until all actin had been sequestered as monomer. overexpression in Dictyostelium [21] stimulates membrane
Instead, on binding to the actin filaments, cofilin depoly- ruming and movement.
merized maximally about 2 J.lM actin. There was no increase
in G-actin above this level when the cofilin concentration was
increased further. In other words, cofilin apparently raised the Profilin, strategy to convert monomers into polymers at a
critical concentration from O.OS-o.l J.lM to about 2 J.lM by high rate associated with a decrease in the apparent
forming a new cofilin-actin polymer system (the critical critical concentration
concentration of ADP-actin was increased to about 4.S J.lM).
The interesting feature of this cofilin-complexed actin-
polymer system is the acquisition of a new set of rate Profilin interactions with actin
constants. The rate constant of depolymerization at the Profilin has been shown to greatly increase the rate of
pointed end (but not at the barbed end) was found to be filament elongation at low conentrations of free G-actin and
increased on the addition of saturating cofilin. The 20 fold to lower the critical concentration in the process. First, a short
increased rate constant of depolymerization at the pointed overview of the various interactions between proffin and actin
ends resulted in an about 20 fold increase in the rate of ([22], review)
treadmilling at steady state. At steady state, when depoly- Profilin is a small actin binding protein of 12-16 k, found
merization and elongation rates are equal, the matching in all cells and essential for all organisms. Profilin binds to
increase in the elongation rate was achieved by a 20 fold domains 1 and 3 of actin; its actin binding constant (Kd = 0.1
increase in the steady state G-actin concentration, from 0.1 J.lM) is about 4-S times higher than that ofthymosin-~4 (Kd
J.lM to about 2 J.lM (the pointed end critical concentration was = 0.4-0.7 J.lM). Under conditions where all barbed filament
about 3.5 J.lM). ends are capped (see below), the profilin-complexed-actin
The high rate constant of depolymerization at the pointed monomer pool serves as a monomer reservoir for future
ends would account for the rapid disappearance ofpolymerized actin polymerization. This pool is usually much smaller than
actin from lamellipodia [6, 17] and Listeria monocytogenes the thymosin-~4 -sequestered pool since the total con-
actin tails ([ 18], review), which is much faster than compatible centration of profilin is usually much lower (l0-50 J.lM)
with the rate constant for depolymerization of unmodified than the thymosin-~4concentration (100-600 J.lM). Further-
actin. A recent paper from Mitchison's lab [19] is very more, in some cells, such as resting platelets, a large fraction
suggestive of a depolymerizing function of cofilin in vivo. of profilin is thought to be bound to the membrane phos-
The authors report that the tails of Listeria in Xenopos egg pholipid, PIP 2 •
extract are much shorter (also observed by Carlier and Profilin greatly increases the rate of nucleotide exchange
Pantalone, [IS]) and contain about 20 x less actin in the of the actin monomer. In the cell, profilin is probably
presence of indigenousXenopos cofilinlactin depolymerizing responsible for rapidly renewing theATP charge ofthe (ADP
factor than after removal of the cofilin by treatment with containing) actin monomers coming off the pointed filament
inactivating antibody. ends. (Thymosin-~4' which blocks nucleotide exchange,
This finding indicates a depolymerizing action of cofilin. severely discriminates against ADP-G-actin and therefore
This becomes clear when one considers the composition of does not interfere with the nucleotide exchange by profilin).
the Listeria tail. The bulk of the tail consists of barbed-end Profilin also caps the barbed filament end. However, its
capped actin filaments which are depolymerizing. They are affinity for the barbed filament is fairly low (Kd has been
replaced by new filaments which are polymerizing at the estimated to be around 10 J.lM).
beginning of the tail, directly behind the bacterium. The
barbed ends of these filaments are free until they capture a Profilin-actin monomer binding to the filament ends and
capping protein from the medium. The tail is at steady state the lowering ofthe barbed end critical concentration by
when the amount of depolymerizing actin (from the capped profilin
filaments) equals the amount of actin polymerizing into Profilin-actin complexes,just like gelsolin-actin complexes,
filaments behind the bacterium. If the rate constant of cannot bind to the pointed filament ends because the inter-
depolymerization is high, the number ofcapped filaments that acting actin surface is blocked by profilin. But profilin-actin
accumulate before the two rates become equal is small. If the complexes, just like gelsolin-actin complexes, bind to the
rate constant is low, many capped filaments accumulate barbed filament ends. In the case of the gelsolin-actin
before the rate of depolymerization equals the rate of poly- complexes, binding is restricted to the addition of a single
merization, and the Listeria actin tails become longer and complex. After that the filament end is capped by gelsolin
denser. which remains tightly bound even in the absence of calcium.
 73

By contrast, profilin-actin complexes elongate actin filaments concentration for uncapped actin filaments, increased amounts
from their barbed ends because profilin rapidly dissociates of actin are released from the thymosin-~4-actin pool, bound
from the barbed end after incorporation of the actin molecule to profilin (which binds actin 4-7 times more strongly than
into the filament. Since the affinity ofprofilin for the barbed does thymosin-~4) and incorporated into actin filaments.
end is quite low, as mentioned above, profilin caps only when Actually, the observation (in vitro) that profilin decreases the
the free profilin concentration is high. amount ofthymosin-~4 -sequestered actin led Pantalone and
When elongation eventually stops, the concentration of Carlier [23] to the discovery that profilin lowers the critical
free actin monomers is well below the normal value for the concentration.
critical concentration [23]. This profilin-induced lowering of In the cell, the primary function of profilin quite likely is
the critical concentration requires energy input provided by the promotion of rapid actin filament assembly at low free
the hydrolysis of actin-bound ATP. This follows from G-actin concentrations. Tilney and Inoue [25, 26] observed
thermodynamic reasoning and is confirmed by experiment: incredibly fast rates of acrosome elongation in Thyone
when ATP is substituted by ADP, profilin no longer lowers acrosomes. Rates around 20 11m/sec are more than 1000 times
the critical concentration. too high to be accounted for by polymerization of free actin
Without ATP hydrolysis, when all actin molecules are the monomers whose concentration did not exceed the critical
same, containing for instance nothing but ADP, elongation concentration for the pointed filament ends (there was no
stops when equilibrium between the different components has elongation from the pointed actin filament ends in the
been reached at the normal critical concentration for un- acrosomes). (0.5 11M free G-actin limits the filament elonga-
capped actin filaments [24]. The reason is that at equilibrium, tion rate to about 0.017 11m/sec or a little less than 1 11m per
i.e. when all reactions are fully reversible, the binding min.) Tilney [27] proposed that profilin, present in the
constants for all interactions are interdependent. By contrast, acrosome in large amounts equal to its actin content, might
filament elongation by profilin-actin complexes would deliver actin to the barbed filament ends. Pollard and Cooper
continue until all monomeric actin had been exhausted, if the [28] and Kaiser et al. [29] demonstrated in vitro that profilin
affinity of profilin for monomeric actin and the affinity of was able to do just that.
profilin-actin complexes for the barbed filament end were High speed actin filament assembly has been observed in
infinite, while binding offree profilin to the barbed filament other cells: in the lamellipodia of leukocytes [17], and the
ends were zero. This is completely ruled out in a system of actin tails of bacteria and some virus, described most ex-
reversible reactions by the thermodynamically required tensively for Listeria tails [18]. While it has not yet been
interdependence of equilibrium constants. proved, with the exception of the Thyone acrosome, that
Even ATP hydrolysis cannot create such an extreme profilin-actin complexes are responsible for the rapid elonga-
separation ofbinding constants but it can approximate it. The tion in these cases, it can be calculated that there is enough
reason is the irreversible conversion of the ATP-actin profilin in these cells (about 20-50 11M) to produce the
molecule in the profilin complex to anADP.P-actin molecule observed rates of about 15 11m/min (Listeria tails) to 30 or
after binding to the filament end. This irreversible conversion even 60 11m/min (leukocyte speed, platelet filopodia ex-
of one actin species to another allows the affinity of the tensions). With a diffusion-limited on-rate constant of about
profilin-actin complex for the barbed end to be too high (in 10 7 M-I S-I (equal to that of free actin monomers), 50 11M total
terms of reversible reactions) in relation to the affinity offree profilin, forming 42 11M profilin-actin complexes (in equi-
profilin for the barbed end. (The disparity between binding librium with a free monomer concentration of 0.5 11M, the
constants is limited to a factor of 106 M~I, corresponding to critical concentration for pointed filament ends), would add
the free energy of ATP hydrolysis). about 420 actin molecules per sec to each barbed filament
During filament elongation, both, the concentrations of end, resulting in an elongation rate of about 1.2 11m/sec or
profilin-actin complexes and offree actin monomer (both in 72 11m/min.
rapid equilibrium with each other) decrease continously, The Thyone acrosome differs from the other systems by
while the free profilin concentration increases. As a result, its enormous amount of profilin which is equimolar to the
the rate of filament elongation decreases over time, first, actin, and the incredible speed of acrosome elongation.
because there is less profilin-actin and second, because Elongation rates of 20 11m/sec are caused by the addition of
filament capping by profilin increases. Net actin poly- about 7000 actin molecules per sec to the barbed ends which
merization ceases when the rate of net polymerization by requires a concentration of about 700 11M profilin-actin
profilin-actin complexes at the uncapped (by profilin) barbed complexes.
ends equals the rate of net depolymerization of actin at the Compatible with both profilin functions described here,
pointed ends of all actin filaments. profilin deletions generally produce a phenotype deficient in
Since during this process of filament elongation the free essential actin filament structures, such as the actin network
monomer concentration is reduced to below the critical restraining nuclei in the nursing cells of the drosophila egg
74

[30], or the actin cables ofyeast necessary for successful bud 14. Moon AL, Janmey PA, Louie KA, Drubin DG: Cofilin is an essential
fonnation [31], to name just two. component of the yeast cortical cytoskeleton. J Cell Bioi 120: 421-
435,1993
15. Carlier M-F, Laurent V, Santolini J, Melki R, Didry D, Xia G-X,
HongY, Chua WK, Pantaloni D: Actin depolymerizing factor (ADFI
Acknowledgements cofilin) enhances the rate of filament turnover: Implication for
actin-based motility. J Cell Bioi 136: 1307-1322, 1997
16. Theriot JA, Mitchison TJ: Actin microfilament dynamics in locomoting
Supported by NIH program grant to the Pennsylvania Muscle cells. Nature 352: 126-131, 1991
Insitute and by ROI-GM-53029. 17. Zigmond SH: Recent quantitative studies of actin filament turnover
during cell locomotion. Cell Motil Cytoskeleton 25: 309-316, 1993
18. Pollard TD: Actin cytoskeleton. Missing link for intracellular bacterial
motility. CUIT Bioi 5: 837-840, 1995
References 19. Rosenblatt J,Agnew BJ,Abe H, Bamburg JR, Mitchison TJ: Xenopos
actin depolymerizing factor/cofilin (XAC) is responsible for the
turnover of actin filaments in Listeria monocytogenes tails. J Cell
I. Oosawa F, Asakura S: Thermodynamics of the polymerization of BioI 136: 1323-1332, 1997
proteins, Academic Press New York, 1975 20. Agnew BJ, Minamide LS, Bamburg JR: Reactivation of phos-
2. Carlier M-F: Actin: Protein structure and filament dynamics. J Bioi phorylated actin depolymerizing factor and identification of the
Chern 266: 1-4,1991 regulatory site. J BioI Chern 270: 17582-17587, 1995
3. Pollard TD, Cooper JA: Actin and actin binding proteins. A critical 21. Aizawa H, Sutoh K, Yahara I: Overexpression ofcofilin stimulates
evaluation of mechanisms and functions. Ann Rev Biochem 55: 987- bundling of actin filaments, membrane ruffling, and cell movement in
103,1986 Dictyostelium. J Cell BioI 132: 335-344, 1996
4. Wegner A: Treadmilling of actin at physiological salt solutions. J Mol 22. Sohn RH, Goldschmidt-Clemont PJ: At the crossroads of signal
Bioi 161: 607--615,1982 transduction and the actin cytoskeleton. Bioassays 16: 465-472, 1994
5. Young CL, Southwick FS, Weber A: Kinetics of the interaction of a 23. Pantaloni D, Carlier MF: How profilin promotes actin filament
41 - kilodalton macrophage capping protein with actin: Promotion of assembly in the presence ofthymosin-~4' Cell 75: 1007-1014,1993
nucleation during prolongation of the lag period. Biochemistry 29: 24. Pring M, WeberA, Bubb MR: Profilin-actin complexes directly elongate
2232-2240, 1990 actin filaments at the barbed end. Biochemistry 31: 1827-1836, 1992
6. Schafer DA, Cooper JA: Control of actin assembly at filament ends. 25. Tilney LC, Inoue S: Acrosomal reaction of Thione sperm III. The
Ann Rev Cell Dev BioI II: 497-514, 1995 relationship between actin assembly and water influx during extension
7. Weber A, Pennise CR, Babcock Gary G, Fowler VM: Tropomodulin of the acrosomal process. J Cell Bioi 100: 1273-1283, 1985
caps the pointed ends of actin filaments. J Cell BioI 127: 1627-1635, 26. Tilney LC, Inoue S: Acrosomal reaction of Thione sperm II. The
1994 kinetics and possible mechanism of acrosomal process elongation. J
8. Fowler VM: Identification and purification of a novel M, 43,000 Cell Bio191: 820-827,1982
tropomyosin binding protein from human erythrocyte membranes. J 27. Tilney LC, Bonder EM, Coluccio LM, Mooseker MS: Actin from
BioI Chern 262: 12792-12800,1987 Thyone sperm assembles on only one end of an actin filament: A
9. Fowler VM: Regulation of actin filament length in erythrocytes and behaviour regulated by profilin. J Cell BioI 97: 112-124,1983
muscle. Curr Opin Cell BioI 8: 86-96, 1996 28. Pollard TD, Cooper JA: Quantitative analysis of the effect of
10. Detmers P, WeberA, Elzinga M, Stephens RE: 7-chloro-4-nitrobenzeno- Acanthamoeba profilin on actin filament nucleation and elongation.
2-oxa-I,3-diazole actin as a probe for actin polymerization. J Bioi Biochemistry 23: 6631--6641, 1984
Chern 256: 99-105,1981 29. Kaiser DA, Sato M, Ebert RF, PollardTD: Purification andcharacteriza-
II. Weber A, Nachmias VT, Pennise CR, Pring M, Safer D: Interaction tion of two isoforms of acantharroeba profilin. J Cell BioI 102: 221-
of thymosin ~4 with muscle and platelet actin: implications for actin 226, 1986
sequestration in resting platelets. Biochemistry 31: 6179--6185, 1992 30. Cooley L, Verheyen E, Ayers K: Chicadee encodes a profilin required
12. Goldschmidt-Clemont PJ, Machesky LM, Baldassare JJ, Pollard TD: for intercellular cytoplasm transport during Drosophila oogenesis. Cell:
The actin-binding protein profilin binds to PIP, and inhibits its 69: 173-184, 1992
hydrolysis by phospholipase C. Science 247: 1575-1578, 1990 31. Haarer BK, Lillie SH, AdamsAEM, Magdolen V, Bandlow W, Brown
13. Sun HQ, Kwiatkowska K, Yin HL: Actin monomer binding proteins. SS: Purification ofprofilin from Saccharomyces cerevisiae and analysis
CUIT Opin Cell Bioi 7: 102-110, 1995 of profilin-deficient cells. J Cell BioI 110: 105-114, 1989
Molecular and Cellular Biochemistry 190: 75-78, 1999.
© 1999 Kluwer Academic Publishers.

A phosphorus-31 nuclear magnetic resonance

study on the complex of chicken gizzard myosin
subfragment 1 with adenosine diphosphate

Masaru Tanokura and Yoshikazu Suzuki

Biotechnology Research Center, University of Tokyo, Yayoi, Bunkyo-ku, Tokyo, Japan

Abstract
The complex of Mg-ADP with chicken gizzard myosin subfragment 1 (S 1), obtained by the treatment with Staphylococcus
aureus V8 proteinase, was observed with 31PNMR at various temperatures between 0 and 25°C. The signal ofSI·ADPcomplex
was observed at-2 to-3 ppm as a rather broad peak. As compared with the results for rabbit skeletal muscle SI·ADP complex
(Tanokura M, Ebashi S: J Biochem 113: 19-21,1993), the signal was assigned to ~-phosphate ofADP in the SI·ADPcomplex.
The signal of the complex was so broad and weak that the dependences on temperature and magnetic field strength were not
clear. The observation suggests the tight interaction ofS 1 with the phosphate moieties ofADP in the complex and the extremely
anisotropic distribution of electrons around phosphorus nuclei. (Mol Cell Biochem 190: 75-78, 1998)

Key words: myosin, subfragment I, 31p NMR, smooth muscle, chicken gizzard

Introduction of the complexes, i.e., high-temperature form, which is

dominant at higher temperatures, and low-temperature form,
Myosin is a contractile protein of muscles and its ATPase which is dominant at lower temperatures.
plays a key role in the energy transduction of chemical energy In the present study, we studied the complex formation of
of ATP to physical force of contraction [I]. In order to chicken gizzard smooth muscle S I with ADP at the various
elucidate the mechanism of energy transduction in muscle the temperatures by 31p NMR. A very weak signal has been
conformational change of myosin heads accompanying its observed at the chemical shift corresponding to the signal of
ATP hydrolysis has been studied by various methods. Among the ADP in the complex with rabbit skeletal muscle S I. The
them, X-ray crystallographic studies have unambiguously results will be discussed in terms of the environments
indicated that there are two conformations of ADP-form and surrounding phosphate moieties of ADP bound to S 1.
ATP-form in myosin heads [2, 3]. However, the dynamics of
myosin heads remain to be elucidated for understanding the
molecular mechanism of contraction in detail. Materials and methods
Since NMR is a good probe for the microenvironments of
an atom and the interactions among molecules, 31p NMR has Materials
been applied to study the conformation of rabbit skeletal
muscle myosin subfragment-I (S 1)-ADP, S I'pyrophosphate Myosin was extracted from chicken gizzard smooth muscle [7]
(PPi), and S 1·5'-adenylylimidodiphosphate (AMPPNP) and then digested to SI by the treatment with Staphylococcus
complexes [4-6]. The signal of SI·PPi complex was in- aureus V8 protease (1/100 w/w, Sigma) for 45 min at 25°C
dependent of temperature, whereas those of SI·ADP and [8]. The digestion was terminated by the addition of benzyl-
SI·AMPPNP were highly dependent on temperature. The sulfonylfluoride (PMSF) and MgCI 2 was added to be 15 mM.
observations lead to the conclusion that there exist two forms The solution dialyzed against the buffer solution without KCI

Address for offprints: M. Tanokura, Biotechnology Research Centre, University of Tokyo, Bunkyo-ku, Tokyo 113-8657, Japan
76

was centrifuged and the supernatant was applied onto

Sephaeryl S-300 gel filtration. The SI-containing fractions
was concentrated by ammonium sulfate precipitation and
dialyzed against the solution for NMR measurements. The
purity of S I was checked by sodium dodecyl sulfate (SDS)/
68k-
polyacrylamide gel electrophoresis. Protein concentrations
were determined by using an extinction coefficient of 0.70
ml/(mg'cm) at 280 nm and a molecular weight of 131,000 [8].
ADP and P1'Ps-di(adenosine-5')pentaphosphate (APsA)
were obtained from Sigma. They were checked for purity
with 31p NMR. Dp was obtained from Merck and Isotec.
26k-
Preparation ojNMR samples
20k-
Sample solutions for NMR measurements contained 0.3-1.5 17k-
mM S1, ADP at two to three fold molar excess ofS 1, 10 mM
MgCI 2 , 0.3-0.6 M KCI, 40-200 11M APsA, 1 or 2 mM
dithiothreitol (DTT), and 50 mM buffer in about 50% Dp.
Fig. I. SDS/polyacrylamide gel electrophoresis of smooth muscle myosin
APsA was added to inhibit the adenylate kinase activity
SI from chicken gizzard obtained as described in Materials and methods.
contaminating Sl. The inhibitory effect of APsA appeared The concentration of the gel was 13.5%.
insufficient even at the concentrations of 200 11M. Dp was
used as a lock signal for NMR measurements and was con-
firmed to have no effect on the 31 PNMR spectraofSl·Mg-ADP Therefore, chicken gizzard S1 processed with V8 proteinase
complex (hereafter described simply as S1·ADP complex). was used in the present study.

NMR spectroscopy lip NMR spectra ojSIADP complex

31 P NMR spectra were recorded on Bruker AM400 and lEOL As a representative example, the 31p NMR spectrum of
a500 NMR spectrometers operated at 162 MHz and 202 Mg'ADP in the presence ofa half fold molar chicken gizzard
MHz, respectively. Typical spectra were obtained with a pulse SI at pH 7.0 and 10°C. In the spectrum, a new broad peak
delay of2.0 sec and with a flip angle of 60°. Chemical shifts was observed at -2 to -3 ppin in addition to the signals of
were measured downfield from external 85% Hl0 4 • freeADP, AMP, inorganic phosphate, andApsA (Fig. 2). The
signal at -2 to -3 ppm was rather broad, which was never
observed in the 31 PNMR spectra of ADP without SI but was
Results and discussion definitely observed in some op1p NMR spectra of ADP in
the presence ofS 1. In addition, the chemical shift ofthe signal
Chicken gizzard myosin Sl is similar to that of~-phosphate of ADP in the rabbit skeletal
muscle S 1·ADP complex [6]. Therefore, the signal at -2 to
Figure 1 shows the results of SDS/polyacrylamide gel -3 ppin is assigned to ~-phosphate of ADP complexed with
electrophoresis ofchicken gizzard smooth muscle myosin SI chicken gizzard myosin S 1. The low-field shift of the
as obtained by the treatment with Staphylococcus aureus V8 ~-phosphate signal of ADP indicates that the density of
proteinase. The prepared S1 consisted of 68 k and 26 kDa electrons surrounding ~-phosphate was decreased upon the
peptide fragments from heavy chain in addition to the 20 k complex formation with S I; the electrons should move
and 17 kDa light chains, which was the same as the result toward the residues of myosin head.
described in a previous paper [8]. The molecular weight of In the 31p NMR spectra of ADP in the presence of rabbit
S I obtained in the present study is larger than that of rabbit skeletal muscle S 1 [6], the a-phosphate signal of ADP
skeletal muscle SI [6], and NMR signals become broader in complexed with S1 was apparently overlapped by that offree
general in a higher molecular weight protein than in a lower ADP, but seemed broader than the ~-phosphate signal of the
molecular weight protein. Among the treatments with various complex. In the spectra of ADP in the presence of chicken
proteinases including trypsin, chymotrypsin, and papain, gizzard S1, the a-phosphate signal of the complex was not
however, the V8 proteinase treatment gave the best S1 as observed, which was probably due to the extreme broadening
judged by the purity and fragmentation (data not shown). of the signal.
 77

AMP

Pi

! /~
Ap5A

I I I
12.0 6.0 0.0 -6.0 -12.0 -18.0 -24.0
Chemical shift / ppm

Fig. 2. IIp NMR spectrum of the mixture ofMg'ADP and chicken gizzard myosin S I recorded at 162 MHz and at 10°e. Signals are assigned to the indicated
phosphorus nuclei: 0. and ~, 0.- and ~-phosphates of free ADP, respectively; and Pi, inorganic phosphate. The arrow indicates the ~-phosphate of ADP
complexed with the S I'sample contained 1.4 mM S I. 2.8 mM ADP, 10 mM MgCI" 0.3 M KCI, 40 J.IM Ap,A, 2 mM DTT, 0.1 mM EGTA, and 50 mM
imidazole-HCl (pH 7.0). The spectrum was accumulated for 5,000 transients with 60° pulses at the interval of 3 sec.

Two forms ofrabbit skeletal muscle Sl as observed by 31 P ties of ADP to make anisotropic the distribution ofelectrons
NMR around the phosphates. On the other hand, the distribution
of electrons surrounding the phosphorus nuclei of ADP may
We have observed the temperature dependence of the 31p be rather isotropic in the high-temperature form. The tem-
NMR signals of the complex of ADP and rabbit skeletal perature-independent line width indicates that the high-
muscle S I. At 25°e, the ~-phosphate signal of the ADP-S I temperature form (observed typically at 25°C) to the low-
complex was observed at -1.8 ppm, shifting about 4 ppm temperature form in slow exchange as the temperature is
to lower field from that offreeADP. When the temperature lowered.
was lowered, the signal of the S1·ADP complex gradually
reduced its intensity without further broadening and was not
observed at 5 and aoe. The temperature-dependent NMR Comparison ofchicken gizzard Sl AD? complex with
spectra ofS I·ADP complex was interpreted by the idea that rabbit skeletal muscle Sl ADP complex
the phosphorus signals of S I· ADP complex should be very
broad and not observed at a low temperature, which leads The chemical shift of the ~-phosphate of the complex is
to the following conclusions. There exist two different similar in chicken gizzard S1 and rabbit skeletal muscle S I
forms ofSl·ADP complex: one is a form existing mainly [6]. This indicates that the electrostatic environments around
at higher temperatures (high-temperature form) and the ~-phosphate do not largely differ in chicken gizzard S1 from
other is predominant at lower temperatures (low-tempera- rabbit skeletal muscle S1.
ture form). In the low-temperature form, the amino acid The signals of ADP complexed with chicken gizzard S1
residues of S 1 may interact strongly with phosphate moie- are broader than those with rabbit skeletal muscle S1. This
78

may be due to tighter interaction of ADP with chicken gizzard References

S1 and more anisotropic environments around phosphorus
nuclei, or due to simply to the larger molecular weight of I. Szent-Gyorgyi A: Chemistry of Muscular Contraction 2nd Ed.,
chicken gizzard S1. Molecular weight is 115,000 for rabbit Academic Press, New York, 1951
skeletal muscle S I and 131,000 for chicken gizzard S 1. Such 2. Rayment I: The structural basis of the myosin ATPase activity. J BioI
Chern 271: 15850-15853, 1996
a relatively small difference would hardly explain the
3. Rayment I, Smith CA, Yount, RG: The active site of myosin. Annu
difference in line widths of the complex signals with chicken Rev Physiol58: 671-702,1996
gizzard S1 and with rabbit skeletal muscle S 1. Therefore, the 4. Shriver JW, Sykes BO: Phosphorus-31 nuclear magnetic resonance
former possibility is more plausible. evidence for two conformations of myosin subfragment-I'nucleotide
complexes. Biochemistry 20: 2004-2012, 1981
5. Shriver JW, Sykes BO: Energeties and kinetics of interconversion of
two myosin subfragment-I'adenosine 5'-diphosphate complexes as
Acknowledgments viewed by phosphorus-31 nuclear magnetic resonance. Biochemistry
20: 6357-6362, 1981
6. Tanokura M, Ebashi S: Complexes of myosin subfragment 1 with
We would like to express our sincere thanks to Professor pyrophosphate and with adenosine diphosphate as studied by phos-
Setsuro Ebashi (National Institute for Physiological Sciences) phorus-31 nuclear magnetic resonance. J Biochem 113: 19-21, 1993
for his encouragement and interest during the experiments. 7. Ebashi S: Muscle contraction and pharmacology. TIPS-inaugural issue:
29-31,1979
This work was supported in part by Grants-in-Aid for
8. Ikebe M, Hartshorne 0: Proteolysis of smooth muscle myosin by
Scientific Research from the Ministry of Education, Science Staphylococcus aureus protease: Preparation of heavy meromyosin
and Culture ofJapan and by the supports ofUehara Memorial and subfragment I with intact 20 OOO-dalton light chains. Biochemistry
Foundation and Sankyo Foundation. 24: 2380-2387,1985
Molecular and Cellular Biochemistry 190: 79-84, 1999.
© 1999 Kluwer Academic Publishers.

Interactions of protein phosphatase type 1, with a

focus on myosin phosphatase*
David J. Hartshorne l and Katsuya Hirano2
IMuscle Biology Group, The University ofArizona, Tucson, AZ, USA; 2Division ofMolecular Cardiology, Research Institute of
Angiocardiology, Faculty ofMedicine, Kyushu University, Fukuoka, Japan

Abstract
It has been established for many years that MLCK is regulated by the intracellular Ca 2+ concentration via the formation of
the Ca2+-calmodulin-MLCK complex. A more recent discovery has been that the myosin phosphatase may also be regulated.
This is manifest at suboptimal Ca 2+ levels where under certain conditions (e.g. stimulation with several agonists) the MP is
inhibited. The net result being that the extent of myosin phosphorylation for a fixed Ca 2+ level is increased, i.e. an enhanced
Ca 2+-sensitivity. Spurred by this intriguing discovery several laboratories began studies on MP with an emphasis to determine
the regulatory, or inhibitory, mechanism. A similar preparation was obtained by 3 laboratories and consisted of a catalytic
subunit, PP 1S, plus a large subunit (M 130/133 for gizzard, M 130 for bladder and M 110 for rat aorta) and a smaller subunit
of 20--2 1 kD. The isolated catalytic subunit has a much lower activity towards phosphorylated myosin than the holoenzyme,
thus the non-catalytic subunits may serve as targeting proteins and in addition may playa regulatory role. Because of the
difference in activities between the catalytic subunit and holoenzyme, one mechanism of regulation may involve dissociation
of the trimeric complex, and such was proposed for the effect of arachidonic acid. Another suggested regulatory mechanism
was that phosphorylation of the large subunit in its C-terminal half caused inhibition of phosphatase activity. The two
mechanisms need not be mutually exclusive and in addition several kinases could influence the activity of the myosin
phosphatase. In order to understand the molecular basis of phosphatase regulation it is necessary to determine the topography
of the holoenzyme and identify sites of interaction between subunits and substrate. This work is in progress. Using various
truncation mutants of M130/133 it has been determined that the binding sites for both PPlc and substrate are located within the
N-terminal part of the molecule. The M20 subunit binds to the C-terminal end, although the functional significance of this is
not established.
Many questions remain to be answered concerning the biochemistry of the myosin phosphatase. An exciting and challenging
focus will be to determine the mechanism(s) of regulation and to unravel the signaling cascade(s) that are initiated by
agonist-receptor complex formation. In addition, the location of the MP is not known and it is important to establish which (if
any) of the cytoskeletal elements are involved in binding to MP. Finally, it is assumed that the trimeric phosphatase, as discussed
above, is specific for myosin dephosphorylation and does not act on other substrates. Because of the breadth of its distribution
in different tissues and the wide range of proteins interacting with the ankyrin repeats it is possible that this phosphatase, or
variants thereof, has roles in other cellular processes. (Mol Cell Biochem 190: 79-84, 1999)

Key words: calmodulin, myosin subunits, binding proteins

Introduction which a molecular focus for Ca 2+ was recognized was

skeletal muscle and Prof. Ebashi made the crucial discovery
It is now accepted that Ca 2+ plays a central role in the that troponin was the Ca 2+ target and regulated contractile
regulation of cellular processes. One of the first tissues in activity [I]. Subsequently, it was realized that the Ca 2+-
binding component was one of the 3 troponin subunits,
*This article is dedicated to Professor Setsuro Ebashi in honor of his pioneer- namely troponin C. The latter is specific to striated muscle
ing studies on the role ofCa'· in cellular processes. The contributions of Prof. but is related to the ubiquitous Ca 2+ -binding protein,
Ebashi and his colleagues were vital to the development of muscle research. calmodulin. Research on the role ofCa 2+ in cellular function

Address for offprints: D.l. Hartshorne, Muscle Biology Group, University of Arizona, Tucson, AZ 85721, USA
80

has blossomed into an enonnous discipline and one of the physiologically. But in an attempt to restrict the number of
critical factors in the development of this area was the possibilities two criteria were mandated: first, the phos-
discovery of troponin. phatase should be effective with P-myosin as substrate, since
Our research interests over the last few years have focussed there is a possibility that some phosphatases will dephos-
on the regulatory mechanisms in smooth muscle and here the phorylate isolated light chains but not native myosin; and
impact of troponin also was felt. Many of the earlier studies second, that the phosphatase should be associated with the
screened smooth muscle preparations for the presence of contractile apparatus. It is known that in skinned fiber
troponin [2]. These attempts proved unsuccessful and it was preparations the phosphatase remains in the fiber and thus
realized that troponin is specific to striated muscle. Another binding to some cytoskeletal element is not unreasonable.
mechanism was required for smooth muscle and indeed for Following these restrictions three laboratories isolated a MP
non-muscle cells and subsequently this was shown to be that was similar. Two preparations were from gizzard [11, 12]
phosphorylation of the regulatory light chains (LC20) of and one from pig bladder [13]. Each was composed of three
myosin (review [3]). subunits: a catalytic subunit of approximately 38 kD, that was
identified [12] as the PP 10 isofonn (also referred to as the
~-isoform [11]); a large subunit of about 100 kD; and a
Phosphorylation mechanism smaller subunit of20 kD. cDNA's for each subunit have been
cloned and derived sequences obtained. The large subunit
The phosphorylation mechanism involves two key regulatory from gizzard has two isofonns, tenned M 130 and M 133 [12],
proteins: myosin light chain kinase (MLCK) and myosin and these are similar but distinct from the rat aorta M 110 [14].
phosphatase (MP). The link to the intracellular Ca2+transients Another isofonn of this subunit has been detected from a rat
is via calmodulin and the Ca2+-calmodulin complex activates kidney cDNA library [15].
the MLCK apoenzyme [3]. The contractile activity ofsmooth A crucial concern in the coordination of cellular activities
muscle is thought to be controlled by the level of myosin is the control or regulatory mechanisms for kinases and
phosphorylation that in tum reflects the balance of activities phosphatases. One intriguing mechanism is that the enzyme
of the two enzymes. Since there was no evidence to suggest is selectively restricted to a cellular location by targeting
regulation of the phosphatase, it was assumed initially that molecules. This could take the fonn of several enzymes in a
the level of myosin phosphorylation tracked the Ca 2+ tran- common pathway linked together in a scaffolding arrangement,
sients and hence the activity of MLCK. However, with the as suggested for the distinct MAP kinase cascades [16], but
increased application of fluorescent Ca2+ indicators it was also may be a simpler arrangement of targeting via an
found that the Ca 2+-phosphorylation relationship could anchoring protein. In the case of the MP holoenzyme it was
change, and on stimulation by certain agonists the level of suggested that the two non-catalytic subunits may be target
myosin phosphorylation at a given suboptimal Ca 2+ con- molecules [11] and would bind to the substrate, P-myosin, and
centration could increase [4]. This increased Ca2+sensitivity also to PP 1O. In addition to co-localizing the phosphatase and
was traced to an inhibition of phosphatase activity. The substrate the interaction of M 130 and PP 10 may also playa
individual steps linking receptor (agonist) occupancy to regulatory role, since it has been shown that phosphorylation
phosphatase inhibition are not defined, but it has been of M 130 inhibits PP 10 [8]. The complexity of the 'target'
suggested to involve heterotrimeric [4] and monomeric G complex is not known. It is unlikely that MLCK also is part of
proteins [4-{)]. In addition, it was proposed that the molecular the complex, since there is no evidence for MLCK-MP
basis for inhibition involved either dissociation of the interaction and the concentrations ofMLCK (about 4 11M) and
phosphatase complex by arachidonic acid [7] or phos- MP (about 0.7 11M [11]) are different. Possibly the kinase
phorylation of one of the phosphatase subunits [8]. responsible for M 130 phosphorylation fonns part of this
complex and it has been suggested that this is a Rho-activated
kinase [17]. Whether M130 or M20 playa dominant role in
Myosin phosphatase targeting is not established but it is likely that many of the
interactive properties are carried by the larger subunit.
A frequently used classification of protein phosphatases
identifies two classes, type I (PP I) and type 2 (PP2), with
the latter subdivided into 3 categories, A, Band C [9]. There Structure of the non-catalytic MP
have been several reports on phosphatase preparations from
different smooth muscles that were effective with phos- subunits
phorylated light chains or phosphorylated myosin as sub-
strates [10] and these included PP 1 and PP2A. Thus it is Diagrams of the MI30/133, MIlO and M20 subunits are
difficult to claim that only one phosphatase is effective shown in Fig. 1. These are based on the derived sequences
 81
0 200 400 600 800 1000
I I
chicken T695 srr
gizzard *
M133 ANK repeats AlB AlB

0 200 500 600 800 1000

Rat
Aorta
MIlO
T641 LZ

o 200
I I I
chicken
gizzard ~
M20
LZ

Fig. I. Diagrams of the non-catalytic subunits of myosin phosphatase. For the large subunit the chicken gizzard M130/133 and rat aorta MIlO isoforms are
represented. The sequence (numbering) is indicated. Different regions of the molecule are shown: ANK repeats - ankyrin repeats; A - acidic region; 1 -
insert of the M 133 isoform; T695, T641 - phosphorylated threonine of M133 and MIlO, respectively (note, it has not been established that MIlO is
phosphorylated); AlB - ionic clusters; SIT - serine threonine-rich cluster; LZ - leucine zipper. The C-terminal part (boxed in) as indicated is similar in
MI30/133, MIlO and M20.

from cDNA. The M130/133 [12] and M20 [14] are from The functions for different regions of the molecule are not
chicken gizzard and MIlO from rat aorta [14]. For chicken established, although it is known that interaction with
gizzard two isoforms ofthe large subunit were detected. These substrate and PP 18 is located in the N-terminal half [15, 18,
are designated M130 and M133. The difference between the 19]. This region contains the ankyrin repeats and it is
two is the central insert, residues 512-552 in M133. The rat speculated that this might provide a platform for interaction
aorta MIlO is similar, especially in the N-terminal region, with other proteins. This terminology arose from the
residues 1-373 where there is 95% identity. For the first 257 presence of multiple repeats in erythrocyte and brain
residues only 9 are different. Other regions of the molecule ankyrins where it is suggested that the repeats form sub-
show varying degrees of similarity. The sequence 545-603 of domains of 6 repeats each [20]. In addition, these repeats
M133 is not present in MilO and in addition the C-terminal are found in the products of yeast cdc 10 and SW 16 genes
sequence of M 11 0 contains 4 leucine zipper sequences, each (thus these repeats may also be refered to as cdc10/SW16
of 7 residues, not found in Ml30/l33. However, the major repeats) and in several proteins involved in development,
distinctive features are present in both Ml30/133 and MllO including Notch in Drosophila [21] and fem-l, a factor
(Fig. 1). These are: the ankyrin repeat region in the N-terminal involved in sex determination in C. elegans [22]. A list of
part of the molecule. For M130/133 eight repeats are shown proteins in which the ankyrin repeat is found is given by
and seven repeats for MIlO. The difference is the 5th repeat Michaely and Bennett [20]. The number of repeats varies,
which shows less similarity to the consensus sequence and was from 1 in glutaminase to 24 in erythrocyte ankyrin. The
included for Ml30/133, but not for MilO; an acidic cluster localization of these proteins also is varied and they are
(residues 326-372 ofM130/l33); a SIT rich cluster (residues found as an extracellular toxin (the a-Iatrotoxin of black
769-793 ofM133); and ionic clusters (residues 719-755 and widow spider), associated with the plasma membrane, in the
814-848 ofM133). The residue that is phosphorylated and cytoplasm and in the nucleus. It is thought that the general
causes inhibition ofphosphatase activity is T654 ofM 130 or function of the ankyrin repeat is to provide a platform for
T695 ofM133 (Fig. 1). binding to other proteins and possibly other macromolecules.
82

The binding may not be restricted to a single partner and Interactions between phosphatase
ankyrin, for example, binds to several proteins [20]. Because
of the variety of interacting species there is no distinctive
subunits
sequence recognizable as an 'ankyrin-binding motif and the
structural basis underlying the ability of ankyrin repeats to A detailed description of the subunit interactions in the
interact with a broad range of partners is not known. trimeric MP is not available. However, evidence is accumu-
lating and some of the key features can be identified. One of
the phosphatases isolated from chicken gizzard consisted of
Catalytic subunit 58 and 38 kD subunits [18]. The 58 kD component sub-
sequently was shown to be an N-terminal fragment ofMI30
Earlier studies using the phosphatase inhibitors, okadaic [28]. Since the 58-38 complex retained the ability to bind to
acid and calyculin A, suggested that the phosphatase myosin and PPIO [18] it was argued that both interactive
associated with gizzard actomyosin was the type I enzyme properties were associated with the N-terminal part ofM 130.
[23]. Five isoforms, the products of3 PP Ic genes, have been In addition, the 58-38 complex did not contain M20 and thus
described, namely ex l , ex 2, Y" Y2, and 0 [24, 25] (PP lois a critical role for the latter subunit was not obvious. However,
sometimes referred to as PPI~ [lID. The isoforms differ the 58-38 complex does not contain the inhibitory phos-
mainly in the C-terminal sequences, for about 30 residues. phorylation site and thus a role for M20 in regulation of
Some variation also is found in the N-terminal sequences, phosphatase activity cannot be eliminated.
but the bulk of the molecule is highly conserved between Subsequent studies focused on the use ofseveral truncation
isoforms with over 95% identity. Differences in function mutants to progressively dissect various properties of the
associated with the N- and C-terminal sequences have not M 130/133 subunit. The M 133 1-{j74 (sequence indicated by
been identified and Zhang et al. [26] found that for 4 superscript) mutant activated PPlc activity towards P-LC20
isoforms (ex l , Y" Y2 and 0) several properties were identical. and P-myosin. Thus it was assumed that the sequence 1-674
There is some difference in tissue distribution; for example, contained binding sites for both PP Ic and substrate. The
PPly" is high in brain and PPIY2 is high in testis [27]. The M 130 1-{j33 mutant also was shown to bind P-HMM, P-LC20
PP I0 isoform is found in smooth muscle and was detected and PPlc [19].
bound to myosin [28]. Subsequent studies used shorter mutants to more precisely
The crystal structure ofPPlex l complexed with microcystin- define the interaction sites. Different techniques were used
LR (a phosphatase inhibitor) has been determined [29]. to detect binding and included the yeast two-hybrid system
Residues 7-300 form a compact structure, referred to as the [30] and phosphatase assays [30, 31]. For the binding ofPPIc
catalytic domain, which is composed of two substructures. to M 130, at least two sites were detected. One in the ankyrin
Two metal ions were detected and because of the importance repeat region [31] probably in the C-terminal repeats [30] and
of the coordinating residues to enzymatic activity, it was surprisingly a site in the N-terminal sequence 1-38 [30, 31].
proposed that the metal ions form part of the active site. The The latter site was dominant and mutants lacking this sequence
identity of the metals was not established. The presence of did not activate PP Ic activity [30, 31] nor bind to PP Ic in
the two ions plus conserved arginine residues is thought to the two-hybrid assays. Recently, Endo et al. [32] found that
create a region of positive electrostatic potential on the an N-terminal peptide of inhibitor-I, KIQF, was necessary
surface of the molecule that would facilitate nucleophilic for full inhibition in addition to the phosphorylated T35. A
attack on the phosphate ester (substrate). The active site and similar sequence, KVKF is found in M130/133 at residues
the electropositive patch are localized at the branch point of 35-38. Thus, it is possible that this short sequence is involved
a Y -shaped groove. One arm of the groove is hydrophobic with interaction of PP Ic.
and the other, within the C-terminal subdomain, is acidic. The situation with respect to binding of substrate (P-LC20
Microcystin bound to the PP Ic molecule at the active site and or P-myosin) is more complex. In the absence of ATP,
within the hydrophobic groove and part of the third arm of dephosphorylated myosin binds to the N-terminal half of
the Y, termed the C-terminal groove. A model for the M 130. However, in the presence ofATP only phosphorylated
interaction of inhibitor-I and DARPP-32 with PPlc involved substrate binds [19]. An interaction between PPlc and
the interaction of the phosphorylated threonine residue with substrate is expected, but this is relatively weak and the
the active site and binding of the basic residues (on the binding ofPPlc to a thiophospborylated myosin column was
N-terminal flank of the phosphoT) in the acidic groove. In barely detected. Under the same conditions binding ofthe MP
addition, hydrophobic residues on the C-terminal flank of the holoenzyme was much stronger [19]. Thus a second site
phosphoT may interact in the hydrophobic groove. Why the exists on M130/133 for the binding of phosphorylated
phosphoT does not represent a substrate for the phosphatase substrate. Using P-LC20 overlays it was suggested that this
is not known. site was in the ankyrin repeat region, again probably in the
 83

C-terminal repeats [30]. TheATP-dependence of the inter- although the functional significance of this binding is not
action is probably due to a conformational change in known.
myosin (or HMM). In the absence of ATP the site for
interaction with M130 is available, whereas in the presence
of ATP this site is masked with dephosphorylated myosin Other PPlc-binding proteins
and open only with the phosphorylated myosin. There are
no data available on the location of this site(s) on the PPlc is involved in many physiological processes [33]. If
myosin molecule. binding (target) proteins are involved in the regulation of
A potential regulatory mechanism for MP involves phos- these mechanisms then the complexity of the situation
phorylation of M 130/133 and previously it was shown that becomes apparent. In order to search for PP 1c-binding
phosphorylation ofT654 on MI30 (T695 ofM133) inhibited proteins in smooth muscle we applied the yeast two-hybrid
phosphatase activity [8]. Since PPlc is thought to be anchored system with PP 10 as bait to screen a chicken gizzard cDNA
to the N-terminal end ofM130 some folding ofM130 would library. Thirty five positive clones were detected and grouped
be expected to facilitate interaction ofthe phosphoTand PPlc. into 8 gene groups. These are shown in Table 1. The sequences
The mechanism of inhibition is not established, but one of several of these were determined and compared to se-
possibility is that the phosphorylated residue interacts at the quences in the data bank. Three of the clones were matched
active site ofPPlc in a manner similar to the phosphorylated with known proteins: clone III was the chicken glycogen-
threonines of inhibitor-I and DARPP-32. The sequences binding subunit; clone IV, was a splicing factor (PSF,
flanking phosphoT for inhibitor-l and DARPP-32 are:- polypyrimidine tract-binding protein-associated splicing
RRRRPpTPATLVLT- and -RRRRPpTPAMLFR-, respe- factor [35]); and clone V, the ribosomal protein, L5 [36].
ctively. For M 130/133 the corresponding sequence is: Proteins similar to the other clones are listed in Table 1. Clone
-QSRRSpTQGVTLT-. This is similar to the other inhibitors I was not identified, but when expressed as a hexahistidine
in that it contains basic residues on the N-terminal flank (to fusion protein it inhibited PP1c activity towards phos-
interact with the acidic groove of PPlc) and hydrophobic phorylase a but not towards P-LC20.
residues on the C-terminal site (to interact with the hydro- The PP 1o-binding proteins detected by the two-hybrid
phobic groove). A weaker interaction of MI30 and PPlc screen represent only a partial listing. It is likely that many
(compared to inhibitor-I and PP 1c) might be expected since more PP10 targets remain to be identified. For example, the
the extent of inhibition is only about 80%. However, this is M130 subunit was not detected, even though with specific
speculative and the mechanism of inhibition remains to be pairs ofPP 10 and M 130 (as prey and bait) binding is detected.
established. The reason that M130/133 was not detected in the library
Finally, it was shown using the two-hybrid system that screen is not known, but may be due to relatively stringent
M20 interacts with the C-terminal part of M 130 [30] screening conditions.

Table I. Summary of PPI S-binding proteins detected by two-hybrid screen of a chicken gizzard EDNA library

Number of Size of cDNA insert Protein product Similar proteins in the data basel
clones' (excluding adaptors) size (+ adaptors) pI'

8 1435 base pairs 426 (+5) residues 5.31 yeast 67kD protein (X87611)
FK506 binding protein
malaria antigen
II 7 1142 268 (+5) 5.16 mouse unknown gene product (Z78141)
C32A3.3 C. elegans coiled-coil domain (Z4824I )
mitosin
CENP-F kinetochore protein
myosin heavy chain
III 5 1010 258 (+5) 9.79 hepatic glycogen-binding subunit
skeletal G-subunit
IV 2 726 242 (+12) 5.35 PSF (U35610, X70944)
V I 376 96 (+5) 10.42 L5 (U31753, X57016)
VI I 345 115(+5) 4.34
VII I -700 43 8.53
VIII I -400 3 (+5) n.d.

'Number of primary clones containing identical insert. 'Theoretical isoelectric point; n.d., not determined. lSimilar proteins obtained by Blast search (34).
Bold, Ilroteins showing score higher than 100; Parenthesis, GeneBank accession number.
84

References 18. Okubo S, Erdiidi F, Ito M, Ichikawa K, Konishi T, Nakano T,

Kawamura T, Brautigan OL, Hartshorne DJ: Characterization of a
myosin-bound phosphatase from smooth muscle. Adv Prot Phos-
I. Ebashi S, Kodama A, Ebashi, F: Troponin I. Preparation and phatases 7: 295-314, 1993
physiological function. J Biochem 64: 465--477, 1968 19. Ichikawa K, Hirano K, Ito M, Tanaka J, Nakano T, Hartshorne OJ:
2. Driska SP, Hartshorne DJ: The contractile proteins of smooth muscle. Interactions and properties of smooth muscle myosin phosphatase.
Properties and components of a Ca'+-sensitive actomyosin from Biochemistry 35: 6313--6320, 1996
chicken gizzard. Arch Biochem Biophys 167: 203-212, 1975 20. Michaely P, Bennett V: The ANK repeat: A ubiquitous motif involved
3. Hartshorne DJ: Biochemistry of the contractile process in smooth in macromolecular recognition. Trends Cell Bioi 2: 127-129, 1992
muscle. In: L.R. Johnson (ed). Physiology of the Gastrointestinal Tract. 21. Wharton KA, Johansen KM, Xu T, Artavanis-Tsakonas S: Nucleotide
Raven Press, New York, 1987, pp 423--482 sequence from the neurogenic locus notch implies a gene product that
4. SomlyoAP, SomlyoAV: Signal transduction and regulation in smooth shares homology with proteins containing EGF-like repeats. Cell 43:
muscle. Nature 372: 231-236, 1994 567-581, 1985
5. Hirata K, Kikuchi A, Sasaki T, Kuroda S, Kaibuchi K, Matsuura Y, 22. Spence AM, Coulson A, Hodgkin J: The product of Jem-I, a
Seki H, Saida K, Takai Y: Involvement of rho p21 in the GTP-enhanced nematode sex determining gene, contains a motiffound in cell cycle
calcium ion sensitivity of smooth muscle contraction. J Bioi Chern control proteins and receptors for cell-cell interactions. Cell 60: 981-
267: 8719-8722, 1992 990,1990
6. Gong MC, Iizuka K, Nixon G, Browne JP, Hall A, Eccleston JF, Sugai 23. Ishihara H, Martin BL, Brautigan DL, Karaki H, Ozaki H, Kato Y,
M, Kobayashi S, SomlyoAV, SomlyoAP: Role of guanine nucleotide- Fusetani N, Watabe S, Hashimoto K, Uemura D, Hartshorne DJ:
binding proteins ras-family or trimeric proteins or both in Ca2+ Calyculin A and okadaic acid: inhibitors of protein phosphatase
sensitization of smooth muscle. Proc Natl Acad Sci USA 93: 1340- activity. Biochem Biophys Res Commun 159: 871-877,1989
1345,1996 24. Sasaki K, Shima H, Kitagawa Y, Irino S, Sugimura T, Nagao M:
7. Gong MC, Kinter MT, Somlyo AV, Somlyo AP: Arachidonic acid and Identification of members of the protein phosphatase I gene family
diacyl glycerol release associated with inhibition of myosin light chain in the rat and enhanced expression of protein phosphatase In gene in
dephosphorylation in rabbit smooth muscle. J Physiol486: 113-122, 1995 rat hepatocellular carcinomas. Jpn J Cancer Res 81: 1271-1280, 1990
8. Ichikawa K, Ito M, Hartshorne DJ: Phosphorylation of the large subunit 25. Ourfee T, Becherer K, Chen P-L, Yeh S-H, Yang Y, Kilburn AE, Lee
of myosin phosphatase and inhibition of phosphatase activity. J Bioi W-H, Elledge ST: The retinoblastoma protein associates with the protein
Chern 271: 4733--4740, 1996 phosphatase type I catalytic subunit. Genes Dev 7: 555-569, 1993
9. Ingebritsen TS, Cohen P: The protein phosphatases involved in cellular 26. Zhang Z, Bai G, Shima M, Zhao S, Nagao M, Lee EYC: Expression
regulation I. Classification and substrate specification. Eur J Biochem and characterization of rat protein phosphatases-I n, -I Y" -I y2, and
132: 255-261, 1983 -10. Arch Biochem Biophys 303: 402--406,1993
10. Erdiidi F, Ito M, Hartshorne DJ: Myosin light chain phosphatase. In: 27. Shima H, Hatano Y, Chun Y-S, SugimurAT, Zhang Z, Lee EYC, Nagao
M. Baniny (ed). Biochemistry of Smooth Muscle Contraction. M: Identification of PP I catalytic subunit isotypes PP Iyl, PP I0 and
Academic Press, San Diego, 1996, pp 131-142 PPln in various rat tissues. Biochem Biophys Res Commun 192: 1289-
II. Alessi D, MacDougall LK, Sola MM, Ikebe M, Cohen P: The control 1296, 1993
of protein phosphatase-I by targeting subunits. The major myosin 28. Okubo S, Ito M, Takashiba Y, Ichikawa K, Miyahara M, Shimizu H,
phosphatase in avian smooth muscle is a novel form of protein Konishi T,'Shima H, Nagao M, Hartshorne DJ, Nakano T: A regulatory
phosphatase-I. Eur J Biochem 210: 1023-1035, 1992 subunit of smooth muscle myosin bound phosphatase. Biochem
12. Shimizu, H, Ito M, Miyahara M, Ichikawa K, Okubo S, Konishi T, Biophys Res Comm 200: 429--434, 1994
Naka M, Tanaka T, Hirano K, Hartshorne DJ, Nakano T: Charact- 29. Goldberg J, Huang H-B, Kwon Y-G, Greengard P, NairnAC, Kuriyan
erization of the myosin-binding subunit of smooth muscle myosin J: Threedimensional structure of the catalytic subunit of protein serine/
phosphatase. J Bioi Chern 269: 30407-30411, 1994 threonine phosphatase-I. Nature 376: 745-753, 1995
13. Shirazi A, Iizuka K, Fadden P, Mosse C, Somlyo AP, Somlyo AV, 30. Hirano K, Phan BC, Hartshorne DJ: Interactions of the subunits of
Haystead TAJ: Purification and characterization of the mammalian smooth muscle myosin phosphatase, in press.
myosin light chain phosphatase holoenzyme. The differential effects 31. Johnson DF, Moorhead G, Caudwell FB, Cohen P, Chen YH, Chen
of the holoenzyme and its subunits on smooth muscle. J BioI Chern MX, Cohen PTW: Identification of protein-phosphatase-I-binding
269: 31598-31606, 1994 domains on the glycogen and myofibrillar targeting subunits. Eur J
14. Chen YH, Chen MY, Alessi DR, Campbell DG, Shanahan C, Cohen Biochem 239: 317-325, 1996
P, Cohen PTW: Molecular cloning of eDNA encoding the 110 and 21 32. Endo S, Zhou X, Connor J, Wang B, Shenolikar S: Multiple structural
kDa regulatory subunits of smooth muscle protein phosphatase 1M. elements define the specificity of recombinant human inhibitor-l as a
FEBS Lett 356: 51-55,1994 protein phosphatase-l inhibitor. Biochemistry 35: 5220-5228, 1996
15. Haystead CMM, Gailly P, Somlyo AP, Somlyo AV, Haystead TAJ: 33. Wera S, Hemmings BA: Serine/threonine protein phosphatases.
Molecular cloning and functional expression of a recombinant 72.5 Biochem J 311: 17-29, 1995
kDa fragment of the 110 kD regulatory subunit of smooth muscle 34. Altschul SF, Gish W, Miller W, Myers EW, Lipman, DJ: Basic local
protein phosphatase 1M. FEBS Lett 377: 123-127,1995 alignment search tool. J Mol Bioi 215: 403--410, 1990
16. Faux MC, Scott JO: Molecular glue: Kinase anchoring and scaffold 35. Hirano K, Erdiidi F, Patton JG, Hartshorne OJ: Interaction ofprotein phos-
proteins. Cell 85: 9-12,1996 phatase type I with a splicing factor. FEBS Lett 389: 191-194, 1996
17. Kimura K, Ito M, Amano M, Chihara K, Fukata Y, Nakafuku M, 36. Hirano K, Ito M, Hartshorne OJ: Interaction of the ribosomal protein,
Yamamori B, Feng J, Nakano T, Okawa K, Iwamatsu A, Kaibuchi K: L5, with protein phosphatase type I. J Bioi Chern 270: 19786-19790,
Regulation of myosin phosphatase by rho and rho-associated kinase 1995.
(rho-kinase). Science 273: 245-248,1996
Molecular and Cellular Biochemistry 190: 85-90, 1999.
© 1999 Kluwer Academic Publishers.

Myosin light chain kinase from skeletal muscle

regulates an ATP-dependent interaction between
actin and myosin by binding to actin

Koichiro Fujita, 1,2 Li-Hong Ye, l Manabu Sato 1 Tsuyoshi Okagaki, 1

Yukio Nagamachi2 and Kazuhiro Kohama 1
'Department ofPharmacology, Gunma University School ofMedicine, Maebashi, Gunma; 2First Department ofSurgery,
Gunma University School ofMedicine, Maebashi, Gunma, Japan

Abstract
Myosin light chain kinase (MLCK) has been purified from various muscles as an enzyme to phosphorylate myosin light chains.
While the regulatory role of smooth muscle MLCK is well understood, the role of skeletal muscle MLCK in the regulation of
contraction has not been fully characterized. Such characterization of skeletal muscle MLCK is difficult because skeletal muscle
myosin interacts with actin whether or not the myosin is phosphorylated. Taking the hint from our recent finding that smooth
muscle MLCK inhibits the actin-myosin interaction by binding to actin (Kohama et al., Biochem Biophys Res Commun 184:
1204-1211, 1992), we investigated the regulatory role of the actin-binding activity of MLCK from chicken breast muscle in
the actin-myosin interaction. The amount ofMLCK that bound to actin increased with increases in the concentration ofMLCK.
However, MLCK hardly bound to myosin. The actin-binding activity of MLCK was affected when Ca2+and calmodulin (Ca 2+-
CaM) were present. The effect of MLCK on the actin-myosin interaction was examined by an in vitro motility assay; the
movement of actin-filaments on a myosin-coated glass surface was inhibited by increasing the concentration ofMLCK. When
CaM was present, the inhibition was overcome in a Ca 2+-dependent manner at 11M levels. The inhibition of the movement by
MLCK and the recovery from the inhibition by Ca 2+--CaM were not altered whether we use phosphorylated or unphosphorylated
myosin for the assay, ruling out the involvement of the kinase activity of MLCK. (Mol Cell Biochem 190: 85-90, 1999)

Key words: skeletal muscle, myosin light chain kinase, actin-binding, actin-myosin interaction

Introduction phosphorylated. MLCK is responsible for this phosphoryla-

tion together with calmodulin (CaM), which binds Ca2+ to
Contraction of skeletal muscle is regulated by the troponin activate the kinase function ofMLCK ([3] for a review).
tropomyosin complex ([ I] for a review). TheATP-dependent MLCK is present in skeletal muscle cell, and its activity
interaction between actin and myosin was inhibited by this was first demonstrated in skeletal muscle [4]. This kinase was
regulatory complex in the absence of Ca2+. When Ca2+ was subsequently purified from skeletal muscle [5] and its amino
elevated to 11M levels, Ca2+binds to troponin C in the complex, acid sequence was determined [6]. However, skeletal muscle
leading to the actin-myosin interaction that causes the muscle myosin interacts with actin to the same extent whether or not
to contract. Smooth muscle also contracts when the intra- its light chain has been phosphorylated by MLCK [7]. Thus
cellular concentration of Ca2+ is elevated [2]. However, the the physiological significance of MLCK in skeletal muscle
details of smooth muscle contraction are quite different from has been an enigma. For example, Stull and his colleagues
those ofskeletal muscle contraction. Smooth muscle contracts suggested recently that MLCK might be involved in the
only when the light chain of smooth muscle myosin is post-tetanic potentiation of skeletal muscle contraction [8].

Address for offprints: K. Kohama, Department of Pharmacology, Gunma University School of Medicine, Maebashi, Gunma 371, Japan
86

It was reported some time ago that MLCK exists in skeletal that had been purified from chicken gizzard smooth muscle
muscle cells in association with actin [9]. Such localization of as described by Hayakawa et al. [17], namely, by a modified
MLCK seems incompatible with the kinase activity ofMLCK version of the method of Adelstein and Klee [18]. The
and it seems possible that the actin-binding property ofMLCK reaction mixture after incubation was mixed with 9 volumes
might play some unknown role in regulating actin-myosin of cold distilled water and then subjected to centrifugation.
interaction. We succeeded recently in identifying the regulatory The pellet was suspended in a solution of 0.1 M KCl and 20
activity of this actin-binding property of MLCK of smooth mM Tris-HCl (PH 7.5) and used as phosphorylated myosin.
muscle [10, 11]. MLCK binds to actin and inhibits the ATP- The extent of phosphorylation of DTNB-light chain was
dependent interaction between actin and myosin. Neither Ca2+ measured by polyacrylamide gel electrophoresis (PAGE) in
nor CaM is required for the inhibition. CaM in the presence of the presence of urea [19]. Actin was purified in the mono-
Ca 2+(Ca2+-eaM) inhibited the actin-binding activity ofMLCK meric form from an acetone powder ofchicken breast muscle
to overcome the inhibition of the ATP-dependent interaction [20], and then it was polymerized and used as actin filaments.
[12-14]. In this recovery, phosphorylation of the myosin light The concentration of actin filaments is expressed in terms
chain plays no part at all. This mode of regulation by the of that of monomeric actin. CaM from bovine brain was pur-
MLCK-CaM complex resembles the mode of action of the chased from Sigma (St. Louis, MO, USA). The purity of pre-
troponin-tropomyosin complex. Therefore, we decided to parations ofMLCK, myosin and actin was routinely checked
examine whether or not such a mode of action, involving the by sodium dodeecyl sulfate-PAGE (SDS-PAGE; see below).
actin-binding property of MLCK, might function to regulate
the contraction of skeletal muscle. In this study, we purified
MLCK from skeletal muscle and characterized it as an actin- Binding experiments
binding protein that inhibits the actin-myosin interaction.
To measure the actin-binding and myosin-binding activities of
MLCK, MLCK was mixed with 1 /-lM actin filaments or 1 /-lM
myosin and 1.57 /-lM CaM in 60 mM KCl, 20 mM Tris-HCl
Materials and methods (PH. 7.5) and 0.1 mm [ethylenebis(oxyethylenenitrilo)]
tetraacetic acid (EGTA). In experiments designed to test the
Proteins
effects of Ca 2+-CaM, EGTA was replaced by 0.1 mM CaCI 2•
After standing for 30 min at room temperature, the mixture
Myosin light chain kinase was purified from chicken breast
was centrifuged at 130,000 x g for 45 min in a Beckman
skeletal muscle by a modified version of the method of
Airfuge. The supernatant and the material in the pellet were
Nunnally et al. [15]. In brief, we prepared a crude extract of
separately subjected to SDS-PAGE (see below). To calculate
the muscle and subjected it to ammonium sulfate fractionation
the amounts ofMLCK that had bound to actin or myosin, the
as described by Nunnally et al. [15]. Then MLCK was purified
amounts of myosin, actin and MLCK on gels were de-
by sequential column chromatography on Super Q Toyopearl
termined by densitometry as described by Ishikawa et al. [21].
650M (Tosoh, Tokyo, Japan), calmodulin-Agarose (Sigma, St.
Louis, MO, USA), and hydroxyapatite (Seikagaku Corp.
Tokyo, Japan) in that order. We confirmed that the MLCK In vitro motility assay
prepared in this way phosphorylated the 5,5-dithiobis (2-
nitrobenzoic acid)-light chain (DTNB-light chain) of myosin The effect of MLCK on the ATP-dependent interaction
(see below). When the DTNB-light chain was phosphorylated between actin and myosin was analyzed in an in vitro motility
by MLCK from skeletal muscle or smooth muscle (see below) assay on myosin-coated surfaces, as described by Okagaki
in the presence of Ca2+-eaM, the former phosphorylated the et al. [22], who modified the original method of Kron and
light chain much more effectively than the latter. MLCK was Spudich [23]. In brief, 3 nM actin filaments labelled with
from skeletal muscle unless otherwise specified. thodaminephalloidin (Molecular Probes, Eugene, OR, USA)
Myosin was prepared from chicken breast muscle as were mixed with MLCK at specified concentrations and were
described by Kohama [16]. We used for experiments without placed on the myosin-coated coverslips in 3 mM MgCI 2, 2
subjecting it to the procedure ofphosphorylation (see below) mM ATP, 60 mM KCl, 30 mM imidazole (pH 7.6) and 0.1
unless otherwise specified. The concentration of myosin is mM EGTA. To see the effect of Ca 2+-eaM, 0.1 mM EGTA
expressed in terms of the myosin heavy chain monomer. In were replaced by 0.1 mM CaCl 2 and specified concentrations
some experiments, we used myosin after subjecting it to the of MLCK. The mounted coverslips were examined under a
procedure ofphosphorylation as follows; Myosin (4 /-lM) was fluorescence microscope equipped with a video camera and
incubated at 25°C for 10 min in a solution of 0.1 M KCl, 10 recorder, and then the mean vellocity (/-lm/sec) of movements
mM MgCI 2 1.1 mM ATP, 20 mM Tris-HCl (pH 7.5) and 0.1 was determined by tracking the movements of 20 actin
mM CaCl 2 together with 4.7 /-lM CaM and 0.4 /-lM MLCK filaments and expressed as mean ± S.D.
 87

Other procedures of MLCK giving half-maximal binding of MLCK to actin

was 0.22 11M. The presence of Ca2+-CaM (closed circles in
The actin-activated ATPase activity of myosin was deter- Fig. lA) affected the actin-binding activity of MLCK. The
mined by the malachite green method [24] in duplicate or maximal extent of binding was reduced to 0.35 mol ofMCK
triplicate assay as follows.ATPwas hydrolyzed by incubation per mol of actin, and the concentration for half-maximal
with 0.1 mM myosin at 25°C for 10 min in the presence binding was increased to 0.39 11M.
MLCK (0-100 nM), O. 95 mM actin filaments, 0.35 11M CaM, We also examined the myosin-binding activity of MLCK
10 mM MgCI 2, 0.5 mM ATP, 60 mM KCI, 20 mM Tris-HCI under conditions similar to those used for the examination of
(pH 7.5) and 0.1 mM CaCI2 or 0.1 mM EGTA. We ensured actin-binding activity. As shown in Fig. IB, MLCK hardly
that the rate of ATP hydrolysis was constant and that the bound to myosin, indicating that the affinity to myosin is
specific activity of the ATPase did not differ from our negligible. The negligible affinity is compatible with the
previously reported values [25]. previous report that MLCK exists in association with actin
SDS-PAGE was performed as described elsewhere [26] by filaments, rather than myosin filaments, in skeletal muscle
a modified version of the method of Laemmli [27]. Protein cells [9].
concentrations were determined as described by Bradford
[28] with bovine serum albumin as the standard.
For calculations of molarities of solution of proteins, the Effects ofMLCK as examined by the in vitro motility assay
following molecular masses were used: myosin heavy chain,
222,972 Da [29]; actin monomer, 41,785 Da [30] and MLCK, The velocity of movement of actin filaments on a myosin-
87,064 Da [6]. coated glass surface was monitored in the presence ofvarious
concentrations of MLCK.
In the absence of Ca 2+-CaM, as indicating by the open
Results circles in Figure 2, the velocity of movement of actin
filaments decreased with increases in the concentration of
Actin-binding and myosin-binding activities ofMLCK MLCK. We failed to observe any movement of actin fila-
ments in the presence of 90 nM MLCK. The concentration
Figure I shows the details of the binding of MLCK to actin of MLCK for half-maximal inhibition was about 60 nM. In
filaments. In the presence of EGTA, when CaM was non- the presence of Ca 2+-CaM (filled circles in Fig. 2), the
functional, the amounts of MLCK, as shown by the open inhibition by MLCK was obscured. Similar inhibition and
circles in Figure 1A. The maximal extent of binding was relief are observed when we examined the smooth muscle
about 0.7 mol of MLCK per mol of actin. The concentration MLCK using the same assay [13].

A 06 B 08

c
~
0.6
o 0.6

~
.2 ~
<G
II: C 0.4

'E~
-
lG
~
0.2 ~ 0.2
g
o -f6.-!!!::="'_~-----,r------'T---"
:!:
o+-__ -s:~---.c~;;....-_,-----,
0----0-
o 0.2 0.4 0.6 0.8 o 0.2 0.4 0.6 0.8

MLCK (II M) MLCK (J.I M )

Fig. I. Binding of MLCK to actin and myosin. Actin filaments (A) were incubated with MLCK at various concentrations (0-0.78 flM) and 1.57 flm M CaM
in 60 mM KCl, 20 mM Tris-HCl (pH 7.5) and 0.1 mM Ca'· (closed circles) or 0.1 mM EGTA (open circles). Myosin (B) was also incubated under the same
conditions as those used for actin filaments except that no Ca2+--eaM was included. After a 30 min incubation, samples were centrifuged at 130,000 x g for
45 min. Supernatants and pellets were separately subjected to SDS-PAGE. The amounts of MLCK, actin and myosin were determined by densitometry.
Ordinates, bound amount (molar ratio); abscissae, concentrations (flM).
88

filaments was depressed. However, with the increase in Ca2+

concentration, the velocity of movement was increased. The
half-maximal inhibition was observed at about pCa=6, giving
an apparent activation of the interaction of myosin with actin
filaments.

Is the Ca 2+ -dependency attributable to the kinase activity

ofMLCK?

The myosin preparations so far used for the in vitro motility

assay have not been subjected to the phosphorylation pro-
cedure. The kinase activity of MLCK requires Ca 2+-CaM to
o+---..----.,..---.,..-~:r-..., phosphorylate myosin by its kinase activity [15]. Therefore,
25 50 75 100
the inhibitory effect of MLCK observed in the presence of
EGTA (Fig. 2 filled circles) obviously does not involve the
MLCK (n M)
kinase activity. However, during the recovery from the
inhibition that occurs in the presence ofCa2+-CaM (Fig. 2 open
Fig. 2. Effects ofMLCK as examined by thein vitro motility assay. MLCK
at various concentrations (0-90 nM) was mixed with 0.1 mM EGTA
circles), MLCK is expected to phosphorylate myosin.
containing actin filaments and ATP. The mixture was mounted on a myosin- We coated glass surfaces with myosins with and without
coated glass surface. The velocity (11m/sec) of movement of actin-filaments subjecting the phosphorylation procedures. Then, we com-
were measured under an fluorescent microscope. To examine the effect of pared the effect ofMLCK; control actin filaments, i.e., actin
Ca'+-eaM, 0.1 mM EGTA was replaced by 100 nM CaM and 0.1 mM filaments that were not associated with MLCK moved at the
Cal+. Filled circles, velocity in the presence of Ca'+-eaM; open circles,
velocity in the presence of EGTA. Ordinate, velocity (11m/see); ordinate,
same velocity on both surfaces (Fig. 4 column a). This finding
concentration (11M); error bars, S.D. confirms by in vitro motility assay that interaction of actin
with skeletal muscle myosin is not altered by the phos-

When Ca2+concentration was varied (Fig. 3) in the presence

of CaM, the effect ofMLCK was dependent on Ca 2+. At the .~

low concentrations of Ca 2+(pCa::;8), the movement of actin T

u :.5
Q)
!!!.
E
~

~
'u
o 15
Q5
>

o -L..L..:.--=--,-
0.5 +--..---r---.----.--..---
o
a b c
8 7 6 5 4

Fig. 4. In vitro motility assays with phosphorylated and unphosphorylated

pea myosins. Actin filaments were allowed to move ATP-dependently on
surfaces coated with unphosphorylated (dotted columns) or phosphorylated
Fig. 3. Ca 2+-deperident movement of actin-filament in the presence of myosin (hatched columns), and the velocity of the movement (11m/sec) was
MLCK and CaM. An in vitro motility assay was carried out in the presence expressed by columns. a, actin filaments alone; b, actin filaments mixed
of20 nM MLCK, 100 nM CaM, and Ca'+ the concentration of which was with 20 nM MLCK and 0.1 mM EGTA; c, actin filaments mixed with 20
varied by Ca-EGTA buffer. For other details, see the legend to Fig. 2. pCa nM MLCK, 100 nM CaM, and 0.1 mM Ca'+. For other details, see the
= log Ca'+ (M). legend to Fig. 2.
 89

phylation level of myosin [7]. As shown by column b in Fig. having a calculated pI of 10.52. This sequence includes two
4, MLCK inhibited the movement to the similar extent regions that are homologous at the 40% level to region in
whether the glass surfaces were coated by the phosphorylated smooth muscle MLCK: K253_ p266 and G272_T286 [34]. Al-
or unphosphorylated myosin. Similarly, the recovery from the though the extent of such homology is not very high, we
inhibition was observed with the both glass surfaces to the speculate tentatively that these basic portions of skeletal
same extent (Fig. 4 column c), indicating that myosin muscle MLCK might be sites ofbinding to actin. Biochemical
phosphorylation by the kinase activity of MLCK, if any, is confirmation is now required using proteolytic fragments of
not responsible for the recovery. these regions of skeletal muscle MLCK.
Present study adds a novel aspect to a few view points
about the role of the MLCK in skeletal muscle (e.g., ref. [8]).
Discussion We do not know whether or not the inhibitory effect ofMLCK
observed in vitro has physiological relevance. To discuss the
This study demonstrates that MLCK binds to actin filaments relevance, we need in vitro data concerning the cross-talk of
to inhibit an ATP-dependent interaction between actin and the MLCK-CaM complex with the troponin-tropomyosin
myosin, and this inhibition can be overcome by Ca2+--CaM. complex as well as in vivo data concerning the involvement
Phosphorylation ofmyosin light chains by the kinase activity of the actin-binding property of MLCK to the regulation of
of MLCK was not involved in either the inhibition or the skeletal muscle contraction.
recovery. The conventional assay for the effect may be pro-
vided by measuring the actin-activated ATPase activity of
myosin. However, we deliberately assayed the effect with the Acknowledgements
in vitro motility of actin filaments, because [1] the motility
assay gives direct aspects ofthe actin-myosin interaction [23] This work was supported by the grants from the Mitsubishi
and (ii) because complication of assembly of actin filaments Foundation and the Smoking Research Foundation, and by
due to MLCK [17] will be avoided. The measurements of the grants-in-aid for scientific research from the Ministry of
ATPase activity as described in Materials and methods con- Education, Science and Culture of Japan
firmed that MLCK inhibits the activity and that Ca2+--CaM
obscures the inhibition. However, we detected partial
inhibition. References
This regulatory property of MLCK resembles in that of
I. Ebashi S, Endo M: Calcium ion and muscle contraction. Progr Biophys
troponin-tropomyosin which binds to actin filaments to
Mol Bioi 18: 123-183, 1958
inhibit the actin-myosin interaction [1]. The inhibition is 2. Ebashi S, Iwakura H, Nakamura R, Ooi Y: New structural protein
overcome when Ca 2+ bind to troponin C, a Ca 2+-binding from dog heart and chicken gizzard. Biochem J 345: 201-211, 1966
subunit of troponin. The abolition of the movement of actin 3. Kamm KE, Stull JT: The function of myosin and myosin light chain
filaments by the which complex on a surface coated with kinase phosphorylation in smooth muscle. Annu Rev Pharmacol
myosin is complete when Ca2+are absent [31]. However, the Toxicol 25: 593~20, 1985
4. Pires E, Perry SV, Thomas MAW: Myosin light chain kinase, a new
inhibitory effect of this complex on the actin-activated enzyme from striated muscle. FEBS Lett 41: 292-296, 1974
ATPase activity is only partial [32]. Therefore, it is not 5. Pires E, Perry SV: Purification and properties of myosin light chain
unreasonable that MLCK, which only partially inhibited the kinase from fast skeletal muscle. Biochem J 167: 137-146, 1977
actin-activated ATPase activity (see above). 6. Leachman SA, Gallagher PJ, Herring BP, McPhaul MJ, Stull JT:
Biochemical properties of chimeric skeletal muscle and smooth muscle
The concentration of actin filaments in the motility assay
myosin light chain kinase. J BioI Chern 267: 4930-4938, 1992
is extremely low, i.e. 3 nM. In the presence of60 nM MLCK, 7. Morgan M, Perry SV, Ottaway J: Myosin light chain phosphatase.
18.4% of actin filaments in terms of monomeric actin is Biochem J 157: 687~97, 1976
calculated to be present to be associated with actin if we 8. Sweeny HL, Bowman BF, Stull JT: Myosin light chain phosphorylation
assume the apparent dissociation constant ofMLCK to actin in vertebrate striated muscle: Regulation and function. Am J Physiol
264:CI085-e1095,1993
filament to be 0.22 11M (Fig. 1A). Therefore, partial asso-
9. Guerriero V Jr., Rowley DR, MeansAR: Production and characteriza-
ciation is sufficient for MLCK to exert the inhibition. tion of an antibody to myosin light chain kinase and intracellular
The actin-linked inhibition of MLCK was first observed localization of the enzyme. Cell 27: 449--458,1981
with MLCK from bovine stomach [12] and then with chicken 10. Ye L-H, Hayakawa K, Kishi H, Imamaura M, Nakamura A, Okagaki
gizzard [13] and bovine aorta [14]. The actin-binding domain T, Takagi T, IwateA, Tanaka T, Kohama K: The structure and function
of the actin-binding domain of myosin light chain kinase of smooth
of MLCK from gizzard smooth muscle is located in the
muscle. J BioI Chern 272: 32182-32189,1997
N-terminal portion between M' and M"4 [33], and this region II. Kohama K, Ye L-H, Hayakawa K, Okagaki T: Myosin light chain
has a calculated pI of 10.58. A similarly basic region exists kinase: An actin-binding protein that regulates an ATP-dependent
in the sequence ofskeletal muscle MLCK [6] from K250_N°O, interaction with myosin. Trends Pharmacol 17: 284-287, 1996
90
12. Kohama K, Okagaki T, Hayakawa K, Lin Y, Ishikawa R, Shimmen T, malachite green method for determination of inorganic phosphate. J
Inoue A: A novel regulatory effect of myosin light chain kinase from Biochem 99: 1465-1472, 1986
smooth muscle on the ATP-dependent interaction between actin and 25. Lin Y, Ishikawa R, Okagaki T, Ye L-H, Kohama K: Stimulation of the
myosin. Biochem Biophys Res Commun 184: 1204-1211,1992 ATP-dependent interaction between actin and myosin by a myosin
13. Ye L-H, Hayakawa K, Lin Y, Okagaki T, Fujita K, Kohama K: The binding fragment of smooth muscle caldesmon. Cell Motil Cytoskel
regulatory role of myosin light chain kinase as an actin-binding protein. 29: 250-258,1994
J Biochem 116: 1377-1382, 1994 26. Ishikawa R, Yamashiro S, Matsumura F: Differential modulation of
14. Sato M, Ye L-H, Kohama K: Myosin light chain kinase from vascular actin-serving activity of gelsolin by multiple isoforms of cultured
smooth muscle inhibits the ATP-dependent interaction between actin rat cell tropomyosin: Potention of protective ability oftropomyosins
and myosin by binding to actin. J Biochem 118: 1-3, 1995 by 83-kDa non-muscle caldesmon. J Bioi Chern 264: 7490-7497,
15. Nunnally MH, Rybici SB, Stull JT: Characterization of chicbn 1989
skeletal muscle myosin light chain kinase. J Bioi Chern 260: 1020- 27. Laemmli UK: Cleavage of structural protein during the assembly of
1026,1985 the head of bacteriophage T4. Nature 227: 680--685, 1970
16. Kohama K: Amino acid incorporation rates into myofibrillar proteins 28. Bradford M: A rapid and sensitive method for the quantitation of
of dystrophic chicken skeletal muscle. J Biochem 90: 497-50 I, 1981 microgram quantities of protein utilizing the principle of protein dye
17. Hayakawa K, Okagaki T, Higashi-Fujime S, Kohama K: Bundling of binding. Anal Biochem 72: 248-254,1976
actin filaments by myosin light chain kinase from smooth muscle. 29. Maita T, Yajima E, Nagata S, Miyanishi T, Nakayaka S, Matsuda G:
Biochem Biophys Res Commun 199: 788-91,1994 The primary structure of skeletal muscle myosin heavy chain: V.
18. Adestein RS, Klee CB: Purification and characterization of smooth Sequence of the rod, and the complete 1,938-residue sequence of the
muscle myosin light chain kinase. J Bioi Chern 256: 7501-7509, 1981 heavy chain. J Biochem 110: 75-78,1991
19. Perrie WT, Perry SV: An electrophoretic study of the low-molecular- 30. Elzinga M, Collins JH, Kuehl WM, Adelstein RS: Complete amino-
weight components of myosin. Biochem J 119: 31-38, 1970 acid sequence of actin of rabbit skeletal muscle. Proc Nail Acad Sci
20. Kohama K: Heterogeneity of amino acid incorporation rate in adult USA 70: 2687-2691,1973
skeletal muscle actin. J Biochem 87: 997-999,1980 31. Prochinietwicz E, Yanagida T: Inhibition of sliding movement of
21. Ishikawa R, Okagaki T, Higashi-Fujime S, Kohama K: Stimulation of F-actin by crosslinking emphasizes the role of actin structure in the
the interaction between actin and myosin by Physarum caldesmon like mechanism of motility. J Mol Bioi 216: 761-772,1990
protein and smooth muscle caldesmon. J Bioi Chern 266: 21784-21790, 32. Kohama K, Saida K, Hirata M, Kitaura T, Ebashi S: Superprecipitation
1991 is a model for in vitro contraction superior to ATPase activity. Japan J
22. OkagakiT, Higashi-Fujime S, Ishikawa R, Takano-Ohomuro H, Kohawa Pharmacol 42: 213-260, 1989
K: Requirement of phosphorylation of myosin light chain and effect of 33. Kanoh S, Ito M, Niwa E, Kawano Y, Hartshorne DJ: Actin-binding
tropomyosin and caldesmon. J Biochem 109: 858-866, 1991 peptide from smooth muscle myosin light chain kinase. Biochemistry
23. Kron SJ, Spudich JA: Fluorescent actin filaments move on myosin fixed 32: 8902-8907, 1993
to a glass surface. Proc Nail Acad Sci USA 83: 6272--6276, 1986 34. Olson NJ, Pearson RB, Needlman DS, Hurwitz MY, Kemp BE, Means
24. Kodama T, Fukui K, Kometani K: The initial phosphatase burst in AR: Regulatory and structural motifs of chicken gizzard myosin light
ATP hydrolysis by myosin and subfragment-I as studied by a modified chain kinase. Proc Natl Acad Sci USA 87: 2284-2288, 1990
Molecular and Cellular Biochemistry 190: 91-98, 1999.
© 1999 Kluwer Academic Publishers.

Ca2+-induced Ca2+ desensitization of myosin light

chain phosphorylation and contraction in phasic
smooth muscle

Tetsuya Murahashi, Akikazu Fujita and Toshio Kitazawa

Department ofPhysiology and Biophysics, Georgetown University School ofMedicine, Washington, DC, USA

Abstract
The temporal relationship between Ca 2+-induced contraction and phosphorylation of20 kDa myosin light chain (MLC) during
a step increase in Ca2+was investigated using permeabilized phasic smooth muscle from rabbit portal vein and guinea-pig ileum
at 25°C. We describe here a Ca2+-induced Ca 2+desensitization phenomenon in which a transient rise in MLC phosphorylation
is followed by a transient rise in contractile force. During and after the peak contraction, the force to phosphorylation ratio
remained constant. Further treatment with cytochalasin D, an actin fragmenting agent, did not affect the transient increase in
phosphorylation, but blocked force development. Together, these results indicate that the transient phosphorylation causes the
transient contraction and that neither inhomogeneous contractility nor reduced thin filament integrity effects the transient
phosphorylation. Lastly, we show that known inhibitors to MLC kinase kinases and to a Ca2+-dependent protein phosphatase
did not eliminate the desensitized contractile force. This study suggests that the Ca 2+-induced Ca 2+desensitization phenomenon
in phasic smooth muscle does not result from any of the known intrinsic mechanisms involved with other aspects of smooth
muscle contractility. (Mol Cell Biochem 190: 91-98, 1999)

Key words: Ca2+ion, Ca2+desensitization, phasic smooth muscle, phosphorylation, myosin light chain kinase, calcium/calmodulin
-dependent protein kinase

Introduction plasmic Ca 2+is a primary determinant of this activity ratio,

other physiological factors can regulate the Ca2+sensitivity
It is generally accepted as a fact that Ca 2+ ion is the most in smooth muscle [3,4].
fundamental intracellular signaling messenger in the initiation Depolarization-induced contractile force in intact phasic
of contraction of a wide variety of muscles [I]. Smooth smooth muscle declines to low steady-state levels during
muscle provides no exception as its contraction is also prolonged incubation in high K+ solutions, while the force
mediated by increases in cytoplasmic free Ca 2+. These of tonic smooth muscle in identical conditions is elevated
increases activate the Ca 2+/calmodulin (CaM)-dependent and sustained. However, Ca 2+ levels in the two smooth
myosin light chain kinase (MLCK) and thereby phosphorylate muscle types have not been shown to differ significantly
the 20 kDa MLC at Ser-19, which in tum elicits a contraction during the maintained depolarizations [5]. Therefore, it had
by activating actomyosin ATPase [2, 3]. Subsequent de- been hypothesized that MLC phosphorylation processes in
creases in cytoplasmic Ca 2+ lead to decreases in MLCK phasic smooth muscle were desensitized to cytoplasmic Ca 2+
activity that allow for dephosphorylation of MLC by MLC with time. To test that hypothesis, Kitazawa and Somlyo
phosphatase (MLCP) and then relaxation. Thus the smooth (1991) used a-toxin-permeabilized ileum smooth muscle
muscle contraction/relaxation cycle is ultimately regulated by where the cytoplasmic Ca 2+ could be clamped by a Ca2+-
the phosphorylation level of MLC which is determined by buffer [6]. A Ca2+step-increase induced increases in MLC
the activity ratio of MLCK to MLCR. Although the cyto- phosphorylation and in force, both of which declined

Address for offprints: T. Kitazawa, Department of Physiology and Biophysics, Georgetown University School of Medicine, 3900 Reservoir Road, N.W.
Washington, D.C. 20007, USA
92

spontaneously to low steady-state levels despite [Ca 2+]j A preliminary report of these results has been presented
remaining constant. This confirmed the hypothesis of a at the Annual Meeting of the American Biophysical Society
Ca2+-induced desensitization to Ca2+ofMLC phosphorylation [ 15].
and force development. However, measured levels of MLC
phosphorylation after peak contraction were at only one point
(10 min). The purpose of the present study was to more Material and methods
precisely observe the comparative time courses between
phosphorylation and force during this desensitization in All animal procedures were approved by the Animal Care and
phasic smooth muscle tissue and to furthermore investigate Use Committee of the Georgetown University. The rabbits
the extent to which three separate mechanisms may account and guinea-pigs were killed by inhalation of, respectively,
for it. halothane and CO 2, The portal vein and ileum were isolated
It has been shown that both MLC phosphorylation and from 2-3 kg male white rabbits and 300-400 g guinea-pigs,
active force of smooth muscle characteristically depend on respectively. Small strips (800 /lm wide, 40-80 /lm thick and
the muscle length in intact [7] and skinned tissues [8]; any 3 mm long) of longitudinal smooth muscle from portal vein
movement away from the optimal length decreases both and ileum were carefully dissected and freed of connective
phosphorylation and contraction. This likely occurs in soft, tissue with a small razor blade. To measure isometric force,
phasic smooth muscle due to unavoidable, inhomogeneous strips were tied with monofilament silk to the fine tips oftwo
shortening and stretching of contractile regions and is tungsten needles, one ofwhich was connected to a force trans-
therefore one possible mechanism to explain the desensitiza- ducer (AM801, SensoNor, Horten, Norway), and mounted in
tion. To test that possibility, we applied cytochalasin D, a a well on a bubble plate to allow for rapid solution exchange
fungal metabolite that practically abolishes contraction of by sliding the plate to an adjacent well and for rapid freezing
intact smooth muscle [9] through fragmentation of actin within a second for measurement of MLC phosphorylation.
filaments by capping their barbed ends [10]. By applying
cytochalasin D, we were able to eliminate any inhomo-
geneous shortening and stretching, while still observing the Solutions
time course and levels of the phosphorylation.
The Ca2+-induced phosphorylation ofMLCK at a regulatory The normal external solutions for intact preparations were
site (site A) is a second possible desensitizing mechanism. neutralized with Tris(hydroxymethyl)aminomethane to give
Phosphorylation at site A results in a higher K m ofMLCK for pH of7.4 at 25°C and contained (mM): NaCl (150); KCl (4);
the Ca 2+/CaM complex that leads to a decrease in Ca 2+sensi- calcium methanesulfonate (CaeMs) (2); MgeMs (2); glucose
tivity ofMLCK activity. Hence, the desensitization would be (5.6); and HEPES (N-2-hydroxyethylpiperazine-N'-2-ethane-
explained if elevated Ca 2+were to induce both MLCK phos- sulfonic acid) (5). To prevent the cells from swelling in de-
phorylation at site A and MLCK activation, with the latter polarizing solutions where [K+] was high, these solutions were
occurring more rapidly. MLCK kinases which are known to made by replacing NaCI in the normal external solution with
phosphorylate at site A in vitro are protein kinase A (PKA), an equimolar concentration ofKeMs, thereby maintaining the
protein kinase C (PKC) and Ca 2+/calmodulin-dependent product of [K+] and [Cl-] constant. In Ca2+-free external solu-
protein kinase II (CaMK II) ([11, 12] and see [3] for other tions, CaeMs was omitted and replaced by 2 mM ethyleneglycol-
references). In smooth muscle culture cells, Tansey et al. bis-(p-aminoethylether)N-N'-tetraacetic acid (EGTA).
(1994) have demonstrated that CaMK II inhibitors decrease The normal relaxing solutions for permeabilized pre-
MLCK phosphorylation at site A and thereby increase the parations were neutralized to pH 7.1 with KOH at 25°C and
Ca2+sensitivity of MLCK activity, indicating that CaMK II contained (mM): KeMs (40.9), Mg2+ (2), MgATP (10), EGTA
is involved in the Ca2+desensitization [13]. To test this second (I), creatine phosphate (10), PIPES (piperazine-N,N'-
mechanism, we looked to see if inhibitors for each of the bis(2-ethanesulfonic acid)) (30). In the activating solutions,
aforementioned MLCK kinases altered the desensitization 10 mM EGTA was used, and specific amounts ofCaeMs were
phenomenon. added to yield a desired concentration of free Ca2+ions. Free
A third possibility is that Ca 2+ activates Ca2+-dependent Ca2+ and Mg 2+concentrations were calculated on a basis of
MLCP(s) with slower kinetics or in a delayed manner relative stability constant values for complexes in solutions as
to the Ca 2+-induced activation of MLCK. The only known described in Horiuti [16]. 10 mM MgATP and 10 mM
Ca 2+-regulated phosphatase is calcineurin, or protein phos- creatine phosphate was used to prevent possible ATP de-
phatase 2B, which exists abundantly in the brain and con- pletion inside cells and maintain a constant intracellularATP
siderably in muscle cells as well [14]. To test this contingency, concentration. 0.2 M ionic strength was achieved in activating
we observed the time course ofthe transient phosphorylation and modified relaxing solutions by adjusting the concentra-
after application of potent calcineurin inhibitors. tion ofKeMs.
 93

Cell permeabilization standard relaxing solution containing 10 11M A23187 for 30

min at 25°C. For the usage ofTriton X-I 00, we simply used
Whenever possible, we used Staphylococcus aureus a-toxin 0.1 % in standard relaxing solution for 45 min at 4°C.
(Gibco BRL) to permeabilize smooth muscle from rabbit
portal vein to see the Ca2+-induced phasic contractions. We
selected a-toxin since it minimally affects the membrane Determination of myosin light chain phosphorylation
associated pathways that often are involved with smooth
muscle contraction [17, 18] and we decided upon rabbit portal For determining MLC phosphorylation, permeabilized strips
vein as our source of phasic smooth muscle because rabbit were rapidly frozen in liquid N 2-cooled liquid chloro-
tissue is very sensitive to Gibco a-toxin. However, in order difluoromethane under various conditions. The MLC phos-
to rapidly introduce the desired inhibitor molecules of this phorylation was measured as described previously [17]. After
study which are larger than 1 kDa, p-escin instead ofa-toxin incubation in acetone containing 10% trichloroacetic acid at
was used to permeabilize strips [6]. The pores created by -80°C overnight, the frozen strips were gradually warmed up,
a-toxin do not accommodate the diffusion of high molecular washed first with fresh acetone and then with ether before
weight solutes while those created by p-escin do. One being allowed to dry. The dried strips were thoroughly
complication with p-escin, however, is that it often dampens homogenized in glycerol sample buffer (containing 1% SDS,
phasic contractions of smooth muscle, especially in rabbit 10% glycerol, 20 mM DTT and 0.1 mg/ml BSA). First
portal veins. Hence, when permeabilizing with p-escin to dimensional electrophoresis of an isoelectric focusing
apply the inhibitors, we selected the longitudinal layer of polyacrylamide tube gel containing 5% pH ampholytes 4.5-
smooth muscle from guinea pig ileum; the ileum retains its 5.4 (Pharmacia; Uppsala, Sweden) and 9.2 M urea ran
Ca 2+-induced phasic contractile nature after treatment with overnight, until a steady current was obtained, to separate the
p-escin [6]. Since the sensitivity of Gibco a-toxin was proteins in the supernatant by charge. The second dimen-
several-fold lower in guinea-pig smooth muscle than in sional SDS polyacrylamide gel electrophoresis then separated
rabbit, we did not utilize the former tissues for experiments the proteins by molecular weight; the appropriate portion of
using a-toxin-permeabilized strips in terms of cost. We also the first dimensional gel laid horizontally on the top of the
used the detergent Triton X-I00 to heavily permeabilize SDS stacking gel before it was run. Electrophoretic transfer
strips. Such a permeabilization uncouples receptors from G was carried out from SDS polyacrylamide gels to nitro-
proteins as well as other signal transduction pathways and it cellulose membranes for 5 h in 25 mM Tris, 194 mM glycine
allows small and large soluble, cytosolic molecules to escape and 20% methanol. The membranes were washed with a
from the fibers. The advantage associated with this technique phosphate buffer solution containing 0.3% Tween 20 over-
is that it had allowed us to see more ofa direct effect ofCa 2+ night and then stained finally with colloidal gold (Bio-Rad;
on the contractile and regulatory machinery. Hercules, CA, USA).
Before permeabilizing the cells, we first measured con- The colloidal gold-stained patterns ofMLC were digitized
tractions induced by high K+ (154 mM) and by agonists (100 with a color scanner (Macintosh, Cupertino, CA, USA) and
11M phenylephrine in portal vein and 10 11M carbachol in analyzed with image processing software (Signal Analytics
ileum) at 30°C to examine the integrity of the preparations. Co., Vienna, VA, USA) that permitted the subtraction of
Then, the strips were incubated in relaxing solution for background obtained from regions adjacent to the focused
several minutes. For permeabilization with a-toxin, the strips proteins. The percentage of MLC phosphorylation was cal-
from portal vein were then treated for 30 min at 30°C with culated by dividing the density ofmono- and di-phosphorylated
5,000 units/ml of purified Staphylococcus aureus a-toxin spots by the combined density of un-, mono-, and di-
(Gibco BRL; Gaithersburg, MD, USA) at pCa 6.7 buffered phosphorylated spots [17].
with 10 mM EGTA. This condition facilitated the high-
temperature-dependence of permeabilization and simul-
taneously enabled one to examine the efficiency of the Miscellaneous
permeabilization. When the SR was depleted ofCa2+and the
cytoplasmic Ca 2+ was held constant, strips were further We purchased the Staphylococcus aureus a-toxin from
treated with the Ca2+ionophore 10 11M A23187 (Calbiochem; Gibco/BRL, the A23187 from Calbiochem, the p-escin,
San Diego, CA, USA) for 2(}-25 min in the relaxing solution Triton X-I 00,4,4'-diisothiocyanatostilbene-2,2'- disulfonic
[17]. Thereafter, presence of 10 11M ryanodine, 10 11M acid (DIDS) and cytochalasin D from Sigma, the PKA
thapsigargin or 100 J.lM inositol 1,4,5-trisphosphate did not inhibitor Rp-cAMPS from BioLog (La Jolla, CA, USA), the
affect the time course ofcontraction induced by constant Ca2+. cypermethrin from LC Lab. (Woburn, MA, USA), the protein
For plasma membrane permeabilization with p-escin, the kinase C inhibitor peptide (Peptide 19-31) and all CaMK II
strips were then instead treated with 40 J.lM p-escin in inhibitor peptides from BioMol (Plymouth Meeting, PA,
94

USA). We also obtained other specific inhibitors of CaMK 5 min

II, the isoquinoline sulfonamide derivatives KN-62 [19] and A
KN-93 [20], from Seikagaku (Rockville, MD, USA). The last
two were prepared in 100% dimethyl sulfoxide and H 20,
respectively. However, when an aliquot of their stock solu-
tions was added into the relaxing solution to make final
concentrations of 1-10 11M, they were precipitated, even at
a final concentration of 2% dimethyl sulfoxide or I% fatty
acid-free BSA. Consequently, we could not definitively
observe either effect of KN-62 or KN-93 on the transient
contractions ofa-toxin-permeabilized rabbit portal vein and pCa 6.3
z
p-escin-permeabilized guinea pig ileum smooth muscles. E
The following peptides to inhibit protein kinases were
used; PKC-IP(Arg-Phe-Ala-Arg-Lys-Gly-Ala-Leu-Arg-Gln-
Lys-Asn-Val), CaMK-IP (281-309) (Met-His-Arg-Gln-Glu-
Thr-Val-Asp-Cys-Leu-Lys-Lys-Phe-Asn-Ala-Arg-Arg-Lys-
Leu-Lys-Gly-Ala-Ile-Leu-Thr-Thr-MEt-Leu-Ala), CaMK-IP
[Ala-286](28 1-302) (Met-His-Arg-Gln-Glu-Ala-Val-Asp-
Cys-Leu-Lys-Lys-Phe-Asn-Ala-Arg-Arg-Lys-Leu-Lys-Gly-
Ala), CaMK-IP [Ala-9] autocamtide-2 (AlP; Lys-Lys-Ala-
Leu-Arg-Arg-Gln-Glu-Ala-Val-Asp-Ala-Leu).

Results pCa6
The Ca + induced Ca + desensitization ofcontractile force
l l
Fig. I. Difference between time courses ofCa'+·induced force develop-
and MLC phosphorylation ment of a-toxin- (A) and Triton X-I OO-permeabilized rabbit portal vein
smooth muscle strips (B) at 25°C. (A) a-toxin-permeabilized, A23187-
Figure I graphically illustrates the marked difference between treated strips were first incubated in the relaxing solution containing 0.1
time courses ofCa 2+-induced force development ofa-toxin- mM EGTA for 15 min and activated with Ca 2+(pCa 6.3) containing solution
and Triton X-I OO-permeabilized rabbit portal vein smooth buffered with 10 mM EGT A. The strips were then relaxed in the relaxing
solution containing 10 mM EGTA and no added Ca2+. The transient
muscle strips. A step increase in Ca 2+to pCa 6.3 in a-toxin-
contractile response reproducible for at least several times could not be
permeabilized, A23187-treated rabbit portal vein strips due to a transient increase in cytoplasmic Ca2+ since the strips were treated
caused a rapid rise in force immediately followed by a with 10 11M A23187 and the Ca'+ concentration in the activating solution
spontaneous decline to a low steady state level (Fig. I A), was buffered with 10 mM EGTA. (B) Strips were permeabilized with 0.1 %
similar to that observed in the a-toxin-permeabilized guinea- Triton X- 100 for 45 min at 4°C. Both 0.1 mM EGTA-relaxing and pCa 6
(10 mM EGTA)-activating solutions contained I 11M CaM. All solutions
pig intestinal smooth muscle [6]. In contrast, Triton X-100-
used contained 10 mM MgATP and 10 mM creatine phosphate.
demembranated strips at pCa 6 with I 11M CaM showed a
monotonic increase in force (Fig. IB). Decreasing either CaM
or Ca 2+concentrations in the latter preparations led to a slower a decrease in the creatine phosphate to 5 mM reduced the
development and a lower steady state level of contraction but relaxation rate to 66 ± 8% of control. Strips were pretreated
did not cause the responses to become phasic. with 5 mM DIDS for 30 min in the normal relaxing solution
Substantially high non-specific ecto-ATPase activity before a-toxin-permeabilization and 10 mM NaN 3 was there-
associated with plasma membrane has been demonstrated in after added into the solutions [22]. This protocol significantly
intact, a-toxin- and p-escin-permeabilized, but not Triton reduced the spontaneous relaxation rate to 53 ± 7% ofcontrol
X-IOO-demembranated smooth muscles [21, 22]. Hence, it and raised the final steady state levels of the phasic pCa
was worthwhile to see what effect on the phasic behavior 6.3-induced contraction, but the phasic responses did not
might come about by minimizing change in nucleotide become tonic. Interestingly, treatment with DIDS and NaN 3
concentrations in the activated muscles and lowering the high also decreased to 54 ± 7% of control without treatments the
resting ATPase activity. In a-toxin-permeabilized rabbit rate of relaxation induced by removal of Ca2+and addition
portal vein strips, an increase in creatine phosphate as an of 10 mM EGTA at steady state in the phasic contraction of
ATP-regenerating system from control (10 mM) to 20 mM a-toxin-permeabilized strips. In addition, this treatment did
did not affect the time courses of spontaneous relaxation, but markedly potentiate a submaximal contraction induced by
 95

pCa 6 with 0.1 l-lM CaM in Triton XI00-demembranated declining and steady-state phases of contraction (regression
preparations. coefficient of 0.97) (Fig. 2B). Furthermore, the low steady-
As shown in Fig. 2A, during the first 0.5 min of con- state levels ofMLC phosphorylation at 1.5 min did not differ
traction, increases in contractile force by Ca 2+were preceded significantly from those at 30 min.
by increases in MLC phosphorylation, but there was no linear
correlation between phosphorylation and force in this rising
phase. However, the two were tightly coupled during the Effect ofcytochalasin D

Cytochalasin D, at concentrations ranging from 30-100 l-lM,

-....
A completely abolished the Ca 2+-induced contraction in the
( ,) 100 _ permeabilized portal vein, but had no significant effect on
It)
:::E 70 III
either the time course or the levels ofMLC phosphorylation,

..
ii
0
80 ~
'tii
even at 100 l-lM (Fig. 3).

-
'0 E
0~ SO 60 ~ A
c ')(
0 III
:;:::
III 40 E 100
~ ~----===d '0
e:.
--'#...
0
.l:
80

i
Co
III --0- phosphorylation 20
0 ---.- force
.l:

-
a. Q)
60
10 ",-~-----'--~---'--~_----lO (,)

0 2 3 0
Time (min) Q) 40
>
;:
as
B G)
a: 20
100

80
o 1 2

--
0~ 60 >--0-<
0.25
B
Time (min)

...
Gl
U
0 40
11.

--
~
20 0.17 (min)
'#.
00 20 60 80 c
o
;:
as
Phosphorylation (%)
~
o
r.
Fig. 2. Relationship between MLC phosphorylation and contractile force Q.
during Ca'+-induced Ca'+ desensitization. (A) Transient time courses of ~
MLC phosphorylation (e) and contractile force (0) induced by a r.
0..
step-increase in Ca'+ to pCa 6.3 in a-toxin-permeabilized, rabbit portal
vein smooth muscle. After steady-state maximum contractions were 10 L-_-'-_----Jl....-_~_--.L_~
obtained at pCa 5, the strips were brought back to the relaxed states in the o 1 2
Ca'+-free solutions containing 10 mM EGTA for 5 min and then 0.1 mM
for 15 min. Thereafter, the strips were subjected to a pCa 6.3 - containing Time (min)
solution buffered with 10 mM EGTA at zero min and frozen at various
times. (B) A good correlation between MLC phosphorylation and contractile Fig 3. Effects of 100 11M cytochalasin D (filled symbol) on the contractile
force during the descending and steady-state phases of contraction (.), force (A) and MLC phosphorylation (B) induced by a step-increase in Ca'+.
but not in the ascending phase (0). The equation for solid interpolated line Control data (open symbol) in both A and B are the same as those of Fig 2.
is F =2.09 x P - 41.8 (regression coefficient of 0.974). F and P represent, In the cytochalasin D experiments, the drugs were applied at dark 20 min
respectively, relative force and phosphorylation ofMLC. before and after increases in Ca'+.
96

MLCK regulation MLCP regulation

Via stimulatory action on PKC, phorbol-12, 13-dibutyrate Cypermethrin is an insecticidal type II pyrethroid and is known
(PDBu) and diacylglycerol (DAG) are known to potentiate as a very specific calcineurin inhibitor with an IC so ranging
smooth muscle contractions despite constant Ca 2+[18]. Some from 10-11 to 10-10 M [28]. Our results show that this compound
isoforms ofPKC are Ca 2+-dependent [23] so we chose a PKC at 20 11M had no effect on the amplitude or the time course of
inhibitor peptide (Peptide 19-31) that corresponds to a pCa 6.1 - induced transient contraction (Table 1).
pseudosubstrate region of these Ca 2+-dependent PKC iso-
forms. Peptide 19-31 at 30 11M inhibited such PDBu- and
synthetic short chain DAG (I ,2-sn-dihexanoyIglycerol; Discussion
diC 6 )-induced (PKC-mediated) potentiation to 8 ± 4 and 18
± 2% of control, respectively, thereby displaying Peptide 19- Time courses of Ca 2+-induced MLC phosphorylation and
31 's potent inhibition of PKC. However, Peptide 19-31 did force development in permeabilized phasic smooth muscle
not alter the amplitude or the time course of the transient have revealed that the increases in phosphorylation and force
contractions in the ~-escin permeabilized guinea-pig ileum development during a Ca 2+step-increase are indeed transient.
smooth muscle strips (Table I). The results of this study are in excellent agreement with the
cAMP is known to cause smooth muscle relaxation through hypothesis made from studies in un-permeabilized, intact
its activation of PKA. We selected a PKA inhibitor, the phasic smooth muscle that the contractile system becomes
Rp-isomer of adenosine-3',5'-cyclic monophosphorothioate desensitized to Ca2+[5]. Our results further indicate that Ca2+-
(Rp-cAMPS), which at 200 11M has been shown to decrease desensitization of MLC phosphorylation and force occurs
cAMP-induced (PKA-mediated) relaxation [24]. This com- rapidly. When the phasic smooth muscle was heavily skinned
pound also had no effect on the transient contractions in our with triton X-I 00 (Fig. I) or saponin (unpublished data),
study (Table I). however, the Ca 2+-induced contractions became tonic. This
Three separate peptides that are known to inhibit isolated suggests that some factor(s), capable of diffusing out of the
CaMK II were used. The first inhibitor peptide (Peptide cell through large pores in the plasma membrane, and/or
281-309) has a sequence identical to the pseudosubstrate constituents ofthe membrane of those preparations may play
regulatory region of CaMK II which includes an intact an important role in the transient MLC phosphorylation and
autophosphorylation site Thr-286 and a CaM binding contraction.
domain. This is a potent CaM antagonist with an IC so of 80 An important finding of this study is that during and after
nM and inhibits the active site ofCaMK II with an IC so of2 the peak of transient Ca 2+-induced contractions, MLC phos-
11M [25,26]. The second (Peptide [Ala-286] 281-302) has a phorylation and force development were tightly coupled at
sequence similar but not identical to the first Peptide 281- a constant Ca 2+. This finding indicates that during the
309; the autophosphorylatable Thr-286 is replaced with Ala dephosphorylation phase, neither significant formation of
and the calmodulin-binding site is absent. The IC so for new latch bridges nor significant contribution of thin-
inhibiting CaMK II activity is around 211M [26,27]. The third filament-linked regulation occurs.
peptide ([Ala-9] autocamtide-2; AlP) has a sequence similar Although length-dependent contraction and MLC phos-
to autocamtide-2 except replacement of phosphorylatable phorylation have been demonstrated in skinned smooth
Thr-9 with unphosphorylatable Ala and is a potent and muscle under constant Ca2+ [8], this study indicates that the
specific inhibitor toward CaMK II with an IC so of 40 nM length dependence does not sufficiently explain the de-
without any effects on PKA, PKC and CaMK IV activities sensitization phenomenon of phasic smooth muscle. Pre-
at 10 11M [27]. Interestingly, the latter two peptides at 10-30 vious research has already demonstrated that changing the
11M did not significantly affect the transient contractions at muscle length from 0.8-1.9 times the slack length has no
all (Table I), but the peptide 281-309 dose-dependently effect on the time course of transient contractions [6].
decreased the peak and steady-state amplitudes of transient Furthermore, the results using cytochalasin D in the present
contractions without modifying the transient time course, study demonstrate the point even more conclusively; the
including the time to peak of contraction and the halftime of elimination of contraction/relaxation events did not affect
relaxation (Table I). Such inhibition of both peak and the time course or levels of MLC phosphorylation. There-
steady-state contractions by peptide 281-309 was overcome fore, it is likely that the transient aspect of Ca 2+ -induced
by addition of 3 11M CaM. The possibility that Peptide contractions is brought about by biochemical (phosphoryla-
[Ala-286] 281-302 and [Ala-9] autocamtide-2 were too large tion) events that are independent of mechanical events. It
to diffuse through the pores is unlikely because they are also seems likely from this study that the transient MLC
smaller than both CaM and Peptide 281-309, which both phosphorylation induced by Ca 2+ is independent of thin
modified contractions in the same preparations. filament configuration, contrary to the proposed dependence
 97
Table I. Effects of various inhibitors on transient contractions activated by pCa 6.1 in ~-escin-permeabilized longitudinal smooth muscle of guinea pig
ileum.

Drugs peak force time to peak steady-state force tl/,ofrelaxation

(% of control) (min) (% of peak) (min)

none 100 0.42 ± 0.02 22 ± 2 0.54 ± 0.03

+ 30 J.lM PKC-IP 108 ± 9 0.36 ± 0.03 23 ±6 0.51 ± 0.08
(Peptide 19-31)
+ 200 J.lM Rp-cAMPS 98 ± 0.50 ± 0.04 22 ± 2 0.48 ± 0.04

+ 10 J.lM CaMK-IP 37 ± 10* 0.47 ± 0.04 6 ± 1* 0.50 ± 0.05

(Peptide 281-309)
+ 10 J.lM CaMK-IP 98 ± 2 0.46 ± 0.05 19 ±4 0.56 ± 0,07
(Peptide [Ala286] 281-302)
+ 10 J.lM CaMK-IP 91 ± 6 0.45 ± 0.02 20 ±3 0.56 ± 0.03
([Ala9] autocamtide-2; AlP)
+ 20 J.lM cypermethrin 90 ± 3 0.46 ± 0.03 18 ±2 0.46 ± 0.02

~-escin-permeabilized ileum smooth muscle strips were first incubated in the O. I mM EGTA relaxing solution for 15 min and activated with Ca'+ (pCa
6.1 )-containing solution buffered with 10 mM EGTA at 25°C. Time courses of the transient contractile response in ~-escin-permeabilized ileum smooth
muscle strips were similar to that of o.-toxin-permeabilized rabbit portal vein illustrated in Fig. IA. Results are mean ± S.E.M. IP = inhibitor peptide. n = 3-
6. *significantly different from control (p < 0.01).

ofMLCK on binding to actin [29]. Taken together, these data important to mention here that this DIDS-NaN 3 protocol did
strongly suggest the transient MLC phosphorylation is an not abolish the phasic component of contraction, suggesting
intrinsic process during Ca 2+-induced contraction of phasic that the transient contraction in a-toxin-permeabilized phasic
smooth muscle. smooth muscle is not mainly due to metabolic changes
It has been recently mentioned that plasma membrane- induced by Ca 2+.
associated high ecto-ATPase activity is present in both The regulation of MLCK seems more likely as supported
a-toxin-permeabilized and unpermeabilized intact smooth by the work of Tansey et at. with select CaMK II inhibitors
muscle (21), causing a reduction of intracellular [ATP] and [13]. In the present study, however, the same and other CaMK
an accumulation of [ADP] in the former preparations [22]. II inhibitors neither reduced the rate of spontaneous relaxa-
Those metabolic changes may greatly affect the contractile tion nor increased low steady state levels of the transient
response and could theoretically give rise to a transient contraction (see Table I and also Materials and methods). The
contraction. The presence of creatine phosphate and creatine counterintuitive inhibition of contraction by the longer CaMK
kinase as an ATP-regenerating system has been shown to II inhibitor, Peptide 281-309, and the removal of its inhibition
greatly reduce abnormally high cellular ADP/ATP ratio that by CaM simply demonstrate that Peptide 281-309 acts as a
occurs in the absence of the regenerating system [22]. In a- strong CaM antagonist. Although the inhibition ofCaMK II
toxin-permeabilized strips, an increase in creatine phosphate activity by the inhibitors was not established in the per-
from :s 5 to 10-20 mM led to increasing the rate of spon- meabilized smooth muscle, the data on three separate CaMK
taneous relaxation. This is consistent with the fact that a II inhibitors suggests that the phosphorylation of MLCK by
decrease in [ADP] and/or inADPI ATP ratio cause an increase CaMK II is not a major mechanism in the Ca 2+-induced Ca2+
in the rate of relaxation induced by removal ofCa2+ [30]. As desensitization ofpermeabilized phasic smooth muscle. The
an alternative method to maintaining adequate ATP energy conflicting results between our findings and Tansey et ai. may
supplies, treatment with 5 mM DIDS and the presence of 10 be attributed to the participation ofmultiple mechanisms and/
mM NaN 3 in combination with the presence of creatine or different experimental conditions, namely tissue versus
phosphate have been shown to further improve the metabolic cultured cells. We are currently investigating in our laboratory
profile and maintain high [ATP] and low [ADP] under both whether or not MLCK activity decreases spontaneously
resting and activated conditions [22]. However, this ATP during Ca2+-induced transient contraction ofthe permeabilized
favoring DIDS-NaN 3 protocol unexpectedly increased the smooth muscle.
level of steady state submaximal contraction at constant Ca2+ The possible participation ofPKA, PKC or calcineurin in
even in Triton X-I OO-demembranated strips and reduced the the context of this study, that being their potential activation
rate of relaxation induced by removal of Ca2+at the steady by Ca 2+, had not been tested. In this study, the Ca 2+-induced
state tension level in a-toxin-permeabilized strips. Those transient contractions were not significantly modified by the
effects appear to be not related to the inhibitory effect of the PKA inhibitor Rp-cAMPS, by the PKC inhibitor Peptide 19-
DIDS-NaN 1 protocol on the ecto-ATPase. Nevertheless, it is 31 or by the calcineurin inhibitor cypermethrin. Therefore,
98

it is unlikely that PKA, PKC or calcineurin plays a critical 13. Tansey MG, Lubyphelps K, Kamm KE, Stull JT: Ca2'-dependent
role in the Ca 2+ desensitization in phasic smooth muscle. phosphorylation of myosin light chain kinase decreases the Ca 2'
sensitivity oflight chain phosphorylation within smooth muscle cells.
Taken together, these findings suggest that the Ca2+ -induced 1 Bioi Chern 269: 9912-9920, 1994
Ca 2+ desensitization phenomenon in phasic smooth muscle 14. Walsh MP, Busaan LJ, Fraser DE, Fu SY, Pato MD, Michalak M:
primarily results from unknown intrinsic mechanism(s). Characterization of the recombinant C-terminal domain of dystrophin:
Phosphorylation by calmodulin-dependent protein kinase II and
dephosphorylation by type 2B protein phosphatase. Biochemistry 34:
5561-5568, 1995
Acknowledgments 15. Murahashi T, FujitaA, Kitazawa T: Transient increase in myosin light
chain (MLC) phosphorylation at constant Ca2 ' in permeabilized smooth
I (TK) heartily thank Professor Setsuro Ebashi for his muscle. Biophys J 68: A277 (abstract), 1995
16. Horiuti K: Mechanism of contracture on cooling of caffeine-treated
teaching and encouraging me. Without those days in his frog skeletal muscle fibres. J Physiol (Lond) 398: 131-148, 1988
laboratory in 1969-1975, my career as a scientist could not 17. KitazawaT, Gaylinn BD, Denney GH, SomlyoAP: G-protein-mediated
be thought. We also thank Matthew R. Lee (Department of Ca" -sensitization of smooth muscle contraction through myosin light
Pharmaceutical Chemistry in the School of Pharmacy at the chain phosphorylation. J Bioi Chern 266: 1708-17 I 5, J 991
University of California, San Francisco) for his editorial 18. Masuo M, Reardon S, Ikebe M, Kitazawa T: A novel mechanism for
the Ca" -sensitizing effect of protein kinase C on vascular smooth
assistance. The work was supported by NIH RO 1 HL51824 muscle: Inhibition of myosin light chain phosphatase. 1 Gen Physiol
to TK. 104: 265-286,1994
19. Tokumitsu H, Chijiwa T, Hagiwara M, Mizutani A, Terasawa M,
Hidaka H: KN-62, 1-[N,O-bis(5-isoquinolinesulfonyl)-N-methyl-L-
tyrosyl]-4-phenylpiperazine, a specific inhibitor of Ca"/calmodulin-
References dependent protein kinase II. J Bioi Chern 265: 4315-4320, 1990
20. Sumi M, Kiuchi K, IshikawaT, IshiiA, Higawara M, NagatsuT, Hidaka
I. Ebashi S: Regulation of muscle contraction. Proc Roy Soc Lond, Series H: The newly synthesized selective Ca"/calmodulin dependent protein
B 207: 259-286, 1980 kinase II inhibitor KN-93 reduces dopamine contents in PC 12h cells.
2. Hartshorne OJ: Biochemistry of the contractile process in smooth Biochem Biophys Res Commun 181: 968-975, 1991
muscle. In: L.R. Johnson (ed). Physiology of the Gastrointestinal Tract. 21. Kerrick WG, Hoar PE: Relationship between ATPase activity, Ca'"
Raven Press, New York, 1987, pp 423-482 and force in a-toxin and p-escin-treated smooth muscle. Can J Physiol
3. Kamm KE, Stull JT: Regulation of smooth muscle contractile elements Pharmacol72: 1361-1367,1994
by second messengers. Ann Rev Physiol 51: 299-3 I3, 1989 22. Trinklemulcahy L, Siegman MJ, ButlerTM: Metabolic characteristics
4. Kitazawa T, Gaznabi AKM, Murahashi T: Regulation of Cal' of a-toxin permeabilized smooth muscle. Am J Physiol 266: C I673-
sensitivity of smooth muscle contraction by myosin light chain C1683,1994
phosphatase. In: K. Maruyama, Y. Nonomura, K. Kohama (eds). 23. Nishizuka Y: The molecular heterogeneity of protein kinase C and its
Calcium as Cell Signal. Igakushoin Ltd., Tokyo, 1995, pp 185-193 implications for cellular regulation. Nature 334: 661--665, 1988
5. Himpens B, Matthijs G, Somlyo AP: Desensitization to cytoplasmic 24. Gailly P, Boland B, Paques C, Himpens B, Casteels R, Gillis JM:
Cal' and Cal' sensitivities of guinea-pig ileum and rabbit pulmonary Post-Receptor pathway of the ATP-induced relaxation in smooth
artery smooth muscle. J Physiol (Lond) 413: 489-503,1989 muscle of the mouse vas deferens. Br J Pharmacol 110: 326-330, 1993
6. Kitazawa T, Somlyo AP: Desensitization and muscarinic re- 25. Payne ME, Fong Y-L, Ono T, Colbran RJ, Kemp BE, Soderling TR,
sensitization of force and myosin light chain phosphorylation to Means AR: Calcium/calmodulin-dependent protein kinase II:
cytoplasmic Ca" in smooth muscle. Biochem Biophys Res Commun Characterization of distinct calmodulin binding and inhibitory domains.
172: 1291-1297,1990 J Bioi Chern 263: 7190-7195,1988
7. Hai CM: Length-dependent myosin phosphorylation and contraction 26. Waxham MN, Malenka RC, Kelly PT, Mauk MD: Calcium/calmodulin-
of arterial smooth muscle. Pflilgers Arch 418: 564-571, 1991 dependent protein kinase II regulates hippocampal synaptic trans-
8. Moreland RS, Moreland S, Murphy RA: Dependence of stress on mission. Brain Res 609: 1-8, 1993
length, Ca", and myosin phosphorylation in skinned smooth muscle. 27. Ishida A, Kameshita I, Okuno S, Kitani T, Fujisawa H: A novel highly
Am J Physiol 255: C473-e478, 1988 specific and potent inhibitor of calmodulin-dependent protein kinase
9. Obara K, Yabu H: Effect ofcytochalasin b on intestinal smooth muscle II. Biochem Biophys Res Commun 212: 806-812, 1995
cells. Eur J Pharmacol 255: 139-147, 1994 28. Enan E, Matsumura F: Specific inhibition of calcineurin by type II
10. Cooper JA: Effects of cytochalasin and phalloidin on actin. J Cell synthetic pyrethroid insecticides. Biochem Pharmacol43: 1777-1784,
Bioi 105: 1473-1478, 1987 1992
I I. Hashimoto Y, Soderling TR: Phosphorylation ofsmooth muscle myosin 29. Dabrowska R, Hinkins S, Walsh MP, Hartshorne 01: The binding of
light chain kinase by Ca2+/calmodulin-dependent protein kinase II: smooth muscle myosin light chain kinase to actin. Biochem Biophys
Comparative study of the phosphorylation sites. Arch Biochem Res Commun 107: 1524-1531,1982
Biophys 278: 41-45,1990 30. Khromov A, Somlyo AV, Trentham DR, Zimmermann B, Somlyo A
12. Ikebe M, Reardon S: Phosphorylation of smooth myosin light chain P: The role of MgADP in force maintenance by dephosphorylated
kinase by smooth muscle Ca"/calmodulin-dependent multifunctional cross-bridge in smooth muscle: a flash photolysis study. Biophys J
protein kinase. J Bioi Chern 265: 8975-8978, 1990 69: 2611-2622,1995
Molecular and Cellular Biochemistry 190: 99-104, 1999.
© 1999 Kluwer Academic Publishers.

Growing and differentiating characterization of

aortic smooth muscle cell line, p53LMACOl
obtained from p53 knock out mice

Tsuyoshi Masuda,I,2 Kazuhiro Ohmi, I Hideki Yamaguchi,2,3 Kazuhide

Hasegawa,2 Tomoyasu Sugiyama,2 Yuzuru Matsuda,2 Masamitsu lino l
and Yoshiaki Nonomura2,4
'1st Department ofPharmacology, Faculty ofMedicine, University of Tokyo, Bunkyoku, Tokyo; 2 Vessel Research
Laboratory Co,. Ltd., in Tokyo Research Laboratories, Kyowa Hakko Kogyo Co., Machida-shi, Tokyo; 3Department of
Chemistry, Faculty ofScience and Technology, Sophia University, Chiyodaku, Tokyo; 4Cellular and Molecular Biology
Unit, School ofMedicine, Teikyo University, Hachioji-shi, Tokyo, Japan

Abstract
Recently we have established an aortic smooth muscle cell line, p53LMACOI obtained from p53 knockout mice. This cell line
showed some differentiated properties which were accelerated by 5-azacytidine treatment [I). In this study, further
characterization ofp53LMACOI cell line was investigated according to cell growth and differentiation, and especially focused
into the changes of cell feature, actin filaments' formation, and changes of intracellular calcium concentrations to sympathetic
nerve transmitter, norepinephrine. While the cell feature was changed from flattened shape to extended form during 4 days,
actin filaments were developing, arranging in parallel to longitudinal direction, and gathering under the surface membrane. In
II days many cells died and detached from substrate, while actin filaments became poor except for the surface membrane in
the remained cells. Appearance of calcium response to noradrenalin needed several days after passage as well as a morphological
change of the cells for the extended form and development of actin filaments. The calcium response was maintained on II
days, which coincided with the result that the cells hold actin filaments under the surface membrane. These results suggest that
p53LMACOI cell line maintains several differentiated characters of adult smooth muscle cell and that their expression needs
several days after passage. (Mol Cell Biochem 190: 99-104, 1999)

Key words: cell line, vascular smooth muscle, actin filaments' formation, expression ofh-caldesmon, expression ofa-receptor

Introduction knock out mice which was screened by expression of a-

smooth muscle actin as a smooth muscle marker. This cell
Developmental study ofvascular smooth muscle cell has been line expressed alpha smooth muscle actin, h-caldesmon, and
a subject of considerable interest as a vascular development calponin which were contractile type smooth muscle cell
itself and as a base in a study of a pathogenesis of athero- markers, and the latter two expressions were increased by 5-
sclerosis [2). But progress in this field has been hampered for azacytidine treatment [I). In this study, we describe further
a long time by a lack of the cultural system of well differen- morphological and functional characterization of this cell
tiated smooth muscle cells [3]. Recently we have establish line, especially focused into a relationship of cultivating time
an aortic smooth muscle cell line (p53LMACO I) from p53 in relation to cell differentiation.

Address for offprints: Y. Nonomura, Teikyo University, School of Medicine, 359 Otsuka, Hachioji-shi, Tokyo 192-0395, Japan
100

Materials and methods Immunofluorescent microscopy

p53LMACO I cells were grown on glass coverslips coated with

Cell culture
collagen type 1. These cells were washed with phosphate buffer
saline(PBS), fixed with I% formal dehyde (Polyscienses) in
p53LMACOI cells were cultured in collagen type I coated
PBS for 5 min, washed twice with PBS, permeabilized in 0.1 %
dishes (IWAKI GLASS) with Oulbecco's modified eagle
Triton X- 100 in PBS for I min, and washed with PBS
medium (OMEM) (GIBCO) supplemented with 10% fetal
sufficiently. Then the samples were incubated with I: 100
bovine serum (Hyclone). Medium were containing 200 l!gl
dilution of FITC conjugated phalloidine (Molecular Probe)
ml L-glutamine, 100 U/ml penicillin (GIBCO), and 100 l!gl
for 30 min, rinsed with PBS three times and mounted in I,
ml streptomycin (GIBCO). Cell numbers were adjusted to 1
4-Oiazabicyclo (2. 2. 2)octane (OABCO). Observation was
x 10 5 cells per 10 cm diameter dish. The cells were incubated
carried out under a fluorescent microscope (Olympus,
in humidified 5% CO 2 - 95% air atmosphere and passaged
VANOX) equipped with epifluorescent optics.
when they reached 80-90% confluence, approximately I x
106 cells.
Measurement ofcytosolic calcium alteration

Cell growth Using a calcium sensitive dye fura-PE3 acetoxymethylester

(Molecular Probe) [4], intracellular calcium change was
For measurement of growth curve ofp53LMACO I, the cells measured in response to norepinephrine. Cells were harvested
were plated at 5 x 104 cells per 35 mm dishes. These cells at a density of I x 10 5 cells per 3.5 cm glass bottom culture
were trypsinized every 24 h and the numbers of these cells dishes (Meridian), washed with balanced salt solution (BSS,
were counted under a phase contrast microscope using a 130 mM NaCl, 5.4 mM KCI, 1.8 mM CaCI 2 , 0.8 mM MgCI 2,
hemocytometer. For measurement of doubling time, the cells 53 mM glucose, 10 mM HEPES pH 7.3) twice, loaded with
were seeded at 2.5 x 104 cells per 35 mm dishes and counted 5 l!M fura-PE3 acetoxymethylester for more than 20
every 12 h. Growth curve were fitted exponentially from 12- minutes, washed with BSS three times for 5 min. Fluores-
72 h and the doubling time was calculated. cence measurements were carried out with ARGUS50

6 6

B
........
-
Cl
OJ
-en
0 .2
en
.....
..... Q.)
Q.) .0
.0
E E
::J
::J 5 C 5
C
Q)
Q.)
u u

o 24 48 72 96 120 144 12 24 36 48 60 72
Time (hours) Time (hours)

Fig. 1. Cell growth curve of p53LMACO 1 cells. The cells were cultured in the DMEM supplemented with 10% fetal calf serum. On every 24 h (A) or 12 h
(B), the cells were tripsinized and counted by a hefflocytometer.
 101

(Hamamatsu Photonics) using a fluorescent microscope reached confluent state at 72 h. From 96 h the numbers of
(Nikon, DIAPHOT), according to Kudo's method [5]. the cells were decreasing (post-confluent state) as shown in
Fig. lA. At the logarithmic phase, the cell growing rate was
remarkably high. Doubling time was measured at every 12 h
Results interval (Fig. IB), and it was approximately 14 h.

Cell growth
General view ofthe cell
Seeded p53L-MACOI cells were grown through lag phase
(from 0-20 h) and logarithmic phase (from 20-72 h), and Morphological changes of the cells through the procedure of
cell cultures were observed under a phase contrast microscope.
In I day culture, many cells showed flattened shape like
fibroblast, but a few cells were somewhat small spindle shape.
When they were cultured for 4 days, almost all cells revealed
elongated and bipolar shape and a small number of the cell
remained flattened. The cells with elongated shape were
arranged in parallel arrays each other. After 4 days cell numbers
were gradually decreasing and cell attachment to the substrate
become worse. Cells were easily detached by the treatment for
fixation, washing and staining. At II days, cell density was re-
duced for detachment from the substrate but long bipolar shaped
cells were remained and observed. Although we did not show
these phase contrast images, those could be easily understood
from fluorescent views (Fig. 2) described in next session.

Development ofactin microfilaments 'formation

Developmental state ofactin filaments' formation in the cell

was investigated under a fluorescent microscope. FITC
conjugated phalloidine stained cells revealed the develop-
mental state ofactin filaments clearly. Many cells after I day
passage contained a lot of stress fibers (Fig. 2A). On 4 day
after passage, actin filaments were rich and run longitudinally
in the cytoplasma of the cells elongated and parallely lined
up. Structure of actin filament was not so clear because of
deconstruction of stress fiber, but diffuse fluorescent image
run alongside the longitudinal direction (Fig. 2B). Under the
surface membrane strong fluorescent image could be ob-
served, which means that many actin filaments gathered at
just inner side of the surface membrane. On 11 days, the
longitudinally run fluorescent image became poor though the
fluorescent image under the surface membrane remained
markedly (Fig. 2e). These results suggest that the cells
needed several days for development to the morphologically
differentiated state from a stand point of actin filaments'
formation, and after 4 days actin filaments in the cytoplasm
began to decompose except those of the surface membrane.

Fig. 2. Cell feature and actin microfilaments' formation in p53LMACOI

Development ofcalcium response to norepinephrine
cells by fluorescent microscopy. Cells were stained with FITC conjugated
phalloidine as described in Materials and methods. The cells cultured I In the previous work, we reported the p53LMACOI cells
day (A), 4 days (8), and 11 days (C). Magnifications, 2000X. responded to norepinephrine, revealing increase of intra-
102

cellular calcium ion and calcium oscillation dose-depen- phrine. When phenoxybenzamine was washed out, the
dently. Here we followed the change of calcium transient calcium response was gradually recovered. This result
response to norepinephrine, according to the development of suggests that the response to norepinephrine is due to an a-
the cells. Cells in I day culture did not show any calcium adrenergic receptor mediated passway.
transient response to a high dose (lO-4M) of norepinephrine
(Fig. 3A). When 4 days has passed after passage, 10-3 M
norepinephrine treatment resulted in an elevation of intra- Discussion
cellular calcium concentration (Fig. 3B). In addition to such
Ca elevation, Ca oscillation was also observed like adult When we reported first a novel vascular smooth muscle cell
smooth muscle cells at high concentrations ofnorepinephrine line p53LMACOI [I], there was remained an unexplained
IQ-4M (Fig. 3D) as described in the previous report [I]. Next, evidence; marker proteins for differentiation, h-caldesmon
we examined Ca transient response to norepinephrine for a and calponin disappeared in post-confluent state. This is a
long term-cultured cells on II days after passage. Almost all strange evidence, because differentiation is rather considered
the cells responded to norepinephrine from a low dose to high to be going on after post-confluent state in vascular smooth
dose as well as on 4 day cultured cells (Fig. 3C). These results muscle cell culture. In this work we have observed the cell
mentioned above showed the typical case ofone cell selected. according to cell growth for relatively long time and found
Then we measured calcium response from many cells and the several new evidences about characters of the cell. Thus, we
result of the typical case was confirmed generally from the could explained several disputed points concerning this cell
results of many cells as shown in Table I. line including the previous unsolved problem.
Figure 4 shows the effect of phenoxybenzamine, a selec-
tive a receptor inhibitor, for a change of intracellular calcium
in the cells by norepinephrine. The cells was able to respond Cell growth
again for the repeated stimulation of 10--6 M norepinephrine
after washing (data not shown). 10-7 M phenoxybenzamine p53 deficient mice was born normally and was viable em-
was able to inhibit calcium elevation induced by norepine- bryonically [6], but some type of cells, fibroblast [7, 8],

1.8 1.8
0
~ 1.6
....
ttl
A 0
:;:; 1.6 B
0
....
ttl

co
C")
1.4 0
co 1.4
u.. C")
.....
0
-.t
C")
u..
1.2
--
u..
0
-.t
C")
1.2

u..
0.8 0.8
0 100 200 300 0 100 200 300
Time (s) Time (s)

1.8 1.8
.2 0
:;:;
(tj 1.6
.... ttl.... 1.6
0 0
co
C")
1.4 co
C") 1.4
u..
.....
0
-.t
1.2
--
u..
0
-.t 1.2
E2
C")
u..
0.8
0 100 200 300 0 100 200 300
Time (s) Time (s)
Fig. 3. Change of cytosolic calcium induced by norepinephrine in p53LMACOI cells. Calcium response to 10-4 M norepinephrine(A and D), 10-5 M
norepinephrine (B and C). The cells were cultured 1 day (A), 4 days (B and D) and II days (C). Inserted bars show application time of norepinephrine.
 103

Morphological change ofp53LMACOl cells in culture

It is well known that many of cell lines will change their

characteristics when their growth are ceased. In smooth
muscle cell lines, A75 cells, which are derived from rat
thoracic aorta, have flat shape when they are proliferating.
This cells become thinner and spindle shaped and line up in
parallel arrays when they reach confluence and cease growth
[10, II]. p53LMACOI cells change their cell feature from
flattened shape to relatively long shape, and cease prolifera-
5,00 10,00 13,50 tion when they differentiate in culture medium containing 5-
Time (min)
azacytidine [I]. Such morphological change was also observed
in many cells when they cultured 4 days after passage without
Fig. 4, Effect of phenoxybenzamine for Ca response by norepinephrine, treatment of 5-azacytidine (Fig. 2B) and this change pro-
Filled box: norepinephrine IO-6M, Hatched box: I(T' M phenoxybenzamine,
4 day cultured cells were treated with I(T6M norepinephrine(box I), washed
gressed gradually and stopped growing at confluence.
with BSS, pre-treated with I(T' M phenoxybenzamine, treated with 10-6M Another characters of this cell after confluence were easily
norepinephrine (box 2) with existence of 10-' M phenoxybenzamine, rinsed detached from the substrate because somewhat cell death
with BSS and then treated with I Q-6 M norepinephrine (box 3), Norepine- might have occurred spontaneously.
phrine was applied for 30 sec, These morphological changes were accompanied with the
changes of actin filaments in the cell. Sometimes even in the
same culture dish, elongated and alined cells with rich actin
astrocyte [9], epithelial cells oflens [7], mammary glands [7],
filaments, and flat shaped cells with poor actin filaments and
seminal bone marrow cells [7] have advantages for growth
stress fiber coexisted (data not shown). This suggested that
in vitro cell culture. Fibloblast cultures derived from p53 cells in culture were developing gradually and the morpho-
lacking mouse showed growing characteristics as faster
logical changes indicated the development of the cells
dividing speed [I], higher confluent densities [7], higher
together with some differentiated characteristics.
plating efficiency [7, 8] and loss ofsenescent period [8]. These
In smooth muscle, differentiation involves the coordinate
cells grew rapidly, but remained contact inhibition and formed
expression of a number of smooth muscle isoforms of myosin
monolayers [8], indicating they were not transformed. So that
heavy chain [12, 13], actin (14-16] caldesmon (16-18],
such mice would be useful for establishing a immortal cell line.
calpoinin [14], SM22 [14], and tropomyosin (19, 20].
The cell line p53LMACOI derived from the vascular smooth
Expression of h-caldesmon and calponin were increased
muscle of p53 deficient mice also showed shorter doubling
gradually through culture days (data not shown) or 5-
time than other smooth muscle cell lines (Fig. I B) and
azacytidine treatment [I], but reduced when they entered into
stopped growing after confluence (Fig. I A). This result may
the post-confluent stage (1]. In this study, we showed that in
reflected a character of the lack of p53 gene. This higher
II days many cells became detached and that actin filaments
growing rate may be convenient for research of cell culture. in the cytoplasm were reduced in remained cells (Fig. 2C).
However, the life span ofthis cell looked shorter as described This result suggested the deconstruction of cytoskeleton like
next session. actin, which was accompanied with the decrement of other
cytoskeletal proteins such as h-caldesmon.
In addition to the protein markers, smooth muscle cells
Table I Calcium responses induced by norepinephrine in p53LMACOI
cells on different cultured days expressed a number of receptors and ion channels required
for the contractile and metabolic functions [3]. Intracellular
Days NA No N % calcium response of p53LMACO I cells to norepinephrine
1 100 tiM 112 1 0,9 were developed through culture days from 1--4 days (Fig. 2
4 100 tiM 108 98 90,7 and Table I). That was consistent with the morphological
10 tiM 101 86 85, I change and the development of actin filaments' formation in
II 100 tiM 95 87 91.4
94,7
p53LMACOI cells (Fig. 2). These data suggest that the cells
10 tiM 94 89
developed through culture days. Although the expression of
NA - Concentration of norepinephrine; No - Numbers of total measured h-caldesmon and actin filaments were reduced at II days, the
cells, N - Numbers of calcium-responded cells, % - NINo x 100,
response to norepinephrine remained as well as that at 4 days
p53LMACOI cells were cultured for 1,4, and II days after passage and
measured intracellular carcium alteration as described in Materials and (Fig. 2 and Table I). At II days, fluorescence of FITC
methods. Responded cells were counted after 30 sec of norepinephrine phalloidine under the surface membrane were still strong
treatment. (Fig. 2C). These results suggest that the life span of actin
104

filaments in cytoplasm and the surface membrane were Biophys J 69(5): 2112-2124, 1995
independent. On the basis of this evidence, the function of 5. Kudo Y, Takita M, Nakamura K, Sugaya K, Ogura A: Heterogenous
distribution of functional glutamate receptor subtypes in organotypic
a-receptor might survive in spite of disappearance of h- slice culture of rat hippocampus revealed by calcium fluorometry.
caldeamon and calponin [1]. Bioimage I: 159-166,1993
Finally the specific characteristic ofa novel vascular smooth 6. Donehower LA, Harvey M, Slagle BL, McArthur MJ, Montgomery
muscle cell line from p53 knock out mice, p53LMAC01, is a Jr. CA, Butel JS, BradleyA: Mice deficient for p53 are developmentally
cell of which the life span runs to end rapidly. During the time normal but susceptible to spontaneous tumors. Nature 356: 215-221,
1992
it shows several features and characters of differentiated 7. Tsukada T, Tomooka Y, Takai S, Ueda Y, Nishikawa S, Yagi T,
smooth muscle; elongated bipolar shape, abundant actin fila- Tokunaga T, Takeda N, Suda Y, Abe S, Matsuo I, Ikawa Y, Aizawa S:
ments' fonnation longitudinally, appearance ofh-caldesmon, Enhanced proliferative potential in culture of cells from p53-deficient
calponin and smooth muscle a-actin, and expression of a- mice. Oncogene 8: 3313-3322, 1993
adrenergic receptor. This cell line should be appreciated for 8. Harvey M, Sands AT, Weiss RS, Hegi ME, Wiseman RW, Pantazis P,
Giovanella BC, Tainsky MA, Bradley A, Donehower LA: In vitro
such excellent differentiating ability which has not been shown growth characteristics of embryo fibroblasts isolated from p53-
by other vascular smooth muscle cell line yet. As for vascular deficient mice. Oncogene 8: 2457-2467, 1993
smooth muscle cell in culture, there is misunderstanding widely 9. Bogler 0, Huang H-J Su, Cavenee WK: Loss of wild-type p53 bestows
accepted, in which cell differentiation begins after arriving at a growth advantage on primary cortical astrocytes and facilities their
confluence, piling up, fonning hill and valley without contact in vitro transformation. Cancer Res 55: 2746-2751, 1995
10. Charnley-Campbell J, Campbell GR, Ross R: The smooth muscle cell
inhibition. On the contrary to such images from a common in culture. Physiolog Rev 59: 1--61, 1979
sense, our cell showed some differentiating characters on the II. Kimes BW, Brandt BL: Characterization of two putative smooth
growing way, and died after the confluence. We are now muscle cell lines from rat thoracic aorta. Exp Cell Res 98: 349-366,
preparing another paper in which these characters like hill and 1976
valley mentioned above do not come from differentiating cells, 12. Zanellato AM, Borrione AC, Giuriato L, Tonello M, Scannapieco G,
Pauletto P, Sartore S: Myosin isoforms and cell heterogeneity in
but come from undifferentiating cells and its stem cells [21]. vascular smooth muscle, I: developing and adult bovine aorta. Dev
Of course our cell line is not perfect, but does not express other Bioi 141:431-446, 1990
receptors existing in adult smooth muscle except for a- 13. Kuro-o M, Nagai R, Nakahara K, Katoh H, Tsai RC, Tsuchimochi H,
adrenergic receptor (not shown). Now we are subcloning Yazaki Y, Ohkubo A, Takaku F: eDNA cloning of a myosin heavy
p53LMACO 1 to get more stable and differentiating characters chain isoform in embryonic smooth muscle and its expression during
vascular development and in arteriosclerosis. J Bioi Chern 266: 3768-
utilizing 5-azacytidine treatment and other agents. 3773,199\
14. McHugh KM, Ussard JL: The development expression of the rat alpha
vascular and gamma-enteric smooth muscle isoactins: Isolation and
Acknowledgements characterization ofa rat gamma-enteric actin eDNA. Mol Cell Bioi 8:
5224-5231, 1988
15. Owens GK, Thompson MM: Developmental changes in isoactin
We, especially Y. N., would like to tell heartful thanks to the expression in rat aortic smooth muscle cells in vivo: Relationship
Emeritus Prof. S. Ebashi for his kind guidance and initiation between growth and cytodifferentiation. J Bioi Chern 261: 13373-
about muscle works for a long time. We also thank Prof. Y. 13380,1986
16. Duband J L, Gimona M, Scatena M, Sartore S, Small JV: Calponin
Kudo, Tokyo Phannaceutical University for his kind intro- and SM22 as differentiation markers of smooth muscle: Spatiotemporal
duction of intracellular calcium measurement. This work was distribution during avian embryonic development. Differentiation 55:
supported by a grant-in-aid for scientific reserch from the 1-11,1993
Ministry of Education, Culture, Science and Sports of Japan. 17. Frid MG, Shekhonin BV, Koteliansky VE, Glukhova MA: Phenotypic
changes in human smooth muscle cells during development: Late expres-
sion of heavy caldesmon and calponin. Dev Bioi 153: 185-193, 1982
18. Glukhova MA, Frid MG, Koteliansky VE: Developmental changes in
References expression of contractile and cytoskeletal proteins in human aortic
smooth muscle. J Bioi Chern 265: 13042-13046, 1990
I. Ohmi K, Masuda T, Yamaguchi H, Sakurai T, Kudo Y, Katsuki M, 19. Fatigati V, Murphy RA: Actin and tropomyosin variants in smooth
Nonomura Y: A novel aortic smooth muscle cell line obtained from muscles: dependence on tissue type. J Bioi Chern 259: 14383-14388,
p53 knock out mice expresses several differentiated characters. 1984
Biochem Biophys Res Comm 238: 154-158, 1997 20. Wieczorek DE, Smith CW, Nadal-Ginard B: The rat alpha tropomyosin
2. Ross R: The pathogenesis of atherosclerosis: A perspective for the gene generates a minimum of six different mRNAs coding for striated,
1990s. Nature 362: 801-809,1993 smooth, and nonmuscle isoforms by alternative splicing. Mol Cell Bioi
3. Owens GK: Regulation of differentiation of vascular smooth muscle 8: 679--694,1988
cells. Physiolog Rev 75: 487-517, 1995 21. Ohmi K, Sakurai T, Nonomura Y: Discovery of a new stem cell
4. Vorndran C, Minta A, Poenie M: New fluorescent calcium indicators producing undifferentiated vascular smooth muscle. (preparing the
disigned for cytosolic retention or measuring calcium near membranes. manuscript) 1997
Molecular and Cellular Biochemistry 190: 105-118, 1999.
© 1999 Kluwer Academic Publishers.

Expressional regulation of smooth muscle

cell-specific genes in association with phenotypic
modulation

Kenji Sobue, Ken'ichiro Hayashi and Wataru Nishida

Department ofNeurochemistry and Neuropharmacology, Biomedical Research Center, Osaka University Medical School,
Osaka, Japan

Abstract
Phenotypic modulation of smooth muscle cells (SMCs) plays an integral role in atherosclerosis, hypertension and leiomyogenic
tumorigenicity. The morphological, functional, and biochemical characteristics of SMCs in different phenotypes such as
differentiated and dedifferentiated states have been well studied. Recent researches have focused on the expressional regulation
ofSMC-specific marker genes in association with phenotypic modulation ofSMCs. The SMC-specific marker genes are regulated
at the levels of transcription and splicing. The caldesmon, smooth muscle myosin heavy chain, a-smooth muscle actin, calponin,
SM22, a- and ~-tropomyosins, and al integrin genes are transcriptionally regulated; transcription of these genes except for the
a-smooth muscle actin gene is upregulated in differentiated SMCs, but is downregulated in dedifferentiated SMCs. The expression
pattern of a-smooth muscle actin is opposite in vascular and visceral SMCs. In almost all promoter regions of these genes, the
CArG box and serum response factor (SRF) are involved in as the positive cis-element and the trans-acting factor, respectively.
Isoform changes of caldesmon, a-tropomyosin, vinculinlmetavinculin, and smooth muscle myosin heavy chain are regulated by
alternative splicing in a SMC phenotype-dependent manner. Among them, isoform interconversions ofcaldesmon and a-tropomyosin
are completely coordinated with phenotype ofSMCs. The purpose ofthis paper is to summarize current knowledge ofthe expressional
regulation of SMC-specific marker genes in different phenotypes of SMCs. (Mol Cell Biochem 190: 105-118, 1999)

Key words: smooth muscle cell, smooth muscle cell-specific markers, phenotypic modulation, transcription, splicing

Introduction In addition to SMC lineages, phenotypic modulation ofSMCs

is known to play an integral role in atherosclerosis, hyper-
Little is known regarding the origins of the precursor cells tension and leiomyogenic tumorigenicity. Recent progress
that give rise to smooth muscle cells (SMCs). At least three has been made in identifying molecular markers to express
different lineages (cardiac neural crest, nodosa placode, and phenotype of SMCs. In this paper, we focused on the expres-
lateral mesoderm) are candidates to produce SMCs. In the sional regulation of SMC-specific marker genes in associa-
case of vascuiogenesis, vascular precursor cells, angioblasts, tion with phenotypic modulation.
initially differentiate from mesoderm to form endothelial
cells, which then coalesce to form a single layer of endothelial
tubes embedded in extracellular matrix. During development, Ca2+ -dependent regulation of smooth
these initial vessels are involved in the recruitment of SMC muscle contraction
precursor cells into the surrounding of endothelial tubes, with
subsequent morphogenesis of the blood vessel (reviewed in The SMCs are involved in controlling the pressure of blood
ref. [1,2]). Similarly, each SMC type responds very differently vessels, peri stasis of enteric organs and contraction of
to local morphogenic cues, reaching to SMC containing tissues. bronchus, uterus, bladder and vas deference. Electron

Addressfor offPrints: K. Sobue, Department of Neurochemistry and Neuropharmacology, Biomedical Research Center, Osaka University Medical School,
2-2 Yamadaoka, Suita, Osaka 565, Japan
106

microscopy revealed the two prominent electron dense Table I. Expressional and isoformal changes ofSIVIC-specific molecular
structures of SMCs localized in the cytoplasm and at the markers
cell-cell contact; the dense body and the dense membrane differentiated SMC dedifferentiated SMC
(dense plaque), respectively. The dense body plays a role as
the termination site of myofibrils (contractile apparatus) caldesmon fr h-CaD I-CaD ~
a-tropomyosin fr a-TM-SM a-TM-Fl and a-TM-F2 ~
corresponding to the Z line of striated muscle. The dense
j3-tropornyosin fr ~
membrane is the heterophilic cell adhesion structure ofSMCs myosin heavy chain ~ SMemb SMemb fr
mediated by extracellular matrix components, which are fr SMI SMI ~
mainly composed of collagen type IV, laminin, and elastin fr SM2 SM2 ~
[3, 4]. The ~ 1 integrin and a member of membraneskeletal
a-smooth muscle actin
proteins such as vinculin, a-actinin, and talin are localized
vascular SMC fr ~
in the dense membrane [5, 6]. visceral SMC ~ fr
The principal function of SMCs is contraction. The
actomyosin system consisting of contractile proteins, which metavinculin fr ~

converts the chemical energy ofATP into mechanical force,

SM22 fr ~
is the molecular basis for contraction of SMCs and various
contractile and motile events in nonmuscle cells. In smooth calponin fr ~
muscle and nonmuscle cells, actomyosin interaction is con-
trolled by the myosin-linked and actin-linked regulations. The al integrin fr ~

myosin-linked regulation is based on phosphorylation of the The direction of arrowheads indicates the up- or downregulation of
20 kDa regulatory light chain of myosin by Ca 2+/calmodulin- SMC-specific marker genes in different phenotypes of SMCs.
dependent myosin light chain kinase, and dephosphorylation
by myosin phosphatase [7]. The actin-linked regulation is the
functional regulation of actin filament by calmodulin, that the dense body and the dense membrane are clearly
caldesmon, and tropomyosin in a Ca 2+-dependent manner [8]. detected in 10-12 day old chick embryo gizzard, and the
An alternative function ofI-caldesmon in nonmuscle cells has number and size of both structures increase during develop-
been proposed, whereby I-caldesmon and tropomyosin are ment [12]. On the contrary, these characteristic structures
coordinately involved in the stabilization of actin filament disappear dramatically during dedifferentiation. The dense
against the severing activity ofgelsolin [9]. Calponin has been body and the dense membrane are therefore morphological
also proposed as a regulatory protein in the actin-linked characteristics ofSMC phenotype. In dedifferentiated SMCs,
regulation (reviewed in ref. [10]). However, the functional rough endoplasmic reticulum and large Golgi complex are well
significance of this protein is not fully characterized. developed [13]. These SMCs also assume the proliferative
activity in response to mitogens and secrete extracellular matrix
components, while lose contractile ability.
Phenotypic modulation of SMCs Biochemically, expression and isoform changes in con-
tractile proteins, membraneskeletal proteins, and cell adhesion
Although critical pathogens of atherosclerosis have been molecules and their receptors (integrins), are found in
unidentified, one ofthe major events is the onset ofphenotypic association with phenotypic modulation of SMCs. Since their
modulation of vascular SMCs (reviewed in ref. [11]). During expressions are regulated at the gene levels, such as at
this process, vascular SMCs undergo a transition in phenotype transcription and splicing, these proteins, listed in Table 1,
from a contractile (differentiated) to synthetic (dedifferen- are suitable molecular markers for phenotype of SMCs. The
tiated) state. Vascular SMCs thus gain two alternative abilities: caldesmon, smooth muscle myosin heavy chain, a-smooth
cell proliferation and migration. Proliferated cells then migrate muscle actin, calponin, SM22, a- and ~-tropomyosins, and
into the intima, causing intimal thickening. Finally, progression al integrin genes are transcriptionally regulated; transcription
of atheroselerotic lesions in the intima is characterized by the of these genes except for the a-smooth muscle actin gene is
accumulation of alternating layers of dedifferentiated SMCs upregulated in differentiated SMCs, but is downregulated in
and lipid-laden macrophages. Therefore, phenotypic modula- dedifferentiated SMCs. The expression of a-smooth muscle
tion of SMCs is one topic in current vascular biology. actin is opposite in vascular and visceral SMCs. Isoform
The morphological and biochemical characteristics of changes of caldesmon, a-tropomyosin, vinculinlmetavinculin,
SMCs in different phenotypes are summarized in Table 1. The and smooth muscle myosin heavy chain are regulated by
dense body, the dense membrane, and myofibrils composed alternative splicing in a SMC phenotype-dependent manner.
of thin filaments and myosin thick filaments, are well The details of these findings will be discussed in part V.
developed in differentiated SMCs. Chou et al. have reported (Molecular markers for phenotype of SMCs).
 107

SMC culture system activity. These results indicate that IGF-IIIGF-I receptor and
laminin are essential for a differentiated phenotype ofSMCs.
Primary cultured SMCs prepared by either enzyme dispersion Furthermore, IGF-I1IGF-I receptor- and/or laminin-generated
or explant rapidly induce dedifferentiation under normal signallings are completely suppressed by the addition of
culture conditions [13]. This may be due to the disruption of tyrosine kinase inhibitors (Fig. 1), suggesting the involve-
SMC adhesion structures including extracellular matrices ment of tyrosine phosphorylation pathway in maintaining a
during culture procedures, the deletion offactors maintaining differentiated phenotype of SMCs.
a differentiated phenotype of SMCs, andlor the addition of
factors inducing SMC dedifferentiation in culture medium.
A culture system, which maintains a differentiated phenotype Molecular markers for phenotype of
of SMCs and controls the phenotypic conversion from a SMCs
dedifferentiated to differentiated state, has not been reported.
The studies using primary cultured SMCs have therefore only Caldesmon
characterized the dedifferentiation process. Moreover, there
are no available SMC lines maintaining a fully differentiated The two caldesmon (CaD) isoforms have been identified:
phenotype [14]. Under these circumstances, it is necessary h-CaD (high Mr-form) is dominantly expressed in differen-
to establish a culture system controlling a phenotype of tiated SMCs, while I-CaD (low Mr-form) widely distributes
SMCs. As described above, collagen type-IV and laminin are in nonmuscle tissues and cells [22, 23]. In particular, the
the main components of the extracellular matrix in the SMC isoform interconversion of CaD is strongly associated with
layer [3,4]. s-Laminin, which is an isoform of laminin ~2 phenotypic modulation ofSMCs, in which the CaD isoform
chain, is uniquely expressed in matured vascular SMC layer, converts from the 1- to h-form during differentiation and vice
but not in visceral SMCs [15]. High levels of a 1 and 0.3 versa [23]. The expression changes of CaD isoforms also
integrins, receptors for collagen and laminin, are detected in link to phenotype of SMCs; the expression of h-CaD in
matured SMC containing tissues [16, 17]. In particular, the differentiated SMCs is higher than that ofI-CaD in dedifferen-
expression ofal integrin depends on phenotype ofSMCs; its tiated SMCs [23]. CaD is therefore considered to be a
expression in SMCs is upregulated in developing embryos, favorable molecular marker for studying such phenotypic
while downregulated in dedifferentiated SMCs under normal modulation of SMCs. The CaD gene is regulated at the levels
culture conditions [17,18]. Simultaneously, the expression of transcription and splicing. The chicken and human CaD
of aV and as integrins increases in dedifferentiated SMCs genes are composed of 14 and 16 exons, respectively [24, 25].
[19]. These findings suggest that the extracellular matrices Exon 3 of both genes encodes the unusual and unique
playa vital role in the regulation ofSMC phenotype. Thyburg domain. The common domain of h- and I-CaDs is encoded
et al. have reported that laminin and collagen type-IV delay in exon 3a, whereas the h-CaD-specific central repeating
the progression of dedifferentiation in primary cultured domain is completely encoded in exon 3b. Therefore, matura-
vascular SMCs, whereas fibronectin stimulates dedifferentia- tion of mRNAs for CaD isoforms is determined by a unique
tion [20]. Consequently, Hayashi et al. have investigated the splicing: the expression ofh- or I-CaDs depends on a selection
effects of extracellular matrix components on phenotype of of two 5'-splice sites within exon 3 (Fig. 2). The SMC
SMCs under serum-free culture conditions [21]. Among phenotype-dependent splicing of CaD has scarcely been
several extracellular matrix components, laminin has the investigated. Recently, it has been reported that repeating
potency to maintain a differentiated phenotype of SMCs. purine-rich motifs within exon 3b act as an exon enhancer
Collagens type-I and type IV and fibronectin do not possess element, causing predominant selection of distal 5'-splice site
such activity. Laminin is, however, not sufficient to maintain in myofibroblasts and nortmuscle cells [26]. On the other hand,
a differentiated phenotype of SMCs for a long culture, a different result has been obtained whereby the intron
suggesting the requirement of additional factors. A variety sequence between exons 3b and 4 might be involved in
of growth factors and cytokines, such as platelet-derived alternative selection ofdistal and proximal 5'-splice sites within
growth factor-AA and -BB (PDGF-AA and -BB), basic the exon 3. The selection ofproximaI5'-splice site is inhibited
fibroblast growth basic fibroblast growth factor (bFGF), in the transcripts from the CaD minigene constructs carrying
angiotensin II, Arg-vasopressin (AVP), epidermal growth deleted or mutated intron sequence in differentiated SMCs.
factor (EGF), and tumor growth factor ~ 1 and ~2 (TGF ~ 1 The expressional amount of CaD is determined by the
and ~2), induce dedifferentiation of SMCs even when regulation of promoter activities. The CaD gene contains two
cultured on laminin. Surprisingly, insulin-like growth factor CaD promoters, gizzard- and brain-types (Fig. 2), in which
I (IGF-I), II (IGF-II), and insulin can maintain a differentiated the gizzard-type promoter shows extremely high activity
phenotype of SMCS for a long culture only in the presence compared with the brain-type promoter [27]. The tran-
of laminin. Among them, IGF-I shows the most potent scriptional activity of the gizzard-type promoter in SMCs is
108

much higher than that in other cell types such as C2C 12 and
HeLa cells, suggesting cell type-specificity. The promoter
G>-
Q,f activity in differentiated SMCs is also higher than that in
~Ql80 dedifferentiated SMCs. These results are consistent with the
O~
c::ClI
high expression of h-CaD in differentiated SMCs compared
-ali 60
QlE
'C-
with the low expression of I-CaD in dedifferentiated cells.
Ql~ The gizzard-type promoter contains multiple E boxes, CArG
m..!40
0:0
box, CCAAT box, TATA box and Sp I. binding site. Promoter
iE analysis using a series ofdeletion mutant ofthe CaD promoter
~020 has revealed that the cell type-specific transcription of the
:Z:::E
=s~ CaD gene depends on a single cis-element, CArG box
O'-.........._ _ ~_~r--_~ _ _~_ _~ ~_ _
(CCAAAAAAGG), located between -309 to -300, and
o 3 5 7 9 11
multiple E boxes are not directly involved in this event [28].
culture days
Using gel-shift assay, serum response factor (SRF) has been
identified as a core trans-acting factor for the CArG box in
Fig. 1. Laminin and IGFI playa vital role in maintaining a differentiated the CaD promoter. Moreover, it has been demonstrated that
phenotype of cultured SMC. Progressive changes in relative ratios (%) of the CArG binding activity ofSRF in differentiated SMCs is
h-CaD mRNA to total CaD mRNAs in primary cultured SMCs are shown. much higher than that in dedifferentiated SMCs, whereas
The SMCs are cultured on laminin-coated (LN(+)) or non-coated plate
there is less significant difference of the SRF expression in
(LN(-)) under serum-free (lGF (-)) or IGFI-stimulated (lGFI (+)) conditions.
Open circles, IGFI (+) / LN (+); closed circles, IGFI (+) / LN (-); open both phenotypes of SMCs [29]. This result suggests the
triangles, IGFI (-) / LN (+); and closed triangles, IGFI (+) or (-) / LN (+) or presence of another factor which confers the SMC-specific
(-) I tyrosine kinase inhibitors (+). phenotype-dependency on SRF-CArG interaction.

Exon 1a-1 1a-2 18-3 1b 2 3 4 5 6 7 8 9 10111213 14

ATG ATG a b TAG

gizzard-type promoter

ATG TAG
h-CaD

ATG TAG
I-CaD

Fig. 2. Genomic structure of the CaD gene and alternative splicing pathways. Genomic structure of the chicken CaD gene is schematically shown. Boxes
indicate exons in the CaD gene, and the short NH,-terminal sequences of the gizzard- and brain-type CaDs are encoded in exons I a-3 and Ib, respectively.
Two different promoters, the gizzard- and brain-type promoters, are shown by arrowheads and the initiation and termination codons for translation are
indicated by ATG and TAG, respectively. In general, the gizzard-type promoter is extremely dominant. Two 5'-splice sites within exon 3 are indicated by
arrows. Alternative selection of these 5'-splice sites determines the expression ofh- or I-CaDs, and such selection is regulated in an SMC phenotype-dependent
manner. Exon 3 encodes a unique and unusual domain; the common domain of h- and I-CaDs is encoded in exon 3a and the h-CaD-specific central repeating
domain is in exon 3b. Exon 4 is spliced in the chicken h-CaD mRNA [24], whereas human I-CaD isoform carrying the sequence encoded in exon 4 has been
detected [25]. Thus, alternative selection of exon 4 is not an SMC-specific event.
 109

Tropomyosin ( differentiated SMC) ( dedifferentiated SMC)

ill-:
exon 3
Tropomyosin (TM) is a predominant helical protein which
binds to actin grooves. In striated muscle, TM is known to CaD gene
mediate the Ca 2+ response of troponin complex to actin 3 / , 3

---GEt- -EJ--
filaments (reviewed in ref. [30]). The TM and CaD in smooth
muscle and nonmuscle cells are coordinately involved in the
Ca 2+-dependent regulation of actin-myosin interaction as exon 2a 2b
actin-linked regulatory proteins [8, 31].
One- and two-dimensional gel electrophoreses revealed a-TM gene [ } - { ]
multiple isoforms of TM associated with morphological 2a..l'" ............ 2b
changes and tumorigenic transformation [32-34]. Genomic
analyses have also shown a diversity ofTM mRNAs genera-
ting from only limited genes by a combination of differential
-{}- =-0-
....................................................................
exon 18 19 20
promoter usage and alternative splicing (reviewed in ref. [35]).
Regarding the a-TM gene, the exon selection among a couple vinculin gene~
of mutually exclusive exon pairs (exon 2a and 2b, and exons
6a and 6b) is regulated in a tissue-specific manner. A pair of
internal exons (exons 6a and 6b) is spliced in a mutually
exclusive manner and their utilization is regulated during the 18 19
/ 20
'\ 18 20
differentiation process of skeletal muscle cells; exons 6a and
6b are used in myoblasts and myotubes, respectively [36, 37]. ~ --[}{}-
Exon 2b rather than exon 2a (termed exons 3 and 2, (metavinculin) (vinculin)
respectively, in ref. 38) is spliced in the a-TM mRNAs in all
cell types except for SMCs [38, 39]. The selection of exon
Fig. 3. SMC phenotype-dependent alternative splicing in the CaD, u-TM,
2a is therefore an SMC-specific event. and vinculin genes. Alternative splicings within exon 3 in the CaD gene
Mutually exclusive splicing between exons 2a and 2b in and between exons 2a and 2b in the u-TM gene are coordinately regulated
the a-TM gene has been intensively studied. The possible in an SMC phenotype-dependent manner. In differentiated SMCs, selections
explanation ofthis event is that predominant selection ofexon of exon 3ab in the CaD gene and exon 2a in the u-TM gene result in
2b may be based on a competition between exons 2a and 2b; expression of h-CaD and IX- TM-SM. In contrast, expressions of I-CaD and
u-TM-F I and u-TM-F2 in dedifferentiated SMCs are arisen by selections
the branch point/pyrimidine tract elements in the downstream of exon 3a in the CaD gene and exon 2b in the IX- TM gene, respectively. In
intron of exon 2b is stronger than those of exon 2a [40]. The the vinculin gene, exon 19 encoding metavinculin-specific sequence is
functional strength of exon selection depends on an affinity specifically selected in differentiated SMCs. However, this splicing is not
of splicing factor, U2AF, with the pyrimidine tract for exon specific to SMCs because metavinculin is also expressed in skeletal and
2b, or that for exon 2a [41]. Alternatively, the SMC-specific cardiac muscles.
selection of exon 2a may be due to inhibition of the exon 2b
selection; two conserved elements in each flanking of exon
2b are essential for such SMC-specific inhibition [42]. The unclear whether or not common factor(s) are involved in such
third possibility is that the splicing machineries themselves exon selection in the a-TM and CaD genes.
may be regulated in a SMC phenotype-dependent fashion. Chicken 13-TM is downregulated in dedifferentiated SMCs,
The SMC phenotype-dependent isoform interconversion of but is upregulated in differentiated SMCs [43]. While the
TM has been studied [43]. During dedifferentiation ofSMCs, transcriptional machineries of the TM genes have not been
the SMC-type a-TM (a-TM-SM) converts to fibroblast-type well characterized, the upstream promoter of the 13- TM gene
1 and 2a-TM isoformsa-TM-Fl anda-TM-F2) by a change has been partially analyzed using skeletal muscle cell line; a
in the exon selection from exon 2a to exon 2b, while the CArG box is identified as one of the essential cis-elements
SMC-type 13- TM (13- TM-SM) is downregulated. Under necessary for skeletal muscle-specific transcription [44].
culture conditions in which differentiated SMCs are caused
to dedifferentiate by serum or PBGF-BB, isoform change of
CaD from a h- to i-form is consistent with that ofTM. In situ a-smooth muscle actin
hybridization revealed coexpression ofa-TM-SM and h-CaD
mRNAs in developing aorta, gizzard and intestine. Taken The actin isoforms are divided into two classes, muscle and
together, the a-TM and CaD genes are coordinately regulated nonmuscle actins, and are designated according to their
in a SMC phenotype-dependent manner (Fig.3). It remains isoelectric points as a, 13, or y. Of these, a-smooth muscle
110

(a-SM) actin comprises a major portion of actin isoforms the negative element in spite of the presence of positive
expressed in vascular SMCs [45,46]. Expression of actin elements. By contrast, the promoter activities in dedifferen-
isoforms in rat and human aortic SMCs has been shown to tiated visceral and differentiated vascular SMCs mainly
be developmentally regulated; l3-nonmuscle actin is sub- depend on the positive elements [55]. Taken together,
stituted bya-SM actin [47, 48]. Transcription ofa-SM actin expression of a-SM actin is paradoxically regulated by a
mRNA increases in developing aorta, to about 90% of total combination of multiple cis-element-trans-acting factor
actin mRNAs in adult tissue [14]. By contrast, aortic SMCs interactions.
in primary culture exhibit a slight decrease in the a-SM actin It has been reported that the region between the E box and
expression in association with phenotypic modulation [49). the purine-rich motifacts as a negative element in mouse em-
Unlike vascular SMCs, y-actin is the dominant isoform in bryonic fibroblasts, and is a target sequence for VACssBF1
visceral SMCs [50). Immunostaining ofa-SM actin in adult and VACssBF2 [56, 57]. The VACssBF1 and VACssBF2
chicken gizzards showed negative in parenchymal cells specifically bind to single-stranded(ss) DNA spanning the
except for blood vessels [51], indicating that a-SM actin is purine-rich motif, and these protein factors might function
not a specific molecular marker for all SMCs. Ectopic as repressors of mouse actin promoter [56-58], but c-myc
expression of a-SM actin has been observed in nonmuscle gene single-stranded binding protein-I, MSSP-1, negatively
stromal cells [52], such as massively proliferated mesangial regulates this promoter [55]. The activation of the promoter
cells [53] and myofibroblasts [54] under certain pathological in dedifferentiated visceral SMCs is strongly dependent on
conditions. the three positive elements, such as the purine-rich motif
Curiously, a-SM actin is expressed in undifferentiated plus the CArGs A and B [55]. It has been reported that
visceral SMCs of early chick embryo, but this expression TEF-1 recognizes the purine-rich motif to regulate serum-
decreases thereafter [51]. Consistent with this, a-SM actin stimulated transcription in AKR-2B fibroblasts [56]. In
is upregulated during dedifferentiation of visceral SMCs. The visceral SMCs, TEF-1 would be also involved in a specific
expressional pattern of a-SM actin in visceral SMCs is DNA-protein complex in the purine-rich motif[55]. Shimizu
therefore opposite to that in vascular SMCs. Such paradoxical et al. have reported that SRF or SRF-like factor(s) from
expression of a-SM actin is regulated by positive and nuclear extracts of dedifferentiated aortic SMCs bind to the
negative elements of the a-SM actin promoter in both CArGs A and B ofrata-SM actin promoter [59). In agreement
phenotypes of vascular and visceral SMCs (Fig. 4). In with these findings, gel-shift analysis revealed that the CArGs
differentiated visceral and dedifferentiated vascular SMCs, A and B ofchicken a-SM actin are also core elements for SRF
the overall promoter activities are dominantly suppressed by binding [55].

a-SM actin
dedifferentiated expression
vascular SMCs
differentiated
visceral S.M_C_S_ _ P_f'
NE ~_I_uP_E----,HcArG BH CArG
f
A~
dedifferentiated
visceral SMCs

differentiated
;
..........
~."""
E box

:
~.i
-193
~
~
i,
""'~
-176
.j
: ~"""
~-70
ti,
",,_
-61
TATA

vascular SMCs : GAGC'l'CTGCGTTGGAATO : - CC'l'TG'l'TTGG

-232 -219
TOTTTATCTTACAC
P u-rlc
0 h motOf :
I ; _
-120 -111
CCCTATATGG

Fig. 4. Paradoxical transcriptional regulation of the chicken o:-SM actin promoter in different phenotypes of vascular and visceral SMCs. Negative and
positive cis-elements involved in the phenotype-dependent transcription in SMCs are indicated by boxes in the alignment map of the promoter region of
o:-SM actin gene. NE, Pu and CArGs A and B indicate the negative element, the purine-rich motif, and two CArG boxes, respectively. The starting site of
transcription is shown by a bended arrow. The elements which strongly or weakly function in different phenotypes of vascular and visceral SMCs are marked
by closed and open bars, respectively.
 III

Myosin heavy chains 10 days postnatally and gradually increases to adult levels,
whereas the SMI as well as NMHCs are expressed in fetal
Myosin, a hexametric motor-enzymic protein, plays a arteries [73].
central role in the development of contractile force in SMCs To elucidate the trans-acting factors activating the SMHC
[60,61]. Two-headed myosin (myosin II) is composed of gene, several investigators have studied the SMHC promoter.
two identical 200 kDa heavy chains noncovalently asso- Babij et a/. found a positive element in 107 bp DNA
ciated with two pairs of light chains, the 20 kDa regulatory fragment between-I,332 and-I,225 of the rabbit SMHC
(LC20, phosphorylatable) and 17 kDa alkali (LC 17, non- gene. This fragment containing a single CArG box shows
phosphorylatable) subunits. The N-termini of the heavy the SMC-specific transcriptional activity [74]. White et al.
chain dimers form two globular heads consisting of the also reported that a region between -1,317 and -1,249 of
ATPase and actin-binding sites. The LC20 and LC 17 bind to the rat SMHC gene is important for the transcription in
the neck region of the heads. The remaining heavy chain SMCs, and that the sequences around the CArG box are
forms an a-helical coiled-coil tail, which is involved in completely conserved in both the rabbit and rat SMHC
filament formation. promoters [75]. In contrast to these findings, Nagai et a/.
Multiple isoforms of myosin heavy chains have been confined the positive element with tandem repeats of
reported in striated and smooth muscles and nortmuscle cells CCTCCC sequence located between -89 and -61 of the
[62, 63]. In SMCs, two variants with molecular weights of mouse SMHC gene [76].
204 kDa (MC204, SM I) and 200 kDa (MHC200, SM2) have
been reported. They are encoded in a single gene (SMHC
gene) and produced by alternative splicing [64]. The SMI and Myosin light chains
SM2 diverge in their C-terminus. The SM2 selects a unique
exon encoding 9 amino acid residues at the C-terminus. The There are two smooth muscle variants ofLC17: LCI7a and
SM I excludes the unique exon of SM2, resulting in a longer LC 17b [77-79]. They are encoded in a single gene and
C-terminus containing 43 residues. The unique C-terminal transcribed via alternative splicing [80, 81]. Both the LCI7a
region ofSMI can be phosphorylated in vivo by casein kinase and LC 17b are present in vascular SMCs, whereas the
II [65]. Although the functional difference ofSMI and SM2 LC 17a is exclusively expressed in gastrointestinal tract as
is unclear, there is some evidence that their expressions are well as nonmuscle cells [78]. The functional significance
tissuespecific and are developmentally regulated [66, 67]. of the two isoforms is yet unknown, but myosin containing
Two additional isoforms of smooth muscle myosin heavy LC 17a possesses high ATPase activity compared with that
chain, SMIA and SMIB, that differ in the N-terminal containing LC 17b [77]. The LC20 has multiple isoforms in
globular head region ofSMI, have been identified [68, 69]. arterial SMCs such as L20A (smooth muscle variant) and
The SM IB includes an additional 7 amino acids in a region L20B (nonmuscle variant) [66], but the difference in
near the ATP binding pocket, resulting in an augmentation of biochemical and physiological characteristics between them
actomyosin ATPase activity in vitro [68]. The SMIA and the remains unknown.
SM IB also show distinct characteristics in their expression in
tissues. The SMIA is present in almost all SMC containing
tissues, whereas the SM IB is dominant in intestinal and urinary Myosin light chain kinase/te/okin
bladder SMCs but not in vascular SMCs. These results suggest
that the 7-amino acid insertion ofSMIB is important for the A key molecule in the myosin-linked regulation is myosin
contractile property of non-vascular SMCs. In addition to light chain kinase (MLCK) which is expressed in both smooth
smooth muscle variants, vascular SMCs express two non- muscle and norimuscle cells. The central catalytic domain of
muscle isoforms, NMHC-A and NMHC-B, under normal and MLCK is homologous to that of other protein kinases. The
pathological conditions. The NNMC-A and NMM-B from regulatory calmodulin-binding domain involved in the
chicken non-muscle tissues show different electrophoretic activation of kinase by Ca 2+/Calmodulin is located in the
mobilities with Mr of 196 and 198 kDa, respectively [70]. C-terminal region of the catalytic domain [82].
These nonmuscle isoforms are the products of different genes The C-terminus ofMLCK is known to be expressed as an
[70, 71]. The NMHC-B is identical with SMemb that is independent protein named "telokin", a 24 kDa acidic protein
expressed in embryonic aorta and neointimal lesion [72]. [83]. cDNA and genomic analyses of MLCK and telokin
During development, the embryonic/fetal non-muscle type revealed that the telokin mRNA is transcribed from a pro-
MHC may be replaced with adult smooth muscle type variants, moter which is located within the intron of the MLCK gene
and the coexistence ofthern is granted even in the adult [73]. [84]. Unlike MLCK, the expression of telokin is restricted
The SM2 is thought to be a marker for a late stage of dif- in some SMC containing tissues, but not in aortic SMCs and
ferentiation because it is firstly detected in vascular SMCs at nortmuscle tissues [84].
112

SM22 in both smooth muscle and nonmuscle tissues such as brain,

heart, stomach, and kidney, as well as dedifferentiated
There are three isofonns of SM22; the a isoform is most vascular SMCs [95). Two variant calponins are encoded in
dominant [85-87]. SM22a is abundantly expressed in all separate genes in human [96].
SMC containing tissues, while the biological and physio- Genomic structure of basic calponin has been recently
logical functions of this protein remain unclear. SM22a reported [97, 98). The calponin gene is composed of7 exons
contains a motif which appears in basic calponin with three and is about 10 kbp in length. The TATA box is not present
times. The SM22a gene is 6-7 kbp in length and composed in the promoter region and no more than 549 bp of 5'
offive exons. The 5'-flanking 445bp sequences preceding the sequence is needed for the augmentation ofpromoter activity
transcription initiation site of the mouse SM22a gene are in dedifferentiated vascular SMCs [97). In the core promoter
sufficient to activate transcription in cultured cells [88]. The region, E box, GC box, CCAAT box andAp2 site are granted,
SM22a promoter contains two CArG boxes and a MEF-2 but precise definition of cis element has yet to be elucidated.
binding site. The fact that deletion of the CArG boxes results Although the expression of calponin is restricted in SMCs,
in loss of transcriptional activity suggests that these CArG calponin mRNA transiently appears in heart during mouse
boxes play an important role in the expression of SM22a. embryogenesis [97]. The diverse expression of calponin in
Experiments using transgenic mice revealed that the 5'- embryonic heart and SMC containing tissues suggests that
flanking 445bp sequences of the SM22a promoter are common mechanism(s) may be involved in the development
essential for the tissue-specific expression of SM22a in of heart and SMC containing tissues.
mouse embryos [88]. The SM22a endogene is transiently
expressed in the heart and skeletal muscle cells in the
myotomal compartment ofthe somites during mouse embryo- Vinculinlmetavinculin
genesis, and comes to be expressed in all SMCs only at
adulthood [89). The expressional pattern of SM22a-LacZ Vinculin is a 117 kDa cytoskeletal protein located at cell-cell
transgene corresponds well to that of the SM22 endogene, and cell-matrix adherence, and is involved in linking F-actin
but the transgene expression is detected in neither venous nor to the cytoplasmic domain of cell-adhesive receptors such as
visceral SMCs [88, 90]. These results suggest that the the cadherin and integrin families (99). It is widely expressed
5'-flanking 445bp sequences of the SM22a promoter can in various tissues and cells. Three antigenetically indistin-
only respond to arterial SMCs and other elements in the guishable isoelectrophoretic isoforms (a, p, and y) have been
promoter may control the expression of SM22a in the venous reported: a- and p-forms are found in all cell types and a
and visceral SMCs. y-form is selectively expressed in cardiac and smooth muscles
[100). Additionally, a high Mr form of vinculin (130-150
kDa) called metavinculin is also expressed in smooth,
Calponin skeletal, and cardiac muscles [100, 10 I). In SMCs, both
vinculin and metavinculin are located in the dense membrane.
In vitro analysis revealed that calponin inhibits the actin-
The expression ofvinculin isofonns in SMCs is regulated in
activated myosin Mg2+ -ATPase activity and blocks movement
a SMC phenotype-dependent manner; the metavinculin
of actins in motility assay (reviewed in ref. [10)). According
expression is developmentally upregulated, but is down-
to these findings, calponin has been thought to be a possible
regulated during dedifferentiation [48, 100, 102). Char-
regulator of smooth muscle relaxation, whereas the localiza-
acterization of genomic structure of vinculin revealed that the
tion of this protein in SMCs does not accurately correspond
expression of vinculin and metavinculin is regulated by
to the actomyosin system [91).
alternative splicing from a single gene [103]; exon 19 encoding
Two major variants of calponin, basic/M calponin and
metavinculin-specific amino acid sequence is selected in
acidic/h2 calponin, have been demonstrated by cDNA
differentiated SMC. The metavinculin expression is therefore
cloning [92, 93]. Avian basic/hi calponin contains three
SMC phenotype-dependent. However, metavinculin is also
characteristic repeats in its C-terminus. Basic calponin is
expressed in other mesodermal originated muscles indicating
exclusively expressed in SMC containing tissues and its
that metavinculin is not a specific molecular marker for SMCs.
expression is downregulated in dedifferentiated SMCs [94].
7hese findings suggest that splicing machinery ofthe vinculin
Another novel variant of calponin, acidic/h2 calponin, has
gene seems to be different from those of the CaD and a-
been isolated from rat aorta [95). Acidic calponin contains
tropomyosin genes (Fig.3). The vinculin gene is known as a
three tandem repeats like basic calponin. The initial 273
member ofimmediate early response genes whose expressions
residues are highly homologous to basic calponin, but the
are rapidly stimulated by serum growth factors [104, 105).
remaining 57 residues at its C-tenninus are unique and quite
Indeed, the vinculin promoter contains the CArG box and
acidic. Unlike basic calponin, acidic calponin is expressed
partial promoter analysis demonstrated serum-inducibility,
 113

suggesting that the CArG box in the vinculin gene would act CArG
as a functional serum response element [103].

al integrin
CArG
SMMHC
Integrins, heterodimeric transmembrane proteins, consist ofa promoter
and ~ subunits and playa role in cell-to-cell and cell-to-extra-
cellular matrix adhesions as well as intracellular signal trans-
duction [106]. Cell adhesion mediated by integrins is important
for cell differentiation, proliferation, migration, wound SM22
healing, and metastasis. There are at least IS a and 8 ~ promoter
subunits, and their combinations generate various integrins
that are widely distributed in most cell types. Among varia- CArG
tions of the integrin family, a I ~1 integrin is a receptor for a-SM actin CCTTGTTTGG
GGAACAAACC
both laminin [107] and collagens [108]. During quail promoter
embryogenesis, al integrin is transiently expressed in
nervous tissues and skeletal and cardiac muscles, and CArG
constantly in microvascular endothelium. It is dramatically
increased in SMC containing tissues during development
~-TM
promoter 1:im~~TATA~
[109]. In the adult quail, the a 1 integrin expression is
restricted in visceral and vascular SMCs and microvascular
endothelium, whereas it is absent from most epithelial a1 integrin
CArG
promoter -----~Mc~~--~C
I I
tissues. Additionally, the al integrin expression is down-
regulated in serum-induced dedifferentiation ofSMCs [17,
11 0] and in some leiomyosarcomas [19]. Thus, the a 1
integrin expression is closely associated with phenotype of Fig. 5. The CArG box is common positive elements in the promoter regions
SMCs. of SMC-speciflic marker genes. The sequences of the CArG boxes in the
promoter regions ofSMC-specific marker genes are schematically shown.
Transcriptional machineries of the integrin genes such as The CArG box is the only c&-element conserved in all SMC-specific marker
a2 [Ill], a4 [112-115], CDlla (aL) [116], COlic (aX) genes listed here.
[117] have been reported. In general, the integrin promoters
lack TATA and CAAT boxes but contain multiple Spl, ApI
and/or AP2 binding sites. In the case of al integrin, its Conclusion
expression only depends upon the CArG box in spite of the
presence of Ap I, Ap2 and Sp I binding sites in its promoter As described in part V. (Molecular markers for phenotype of
region. The SRF is also a core trans-acting factor for the SMCs), the essential cis-element and trans-acting factor of
CArG box of the a 1 integrin promoter [118). SMC-specific marker genes have been identified as the CArG
In situ hybridization revealed that the al integrin mRNA box or CArG box-like motif (Fig. 5) and the SRE respectively.
is dominantly expressed in vascular and visceral SMCs at late Interaction between CArG box and SRF might be involved
stages of chick embryo; it is detected in 8-day old embryo in the transcriptional activation of SMC-specific marker
and is gradually increased in proportion to the expression of genes. The SRF is also observed to increase in mesoderm-
h-caldesmon mRNA thereafter [118). Such an expressional originated muscles such as skeletal, cardiac, and smooth
profile of the al integrin mRNA almost completely coincides muscles during embryonic development [120). In skeletal
with its protein level in developing embryos [17]. In addition muscle cells, the expression of SRF dramatically increases
to SMCs, low but significant amounts of al integrin mRNA in association with differentiation from myoblast to myo-
are detected in the kidney. Korhonen et at. [119] also found tube [120). In addition, transient transfection ofSRF cDNA
al integrin expression in mesangial and endothelial cells in induces activation of the skeletal a-actin promoter [121).
fetal and adult human kidney. The a I integrin is dramatically These findings support the possible involvement of SRF in
downregulated during dedifferentiation of primary cultured a wide variety of muscle differentiation-related gene
SMCs by serum-stimulation [17). These results indicate that expression.
the expression of the al integrin gene is transcriptionally The SRF was originally discovered as the SRE (serum
regulated in a SMC phenotype-dependent manner and is response element) or CArG box binding factor of immediate
available as a molecular marker for SMC differentiation. early response gene, c-fos [122). Subsequently, it has been
114

therefore reasonable to consider that SRF is a core tran-

scription factor for the CArG box binding. In order to resolve
this missing-link, other unknown factors which confer SMC
specificity might be required for the SMC-specific tran-
scriptional activation. We tentatively named the putative
genes encoding these unknown factors as SMC regulator

, ,
tyrosine phosphorylation genes (Fig. 6). In relation to our notion, some transcription
: pathway: factors have been identified in SMCs as possible candidates
for such SMC regulator genes. In vertebrates, skeletal,
SMC regulator genes? cardiac, and smooth muscles contain functionally saturating
l :

(transcription factors) ( splicing factors) levels of MEF2 trans-acting factors that are absent in non-

\
muscle cells. However, MEF2 alone is insufficient to produce

CaD
~
~ exon3
the full muscle phenotype [125]. In a loss-of-function of the
mej2 gene in Drosophila embryos, somatic, cardiac, and
visceral muscles do not differentiate [126], suggesting that
~~ ~
-----BB •
t..:::..:J CaD D-mej2 acts at a relatively late stage within different myo-
SM MHC genic lineages to control differentiation. The homeobox gene
SMC-speclflc exon tinman is coexpressed with D-mej2 in the ventral mesoderm
SM22
~ CArG CArG TATA
exon
2 and subsequently becomes restricted to the dorsal vessel. In
tinman mutant embryos, heart and visceral muscles fail to
~c~m -1cArGHcArG~TATA~ aTM - [ } - [ } - form [127, 128]. Therefore, tinman seems to be one of the
~ 2a 2b earliest genes required for the heart and visceral muscle
~TM e5 ~ • development. Homeobox sequences also isolated from an
a1 ~ SMC-speclflc exon
adult rat vascular SMC eDNA library including Hoxl.ll,
integrin e5a~
Hox-I.4, and Hox-l. 3 [129]. The interaction ofother homeo-
box protein, Mhox/Phox, with the SRF at the CArG box is
Fig. 6. Summary of expressional regulation of the SMC-specific marker
significant because this homeobox protein could potentially
genes. Signal transductions mediated by IGFI/IGFI receptor and laminin regulate the gene expression [130]. Despite ofthese evidences,
plays a vital role in the expressional regulation of SMC-specific marker the expression of MEF2 and homeobox proteins described
genes. Tyrosine phosphorylation is involved in this signal transduction above is not restricted within a lineage of SMCs. Future
pathways. Signaling pathways control the putative 'SMC regulator genes'
studies are required for isolating the putative SMC-specific
which may confer the SMC specificity on transcription and splicing factors.
For example, some unknown factor encoded by the SMC regulator gene
regulator genes which also govern the SMC-specific splicing-
may activate SRF-CArG interaction, resulting in enhancement of SMC- such as CaD and a.-tropomyosin (Fig. 6). Progress has been
specific marker gene driving. made in establishing a SMC culture system maintaining a
differentiated phenotype. This culture system will provide a
valuable system with which to identify and isolate key
demonstrated that SRF-CArG(SRE) interaction is involved signalling molecules and SMC regulator genes involved in
in other immediate early response genes [123]. In the case control of SMC phenotype.
of SRF-linked c-fos transactivation, the signal transduction
pathways have been well studied [124]. One potent pathway
is mediated by Ras, Raf, MEK and ERK. Another pathway Acknowledgements
is the Rho familymediated MAP kinase cascade. Both
pathways regulate the function of ternary complex factors We wish to thank Dr. Setsuro Ebashi for encouraging and
(TCFs) or SRF by phosphorylation, finally activating the supporting our studies for a long time. This work is supported
SRF-mediated c-fos driving. to K.S. By Grants-in-aid for COE research from the ministry
Taken together, SRF-CArG interaction might be involved of Education, Science, and Culture ofJapan.
in distinctively different biological activities; cell prolifera-
tion and differentiation. As described here, SRF-CArG
interaction directly links to the transcriptional activation of
SMC-specific marker genes. However, SRF is a ubiquitous
References
transcription factor involved in a variety ofcell activities. The I. Poole v, Coffin JD: Vasculogenesis and angiogenesis: Two distinct
transcriptional activation of SMC-specific marker genes morphogenetic mechanisms establish embryonic vascular pattern. J
cannot be simply explained by SRF-CArG interaction. It is Exp Zool 251: 224-231, 1989
 115

2. Noden DM: Embryonic origins and assembly of blood vessels. Am 22. Sobue K, Tanaka T, Kanda K,Ashino N, Kakiuchi S: Purification and
Rev Respir Dis 140: 1097-1103, 1989 characterization of caldesmon77: A calmodulin-binding protein that
3. Carey DJ: Control of growth and differentiation of vascular cells by interacts with actin filaments from bovine adrenal medulla. Proc Nat!
extracellular matrix proteins. Annu Rev Physiol 53: 161-177, 1991 Acad Sci USA 82: 5025-5029, 1985
4. Risau W, Lemmon V: Changes in the vascular extracellular matrix 23. Ueki N, Sobue K, Kanda K, Hada T, Higashino K: Expression of high
during embryonic vasculogenesis and angiogenesis. Dev Bioi 125: and low molecular weight caldesmons during phenotypic modulation
441--450, 1988 ofsmooth muscle cells. Proc Nat! Acad Sci USA 84: 9049-9053,1987
5. Geiger B, Dutton AH, Tokuyasu KT, Singer SJ: Immunoelectron 24. Haruna M, Hayashi K, Yano H, Takeuchi 0, Sobue K: Common
microscope studies of membrane-microfilament interactions: distribu- structural and expressional properties of vertebrate caldesmon genes.
tions of alpha-actinin, tropomyosin, and vinculin in intestinal epithelial Biochem Biophys Res Commun 197: 145-153,1993
brush border and chicken gizzard smooth muscle cells. J Cell Bioi 91 : 25. Hayashi K, Yano H, Hashida T, Takeuchi R, Takeda 0, Asada K,
614-628,1981 Takahashi E, Kato I, Sobue K: Genomic structure of the human
6. Draeger A, Stelzer EH, Herzog M, Small JV: Unique geometry of caldesmon gene. Proc Nat! Acad Sci USA 89: 12122-12126,1992
actin-membrane anchorage sites in avian gizzard smooth muscle cells. 26. Humphrey MB, Bryan J, Cooper TA, Berget SM: A 32-nucleotide
J Cell Sci 94: 703-711, 1989 exon-splicing enhancer regulates usage of competing 5'-splice sites
7. Sellers JR, Adelstein RS: The mechanism ofregulation ofsmooth muscle in a differential internal exon. Mol Cell Bioi 15: 3979-3988, 1995
myosin by phosphorylation. Curr Top Cell Regul 27: 51--62, 1985 27. Yano H, Hayashi K, Haruna M, Sobue K: Identification of two distinct
8. Sobue K, Sellers JR: Caldesmon, a novel regulatory protein in smooth promoters in the chicken caldesmon gene. Biochem Biophys Res
muscle and noninusele actomyosin systems. J Bioi Chern 266: 12115- Commun 201: 618--626,1994
12118,1991 28. Yano H, Hayashi Y, Momiyama T, Saga H, Haruna M, Sobue K:
9. Ishikawa R, Yamashiro S, Matsumura F: Differential modulation of Transcriptional regulation of the chicken caldesmon gene. Activation
actin-severing activity of gelsolin by multiple isoforms of cultured rat of gizzard-type caldesmon promoter requires a CArG box-like motif.
cell tropomyosin. Potentiation of protective ability of tropomyosins J Bioi Chern 270: 23661-23666, 1995
by 83-kDa non muscle caldesmon. J Bioi Chern 264: 7490-7497, 1989 29. Momiyama T, Hayashi K, Obata H, Chimori Y, Nishida T, Ito T,
10. Winder SJ, Walsh MT: Calponin. CurrTop Cell Regul 34: 33--61, 1996 Kamiike W, Matsuda H, Sobue K: Functional involvement of serum
II. Ross R: The pathogenesis of atherosclerosis: A perspective for the response factor in the transcriptional regulation of caldesmon gene.
1990s. Nature 362: 801-809, 1993 Biochem Biophys Res C:ommun 242: 429--435, 1998
12. Chou RG, Stromer MH, Robson RM, Huiatt TW: Assembly of 30. Ebashi S: The Croonian lecture, 1979: Regulation of muscle con-
contractile and cytoskeletal elements in developing smooth muscle traction. Proc R Soc Lond B Bioi Sci 207: 259-286, 1980
cells. Dev Bioi 149: 339-348,1992 31. Huber PA, Fraser ID, Marston SB: Location of smooth-muscle myosin
13. Charnley-Campbell J, Campbell GR, Ross R: The smooth muscle cell and tropomyosin binding sites in the C-terminal288 residues of human
in culture. Physiol Rev 59: 1--61, 1979 caldesmon. Biochem J 312: 617--625, 1995
14. Campbell JH, Koeher 0, Skalli 0, Gabbiani G, Campbell GR: 32. Yamashiro-Matsumura S, Matsumura F: Characterization of 83-
Cytodifferentiation and expression of alpha-smooth muscle actin kilodalton nonmuscle caldesmon from cultured rat cells: Stimulation of
mRNA and protein during primary culture of aortic smooth muscle actin binding of nonmuscle tropomyosin and periodic localization along
cells. Correlation with cell density and proliferative state. Arterio- microfilaments like tropomyosin. J Cell Bioi 106: 1973-1983, 1988
sclerosis 9: 633--643, 1989 33. Hendricks M, Weintraub H: Multiple tropomyosin polypeptides in
15. Iivanainen A, Wolteenaho R, Sainio K, Eddy R, Shows TB, Sariola chicken embryo fibroblasts: Differential repression of transcription by
H, Tryggvason K: The human laminin beta 2 chain (S-laminin): Rous sarcoma virus transformation. Mol Cell Biol4: 1823-1833, 1984
Structure, expression in fetal tissues and chromosomal assignment of 34. Lin JJ, Helfman DM, Hughes SH, Chou CS: Tropomyosin isoforms in
the LAMB2 gene. Matrix Bioi 14: 489--497,1995 chicken embryo fibroblasts: Purification, characterization, and changes
16. Glukhova M, Koteliansky V, Fondacci C, Marotte F, Rappaport L: in Rous sarcoma virus-transformed cells. J Cell Bioi I00: 692-703, 1985
Laminin variants and integrin laminin receptors in developing and adult 35. Lees-Miller JP, Helfman DM: The molecular basis for tropomyosin
human smooth muscle. Dev Bioi 157: 437--447,1993 isoform diversity. Bioessays 13: 429--437, 1991
17. Belkin VM, Belkin AM, Koteliansky VE: Human smooth muscle 36. Helfman DM, Cheley S, Kuismanen E, Finn LA, Yamawaki-Kataoka
VLA-I integrin: Purification, substrate specificity, localization in Y: Nonmuscle and muscle tropomyosin isoforms are expressed from
aorta, and expression during development. J Cell Bioi III: 2159- a single gene by alternative RNA splicing and polyadenylation. Mol
2170,1990 Cell Bioi 6: 3582-3595, 1986
18. Duband JL, Belkin AM, Syfrig J, Thiery JP, Koteliansky VE: 37. Libri D, Lemonnier M, Meinnel T, Fiszman MY: A single gene codes
Expression of alpha 1 integrin, a laminin-collagen receptor, during for the beta subunits of smooth and skeletal muscle tropomyosin in
myogenesis and neurogenesis in the avian embryo. Development 116: the chicken. J Bioi Chern 264: 2935-2944, 1989
585--600, 1992 38. Wieczorek DF, Smith CW, Nadal-Ginard B: The rat alpha-tropomyosin
19. Mechtersheimer 0, Barth T, Quentmeier A, Moller P: Differential gene generates a minimum of six different mRNAs coding for striated,
expression of beta I integrins in nonneoplastic smooth and striated smooth, and nonmuscle isoforms by alternative splicing. Mol Cell Bioi
muscle cells and in tumors derived from these cells. Am J Pathol 144: 8: 679--694, 1988
1172-1182, 1994 39. Lemonnier M, Balvay L, MoulyV, Libri D, Fiszman MY: The chicken
20. Hedin U, Bottger BA, Forsberg E, Johansson S, Thyberg J: Diverse gene encoding the alpha isoform oftropomyosin offast-twitch muscle
effects offibronectin and laminin on phenotypic properties of cultured fibers: Organization, expression and identification of the major proteins
arterial smooth muscle cells. J Cell Bioi 107: 307-319, 1988 synthesized. Gene 107: 229-240, 1991
21. Hayashi K, Saga H, Chimori Y, Kimura K, Yamanaka Y, Sobue K: 40. Mullen MP, Smith CW, Patton JG, Nadal-Ginard B:Alpha-tropomyosin
Differentiated phenotype of smooth muscle cells depends on signaling mutually exclusive exon selection: competition between branchpoint!
pathways through insulin-like growth factors and phosphatidylinositol polypyrimidine tracts determines default exon choice. Genes Dev 5:
3-kinase. J Bioi Chern 1998 (In Press) 642--655, 1991
116

41. Zamore PD, Patton JG, Green MR: Cloning and domain structure of 57. Sun S, Stoflet ES, Cogan 10, Strauch AR, Getz MJ: Negative regulation
the mammalian splicing factor U2AF. Nature 355: 60~14, 1992 of the vascular smooth muscle alpha-actin gene in fibroblasts and
42. Gooding C, Roberts GC, Moreau G, Nadal-Ginard B, Smith CW: myoblasts: Disruption ofenhancer function by sequence-specific single-
Smooth muscle-specific switching of alpha-tropomyosin mutually stranded-DNA-binding proteins. Mol Cell BioI 15: 2429--2436,1995
exclusive exon selection by specific inhibition of the strong default 58. Keirn RJ Jr., Sun S, StrauchAR, Getz MJ: Repression of transcriptional
exon. EMBO J 13: 3861-3872, 1994 enhancer factor-I and activator protein- I -dependent enhancer activity
43. Kashiwada K, Nishida W, Hayashi K, Ozawa K, Yamanaka Y, Saga H, by vascular actin single-stranded DNA binding factor 2. J Bioi Chern
Yamashita T, Tohyama M, Shimada S, Sato K, Sobue K: Coordinate 271: 24278-24285,1996
expression of alpha-tropomyosin and caldesmon isoforms in associa- 59. Shimizu RT, Blank RS, Jervis R, Lawrenz-Smith SC, Owens GK: The
tion with phenotypic modulation of smooth muscle cells. J Bioi Chern smooth muscle alpha-actin gene promoter is differentially regulated
1997 in smooth muscle versus non-smooth muscle cells. J Bioi Chern 270:
44. Toutant M, Gauthier-Rouviere C, Fiszman MY, Lemonnier M: 7631-7643,1995
Promoter elements and transcriptional control of the chicken tropo- 60. Hathaway DR, March KL, Lash JA,Adam LP, Wilensky RL: Vascular
myosin gene. Nucl Acids Res 22: 1838-1845, 1994 smooth muscle. A review of the molecular basis of contractility.
45. Vandekerckhove J, Weber K: At least six different actins are Circulation 83: 382-390, 1991
expressed in a higher mammal: An analysis based on the amino acid 61. Kiehart DP: Molecular genetic dissection of myosin heavy chain
sequence of the amino-terminal tryptic peptide. J Mol Bioi 126: 783- function. Cell 60: 347-350, 1990
802, 1978 62. Swynghedauw B: Developmental and functional adaptation of
46. Skalli 0, Ropraz P, Trzeciak A, Benzonana G, Gillessen D, Gabbiani contractile proteins in cardiac and skeletal muscles. Physiol Rev 66:
G: A monoclonal antibody against alpha-smooth muscle actin: A new 710-771,1986
probe for smooth muscle differentiation. J Cell Bioi 103: 2787-2796, 63. Umemoto S, Sellers JR: Characterization of in vitro motility assays
1986 using smooth muscle and cytoplasmic myosins. J Bioi Chern 265:
47. Owens GK, Thompson MM: Developmental changes in isoactin 14864-14869, 1990
expression in rat aortic smooth muscle cells in vivo. Relationship 64. Nagai R, Kuro-o M, Babij P, Periasamy M: Identification of two types
between growth and cytodifferentiation. J Bioi Chern 261: 13373- of smooth muscle myosin heavy chain isoforms by cDNA cloning and
13380, 1986 immunoblot analysis. J Bioi Chern 264: 9734-9737, 1989
48. Glukhova MA, Frid MG, Koteliansky VE: Developmental changes in 65. Kelley CA, Adelstein RS: The 204-kDa smooth muscle myosin heavy
expression of contractile and cytoskeletal proteins in human aortic chain is phosphorylated in intact cells by casein kinase II on a serine
smooth muscle. J BioI Chern 265: 13042-13046, 1990 near the carboxyl terminus. J Bioi Chern 265: 17876-17882, 1990
49. Blank RS, McQuinn TC, Yin KC, Thompson MM, Takeyasu K, 66. SomlyoAP: Myosin isoforms in smooth muscle: How may they affect
Schwartz RJ, Owens OK: Elements of the smooth muscle alpha-actin function and structure? J Muscle Res Cell Motil 14: 557-563, 1993
promoter required in cis for transcriptional activation in smooth muscle. 67. Sartore S, Scatena M, Chiavegato A, Faggin E, Giuriato L, Pauletto P:
Evidence for cell type-specific regulation. J Bioi Chern 267: 984-989, Myosin isoform expression in smooth muscle cells during physiological
1992 and pathological vascular remodeling. J Vasc Res 31: 61-81, 1994
50. Fatigati V, Murphy RA: Actin and tropomyosin variants in smooth 68. Kelley CA, Takahashi M, Yu JH, Adelstein RS: An insert of seven
muscles. Dependence on tissue type. J BioI Chem 259: 14383~14388, amino acids confers functional 'differences between smooth musclc
1984 myosins from the intestines and vasculature. J Bioi Chern 268: 12848-
51. Saga H, Kimura K, Hayashi K, Gotow T, Uchiyama Y, Momiyama T, 12854, 1993
Tadokoro S, Kawashima N, JimbouA, Sobue K: Phenotype-dependent 69. White S, Martin AF, Periasamy M: Identification of a novel smooth
expression of a-smooth muscle actin in visceral smooth muscle cells. muscle myosin heavy chain cDNA: Isoform diversity in the S I head
Exp Cell Res 1998 (In Press) region. Am J Physiol 264: CI252--e1258, 1993
52. Lazard D, Sastre X, Frid MG, Glukhova MA, Thiery JP, Koteliansky 70. Kawamoto S, Adelstein RS: Chicken nonmuscle myosin heavy chains:
VE: Expression of smooth muscle-specific proteins in myoepithelium differential expression of two mRNAs and evidence for two different
and stromal myofibroblasts of normal and malignant human breast polypeptides. J Cell Bioi 112: 915-924, 1991
tissue. Proc Natl Acad Sci USA 90: 999-1003, 1993 71. Simons M, Wang M, McBride OW, Kawamoto S, Yamakawa K, Gdula
53. Johnson RJ, Eda H, Alpers CE, Majesky MW, Schwartz SM, Pritzi P, D, Adelstein RS, Weir L: Human nonmuscle myosin heavy chains are
Gordon K, Gown AM: Expression of smooth muscle cell phenotype encoded by two genes located on different chromosomes. Circ Res
by rat mesangial cells in immune complex nephritis. Alpha-smooth 69: 530-539, 1991
muscle actin is a marker of mesangial cell proliferation. J Clin Invest 72. Kuro-o M, Nagai R, Nakahara K, Katoh H, Tsai RC, Tsuchimochi H,
87: 847-858, 1991 Yazaki Y, OhkuboA, Takaku F: cDNA cloning of a myosin heavy chain
54. Kirk TZ, Mark ME, Chua CC, Chua BH, Mayes MD: Myofibroblasts isoform in embryonic smooth muscle and its expression during vascular
from scleroderma skin synthesize elevated levels ofcollagen and tissue development and in arteriosclerosis. J Bioi Chern 266: 3768-3773, 1991
inhibitor ofmetalloproteinase (TIMP- I) with two forms ofTIMP-I. J 73. Kuro-o M, Nagai R, Tsuchimochi H, Katoh H, Yazaki Y, Ohkubo A,
Bioi Chern 270: 3423-3428, 1995 Takaku F: Developmentally regulated expression of vascular smooth
55. Kimura K, Saga H, Hayashi K, Obata H, Chimori Y, Ariga H, Sobue muscle myosin heavy chain isoforms. J Bioi Chern 264: 18272-18275,
K: c-Myc gene single-strand binding protein-I, MSSP-I, suppresses 1989
transcription ofa-smooth muscle actin gene in chicken visceral smooth 74. Kallmeier RC, Somasundaram C, Babij P: A novel smooth muscle-
muscle cells. Nucl Acids Res 26: 2420-2425, 1998 specific enhancer regulates transcription of the smooth muscle myosin
56. Cogan JG, Sun S, Stoflet ES, Schmidt U, Getz MJ, Strauch AR: heavy chain gene in vascular smooth muscle cells. J Bioi Chern 270:
Plasticity of vascular smooth muscle alpha-actin gene transcription. 30949--30957, 1995
Characterization of multiple, single-, and double-strand specific DNA- 75. White SL, Low RB: Identification of promoter elements involved in
binding proteins in myoblasts and fibroblasts. J Bioi Chern 270: 11310- cell-specific regulation of rat smooth muscle myosin heavy chain gene
11321,1995 transcription. J Bioi Chern 271: 15008-15017, 1996
 117

76. Watanabe M, Sakornura Y, Kurabayashi M, Manabe I, Aikawa M, 95. Applegate 0, Feng W, Green RS, Taubman MB: Cloning and
Kuro-o M, Suzuki T, Yazaki Y, Nagai R: Structure and characterization expression of a novel acidic calponin isoform from rat aortic vascular
of the T-flanking region of the mouse smooth muscle myosin heavy smooth muscle. 1 Bioi Chern 269: 10683-10690, 1994
chain (SM 1/2) gene. Circ Res 78: 978-989, 1996 96. Maguchi M, Nishida W, Kohara K, KuwanoA, Kondo I, Hiwada K:
77. Hasegawa Y, Morita F: Role of 17-kDa essential light chain isoforms Molecular cloning and gene mapping of human basic and acidic
of aorta smooth muscle myosin. 1 Biochem (Tokyo) III: 804-809, calponins. Biochem Biophys Res Commun 217: 238-244, 1995
1992 97. Mano 1M, Olson EN: Expression of the smooth muscle cell calponin
78. Helper 01, Lash lA, Hathaway DR: Distribution ofisoelectric variants gene marks the early cardiac and smooth muscle cell lineages during
of the 17,000-daiton myosin light chain in mammalian smooth muscle. mouse embryogenesis. 1 Bioi Chern 271: 7095-7103, 1996
1 Bioi Chern 263: 15748-15753, 1988 98. Samaha FF, Ip HS, Morrisey EE, Seltzer 1, Tang Z, Solway 1,
79. Malmqvist U, Arner A: Correlation between isoform, composition of Parmacek MS: Developmental pattern of expression and genomic
the 17 kDa myosin light chain and maximal shortening velocity in organization of the calponin-h I gene. A contractile smooth muscle
smooth muscle. Pflugers Arch 418: 523-530, 1991 cell marker. 1 Bioi Chern 271: 395--403, 1996
80. Lenz S, Lohse P, Seidel U, Arnold HH: The alkali light chains of 99. Burridge K, Fath K, KellyT, Nuckolls 0, Turner C: Focal adhesions:
human smooth and nonmuscle myosins are encoded by a single gene. transmembrane junctions between the extracellular matrix and the
Tissue-specific expression by alternative splicing pathways. 1 Bioi cytoskeleton. Annu Rev Cell Bioi 4: 487-525, 1988
Chern 264: 9009-9015,1989 100. Belkin AM, Ornatsky 01, Kabakov AE, Glukhova MA, Koteliansky
81. Nabeshima Y, Nonornura Y, Fujii-Kuriyama Y: Nonmuscle and smooth VE: Diversity of vinculin/meta-vinculin in human tissues and
muscle myosin light chain mRNAs are generated from a single gene cultivated cells. Expression of muscle specific variants of vinculin
by the tissue-specific alternative RNA splicing. 1 Bioi Chern 262: in human aorta smooth muscle cells. 1 Bioi Chern 263: 6631--6635,
10608-10612, 1987 1988
82. Olson Nl, Pearson RB, Needleman OS, Hurwitz MY, Kemp BE, Means 101. Sificiano ID, Craig SW: Properties of smooth muscle meta-vinculin.
AR: Regulatory and structural motifs of chicken gizzard myosin light 1 Cell Bioi 104: 473--482, 1987
chain kinase. Proc Natl Acad Sci USA 87: 2284-2288, 1990 102. Glukhova MA, Kabakov AE, Frid MG, Ornatsky 01, Belkin AM,
83. Ito M, Dabrowska R, Guerriero V lr., Hartshorne 01: Identification in Mukhin DN, Orekhov AN, Koteliansky VE, Smimov VN: Modulation
turkey gizzard of an acidic protein related, to the C-terminal portion of human aorta smooth muscle cell phenotype: A study of muscle-
of smooth muscle myosin light chain kinase. 1 Bioi Chern 264: 13971- specific variants of vinculin, caldesmon, and actin expression. Proc
13974, 1989 Nat! Acad Sci USA 85: 9542-9546, 1988
84. Gallagher Pl, Herring BP: The carboxyl terminus of the smooth muscle 103. Moiseyeva EP, Weller PA, Zhidkova NI, Corben EB, Patel B, lasinska
myosin light chain kinase is expressed as an independent protein, I, Koteliansky VE, Critchley DR: Organization of the human gene
telokin. 1 Bioi Chern 266: 23945-23952, 1991 encoding the cytoskeletal protein vinculin and the sequence of the
85. Lees-Miller IP, Heeley DH, Smillie LB, Kay CM: Isolation and vinculin promoter. 1 Bioi Chern 268: 4318--4325, 1993
characterization of an abundant and novel 22-kDa protein (SM22) from 104. AlmendraUM, Sommer D, Macdonald-Bravo H, Burckhardt 1, Perera
chicken gizzard smooth muscle. 1 Bioi Chern 262: 2988-2993, 1987 1, Bravo R: Complexity of the early genetic response to growth factors
86. Lees-Miller IP, Heeley DH, Smillie LB: An abundant and novel protein in mouse fibroblasts. Mol Cell Bioi 8: 2140-2148, 1988
of22 kDa (SM22) is widely distributed in smooth muscles. Purification 105. Rollins Rl, Stiles CD: Serum-inducible genes. Adv Cancer Res 511-
from bovine aorta. Biochem 1244:705-709, 1987 32, 1989
87. Pearlstone lR, Weber M, Lees-Miller IP, Carpenter MR, Smillie LB: 106. Hemler ME: VLA proteins in the integrin family: Structures,
Amino acid sequence of chicken gizzard smooth muscle SM22 alpha. functions, and their role on leukocytes. Annu Rev Immunol 8: 365-
1 Bioi Chern 262: 5985-5991, 1987 400, 1990
88. Li L, Miano 1M, Mercer B, Olson EN: Expression of the SM22 alpha 107. Hynes RO: Integrins: Versatility, modulation, and signaling in cell
promoter in transgenic mice provides evidence for distinct tran- adhesion. Cell 69: 11-25, 1992
scriptional regulatory programs in vascular and visceral smooth muscle 108. Ignatius Ml, Reichardt LF: Identification of a neuronal laminin
cells. 1 Cell Bioi 132: 849-859, 1996 receptor: An Mr 200Kll20K integrin heterodimer that binds laminin
89. Li L, Miano 1M, Cserjesi P, Olson EN: SM22 alpha, a marker of adult in a divalent cation-dependent manner. Neuron I: 713-725, 1988
smooth muscle, is expressed in multiple myogenic lineages during 109. Kramer RH, Marks N: Identification of integrin collagen receptors
embryogenesis. Circ Res 78: 188-195, 1996 on human melanoma cells. 1 Bioi Chern 264: 4684--4688, 1989
90. Moessler H, Mericskay M, Li Z, Nagi S, Paulin D, Small IV: The SM 110. Skinner MP, Raines EW, Ross R: Dynamic expression of alphal
22 promoter directs tissue-specific expression in arterial but not in beta!. and alpha2 betal integrin receptors by human vascular smooth
venous or visceral smooth muscle cells in transgenic mice. Develop- muscle cells. alpha2 beta I integrin is required for chemotaxis across
ment 122: 2415-2425, 1996 type I collagen-coated membranes. Am 1 Pathol 145: 1070-1081,
91. Mabuchi K, Li Y, Tao T, Wang CL: Immunocytochemical localization 1994
of caldesmon and calponin in chicken gizzard smooth muscle. 1 Muscle III. Zutter MM, Santoro SA, PainterAS, Tsung YL, Gafford A: The human
Res Cell Motil 17: 243-260, 1996 alpha2 integrin gene promoter. Identification of positive and negative
92. Takahashi K, Nadal-Ginard B: Molecular cloning and sequence regulatory elements important for cell-type and developmentally
analysis of smooth muscle calponin. 1 Bioi Chern 266: 13284-13288, restricted gene expression. 1 BioI Chern 269: 463--469, 1994
1991 112. Audet IF, Masson lY, Rosen GD, Salesse Q Guerin SL: Multiple
93. Strasser P, Gimona M, Moessler H, Herzog M, Small IV: Mammalian regulatory elements control the basal promoter activity of the human
calponin. Identification and expression of genetic variants. FEBS Lett alpha4 integrin gene. DNA Cell Bioi 13: 1071-1085,1994
330: 13-18, 1993 113. De Meirsman C, Schollen E, laspers M, Ongena K, Matthijs G,
94. Shanahan CM, Weissberg PL, Metcalfe lC: Isolation of gene markers Marynen P, Cassiman 11: Cloning and characterization of the promoter
of differentiated and proliferating vascular smooth muscle cells. Circ region ofthe murine alpha4 integrin subunit. DNA Cell Bioi 13: 743-
Res 73: 193-204, 1993 754, 1994
118

114. Rosen GD, BirkenmeierTM, Dean DC: Characterization of the alpha4 Oncogene 10: 1361-1370, 1995
integrin gene promoter. Proc NatiAcad Sci USA 88: 4094-4098, 1991 122. Treisman R: Transient accumulation of c-Ios RNA following serum
115. Rosen GD, Barks JL, lademarco ME, Fisher RJ, Dean DC: An stimulation requires a conserved 5' element and c-fos 3' sequences.
intricate arrangement of binding sites for the Ets family of tran- Cell 42: 889-902,1985
scription factors regulates activity of the alpha4 integrin gene 123. Treisman R: The SRE: A growth factor responsive transcriptional
promoter. J Bioi Chern 269: 15652-15660, 1994 regulator. Semin Cancer Bioi I: 47-58,1990
116. Nueda A, Lopez Cabrera M, Vara A, Corbi AL: Characterization of 124. Hill CS, Wynne J, Treisman R: The Rho family GTPases RhoA, Racl,
the CDlla (alphaL, LFA-I alpha) integrin gene promoter. J Bioi Chern and CDC42Hs regulate transcriptional activation by SRF. Cell 81:
268: 19305-19311, 1993 1159-1170,1995
117. Lopez Cabrera M, Nueda A, Vara A, Garcia Aguilar J, Tugores A, 125. Yu YT, Breitbart RE, Smoot LB, Lee Y, Mahdavi V, Nadal Ginard B:
Corbi AL: Characterization of the p150,95 leukocyte integrin alpha Human myocyte-specific enhancer factor 2 comprises a group of
subunit (CD II c) gene promoter. Identification ofcis-acting elements. tissue-restricted MADS box transcription factors. Genes Dev 6: 1783-
J Bioi Chern 268: 1187-1193, 1993 1798,1992
118. Obata H, Hayashi K, Nishida W, Momiyarna T, Sobue K: Smooth 126. Lilly B, Zhao B, Ranganayakulu G, Paterson BM, Schulz RA, Olson
muscle cell phenotype-dependent transcriptional regulation of the EN: Requirement of MADS domain transcription factor D-MEF2
alpha I integrin gene. J Bioi Chern 272: 26648-26651, 1997 for muscle formation in Drosophila. Science 267: 688-693, 1995
119. Korhonen M, Ylanne J, Laitinen L, Virtanen I: The alphal-alpha6 127. Azpiazu N, Frasch M: finman and bagpipe: Two homeo box genes
subunits of integrins are characteristically expressed in distinct that determine cell fates in the dorsal mesoderm of Drosophila. Genes
segments of developing and adult human nephron. J Cell Bioi III: Dev 7: 1325-1340,1993
1245-1254,1990 128. Bodmer R: The gene tinman is required for specification of the heart
120. Croissant JD, Kim JH, Eichele G, Goering L, Lough J, Prywes R, and visceral muscles in Drosophila. Development 118: 719-729, 1993
Schwartz RJ: Avian serum response factor expression restricted 129. Patel CV, Gorski DH, LePage DF, Lincecum J, Walsh K: Molecular
primarily to muscle cell lineages is required for alpha-actin gene cloning of a homeobox transcription factor from adult aortic smooth
transcription. Dev Bioi 177: 250-264, 1996 muscle. J Bioi Chern 267: 26085-26090,1992
121. Bushel P, Kim JH, Chang W, Catino JJ, Ruley HE, Kumar CC: Two 130. Gmeneberg DA, Natesan S, Alexandre C, Gilman NIZ: Human and
serum response elements mediate transcriptional repression of human Drosophila homeodomain proteins that enhance the DNA-binding
smooth muscle alpha-actin promoter in ras-transformed cells. activity of serum response factor. Science 257: 1089-1095, 1992
Molecular and Cellular Biochemistry 190: 119--124, 1999.
© 1999 Kluwer Academic Publishers.

Roles of intracellular Ca2+ receptors in the

pancreatic ~-cell in insulin secretion

Ichiro Niki and Hiroyoshi Hidaka

Department ofPharmacology, Nagoya University School ofMedicine, Nagoya, Japan

Abstract
Ca2+is the central second messenger in the regulation of insulin release from the pancreatic ~-cell; and intracellular Ca2+-binding
proteins, classified into two groups, the EF hand proteins and the Ca2+j phospholipid binding proteins, are considered to mediate
Ca2+signaling. A number of Ca binding proteins have been suggested to participate in the secretory machinery in the ~-cell.
Calmodulin, the ubiquitous EF hand protein, is the predominant intracellular Ca2+receptor that modulates insulin release via
the multiplicity of its binding to target proteins including protein kinases. Other Ca binding proteins such as calcyclin and the
Ca2+jphospholipid binding proteins may also be suggested to be involved. Ca 2+influx from the extracellular space appears to
be responsible for exocytosis of insulin via Ca2+-dependent protein/protein interactions. On the other hand, intracellular Ca 2+
mobilization resulting in secretory granule movement may be controlled by Ca 2+jcalmodulin-dependent protein phosphorylation.
Thus, Ca 2+exerts versatile effects on the secretory cascade via binding to specific binding proteins in the pancreactic ~-cells.
(Mol Cell Biochem 190: 119-124, 1999)

Key words: Ca 2+binding proteins, secretory granules, protein kinases

Introduction are going to overview the Ca binding proteins relevant to

transduction of Ca 2+signals in the pancreatic ~-cell.
Insulin secretion from the pancreatic ~-cell is finely regulated
in response to changes in blood glucose in vivo to maintain
The EF hand proteins
homeostasis of nutrient metabolism and its impairment is
involved in the pathology of NIDDM. Glucose induces The cytoplasmic Ca binding proteins comprise a large family
insulin release via several mechanisms involving intracellular classified into two groups; EF hand proteins and Ca 2+j
second messengers such as Ca 2+, cAMP and phospholipid phospholipid binding proteins. The former constitute a group
metabolites. Among these, Ca2+plays the most critical role of Ca binding proteins with 2--6 EF hand structures where
[I]. Thus, the secretory pathway in the ~-cell is composed Ca 2+binds. The EF hand proteins are further classified on the
of multiple steps, i.e., synthesis/processing of insulin, basis of the numbers of such binding structures (for a review,
secretory granule translocation and docking, as well as see Refs [3,4]), first discovered by Kretsinger in parvalbumin
priming and exocytosis, and Ca2+appears to act at more than and then demonstrated to be conserved in a variety of Ca
one stages through interactions with specific Ca binding binding proteins [5]. Some ofEF hand proteins are considered
proteins. Studies on Ca binding proteins originated with the to participate in the insulin secretory process.
discovery of troponin C by Dr Setsuro Ebashi, who stressed
that 'Ca research is to some extent a matter of Ca binding
proteins' [2]. The mechanisms of control of intracellularCa2+ Calmodulin
in the pancreatic ~-cell are now relatively well studied, but
we still only have limited knowledge of the processes Calmodulin, the most ubiquitous and multifunctional 4 EF
involved in its control of insulin release. In this review, we hand protein, was first identified as a Ca 2+-dependent

Address for offprints: H. Hidaka, Department of Pharmacology, Nagoya University School of Medicine, 65 Tsuruma-cho Showa-ku, Nagoya 466-8550, Japan
120

activator of cyclic nucleotide phosphodiesterase in brain, and of CaM kinases have been identified [18], and recent studies
later found to be distributed in alI eukaryotic celIs examined. have demonstrated the presence of activator kinases which
The presence of calmodulin in the pancreatic p-celI was phosphorylate and activate CaM kinase I and IV in a Ca2+j
first reported by Ashcroft and his colIeagues [6], and an active calmodulin-dependent manner [19]. Among these CaM
role in the control of secretory process was supported by the kinases, myosin light chain kinase (MLCK) was the first to
findings that various types of calmodulin antagonists inhibit be discovered CaM kinase in smooth muscle celIs [20], and
hormone release in response to Ca2+. Thus, insulin release was the first to be identi fied in the pancreatic p-celI [21]. MLCK
found to be inhibited by such antagonists as trifluoroperazine specificalIy phosphorylates the regulatory light chain of
[7] and W-7 [8], although there is argument concerning the myosin (MLC) and activates its actin-dependent ATPase
conclusion to be drawn [9]. The actions of calmodulin in the activity, this appearing to participate not only in smooth
p-celI have also been investigated in permeabilized celIs, but muscle contraction but also in motile events in non-muscle
simple addition did not alter insulin secretion from streptolysin- tissues [22]. In the p-celI, inhibition of this kinase by its
a-treated pancreatic islets [10], although it was reported that selective inhibitor, ML-9, has provided evidence that MLCK
exogenous calmodulin increased release from digitonin- may participate in the control of insulin secretion [23].
permeabilized islets in the absence of effective Ca2+ [II]. A Studies using permeabilized chromaffin celIs suggested that
possible involvement of calmodulin in the Ca2+-dependent it controls hormone secretion via its action on a proximal step
secretory processes, however, cannot be precluded, since its in the secretory cascade [24]. This idea was also supported
target protein(s) could be lost through the membrane pores by the finding that microinjection of anti-MLCK antibody
formed by the permeabilization treatment. In addition, it into superior cervical ganglion neurons resulted in retardation
might act in a non rate-limiting step. Microinjection of of neurotransmitter release [25]. Our recent findings using
calmodulin antibodies into cultured chromaffin celIs has pancreatic p-cells treated with streptolysin-O further indicate
suggested some roles in catecholamine secretion [I2], but this that protein phosphorylation by MLCK is involved in the
approach has not yet been employed with pancreatic p-celIs. control of (a) proximal and non-rate limiting step(s), possibly
It should be noted that the calmodulin content of the p-celI including intracelIular movement of secretory granules ([ 10]
is fairly high (36 JlM in Ref [5],50 JlM in Ref [13]) as and discussed below).
compared with the Km values for activation of related Some possible mechanisms have been proposed for CaM
enzymes: the Km values for calmodulin are 0.2 JlM with kinases and their substrates. The synapsins abundant in neural
MLCK, 0, I JlM with adenylate cyclase, and 10-50 JlM with celIs, found concentrated at the nerve terminals and bound
Ca 2+-ATPase. It is possible that other EF hand proteins may to the synaptic granules at the cytoplasmic surface, are
modulate some of its effects. For example, the glial-specific common substrates of CaM kinases I, II & IV and protein
S I00 protein which possesses two EF hand structures has kinases A & C. Phosphorylation of the synapsins occurs in
been reported to inhibit protein phosphorylation by Ca 2+j parallel to the release of neurotransmitters, suggesting that
calmodulin in brain [14]. A heterogeneous distribution of this process is involved in the secretory events in the neural
calmodulin has been noted in smooth muscle celIs [15]. cells [26]. A synapsin I-like protein has recently been
identified in a pancreatic p-cell line, MIN 6 [27] and CaM
kinase II has been found in pancreatic islets [23,28]. More-
How does calmodulin regulate insulin release? over, there is evidence that it is activated in the p-cell by
glucose and high K+ [29]. Therefore, phosphorylation of the
The multiple functions of calmodulin are due to the variety synapsin-like protein by CaM kinase II may function in the
of its target proteins. In the pancreatic islets, cyclic nucleotide secretory process of the p-celI.
phosphodiesterase and adenylate cyclase, which respectively There are also reports pointing to the presence of other
hydrolyze and synthesize cAMP, were found to be calmodulin- calmodulin-target proteins in pancreatic P-cells. Ca 2+-
dependent [5, 13]. Since glucose increases the p-celI cAMP transport in the p-cell was reported to be calmodulin-
level in an extracelIular Ca2+-dependent manner, this could, dependent [30] and related isoforms ofCa2+-ATPase (PMCAs
at least in part, be achieved via calmodulin-dependent 2b and 4b) have been identified [31], though their exact
adenylate cyclase [16]. Elevation of cAMP and resultant significance remains to be elucidated. Goosefish islet celIs
activation ofprotein kinase A, increase insulin release by direct contain at least three calmodulin-binding proteins in the
activation of the voltage-dependent Ca 2+ channel and by granule fraction [32]. Among them, the 65 kDa protein seems
increasing the sensitivity of the secretory process(es) to Ca2+ to be an isoform of synaptotagmin, a putative Ca2+ sensor
[17], although the responsible substrates are yet to be identified. protein, which forms a complex with other secretion-related
Another possible mode of calmodulin regulation of insulin proteins (see below). Interestingly, synaptotagmin is a
release is via protein phosphorylation by Ca 2+j calmodulin- substrate of CaM kinase II [33]. Therefore, it is tempting to
dependent kinases (CaM kinases). At least six di fferent types speculate that it might be involved in the regulation of the
 121

secretory process by Ca2+/calmodulin, although direct evi- Ca 2+-stimulated insulin release among those tested. This
dence for this is not available at present. suggests that cytoplasmic calcyclin responds to Ca 2+signal
to promote insulin secretion from the ~-cell. It may also be
involved in lactogen-II release from placental tissue [36],
Myosin (myosin light chain) although in this case the function could be extracellular.
One interesting property of S I00 proteins is their inter-
Myosin is a heteromer protein composed of two heavy chains action with other Ca binding proteins, including calmodulin,
and two sets of regulatory and essential light chains. These as discussed above. Some can also interact with the Ca 2+/
light chains possess four EF hand structures. In addition to phospholipid binding proteins, annexins in a specific manner,
its well-known role in muscle contraction, myosin and its for example, annexin I and calgizzarin, or annexin II and pi 0
isoforms are considered to participate in other cell functions [37, 38]. Indeed, annexin XI was originally identified as a
including the secretory processes in a varity of tissues. specific binding protein for calcyclin [39]. Since calcyclin is
Interaction ofmyosin and secretory granules in the cytoplasm reported to possess binding activity for other annexins, it is
has been described [34], and therefore, this Ca binding possible that it acts on the secretory process via such specific
protein has been suggested to playa role in translocation of binding.
granules (discussed below) though it is not clear whether its
binding to Ca2+mediates Ca 2+signaling in the ~-cell.
Annexins and insulin release
SIOO proteins
Annexins are a group of Ca binding proteins which in
The S I 00 proteins, originally designated from the bio- addition bind phospholipids in a Ca 2+-dependent manner.
chemical properties that they remain soluble in saturable They share a common structure called the core-domain in
(100%) ammonium sulphate. We have examined the effects their C-terminals, where four repeats of the 'endonexin fold',
ofpurified preparations on insulin secretion using streptolysin- the Ca 2+-binding site, are located. Thirteen members have so
O-permeabilized islets to introduce hydrophilic molecules far been reported to belong to this family [40]. Some have
such as Ca 2+, and large molecules such as Ca binding proteins been suggested to mediate Ca2+signals in secretory cells [40],
directly into the intracellular space [35]. As shown in Fig. I, since (I) they bind phospholipids in a Ca 2+-dependent
calcyclin was found to be the only EF hand protein to increase manner, (2) annexins I, II and VII cause Ca 2+-dependent

Ca2+ Additions

10-7

10-5

10- 5 calcyclin (100nM)

10-5 calgizzarin(265nM)

10- 5 calvasculin (545nM)

10- 5 neurocalcin (160nM)

o 200 400 600 800 1000 1200

Insulin release
(ng/5 islets/45min)

Fig. 1. Effects of EF hand proteins on Ca2+-induced insulin release from STLO-permeabilized pancreatic islets. Pancreatic islets permeabilized with STLO
were incubated with 10 f.IM Ca 2+in the presence or absence of purified Ca 2+-binding proteins. Released insulin during the 45min incubation was collected
and assayed by radioimmunoassay. Data are expressed by the mean ± S.E. for 4-8 observations.
122

aggregation and fusion ofvesicles and (3) addition ofannexin considered to be specific to neuronal cells in early studies,
II increases or restores Ca2+-induced catecholamine release but was later proven to be ubiquitous in secretory tissues.
from leaky chromaffin cells. As noted before, annexins Indeed, the pancreatic ~-cell was the site where a non-neural
possess binding activity to specific S I00 proteins. In this isoform was first identified [51]. Synaptotagrnin isoforms share
context, it is of interest that intracellular translocation of a characteristic structure with two C2 region structures in the
annexin II, which occurs in chromaffin cells in response to same molecule, and their binding of syntaxin in a Ca 2+-
nicotine, requires lower concentrations of Ca 2+ in the co- dependent manner seems to be responsible for mediation of
presence of an EF-hand protein, plO [41]. Recently it was Ca2+-signaling in the secretory machinery. Synaptotagmins also
reported that annexin VII could be a common target protein interact with clathrin-AP2, but independently ofCa2+[52].
not only for Ca 2+but also GTP in chromaffin cells [42]. Other C2 region proteins have also been suggested to
To our knowledge, only annexin I has been suggested to participate in the control of insulin secretion. For example,
participate in the secretory events in the pancreatic ~-cell [43], phospholipase isozymes like cytosolic phospholipase A2 and
being demonstrated to be localized in the secretory granules the ~, yand () isoforms of phospholipase C, exhibit the C2
and also phosphorylated in response to glucose stimulation regions and are activated by Ca 2+ increase, resulting in
by protein kinase(s) sensitive to H-7. Whether other annexins production of phospholipid metabolites which may alter the
are also involved in the secretory machinery remains to be secretory output of insulin. Another C2 region protein,
elucidated. rabphillin 3A, an effector of the secretion-related small G
protein, Rab3A [53]. Since peptides of the effector domain
ofRab proteins increase secretion from various permeabilized
Possible roles of C2 region proteins in the pancreatic endocrine and exocrine cells [54,55], interaction ofRab and
f3-cell rabphillin is considered to participate in the secretory events,
including insulin release [56].
The C2 region was originally designated from one of the
conserved domains ofprotein kinase C (PKC) isoforms which
Ca 2+and phospholipids bind. Activation of PKC increases Ca l + influx and Ca l + mobilization: to which do Ca l +
secretion of various hormones including insulin, but it is not binding proteins respond?
clear whether the Ca2+-binding nature of PKC is responsible
for mediating the Ca 2+ signal in the ~-cell. Since diacyl- Cytoplasmic Ca 2+in the pancreatic ~-cell is kept below 100
glycerol formation by phospholipase C is Ca 2+-dependent, nM, whereas extracellular Ca 2+concentration is in the mM
PKC might be activated by diacylglycerol produced by an order, and even higher concentrations ofCa (free and bound)
increase of intracellular Ca 2+. However, it seems that, in exist in the endoplasmic reticulum or secretory granules. Such
contrast to protein kinase A, insulin release due to PKC a heterogeneous distribution of Ca 2+is achieved by the Ca 2+
activation does not require stimulatory concentrations ofCa 2+ pump located in the plasma and intracellular membranes with
[44,45]. a contribution by Ca 2+ buffering proteins in the cytoplasm.
Whether activation ofPKC is indeed involved in glucose- Cytoplasmic Ca 2+ can be raised by two mechanisms; Ca 2+
induced insulin release is a matter of debate, although a few influx from the extracellular space through Ca 2+channels in
inhibitors have been reported to block the secretory process. the plasma membranes and Ca 2+mobilization from the intra-
H-7 is one of such inhibitor, but effects were only observed cellular store sites, mainly from the endoplasmic reticulum.
with glucose stimulation at rather high concentrations [46]. Secretory granules are another possible source of Ca 2+,
The second phase of the biphasic secretory response to though further work is needed to verify that this is the case
glucose is preferentially decreased by down-regulation of in practice.
PKC with 12-0-tetradecanoyl-phorbol 13-acetate (TPA) In the pancreatic ~-cell, Ca2+influx is primarily responsible
[47]. On the basis of findings with H-7, we previously for the release of insulin, because (I) depletion of extra-
proposed that this kinase might be involved in the priming cellular Ca2+or blockers of voltage-dependent Ca2+channels
effect of glucose on insulin secretion [48]. This is equivocal, inhibit insulin release from the ~-cell by many secretagogues,
however, because PKC-depleted islets retain their secretory (2) high K+ depolarization or sulphonylureas, which provoke
response to glucose [49]. The inconsistency may partly result Ca2+influx through voltage-dependent Ca2+channels, stimulate
from differences in the natures of the PKC expressed, since insulin release, (3) acetylcholine, which mainly mobilizes
multiple isoforms were recently identified in an insulinoma intracellular Ca2+via muscarinic activation of phospholipase
cell line, MIN 6 [50]. Obviously, the ability to specifically C, causes only a small increase in release.
inhibit the activity of each PKC isoform must be required to We have recently investigated intracellular movement of
solve this problem. Synaptotagmin, originally identified as insulin granules in living insulinoma cells (HIT T 15) under
a 65 kDa protein abundant in the synaptosome fractions, was a phase-contrast microscope [57]. The movement of the
 123

o 2. Ebashi S: Ca'+ in biological systems. Experientia 41: 978-981, 1985

3. Schafer BW, Heizmann CW: The SIOO family ofEF-hand calcium-
00 mobilized Ca 2 + binding proteins: Functions and pathology. Trends Biochem Sci 21:

o 00 134-140,199
4. Niki I, Yokokura H, Sudo T, Kato M, Hidaka H: Ca'+ signaling and intra-
cellular Ca binding proteins. 1 Biochem (Tokyo), 120: 685-698, 1996

111-
Granule 5. Kretsinger H: Structure and evolution of calcium-modulated proteins.
Protein
translocation CRC Crit Rev Biochem 8: 119-174,1980
phosphorylation 6. Sugden MC, Christie MR, Ashcroft SlH: Presence and possible role
ofcalcium-dependent regulator (calmodulin) in rat islets of Langerhans.
Docking FEBS Lett 105: 95-100,1979
7. Krauz Y, Wollheim CB, Siegel E, Sharp OWC: Possible role for
Primi :g Protei n/ protei n calmodulin in insulin release. 1 Clin Invest 66: 603-607, 1980
8. Niki H, NikiA, Hidaka H: Effects ofa new calmodulin inhibitor (W-7)
interaction
......
on glucose-induced insulin release and biosynthesis. Biomed Res 2:
Exocytosis
......
...... t
Ca 2 + influx
413--417,1981
9. Yaney OC, Sharp OWC: Calmodulin and insulin secretion: Use ofnaph-
thalenesulphonamide compounds. Am 1 Physiol259: E856-E864, 1990
10. Iida Y, Senda T, Matsukawa Y, Onoda K, Miyazaki 1-1, Sakaguchi H,
Nimura Y, Hidaka H, Niki I: Myosin light chain phosphorylation
Fig. 2. Putative model for mediation ofCa'+ signals by Ca binding proteins
controls insulin secretion at a proximal step in the secretory cascade.
in the pancreatic ~-cell.
Am 1 Physiol 273 (Endocrinol Metab 36): E782-E789, 1997
II. Colca lR, Wolf BA, Comens PO, McDaniel ML: Protein phos-
phorylation in permeabilized pancreatic islet cells. Biochem 1 228:
secretory vesicles is regulated by a Ca2+-dependent mechanism 529-536, 1985
12. Kenigsberg RL, Trifaro 1M: Microinjection of calmodulin antibodies
since it can be abolished by treatment with BAPTA and
into cultured chromatin cells blocks catecholamine release in response
appears to be distinct from that responsible for exocytosis. to stimulation. Neuroscience 14: 335-347,1985
Motility was also noted in non-stimulated ~-cells, which may 13. Valverde I, Vandermeers A, Anjaneyulu R, Malaisse Wl: Calmodulin
suggest that the movement of insulin granules requires less activation of adenylate cyclase in pancreatic islets. Science 206: 225-
Ca 2+ than the release process. Interestingly, IP 3-induced Ca 2+ 227, 1979
14. Patel 1, Marangos Pl, Heydorn WE, Chang 0, Verma A, Jacobowitz
mobilization from the intracellular Ca 2+ stores may be
D: S-IOO-mediated inhibition of brain protein phosphorylation. J
particularly important for activation of movement, firstly Neurochem 41: 1040--1045, 1983
because acetylcholine increases the number of moving 15. Luby-Phelps K, Hori M, Phelps JM, Won D: Ca'+-regulated dynamic
granules even in the co-presence of the Ca 2+ channel compartmentalization of calmodulin in living smooth muscle cells. J
blocker, nifedipine, and secondly because high k+, sul- BioI Chern 270: 21532-21538, 1995
16. Sharp OWC: The adenylate cyclase-cyclic AMP systeirn in islets of
phonylurea, and TPA failed to mimic this effect. Taking this
Langerhans and its role in the control of insulin release. Diabetologia
observation and the nature of MLCK discussed above into 16: 287-296, 1979
consideration, it is suggested that the calmodulin/ MLCK- 17. Amiilii C, Ashcroft FM, Rorsman P: Calcium-independent potentiation
related system may work as an intracellular receptors for of insulin release by cyclic AMP in single ~-cells. Nature 363: 356-
mobilized Ca 2+ from the endoplasmic reticulum [10]. In 358, 1993
18. Nairn AC, Picciotto MR: Calcium/calmodulin-dependent protein
contrast to the motile event, exocytosis, the final and
kinases. Cancer Bioi 15: 295-303,1994
rate-limiting step in the secretory cascade, may be controlled 19. Tokumitsu H, Brickey DA, Clod 1, Hidaka H, Sikela J, SoderiingTR:
by multiple proteins including synaptotagmin and possibly Activation mechanisms for Ca'+/calmodulin-dependent protein kinase
annexins, in response to Ca2+ entering from the extracellular IV. 1 BioI Chern 269: 28640--28647, 1994
space through Ca 2+ channels at the plasma membrane (Fig. 20. Dabrowska R,Aromatorio D, Sherry JMF, Hartshorne DJ: Modulator
protein as a component of the myosin light chain kinase from chicken
2). The function of additional Ca binding proteins (Ca 2+
gizzard. Biochemistry 17: 253-258, 1978
buffering proteins) might be to avoid disorganization of the 21. Penn EJ, Brocklehurst KW, SopwithAM, Hales CN, Hutton JC: Ca'+-
Ca2+ signal by keeping changes in intracellular Ca2+ localized. calmodulin-dependent myosin light-chain phosphorylating activity in
Thus Ca2+ signaling appears to be conducted by a teamwork insulin-secreting tissues. FEBS Lett 139: 4-8, 1982
of Ca binding proteins. 22. Tan lL, Ravid S, Spudich lA: Control of nonmuscle myosins by
phosphorylation. Annu Rev Biochem 61: 721-759, 1992
23. Niki I, Okazaki K, Saitoh M, Niki A, Niki H, Tamagawa T, Iguchi A,
Hidaka H: Presence and possible involvement of Ca/calmodulin-
References dependent protein kinases in insulin release from the rat pancreatic ~
cell. Biochem Biophys Res Commun 191: 255-261, 1993
24. Kumakura K Sasaki K Sakurai T, Ohara-Imaizumi M, Misonou H,
I. Ashcroft FM, Proks P, Smith PA, Amiilii C, Bokvist K, Rorsman P: Nakamura S, Matsuda Y, Nonomura Y: Essential role of myosin light
Stimulus-secretion coupling in pancreatic ~ cells. 1 Cell BioI 55S: chain kinase in the mechanism for MgATP-dependent priming of
54-65, 1994 exocytosis in adrenal chromaffin cells. J Neurosci 14: 7695-7703, 1994
124

25. Mochida S: Role of myosin in neurotransmitter release: Functional 43. Ohnishi M, Tokuda M, Masaki T, Fujimura T, Tai Y, Itano T, Matsui
studies at synapses formed in culture. J Physiol (Paris) 89: 83-94, H, IshidaT, Konishi R, Takahara J, Hatase 0: Involvement ofannexin-I
1995 in glucose-induced insulin secretion in rat pancreatic islets. Endo-
26. Valtorta F, Benfenati F, Greengard P: Structure and function of the crinology 136: 2421-2426,1995
synapsins. J Bioi Chern 267: 7195-7198,1992 44. Tamagawa T, Niki H, Niki A: Insulin release independent ofa rise in
27. Matsumoto K, K Fukunaga, J Miyazaki, M Shichiri, E Miyamoto: Ca"/ cytosolic free Ca" by forskolin and phorbol ester. FEBS Letts 183:
calmodulin-dependent protein kinase II and synapsin I-like protein in 430--432,1985
mouse insulinoma MIN 6 cells. Endocrinology 136: 3784-3793,1995 45. Komatsu M, Schermerhorn T, Aizawa T, Sharp GWG: Glucose
28. Hughes SJ, Smith H, Ashcroft SJH: Characterization of Ca"/ stimulation of insulin release in the absence of extracellular Ca" and
calmodulin-dependent protein kinase in rat pancreatic islets. Biochem in the absence of any increase in intracellular Ca" in rat pancreatic
J 289: 795-800, 1993 islets. Proc Nat! Acad Sci USA 92: 10728-10732,1995
29. Wenham RM, Lancit M, Eason RA: Glucose activates the multi- 46. Malaisse WJ, Sene A: Inhibition by 1- (5-isoquinolinesulfonyl)-2-
functional Ca2+/calmodulin-dependent protein kinase II in isolated rat methylpiperazine (H-7) of protein kinase C activity and insulin release
pancreatic islets. J Bioi Chern 269: 4947--4952, 1994 in pancreatic islets. IRCS Med Sci 13: 1239-1240, 1985
30. Pershadsingh HA, McDaniel ML, Landt M, Bry CG, Lacy PE, 47. Thams P, Capito K, Hedeskov CJ, Kofod H: Phorbol-ester-induced
McDonald JM: Ca"-activated ATPase and ATP-dependent cal mod down-regulation of protein kinase C in mouse pancreatic islets.
ulin-stimualted Ca" transport in islet cell plasma membrane. Nature Potentiation of phase 1 and inhibition of phase 2 of glucose-induced
288:492--495,1980 insulin release. Biochem J 265: 777-787,1990
31. Varadi A, Molnar E, Ashcroft SJH: A unique combination of plasma 48. Niki I, Tamagawa T, Niki H, Niki A, Koide T, Sakamoto N: (1988)
membrane Ca"-ATP ase isoforms is expressed in islets of Langerhans Possible involvement of diacylglycerol-activated, Ca"-dependent
and pancreatic ~-celliines. Biochem J 314: 663--669, 1996 protein kinase in glucose memory of the rat pancreatic B-cell. Acta
32. Watkins D, White BA: Identification and characterization of Endocrinologica (Copenh) 118: 204-208, 1988
calmodulin-binding proteins in islet secretion granules. J Bioi Chern 49. Hii CS, Jones PM, Persaud SJ, Howell SL: A re-assessment of the
260: 5161-5165,1985 role of protein kinase C in glucose-stimulated insulin secretion.
33. Popoli M: Synaptotagmin is endogenously phosphorylated by Ca"/ Biochem J 246: 489--493, 1987
calmodulin protein kinase II in synaptic vesicles. FEBS Lett 317: 85- 50. Tian YoM, Urquidi V, Ashcroft SJH: Protein kinase C in beta-cells:
88, 1993 Expression of multiple isoforms and involvement in cholinergic
34. Burridge K, Phillips JH: Association ofactin and myosin with secretory stimulation of insulin secretion. Mol Cell Endocrinol 119: 185-193,
granule membranes. Nature 254: 526-529, 1975 1996
35. Okazaki K, Niki I, Iino S, Kobayashi S, Hidaka H: A role of calcyclin, 51. Mizuta M, Inagaki N, Nemoto Y, Matsukura S, Takahashi M, Seino S:
a Ca" -binding protein, on the Ca" -dependent insulin release from the Synaptotagmin 1II is a novel isoform of rat synaptotagmin expressed
pancreatic ~ cell. J Bioi Chern 269: 6149--6152, 1994 in endocrine and neuronal cells. J Bioi Chern 269: 11675-11678, 1994
36. Thordarson G, Southard IN, Talamantes E: Purification and character- 52. Li C, Ullrich B, Zhang JZ, Anderson RGW, Brose N, Siidhof TC:
ization of mouse decidual calcyclin: a novel stimulator of mouse Ca2+-dependent and -independent activities of neural and non-neural
placental lactogen-II secretion. Endocrinology 129: 1257-1265, 1991 synaptotagmins. Nature 375: 594-599, 1995
37. Glenney JR Jr, Boudreau M, Galyean R, HunterT, Tack B: Association 53. Shirataki H, Kaibuchi K, Sakoda T, Kishida S, Yamaguchi T, Wada K,
of the S-I OO-related calpactin I light chain with the NH2-terminal tail Miyazaki M, Takai Y: Rabphilin-3A, a putative target protein for
of the 36-kDa heavy chain. J Bioi Chern 261: 10485, 1986 smgp25A/rab3A p25 small GTP-binding protein related to synapto-
38. MailliardWS, Haigler HT, Schlaepfer DD: Calcium-dependent binding tagmin. Mol Cell Bioi 13: 2061-2068,1993
of S IOOC to the N- terminal domain of annexin I. J Bioi Chern 271: 54. Senyshyn J, Balch WE, Holz RW: 1992 Synthetic peptides of the
719,1996 effector - binding domain of rab enhance secretion from digitonin-
39. Minami H, Tokumitsu H, Mizutani A, Watanabe Y, Watanabe M, permeabilized chromaffin cells. FEBS Letts 309: 41--46,1992
Hidaka H: Specific binding of CAP-50 to calcyclin. FEBS Letts. 305: 55. Padfield PJ, Balch WE, Jamieson JD: 1992 A synthetic peptide of
217-219,1992 the rab3a effector domain stimulates amylase release from per-
40. Raynal P, Pollard HB:Annexins: The problem ofassessing the biological meabilized pancreatic acini. Proc Nat! Acad Sci USA 89: 1656-1660,
role for a gene family of multifuncitional calcium- and phospholipid- 1992
binding proteins. Biochim Biophys Acta 1197: 63-93, 1994 56. Li G, Regazzi R, Balch WE, Wollheim CB: Stimulation of insulin
41. Chasserot-Golaz S, Vitale N, Sagot I, Delouche B, Dirrig S, Pradel release from permeabilized HIT-TI 5 cells by a synthetic peptide
LA, Henry J-P, Aunis D, Bader M-R: Annexin II in exocytosis: corresponding to the effector domains of the small GTP-binding protein
Catecholamine secretion requires the translocation of p36 to the rab3. FEBS Letts 327: 145-149, 1993
subplasmalemmal region in chromatin cells. J Cell Bioi 133: 1217- 57. Hisatomi M, Hidaka H, Niki I: Ca"/ calmodulin and cyclic 3,5'
1236,1996 adenosine monophosphate control movement of secretory granules
42. Caohuy H, Srivatrava M, Pollard HB: Membrane fusion protein through protein phosphorylation/ dephosphorylation in the pancreatic
synexin (annexin VII) as a Ca2+/ GTP sensor in exocytotic secretion. ~-cell. Endocrinology 137: 4644--4649, 1996
Molecular and Cellular Biochemistry 190: 125-131, 1999.
© 1999 Kluwer Academic Publishers.

Organization of connectin/titin filaments in

sarcomeres of differentiating chicken skeletal
muscle cells

Yoshinori Soeno, Hirohiko Yajima, Yuuki Kawamura, Sumiko Kimura,

Koscak Maruyama and Takashi Obinata
Department ofBiology, Faculty ofScience, Chiba University, Chiba, Japan

Abstract
Very long, elastic connectin/titin molecules position the myosin filaments at the center of a sarcomere by linking them to the
Z line. The behavior of the connectin filaments during sarcomere formation in differentiating chicken skeletal muscle cells
was observed under a fluorescent microscope using the antibodies to the N terminal (located in the Z line), C terminal (M line),
and C zone (myosin filament) regions of connectin and was compared to the incorporation ofa-actinin and myosin into forming
sarcomeres. In early stages of differentiating muscle cells, the N terminal region of connectin was incorporated into a stress
fiber-like structure (SFLS) together with a-actinin to form dots, whereas the C terminal region was diffusely distributed in the
cytoplasm. When both the C and N terminal regions formed striations in young myofibrils, the epitope to the C zone of A-band
region, that is the center between the A-I junction and the M-line, initially was diffuse in appearance and later formed definite
striations. It appears that it took some time for the Nand C terminal regions of connectin to form a regular organization in a
sarcomere. Thus the two ends of the connectin filaments were first fixed followed by the specific binding of the middle portion
onto the myosin filament during sarcomere formation. (Mol Cell Biochem 190: 125-131, 1999)

Key words: connectin/titin, a-actinin, myosin, sarcomere formation, chicken skeletal muscle

Introduction served to encourage TO in his research career to elucidate the

mechanism of muscle differentiation. All three authors join
It is a great pleasure for three of the present authors (SK, KM, in thanking Professor Ebashi for his warm friendship and
and TO) to celebrate Professor Setsuro Ebashi 's irreplaceable share an admiration of his neverending fighting spirit.
contribution to biological science: the first recognition ofthe In the sarcomere of differentiating myofibrils, actin and
role ofcalcium ion as cell signal. KM first met Professor Ebashi myosin filaments of uniform lengths are organized in a
in the spring of 1953, when Dr. S. V. Perry gave a seminar on hexagonal lattice with right polarity and ordered spatial
myosin ATPase at the Faculty of Medicine, University of position[ 1, 2]. The molecular mechanism by which the
Tokyo (Perry visited Japan at that time as Cambridge rugby filaments are arranged into a mature sarcomeric pattern has
team manager, not as a scientist). Since then the two have been investigated in various experimental systems by dif-
maintained close personal contact and KM has constantly been ferent methodologies, but it is still poorly understood (for
stimulated by his unwavering pursuit ofacademic excellence. reviews, see [3,4]).
KM contributed to the promotion of Prof. Ebashi 's discovery Connectin/titin is a giant filamentous protein (3,000 kDa)
ofa-actinin by showing its actin-gelation activity (1965). KM ofstriated muscles positioning the myosin filament at the center
and SK followed the Ebashi path oftroponin research as they of a sarcomere by linking the filament to the Z line [5, 6].
continued connectin/titin work. This pioneering path also Thus connectin plays a key role in sarcomere architecture.

Address for offprints: T. Obinata, Department of Biology, Faculty of Science, Chiba University, Yaoyoi-cho, Inage-ku, Chiba 263, Japan
126

It is interesting how connectin filaments are involved in 5) FL-1: Polyclonal antibodies against myosin light chain
the organization ofsarcomere assembly during differentiation [13]. Tests have shown that the staining pattern byA4.1025
of striated muscle cells, and a number of references on this is identical to that by FL-1 in cultured myocytes.
subject are available [7-9]. Furst and his associates [9] recently
observed the behaviors ofconnectin filaments in differentiating As secondary antibodies, horseradish peroxidase (HRP)-
human muscle cells using several kinds of antibodies specific labeled goat anti-mouse IgG (GAM) and HRP-labeled goat
to different regions of a connectin molecule. anti-rabbit IgG (GAR) were purchased from Bio-Rad
In the present study, the behaviors of connectin filament (Richmond, California, USA), and fluorescein (FITC)-
were recognized in differentiating chicken skeletal muscle labeled GAM and GAR, tetramethylrhodamine (TRITC)-
cells in culture together with a-actinin and myosin using labeled GAM and GAR from Tago (Burlingame, California,
several kinds of antibodies specific to the N-terminal, C USA), respectively.
terminal and myosin-binding regions of connectin. The
primary culture ofchicken skeletal muscle is ideal for detailed
analysis of the localization of each connectin epitope in the Muscle cell culture
structure of developing myofibrils, since sarcomeric struc-
tures are well formed in this culture. The N terminal region Chicken mononucleated myogenic cells were dissociated
of connectin binds to the a-actinin bands at almost the same from breast muscles of 12-day-old chicken embryos by
time as does the C-terminal region to the M line region of the mechanical dissociation [14] and filtered through 10 layers
myosin filaments. Thereafter, the myosin binding domains oflens paper. They were then plated on glass coverslips coated
bind firmly to the specific sites of A-bands. with collagen in 60-mm tissue culture dishes at a density of 5
x 106 cells per dish. The culture medium consisted of 85%
Eagle's minimum essential medium (MEM: Nissui Co.
Materials and methods Tokyo) supplemented with 2 mM L-glutamine, 15% horse
scrum, and 4% chick embryo extract. Cultures were main-
Antibodies tained in a humidified atmosphere with 5% CO 2 and 95% air
at 37°C.
The mouse monoclonal antibody (2C8) to connectin, which
recognizes several epitopes ofthe connectin portion localized
in A-band between the A-I junction and the M-Iine (see Fig. Immunoblots and immunoelectron microscopy
I), was prepared by a standard procedure using purified /3-
connectin from chicken breast muscle as an immunogen. Poly- A total SDS extract of chicken breast muscle was electro-
clonal antibodies against the C terminal region of connectin phoresed using 2.3--4% gradient polyacrylamide gels.
(PcnC) were raised in a rabbit using a recombinantconnectin Immunoblot detection was carried out as described [15].
peptide as an immunogen; this was produced in an Escherichia Immunoelectron microscopy was performed using a JEM
coli expression system with the cDNA encoding the C-terminal 100S electron microscope as previously reported [16].
region (654 amino acids) ofchicken skeletal muscle connectin
[10]. The fusion protein was obtained as a 6xHis tagged protein
and purified on a Ni-NTA-agarose column according to the Immunofluorescent labeling ofcells
manufacturer's direction (WAGEN). The antibody stained the
M line region in mature myofibrils (see Fig. 1). Cells grown on coverslips were fixed for 10 min in PBS
The other antibodies used as primary antibodies were: (phosphatebuffer saline: 0.15 M NaCI and 10 mM phosphate
buffer, pH 7.0) containing 4% paraformaldehyde at room
1) PcCOMl: Polyclonal antibodies against the N terminal temperature, permeabilized for 10 min in PBS containing 0.1 %
motif II of chicken skeletal muscle connectin, which stains Triton X-lOO at room temperature and washed with PBS.
the Z-line in mature myofibrils ([11] see Fig. 2). The fixed cells were blocked with PBS containing 1% BSA
2) mAb20: Mouse monoclonal antibody against zeugmatin. and 0.05% NaN 3 at 4°C overnight. They were then incubated
Zeugmatin is known to be identical to the N terminal re- with the primary antibodies, various combinations of mono-
gion ofconnectin [12] (Developmental Studies Hybridoma clonal and polyclonal antibodies against different regions of
Bank) (Fig. 2). connectin molecules or other myofibrillar proteins for 90
3) A7811: Mouse monoclonal antibody against skeletal min at room temperature. After extensive washing for 30
muscle a-actinin (Sigma, St. Louis, USA). min in PBS, they were incubated with a mixture of FITC-
4) A4.1025: Mouse monoclonal antibody against myosin GAM and TRITC-GAR for 60 min at room temperature.
heavy chain (Developmental Studies Hybridoma Bank). After washing several times, specimens were mounted in
 127

A B
abc z M z
a a
l3

N
b

MHC
c

Fig. I. Specificity of the antibodies and localization of their epitopes in a sarcomere of chicken breast muscle. (A) Immunoblot tests ofa total SDS extract
of chicken breast muscle. a - Amido Black stain; b - treated with Pc72C (polyclonal antibodies to connectin); c - treated with 2C8 (monoclonal antidody to
connectin). a - a-connectin; p - p-connectin; N - nebulin; MHC - myosin heavy chain. 2.3-4% polyacrylamide gels were used. (B) Immunoelectron
micrographs of chicken breast muscle sarcomeres treated with the antibodies. a - control without the first antibody treatment; b - treated with pc72e; c-
treated with 2C8. M - M line; Z - Z line. Bar, 0.5 flm.

anti-fader (l mg/ml p-phenylendiamine, 50% glycerol in According to Obermann et al. [17], the connectin filaments
PBS, pH 8.0) and examined under a ZEISS Axioskop epi- from the two Zlines in a sarcomere overlap at the M line.
fluorescence microscope. When treated with 2C8, the myofibrils exhibited three labeled
stripes in each half of the A-band (Fig. I-B, c). They are
located at the center between the A-I junction and the M-line
Results where C-protein is also located, although the localization has
not been exactly compared with the C-protein stripes. It is to
Characterization ofantibodies be noted that the two epitopes to 2C8 are almost the same as
the epitopes to no [9] (Fig. 2).
When the total SDS extract of chicken skeletal muscle was
examined by immunoblotting combined with SDS-PAGE,
both 2C8 and Pc72C interacted with the a- and ~-connectin a-actinin and the N terminal region of connectin
bands as shown in Fig. I-A. The bands above the nebulin
band were faintly stained when the combination ofPc72C and It is generally accepted that the periodic deposition of a-
HRP-GAR was used (Fig. I-A, b), but we conclude that these actinin into a stress fiber-like structure (SFLS) is the first sign
were due to staining by the second antibody, since the second of sarcomeric formation [18]. a-actinin is the main com-
antibody alone also showed these bands. ponent of the Z line. As shown in Fig. 3a, a-actinin appeared
To determine the locality of the epitopes to the two as lines of dots in SFLS of the 2nd day myocytes, whereas
antibodies in the myofibrillar structure, skinned muscle fibers the deposits of the N terminal region of connectin were
prepared from chicken breast muscle were treated with the unclear dots (Fig. 3b). On the 3rd day of culture, the a-actinin
antibodies and deposits of the antibodies were examined deposits became more or less Z line-like bands (Fig. 3c) and
under an electron microscope. When the skinned fiber was the bands of the N terminal region of connectin were some-
treated with Pc72C, deposits of the antibody were detected what wider at the same position as the a-actinin bands (Fig.
on both sides of the M-line approximately 100 nm apart, 3d). Interestingly, the deposits of the antibodies to the N
indicating that the epitopes of this antibody are located near terminal region of connectin attached to both sides of the
the C-terminus of the connectin molecule (Fig. I-B, b). Zline on the 4th day (Fig. 3f). On the 5th day, the bands of
128

Z line Myosin filament M line

~J1
. . . ,, , , ,, , ,,

/j~
" ~ "

·· .
" "
,,
I'
,,

1
N j DmC
PcCOM1 mAb20 (T12) (9Dl0) 2C8 (T30) (T31) Pc72C

Fig. 2. Schematic representation of the localization of connectin in a half sarcomere, indicating the positions of the connectin epitopes used in this study.
For T12, 9DI 0, T 30, T31, see van der Loop et al. [9].

connectin N-tenninal region became singular but were still The Nand C terminal regions ofconnectin
wider than the a-actinin bands (Fig. 3g, h).
The above observations suggest that the N tenninal portion Comparison of the assembly of the Nand C tenninal regions
ofconnectin associates with deposited a-actinin dots in SFLS of connectin in the same myofibrils was carried out by dual
in early phases of sarcomere fonnation. staining with the monoclonal antibody to the N terminal
region (mAb20) [12] and the polyclonal antibodies to the C
terminal region (PcnC). On the first day of culture, the
deposits of the monoclonal antibody to the N tenninal region
of connectin (mAb20) [12] were rather diffuse, although
some were on fibrous structures in mononucleated myoblasts
as seen in Fig. 4a. On the contrary, the deposits of polyclonal
. ....u::- • antibodies, pcnc, were diffuse in the cytoplasm of myo-
}) .
blasts (Fig. 4b). On the second day, the deposits of the
.,:"lIr, . .
~ . ........ antibody, mAb20, were lined dots in SFLS (Fig. 4c), while
the pcnc deposits were filamentous but not yet dotted in
SFLS (Fig. 4d). On the 3rd day of culture, both deposits were
periodically striated in young myotubes (Fig. 4e, f). As
myotubes developed, staining by mAb20 and pcnc was
restricted to the Z-line region and the M-line region at the
center of the A-band, respectively, although staining by
mAb20 was a little wider (Fig. 4g, h). In well developed
myotubes, mAb20 and pcnc stained Z-line and M-line
regions as narrow clear bands, respectively (Fig. 4i,j). These
observations strongly suggest that N-tenninal and C-tenninal
ends of connectin are fixed to specific regions of the sarco-
meres at an early phase of myofibrillogenesis.
.,
J') --~- ' The C terminal region ofconnectin and myosin

On the second day of culture, the monoclonal antibody to

Fig. 3. Immunofluorescence micrographs ofcultured chicken skeletal muscle
myosin heavy chain (A4.1 025) stained filamentous structures
cells stained with anti-a-actinin (A 7811) and antibodies to the C terminal
region of connectin (PcCOM I). a, c, e, g - treated with A7811; b, d, f, h - in SFLS without periodic structures (Fig. 5a). In contrast,
staining by the antibody to the C tenninal region of connectin
°
treated with PcCOM 1. a, b - 2nd day of culture; c, d - 3rd day; e, f - 4th day;
g, h - 5th day. Arrowheads indicate the position of Z-line. Bar, 1 J.lrn. was detected mostly in a diffused pattern (Fig. 5b). On the
 _
129

.._.~.
}) ~

Fig. 5. Immunofluorescence micrographs of cultured chicken muscle cells

stained with anti-myosin heavy chain antibody (A4.1 025) and antibodies
to the C terminal region of connectin (PcnC). a, c, e, g - treated with
A4.1025; b, d, f, h - treated with pcne. a, b - 2nd day of culture; c, d -
3rd day; e, f - 4th day; g, h - 5th day. Arrowheads in c and d indicate the
position of M-line and those in g and h indicate the position of Z-line. Bar,
Fig. 4. Immunofluorescence micrographs of cultured chicken skeletal 10 ~m.
muscle cells stained with antibodies to the N terminal (mAb20) and C
terminal (PcnC) regions of connectin. a, c, e, g, i-treated with mAb20;
b,d, f, h, j - treated with pcnc. a, b - Ist day of culture; c, d - 2nd day; e,
of culture, the deposits of pcnc fonned some periodic stripes
f - 3rd day; g, h - 4th day; i, j - 5th day. Arrowheads indicate the position
ofZ-line. Bar, 10 ~m. on differentiating myotubes (Fig. 6b), whereas there were
random dots on the thin bundles in the 2C8-treated myotubes
(Fig. 6a). The most striking difference between the staining
3rd day of culture, in the region where A band-like structures patterns by the two antibodies was observed on the 4th day.
were fonned (Fig. 5c), the C tenninal region ofconnectin was The 2C8 epitopes were clearly seen on the entire A band (Fig.
localized to form stripes in the nascent myofibrils (Fig. 5d, 6c), while pcnc was deposited only at the center ofthe A band
arrowheads) at the center of the A-bands (Fig. 5c, arrow- as clear-cut stripes (Fig. 6d), as seen in mature sarcomeres (Fig.
heads). When the A band was fonned clearly in myotubes 6f). On the 5th day, the 2C8 epitopes became doublet bands
on the 4th day of culture (Fig. 5e), the C tenninal region of on the A band (Fig. 6e), suggesting that the antibodies bound
connectin was largely deposited on the M line region of this to the epitope sites seen in the mature sarcomeres. Thus it is
band, but the stripes were irregular (Fig. 5f). Completed evident that the antibodies to the epitopes in the middle portion
myofibrils show the colocalization of myosin and the C of connectin filaments bind to specific sites on the myosin
tenninal region of connectin in sarcomeres (Fig. 5g, h). filament after the two ends of the filaments are fixed.

The C terminal and myosin binding regions ofconnectin The myosin binding region of connectin and myosin

Assembly patterns of the C terminal and myosin binding In early stages of differentiating muscle cells (the 2nd day
regions of connectin were compared by dual staining of of culture), the localization of both 2C8 epitope and myosin
developing myotubes with the polyclonal antibodies to the C was obscure (Fig. 7a, b).
tenninal region (PcnC) and the monoclonal antibody to the On the 3rd day, myosin became assembled into sarcomeric
myosin filaments halfway to the M line (2C8). On the third day structures (Fig. 7d), but the 2C8 epitopes were still mostly
130

Fig. 6. Immunofluorescence micrographs of cultured chicken skeletal

muscle cells stained with antibodies to the myosin binding region (2C8)
and the C-terminal region (PcnC) of connectin. a, c, e - treated with 2C8;
Fig. 7. Immunofluorescence micrographs of cultured chicken muscle cells
b, d, f - treated with pcne. a, b - 3rd day of culture; c, d - 4th day; e, f-
stained with antibodies to the myosin binding region of connectin (2C8)
5th day. Arrowheads indicate the position of Z-line. Bar, I0 ~m.
and anti-myosin light chain antibodies (FL-I). a, c, e, g - treated with 2C8;
b, d, f, h - treated with FL-I. a, b - 2nd day of culture; c, d - 3rd day; e, f
- 4th day; g, h - 5 th day. Arrowheads indicate the position of Z-line. Bar,
diffused, although they were detected in some areas as dots 10 ~m.
along differentiating myofibrils (Fig. 7c). When theA bands
were periodically aligned (Fig. 7f), the 2C8 epitopes were
also visible in periodic structures and the stripes stained by the myosin and actin filaments are arranged with accurate
2C8 were almost the same as that of myosin, much wider than polarity and ordered spatial position into a mature sarcomeric
those in mature myofibrils (Fig. 7e). As muscle development pattern (for reviews, see [3,4])
progressed further, the staining pattern of anti-myosin The dynamic role of connectin in sarcomere organization
antibody remained unchanged (Fig. 7h), but the stripe pattern has recently been emphasized [7, 9]. Connectin is synthesized
of 2C8 became significantly narrower just as in mature in mononuclear myoblasts and can be seen as diffuse dots in
myofibrils (Fig. 7g). Thus the organized localization of the the cytoplasm by immunofluorescence microscopy (cf. Fig.
connectin filament on the myosin filament appears to occur 4b). Connectin is incorporated into SFLS, possibly guided
after the formation of the A band. by desmin intermediate filaments [23]. The present work has
shown that the N terminal region of connectin binds to SFLS,
very possibly via a pre-formed a-actinin structure in SFLS
Discussion (Fig. 3).
The C terminal region of connectin bound to SFLS at
It is generally accepted that myofibrillogenesis is initiated by almost the same time as the N terminal region, or perhaps
the formation of a stress-fiber-like structure (SFLS) com- slightly later (Fig. 4). This C terminal region bound to the
posed primarily of actin filaments and followed by the center of the A band already formed in sarcomeres (Fig. 5).
incorporation of myosin filaments in both developing and It is to be noted that, although both Nand C terminal regions
cultured muscle cells [7, 8]. Actin-binding proteins and of connectin bound early to SFLS, time was required to form
myosin-binding proteins are both involved in the assembly their regular structures.
offilaments and their organization into sarcomeric structures The behavior of the myosin binding domain of connectin
[3, 4, 19-22]. However, it is still not well understood how is of interest: As seen in Fig. 6, when the C terminal end
 131

completely bound to the M line of the A band (Fig, 6d), the 10. Yajima H, Ohtsuka H, Kume H, Endo T, Maruyama E, Kimura S,
antibody 2C8 bound to the whole A band (Fig. 6c). Later, 2C8 Maruyama K: Molecular cloning of a partial cDNA clone encoding
the C terminal region of chicken breast muscle connectin. Zool Sci
formed regular doublet striation in the A band (Fig. 6e). The 13: 119-123,1996
binding of 2C8 to the entire A band was confirmed by double 11. Yajima H, Ohtsuka H, Kawamura Y, Kume H, Murayama T, Abe H,
staining with antibodies to myosin light chains (Fig. 7), and Kimura S, Maruyama K: A 11.5 kb 5'-terminal cDNA sequence of
was due to the reorganization of the 2C8 epitope region of chicken breast muscle connectin/titin reveals its Z line binding region.
connectin on the myosin filament. This delay in reorganiza- Biochem Biophys Res Comm 223: 160-164, 1996
12. Turnacioglu KK, Mittal B, Sanger 1M, Sanger JW: Partial char-
tion of the 2C8 epitope localization was reported earlier by acterization ofzeugmatin indicates that it is a part of the Z band region
van der Loop et al. [9] using the monoclonal antibody no, oftitin. Cell Motil Cytoskel34: 108-121,1996
the epitopes of which are localized near those of 2C8 (Fig. 13. Obinata T, Masaki T, Takano H: Immunochemical comparison of
2). The present work first showed the incorporation of the N myosin light chains from fast skeletal, slow skeletal and cardiac muscle.
and C terminal regions of connectin into SFLS together with J Biochem 86: 131-137, 1979
14. Ii I, Kimura I, Ozawa E: A myotropic protein from chick embryo
a-actinin. extract; its purification, identity to transferrin, and indispensability
for avian myogenesis. Dev Bioi 94: 366-377, 1982
15. Kimura S, Matsuura T, Ohtsuka S, Nakauchi Y, MatsunoA, Maruyama
Acknowledgments K: Characterization and localization ofa-connectin (titin I): An elastic
protein isolated from rabbit skeletal muscle. J Muscle Res Cell Motil
13: 39-47, 1992
This research was supported in part by research grants from 16. Itoh Y, Suzuki T, Kimura S, Ohashi K, Higuchi H, Sawada H, Shimizu
the Ministry of Education, Science and Culture, and the T, Shibata M, Maruyama K: Extensible and less-extensible domains
ofconnectin filaments in stretched vertebrate skeletal muscle sarcomeres
National Center ofNeurology and Psychiatry (NCNP) of the as detected by immunofluorescence and immunoelectron microscopy
Ministry of Health and Welfare of Japan. using monoclonal antibodies. 1 Biochem 104: 504-508, 1988
17. Obermann WM1, Gaut;1 M, Steiner F, van der Yen PFM, Weber E,
Furst DO: The structure of the sarcomeric M band: Localization of
defined domains of myomesin, M-protein, and the 250-kD carboxy-
References terminal region of titin by immunoelectron microscopy. J Cell Bioi
134: 1441-1453,1996
I. Fischman DA: An electron microscope study of myofibril formation 18. Hill CS, Duran S, Zhong XL, Weber K, Holtzer H: Titin and myosin,
in embryonic chick skeletal muscle. J Cell Bioi 32: 557-575, 1967 but not desmin are linked during myofibrillogenesis in postmitotic
2. Shimada Y, Obinata T: Polarity of actin filaments at the initial stage of mononuclear myoblasts. 1 Cell BioI, 103: 2185-2196, 1986
myofibril assembly in myogenic cells in vitro. J Cell Bioi 72: 777- 19. Gilbert R, Kelly MG, MikawaT, Fischman DA:The carboxyl terminus
785, 1977 of myosin binding protein C(N 1Y13P-C, C-protein) specifies in-
3. Epstein HF, Fischman DA: Molecular analysis of protein assembly in corporation into the A-band of striated muscle. 1 Cell Sci 109: 101-
muscle development. Science 251: 1039-1044, 1991 111,1996
4. Obinata T: Contractile proteins and myofibrillogenesis. Int Rev Cytol 20. McKim KS, Matheson C, Marra MA, Wakarchuk MF, Baillic DL The
143: 153-189,1993 Caenorhabditis elegans unc-60 gene encodes proteins homologous to
5. Maruyama K: Connectin, an elastic protein of striated muscle. Biophys a family of actin-binding proteins. Mol Gen Genet 242: 346-357, 1994
Chern 50: 73-85, 1994 21. Nagaoka R, Minami N, Hayakawa E, Abe H, Obinata T: Quantitative
6. Maruyama K: Connectin/titin, giant elastic protein of muscle. FASEB analysis of low molecular weight actin-binding proteins, ADF and
J, 1997, in press profilin, expressed in developing and degenerating chicken skeletal
7. Shimada Y, Korniyarna M, Begurn S, Maruyama K: Development of muscles. J Muscle Res Cell Motil 17: 463--473, 1996
connectin/titin and nebulin in striated muscles of chicken. Adv Biophys 22. Obi nata T, Nagaoka-Yasuda R, Ono S, Kusano E, Mohri K, Ohtaka Y,
33: 223-234, 1996 Yamashiro S, Okada K,Abe H: Low molecular-weight G-actin binding
8. van der Loop FTL: Cell biological aspects of muscle cell dif- proteins involved in the regulation of actin assembly during myo-
ferentiation, Thesis, Univ Maastricht, 1996 fibrillogenesis. Cell Str Func 1997
9. van der Loop FTL, van der Yen PFM, Furst DO, Gautel M, van Eys 23. van der Yen PFM, Schaart G, Croes WE, lap PHK, Ginsel LA,
EM, Ramackero FCS: Integration of titin into the sarcomeres of Ramaekers FCS: Titin aggregates associated with intermediate
cultured differentiating human skeletal muscle cells. Eur 1 Cell BioI filaments align along stress fiber-like structures during human skeletal
69: 301-307, 1996 muscle differentiation. J Cell Sci 106: 749-759, 1993
Molecular and Cellular Biochemistry 190: 133-141, 1999.
© 1999 Kluwer Academic Publishers.

Detection of a sequence involved in actin-binding

and phosphoinositide-binding in the N-terminal
side of coIDin
Ken-ichi Kusano, I Hiroshi Abe 2 and Takashi Obinata l , 2
1Department ofBiological Sciences, Graduate School ofScience, University of Tokyo, Hongo, Bukyo-ku, Tokyo;

2Department ofBiology, Faculty ofScience, Chiba University, Yayoi-cho, Inage-ku, Chiba, Japan

Abstract
Cofilin is an actin-binding protein of low molecular weight which is widely distributed in eukaryotes and is deeply involved
in the dynamics of actin assembly in the cytoplasm. The actin-binding ability of cofilin is inhibited by inositol phosphates
(PIP 2), and the PIP 2- and actin-binding site(s) has been localized in residues W104 - Mils of the cofilin primary sequence
(Yonezawa et al. 1991). In the present study, in order to further clarify the functional domains in cofilin molecule, we constructed
expression vectors containing cDNAs of different size with deletion at the 3'-region of the open reading frame. The truncated
cofilin molecules produced in E. coli were purified and examined for their actin-binding and PIP 2-binding ability. We found
that the truncated cofilin molecule without C-terminal residues #100-#166 including the previously-described actin-binding
site could be cross-linked with actin by EDC, a zero-length cross-linker. In addition, these truncated peptides as well as synthetic
peptides corresponding to the N-terminal sequence of cofilin suppressed the inhibitory action of PIP 2 on actin-cofilin interaction.
These results strongly suggest that additional actin- and PIP 2-binding sites exist in the N-terminal region ofcofilin. (Mol Cell
Biochem 190: 133-141, 1999)

Key words: cofilin, actin-binding protein, actin, inositol phosphate

Abbreviations: r-cofilin - recombinant cofilin; SDS-PAGE - SDS-polyacrylamide gel electrophoresis; EDC - l-ethyl-3-(3-
dimethylaminopropyl) carbodiimide

Introduction organization; for example, mutations that inactivate the cofilin/

ADF genes in Drosophila melanogaster [9], Caenorhabditis
Cofilin is an actin-binding protein of about 20 kDa which elegans [10], and Saccharomyces cerevisiae [II, 12] severely
binds to both G- and F-actin, inhibits binding of tropomyosin damage their viability. Cofilin seems to be deeply involved
to F-actin, and inhibits actin-myosin interaction [I]. Actin in actin dynamics during cytokinesis, since it is accumulated
depolymerizing factor (ADF) with a slightly smaller size is in the contractile ring [13, 14]. This protein is also important
also known to have a similar function [2]. The two proteins as a regulator for reorganization of actin in developing and
have been categorized in the same protein family, the cofilin/ degenerating muscle cells [15-17].
ADF family, because of their high sequence homology, [3-5] It has been demonstrated that the activity of cofilin can be
and functional similarity [I, 6, 7]. Proteins of this family have regulated in several ways. For example, cofilin remains
been detected in both muscle and non-muscle cells of a associated with F-actin at neutral pH but disassembles
variety of eukaryotes (for review, see [8]) and recently F-actin rapidly into G-actin at alkaline pH [I, 18]. Further,
considerable information has accumulated regarding their phosphoinositides, especially PIP 2, have been considered as
possibly essential role for the structure and function of actin a suppresser of cofilin activity [17, 19]. Phosphorylation is

Address for offprints: T. Obinata, Department of Biology, Faculty of Science, Chiba University, Yayoi-cho, Inage-ku, Chiba 263, Japan
134

another major way to modulate the activity of cofilin and digested and blunted withXbal and Klenow polymerase. For
ADF. In the oocytes of Xenopus laevis, almost 100% of CD66-C, CMC-16 was digested with £CoO 1091 and blunted
Xenopus ADF/cofilin (XAC) is phosphorylated, but more with Klenow polymerase. The fragment excised by subsequent
than 60% of the protein is dephosphorylated within 30 min digestion with Pstl was ligated into pCMC-16, the insert of
after fertilization (13]. The phosphorylated form of cofilin is which was removed by digestion withPstl andNeol, and the
free from actin filament in vivo and lacks binding ability to Nco 1 site was blunted with Klenow polymerase. The truncated
actin in vitro [20]. Biochemical analyses revealed that cDNAs in pBluescript IIKS+ vector were digested with Nco I
phosphorylatedADF neither binds to G-actin nor affects the and BamHl and ligated into the Ncol and BamHl sites of
rate or extent ofactin assembly, but removal ofphosphate from pET-3d vector. The constructs were confirmed by sequencing
phosphorylated ADF restores full activity to depolymerize and restriction enzyme mapping.
F-actin [21].
The basic functional domains have been localized in the
cofilin and ADF sequences; that is, it was established that the Expression and purification ofr-cojilin
sequence Trp I 04-Met l15 is responsible for actin-binding and that
Lys ll2 and Lysl14 in this sequence are crucial for this [22, 23). BL2l (DE3) pLysS cells transformed with expression con-
The sequence Trplo4-Met ll5 is also known to interact with structs were grown at 37°C in an LB medium containing
phosphoinositides [23]. A stretch of basic amino acids, 5 mg/ml of ampicillin. IPTG was added to a final con-
K30KRKK34, functions as a signal for heat shock- or dimethyl centration of 0.5 mM, when OD600 of the cell suspension
sulfoxide-induced nuclear translocation ofcofilin [24, 25). The reached 0.6. Cells were harvested at 3 h post-induction by
functional phosphorylation site of both ADF and cofilin was centrifugation at 10,000 g for 30 min. The cells were washed
recently determined as SerJ in the N-terminal side [20, 21). In once with TEN solution (0.1 M NaCI, 20 mM EDTA, 20 mM
the primary structure, however, the phosphorylation site is Tris-HCI, pH 8.0) and resuspended in a solution composed
considerably apart from the actinlphosphoinositides-binding site. of 10% sucrose, 20 mM EDTA, 50 mMTris-HCI, pH 8.0, and
In the present study, assuming that additional sequences incubated for 30 min at 40°C. The cells were then lysed in a
involved in the interaction with actin might exist in a cofilin solution containing 5% Triton X-I 00, 31 mM EDTA, 1 mM
molecule to enable the complicated regulation of cofilin- PMSF and 25 mM Tris-HCI, pH 8.0 at 40°C.
actin interaction, we examined the sequence in the N- Wild type r-cofilin was purified by the method described
terminal side from a functional viewpoint. We report here previously [17] with slight modification. Briefly, the lysate
that a sequence responsible for actin-binding and phospho- was centrifuged at 20,000 g for 1 h at 40°C and r-cofilin was
inositide-binding is present in the N-terminal side of cofilin, precipitated from the supernatant with ammonium sulfate of
near the phosphorylation site. 60-80% saturation. The pellet was dissolved in a solution
containing 0.2 M KCI, 0.1 mM DTT, 20 mMTris-HCI, pH 7.5,
dialyzed against the same buffer at 40°C, and the protein
Materials and methods solution was applied to a hydroxylapatite column equilibrated
with the same solution. r-cofilin was eluted with 0.6 M KCI
Construction ofexpression vectors - 10 mM K-phosphate buffer, pH 7.0. The eluate was con-
centrated with a YM3 membrane (Amicon) and subjected to a
Full-length chicken cofilin cDNA cloned into pBluescript Sephadex G-75 column equilibrated with 0.1 M KCI, 0.1 mM
IIKS+ (pCMC-16) [3] was digested with Nco 1 andBamHl. The DTT, 0.01% NaN 3, 20 mM Tris-HCl, pH 7.5. The fractions
construct for expression of wild-type cofilin (pET-3d-COF) containing cofilin ofhigh purity were collected by monitoring
was obtained by ligating the cDNA into pET-3d vector by SDS-PAGE. The purified r-cofilin was dialyzed against an
linearized with Ncol and BamHl as described previously assaybuffer(O.l mM DTT, 0.01% NaN3, 2 mM HEPES-KOH,
[17]. Four C-terminal truncated mutant cDNAs (CDN-66, pH 7.0). The protein was further purified by centrifugation
CDN-llO, CDN-116, and CDN-146) and an N-terminal at 130,000 g for 2 h at 40°C before use.
truncated mutant cDNA (CD66-C) of cofilin were created by The truncated r-cofilin became mostly insoluble by forming
using restriction enzymes and Klenow polymerase (see Fig. inclusion bodies when expressed in E. coli. Therefore, the
3); CMC-16 was digested with Ec047 1 for CDN-66 and bacterial cells were sonicated after being lysed as described
CDN-146, with SauAl for CDN-116, and with Dral for above, and centrifuged at 20,000 g for 30 min at 40°C. The
CDNl16, and blunted with Klenow polymerase. The frag- pellet was washed repeatedly with a solution containing 0.1
ments were excised by digesting with Kpn 1 for insertion into M NaCI, 2% Triton X-lOO, 20 mM EDTA, and 20 mM
the Bani 11/Kpnl site ofpBluescript IIKS+ vector. The CDN- Tris-HCI, pH 8.0, and then with a solution containing 50 mM
110 was created by ligating the Dral-EcoRV fragment of KCI, 5 mM EDTA, 10 mM DTT, and 20 mMTris-HCI, pH 8.0.
CMC-16 into pBluescript IIKS+ vector which had been The inclusion bodies were solubilized in a solution of8 M urea,
 135

50 mM KCI, 10 mM EDTA, 20 mM Tris-HCI, pH 8.3. After Co-sedimentation assay

incubation for 2 h at room temperature, the mixture was diluted
four times with 50 mM KCI, 5 mM EDTA, and 20 mM The effects of truncated cofilin and synthetic peptides on
Tris-HCI, pH 8.0, to obtain 2 M urea concentration, and then the PIP z-dependent regulation of cofilin-actin interaction
centrifuged at 185,000 g for 2 h at 40°C. The resultant were examined in a solution containing 40 mM KCI, 0.04
supernatant was dialyzed against 50 mM KCI and centrifuged mM CaCl z, 0.1 mM DTT, 0.01% NaN 3' and 20 mM HEPES-
at 190,000 g for 2 h at 40°C. Before use, the supernatant was KOH, pH 7.0, as follows; PIP z (200 J..lM) preincubated with
dialyzed against the assay buffer and further purified by truncated cofilin (CDN66) (80 J..lM) or synthetic peptides (1
centrifugation at 190,000 g for 2 h at 40°C. mM) for 15 min at 25°C, or untreated PIP z (200 J..lM) was
added to r-cofilin (4 J..lM), and the mixture was incubated
for 5 min at 25°C. Then, F-actin (4 J..lM) was added to the
Preparation ofsynthetic peptides and other proteins mixture and incubated for an additional 90 min at 25°C. The
mixture was centrifuged at 386,000 g for 20 min at 20°C.
Partial peptide sequences of chicken cofilin, D9EVIKVFND- The proteins in the supernatants and pellets were examined
MKVRKSST z5 , CZ6PEEIKKRKKAV36, andN3LYDATYET- by SDS-PAGE.
KESKKEDL99, were synthesized by standard solid phase
methods on a peptide synthesizer (Applied Biosystems 430A)
and purified by reverse phase HPLC using a gradient of Other methods
acetonitrile in 0.1 % TFA. The purity of the synthetic peptides
was confirmed by amino acid analyzer. SDS-PAGE was carried out using 13.5 or 15% polyacrylamide
Cofilin in embryonic chicken skeletal muscle was purified gel in a discontinuous Tris-glycine buffer system according
as described [15]. G-Actin was prepared from acetone-dried to Laemmli [28]. Protein concentrations were determined by
powder of rabbit skeletal muscle by the method of Spudich the method ofItzhaki and Gill [29] using BSA as a standard.
and Watt [26] and purified by gel filtration on a Sephadex PIP z was dissolved at a concentration of2 mg/ml in 0.1 mM
G-IOO column. DTT, 2 mM HEPES-KOH, pH 7.0 and stored at-80°C. The
solutions were quickly thawed and sonicated on ice for I min
before use.
Assay ofactin polymerization

G-actin alone or the mixture of G-actin and cofilin was Results

incubated in the buffer of neutral or alkaline pH for 15 min
at room temperature, and then actin polymerization was Preparation of recombinant cofilin
initiated by adding KCI and MgClz- The detailed conditions
for polymerization are described in the figure legends. Actin Recombinant cofilin (r-cofilin) containing the entire sequence
polymerization was measured as the increase in absorbance at of chicken cofilin was generated in an E. coli expression
237 nm (Am) using a Shimadzu UV-265 spectrophotometer system and purified by a combination of ammonium sulfate
according to Higashi and Oosawa [27]. fractionation and column chromatography with hydroxyl-
apatite and Sephadex G-75 (Fig. I). The protein obtained
migrated as a single band in SDS-PAGE (Fig. I, d) and gave
Chemical cross-linking a single spot in two dimensional electrophoresis, a combination
ofnonequilibrium pH gradient gel electrophoresis (NEpHGE)
Actin and r-cofilin or truncated cofilin were incubated in a and SDS-PAGE (data not shown).
solution containing 0.1 M KCI, 20 mM l-ethyl-3(3-dimethyl- We confirmed that r-cofilin was functionally active just as
aminopropyl) carbodiimide (EDC), 2 mM MgCl z, 0.1 mM the authentic cofilin from embryonic chicken skeletal muscle
DTT, 0.01% NaN 3, and 20 mM HEPES-KOH (pH 7.0), for [15] by measuring its effect on actin polymerization. Either
2 h at 25°C. For examining the inhibitory action of synthetic Hofilin (Fig. 2A) or the authentic protein (Fig. 2B) was added
peptides on actin-cofilin interaction, synthetic peptides, to G-actin at 1: 1 molar ratio (actin concentration, 5 mM),
r-cofilin and actin were mixed under the same conditions as incubated for 15 min, and actin polymerization was induced
above. The cross-linking reaction was terminated by adding by adding 0.1 M KCI and 2 mM MgCl z. The polymerization
an SDS-Iysis buffer containing 4% SDS, 12% glycerol, 2% process was monitored by measuring the increase in ab-
2-mercaptoethanol, 0.003% bromphenol blue (BPB), and sorbance at 237 nm.As shown in Fig. 2, the effects ofr-cofilin
50 mM Tris-HCI, pH 6.5. The mixtures were boiled for 5 min on actin polymerization were indistinguishable from those of
and subjected to SDS-PAGE. the authentic cofilin; at pH 8.0, both r-cofilin and the
136

a b c d I1A237 A
0090

94
67
a
43 0045 b

30 c

20 -

30 60 90 min
14 -

I1A237 B
Fig. I. Expression and purification of recombinant cofilin (r-cofilin). The 0090
fractions at each step of the purification of r-cofilin were analyzed by
SDS-PAGE with 13.5% polyacrylamide gel. (a) the whole lysate of E. coli
treated with IPTG to induce cofilin expression; (b) the precipitate with
ammonium sulfate at 60-80% saturation; (c) the cofilin fraction by
hydroxylapatite chromatography; (d) the final cofilin preparation obtained a
by gel filtration with Sephadex G-75. b

c
authentic cofilin prolonged the duration of the lag phase,
corresponding to a nucleation process, and kept the steady
state absorbance lower, whereas at pH 7.0, they accelerated
the early phase of polymerization and shortened the time
30 60 90 min
required for its termination. From these results, we conclude
that r-cofilin possesses basically the same biological activities
as the authentic chicken cofilin. Fig. 2. Effects ofr-cofilin (A) and authentic cofiJin from embryonic skeletal
muscle (B) on actin polymerization. Actin polymerization was initiated in
the absence (a) or presence (b, c) ofcofilin by adding KCI and MgCI, at
20°C; the polymerization process was monitored by UV absorption at
Preparation ofdeletion mutants ofcojilin 237 nm. The abscissa indicates time (min) after the addition of salts. The
final actin solution contained 4.8 11M actin, 4.8 11M cofilin (b, c), 0.1 M
In the cofilin sequence, functional domains responsible for KCI,2 mM MgCI,,20 11M CaCI" 0.1 mM and 10 mM PIPES at pH 7.0 (a,
b) or 10 mM Tris-HCI at pH 8.3 (c). Polymerization of actin without cofilin
actin-binding [22], nuclear localization [24, 25] and phos-
was not affected significantly by pH.
phorylation [20, 21] have been identified. Since the actin-binding
domain is on the C-terrninal side and the phosphorylation site
involved in the regulation of the activity is Ser 3 at the N-terrninal side, were soluble in a salt solution without urea
N-terrninus, we assumed that an additional sequence involved (see details in Materials and methods).
in actin-cofilin interaction might exist on the N-terrninal side
near the phosphorylation site. Therefore, we designed
several cofilin mutant molecules which were truncated at Binding ofcojilin deletion mutants to actin
the C-terrninal side by various lengths and the molecule with
deletion at the N-terrninal side as well, as shown in Fig. 3A. Binding ability of the cofilin deletion mutants to actin was
When they were produced in an E. coli expression system, examined by using a zero-length cross-linking reagent,
however, all of them were detected as insoluble forms in the l-ethyl-3-[3-(dimethylamino)propyll- carbodiin-dde (EDC).
lysates of the bacteria, so called 'inclusion bodies'. By The mixtures of G-actin and the cofilin mutants or wild-type
collecting the insoluble proteins by centrifugation and r-cofilin were incubated with 20 mM EDC for 3 h at 20°C in
dissolving them in 8 M urea in the presence ofDTT, we could 20 mM HEPES-KOH (pH 7.5), and then the mixtures were
obtain fairly pure proteins (Fig. 3B). After purification, the subjected to SDS-PAGE. As shown in Fig. 4, new protein
proteins except CD66C (see Fig. 3), a protein truncated at the bands of larger molecular weight (marked with asterisks)
 137

A 0;
-0
<: r::::
B
-g§ -g

_.
1ii~"S
LijLij aVi:<;: Lij
a b c d e
NlS .<lInlP,P....""i1g 94 -
r-cofilin
...,.
~
104-115122·123 ,eejM)
67 -
43-
CDN146 ~ -J
~ 30 -

CDNl16 ~

CDN110
_
• 20 -

14 -

CDN66 ~

CD66C

Fig. 3. Generation and purification of mutated cofilin molecules by using an E. coli expression system. (A) Schematic representation of the mutated cofilin
molecules generated. The restriction sites ofcofilin cDNA used for generating the deletionmutant molecules are shown on the top. The wild-type molecule
(r-cofilin) is shown as a reference. Closed and hatched boxes indicate the location of the actin/PIP,-binding and nuclear localization signal (NLS) sequences,
respectively, which were previously described [24, 25]. (8) SDS-PAGE patterns of purified deletion-mutants of cofilin. a - CDN66; b - CDN II 0; c - CDN 116;
d - CDN 146; e - r-cofilin.

appeared by SDS-PAGE after treating with EDC, and they conditions as described for Fig. 4 to examine the generation
were equivalent to the sum of the molecular weights of actin of the cross-linked product of cofilin and actin. As shown
and the respective mutants or wild-type r-cofilin. The bands in Fig. 6, cross-linking of cofilin and actin by EDC was
became more evident when cofilin or cofilin mutants was diminished remarkably in the presence of peptide a, D 9EVIK-
adde~ to G-actin at a higher concentration (Fig. 4, g-j). These VFNDMKVRKSST25, when added at a high concentration,
bands did not appear at al1 when EDC was applied to cofilin and weakly by peptide b, C26PEEIKKRKKAV36, including
alone or actin alone (data not shown). These results indicate a nuclear localization signal. A recombinant N-terminal
that al1 the cofilin mutants used as wel1 as wild-type r-cofilin fragment of cofilin, CDN66, scarcely interfered with the
form a one-to-one complex with G-actin molecules. It should cross-linking ofcofilin and actin, although the fragment itself
be stressed that even the partial sequence at the N-terminal was clearly cross-linked with actin. This was probably due
side (CDN66) without the previously described actin-binding to the fact that the amount ofCDN66 added was much smal1er
sequence could make a complex with actin. Therefore, it is than that of the synthetic oligopeptides; the molar ratio of
suggested that an additional sequence responsible for actin- CDN66 to actin (or cofilin) was 2: 1 to 5: 1, while the ratio of
binding exists in the N-terminal side of cofilin molecules. the synthetic peptides to actin (or cofilin) was 200: 1 to 500: 1.

Effects ofcofilin N-terminal peptides on actin-cofilin Effects ofthe N-terminal peptides on PIP2-dependent
interaction regulation ofactin-cofilin interaction

In order to further clarify the region at the N-terminal side It has been demonstrated that the previously described
of cofilin molecule which is involved in the interaction with actin-binding site W104APESAPLKSKM l1 5is also the site for
actin, oligopeptides of the cofilin sequence at the N-terminal PIP 2-binding, and therefore the interaction of cofilin with
side were synthesized as shown in Fig. 5. Cofilin was mixed actin is suppressed by inositol phosphate (PIP2) [19]. We were
with actin in the presence or absence ofthe synthetic peptides, interested in whether the cofilin sequence at the N-terminal
and then the mixture was treated with EDC under the same side which showed actin-binding ability was also sensitive
138

a b c d e f 9 h J A
NLS..- ...:.A.:.:,cllnlPlP2-blndlng

10.---'--1
94 -
67 - ~
43 - 26·36 104·115 122·126 166 (AA)

30 -
aD 9 25
20 -
b~
cD
26 36
14 -
63 99

B a:DEVIKVFNDMKVRKSST

b:C-PEEIKKRKKAV

Fig. 4. Cross-linking ofrecombinant cofilin (r-cofilin) and its truncated mutants

C:ALYDATYETKESKKEDL
to actin. G-actin (I 0 ~M) was cross-linked with r-cofilin (b-l 0 ~M) or truncated
mutants (c-f-IO ~M; g-j -20 ~M) in the presence of20 mM EDC at 25°C Fig. 5. Primary structures ofsynthetic peptides and their location in cofilin
for 2 h in a solution containing 0.1 M KC1, 2 mM MgCl" 0.1 mM DTT, sequence. (A) Schematic representation of the location of synthetic peptides
0.01 % NaN 3, and 20 mM HEPES-KOH (pH 7.0). The cross-linked products (a-e) in cofilin sequence. Closed and hatched boxes indicate the location
were analyzed by SDS-PAGE with 15% acrylamide gel. a - actin alone; b of the actin/ PIP 2-binding and nuclear localization signal (NLS) sequences,
- actin plus r-cofilin; c and g - actin plus CDN66; d and h - actin plus respectively. a - Asp9-Thr"; b- Pro'6-VaP6; c - Ala 83-Leu 99 . In peptide b,
CDNIIO; e and i-actin plus CDNI16; f and j - actin plus CDN146. an additional Cys residue is present at the N-terminus. (B) amino acid
Positions of the cross-linked product of actin-cofilin and the cross-linked sequences of the synthetic peptides.
products of actin-mutant cofilin are indicated by asterisks. M, standards
(x 10-3) are indicated at the left margin.

The other peptide at the central region of the cofilin sequence

to PIP 2. In order to assay the binding of cofilin to F-actin, (peptide c) had no effect on the PIP 2-dependent regulation
r-cofilin and F-actin were mixed at I: I molar ratio in the of cofilin-actin interaction (Fig. 7P, g).
presence or absence of PIP 2 and centrifuged at high speed.
Both cofilin and actin were precipitated stoichiometrically in
the absence of PIP 2, but in its presence only actin was Discussion
precipitated, while cofilin remained in the supernatant as
previously demonstrated [19], indicating that cofilin-actin It has been established that the sequence at the C-terminal
interaction was suppressed by PIP 2 (Fig. 7P, b and c). When side of cofilin/ADF, Trp l04-Met 11 S, is responsible for actin-
an excess amount of cofilin N-terminal fragment (CDN66) binding [23, 31]. Since the actin-binding ability of cofilin
or the peptides ofN-terminal region (peptide a or b, see Fig. is eliminated by the mutagenesis of Lys l12 and Lys 114, this
5) was added to the mixture, cofilin together with actin sequence is of primary importance for actin binding [32].
became precipitated by centrifugation irrespective of the However, an additional actin-binding site may be needed
presence ofPIP 2(Fig. 7P, d-f). These results strongly suggest to enable the rapid depolymerization ofF-actin at alkaline
that the cofilin fragment (CDN66) or the peptides of the N- pH and the complex regulation of actin-cofilin interaction
terminal region interacted with PIP2' so that the inhibitory in several different ways, namely control by pH, phos-
action of PIP2 on cofilin-actin interaction was removed. As phorylation, and PIP 2-binding. The sequence homologous
described above, both CDN66 and peptide a showed the to tropomyosin, D I22 AIKKKFI28, which is located near the
ability to interact with actin, but peptide b was less sensitive essential actin-binding site, Trp I04-Met 115. is also known to
to actin as judged by the weak inhibitory action on actin- have the ability to interact with actin [31]. In the present
cofilin binding but were sensitive to PIP2' We assume that study, we further examined the cofilin sequence in terms of
peptide b interacts with PIP2 and actin non-specifically actin-binding ability, and we found an additional sequence
because of the presence of a high concentration of basic having the ability to interact with actin in the fragments at
amino acids in the sequence, as previously suggested [30]. the N-terminal side near the phosphorylation site. The basis
 139

a b c d e f 9 h s p
abc d e 9 'a bed e 9

------~A

Fig. 7. Effects of cofilin peptides on PIP,-dependent regulation of actin-

Fig. 6. Effects of synthetic peptides of cofilin on cofilin-actin binding. cofilin interaction. Synthetic oligopeptides of cofilin (I mM) or truncated
Synthetic oligopeptides of cofilin (2 mM) or truncated recombinant cofilin recombinant cofilin without C-terminal side (MN66) (80 J.IM) were added
without C-terminal side (CON66) (20 J.IM) was added to r-cofilin. Then, to the mixture of actin (4 J.IM), cofilin (4 J.IM) and PIP, (200 J.IM) in a
cofilin (10 J.IM in lHl or 4 J.IM in f-j) was mixed with actin (I 0 J.IM in a--e or solution containing 40 mM KCl, 0.04 mM CaCl" 0.1 mM DTT, 0.01%
4 J.IM in f-j) in a solution containing 20 mM EOC, 0.1 M KCI, 2 mM NaN, and 20 mM HEPES-KOH (pH 7.0). The mixture was incubated for
MgCl" 0.1 mM OTT, 0.01% NaN" and 20 mM HEPES-KOH (pH 7.0). 1.5 h at 25°C, and then centrifuged at 386,000 x g for 20 min. The resultant
The mixture was incubated for 2 h at 25°C. Generation of the cross-linked supernatants (S) and precipitates (P) were subjected to SOS-PAGE. a-
product of actin and cofilin was analyzed by SOS-PAGE with 15% actin alone; b -actin + cofilin; c -actin + cofilin + PIP,; d-g - the same as
polyacrylamide gel. a and f - actin alone; band g - actin + r-cofilin; c and c except CON66 (d), peptide a (e), peptide b (f) or peptide c (g) was included,
h - actin + r-cofilin + MN66; d and i-actin + r-cofilin + peptide a (Fig. 5); respectively. Actin and cofilin bands were marked by A and C, respectively.
e and j - actin + rcofilin + peptide b (Fig. 5). The bands corresponding to
actin-cofilin complex, actin and cofilin were marked by A-C, A and C,
respectively.
apart from the previously described actinlphosphoinositides-
binding site, Trp'04-Met"5, but close to the actin-binding
for our conclusion that the sequence at the N-terminal side sequence at the N-terminal side described here. In the recently
has actin-binding-ability is that (1) the N-terminal fragments determined tertiary structure of destrin (ADF), the phos-
can be cross-linked to actin by a chemical cross-linker phorylation site is spatially close to the actin-binding site
(EDC) and (2) the fragments compete with cofilin in Trplo4-Met '15 [35]. Therefore, phosphorylation ofSer3 could
actin-binding. In the N-terminal region of cofilin having modulate the activity of both actin-interacting sequences at
actin-binding ability, we can detect a sequence where the N-terminal and C-terminal regions of cofilin molecule.
several Lys residues are concentrated, namely amino acid We determined that the N-terminal actin-binding site is
residues #9---#25 (Fig. 8). A stretch of Lys (or Arg) residues sensitive to PIP 2 just as the previously described C-terminal
in this sequence is highly conserved among various actin- actin-binding site, and it is highly likely that the activity of
binding proteins (Fig. 8). Since Lys residues are involved both actin-binding sites is similarly controlled by phospho-
in actin-binding in many cases, it seems likely that the Lys inositides.
residues in this sequence may also be responsible for We previously demonstrated that cofilin exhibits drastic
actin-binding. The sequence of cofilin at this region could effects on the actin cytoskeleton when introduced into
also bind PIP 2' since the phosphoinositide-binding site of cultured cells in large amounts; actin filaments are disrupted
gelsolin is quite similar to this cofilin sequence [33]. and actin-cofilin rods are formed in the cytoplasm [17]. The
It is a matter of particular interest how the two actin- N-terminal fragment with actin-binding ability (CDN-66)
binding sites are involved in the complex regulation of actin could not disassemble cytoplasmic actin filaments when
assembly by cofilin. In the case of the gelsolin family, it is introduced into cultured cells, but interfered with the effect of
widely accepted that a single actin-binding domain cannot wild-type cofilin on the actin cytoskeleton in a dose-dependent
sever actin filaments and that cooperativity between the manner when introduced into the cytoplasm together with the
F-actin-binding and G-actin-binding domains is essential wild-type cofilin (unpublished data). Thus, the N-terminal
[34]. Cofi lin is known to sever actin filaments at alkaline pH, actin-interacting sequence is suggested to be functional in the
and the two actin-binding sites may be needed for this. In the cytoplasm.
primary structure of cofilin, the phosphorylation site which is
deeply involved in the regulation of its activity is considerably
140

mouse thymosin ~4 7AEIE~---DKS~ET22

rat thymosin ~lO 7GEIASF---DKA!r&ET22

chicken cofilin 9 DEVI SST 25

mouse m-cofilin 9 DEVI SST 25
mouse nm-cofilin 9DGVI F-ND- SST 25
chicken ADF 9DEVC" IF-YD- ST 25
porcine ADF 9DEVC" IF-YD- ST 25
yeast COFI 1 0DESLTAF-ND-L~'UL~YK2 6
amoeba actophorin 24
9DcvlF--NE-L HR
lily ADF 13 ECK --ME-L FR 28
starfish depactin 12 KEE " KMDQS PWM 30

human pl-gelsolin 158 LGYF SG----L

human cp-gelsolin 184 VVVQ"
human CapG 134AAI
chicken villin 134 yNV

consensus ----*-------*x**---
Fig. 8. Comparison of the amino acid sequence at the N-terminal side ofcofilin responsible for actin-binding with other G-actin-binding proteins. Residues
7-22 of mouse thymosin 134 [36] and rat thymosin 131 0 [37], residues 9-25 of chicken cofilin [3], mouse muscle-type(m) cofilin [38], mouse non-muscle-type
(nm) cofilin [39], chicken ADF [3, 4] and porcine ADF [40], residues 10-26 of yeast COFI [12], residues 9-24 of amoeba actophorin [41], residues 13-28
oflily ADF [42], residues 12-30 of starfish depactin [7], residues 158-172 of human plasma (pi) gelsolin [43], residues 184-199 of human cytoplasmic (ep)
gelsolin [43], residues 134-149 of human CapG [44], and residues 134-149 of chicken villin [45] arc aligned. Putative consensus sequence is shown at the
bottom.

Acknowledgments 2. Bamburg JR, Harris HE, Weeds AC: Partial purification and char-
acterization of an actin depolymerizing factor. FEBS Lett 121: 178-
181,1980
One of the authors, TO, is greatly indebted to Dr. Setsuro 3. Abe H, Endo T, Yamamoto K, Obi nata T: Sequence of cDNAs encoding
Ebashi, Professor Emeritus of the University ofTokyo and of actin depolymerizing factor and cofilin of embryonic chicken skeletal
National Institute for Physiological Sciences, for his continuing muscle: Two functionally distinct actin-regulatory proteins exhibit high
encouragement and stimulating advise through the author's structural homology. Biochemistry 29: 7420-7425, 1990
studies on muscle cell differentiation and comparative bio- 4. Adams ME, Minamide LS, Deuster GL, Bamburg JR: Nucleotide
sequence and expression of a cDNA encoding chick brain actin
chemistry ofmuscle regulatory proteins. 1t is his great pleasure depolymerizing factor. Biochemistry 29: 7413-7419,1990
to dedicate this paper to Professor Setsuro Ebashi. 5. McKim KS, Matheson C, Marra MA, Wakarchuk MF, Baillie DL:
This research was supported in part by grants from the The Caenorhabditis elegans unc-60 gene encodes proteins homologous
Ministry of Education, Science and Culture, the National to a family of actin-binding proteins Mol Gen Genet 242: 346-357,
Center ofNeurology and Psychiatry (NCPN) of the Ministry 1994
6. Hayden SM, Miller PS, Brauweiler A, Bamburg JR: Analysis of the
of Health and Welfare of japan, and the Tsuchiya Culture interactions of actin depolymerizing factor with G- and F-actin.
Foundation, Japan. Biochemistry 32: 9994-10004, 1993
7. Takagi T, Konishi K, Mabuchi I: Amino acid sequence of starfish
oocyte depactin. J Bioi Chern 263: 3097-3102,1988
References 8. Sun H-Q, Kwiatkowska K, Yin HL: Actin monomer binding proteins.
CurrOpinCell Bioi 7: 102-110,1995
9. Gunsalus KC, Bonaccorsi S, Williams E, Vemi F, Gatti M, Goldberg
1. Nishida E, Maekawa S, Sakai H: Cofilin, a protein in porcine brain ML: Mutation in twinster, a Drosophila gene encoding a cofilin/ADF
that binds to actin filaments and inhibits their interactions with myosin homologue, results in defects in centrosome migration and cytokinesis.
and tropomyosin. Biochemistry 23: 5307-5313, 1984 J Cell Bioll31: 1243-1259, 1995
 141

10. Matsuzaki F, Matsumoto F, Yahara I, Yonezawa N, Nishida E, Sakai 28. Laemmli UK: Cleavage of structural proteins during the assembly of
H: Cloning and characterization of porcine brain cofilin cDNA. J BioI the head of bacteriophage T4. Nature 227: 680-685, 1970
Chern 263: 11564-11568,1988 29. Itzaki RF, Gill OM: A micro-biuret method for estimating proteins.
I I. Iida K, Moriyama K, Matsumoto S, Kawasaki H, Nishida E, Sakai H: Analyt Biochem 9: 401--410,1964
Isolation of a yeast essential gene, COF I, that encodes a homologue 30. Janmey PA, Lamb J, Allen PC, Matsudaira PT: Phosphoinisitide-
of mammalian cofilin, a low-M, actin binding and depolymerizing. binding peptides derived from the sequences of gelsolin and villin. J
Gene (Arnst) 124: 115-120,1993 Bioi Chern 267: 11818-11823, 1992
12. Moon AL, Janmey P, Louie KA, Drubin 0: Cofilin is an essential 31. Yonezawa N, Nishida E, Ohba M, Seki M, Kumagai H, Sakai H: An
component of the yeast cortical cytoskeleton. J Cell BioI 120: 421- actin-interacting peptide in the cofilin sequence. Eur J Biochem 183
435,1993 235-238, 1989
13. Abe H, Obinata T, Minamide LS, Bamburg JR: Xenopus laevis actin- 32. Moriyama K Yonezawa N, Sakai H, Yahara I, Nishida E: Mutational
depolymerizing factor/cofilin: A phosphorylation-regulated protein analysis of an actin-binding site of cofilin and characterization of
essential for development. J Cell Bioi 132: 871-885, 1996 chimeric proteins between cofilin and destrin. J Bioi Chern 267: 7240-
14. Nagaoka R, Abe H, Kusano K, Obinata T: Concentration of cofilin, a 7244,1992
small actin-binding protein, at the cleavage furrow during cytokinesis. 33. Yu F-X, Sun HQ, Janmey PA, Yin HL: Identification of a poly-
Cell Moti1 Cytoskeleton 30: 1-7,1995 phosphoinositide-binding sequence in an actin monomer-binding
15. Abe H, Ohshima S, Obinata T: A cofilin-like protein is involved in the domain of gelsolin. J Bioi Chern 267: 14616-1462 I, 1992
regulation ofactin assembly in developing skeletal muscle. J Biochem 34. Matsudaira P, Janmey P: Pieces in the actin-severing protein puzzle.
(Tokyo) 106: 696-702,1989 Cell 54: 139-140, 1988
16. Obinata T: Contractile proteins and myofibrillogenesis. Int Rev Cytol 35. Hatanaka H, Ogura K, Moriyama K, Ichikawa S, Yahara I, Inagaki F:
143: 153-189, 1993 Tertiary structure of destrin and structural similarity between two
17. Nagaoka R, Kusano K Abe H, Obinata T: Effects of cofilin on actin actin-regulating protein families. Cell 85: 1047-1055,1996
filamentous structures in cultured muscle cells: Intracellular regulation 36. Rudin CM, Englaer P, Storb U: Differential splicing of thymosin ~4
of cofilin action. J Cell Sci 108: 581-593, 1995 mRNA. J Immunol 144: 4857--4862,1990
18. Yonezawa N, Nishida E, Sakai H: pH control of actin polymerization 37. Erickson-Viitanen S, Ruggieri S, Natalini P, Horcker BL: Distribution
ofcofilin. J Bioi Chern 260: 14410-14412, 1985 ofthymosin ~4 in vertebrate classes. Arch Biochem Biophys 221 : 570-
19. Yonezawa N, Nishida E, !ida K, Yahara I, Sakai H: Inhibition of the 576, 1983
interactions of cofilin, destrin, and deoxyribonuclease I with actin by 38. Ono S, Minami N,Abe H, Obi nata T: Characterization ofa novel cofilin
phosphoinositides. J Bioi Chern 265: 8382-8386, 1990 isoform that is predominantly expressed in mammalian skeletal muscle.
20. Moriyama K, !ida K, Yahara I: Phosphorylation of Ser-3 of cofilin J Bioi Chern 269: 15280-15286, 1994
regulates its essential function on actin. Genes to Cells I: 73-86, 1996 39. Moriyama K, Matsumoto S, Nishida E, Sakai H, Yahara I: Nucleotide
21. Agnew BJ, Minamide LS, Bamburg JR: Reactivation of phos- sequence of mouse cofilin cDNA. Nucl Acids Res 18: 3053, 1990
phorylated actin depolymerizing factor and identification of the 40. Moriyama K, Nishida E, Yonezawa N, Sakai H, Matsumoto S, lida K,
regulatory site. J Bioi Chern 270: 17582-17587, 1995 I: Destrin, a mammalian actin-depolymerizing protein, is closely
22. Yonezawa N, Nishida E, lida K, Kumagai H, Yahara I, Sakai H: related to cofilin. Cloning and expression of porcine brain destrin
Inhibition of actin polymerization by a synthetic dodecapeptide cDNA. J Bioi Chern 265: 5771-5773,1990
patterned on the sequence around the actin-binding site of cofilin. J 41. Quirk S, Maciver SK, Ampe C, Doberstin SK, Kaise DA, VanDamme
Bioi Chern 266: 10485-10489, 1991 J, Vadekerckhove JS, Pollard TD: Primary structure of and studies
23. Yonezawa N, Homma Y, Yahara I, Sakai H, Nishida E:A short sequence on Acanthamoeba actophorin. Biochemistry 32: 8525-8533,1993
responsible for both phosphoinositide binding and actin binding 42. Kim S-R, Kim Y, An G: Molecular cloning and characterization of
activities of cofilin. J Bioi Chern 266: 17218-17221, 1991 antherpreferential cDNA encoding a putative actin-depolymerizing
24. Abe H, Nagaoka R, Obinata T: Cytoplasmic localization and nuclear factor. Plat Mol Bioi 21: 39--45, 1993
transport of cofilin in cultured myotubes. Exp Cell Res 206: 1-10, 1993 43. Kwiatowski OJ, Stossel TP, Orkin SH, Mole JE, Colten HR, Yin
25. Iida K, Matsumoto S, Yahara I: The KKRKK sequence is involved in HL: Plasma and cytoplasmic gelsolins are encoded by a single gene
heat shock-induced nuclear translocation of the 18-kDa actin-binding and contain a duplicated actin-binding domain. Nature 323: 455-
protein, cofilin. Cell Struct Funct 17: 39--46, 1992 488, 1986
26. Spudich JA, Watt S: The regulation of rabbit skeletal muscle 44. Dabiri GA, Young CL, Rosenbloom J, Southwick FS: Molecular
contraction. I. Biochemical studies of the tropomyosin-troponin cloning ofhuman macrophage capping protein cDNA.A unique member
complex with actin and the proteolytic fragment of myosin. J Bioi of the gelsolinlvillin family expressed primarily in macrophage. J Bioi
Chern 246: 4866--4871, 1971 Chern 267: 16545-16552,1992
27. Higashi S, Oosawa F: Conformational changes associated with 45. Bazari WL, Matsudaira P, Wallek M, Smeal. T, Jakes R, Asmed Y:
polymerization and nucleotide binding in actin binding. J Mol Bioi Villin sequence and peptide map identify six homologous domains.
112: 843-865,1965 Proc Natl Acad Sci USA 85: 4986--4990, 1988
Molecular and Cellular Biochemistry 190: 143-151, 1999.
© 1999 Kluwer Academic Publishers.

Creatine kinase, cell membrane and Duchenne

muscular dystrophy*

Eijiro Ozawa, Yasuko Hagiwara and Mikiharu Yoshida

National Institute ofNeuroscience, NCNp, 4-1-1 Ogawahigashi-cho Kodaira, Tokyo, Japan

Abstract
In 1958 Professor Setsuro Ebashi found that serum creatine kinase activity is increased in patients suffering from various muscular
dystrophies, especially Duchenne muscular dystrophy (DMD). He and others proposed that creatine kinase passes through the
cell membrane as it is released from DMD muscle fibers.
Since then, it has been found that dystrophin and dystrophin-associated proteins are connected to several other components,
including the basal lamina and subsarcolemmal cytoskeletal networks on the cell membrane, while dystrophin anchors these
dystrophin-associated proteins to the actin filaments inside the muscle cell. In DMD muscle, dystrophin has been found to be
absent and dystroglycans and sarcoglycans decreased. However, how creatine kinase molecules can pass through the DMD
muscle cell membrane still remains unanswered.
On the basis of recent findings on the structure ofthe protein layers which sandwich the lipid bilayer of muscle cell membranes,
this essay stresses the importance ofthese lipid bilayers in protecting creatine kinase release from protoplasma in normal muscle.
It further indicates the possibility that the absence of dystrophin in DMD muscle during muscle contraction may result in temporal
damage to the lipid bilayer. (Mol Cell Biochem 190: 143-151, 1999)

Key words: dystrophin, dystroglycans, sarcoglycans

Introduction the national economy of defeated Japan was too weak to keep
laboratories well equipped. Indeed, in order to obtain ATP,
When dystrophin [1-6], a component of the subsarcolemmal one of the substrates of creatine kinase, Professor Ebashi and
cytoskeletal network, is absent from and irregularly present his colleagues had to purify ATP from muscle by their own
on the subsarcolemmal undercoat, Duchenne muscular hands. This work was summarized in a short note published
dystrophy (DMD) and its milder form, Becker muscular in the Journal of Biochemistry, Tokyo [10], which was not
dystrophy (BMD) [7, 8] occur, respectively. Therefore, widely distributed at that time. This discovery, however, was
DMD and BMD together are termed dystrophinopathy. soon confirmed by many researchers all over the world [12,
From the beginning of studies on dystrophinopathy in the 13]. Measurement of serum creatine kinase activity was then
middle of the last century, they were mostly restricted to widely adopted as the most important biochemical test for
semiology (classification of the diseases), classical human muscular dystrophy, including carrier determination [14], and
genetics and pathology [9], until dystrophin gene located at has been still routinely used as a laboratory test.
Xp21 was cloned in 1987 by Kunkel's group [1,3]. In 1959, The increased serum creatine kinase levels in DMD have
however, Professor Ebashi (Ebashi et al. [10] and Okinaka been attributed to increased creatine kinase released from
et al. [11]) found that creatine kinase activity is increased in skeletal muscle fibers judging from the isoform patterns of
the serum of muscular dystrophy patients, especially DMD the enzyme [15-17]. The increase of creatine kinase is
patients. This discovery was made under difficult conditions speculated to result from leakage ofthe enzyme from diseased
at a time relatively close to the end of World War II, when muscle fibers. However, the mechanism of the enzyme

*Dedicated to Professor S. Ebashi

Address for offprints: E. Ozawa, National Institute of Neuroscience, NCNP, 4-1-1 Ogawahigashi-cho Kodaira, Tokyo 187, Japan
144

leakage is not known. Defects of DMD muscle membrane contrast, the outer layer can slide freely against the lipid
were thoroughly reviewed by Rowland in 1984 [18]. bilayer and this may undermine its protective function ifthere
Therefore, it is not necessary to review here the classical is no fixatives between the inner and outer layers. Dystrophin
'membrane theory' of DMD. and dystrophin-associated proteins (DAPs) connecting the
Since the cloning ofdystrophin eDNA in 1987, its cell and outer and inner layer may serve as props [20] which prevent
molecular biology of DMD muscle cells has been well the sliding.
developed [19-21]. The structure of subsarcolemmal cyto-
skeletal network including dystrophin and the relationship
between the network and the basal lamina, the structure Dystrophin
formed by dystrophin and dystrophin-associated proteins
(DAPs), have been fairly well elucidated in normal muscle.
The dystrophin molecule has a molecular mass of 427 kDa,
However, it is still too early for us to answer the question
is long and slender and is composed offour domains, namely,
regarding how creatine kinase is released from muscle fibers,
N-terminal actin-binding, rod, cysteine-rich and C-terminal
which was raised by Professor Ebashi about 40 years ago. The
domains [3]. Immunohistochemically, the cell membrane is
purpose of this essay is to draw the researcher's attention to
stained for dystrophin in normal muscle but not in DMD
the studies which need to be carried out to solve this problem.
muscle [5, 23, 24]. This suggests that dystrophin is absent
from DMD muscle cell membrane. Immunohistochemical
electron microscopy revealed that dystrophin is present on
Structure of muscle fiber the protoplasmic surface of the cell membrane [6, 25].
Dystrophin possesses neither hydrophobic transmembrane
A muscle fiber is a long slender cell, 50-100 11m in diameter
domains [3] nor sugar chains [26]. Thus, dystrophin is a
and several ten em in length in the case of the longest fibers
peripheral membrane protein which is present close to but not
[22]. It contains numerous myofibrils covered by the cell
integrated in the lipid bilayer. Absence of dystrophin from
membrane or sarcolemma. The muscle cell membrane is
DMD muscle cell membrane implies that some structural
composed of a lipid bilayer in which the lipid molecules are
abnormalities are present at the subsarcolemmal undercoat
arranged putting their hydrophobic group faces inside and
in DMD muscle, which are likely to be pathologically
hydrophilic faces outside of the membrane. The lipid bilayer
important [20].
serves as an impermeable barrier to the flow of most large
polar hydrophilic molecules such as water-soluble enzymes
including creatine kinase. However, the lipid bilayer may
not stand up to severe mechanical agitation. The cell Dystrophin-associated proteins
membrane of muscle fibers is subjected to strong tension
on contraction, at that time the lipid bilayer may be sus- The dystrophin fraction of muscle cell membrane solubilized
ceptible to micro-trauma, unless it is mechanically protected with digitonin, contains a number ofproteins. About 10 proteins
by some additional layers. in this fraction were shown to be associated with dystrophin
Extracellularly, the muscle fiber is covered with the basal following further fractionation by sucrose density gradient
lamina (outer layer), which is a thick and mechanically tough centrifugation or gel filtration chromatography [27, 28].
network composed of collagen, laminin, heparan sulfate and The DAPs thus identified are listed in Table I. Since these
so forth. Intracellularly, beneath the lipid bilayer there are proteins were identified by Campbell's group [27] and
cytoskeletal networks (inner layer) composed of actin, Yoshida and Ozawa [28] independently and almost sim-
vimentin and so forth. These extracellular outer and intra- ultaneously in 1990, these two groups at first named them
cellular inner layers sandwich the lipid bilayer and are differently. The lists of DAPs from the two groups were
composed of networks of long filaments, and most soluble similar but not identical. In the list of Yoshida and Ozawa [28],
molecules pass freely through them. These outer and inner I56DAG (a-dystroglycan) was not present, whereas in the list
layers do not serve as barriers against creatine kinase leakage presented by Campbell's group [27],A3b (~-sarcoglycan) and
from the cell. Therefore, it must be the lipid bilayer which AO (dystrobrevin) were missing. Thus, when the lists
prevents the creatine kinase molecules from leaking from the presented by the two groups are combined, we have a
muscle fiber. complete ofDAPs identified except for8-sarcoglycan [29-
One of the roles of these inner and outer layers must be to 31], which was discovered in 1996. It is now known that some
protect the lipid bilayer from mechanical agitation. Ofthe two DAPs such as syntrophins exist as a number of isoforms
layers, the outer one is very important as a protector against distinguishable by 2-D PAGE [32].
a force from outside the muscle fiber. The inner layer is fixed In 1994, Yoshida et al. [33] fractionated DAPs into 3
to other intracellular structures and does not move freely. In complexes by treating dystrophin-DAP complex with
 145

Table I. Dystrophin-associated proteins [34]. Thus, ~-dystroglycan and ~-sarcoglycan were sometimes
confused. Their cDNAs have been cloned, and their genes have
DAP gene locus molecular mass former names
been mapped to different loci [34, 37, 38]. These proteins are
a-dystroglycan 3p21 156 kD 156DAG known to belong to different complexes [33] and do not
~-dystroglycan 3p21 43 43DAG,A3a cross-react with each other immunologically [33].
a-sarcoglycan l7q2l A2, 50DAG, adhalin
The sarcoglycan complex with special reference to
50
~-sarcoglycan 4ql2 43 A3b, (43DAG) Duchenne-like autosomal recessive muscular dystrophy has
y-sarcoglycan 13ql2 35 A4,35DAG been extensively reviewed elsewhere [40] and is not reviewed
8-sarcoglycan 5q33 35 here. In brief, the sarcoglycan complex is almost specific to
sarcospan 12pII.2 25 A5,25DAP striated muscle [41,42]. Loss of the sarcoglycan complex
a-syntrophin 20q 11.2 aAI,60DAG
from the sarcolemma causes sarcoglycanopathy, a group of
60
~,-syntrophin 8q23-24 60 ~Al,60DAG autosomal recessive muscular dystrophies, whose phenotype
~2-syntrophin 16q22-23 60 ~Al,60DAG is very similar to that ofdystrophinopathy. Each sarcoglycan
gene is a responsible gene of sarcoglycanopathy [43].
dystrobrevin 90, etc* AO, 87K protein Three syntrophins, a, ~I and ~2 [32,44--49], have been
*There are many isoforms. identified and their cDNAs have been cloned. These are
non-glycosylated proteins located at the inner surface of
muscle cell membrane. Although syntrophins are expressed
n-octyl-~-D-glucoside and termed these complexes the in diverse tissues [49], a-syntrophin is mainly expressed in
dystroglycan complex after Ibraghimov-Beskrovnaya et al. striated muscle. Dystrobrevin, formerly caliedAO [28] or 87
[34], sarcoglycan complex and others. Dystroglycan and K protein [50, 51], is also a non-glycosylated protein [26, 32]
sarcoglycan complexes are glycoproteins and together are and its cDNA has been cloned [51,52]. Complicated isoforms
sometimes called the glycoprotein complex (Table I). of dystrobrevin are due to alternative splicing [53, 54].
The dystroglycan complex is composed of a- and ~­ Syntrophins [55-58] and dystrobrevin [55] bind to the second
subunits as shown by Ibraghimov-Beskrovnaya et al. [34] in half of the C-terminal domain of dystrophin. At present, the
1992 when they cloned dystroglycan cDNA encoding both roles played by syntrophins and dystrobrevin in muscular
subunits in a single gene. This mRNA encodes 4,200 dystrophy are not known.
nucleotides containing an open reading frame of 2,685
nucleotides which encodes 895 amino acids residues with
a calculated molecular weight of 97,029. This 97-kDa Fixation mode of dystrophin
polypeptide has a signal sequence at the N-terminal and a
hydrophobic amino acid sequence which is considered to be As soon as the presence of dystrophin at the protoplasmic
a transmembranous domain at about one-fifth from the surface of the cell membrane was determined in 1988, when
C-terminal. This polypeptide is cleaved into 2 parts most existence ofDAPs was not yet known, Hoffman and Kunkel
likely after glycosylation, but the exact cleavage site has not [4] proposed a model in which dystrophin binds to unknown
been determined. a- and ~-dystroglycans are glycosylated, membrane-integrated protein(s) (MIP: This abbreviation is
but a-dystroglycan is particularly heavily decorated with referred to only the proteins in the Hoffman-Kunkel model).
sugar chains. a- and ~-dystroglycans bind to each other and According to their model, dystrophin binds to actin filament
form a complex after being separated into two molecules. at its N-terminus and to the MIP at its C-terminus. They further
When Yoshida et al. [33] identified the sarcoglycan hypothesized that dystrophin is present as an antiparallel
complex, it was considered to be composed of3 components, homodimer because of the structural similarity of the rod
but it is now known to be composed of 4 components, a-, domain to the corresponding domain ofa-actinin which forms
~-, y- and 8-sarcoglycans [29-31]. Their cDNAs have been an antiparallel homodimer. According to their model, the
cloned and each protein has been found to have a small C-terminus of dystrophin should be present at both ends of the
intracellular and a large extracellular domain interrupted by dimer dystrophin which binds to MIP. If dystrophin is present
a transmembranous domain. The genes encoding these in antiparallel dimer, dystrophin is considered to be present
proteins have been mapped on the different chromosomes close to and in parallel to the surface of the lipid bilayer. In
[29, 35-39]. All sarcoglycans are glycosylated and form a 1988, however, questions regarding (1) whether MIP exists,
complex. ~-dystroglycan and ~-sarcoglycan are different (2) what membrane proteins represent MIP (3) to what locus
molecules in spite of their close similarity in molecular of dystrophin MIP binds and (4) whether dystrophin is
weight. This must be emphasized because their molecular present as an antiparallel homodimer were not answered.
weights are very similar (43,000) and it was reported that Campbell's group mainly studied the first question and
anti-~-dystroglycan antibody reacted with ~-sarcoglycan mainly Ozawa's group and also other groups answered the
146

third question at almost the same period. The second problem column, short fragments of dystrophin were collected with
was answered by both groups and by other groups. Some dystroglycan and sarcoglycan complexes. These dystrophin
other questions remain unanswered. fragments were not mere contaminants in the fraction, but
were bound to these glycoproteins. By use of antibodies
against various epitopes on dystrophin, a common region
Are DAPs MIP? included in these dystrophin fragments was determined to be
a region spanning the cysteine-rich domain and the first half
In 1991 when the nature of DAPs was not well understood, of the C-terminal domain of dystrophin, suggesting that the
Ervasti et al. [59] and Ervasti and Campbell [26] immuno- glycoproteins bind to these domains. The glycoprotein-
histochemically showed that some DAPs are present at the binding locus corresponds well to a region whose absence
muscle cell membrane, similar to dystrophin. This was results in severe DMD phenotypes, which was determined
supported by the biochemical finding that the dystrophin- by gene analysis [62]. This correspondence suggested that
glycoprotein complex is highly enriched in isolated muscle loss of interaction between dystrophin and glycoproteins
cell membrane [27, 28, 59, 60]. results in loss of dystrophin and development of severe
Campbell's group [26, 60] showed that upon treatment DMD symptoms. This further explains why truncation of
with alkaline solution (pH 11), dystrophin is released from dystrophin due to premature stop codon in the gene in the
the membrane fraction, whereas glycoproteins are not. This presence of an out-of-frame deletion or duplication results
finding together with the previous electron microscopic in the DMD phenotype [63]. The presence of a premature
findings of Kunkel's group [6] on the location of dystrophin stop codon in hot spots [3, 64] localized in regions over-
in myofibers suggests that dystrophin is not a membrane- lapping the first 20 exons and encompassing exons 44-53
integrated protein but a membrane peripheral protein as of the dystrophin gene results in loss of the glycoprotein-
postulated in the Hoffman-Kunkel model. In a more alkaline binding locus and inability of dystrophin to fix to the
solution (pH 12), a-dystroglycan was released from the subsarcolemmal undercoat.
membrane fraction. Because a-dystroglycan binds to In 1994, Suzuki et al. [65] showed, using an overlay
laminin this result suggested that a-dystroglycan is an binding assay method, that l3-dystroglycan binds to the
extracellular protein not integrated into the muscle cell cysteine-rich domain of dystrophin and that the binding is
membrane. Even after treatment with a pH 12 solution, a- strengthened when the first half of the C-terminal domain of
and y-sarcoglycans and l3-dystroglycan remain in the dystrophin is connected in tandem to the cysteine-rich
membrane. These proteins and 25DAP were labeled with domain. Later, lung et al. [66] confirmed this finding. They
[125I] 3-(trifluoromethyl)-3-(m-iodophenyl) diazirine which further localized the dystrophin-binding domain of 13-
is specifically incorporated into the hydrophobic segments dystroglycan to proline-rich 15 amino acid residues of the C-
of the proteins, suggesting that they are transmembranous. terminal region starting from amino acid number 880.
Therefore, a- and y-sarcoglycans and l3-dystroglycan were
considered to be membrane-integrated. The presence of a
transmembranous domain in these proteins was later confirmed Dystrophin axis connecting the outer
by analysis of their amino acid sequences deduced from their and inner layers
cDNA sequences [30, 31, 35].
Together with the findings that these transmembranous Based on the above and other evidence, we can now delineate
proteins associate with dystrophin, it was assumed that the axis connecting the outer and inner layers, respectively
membrane-integrated dystrophin-associated glycoproteins (Fig. 1).
serve as MIP. However, the problems which proteins directly Laminin in the outer layer binds to a-dystroglycan [34].
interact with dystrophin and which loci of dystrophin bind The laminin molecule is composed of a, 13, and y subunits.
to these glycoproteins were not answered. In the muscle basal lamina, except the neuromuscular
junctional region, az131YI laminin is present (a z was formerly
called merosin) [67, 68]. In cells other than muscle fibers,
GPC-binding sites on dystrophin the basal lamina contains a l in place of a z. In the presence
of Ca ions, a z subunit binds to the sugar chain of a-
In 1992, Suzuki et al. [61] treated the dystrophin-DAP dystroglycan [21]. Extracellular a-dystroglycan binds to
complex with calpain, and found that glycoproteins remained transmembranous l3-dystroglycan which, in turn, binds to
undigested as determined by electromobility on SDS-PAGE dystrophin mainly in the cysteine-rich domain [33, 61, 65].
gels, but dystrophin was digested. When glycoproteins were The details of the topographical relationship between
collected using wheat germ agglutinin (WGA) affinity sarcoglycan complex and the dystrophin-dystroglycan
 147

laminin

basal lamina

Z'l-t ~~12-t ~ H~12-~ ~ 1l-~~t ~

lipid bilayer

m Sarcoglycan complex
_ Oystroglycan complex

~ Syntrophin complex
Dystrobrevin

Fig. I. Molecular architecture of dystrophin and OAPs on cell membrane [97].

complex are not known. However, it has been reported that Muscular dystrophy and the protecting
the concentration of dystrophin is decreased in the heart
muscle of the dystrophic hamster which lacks the sarco-
system
glycan complex suggesting a role for the sarcoglycan
complex in fixing dystrophin [69,70). However, there is no The basal lamina is 'worn off' in merosin-free congenital
direct evidence regarding whether sarcoglycan complex muscular dystrophy (MCMD) muscle fibers [71] in which the
serves to bind dystrophin tightly. merosin (0. 2 subunit of laminin) gene is defective [72, 73].
The region around the C-terminus of dystrophin is located MCMD patients are scarcely mobile, and have a mediate
very close to the cell membrane, because this region binds increase in the serum creatine kinase level. These findings
to membrane proteins. There has been discussion that suggest that destruction of the outer layer of the protecting
dystrophin forms an antiparallel homodimer [4]. If this is the system of the lipid bilayer may be important in giving rise to
case, the dimer has a C-terminal region at both ends. Thus, dystrophic changes.
dystrophin molecule is expected to be located very closely to In DMD muscle fibers, the outer layer remains almost
the membrane. In immunohistochemical electron microscopic normal as determined morphologically. Although dys-
analysis, it has been shown that gold particles bound to the troglycans is decreased in DMD muscle fibers but distinctly
antibody against the C-terminal of dystrophin are scattered preserved immunohistochemically [74], dystrophin is absent.
close to the cell membrane whereas ones bound to another However, the expression of utrophin [75], an autosomal
antibody against the middle of the rod domain are arranged homologue of dystrophin, is upregulated in DMD muscle [76,
deep inside the membrane. These findings suggest that 77). Utrophin is present in only small amounts at the sub-
dystrophin is not an antiparallel homodimer but a monomer sarcolemmal undercoat in normal muscle [78] except for at
or a parallel oligomer. Dystrophin is thought to be present the neuromuscular junction where utrophin is present in very
rather perpendicular to the cell membrane (Saito S., Nonaka large amounts [79]. The glycoprotein-binding and actin-
I and Ozawa E, unpublished data). binding regions of utrophin are highly homologous to the
The N-terminal domain of dystrophin binds to actin corresponding regions of dystrophin, and utrophin binds to
filaments present in the inner layer. Therefore, 0.- and p-dystroglycan and actin filaments in the similar manner to
P-dystroglycans and dystrophin form an axis connecting the dystrophin [80, 81). It is plausible that in DMD muscle fibers
outer and inner layers. The axis is socketed with laminin and utrophin and dystroglycans form a utrophin axis resembling
actin filaments and the sarcoglycan complex might strengthen the dystrophin axis, and such a utrophin axis may be fixed
the axis. Since the axis is composed of 0.- and P-dys- in a casette formed by laminin and actin filaments [15].
troglycans and dystrophin, this complex is called the However, it is unknown to what extent utrophin can fun-
dystrophin axis here. ctionally replace dystrophin. Utrophin is used in various
148

non-contracting cells other than muscle fiber [82]. Although of the lipid bilayer in the cell membrane of DMD muscle,
utrophin expression is upregulated in DMD muscle, the where the degree ofmechanical protection ofthe lipid bilayer
muscle fibers ultimately succumb to dystrophic change [76, by the outer and inner layers is decreased.
77]. Therefore, it seems that the utrophin axis present in DMD In this context, it is tempting for us to speculate that in
muscle may not be mechanically so strong as dystrophin. DMD muscle fibers the lipid bilayer is torn by mechanical
The mdx mouse has a point mutation in dystrophin gene agitation on contraction, resulting in formation of tiny clefts
resulting in the loss of dystrophin from skeletal muscle [83, in the lipid bilayer. Is resealing ofthe lipid bilayer following
84]. When Dp71, a short dystrophin gene product spanning mechanical agitation possible? We have no experimental
roughly from the cysteine-rich domain to the C-terminal evidence indicating that it is. However, in electron phy-
domain but not having actin-binding domain [85, 86], is siological studies when an electrode is inserted into a muscle
transgenically expressed in the mdx mouse, DAPs are fiber, the membrane potential reaches a steady resting level
recruited onto the cell membrane but muscular dystrophy following the exhibition of an artefactual wave. This may
symptoms are not alleviated and serum creatine kinase activity indicate resealing of the lipid bilayer after mechanical injury.
remains high [86, 87]. Because Dp71 has neither actin-binding From the tiny clefts which may appear over a very short
domain nor most ofthe rod domain, dystroglycans-Dp71 may period of time, creatine kinase would be released into the
connect to outer layer but may not to the inner layer. This is extracellular space. Following repeated or continuous heavy
likely why the DMD symptoms of the DP71-transgenic mdx exercise, a certain amount of creatine kinase may be released
mouse are not alleviated [17]. When dystrophin or utrophin is and accumulate in the serum. If such clefts appear due to the
overexpressed in transgenic mdx mice, the DMD symptoms contraction-relaxation cycle, many known and unknown
are alleviated and serum creatine kinase activity is decreased protoplasmic substances would be lost into the extracellular
[88, 89]. In these cases, both dystrophin and utrophin seem space. At the same time, some toxic substances such as Ca
to be involved in the axis formation. This suggests that the ions in the extracellular fluid may enter into the protoplasm.
utrophin axis when present in very large amounts can The efflux and influx of these substances, when repeated for
functionally replace the dystrophin axis. a long period, might cause or facilitate the dystrophic changes
Taking above into consideration, preservation of the in muscle fibers.
mechanical protecting systems composed of the outer and Therefore, if the membrane clefts were not resealed and
inner layers and the dystrophin or utrophin axis is essential were present continuously, the muscle fibers would be
for preventing the increase in serum creatine kinase level. destroyed and the muscle mass would be decreased.
Concerning the appearance of the clefts following exercise,
the following observations are noteworthy. Many pediatric
We still do not know how creatine neurologists have observed that in DMD children, serum
creatine kinase activity is increased following walking for some
kinase molecules are released in the distance but decreased following admission of patients to
serum from muscle fibers hospital which necessarily restricts the patients' movements.
Following vigorous exercise, DMD children sometimes suffer
Nearly 40 years have elapsed since Professor Ebashi dis- from muscle contracture in their legs, which likely be resulted
covered the increase in serum creatine kinase activity in DMD from the entrance of Ca ions into muscle fibers.
patients [10]. Before dystrophin was discovered, various
studies were carried out based on the viewpoint that there
must be lesions in muscle cell membrane (reviewed by Epilogue
Rolland [18]). After the discovery of dystrophin, the
molecular structure of the outer and inner layers connected On the basis of knowledge regarding the molecular structure
with dystrophin and DAPs has been fairly well elucidated. of dystrophin and DAPs, we speculated on the temporary
In spite of these studies, we are still unable to explain how appearance of tiny clefts in the lipid bilayer which might be
creatine kinase is released from muscle fiber by passing resulted of contraction and may playa crucial role in dev-
through the muscle cell membrane barrier. elopment ofDMD changes in muscle fibers. Needless to say,
It is probably because that the barrier is not composed of there is no reliable experimental evidence to support this
proteins of the basal lamina or subsarcolemmal cytoskeletal speculation. We believe that the lipid bilayer must be
networks, but is the lipid bilayer. There have been almost no extensively studied in the near future.
data available, however, concerning changes in the lipid Dystrophin may have functions other than the mechanical
bilayer in DMD muscle. Are there any changes in the fixation of outer and inner layers in the protective system of
chemical composition of the bilayer? We cannot answer this. lipid bilayer. Very recently, some molecules which may be
But we would like to focus our attention on the physical status related to signal transduction have been reported to bind to
 149

dystrophin or DAPs. These molecules are nitrogen oxide 14. Sugita H, Tyler FH: Pathogenesis of muscular dystrophy. Trans Assoc
synthase [90], aciculin [91-93], Grb 2 [94], and caveolin 3 Amer Physicians 76: 231-243, 1963
15. Goto I, Nagamine M, Kathuki S: Creatine phosphokinase isozymes
[95, 96]. Therefore, dystrophin and DAPs might participate in-muscle. Arch Neurol 20: 422--429, 1969
in signal transduction. Our present speculation outlined above 16. Somer H, Donner M, Murros J, Konttinen A: A serum isozyme study
does not exclude the possibility that dystrophin plays various in muscular dystrophy. Arch Neurol 29: 343-345, 1973
roles other than the mechanical fixation of the outer and inner 17. Takahashi K, Shutta K, Matsuo B, Takai T, Takao H, Imura H: Serum
layers, which is no doubt one of the most important functions creatine kinase isoenzymes in Duchenne muscular dystrophy. Clinica
Chimica Acta 75: 435--442, 1977
of dystrophin. 18. Rowland LP: The membrane theory of Duchenne dystrophy: Where
In any case, the problem raised by Prof. Ebashi long ago is it? Ital J Neurol Sci 3(suppl): 13-28, 1984
regarding what mechanism underlies creatine kinase release 19. Tinsley JM, Blake OJ, Zuellig RA, Davies KE: Increasing complexity
from muscle fiber remains unsolved, but it no doubt suggests of the dystrophin-associated protein complex. Proc NatlAcad Sci USA
the key for solving the nature of DMD. 91: 8307-8313, 1994
20. Ozawa E, Yoshida M, Suzuki A, Mizuno Y, Hagiwara Y, Noguchi S:
Dystrophin-associated proteins in muscular dystrophy. Hum Mol Genet
4: 1711-1716,1995
References 21. Campbell KP: Three muscular dystrophies: Loss of cytoskeleton-
extracellular matrix linkage. Cell 80: 675--679, 1995
I. Koenig M, Hoffman EP, Bertelson CJ, MonacoAP, Feener C, Kunkel 22. Fawcett OW: A Textbook of Histology. Chapman & Hall, New York,
LM: Complete cloning of the Duchenne muscular dystrophy (DMD) 1994, pp 260-308
cDNA and preliminary genomic organization of the DMD gene in 23. Zubrzycka-Gaam EE, Bulman DE, Karpati G, Bughes AHM, Belfall
normal and affected individuals. Cell 50: 509-517, 1987 B, Klamut HJ, Talbot J, Hodges RS, Ray PN, Worton RG: The
2. Hoffman EP, Brown RH Jr, Kunkel LM: Dystrophin: The protein Duchenne muscular dystrophy gene product is localized in sarcolemma
product of the Duchenne muscular dystrophy locus. Cell 51: 919- of human skeletal muscle. Nature 333: 466--469,1988
928,1987 24. Bonilla E, Samitt CE, Miranda AF, Hays AP, Salviati G, DiMauro S,
3. Koenig M, Monaco AP, Kunkel LM: The complete sequence of Kunkel LM, Hoffman EP, Rowland LP: Duchenne muscular dystrophy:
dystrophin predicts a rod-shaped cytoskeletal protein. Cell 53: 219- Deficiency of dystrophin at the muscle cell surface. Cell 54: 447-
228, 1988 452, 1988
4. Hoffman EP, Kunkel LM: Dystrophin abnormalities in Duchenne/ 25. Cullen MJ, Watkins SC: Ultrastructure of muscular dystrophy: New
Becker muscular dystrophy. Neuron 2: 1019-1029, 1989 Aspects. Micron 24: 287-307, 1993
5. Arahata K, Ishiura S, Ishiguro T, Tsukahara T, Suhara Y, Eguchi C, 26. Ervasti JM, Campbell KP: Membrane organization of the dystrophin-
Ishihara T, Nonaka I, Ozawa E, Sugita H: Immunostaining of skeletal glycoprotein complex. Cell 66: 1121-1131, 1991
and cardiac muscle surface membrane with antibody against Duchenne 27. Ervasti JM, Ohlendieck, K, Kahl SO, Gaver MG, Campbell KP:
muscular dystrophy peptide. Nature 333: 861-863, 1988 Deficiency ofa glycoprotein component of the dystrophin complex in
6. Watkins SC, Hoffman EP, Slayter HS, Kunkel LM: Immunoelectron dystrophic muscle. Nature 345: 315-319, 1990
microscopic localization of dystrophin in myofibres. Nature 333: 863- 28. Yoshida M, Ozawa E: Glycoprotein complex anchoring dystrophin to
866, 1988 sarcolemma. J Biochem 108: 748-752, 1990
7. Hoffman EP, Fischbek KH, Brown RH, Johnson M, Medori R, Loike 29. Nigro V, Piluso G, BelsitoA, Politano J, PucaAA, Papparella S, Rossi
JD, Harris JB, Waterston R, Brooke M, Specht L, Kupsky W, E, Viglietto G, Esposito MG,Abbondanza C, Medici N, Molinari AM,
Chamberlain J, Caskey T, Shapiro F, Kunkel LM: Characterization of Nigro G, Puca GA: Identification of a novel sarcoglycan gene at 5q33
dystrophin in muscle-biopsy specimens from patients with Duchenne's encoding a sarcolemmal 35 kDa glycoprotein. Hum Mol Genet 5:
or Becker's muscular dystrophy. New Engl J Med 318: 1363-1368, 1179-1186, 1996
1988 30. Jung 0, Duclos F, Apostol B, Straub V, Lee JC, Allamand V, Venzke,
8. Arahata K, Hoffman EP, Kunkel LM, Ishiura S, Tsukahara T, Ishihara DP, Sunada Y, Moomaw CR, Leveille CJ, Slaughter CA, Crawford
T, Sunohara N, Nonaka I, Ozawa E, Sugita H: Dystrophin diagnosis: TO, McPherson JD, Campbell KP: Characterization of 3-sarcoglycan,
Comparison of dystrophin abnormalities by immunofluorescence and a novel component of the oligomeric sarcoglycan complex involved
immunoblot analyses. Proc Natl Acad Sci USA 86: 7154-7158, 1989 in limb-girdle muscular dystrophy. J Bioi Chern 271: 32321-32329,
9. Emery AEH, Emery MLH: The History ofa Genetic Disease Duchenne 1996
Muscular Dystrophy or Meryon's Disease. The Royal Society of 3 I. Yoshida M, Noguchi S, Wakabayashi E, Piluso G, BelsitoA, Nigro V,
Medicine Press Limited, New York, 1995 Ozawa E: The fourth component of the sarcoglycan complex. FEBS
10. Ebashi S, Toyokura Y, Momoi H, Sugita H: High creatine phos- Lett 403: 143-148,1997
phokinase activity of sera of progressive muscular dystrophy. J 32. Yamamoto H, Hagiwara Y, Mizuno Y, Yoshida M, Ozawa E: Het-
Biochem 46: 103-104, 1959 erogeneity of dystrophin-associated proteins. J Biochem 114: 132-
II. Okinaka S, Sugita H, Momoi H, Toyokura Y, Kumagai H, Ebashi S, 139,1993
Fujie Y: Serum creatine phosphokinase and aldolase activity in 33. Yoshida M, Suzuki A, Yamamoto H, Noguchi S, Mizurto Y, Ozawa E:
neuromuscular disorders. TransactAmer NeurolAssoc 84: 62--64, 1959 Dissociation of the complex of dystrophin and its associated proteins
12. Dreyfus J-C, Schapira G, Demos J: Etude de la creatine-kinase serique into several unique groups by n-octyl ~-D-glucoside. Eur J Biochem
chez les myopathes et leurs families. Revue Fran9aise Etude Clinique 222: 1055-1061,1994
et Biologique 5: 384-386, 1960 34. Ibraghimov-Beskrovnaya 0, Ervasti JM, Leveille CJ, Slaughter CA,
13. EngelAG, Yamamoto M, Fischbeck KH: Dystrophinopathies.ln:A.G. Semett SW, Campbell KP: Primary structure of dystrophin-associated
Engel, C. Franzini-Armstrong (eds). Myology. McGraw-Hill Inc, New glycoproteins linking dystrophin to the extracellular matrix. Nature
York, 1994,pp 1133-1187 355: 696-702, 1992
150

35. Roberds SL, Leturcq F, Allamand V, Piccolo F, Jeanpierr M,Anderson 50. Carr C, Fischbach GD, Cohen 18: A novel 87,000-Mr protein
RD, Lim LE, Lee JC, Tome FMS, Romero NB, Fardeau M, Beckmann associated with acetylcholine receptors in Torpedo electric organ and
JS, Kaplan J-Q Campbell KP: Missense mutations in the adhalin gene vertebrate skeletal muscle. J Cell Bioi 109: 1753-1764, 1989
linked to autosomal recessive muscular dystrophy. Cell 78: 625--633, 51. Wagner KR, Cohen JB, Huganir RL: The 87K postsynaptic membrane
1994 protein from Torpedo is a protein-tyrosine kinase substrate homologous
36. McNally EM, Yoshida M, Mizurto Y, Ozawa E, Kunkel LM: Human todystrophin. Neuron 10: 511-522, 1993
adhalin is alternatively spliced and the gene is located on chromosome 52. Yoshida M, Yamamoto H, Noguchi S, Mizuno Y, Hagiwara Y, Ozawa
17q21. Proc Natl Acad Sci USA 91: 9690--9694, 1994 E: Dystrophin-associated protein AO is a homologue of the Torpedo
37. Lim LE, Duclos F, Broux 0, Bourg N, Sunada Y, Allamand V, Meyer 87K protein. FEBS Lett 367: 311-314, 1995
J, Richard I, Moomaw C, Slaughter C, Tome FMS, Fardeau M, Jackson 53. Blake DJ, Nawrotzki R, Peters MF, Froehner SC, Davies KE: Isoform
CE, Beckmann JS, Campbell KP: p-sarcoglycan: Characterization and diversity of dystrobrevin, the murine 87-kDa postsynaptic protein. J
role in limb-girdle muscular dystrophy linked to 4q12. Nature Genet Bioi Chern 271: 7802-7810, 1996
II: 257-265, 1995 54. Sadoulet-Puccio HM, Khurana TS, Cohen JB, Kunkel LM: Cloning
38. Bonnemann CG, Modi R, Noguchi S, Mizuno Y, Yoshida M, Gussoni and characterization of the human homologue of a dystrophin related
E, McNally EM, Duggan DJ, Angelini C, Hoffman EP, Ozawa E, phosphoprotein found at the Torpedo electric organ post-synaptic
Kunkel LM: p-sarcoglycan (A3b) mutations cause autosomal recessive membrane. Hum Mol Genet 5: 489-496, 1996
muscular dystrophy with loss of the sarcoglycan complex. Nature 55. Suzuki A, Yoshida M, Ozawa E: Mammalian <x1- and pl-syntrophin
Genet II: 266-273,1995 bind to the alternative splice-prone region of the dystrophin COOH
39. Noguchi S, McNally EM, Ben Othmane K, Hagiwara Y, Mizuno Y, terminus. J Cell Bioi 128: 373-381,1995
Yoshida M, Yamamoto H, Bonnermann CG, Gussoni E, Denton PH, 56. Ahn AH, Kunkel LM: Syntrophin binds to an alternatively spliced
KyriakidesT, Middleton L, Hentati F, Ben Hamida M, Nonaka I, Vance exon ofdystrophin. 1 Cell Bioi 128: 363-371, 1995
JM, Kunkel LM, Ozawa E: Mutations in the dystrophin-associated 57. Yang B, Jung D, Rafael JA, Chamberlain JS, Campbell KP: Iden-
protein y-sarcoglycan in chromosome 13 muscular dystrophy. Science tification of <X-syntrophin binding to syntrophin triplet, dystrophin,
270: 819-822,1995 and utrophin. 1 Bioi Chern 270: 4975-4978, 1995
40. Ozawa E, Noguchi S, Mizuno Y, Hagiwara Y, Yoshida M: From 58. Castello A, Brocheriou V, Chafey P, Kahn A, Gilgenkrantz H:
dystrophinopathy to sarcoglycanopathy. Evolution of a concept of Characterization of the dystrophin-syntrophin interaction using the
muscular dystrophy. Muscle Nerve 21: 421-438, 1998 two-hybrid system in yeast. FEBS Lett 383: 124-128, 1996
41. Mizuno Y, Yoshida M, Yamamoto H, Hirai S, Ozawa E: Distribution 59. Ervasti JM, Kahl SD, Campbell KP: Purification of dystrophin from
of dystrophin isoforms and dystrophin-associated proteins 43DAG skeletal muscle. J Bioi Chern 266: 9161-9165,1991
(A3a) and 50DAG (A2) in various monkey tissues. J Biochem 114: 60. Ohlendieck K, Ervasti 1M, Snook 18, Campbell KP: Dystrophin-
936-941, 1993 glycoprotein complex is highly enriched in isolated skeletal muscle
42. Yamamoto H, Mizuno Y, Hayashi K, Nonaka I, Yoshida M, Ozawa sarcolemma. J Cell Bioi 112: 135-148, 1991
E: Expression of dystrophin-associated protein 35DAG (A4) and 61. Suzuki A, Yoshida M, Yamamoto H, Ozawa E: Glycoprotein-binding
50DAG (A2) is confined to striated muscles. 1 Biochem 115: 162- site of dystrophin is confined to the cysteine-rich domain and the first
167,1994 half of the carboxy-terminal domain. FEBS Lett 308: 154-160, 1992
43. Mizuno Y, Noguchi S, Yamamoto H, Yoshida M, Suzuki A, Hagiwara 62. Beggs AH, Hoffman EP, Snyder JR, Arahata K, Specht L, Shapiro F,
Y, Hayashi, YK, Arahata K, Nonaka I, Hirai S, Ozawa E: Selective Angelini C, Sugita H, Kunkel LM: Exploring the molecular basis for
defect ofsarcoglycan complex in severe childhood autosomal recessive variability among patients with Becker muscular dystrophy: dystrophin
muscular dystrophy muscle. Biochem Biophys Res Comm 203: 979- gene and protein studies. Am J Hum Genet 49: 54--67, 1991
983,1994 63. Monaco AP, Bertelson Cl, Liechti-Gallati S, Moser H, Kunkel LM:
44. Adams ME, Butler MH, DwyerTM, Peters MF, MumaneAA, Froehner An explanation for the phenotypic differences between patients bearing
SC: Two forms of mouse syntrophin, a 58 kD dystrophin-associated partial deletions of the DMD locus. Genomics 2: 90--95, 1988
protein, differ in primary structure and tissue distribution. Neuron II: 64. Wapenaar, MC, KievitsT, Hart KA,Abbs S, Blonden LAJ, den Dunne
531-540,1993 JT, Grootscholten PM, Bakker E, Verellen-Dumoulin Ch, Bobrow M,
45. Peters MF, Kramarcy, NR, Sealock R, Froehner SC: p2-syntrophin: van Ommen G18, Pearson PL: A deletion hot spot in the Duchenne
Localization at the neuromuscular junction in skeletal muscle. muscular dystrophy gene. Genomics 2: 101-108,1988
NeuroReport 5: 1577-1580, 1994 65. Suzuki A, Yoshida M, Hayashi K, Mizuno Y, Hagiwara Y, Ozawa E:
46. AhnAH, Yoshida M,Anderson MS, Feener CA, Selig S, Hagiwara Y, Molecular organization at the glycoprotein-complex-binding site of
Ozawa E, Kunkel LM: Cloning of human basic AI, a distinct 59-kDa dystrophin. Three dystrophin-associated proteins bind directly to the
dystrophin-associated protein encoded on chromosome 8q23-24. Proc carboxy-terminal portion ofdystrophin. Eur J Biochem 220: 283-292,
Natl Acad Sci USA 91: 4446-4450, 1994 1994
47. Adams ME, DwyerTM, Dowler LL, White RA, Froehner SC: Mouse 66. lung D, Yang B, Meyer 1, Chamberlain lS, Campbell KP: Identification
<x1- and p2-syntrophin gene structure, chromosome localization, and and characterization ofthe dystrophin anchoring site on P-dystroglycan.
homology with a discs large domain. J Bioi Chern 270: 25859-25865, 1 Bioi Chern 270: 27305-27310, 1995
1995 67. Leivo I, Engvall E: Merosin, a protein specific for basement mem-
48. Yang B, Ibraghimov-Beskrovnaya 0, Moomaw CR, Slaughter CA, branes ofSchwann cells, striated muscle, and trophoblast, is expressed
Campbell KP: Heterogeneity of the 59-kDa dystrophin-associated late in nerve and muscle development. Proc Natl Acad Sci USA 85:
protein revealed by cDNA cloning and expression. J Bioi Chern 269: 1544-1548,1988
6040--6044, 1994 68. Vuolteenaho R, Nissinen M, Sainio K, Byers M, Eddy R, Hirvonen H,
49. Ahn AH, Freener CA, Gussoni E, Yoshida M, Ozawa E, Kunkel LM: Show, T B, Sariola H, Engvall E, Tryggvason K: Human laminin M
The three human syntrophin genes are expressed in diverse tissues, chain (Merosin): Complete primary structure, chromosomal assignment,
have distinct chromosomal locations, and each bind to dystrophin and and expression of the M and A chain in human fetal tissues. J Cell
its relatives. 1 BioI Chern 271: 2724-2730, 1996 Bioi 124: 381-394,1994
 151

69. Iwata Y, Nakamura H, Fujiwara K, Shigekawa M:Altered membrane- of the dystrophin-related gene product and expression in the mdx and
dystrophin association in the cardiomyopathic hamster heart muscle. dy mouse. Proc Natl Acad Sci USA 88: 3243-3247, 1991
Biochem Biophys Res Commun 190: 589-595, 1993 83. Bulfield G, Siller WG, Wight PAL, Moore KJ: X chromosome-linked
70. Mizuno Y, Noguchi S, Yamamoto H, Yoshida M, Nonaka I, Hirai S, muscular dystrophy (mdx) in the mouse. Proc Natl Acad Sci USA 81:
Ozawa E: Sarcoglycan complex is selectively lost in dystrophic hamster 1189-1192,1984
muscle. Am J Pathol146: 530-536, 1995 84. Sicinski P, Geng Y, Ryder-Cook AS, Barnard EA, Darlison MG,
71. Osari S, Kobayashi 0, Yamashita Y, Matsuish T, Goto M, Yanabe Y, Barnard PJ: The molecular basis of muscular dystrophy in the mdx
Migita T, Nonaka I: Basement membrane abnormality in merosin- mouse: A point mutation. Science 244: 1578-1580, 1989
negative congenital muscular dystrophy. Acta Neuropathol 91: 332- 85. Bar S, Barnea E, Levy Z, Neuman S, Yaffe D, Nudel U: A novel product
336, 1996 of the Duchenne muscular dystrophy gene which greatly differs from
72. Tome FMS, Evangelista T, Leclerc A, Sunada Y, Manole E, Estornet the known isoforrns in its structure and tissue distribution. Biochem J
B, Barois A, Campbell KP: Congenital muscular dystrophy with 272: 557-560, 1990
merosin deficiency. C R Acad Sci III 317: 351-357, 1994 86. Cox GA, Sunada Y, Campbell KP, Chamberlain JS: Dp71 can restore
73. Hillaire D, Leclerc A, Faure S, Topaloglu H, Chiannilulchai N, the dystrophin-associated glycoprotein complex in muscle but fails to
Guicheney P, Grinas L, Legos P, Phipot J, Evangelista T, Routon MC, prevent dystrophy. Nature Genet 8: 333-337, 1994
Mayer M, Pellissier JF, Estournet B, BaroisA, Hentati F, Feingold N, 87. Greenberg DS, Sunada Y, Campbell KP, Yaffe D, Nudel U: Exogenous
Beckmann JS, Dubowitz V, Tome FMS, Fardeau M: Localization of Dp71 restores the levels of dystrophin-associated proteins but does
merosin-negative congenital muscular dystrophy to chromosome 6q2 not alleviate muscle damage in mdx mice. Nature Genet 8: 340-344,
by homozygosity mapping. Hum Mol Genet 3: 1657-1661,1994 1994
74. Yoshida M, Mizuno Y, Nonaka I, Ozawa E: A dystrophin-associated 88. Cox GA, Cole NM, Matsumura K, Phelps SF, Hauschka SD, Campbell
glycoprotein, A3a (one of 43DAG doublets), is retained in Duchenne KP, Faulkner JA, Chamberlain JS: Overexpression of dystrophin in
muscular dystrophy muscle. J Biochem 114: 634-639, 1993 transgenic mdx mice eliminated dystrophic symptoms without toxicity.
75. Love DR, Hill DF, Dickson G, Spurr NK, Byth BC, Varsden RF, Walsh Nature 364: 725-729, 1993
FS, Edwards YH, Davies KE: An autosomal transcript in skeletal 89. Tinsley JM, Potter AC, Phelps SR, Fisher R, Tricket JI, Davies KE:
muscle with homology to dystrophin. Nature 339: 55058, 1989 Amelioration of the dystrophic phenotype of mdx mice using a
76. Tanaka H, Ishiguro T, Eguchi C, Saito K, Ozawa E: Expression of a truncated utrophin transgene. Nature 384: 349-353, 1996
dystrophin-related protein associated with the skeletal muscle cell 90. Brenman JE, Chao DS, Xia H, Aldape K, Bredt DS: Nitric oxide
membrane. Histochemistry 96: 1-5,1991 synthase complexed with dystrophin and absent from skeletal muscle
77. Mizuno Y, Nonaka I, Hirai S, Ozawa E: Reciprocal expression of sarcolemma in Duchenne muscular dystrophy. Cell 82: 743-752, 1995
dystrophin and utrophin in muscles of Duchenne muscular dystrophy 91. Belkin AM, Burridge K: Expression and localization of the phos-
patients, female DMD-carriers and control subjects. J Neurol Sci 119: phoglucomutase-related cytoskeletal protein, aciculin, in skeletal
43-52, 1993 muscle. J Cell Sci 107: 1993-2003, 1994
78. Mizuno Y, Yoshida M, Nonaka I, Hirai S, Ozawa E: Expression of 92. BelkinAM, Burridge K: Localization ofutrophin and aciculin at sites
utrophin (dystrophin-related protein) and dystrophin-associated of cell-matrix and cell-cell adhesion in cultured cells. Esp Cell Res
glycoproteins in muscles from patients with Duchenne muscular 221: 132-140, 1995
dystrophy. Muscle Nerve 17: 206-216, 1994 93. Moiseeva EP, Belkin AM, Spurr NK, Koteliansky VE, Critchley DR:
79. thi Man N, Ellis JM, Love DR, Davies KE, Gatter KC, Dickson G, A novel dystrophin/utrophin-associated protein is an enzymatically
Morris GE: Localization of the DMDL gene-encoded dystrophin- inactive member of the phosphoglucomutase superfamily. Eur J
related protein using a panel of nineteen monoclonal antibodies: Biochem 235: 103-113,1996
Presence at neuromuscular junctions, in the sarcolemma ofdystrophic 94. Yang B, Jung D, Motto D, Meyer J, Koretzky G, Campbell KP: SH3
skeletal muscle, in vascular and other smooth muscles, and in domain-mediated interaction of dystroglycan and Grb2 J BioI Chern
proliferating brain cell lines. J Cell BioI 115: 1695-1700, 1991 270: 11711-11714,1995
80. Matsumura K, Ervasti JM, Ohlendieck, K, Kahl SD, Cambpell KP: 95. Song KS, Li S, Okainoto T, Quilliam LA, Sargiacomo M, Lisanti MP:
Association of dystrophin-related protein with dystrophin-associated Copurification and direct interaction ofRas with Caveolin, an integral
proteins in mdx mouse muscle. Nature 360: 588-591, 1992 membrane protein ofcaveolae microdomains. J Bioi Chern 271: 9690-
81. Tinsley, JM, Blake DJ, RocheA, Fairbrother U, Riss J, Byth BC, Knight 9697, 1996
AE, Kendrick-Jones J, Suthers GK, Love DR, Edwards YH, Davies 96. Song KS, Scherer PE, Tang Z, Okamoto T, Li S, Chafel M, Chu C,
KE: Primary structure ofdystrophin-related protein. Nature 360: 591- Kohtz DS, Lisanti MP: Expression of caveolin-3 in skeletal, cardiac,
593, 1992 and smooth muscle cells. J Bioi Chern 271: 15160-15165, 1996
82. Love DR, Morris GE, Ellis JM, Fairbrother U, Marsden RF, Bloomfield 97. Ozawa E. Muscular Dystrophy (in Japanese). Japanese Muscular
JF, Edwards YH, Slater CP, Parry DJ, Davies KE: Tissue distribution Dystrophy Association, 1997
Molecular and Cellular Biochemistry 190: 153-156, 1999.
© 1999 Kluwer Academic Publishers.

Structural basis of the function of endothelin

receptor

Tomoh Masaki, Haruaki Ninomiya, Aiji Sakamoto and Yasuo Okamoto

Department ofPharmacology, Faculty ofMedicine, Kyoto University, Japan

Abstract
Endothelin receptor is a good model for analysis of the function of heptahelical G-protein coupled receptor. In ligand binding
to the heptahelical receptor, the receptor has two functions, i.e. 'message' and 'address' functions. Each function has been
assigned to different domain of the receptor. A different part of the ligand structure also corresponds to each domain of the
receptor. Classically, classification ofreceptor has been done according to the difference ofaddress domain, i.e. affinity difference
of the receptor. However, present results predict that the classification of receptor is also possible according to the message
domain.
After stimulation of ET receptor by a ligand, the receptor transmits a signal to G-protein. Several kinds of G-proteins can
possibly be activated. Different structural domains of the receptor are assigned to the coupling of the different Ga.-protein.
Activated G-protein transmits the message to effector. Each Ga.-protein acts on different target molecules, resulting in different
responses. However, the activation of each Ga.-protein presumably depends on its intracellular level. Even if the same receptor
is activated with the same ligand, resulting final response is different from cell to cell. Therefore, classification of receptor
according to the function of the receptor is difficult. (Mol Cell Biochem 190: 153-156, 1999)

Key words: G-protein, vasoconstriction, ligand selectivity

Introduction variety of tissues and cells in different proportion and mediate

a variety of pharmacological actions.
Endothelin (ET) is a potent vasoconstrictive peptide initially In vascular tissues, ETB receptor is mainly situated on
isolated from conditioned medium of the cultured end- endothelial cells and mediates release of relaxing factor such
othelial cells [I]. It consists of 21 amino acid residues. as prostacyclin and nitric oxide. ETA is present on vascular
Subsequent analysis of endothelin gene revealed existence smooth muscle cells and mediates contraction. Some physio-
of three distinct isoforms, designated ET-I, ET-2 and ET-3 logical factors such as blood flow, or some vasoactive
[2]. Vascular endothelial cells exclusively produce ET-I. substances regulate production of ET-I in endothelial cells.
The resulting ET-I acts on the underlying smooth muscle The released ET-I from endothelium presumably acts on the
cells and induces vasoconstriction. It is now generally surrounding endothelial cells and induces release of relaxing
accepted that ETs act not only on vascular system but also factor(s). The rest of the ET may diffuse further in media. It
on non-vascular system. ETs show a variety of pharm- acts on the smooth muscle cells and elicits contraction. Then
acological actions. Indeed, ET receptor was found in various a question arises whether ET acts as a relaxing factor or a
tissues and cells. constricting factor under the physiological condition.
So far two types of ET-receptor have been reported in Presumably it may act as a constricting factor when it is
mammalian tissues. They are designated as ETA and ETB [3- overproduced, whereas it may act as a relaxing factor at the
5]. ETA has a high affinity to ET-I and ET-2, but low affinity low concentration. However, it depends on the condition of
to ET-3, while ETB has an equal affinity to all of the three the vascular beds [6].
endogenous ETs. Those two receptors are distributed in a

Address for offprints: T. Masaki, National Cardiovascular Centre Research Institute, Fujishiroeai, Suita, Osaka 565-8565, Japan
154

Diversity of ET-induced several physiological and pathophysiological factors, such as

blood flow [16], heart failure [18] or atherosclerosis [17].
vasoconstriction Thus the activity of ET-l-induced contraction ofthe vascular
beds depends on the vascular bed and the surrounding
Several lines ofin vivo experiments provided a strong evidence physiological and pathophysiological condition.
that ET has a physiological role in humans in maintenance of
normal vascular tone with increase in peripheral resistance [7,
8]. A specific ETA antagonist, BQ-123 caused a decrease in
Structural determinant of ligand
vascular resistance when it was infused in human forearm
vascular beds, suggesting an involvement of endogenous ET- selectivity
1 and ETA in mechanism of vascular tone [7]. On the other
hand, ETB antagonist, BQ788, abolished the depressor res- Both ETA and ETB belong to the superfamily of heptahelical
ponse induced by ET-I and induced a transient increase in G-protein coupled receptors. Both receptors exhibit a high
blood pressure when administered in vascular beds suggesting polypeptide sequence identity with each other. On the other
an involvement ofETB in decrease in vascular resistance [9]. hand, a number of synthetic agonists and antagonists for the
Another investigator reported that systemic infusion of non- ET-receptors have been developed recently [19]. Carboxy-
selective ETA/ETB antagonists, TAK-044, decreased also terminus ofthose peptide ligands and endogenous ETs are very
peripheral vascular resistance, suggesting an involvement of similar suggesting that the binding sites of the ET-receptor to
ETB as well as ETA in regulation of vascular tone [10]. those ligands are the same. Nevertheless, the two receptors
In vitro experiments on rabbit jugular and saphenous veins show a clear distinction in ligand binding selectivity. To
and human mammary arteries also demonstrated that ETB is elucidate the subtype-specific determinant ofthe selectivity of
involved in vasoconstriction [11]. Similar results were both receptors, we constructed a series of chimeric receptors
reported by other investigators [12]. The ETB mediating the between those two receptors and characterized those receptors
vasoconstriction is a different ETB from endothelial ETB. This in terms ofligand selectivity [20, 21]. The results showed that
fact was verified, for example, in rabbit saphenous vein. ETB- ET-receptor consisted oftwo distinct domains. Amino-terminal
receptor agonists IRLl620 and S6c as well as ET-l elicited structure of agonist and receptor structure including trans-
contraction in this tissue when endothelium was removed. membrane domains (TMDs) IV-VI determined the selection
This vasoconstriction elicited by S6c appeared to be mediated ofligand-receptor. ETB-selective agonists, ET-3, BQ3020 and
by two types of ETB. This contraction was not sensitive to IRL 1620 bound to a ETB-like chimeric receptor that has the
ETA-selective antagonist BQ-123, but sensitive to non- TMDs IV-VI and adjacent loop regions from ETB inserted into
selective ET-receptor antagonist PDl42893. In addition, the the remaining regions from ETA. An ETA-selective antagonist
major component of S6c-induced contraction was resistant BQ-123 totally inhibited the binding of the ETB-selective
to PD142893, while the IRLl620-induced contraction was ligands to the ETB-like chimera. In contrast, the carboxy-
inhibited by PDl42893 completely. Those results strongly terminal portion of the ligands and the TMDs I, II, III and VII
suggest the existence of two subtypes of ETB receptor, plus intervening loop regions play an important role in the
designated ETBI (PD 142893-sensitive) and ETB2 (PD 142893- ligand-receptor binding and transmission of the ligand
insensitive) respectively. Interestingly, recent reports on ETB- message. Adachi et al. provided further evidence that ET-l
receptor knockout mouse or ETB-receptor gene-deficient rat binding site of ETA receptor resided around consecutive five
revealed that both ETBI and ETB2 receptors were derived amino acid residues located on the border of the second TMD
from the same gene [13, 14]. and the second extracellular loop [22]. Therefore, amino-
To complicate further the ET-l-induced responses of terminal portion of the ligands and the receptor domain
vascular beds, ETA and ETB are distributed heterogeneously including TMD IV-VI can be considered as the 'address'
between different species and between different vessels, even domain ofligand and receptor, and the latter domain containing
in the same blood vessels [15]. Furthermore, plasticity ofET the second TMD and carboxy-terminal portion of ET is
receptor, particularly ETB-receptor in response to exogenous considered to be 'message' domain.
factors complicates the response [16-18]. In human coronary
artery, both ETA and ETB are distributed in the media of
vascular bed. Vasoconstriction oflarge coronary artery seems Structural determinant of G-protein
to be predominately mediated via ETA receptor. However, the coupling
ET-l induced constriction in the coronary artery may possibly
be mediated via ETB particularly in peripheral part of the It has been demonstrated that ET-l activates phospholipase
coronal artery because density ofETB increases towards distal C, and regulated adenylate cyclase activity, i.e. it increases
regions [15]. In addition, expression ofETB is regulated by the cAMP level via ETA and decreases the cAMP level after
 155

Ligand formation in each cell line was tested. When the sequence
C N
from amino-terminus to the third transmembrane domain of
ETA receptor was replaced by that of ETB receptor, activity

/ t of cAMP formation by ET-I did not change. However, when

the sequence from amino-terminus to the second intracellular
loop was replaced by that of ETB receptor, the activity of
cAMP formation decreased. Conversely, when carboxy-
terminus of wild type ETA was replaced by that of ETB until
the seventh transmembrane domain, effect ofET-1 on cAMP
formation was unchanged. However, sequence from carboxy-
terminus to the fifth transmembrane domain was replaced by
that of ETB, the activity decreased dramatically, but in-
complete. If the sequence was replaced by that ofETB from
carboxy-terminus until the second intracellular loop, the
activity was totally lost. These results indicate that intra-
cellular second loop and some portion ofthe third loop ofETA
Gas Gill Gllq are involved in activation of Gas. Similarly, ETB with the
replacement ofthe third intracellular loop to the corresponding
region of ETAfailed to transmit the inhibitory effect of ET-
forskolin stimulation via ETB, suggesting an involvement of I on cAMP level, suggesting that the third intracellular loop
various types of Ga-proteins. In fact, we demonstrated that of ETB was involved in Gai coupling. In addition, recent
both ETA and ETB receptors expressed in COS-7 cells were experiments revealed that carboxy-terminal portion of ET
able to couple to at least Gaq, Gall, Gas and Gai2 [23]. receptor was also involved in activation of Gai [27].
However, subsequent our and other's reports showed that Previous reports demonstrated that ET-I did not induce a
ETA expressed in CHO cells was coupled to Gas but not Gai transient increase in intracellular free calcium level in CHO
[24, 25]. In contrast, ETA expressed in guinea pig atrial cells transfected with carboxy-terminus truncated ETA gene,
myocardium was shown to induce down-regulation of cAMP suggesting a loss of Gaq coupling [28]. An ETA mutant
level, when it was stimulated by ET-I [26] suggesting the receptor which five cystein residues of carboxy-terminus of
coupling of Gai to ETA. The promiscuous nature of the ETA were replaced by serine residues showed no incorporation
coupling of multiple types of G-proteins may be ascribed to of palmitate [29]. CHO-KI cell transferred with the mutant
the status of the intracellular condition which presumably ETA receptor failed to stimulate phospholipase C with ET-I.
modifies the receptor or receptor-related proteins. Similar results were obtained with ETB mutant. Those results
To elucidate these complication problems, we searched the strongly suggest that the modification of carboxy-terminus,
receptor domain(s) that determined the selective coupling of sequence particularly palmitoylation of cystein residue is
the receptor with Gas and Gai [25]. We used again chimeric critical in Gaq coupling.
receptors between ETA and ETB. The chimeric receptors were
expressed on CHO cells, and the effect of ET-I on cAMP
References
I. Yanagisawa M, Kurihara H, Kimura S, Tomobe Y, Kobayashi M, Mitsui
Table 1. Sensitivity of various agonists and antagonists to two subtypes of Y, Yazaki Y, Goto K, Masaki T: A novel potent vasoconstrictor peptide
ETB receptor in rabbit saphenous vein produced by vascular endothelial cells. Nature 322: 411-415,1988
2. Inoue A, Yanagisawa M, Kimura S, Kasuya Y, Miyauchi T, Goto K,
ETBI ETB2 Masaki T: The human endothelin family: Three structurally and
pharmacologically distinct isopeptides predicted by three separate
BQ-123
genes. Proc Nat! Acad Sci USA 86: 2863-2867, 1989
(ETA antagonist)
3. Arai H, Hori S, Aramori 1, Ohkubo H, Nakanishi S: Cloning and
expression of a cDNA encoding an endothelin receptor. Nature 348:
PD-142893
730-732, 1990
(ETA/ETB antagonist) +
4. Sakurai T, Yanagisawa M, Takuwa Y, Miyazaki H, Kimura S, Goto K,
Masaki T: Cloning of a cDNA encoding a non-isopeptide-selective
IRLl620
subtype of the endothelin receptor. Nature 348: 732-735, 1990
(ETB agonist) +
5. Masaki T, Vane J, Vanhoutte P: V. International union of pharmacology
nomenclature ofendothelin receptor. Pharmacol Rev 46: 137-142, 1994
STX-S6c
6. Masaki T: Possible role of endothelin in endothelial regulation of
(ETB agonist) +
vascular tonus. Ann Rev Pharmacol Toxicol 35: 235-255, 1995
156

7. Haynes WG, Webb DJ: Contribution of endogenous generation of 19. Elliott JD, Lago MA, Peishoff CE: Endothelin receptor antagonists.
endothelin-I to basal vascular tone. Lancet 344: 852-854, 1994 In: R.R. Ruffolo Jr. (ed). Endothelin Receptors from the Gene to the
8. HaynesWG, Ferro CJ, O'Kane KPJ, Somerville D, Lomax CC, Webb Human. CRC Press, Boca Raton, New York, London, Tokyo, 1995,
DJ: Systemic endothelin receptor blockade decreases peripheral pp 79-108
vascular resistance and blood pressure in humans. Circulation 93: 20. Sakamoto A, Yanagisawa M, Sawamura T, Enoki T, Ohtani T, Sakurai
1860-1870, 1996 T, Nakao K, Toyo-oka T, Masaki T: Distinct subdomains of human
9. Ishikawa K, Ihara M, Noguchi K, Mase T, Mino N, Saeki T, Fukuroda T, endothelin receptors determine their selectivity to endothelin A-
Fukami T, Ozaki S, Nagase T, Nishikibe M, Yano M: Biochemical and selective antagonist and endothelin B-selective agonists. J Bioi Chern
pharmacological profile of a potent and selective endothelin B-receptor 268: 8547-8553, 1993
antagonist BQ-788. Proc Nat! Acad Sci USA 91: 4892-4896, 1994 21. Sakamoto A, Yanagisawa M, Sakurai T, Nakao K, Toyo-oka T, Yano
10. Plumpton C, Ferro CJ, Haynes WG, Webb DJ, Davenport AP: The M, Masaki T: The ligand-receptor interactions of the endothelin
increase in human plasma immunoreactive endothelin but not big systems are mediated by distinct 'message' and 'address' domains. J
endothelin-I or its C-terminal fragment induced by systemic Cardiovasc Pharmacol 22(suppl 8): S113-S 116, 1993
administration of the endothelin agonist TAK-044. Brit J Pharmacol 22. Adachi M, Hashida K, TrzeciakA, WatanabeT, Furuichi Y, Miyamoto
119: 311-314, 1996 C: Functional domains of human endothelin receptor. J Cardiovasc
II. Nishiyama M, Moroi, K, Shan L, Yamamoto M, Takasaki C, Masaki T, Pharmacol 22(suppl 8): SI21-S124, 1993
Kimura S: Two distinct endothelin B receptor subtypes mediate con- 23. Takigawa M, Sakurai T, Kasuya Y, Abe Y, Masaki T, Goto K:
traction ofthe rabbit saphenous vein. Jpn J Pharmacol68: 235-243, 1995 Molecular identification of guanine-nucleotide-binding regulatory
12. Sudjarwo S, Hori M, Takai M, Urade Y, Okada T, Karaki H: A novel proteins which couple to endothelin receptors. Eur J Biochem 228:
subtype of endothelin B receptor mediating contraction in swine 102-108,1995
pulmonary vein. Life Sci 53: 431-437,1993 24. Aramori I, Nakanishi S: Coupling of two endothelin receptor subtypes
13. Mizuguchi T, Nishiyama M, Moroi K, Tanaka H, Saito T, Masuda Y, to differing signal transduction in transfected Chinese hamster ovary
Masaki T, de Wit D, Yanagisawa M, Kimura S: Two endothelin B cells. J BioI Chern 267: 12468-12474, 1992
receptor subtypes predicted pharmacologically are derived from one 25. Takagi Y, Ninomiya H, Sakamoto A, Miwa S, Masaki T: Structural
gene. Brit J Pharmacol (in press) basis ofG-protein specificity of human endothelin receptor: A study
14. Karaki H, Mitsui-Saito M, Takimoto M, Oda K, Okada T, Ozaki T, with endothelin AlB chimeras. J Bioi Chern 270: 10072-10078,
Kunieda T: Lack of endothelin ETB receptor binding and function in 1995
the rat with a mutant ETB receptor gene. Biochem Biophys Res 26. Ono K, Tsujimoto G, Sakamoto A, Eto K, Masaki T, Ozaki Y, Sakae
Commun 222: 139-143, 1996 M: Endothelin-A receptor mediates cardiac inhibition by regulating
15. Dashwood MR, Timm M, Kaski JC: Regional variations in ETAiETB calcium and potassium currents. Nature: 301-304,1994
binding sites in human coronary vasculature. J Cardiovasc Pharmacol 27. Okamoto Y, Ninomiya H, Tanioka M, Sakamoto A, Miwa S, Masaki
26(suppl 3): S351-S354, 1995 T: Palmitoylation of human endothelin-B. Its critical role in G protein
16. Barber DA, Michener SR, Ziesmer SC, MillerVM: Chronic increases coupling and a differential requirement for the cytoplasmic tail by G
in blood flow upregulate endothelin-B receptors in arterial smooth protein subtypes. J Bioi Chern 272: 21589-21596, 1997
muscle. Am J Physiol 270: H65-H71, 1996 28. Koshimizu, Tsujimoto G, Ono K, Masaki T, Sakamoto A: Truncation
17. Dagassan PH, Breu V, Clozel M, Ktinzli A, Vogt P, Turina M, of the receptor carboxyl terminus impairs membrane signaling but not
Kiowski W, Clozel J: Upregulation of endothelin-B receptors in ligand binding of human ETB endothelin receptor. Biochem Biophys
atherosclerotic human coronary arteries. J Cardiovasc Pharmacol Res Commun 217: 354-362, 1995
27: 147-153, 1995 29. Horstmeyer A, Cramer H, Sauer T, Muller-Esterl W, Schroeder C:
18. Cannan CR, Burnett JC, Lerman A: Enhanced coronary vaso- Palmitoylation of endothelin receptor A. Differential modulation of
constriction to endothelin-B receptor activation in experimental signal transduction activity by post-translational modification. J Bioi
congestive heart failure. Circulation 93: 646-651, 1996 Chern 271: 20811-20819,1996
Molecular and Cellular Biochemistry 190: 157-167, 1999.
© 1999 Kluwer Academic Publishers.

Stimulation of plasma membrane Ca2+-pump

ATPase of vascular smooth muscle by cGMP-
dependent protein kinase: Functional
reconstitution with purified proteins

Yutaka Yoshida, Akira Toyosato, Mohammed Omedul Islam, Takaki

Koga, Satoru Fujita and Shoichi Imai
Department ofPharmacology, Niigata University School ofMedicine, Niigata, Japan

Abstract
A 240-kDa protein isolated from porcine aortic smooth muscle as a substrate for cGMP-dependent protein kinase (cGMP kinase)
whose phosphorylation was in a close association with stimulation of partially purified plasma membrane Ca2+-pumpATPase
by the kinase was later shown to represent splicing variants of type I inositol 1,4,5-trisphosphate (IP3) receptor. To further
clarify the role played by this protein in the stimulation of Ca2+-pump ATPase, it was attempted in the present study to specifically
remove the protein by immunoprecipitation with an antibody specific to type 1 IP 3 receptor. Contrary to expectation, stimulation
of the ATPase by cGMP kinase was still observed after removal of the IP 3 receptor. Furthermore, cGMP kinase stimulated a
highly purified preparation of Ca2+-pump ATPase deprived of IP 3 receptor when the concentrations of the ATPase were low
enough (10-20 nM) to make it retain a monomeric form, while it did not produce stimulation when the concentration of the
enzyme was increased to 40 nM at which the enzyme is known to take an oligomeric, fully activated form insensitive to
activation by calmodulin. Heat-inactivated cGMP kinase and cGMP kinase without cGMP failed to stimulate the highly
purified Ca2+-pump ATPase. In addition, type Ia but not type I~ cGMP kinase was found to stimulate the ATPase. The stimulation
ofCa 2+-pump ATPase by cGMP kinase occurs without any detectable phosphorylation of the ATPase. In conclusion, cGMP
kinase can stimulate the plasma membrane Ca2+-pump ATPase when it is in a monomeric form without phosphorylating the
Ca 2+-pumpATPase and that of the two cGMP kinase isozymes found in the vascular smooth muscle, only type Ia cGMP kinase
participates in the stimulation. (Mol Cell Biochem 190: 157-167,1999)

Key words: plasma membrane Ca 2+-pump ATPase, cGMP-dependent protein kinase isozymes, inositol 1,4,5-trisphosphate
receptor, vascular smooth muscle

Introduction As reduction of intracellular Ca 2+ concentration is an

important component ofcGMP-dependent relaxation [3, 4] and
It is now generally accepted that cGMP mediates vascular stimulation ofplasma membrane Ca2+-pumpATPase by cGMP
smooth muscle relaxation by nitric oxide-generating vaso- kinase was demonstrated in cultured cells [5], crude membrane
dilators, atrial natriuretic peptide, and endothelium-derived preparations [6-8] and partially purified enzyme preparations
relaxing factor [I, 2]. Phosphorylation of a specific protein [9-11], the Ca2+-pumpATPase itself was once presumed to be
or proteins by cGMP-dependent protein kinase (cGMP a substrate protein for cGMP kinase. However, it soon became
kinase) is postulated to represent the molecular event linking apparent that the Ca2+-pump itself was not a substrate for
the cGMP increase with the relaxation [3, 4]. cGMP kinase [10, 12, 13]. We found in porcine aortic smooth

Address for offprints: Y. Yoshida, Department of Pharmacology, Niigata University School of Medicine, No. 757, Asahimachi-dori 1, Niigata 951, Japan
158

muscle a 240-kDa, membrane-bound protein as a substrate for phatidylcholine (PC), and 0.05 mM CaCI 2), and buffer B
cGMP kinase whose phosphorylation was in a close asso- containing 1 M NaCI. After a further wash of the column
ciation with the stimulation ofCa2+-pumpATPase by the kinase with 120 ml of buffer B without added CaCI 2, Ca 2+-pump
[11]. The protein was later shown to be a major inositol ATPase was eluted with buffer C (0.1 % Triton X-I 00,0.16 M
1,4,5-trisphosphate (lP) receptor of the vascular smooth NaCI, 20 mM Hepes-NaOH, pH 7.4, 2 mM dithiothreitol,
muscle structurally distinct from those expressed in the 20% glycerol, 0.2 mg/ml PC, and 1 mM EGTA). The
cerebellum and representing splicing variants ofthe type 1 IP3 fractions containing Ca2+-pumpATPase were pooled, mixed
receptor [14,15]. As the phosphorylation ofIP 3 receptor has with 2 ml of concanavalin A-Sepharose pre-equilibrated with
recently been demonstrated in intact vascular smooth muscle buffer C, and shaken on ice for 2 h. Only IP 3 receptor bound
cells treated with cGMP-elevating drugs [16,17], we have tried to the resin and could be separated from Ca2+-pump ATPase
to further define the role played by the 240-kDa proteinlIP3 by centrifugation (l 00 x g, 5 min). The supernatant containing
receptor in the stimulation of plasma membrane Ca2+-pump Ca2+-pump ATPase was concentrated by re-chromatography
ATPase. Contrary to expectation stimulation of Ca 2+-pump on a calmodulin-Sepharose 4B column, frozen at liquid N2
ATPase by cGMP kinase occurred in the absence of IP 3 temperature, and stored at -80°C until use. The final enzyme
receptor. Furthermore, reconstitution experiments conducted preparation contained 30-40 Ilg/ml protein in buffer C.
with highly purified plasma membrane Ca 2+-pump ATPase
and cGMP kinase demonstrated that the stimulation of the
ATPase could occur without phosphorylation ofthe enzyme. Ca 2 +-pump ATPase assay

Ca 2+-activated activity of plasma membrane Ca 2+-pump

ATPase was determined as described previously [11]. The
Materials and methods assay was performed in a reaction mixture (250 Ill) con-
taining 100 mM KCI, 20 mM Tris-maleate, pH 7.5, 2 mM
Purification ofplasma membrane Ca 2+-pump ATPase MgCI 2, 2 mM ATP, I mM EGTA and CaCI 2 in concen-
trations yielding the required free Ca 2+ concentrations.
A partially purified preparation of plasma membrane Ca 2+- Appropriate aliquots of the enzyme were added to achieve
pump ATPase with co-purified IP 3 receptor was obtained a desired enzyme concentration. The concentration ofTriton
from porcine aorta using the method of Kosk-Kosicka et al. X-IOO was adjusted to 0.015% by adding appropriate
[18] by calmodulin affinity chromatography as described amount of a stock solution of Triton X-I 00. A saturating
previously [II]. The enzyme eluted from the affinity column concentration (200 nM) of purified calmodulin (bovine brain)
was supplemented with L-a-phosphatidyl-L-serine (PS) (0.2 was added when indicated. The reaction was started by addition
mg/ml) to restore its activity, frozen at liquid N2 temperature, of ATP and carried out for 10-30 min at 37°C. Liberated
and stored at-80°C until use. The final enzyme preparation inorganic phosphate was determined by the malachite green
contained 50-60 Ilg/ml protein, 0.1 % Triton X-I 00,0.13 M method as described previously [8]. The difference between
KCI, 10 mM Tris-maleate, pH 7.4, 0.5 mM MgCI 2, 20% the ATPase activities in the presence and absence of added
glycerol, 0.2 mg/ml PS, and 5 mM EGTA. CaCI 2 was taken as Ca 2+-activated ATPase activity.
A highly purified preparation of plasma membrane Ca 2+-
pump ATPase devoid of IP 3 receptor was obtained from
porcine aorta as described previously [11]. A concanavalin lmmunoprecipitation ofIP 3 receptor
A-Sepharose chromatography was added as a final step.
Briefly, calmodulin-depleted microsomes from porcine aorta Immunoprecipitation ofIP 3 receptor from the partially purified
(800 g) was solubilized at 4 mg/ml of protein by continuous Ca 2+-pump ATPase preparation was carried out by using an
stirring at4°C for 30 min in 0.5% Triton X-I00, 0.13 M KCI, affinity purified polyclonal antibody (APISY-l) raised against
10 mM Tris-maleate, pH 7.4, 0.5 mM MgCI 2, 0.05 mM type 1 IP 3 receptor. Rabbit antiserum was produced against a
CaCI 2, and 20% glycerol (buffer A) containing protease peptide corresponding to the C-terminal 14 amino acids of rat
inhibitors (0.1 mM phenylmethylsulfonyl fluoride, 10 units/ type 1 IP 3 receptor, and affinity-purified on peptide-coupled
ml aprotinin, 4 Ilg/mlleupeptin, and 4 Ilg/ml pepstatin A). matrix. On immunoblotting the antibody was shown to
After removal of insoluble material by centrifugation at specifically recognize type 1IP3 receptor ofeither rat or porcine
120,000 x g for 60 min, the supernatant was immediately tissues and did not cross-react with types 2 and 3 IP 3 receptors.
applied to a calmodulin-Sepharose 4B column (1.9 x 7 em). The details ofantibody production and full characterization of
The column was washed sequentially with 120 ml each of the antibody will be published elsewhere (Islam MO, Smyth GE,
buffer A containing 0.1 % Triton X-I 00, buffer B (0.1 % Yano J, Yoshida Y and Imai S, in preparation for submission).
TritonX-100, 0.16 M NaCl, 20 mM Hepes-NaOH, pH 7.4, APISY-l in TBS (0.15 M NaCI, and 20 mM Tris-HCI,
2 mM dithiothreitol, 20% glycerol, 0.2 mg/ml L-a-phos- pH 7.5) was added to the partially purified plasma membrane
 159

Ca2+-pump ATPase preparation (12.5 Ilg/ml of protein) at a Tightly bound cGMP was removed from the kinase with
ratio of21lgAPISY-lll 00 III of the enzyme preparation, and DEAE-cellulose chromatography as described by Corbin
incubated at 4°C for 2 h. Protein A-Sepharose CL-4B was and D0skeland [19]. The activity of the cGMP-free enzyme
then added (15 III ofswollen, packed resinlllg of antibody) and was found to be stimulated 6-8 fold by 1 IlM cGMP. Type
the mixture was incubated at 4°C for 2 h with gentle agitation. IP cGMP kinase was purified from porcine aorta using the
The resin was removed by centrifugation (1,000 x g, 5 min) method as developed by Wolfe et al. [20] for bovine aortic
and the resultant supernatant was immediately assayed for enzymes. The purified enzyme was apparently homo-
Ca2+-activated ATPase activity as described above. After geneous as judged by SDS-PAGE under denaturing con-
washing the precipitates three times with TBS containing ditions followed by silver-staining. Mr of a subunit of the
0.1 % Tween 20, bound protein was eluted into 3 x SDS- type IP was larger, being 80 K as compared with that of the
sample buffer (6% SDS, 0.19 M Tris-HCI, pH 6.8, 6% type la of 76 K on estimation by SDS-PAGE (data not
2-mercaptoethanol, and 30% glycerol) containing 1.2 M urea shown). The 240-kDa cGMP kinase substrate protein or type
by boiling for 5 min, and analyzed by sodium dodecyl 1 IP 3 receptor of porcine aorta was purified as described
sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). previously [14].
Control samples were prepared in the same way except that Protein concentration was determined by a modified
purified rabbit IgG (Jackson Immuno Research Laboratory) Lowry method [21] with bovine serum albumin as a standard.
was used in place of APISY-l. SDS-PAGE was conducted under a denaturing conditions on
a 6% polyacrylamide gel using the Laemmli buffer system
[22]. For silver-staining of proteins separated on SDS-PAGE
Phosphorylation by cGMP kinase gels, a Silver-Stain II kit (Wako Pure Chemicals, Osaka, Japan)
was used. For immunoblotting, proteins were separated by
Phosphorylation by cGMP kinase was conducted under SDS-PAGE under denaturing conditions, and transferred
conditions optimized for detection ofcGMP kinase substrates onto polyvinylidine difluoride membranes (Clear Blot
as described previously [15]. Briefly, 20 III of sample membrane-p, Atto, Tokyo, Japan). The immunoblots were
containing 1.2 Ilg of the highly purified or the partially probed withAPISY-l or P32 (a monoclonal antibody specific
purified preparation of plasma membrane Ca2+-pumpATPase to plasma membrane Ca2+-pumpATPase), and visualized with
was mixed with 20 III of 40 Ilg/ml purified type la or type Ip an enhanced chemiluminescence assay kit (Amersham) as
cGMP kinase and 4 IlM cGMP in TEM buffer (20 mM described previously [11].
Tris-HCI, pH 7.4, 2 mM EDTA, and 25 mM 2-mercapto-
ethanol) and 20 III of phosphorylation buffer (80 mM
Tris-HCI, pH 7.5, 40 mM Mg (CH 3COO)2 and 0.2 mM Materials
3-isobutyl-I-methylxanthine). After pre-incubation for 5 min
at 37°C, the reaction was initiated by addition of 20 III of The sources ofthe materials used in this study were as follows:
400 IlM of [y_3 2P]ATP (5 Ci/mmol) and allowed to proceed concanavalin A-Sepharose and calmodulin-Sepharose 4B
for 10 min. The reaction was terminated by addition of 40 III from Pharmacia Biotek, Uppsala; bovine brain calmodulin
of 3 x SDS-sample buffer. Heating at 37°C for 10 min and proteinA-Sepharose CL-4B from Sigma; [y_3 2P]ATP tetra
followed. A 60-1l1 portion of the sample was separated by (triethyl-ammonium) salt (30 Ci/mmol) from Du Pont-New
SDS-PAGE. The gel was silver-stained, dried, and exposed England Nuclear. P32, a monoclonal antibody raised against
to Fuji RX X-ray film at -80°C. purified porcine aortic plasma membrane Ca2+-pumpATPase,
Phosphorylation ofIP 3 receptor purified from porcine aorta was a generous gift from Dr. H. Kawakatsu, Nippon Shinyaku,
was carried out similarly as described above. Briefly, IP 3 Kyoto, Japan. All other chemicals were of the highest purity
receptor (1.7-2.3 Ilg) was incubated at 37°C for 30 min with grade commercially available.
10 Ilg/ml type la cGMP kinase plus 1 IlM cGMP. This is the
condition under which the incorporation of phosphate into
the IP 3 receptor was shown to be maximal in a previous study, Results
being approximately 1.2 mole of phosphate per mole of IP 3
receptor subunit [15]. Effect ofremoval ofIP j receptor on stimulation ofplasma
membrane Ca 2+-pump ATPase by cGMP kinase

Other methods The partially purified plasma membrane Ca2+-pump ATPase

obtained by a method of Kosk-Kosicka et al. [18] is the
Type la cGMP kinase was purified to an apparent hom- preparation that retains responsiveness to stimulation by cGMP
ogeneity from porcine lung as described previously [11]. kinase [11]. Several proteins present in the preparation were
160

found to be phosphorylated by cGMP kinase. However, only just as it was with the partially purified Ca2+-pump ATPase
the phosphorylation of 240-kDa protein was in a close corre- preparation containing a significant amount of IP 3 receptor
lation with the concentration-dependent stimulation of Ca2+- (Fig. IB).
pumpATPase. Apurified Ca2+ -pumpATPase preparation largely
devoid ofthe 240-KDa protein was not stimulated by the kinase
[II]. The 240-kDa protein was later shown to be identical with Effects ofcGMP kinase on highly purified plasma
a major IP 3 receptor of the porcine aorta and to represent membrane Ca 2+-pump ATPase
splicing variants of the type I IP 3 receptor that are structurally
distinct from the cerebellar type I IP3 receptors [14, 15]. The finding that removal of IP 3 receptor did not affect the
The IP 3 receptor was co-purified with the Ca 2+-pump stimulation of the partially purified Ca 2+-pump ATPase by
ATPase by calmodulin affinity chromatography; the two cGMP kinase was apparently in conflict with our previous
proteins bound to calmodulin in a Ca 2+-dependent manner findings that the kinase did not stimulate Ca2+-pump ATPase
and could not be separated from each other under the purified not to contain IP 3 receptor except for a trace amount
chromatographic conditions employed [15]. [II]. We, therefore, attempted to characterize the properties
As shown in Fig. lA, a polyclonal antibody (APISY-I) of plasma membrane Ca 2+-pump ATPase purified from
raised against type I IP 3 receptor specifically immuno- porcine aortic smooth muscle in relation to stimulation by
precipitated 240-kDa protein. Immunoblotting with the cGMP kinase.
antibody failed to detect any immunoreactive bands in the Combining calmodulin-affinity chromatography with
supernatants (data not shown). concanavalinA-affinity chromatography, we purified plasma
Contrary to expectation, cGMP kinase was still capable of membrane Ca2+-pump ATPase to an apparent homogeneity.
stimulating Ca 2+-pump ATPase in the absence ofIP 3 receptor In calmodulin-affinity chromatography most ofIP 3 receptor

A 8 100
2 2 IP 3 R (- )

-IPJR 80 I-
Control + -

60 -
rl- -
PMCA- Q)
<II
«l
u-

r+
fo-
e:(
40 f- -
a.
E
-+-
~

+
'?-
N«l 20 -
o

o
.. Dye
front
GPK + +
sup ppt

Fig. 1. Immunoprecipitation oflP, receptor from partially purified plasma membrane Ca'+-pump ATPase preparation and effects of cGMP kinase on Ca'+·
pump ATPase activity. The partially purified plasma membrane Ca'+-pump ATPase preparation was prepared by calmodulin affinity chromatography, and
was subjected to immunoprecipitation with APISY-I antibody specific to type I IP 3 receptor. (A) Supernatants (Sup, 40 l.tI each) and immunoprecipitates
(PPT, 60 j.ll each) obtained with either control, normallgG (Lane I) or APISY·I (Lane 2) were separated on 6% SOS-PAGE gels and silver-stained. PMCA
- plasma membrane Ca'+-pump ATPase; IP 3 R - IP 3 receptor. (B) Supernatants from the immunoprecipitation experiments were assayed for Ca'+-pump
ATPase activity in the presence or absence of purified type 10. cGMP kinase (GPK, 10 j.lglml) plus I j.lM cGMP as indicated. The Ca'+-pump ATPase
activities were determined at the free Ca'+ concentration of I j.lM. Each value represents the mean ± S.E. of three experiments. Further experimental details
are described in the Materials and methods section.
 161

was eluted with an extensive washing step with 1 M NaCl- Ca 2+-pumpATPase [23], even with a saturating concentration
containing buffer conducted prior to elution of Ca2+-pump ofcalmodulin and at a very low concentration of the enzyme
ATPase with EGTA [14]. However, to further remove IP 3 the activation of the enzyme was not full. cGMP kinase was
receptor, concanavalinA-affinity chromatography was added, found to stimulate the enzyme when the enzyme concentration
making use ofthe fact that IP 3 receptor is a glycoprotein with was 10 nM, a concentration at which the enzyme was
a high affinity to concanavalin A. Figure 2 depicts an presumed to exist as a monomeric form. The stimulation
SDS-PAGE analysis of the highly purified Ca2+-pump became smaller at higher concentrations of the enzyme,
ATPase. As is evident from the figure, no IP 3 receptor was finally disappearing at 40 nM (Fig. 3B), indicating a profound
detected either by silver-staining or by immunoblotting. influence of oligomerization ofthe enzyme on the stimulation
It was demonstrated with a detergent-solubilized, purified by cGMP kinase. At the low concentration (10 nM) of the
form of the erythrocyte Ca 2+-pump ATPase [23] that the Ca2+-pump ATPase, cGMP kinase stimulated the enzyme in
plasma membrane Ca 2+-pump ATPase took two forms, i.e. a concentration-dependent manner (Fig. 4A). The same
monomeric and oligomeric forms, depending on the con- concentration of bovine serum albumin did not affect the
centrations, half maximal oligomerization occurring between Ca2+-pump ATPase (Fig. 4B). Thus, the possibility that the
10-20 nM [23, 24]. The monomeric form is sensitive to prevention of denaturation of Ca2+-pump resulted from an
activation by calmodulin, while the oligomeric form is fully extreme dilution was prevented by addition of cGMP kinase
activated as it is and insensitive to calmodulin. The two forms was the cause of the observed stimulation by cGMP kinase
are in an equilibrium with each other. may be excluded. Heat-inactivated cGMP kinase had no
Like the erythrocyte enzyme, the highly purified Ca 2+- effect. Furthermore, the stimulation was observed only when
pump ATPase of porcine aorta was found to take two forms the cGMP kinase was in combination with cGMP (Fig. 5).
depending on the concentrations. There was a gradual These findings suggest the necessity of an intact, activated
transition from a calmodulin-sensitive form to a calmodulin- form of cGMP kinase for stimulation of plasma membrane
insensitive form with the increases in enzyme concentrations Ca 2+-pump ATPase by cGMP kinase.
(Fig. 3A). However,just as was the case with the erythrocyte Vascular smooth muscle contains two isozymes of type I
cGMP kinase, designated as 10. and IP [20, 25], that closely
c resemble with each other in many properties but differ in
A 8
sensitivities to cyclic nucleotide analogues [26]. We in-
vestigated the effects of purified 10. and IP cGMP kinases
on the activity of Ca 2+-pump ATPase. As shown in Fig. 6,
240K- l-IP,R
only type 10. cGMP kinase was capable of stimulating the
ATPase activity. The failure of type IP cGMP kinase to
stimulate the ATPase could not be attributed to no or weak
145K-
:=PMCA
phosphorylating activity of the kinase since both isozymes
135K-
of cyclic GMP kinase used in the present study were able
to phosphorylate IP 3 receptor as well as to bring about
80K- autophosphorylation of a similar magnitude as described in
the next section.

Phosphorylation ofplasma membrane Ca 2+-pump ATPase

• Dye
front by cGMP kinase
2 2 2
Confirming our previous observation [13], stimulation of
Fig. 2. SDS-PAGE profiles and immunoblotting analyses of partially plasma membrane Ca 2+-pump ATPase by type 10. cGMP
purified and highly purified plasma membrane Ca'+-pump ATPase kinase occurred without detectable phosphorylation of the
preparations. The partially purified Ca'+-pump ATPase (0.5 ~g) and the
highly purified Ca 2+-pump ATPase (0.5 ~g) was separated on 6% SDS-
Ca 2+-pumpATPase either in a partially purified or in a highly
PAGE. The gels were silver-stained (A), or probed with a monoclonal purified form (Fig. 7). Phosphorylation of the ATPase was
antibody (P32) specific to plasma membrane Ca'+-pump ATPase (B) or not observed even under conditions under which almost
with a polyclonal antibody (APISY-I) specific to type I IP,-receptor (C) maximal, stoichiometric incorporation of phosphate into the
as described in the Materials and methods section. The band with molecular IP 3 receptor could be achieved (1 mol of phosphate/mol of
mass ofabout 80 kDa represents a proteolytic fragment of plasma membrane
Ca'+-pump ATPase produced during purification [II]. PMCA - plasma
IP 3 receptor subunit) with type 10. cGMP kinase [15] and
membrane Ca'+-pump ATPase; IP,R - IP, receptor; ·silver-staining cGMP kinase itself was efficiently autophosphorylated.
artifacts. Phosphorylation by type 10. and type IP cGMP kinases
162

A 8

...
.......
.c
"- C)
20 20
E
"-"0 cGMP kinase (+)
E
::1 15 15
......,
Q)
(I)
co
Q.. 10 10
~
Q.
E
:l 5 5
~
+
N
co
<.>

0 10 20 30 40 50 0 10 20 30 40
PMCA (nM) PMCA (nM)

Fig. 3. Dependency of highly purified plasma membrane Ca 2+-pump ATPase activity and its stimulation by cGMP kinase on enzyme concentration. The
highly purified preparation of plasma membrane Ca 2+-pump ATPase (PMCA) devoid ofIP J receptor was used. The Ca 2+-pump ATPase activity was determined
at the free Ca 2+concentration of I J.lM with various concentrations of the Ca 2+-pump ATPase as indicated. (A) The activity was determined in the presence
(e) or absence (0) of 0.2 J.lM of calmodulin. (S) The activity was determined in the presence (e) or absence (0) of 10 J.lg/ml type In cGMP kinase and I
J.lM cGMP. Each value represents the mean of two experiments, each performed in duplicate. Further experimental details are described in the Materials and
methods section.

A 8

.2
~ 20
"- CJ)

"-

+
E 15
"-"0
CJ)
E
"-"0 15 E
:I.
E
:I.
Q)
10
III
III
a..
i+-
Q)

~
III
III 10 I-
a.. oe{
I-
oe{ a. 5
a. E
:l
E '?-
:l 5 +
N
.'il 0
Ill
N
Ill 0
0
0
cGMP kinase
(10 IJ g/ml) +
0 5 10
BSA
cGMP kinase (IJ g/ml) (10 I.l g/ml) +
Fig. 4. cGMP kinase stimulation of monomeric plasma membrane Ca 2+-pump ATPase. The highly purified plasma membrane Ca 2+-pump ATPase devoid of
IP J receptor was used at a fixed concentration of 10 nM. The Ca 2+-pump ATPase activity was determined at the free Ca2+ concentration of I J.lM. (A) The
enzyme activity was determined in the presence of indicated concentrations of type In cGMP kinase and I J.lM cGMP. (S) The enzyme activity was
determined in the presence or absence of 10 J.lg/ml type In cGMP plus I J.lM cGMP kinase or 10 J.lg/ml bovine serum albumin (SSA) plus I J.lM cGMP as
indicated. Each value represents the mean of two or three experiments, each performed in duplicate.
 163

yielded a similar set of phosphoproteins in partially purified 100

and highly purified Ca 2+-pump ATPase preparation. ...
r-..
.c

Effects ofpurified IP] receptor on plasma membrane Ca z+-

~
E
80
rl-
pump ATPase '"
(3
E
::L 60
As shown in Fig. 8A, IP) receptor purified from porcine aorta
inhibited the highly purified plasma membrane Ca 2+-pump Q)

g:
(J)
ATPase activity in a concentration-dependent manner. The 40
inhibitory action was observed with very low concentrations
ofIP) receptor, suggesting a high affinity interaction between
!;;: r-+- rl-
c.
the IP) receptor and the plasma membrane Ca2+-pumpATPase.
The inhibition may constitute the basis for a slight, but definite
increase in ATPase activity observed when IP) receptor was
.§'?-
N
co
20

(J
specifically removed by immunoprecipitation from the
partially purified Ca2+-pump ATPase preparation (Fig. IB).
o
Phosphorylation of the IP) receptor by cGMP kinase did not Control 10 la
alter the inhibitory effect of the protein on the Ca 2+-pump
ATPase (Fig. 8B). Fig. 6. Effects of type In and type 113 cGMP kinase on activity of plasma
membrane Ca'·-pump ATPase. Effects of type In cGMP kinase purified
from porcine lung and type 113 cGMP kinase purified from porcine aorta on
...
,......
..s::::
the activity of partially purified plasma membrane Ca'·-pump ATPase were
examined. The Ca'·-pump ATPase activities were determined at the free
"-Cl
20 Ca'· concentration of 1 I!M with or without type In cGMP kinase (In) or
E type 113 cGMP kinase (113), each 10 I!g/ml plus I I!M cGMP. Each value
"-
(5
represents the mean ± S.E. of three experiments.

E 15
::L

r-+-
r+- Discussion
Q)
(I)
ell
10
The most unexpected finding in the present study is the
r+-
CL
I- ,-+- ,-t-
« stimulation by cGMP kinase of the partially purified plasma
0.
membrane Ca 2+-pump ATPase from which 240-kDa cGMP
E 5 kinase substrate protein/IP) receptor was specifically removed
::J
0.
I
by imrnunoprecipitation with an antibody (APISY-I) specific
+
'"ell to type lIP) receptor. With SDS-PAGE coupled with highly
U 0 sensitive silver-staining and immunostaining no IP) receptor
Control cGMP kinase 1 cGMP kinase 2 was detected in the supernatant. Thus, the idea that phos-
(10 I! g/ml) (10 I! g/ml) phorylation of IP) receptor constitutes an essential step in
intact ?eat :- cGMP (-) cGMP (+)
stimulation of plasma membrane Ca 2+-pump ATPase by
mactlvated cGMP kinase was not substantiated.
Fig. 5. Requirement cf intact, activated form of cGMP kinase for cGMP
Stimulation by cGMP kinase ofCa2+-pump ATPase devoid
kinase stimulation of plasma membrane Ca'·-pump ATPase. The highly of 240-kDa protein is at variance with our previous finding
purified plasma membrane Ca2+-pump ATPase devoid of IP 3 receptor was that cGMP kinase failed to stimulate the purified Ca2+-pump
used at a concentration of 10 nM. The Ca2+-pump ATPase activity was ATPase [11]. However, recent studies conducted with purified
determined at the free Ca2+ concentration of I I!M. cGMP kinase 1 represents
erythrocyte Ca2+-pump ATPase [23, 24] demonstrated that
a purified type In cGMP kinase prepared by the method in which the
final step of removal of tightly bound cGMP was omitted, and were added
plasma membrane Ca 2 +-pump ATPase took two forms
at 10 l!g1ml along with I I!M cGMP. Heat-inactivation was carried out by depending on the concentrations, i.e. a monomeric form
heating the purified cGMP kinase at 100°C for 5 min. cGMP kinase 2 sensitive to stimulation by calmodulin at lower concentrations
represents a purified type In cGMP kinase preparation from which the and an oligomeric, activated form at higher concentrations
bound cGMP was removed to restore the sensitivity ofthe enzyme to cGMP
that is insensitive to stimulation by calmodulin. The plasma
activation as described in the Materials and methods section. cGMP was
added at I I!M when indicated. Each value represents the mean of two or
membrane Ca2+-pumpATPase purified in the present study
three experiments, each performed in duplicate. with a combination of calmodulin- and concanavalin
164

IP,R· • • •

PMCA: •
•

-Auto G

~Dye
front
2 3 4 5 6 2 3 4 5 6

A. Silver-staining 8. Autoradiogram

Fig. 7. Phosphorylation of partially purified and highly purified plasma membrane Ca'·-pump ATPase preparations by cGMP kinase. The partially
purified and the highly purified plasma membrane Cal> -pump ATPase samples, each containing 1.2 l!g of protein, were phosphorylated by incubation at
37°C for 10 min with 10 l!g/ml of purified type la or type 113 cGMP kinase plus I l!M cGMP, and separated under denaturing conditions on a 6% SDS-PAGE
gel as described in the Materials and methods section. A silver-stained gel is shown in A and its autoradiograin in B. Lanes I and 2, type la and type 113
cGMP kinase alone, respectively; lancs 3 and 4, partially purified Ca'·-pump ATPasc plus type Ia and type 113 cGMP kinase respectively; lanes 5 and 6,
highly purified Ca'·-pump ATPase plus type la and type 113 cGMP kinase, respectively. The position of autophosphorylated cGMP kinase subunit (80 kDa)
was indicated by Auto-G. PMCA - plasma membrane Ca'·-pump ATPase; IPJR -IP J receptor; la - type la cGMP kinase; 113 - type 113 cyclic GMP kinase.
A weakly phosphorylated protein indicated by open triangle was the 138-kDa cGMP kinase substrate, a protein distinct from plasma membrane Ca'·-pump
ATPase [13]. Incubation of either partially purified or highly purified Ca'·-pump ATPase in the absence of cGMP kinase did not produce any phosphorylated
bands (data not shown).

A-affinity chromatographies was found to behave in the occurs within the first two exons. Type lex and type I~ cGMP
same way; the activity and calmodulin sensitivity of the kinases are, therefore, identical in COOH-terminal sequence
enzyme depended on the concentrations of enzyme in assay of about 600 amino acids which constitutes most part of the
mixtures. As the concentration ofthe Ca2+-pumpATPase used kinases, while they differ in their NH 2-terminal region of
in our previous study was high, being 40 nM, a concentration about 100 amino acids where only 36% of the first 103 amino
at which the enzyme took the oligomeric form, we reexamined acids in type I~ are identical to those in type lex [27]. The
the stimulation of purified Ca 2+-pump ATPase by cGMP NH 2-terminal region contains dimerization domains at the
kinase using a wide range of concentrations of the enzyme extreme amino terminus and autophosphorylation/auto-
and found that the purified Ca 2+-pump ATPase of vascular inhibitory domain just carboxyl terminal to the dimerization
smooth muscle could be stimulated by cGMP kinase when domain, while two cGMP-binding domains, a catalytic
the concentration of ATPase was low, while stimulation was domain and a carboxyl-terminal domain ofunknown function
no longer observed at 40 nM in agreement with the results are located in the predominant part common to both isozymes
of our previous study [II]. [27]. The two isozymes share many properties in common
Vascular smooth muscle expresses roughly equal quantities but differ in kinetic properties and sensitivities to cyclic
of two isozymes of type I cGMP kinases, namely lex and I~ nucleotide analogues, implying different roles of the two
[20, 25, 26]. The two isozymes are products of alternative isozymes in cGMP signaling pathways. On the basis of a
splicing ofthe primary transcript from a single gene. Splicing comparison of sensitivities of the two isozymes to various
 165

A B
14 f- -

--rl-
,-..
.c ,-.. 12 -
"-OJ
20 "-
.c
E "-
"- OJ
10 -

+
E -
rl-_
(5
E 15 "-
(5
:::l.
E
(1)
:::l. 8 f-
l/)
ro 10 (1)
0-
-
l/)
f- ro 6
<{ 0-
f-
Q <{
E 5
:J
'?-
Q
E 4 f- -
:J
"'ro '?-
u 0
+
ro -
N

0 5 10 U
2 ,-
IP 3 R (nM)
o
Control

Fig. 8. Effects of non-phosphorylated and phosphorylated IP J receptor on plasma membrane Ca'+-pump ATPase. The highly purified plasma membrane
Ca'+-pump ATPase deprived of IP J receptor was used at a concentration of 10 nM. The Ca'+-pump ATPase activity was determined at the free Ca'·
concentration of I ~M. (A) The ATPase activity was determined in the presence of indicated concentration of IP J receptor (IP JR) purified from porcine
aorta. (B) The ATPase activity was determined in the presence or absence of 4 nM purified IP J receptor subunit either mock-phosphorylated (IPJR) or
phosphorylated (IP JR-P) by type let cGMP kinase. Phosphorylation of IP J receptor was carried out by incubation at 37°C for 30 min with I0 ~g/ml of
type let cGMP kinase plus I 11M cGMP as described in the Materials and methods section and directly added to the assay mixture for Ca'+-pump ATPase
activity. Mock-phosphorylated IPJ receptor was prepared in a similar way to the phosphorylated IP J receptor but without ATP. The final concentration of
cGMP kinase carried over into the ATPase assay mixture was 1.6 Ilg/ml, at which the stimulatory effect of the kinase was presumed to be minimum (not
shown).

cyclic nucleotide analogues and potencies for production of the Ca2+-pumpATPase is not a substrate for cGMP kinase [10,
relaxation of the pig coronary artery, Sekhar et al. [26] 12, 13]. In the present study phosphorylation was not
hypothesized that type Ia cGMP kinase mediates the observed even under conditions optimized for detection of
relaxation of vascular smooth muscle by cGMP elevation. cGMP kinase substrates.
Our present finding that only type Ia but not type l~ cGMP Notwithstanding, in the present study intact, activated form
kinase stimulates the plasma membrane Ca 2+-pump ATPase of type Ia cGMP kinase was necessary for the stimulation
is in accord with the hypothesis. ofCa2+-pumpATPase. Presumably, a certain site ofthe cGMP
Smooth muscle cells express PMCA 1b isoform as a kinase would be exposed on the surface due to a con-
predominant isoform of plasma membrane Ca 2+-pump formational change induced through binding of cGMP to
ATPase [28-30].Among about 20 isoforms so far identified, its binding sites on the kinase and the site thus exposed may
the PMCA 1b is the only isoform that contains the consensus serve as an interaction site(s) with the plasma membrane
sequence for cyclic nucleotide-dependent protein kinase Ca2+-pump ATPase. Alternatively, such an interaction site
[31]. With human erythrocyte Ca 2+-pump ATPase [32] would be generated by autophosphorylation of the kinase.
phosphorylation of this site by cAMP-dependent protein The resultant interaction between the two proteins may lead
kinase (cAMP kinase) and concomitant stimulation of to stimulation of Ca 2+-pump ATPase. The lack of the
Ca 2+-pump ATPase was demonstrated. Since there is a stimulating effect oftype I~ cyclic GMP kinase suggests the
considerable overlap [33] as regards substrate specificity involvement ofthe N-terminal region oftype Ia cGMP kinase
between cAMP kinase and its closely related counterpart, as a site for the interaction, where the structure of the two
cGMP kinase, it may be reasonable to assume that not only isozymes differ greatly as discussed above.
cAMP kinase but also cGMP kinase would phosphorylate the Detailed analysis of oligomerization process conducted
consensus site and lead to stimulation of the Ca 2+-pump with isolated erythrocyte Ca2+-pumpATPase revealed that the
ATPase. However, it has been consistently demonstrated that calmodulin binding domain of the enzyme is the site for
166

oligomerization [34]. The possible interaction site between 8. Imai S, Yoshida Y, Sun, H-T: Sarcolemmal (Ca 2++Mg2+)-ATPase of
the activated type lex cGMP kinase and the Ca 2+-pump vascular smooth muscle and the effects of protein kinases thereupon.
J Biochem (Tokyo) 107: 755-761, 1990
ATPase may be the calmodulin binding domain of the 9. Furukawa K-I, Nakamura H: Cyclic GMP regulation of the plasma
ATPase. membrane (Ca2+_Mg2+) ATPase in vascular smooth muscle. J Biochem
Interestingly a low concentration of type I IP3 receptor, a (Tokyo) 101: 287-290,1987
major subtype ofthe vascular smooth muscle IP 3 receptor, was 10. Vrolix M, Raeymaekers L, Wuytack F, Hofmann F, Casteels R: Cyclic
shown in the present study to inhibit the plasma membrane GMP-dependent protein kinase stimulates the plasmalemmal Ca'+-
pump of smooth muscle via phosphorylation of phosphatidylinositol.
Ca 2+-pump ATPase, suggesting a high affinity interaction Biochem J 255: 855-863, 1988
between the IP3 receptor and the Ca2+-pumpATPase.Although II. Yoshida Y, Sun H-T, Cai J-Q, Imai S: Cyclic GMP-dependent protein
this subtype of IP 3 receptor has been proved to be a good kinase stimulates the plasma membrane Ca'+-pumpATPase of vascular
substrate for cGMP kinase in vitro and in vivo [14-17], the smooth muscle via phosphorylation ofa 240-kDa protein. J Bioi Chern
inhibition was observed not only with the phosphorylated 266: 19819-19825, 1991
12. Baltensperger K, Carafoli E, Chiese M: The Ca'+-pumpingATPase and
receptor but also with non-phosphorylated one. the major substrates of the cGMP-dependent protein kinase in smooth
In conclusion, reconstitution experiments conducted with muscle sarcolemma are distinct entities. Eur J Biochem 172: 7-16, 1988
highly purified proteins indicate that type lex but not type I~ 13. Yoshida Y, Cai J-Q, Imai S: Plasma membrane Ca 2+-pump ATPase is
cGMP kinase can stimulate the plasma membrane Ca2+-pump not a substrate for cGMP-dependent protein kinase. J Biochem (Tokyo)
ATPase of vascular smooth muscle directly without any III: 559-562,1992
14. Koga T, Yoshida Y, Cai J-Q, Islam MO, Imai S: Purification and
detectable phosphorylation of the enzyme when the ATPase characterization of240-kDa cGMP-dependent protein kinase substrate
is in monomeric form. The idea that phosphorylation of of vascular smooth muscle. J Bioi Chern 269: 11640-11647, 1994
240-kDa protein/IP3 receptor by cGMP kinase constitutes an 15. Islam MO, Yoshida Y, Koga T, Kojima T, Kangawa K, Imai S: Isolation
essential step for stimulation ofplasma membrane Ca 2+pump and characterization of vascular smooth muscle inositol 1,4,5-
ATPase by cGMP kinase was not substantiated. trisphosphate receptor. Biochem J 316: 295-302, 1996
16. Komalavilas P, Lincoln TM: Phosphorylation of the inositol 1,4,5-
trisphosphate receptor by cyclic GMP-dependent protein kinase. J Bioi
Chern 269: 8701-8707, 1994
Acknowledgements 17. Komalavilas P, Lincoln TM: Phosphorylation of the inositol 1,4,5-
trisphosphate receptor: Cyclic GMP-dependent protein kinase mediates
cAMP and cGMP dependent phosphorylation in the intact rat aorta. J
We are grateful to Dr. H. Kawakatsu for his generous gift of Bioi Chern 271: 21933-21938,1996
monoclonal antibody (P32). We thank A. Mitomi and H. 18. Kosk-Kosicka D, Scaillet S, Inesi G: The partial reactions in the catalytic
Sakurai for their skillful technical assistance and H. Arakakawa cycle of the calcium-dependent adenosine triphosphatase purified from
for her help in preparing the manuscript. This work was erythrocyte membranes. J BioI Chern 261: 3333-3338, 1986
19. Corbin JD, Doskeland SO: Studies of two different intrachain
supported in part by a Grant-in-Aid for Scientific Research cGMP-binding sites ofcGMP-dependent protein kinase. J Bioi Chern
(B) to S.1. (No. 08457026). 258: 11391-11397,1983
20. Wolfe L, Corbin JD, Francis SH: Characterization ofa novel isozyme
of cGMP-dependent protein kinase from bovine aorta. J Bioi Chern
264: 7734-7741, 1989
References 21. Paterson GL: A simplification of the protein assay method of Lowry et
al. which is more generally applicable. Anal Biochem 83: 346-356, 1977
I. Waldman SA, Murad F: Cyclic GMP synthesis and function. Pharmacol 22. Laemmli UK: Cleavage of structure proteins during the assembly of
Rev 39: 163-194, 1987 the head of bacteriophage T4. Nature 227: 680-685, 1970
2. Warner TD, Mitchell JA, Sheng H, Murad F: Effects of cyclic GMP 23. Kosk-Kosicka D, Bzdega T: Activation of the erythrocyte Ca'+-ATPase
on smooth muscle relaxation. Adv Pharmacol 26: 171-194, 1994 by either self-association or interaction with calmodulin. J Bioi Chern
3. Lincoln TM, Cornwell TL: Intracellular cyclic GMP receptor proteins. 263: 18184-18189, 1988
FASEB J 7: 328-338, 1993 24. Kosk-Kosicka D, Bzdega T, Johnson JD: Fluorescence studies on
4. Lincoln TM, Komalavilas P, Cornwell TL: Pleiotropic regulation of calmodulin binding to erythrocyte Ca2+-ATPase in different olig-
vascular smooth muscle tone by cyclic GMP-dependent protein kinase. omerization states. Biochemistry 29: 1875-1879, 1990
Hypertension 23(part 2): 1141-1147,1994 25. KeilbachA, Ruth P, Hofmann F: Detection of cGMP dependent protein
5. Furukawa K-I, Tawada Y, Shigekawa M: Regulation of the plasma kinase isozymes by specific antibodies. Eur J Biochem 208: 467--473,
membrane Ca 2+-pump by cyclic nucleotides in cultured vascular 1992
smooth muscle cells. J BioI Chern 263: 8058--8065, 1988 26. Sekhar KR, Hatchett RJ, Shabb JB, Wolfe L, Francis SH, Wells IN,
6. Popescu LM, Panoiu C, Hinescu M, Nutu 0: The mechanism of Jastorff B, Butt E, Chakinala MM, Corbin, JD: Relaxation of pig
cGMP-induced relaxation in vascular smooth muscle. Eur J Pharmacol coronary arteries by new and potent cGMP analogs that selectively
107: 393-394,1985 activate type la, compared with type I~, cGMP-dependent protein
7. Rashatwar SS, Cornwell TL, Lincoln TM: Effects of 8-bromo-cGMP kinase. Mol Pharmacol 42: 103-108, 1992
on Ca2+levels in vascular smooth muscle cells: Possible regulation of 27. Francis SH, Corbin JD: Progress in understanding the mechanism
Ca2+-ATPase by cGMP-dependent protein kinase. Proc Natl Acad Sci and function of cyclic GMP-dependent protein kinase. Advances
USA 84: 5685-5689, 1987 Pharmacol 26: 115-170, 1994
 167

28. De Jaegere S, Wuytack F, Eggermont JA, Verboomen H, Casteels, R: 31. Carafoli E: Biogenesis: Plasma membrane calcium ATPase: 15 years
Molecular cloning and sequencing of the plasma-membrane Ca'+-pump of work on the purified enzyme. FASEB J 8: 993-1002, 1994
of pig smooth muscle. Biochem J 271: 655--660,1990 32. James PH, Pruschy M, Vorherr TE, Penniston JT, Carafoli E: Primary
29. Khan I, Grover AK: Expression of cyclic nucleotide-sensitive and structure of the cAMP-dependent phosphorylation site of the plasma
-insensitive isoforms of the plasma membrane Ca'+-pump in smooth membrane calcium pump. Biochemistry 28: 4253-4258, 1989
muscle and other tissues. Biochem J 277: 345-349, 1991 33. Francis SH, Corbin JD: Structure and function of cyclic nucleotide-
30. Hammes A, Oberdorf S, Strehler EE, Stauffer T, Carafoli E, Vetter H, dependent protein kinases. Annu Rev Physiol 56: 237-272, 1994
Neyses L: Differentiation-specific isoform mRNA expression of the 34. VorherrT, KesslerT, Hofmann F, Carafoli E: The calmodulin-binding
calmodulin-dependent plasma membrane Ca'+-ATPase. FASEB J 8: domain mediates the self-association of the plasma membrane Ca'+-
428-435, 1994 pump. J Bioi Chern 266: 22-27,1991
Molecular and Cellular Biochemistry 190: 169-177, 1999.
© 1999 Kluwer Academic Publishers.

Chemical modification of an arginine residue in the

ATP-binding site of Ca2+-transporting ATPase of
sarcoplasmic reticulum by phenylglyoxal*

Hisanori Yamamoto and Masao Kawakita

Department of Physiological Chemistry, The Tokyo Metropolitan Institute of Medical Science (Rinshoken), Bunkyo-ku,
Tokyo, Japan

Abstract
Phenylglyoxal (PGO) was used as a reagent for chemical modification of the ATP-binding site of Caz+-transporting ATPase
of rabbit skeletal muscle sarcoplasmic reticulum (SR-ATPase). When I mM PGO was reacted with SR-ATPase at 30°C at
pH 8.5, PGO was bound to the ATPase molecule in two-to-one stoichiometry with concomitant loss of activity of the ATPase
to form the phosphorylated intermediate (E-P). ATP andADP prevented the binding ofPGO and thereby protected the enzyme
from inactivation. The SR membranes were labeled with [14C]PGO and then digested with pepsin to identify the attachment
site ofPGO. A 14C-labeled peptide (4ozlIe-Arg*-Ser-Gly-Gln406) was purified to homogeneity by Cis-reversed phase HPLC (Arg*
denotes the binding site of [14C]PGO). These results indicate that Arg403 is located in the ATP binding site of the SR-ATPase.
(Mol Cell Biochem 190: 169-177,1999)

Key words: ATPase, ATP-binding site, Ca 2+-transport, phenylglyoxal, sarcoplasmic reticulum

Introduction pursuit of the role of Ca 2+ and of the mediators of the

Ca 2+-effects since then, and his pioneering works along these
The Caz+-transporting ATPase of rabbit skeletal muscle lines were, in fact, forerunners of the calcium era in which
sarcoplasmic reticulum (SR-ATPase) is an intrinsic mem- we are still living [5].
brane protein that is most abundant in the sarcoplasmic He had already left the field ofSR-ATPase long before we
reticulum membrane. It couples the energy derived from entered the field to elucidate the molecular mechanisms
hydrolysis of ATP to the transport of Caz+ ions across the underlying the energy coupling between Ca 2+-transport and
membrane against concentration gradient (see ref. [I] for a ATP splitting [6], and unfortunately, we did not have an
recent review). The finding, made by Dr. Ebashi in early opportunity of intimately collaborating with him. Never-
1960s, that Ca z+ions were transported from the cytoplasmic theless, his devotion to the pursuit of scientific reality and
side to the luminal side of the membrane concomitant with his warm personality have continuously inspired and
ATPhydrolysis [2, 3] was the long sought missing link toward encouraged us greatly during the course of our own struggle,
the understanding of the nature of the muscle relaxing factor for which we feel deeply indebted to him.
and the physiological role of Caz+ ion in the regulation of We have been attempting in these years to understand the
muscle contraction [4]. Dr. Ebashi went further on along the molecular mechanisms ofCaz+-transport: [6-14]. Elucidation
ofthe structure ofATPase is an indispensable step toward this
understanding. A structural diagram of the ATPase molecule
*Dedicated to Dr. Setsuro Ebashi who pioneered the field of Ca'+-mediated based on the primary structure and a secondary structure
regulation of physiological processes. prediction has been proposed [15], and a structural analysis

Address for offPrints: M. Kawakita, Department of Physiological Chemistry, The Tokyo Metropolitan Institute of Medical Science (Rinshoken), 3-18-22
Honkomagome, Bunkyo-ku, Tokyo 113-8613, Japan
170

based on cryoelectron microscopy has provided us with a site, and the results on the identification of the target site
rough sketch of its molecular shape [16], but our knowledge of this reagent.
about the actual folding of the peptide particularly in the
cytoplasmic domain is still quite limited.
Active site mapping by means of site-specific chemical Materials and methods
modification of the ATPase molecule would be a useful
approach toward the understanding of spatial arrangement of Materials
the ATPase peptide and can be complementary to that based
on site-directed mutagenesis. The amino acid residues Sarcoplasmic reticulum (SR) membranes were prepared
known to be located in the ATP binding domain include from rabbit skeletal muscles as previously described [6].
ASP351 [17], Lys492 [12, 18-20], Lys515 [21,22], Thr532.533 [23] Phenylglyoxal was purchased from Aldrich (Milwaukee,
and LYS684 [24, 25]. ASP351 is phosphorylated by interacting USA), and [7_ 14C]PGO from Amersham (Buckinghamshire,
with y-phosphoryl moiety of ATP. Adenosine triphos- UK). Pepsin was obtained from Wako Pure Chemicals (Tokyo,
phopyridoxal (AP 3PL) reacted specifically but differently Japan), and ODS-120T column from Tosoh (Tokyo, Japan).
with Lys492 and LYS684 in the absence and presence of Ca 2+
[12]. The distance between the y-phosphoryl group of the
ATP moiety of AP 3PL and its 4-formyl group, the active Chemical modification with PGO
functional group that react with Lys492 and LYS684' is rather
small. Therefore ASP351 and LYS492/684 are likely located very The reaction mixture contained, unless otherwise specified,
close to each other. Proximity of Lys492 and Arg678 was also 40 mM HEPES-KOH (pH 8.5), 0.1 M KCl, 5 mM MgCI 2,
suggested based on the results of intramolecular cross-linking 50 11M CaCI2, SR membranes at I or 3 mg protein/ml and
[26]. LYS515 that is the FITC (fluorescein isothiocyanate)- I mM nonradioactive or [14C]PGO (4 cpm/pmol) as specified
reactive Lys residue is also supposed to be a part of the ATP in appropriate figure legends. The mixture was incubated
binding site [21,22]. The knowledge of the locations of these for 40 min at 30°C and the reaction was stopped by dilution
amino acid residues on the linear map has significantly with 40 mM HEPES-KOH (pH 8.5).
improved our understanding on how the ATPase peptide is
folded, since these residues must be actually placed to form
a rather narrow space that would accommodate an ATP Pepsin digestion
molecule [23]. For further active site mapping, we tried to
determine other amino acid residues located in the ATP SR membranes labeled with [14C]PGO ([14C]PGO-SR) were
binding site. digested by pepsin (1/100, w/w) in 1% HCOOH (pH 2) at
An Arg residue was also proposed to be a part of the ATP 30°C for 60 min according to the method ofWakabayashi et
binding site, based on the fact that 2,3-butanedione inhibited al. [30]. The solution was then centrifuged at 70,000 x g for
SR-ATPase activity in borate buffer and that the inhibition 30 min. A clear supernatant containing 14C-labeled peptides
was prevented by ATP [27]. We therefore tried to identify the was recovered, filtered through Ultra Free C3 (Millipore,
reactive Arg residue in the ATP-binding site of SR-ATPase. Bedford, USA), and the peptides in the filtrate were ready for
However, 2,3-butanedione did not seem to be the reagent of purification by HPLC.
choice, if the target site were to be determined, since the
resulting addition compound was too unstable in the absence
Purification of labeled peptides by reversed phase HPLC
of borate to allow its identification. Phenylglyoxal (PGO)
is another arginine-specific reagent, and a preliminary HPLC was carried out by using a Tosoh WPM HPLC pump
report on its application to SR-ATPase appeared recently connected to a Shimadzu SPD-6A UV monitor (Shimadzu,
[28]. It reacts with the guanidino group of arginine residue Kyoto, Japan). PGO-labeled peptides were detected by radio-
at pH 7.0-9.0 in a manner similar to 2,3-butanedione, but activity due to 14C-label and absorbance of PGO at 250 nm
the resulting derivative is stable under acidic conditions (£250= 11000). A Tosoh ODS-120Tcolumn (4 x 250 mm) was
below pH 4 [29]. This would allow us to obtain a peptide used throughout the study. Peptides were eluted at a flow rate
fragment with PGO remaining attached to an Arg residue, if of 0.5 ml/min.
suitable care were taken. Furthermore, radioactive [14C]PGO
is commercially available, thus enabling us to trace PGO-
labeled peptides by radioactivity. Characterization of t 4
C}PGO-Iabeled peptides
In an attempt to further extend active site mapping we
describe in this paper the characteristics of the reaction of Amino acid analysis was performed by the PITC method [31]
PGO with SR-ATPase, which is directed to the ATP binding after hydrolysis in 6 N HCI containing 0.1 % phenol for 3 h
 171

at 130°C. PITC-amino acids were detected by absorbance

at 254 om by using a Shimadzu reversed phase column STR 100
ODS-H at 38°C connected to a Shimadzu SPD-6A UV
monitor and a C-R6A Chromatopac data analyzer. z
o
~
~ 50
Other methods
II:
oLL
Protein concentration was determined by the method of 0-
Lowry et ai. [32], using bovine serum albumin as a standard. W
E-P forming activity and the amount of [14C]PGO bound to o'-( I<....L-_...L---'--..JL.-....L:::o~

the SR membranes were determined as described previously o 0.2 0.5 2.0 8.0
[33]. [PHENYLGL YOXAL] (mM)

Fig. 2. Effect of the concentration of PGO on inactivation of SR-ATPase.

SR membranes (I mg protein/ml) were incubated with PGO at various
Results concentrations as indicated for 10 min at pH 8.5, and then the E-P forming
activity was measured. The E-P forming activity in the absence ofPGO (2.9
Effect of PGO on E-P formation by the SR membranes nmol/mg protein) was taken as 100%.

SR membranes were incubated with I mM PGO at 30°C at

Effects of adenine nucleotides and inorganic phosphate
pH 7.5, 8.0 and 8.5, and then the activity to form the E-P
on the inactivation of E-P formation by PGO
intermediate was measured. At pH 8.5, PGO inactivated the
E-P forming activity most effectively, resulting in about
SR membranes were treated with PGO in the presence of
80% inactivation in 30 min (Fig. I). We therefore ex-
various concentrations ofATP andADP, and the E-P forming
amined the reaction between PGO and SR-ATPase at pH 8.5
activity was measured (Fig. 3). Addition of ATP and ADP
in more detail. SR membranes were incubated with various
offered a marked protection on the E-P forming activity.
concentrations of PGO at 30°C for 10 min at pH 8.5. The
This suggested that PGO was directed to the nucleotide
E-P forming activity was decreased by PGO treatment in a
binding site of the ATPase molecule, and thus was excluded
dose-dependent manner as shown in Fig. 2. At concen-
from the site in the presence of excess ATP and ADP.
trations 4 mM or above the activity was almost completely
PGO-treatment was also carried out in the presence of Pi,
lost in 10 min.
since Pi was known to interact with SR-ATPase at the
phosphorylation site in the absence of Ca 2+ to form the E-P
intermediate [34]. Figure 4 shows that 10 mM Pi in the

100
100 rI f - - - - , - - - - , - - - - r - - ,
~
~
Z
0 z
i= o
<{
~
II:
50 i=
<{
~
50 •
0 II:
LL
0-,
oLL
UJ 0-
W
30 60 90 o '-I j'----"-----..I...----'------'

TIM E (min) o 0.2 2 20

[NUCLEOTIDE] (mM)
Fig. I. pH dependence of inactivation of SR-ATPase by PGO-treatment.
SR membranes (I mg protein/ml) were incubated with I mM PGO at pH 7.5, Fig. 3. Effect of ATP and ADP on inactivation ofSR-ATPase by PGO. SR
8.0, and 8.5, and then the E-P forming activity was measured at the times as membranes (I mg protein/ml) were incubated with 1 mM PGO in the presence
indicated. The values ofthe E-P forming activity in the absence ofPGO (2.6, of various concentrations of ATP (0) or ADP (e) for 60 min at pH 8.5, and
2.4 and 2.2 nmollmg protein at pH 7.5 (D), pH 8.0 (0) and pH 8.5 ce), then the E-P forming activity was measured. The E-P forming activity in the
respectively) were taken as 100%. absence ofPGO (3.2 nmol/mg protein) was taken as 100%.
172

absence of Ca 2+protected SR-ATPase almost completely

against the inactivation by I mM PGO. The result is con- -100
• •
'::R.
~
sistent with the idea that PGO is interacting with an amino >-
acid residue that lies in the vicinity of the phosphorylation l- 80
S;
site, ASP351' i=
Q
<t
LU
Differential effects of PGO on the E-P forming and the >
i= •
Ca 2+ -transporting activities of the SR membranes in the <t 20
presence of adenyl-5'-yl imidodiphosphate
.....J
LU
c::
•
(AMP-P(NH)P) 0
0 20 40 60 80 100
SR membranes were treated with 1 mM PGO in the presence TIM E (min)
of 6 mM AMP-P(NH)P, and the E-P forming and the
Ca 2+-transporting activities were measured (Fig. 5). The E-P Fig. 5. Differential effect ofPGO-treatment in the presence of AMP-P(NH)P
forming activity as well as theATPase activity (data not shown on the Ca 2+-transporting and the E-P forming activities ofSR membranes.
and ref. [28]) was very well protected by AMP-P(NH)P, as SR membranes (I mg protein/ml) were incubated with 1 mM PGO in the
presence of6 mM AMP-P(NH)P at pH 8.5. Aliquots were withdrawn from
were in the presence of ATP and ADP as described above.
the reaction mixture at the times as indicated, and the Ca 2+-transporting (.)
In contrast, the Ca2+-transporting activity was not, and was and the E-P forming (e) activities were measured. The Ca 2+-transporting
lost rapidly. Thus, PGO-treatment of the SR membranes in and the E-P forming activities in the absence ofPGO (1.4 J.Imol/min/mg,
the presence of AMP-P(NH)P led to an uncoupling of ATP 1.8 nmol/mg protein, respectively) were taken as 100%.
splitting from Ca 2+-transport. Uncoupling as a result of
chemical modification of SR membranes with N-(3- 6 mM AMP-P(NH)P. At the times as indicated in Fig. 6A,
pyrene)maleimide [14, 35] and maleic anhydride [36] was aliquots of the sample were assayed for the E-P forming
also reported. activity and for the amount of PGO covalently bound to the
membranes. PGO was bound to SR-ATPase rapidly in the
absence of AMP-P(NH)P with concomitant decrease in the
Measurement ofthe amount ofPGO bound to SR-ATPase E-P forming activity. AMP-P(NH)P at 6 mM concentration
protected SR-ATPase from inactivation effectively, and at the
SR membranes (3.0 mg/ml) were incubated with 1 mM same time substantially decreased the binding of PGO.
[14C]PGO at 30°C, at pH 8.5 in the presence or absence of Fraction of PGO that was prevented from binding to the SR
membranes in the presence of AMP-P(NH)P was assumed to
represent the one directed specifically to theATP-binding site,
100 and the amount of PGO bound specifically to this site was
~ estimated from Fig. 6A. Figure 6B shows the relationship
~ between the amount of this specifically bound PGO and the
Z
0 remaining E-P forming activity. The extrapolated value of 14.2
i= nmol PGO bound per mg SR membranes for the complete
~ 50 inactivation corresponds to 2 mol ofPGO per mol ofATPase,
a: assuming that ATPase of Mr. 110,000 accounts for 70% of
0
u..
a..
• the SR proteins. Taking into account that a stoichiometry of
2 moles of PGO to 1 mole of Arg residue was proposed for
uJ • this modification reaction [29], this result was consistent with
o L...f J ' - - ' - - - - ' - - - - ' - - - - ' - - ' the idea that the stoichiometric binding of PGO to an Arg
o 0.01 0.1 1 10 residue in the catalytic site of the SR-ATPase molecule is
[PHOSPHATE] (mM) responsible for the inactivation. We therefore tried to
identify the binding site of PGO, which was rendered
Fig. 4. Effect of inorganic phosphate on inactivation of the E-P forming inaccessible in the presence of AMP-P(NH)P and ATP, on
activity by PGO. SR membranes (I mg protein/ml) were incubated with 1 mM
the ATPase peptide. The fraction ofPGO bound to the site(s)
PGO in the presence of various concentrations of phosphate as indicated
for 30 min at pH 8.5 in the presence of 1 mM EGTA with no externally added
not protected by AMP-P(NH)P, that constitutes about a half
Ca 2+. After the incubation the E-P forming activity was measured. The E-P of the total, may be involved in the selective inactivation of
forming activity in the absence ofPGO (2.0 nmol/mg protein) was taken as the Ca2+-transporting activity as described in the preceding
100%. section.
 173

-e
50 a.
to dryness by using a rotary evaporator, and the peptides were
dissolved in a small volume of distilled water. An aliquot ofthe
Cl
sample was applied to a TSK ODS-120T column (4 x 250 mm)
.E
(5 and eluted with a 20-60% CH 3 CN gradient in 0.1 % tri-
E
25 -
c:: fluoroacetic acid (TFA), and PGO-labeled peptides were
Cl detected by absorbance at 250 nm (Fig. 7). When the peptic
Z peptides obtained from the samples labeled in the presence
::J
oco and absence ofATP were compared, a major difference was
o<!) noted in a peak (Peak a) eluted at about 35% CH 3CN. The
L- ----L -'----' 0 0-
material in this peak was radioactive and markedly diminished
in a sample labeled in the presence of ATP. This strongly
30 60
TIM E (min)
suggested that the peak contained a peptide which was derived
from the ATP-binding site and thus carried a PGO-modified
Arg residue. We therefore saved this peak fraction for
100 B further purification.

z
o Further purification of [I4C]PGO-labeled peptide
i=
<t 50
~
a: The radioactive peptide in 'Peak a' was further purified by
ou.
using the same column with 18% CH3CN in 5 mM phos-
0-
W phate buffer (pH 6.8) containing 20 mM Na 2 S0 4 as an
0
0 5 10 15
AMP-P(NH)P-PROTECTABLE
PGO BINDING (nmol/mg prot.)

0.032
A
Fig. 6. Relationship between the amount of [14C]PGO bound to the SR
membranes and inactivation of the E-P forming activity. (A) Time course
of the binding ofPGO. SR membranes (3 mg protein/ml) were incubated o
."
with I mM ['4C]PGO in the presence (e,.) or absence (0, D) of 6 mM N
c(
AMP-P(NH)P at pH 8.5. At the times as indicated, aliquots of the reaction
mixtures were withdrawn and assayed for the E-P forming activity (e, 0)
and the amount of['4C]PGO bound to the SR membranes (., D). The E-P O ~ - m --:j

forming activity in the absence ofPGO (2.5 nmol/mg protein) was taken as
100%. (B) Relationship between the inactivation ofE-P forming activity and a
the amount of['4C]PGO bound to the AMP-P(NH)p-protectable site on the B
SR-ATPase. Remaining E-P forming activity in (A) on the ordinate was

-]";
0.032
plotted against the difference between the amount of PGO bound in the
absence and the presence of AMP-P(NH)P at corresponding times as
indicated in (A) on the abscissa.
o
."
N
-- - -
c(

Comparison of the peptides derived from SR-ATPase

o 20
labeled with [I4C]PGO in the presence and absence of
ATP on pepsin-digestion o 20 40 60
RETENTION TIME (min)
SR membranes were labeled with ['4C]PGO (4 cpm/pmol)
for 40 min in the presence and absence of 10 mMATP at 30°C Fig. 7. Analysis of peptic peptide fragments obtained from the PGO-treated
at pH 8.5. The labeled membranes were digested with pepsin SR membranes on reversed phase HPLC. SR membranes (3 mg protein/ml)
(l: 100, w/w) for 1 h at 30°C in 1% HCOOH [30]. The were incubated with I mM ['4C]PGO in the presence (A) or absence (B) of
membranous remnants were precipitated by centrifugation, 10 mM ATP for 30 min at pH 8.5. The labeled membranes were digested with
pepsin (1:100, w/w) for 60 min at 30°C, and then the peptides were
and the supernatants, containing more than 80% of the total fractionated by HPLC on ODS-120T. The samples were eluted with 20-60%
4
1 C-radioactivity originally attached to the membranes
gradient of CH,CN in 0.1 % TFA, and the elution of the peptides was
labeled in the presence and absence of ATP, respectively, monitored by absorption at 250 nm to detect PGO-labeled materials
were saved. The ['4C]PGO-labe1ed peptides were evaporated efficiently.
174

eluent (Fig. 8A), and subsequently, with 26% CH 3CN

containing 0.1 % TFA (Fig. 8B). Finally, as shown in Fig. 8B, A B
a radioactive peptide which showed optical absorption at 1.92 0.24
both 220 and 250 nm was obtained. Radioactivity and ab-
sorption at 250 nm indicated that this is a ['4C]PGO-Iabeled
peptide.

1.28 0.16
o
Analysis of the labeled peptide and identification of the CII
CII
target-site of PGO C

Table I shows the amino acid composition of the [14C]PGO- 0.64 0.08
labeled peptide. The amount ofPGO attached to the peptide
subjected to the amino acid analysis was estimated to be
approximately 50 pmol based on its radioactivity. Assuming
that 2 mol of PGO was attached to I mol of Arg, the result o
0.08
Uri
II
o
0.01
1-, I
11.1,
presented in the table implies that I mol of the peptide II I I U'
oII) i
contained I mol each of Glx, Ser, Gly, Ala, and Arg. The
II
.:: III ,!'~ .. ""\
I I' \
R' I Ill.. ~ \ ,-.l,.
CII I
amino acid analysis did not explicitly show the presence of C
o "'J ,
4 \'. \ ,

JU ~
o J " ......,.----
a PGO-modified Arg residue, because PITC-PGO-Arg could >
not be detected by the present analysis [37]. Of all the t: 400 200
>
pentapeptides derivable from SR-ATPase a pentapeptide ~E
corresponding to a stretch from lle402 to G In406 is uniquely
(Ja.
CU
0-
...,..rn.
o l-...JllIlJlllIll1J:;L.....-..JJ o
consistent with the amino acid composition of the purified oC o 20 40 o 20
peptide, and the result indicated that PGO was selectively TIME (min) TIME (min)
a:
reacted withArg403 under the condition used for the chemical
modification. Fig. 8. Further purification of the PGO-Iabeled peptide. (A) The 'Peak a'
fraction obtained by HPLC from the SR membranes labeled in the absence
of ATP as in Fig. 7B was applied to an ODS-120T column equilibrated with
Discussion 18% CHJCN in 5 mM sodium phosphate buffer (pH 6.8) containing 20 mM
Na,S04' and then eluted with the same solution. Elution of peptides with
and without PGO-Iabel was monitored by absorption at 220 and 250 nm and
We have demonstrated in this paper that phenyl glyoxal by radioactivity in 200-111 aliquots. (B) The radioactive peptide recovered
(PGO) interacted with SR-ATPase and caused a loss of the from the HPLC as illustrated in A was applied to the same column equilibrated
with 26% CHJCN in 0.1 % TFA, and then eluted with the same solution.
E-P forming activity of the ATPase. The inactivation was
prevented by the presence ofATP orAMP-P(NH)P, indicating
that PGO was interacting at theATP-binding site and that an
arginine residue was located in that site. Our present result by PGO further suggested the proximity of the target site
indicating that Pi protected the ATPase against the inactivation to the phosphorylation site. Covalent modification by PGO
was not completely prevented byATP andAMP-P(NH)P, but
about a half of the total binding was unaffected by the
Table I. Amino acid composition of the purified PGO-Iabeled peptide. substrate and its analog. In this context, it is noteworthy that
Amino acid analysis was carried out as described under Materials and the Ca 2+-transporting activity ofSR membranes was severely
methods. The amount of the peptide used was 25 pmol as estimated from impaired even in the presence of6 mMAMP-P(NH)P, while
the radioactivity. Note that PGO-modifiedArg is undetectable by the present
the ATPase activity as well as the E-P forming activity was
analysis [37).
maintained. Uncoupling of the ATP splitting from the Ca2+-
Recovery transport accompanying chemical modification with N-
Amino acid pmol Relative (nearest integer) (3-pyrene)maleimide [14, 35] and maleic anhydride [36] was
Glx 14.7 0.79(1)
previously reported. The molecular basis for the uncoupling
Ser 17.2 0.93 (I) is intriguing, but we have not investigated this issue in further
Gly 27.5 1.47(1) detail in the present work.
lie 18.7 1.0 (1) PGO was successfully used in a number of cases for
402 * 406 chemical modification of Arg residues in proteins [37--41].
-lie Arg Ser Gly Gln- The product of the reaction between PGO and an Arg residue,
 175

which contains two PGO moieties per guanidino group, is residues were also assumed to be in the close vicinity of the
sufficiently stable particularly under acidic conditions phosphorylation site of SR-ATPase based on the structure
below pH 4 to allow subsequent separation and analysis of of adenosine triphosphopyridoxal, the reagent used in the
the derivatized peptides. Thus, in the cases ofchroloplast H+- affinity labeling experiments. Collectively, these results
ATPase [42] and carboxypeptidase B [43], PGO-reactive indicate that ASP351' Lys492' LYS684 and Arg 403 are located
sites were identified, but there are many cases in which the close together in the ATP binding site of SR-ATPase, and
target site of PGO remains unidentified yet. This is partly may constitute the site for the interaction of phosphoryl
because an appropriate measure for obtaining PGO-Iabeled moiety of ATP with the ATPase. Positive charges on these
peptide fragments has not been found in those cases. The basic amino acids may be important for the catalytic
PGO derivative is unstable at neutral and alkaline pH where activity or may contribute to the stabilization ofATP in the
many proteolytic enzymes work optimally, and therefore the binding site. It is important that a segment containing Arg403
derivative tended to be decomposed during proteolytic was shown for the first time to be involved in the con-
digestion. In fact, digestion ofPGO-labeled SR membranes struction oftheATP-binding site. Site directed mutagenesis
with either trypsin (at pH 8.5) or thermolysin (at pH 8.0) did of groups thus shown to lie in the ATP-binding site by
not yield a specific PGO-Iabeled peptides in our hands (data means of chemical modification would lead to a better
not shown). understanding oftheir functional significance, as was the case
Bond et ai. obviated this difficulty by using V8 protease with Lys684' which was previously shown to be selectively
and pepsin which were active under acidic conditions [38]. modified by adenosine triphosphopyridoxal in the presence
The latter was also successfully used by Wakabayashi et of Ca 2+ and later shown to be functionally important by
ai. to digest SR-ATPase labeled with 7-chloro-4-nitro- mutagenesis [25].
benzoxadiazole (NBD-Cl) in 1% HCOOH, for obtaining Unfortunately, amino acid residues exposed in other parts
NBD-Cllabeled peptide fragments [30]. Although pepsin is oftheATP-binding site have been only poorly characterized
not usually preferred because of its rather broad specificity, except that LYS515' that is highly reactive with FITC, is
the procedure was proved to be well suited for our purpose supposed to be located close to the adenine-binding portion
of preserving the PGO-Iabel during proteolytic digestion of of the ATP-binding site. More recently Thr 532 and Thr 533
the labeled SR membranes. Digestion of ['4C]PGO-Iabeled were shown to be labeled with 8-azidoADP [23]. However,
SR at 3 mg protein/ml in 1% HCOOH with pepsin (1/1 00, results with 8-azidoadenine nucleotide anologs were
w/w) at 30°C for 60 min effectively released peptide somewhat variable in the literature. Thus, the ATPase protein
fragments including the PGO-Iabeled ones in a soluble form, was not photolabeled significantly or labeled only poorly in
with little [14C]PGO being detached free from the target site some cases [44, 45]. In the latter case 8-azido ADP did not
during the digestion. A single peptide with attached PGO seem to specifically label any SR proteins [45]. During the
label, that was specifically masked by ATP and remained course of the active site mapping study we also examined the
unlabeled in its presence, was identified with HPLC and effect of8-azidoadenosine triphosphopyridoxal (N 3-AP3PL),
eventually obtained in a pure form. Amino acid analysis a photoreactive bifunctional ATP analog, to probe the part
revealed that the peptide fragment repre<;ented the 402lle- of the ATP-binding site distal to the phosphorylation site.
Arg-Ser-Gly-Gln406 stretch of the SR-ATPase molecule, and N3APlL was found attached to Lys492 through its pyridoxal
Arg 403 was concluded to be the PGO-reactive site. The moiety after treatment with NaBH 4. In contrast, irradiation
conclusion is clear-cut, though not consistent with the by UV-light did not lead to significant covalent attachment of
previous suggestion by Corbalan-Garcia et ai. [28] that N3-APlL via adenine moiety of the ATP analog (Yamamoto
labeling on the B-tryptic fragment could be protected by ATP. H., unpublished observation). The result is consistent with
Occurrence of an arginine residue in theATP-binding site those suggesting inefficient labeling with 8-azido ATP as
of SR-ATPase is consistent with previous reports by others cited above, and may imply that the adenine-binding portion
[27, 28], but the target-site of chemical modification by PGO of the ATP-binding pocket is readily accessible to ,vater, in
has not so far been identified. Now we have demonstrated for contrast to the situation with the phosphorylation site at the
the first time that the target-site ofPGO is Arg403 . According other end of the pocket, that has to be inaccessible to water
to the secondary structure prediction Arg403 is supposed to before conformational changes accompanying the E1-P to
be involved in the ~-sheet structure that can be placed in the E2-P transition occur.
neighbor of ASP351' the phosphorylation site. Such an Development of additional reagents that can be reacted
arrangement seems likely since Arg 403 is supposed to be quantitatively with amino acid residues that line the ATP-
located close to the Pi-binding site, as the protection by Pi binding pocket, particularly its adenine- and ribose-binding
against the inactivation by PGO suggested. portions and thus can be utilized in probing for these residues
We have reported previously that Lys492 and LYS684 con- would certainly help us to deepen our understanding on the
stitute a part of the ATP-binding site [12, 24]. These Iysyl mechanisms of SR-ATPase.
176

Acknowledgement 15. Brandl W, Green NM, Korczak B, MacLennan DH: Two Ca'+ ATPase
genes: Homologies and mechanistic implications of deduced amino
acid sequences. Cell 44: 597-607, 1986
This work was supported in part by Grants-in-Aid for 16. Toyoshima C, Sasabe H, Stokes DL: Three dimensional cryo-electron
Scientific Research from the Ministry of Education, Science, microscopy of the calcium ion pump in the sarcoplasmic reticulum
Sports and Culture of Japan. membrane. Nature 362: 469-471, 1993
17. Bastide F, Meissner G, Fleischer S, Post RL: Similarity of the active site
of phosphorylation of the adenosine triphosphatase from transport of
sodium and potassium ions in kidney to that for transport of calcium
References ions in the sarcoplasmic reticulum of muscle. J Bioi Chern 248: 8385-
8391,1973
1. Inesi G, Lewis D, Nikic D, Hussain A, Kirtley ME: Long-range 18. Yamasaki K, DaihoT, KanazawaT: 3'-0-(5-fluoro-2,4-dinitrophenyl)-
intramolecular linked function in the calcium transport ATPase. Adv ATP exclusively labels Lys-492 at the active site of the sarcoplasmic
EnzymolMolBiol65: 185-215, 1992 reticulumCa'+-ATPase. J Bioi Chern 269: 4129-4134,1994
2. Ebashi S: Calcium binding and relaxation in the actomyosin system. 19. Yamagata K, Daiho T, Kanazawa T: Labeling oflysine 492 with pyridoxal
JBiochem48: 150-151,1960 5'-phosphate in the sarcoplasmic reticulum Ca'+-ATPase. Lysine 492
3. Ebashi S, Lipmann F: Adenosine triphosphatase-linked concentration residue is located outside the fluorescein 5-isothiocyanate-binding
of calcium ions in a particulate fraction of rabbit muscle. J Cell Bioi 14: region in or near theATP binding site. J Bioi Chern 268: 20930-20936,
389-400, 1962 1993
4. Ebashi S, Endo M: Calcium and muscle contraction. Prog Biophys Mol 20. Mcintosh DB, Woolley DG, Berman MC: 2',3'-0-(2,4,6-trinitrophenyl)-
Bioi 18: 123-183,1968 8-azido-AMP and -ATP photo label Lys492 at the active site of
5. Ebashi S: Relaxing factor, sarcoplasmic reticulum and troponin-A his- sarcoplasmic reticulum Ca'+-ATPase. J Bioi Chern 267: 5301-5309,
torical survey-. In: S. Fleischer, Y Tonomura (eds). Structure and Function 1992
ofSarcoplasmic Reticulum. Academic Press, Orlando, 1985, pp 1-18. 21. Mitchinson C, WilderspinAF, Trinnaman BJ, Green NM: Identification
6. Kawakita M, Yasuoka K, Kaziro Y: Selective modification of functionally of a labelled peptide after stoichiometric reaction of fluorescein
distinct sulfhydryl groups of sarcoplasmic reticulum Ca'., Mg'+-adeno- isothiocyanate with the Ca'+-dependent adenosine triphosphatase of
sine triphosphatase with N-ethylmaleimide. J Biochem 87: 609-617, 1980 sarcoplasmic reticulum. FEBS Lett 146: 87-92, 1982
7. Kawakita M, Yasuoka-Yabe K, Saito K, Imamura Y: Conformational 22. Pick U, Bassilian S: Modification of the ATP binding site of the
changes of Ca'., Mg'+-adenosinetriphosphatase as detected by site- Ca'+-ATPase from sarcoplasmic reticulum by fluorescein isothiocyanate.
specifically introduced fluorescent and paramagnetic probes and by FEBS Lett 123: 127-130, 1981
susceptibility to tryptic digestion. In: S. Fleischer, Y. Tonomura (eds). 23. Lacapere J-J, Garin J, Trinnaman B, Green NM: Identification of amino
Structure and Function of Sarcoplasmic Reticulum. Academic Press, acid residues photolabeled with 8-azidoadenosine 5'-diphosphate in
Orlando, 1985, pp 63-89 the catalytic site of sarcoplasmic reticulum Ca-ATPase. Biochemistry
8. Kawakita M, Imamura Y, Yamamoto H, Suzuki S, Kawato S: Structural 32:3414-3421,1993
basis for the molecular mechanisms of the calcium-transportingATPase 24. Yamamoto H, Tagaya M, Fukui T, Kawakita M: Affinity labeling of the
of the sarcoplasmic reticulum. In Y. Mukohata (ed). New Era of ATP-binding site of Ca'+-transporting ATPase of sarcoplasmic
Bioenergetics.Academic Press, Tokyo, 1991, pp 47-72 reticulum by adenosine triphosphopyridoxal: Identification of the
9. Yasuoka-Yabe K, Tsuji A, Kawakita M: Studies on conformational reactive Iysyl residue. J Biochem 103: 452--457, 1988
transitions of Ca'., Mg'+-adenosine triphosphatase of sarcoplasmic 25. Vilsen B, Andersen JP, Maclennan DH: Functional consequences of
reticulum. II. Conformational characteristics of stabilized reaction alterations to amino acids located in the hinge domain of the Ca'+-
intermediates as revealed by fluorescent and paramagnetic probes. J ATPase ofsarcoplasmic reticulum. J Bioi Chern 266: 16157-16164,1991
Biochem 94: 677-688, 1983 26. Mcintosh DB: Glutaraldehyde cross-links Lys-492 andArg-678 at the
10. Saito-Nakatsuka K, Yamashita T, Kubota I, Kawakita M: Reactive active site of sarcoplasmic reticulum Ca'+-ATPase. J Bioi Chern 267:
sulfhydryl groups of sarcoplasmic reticulum ATPase. I. Location of a 22328-22335,1992
group which is most reactive with N-ethylmaleimide. J Biochem 101: 27. Murphy AJ: Arginyl residue modification of the sarcoplasmic reticulum
365-376, 1987 ATPase protein. Biochem Biophys Res Commun 70: 1048-1054, 1976
11. Imamura Y, Kawakita M: Purification of limited tryptic fragments of 28. Corbalan-Garcia S, Teruel JA, Gomez-Fernandez JC: Chemical mod-
Ca'., Mg'+-adenosine triphosphatase of the sarcoplasmic reticulum ification of Ca'+-ATPase from sarcoplasmic reticulum with phenyl-
and identification of conformation-sensitive cleavage sites. J Biochem glyoxal. Biochem Soc Trans 22: 381 S, 1994
105:775-781,1989 29. Takahashi K: The reaction of phenylglyoxal with arginine residues in
12. Yamamoto H, Imamura Y, Tagaya, M, Fukui T, Kawakita M: Ca'+- proteins. J Bioi Chern 243: 6171-6179,1968
dependent conformational change of the ATP-binding site of Ca2+_ 30. Wakabayashi S, Imagawa T, Shigekawa M: Does fluorescence of 4-
transporting ATPase of sarcoplasmic reticulum as revealed by an nitrobenzo-2-oxa-1 ,3-diazole incorporated into sarcoplasmic reticulum
alteration of the target-site specificity of adenosine triphospho- ATPase monitor putative E,-E, conformational transition? J Biochem
pyridoxal.J Biochem 106: 1121-1125, 1989 107:563-571,1990
13. Imamura Y, Kawakita M: Location of the Ca'+-transport site ofCa2+- 31. Heinrikson RL, Meredith SC: Amino acid analysis by reverse-phase
transporting ATPase of the sarcoplasmic reticulum as determined by high performance liquid chromatography: Precolumn derivatization with
analysis of paramagnetic interaction between Gd 3+ ions bound at the phenylisothiocyanate. Anal Biochem 136: 65-74,1984
transport site and membrane-embedded nitroxide spin probes. J 32. Lowry OH, Rosebrough NJ, FaIT AL, Randall RJ: Protein measurement
Biochem 110: 220-225, 1991 with Folin phenol reagent. J Bioi Chern 193: 265-275, 1951
14. Suzuki T, Kawakita M: Sites oflabeling with N-(3-pyrene)maleimide on 33. Yasuoka K, Kawakita M, Kaziro Y: Interaction of adenosine-5'-0-(3
Ca'+-transporting ATPase of the sarcoplasmic reticulum. J Biochem thiotriphosphate) with Ca'., Mg'+-adenosine triphosphatase of
117:881-887,1995 sarcoplasmic reticulum. J Biochem 91: 1629-1637,1982
 177

34. Kanazawa T, Boyer PD: Occurrence and characteristics of a rapid 40. Raess BU, Record DM, TunnicliffG: Interaction of phenylglyoxal with
exchange of phosphate oxygens catalyzed by sarcoplasmic reticulum the human erythrocyte Ca2 ++ Mg2+-ATPase. Evidence for the presence
vesicles. J Bioi Chern 248: 3163-3171, 1983 ofan essential arginyl residue. Mol Pharmacol27: 444-450, 1985
35. Suzuki T, Kawakita M: Uncoupling of ATP Splitting from Ca 2+_ 41. Kasher JS, Alien KE, Kasamo K, Siayman CW: Characterization ofan
transport in Ca 2+-transportingATPase of the sarcoplasmic reticulum essential arginine residue in the plasma membrane H+-ATPase of
as a result of modification by N-(3-pyrene)maleimide: Activation ofa Neurospora crassa. J Bioi Chern 261: 10808-10813, 1986
channel with a specificity for alkaline earth metal ions. J Biochem 42. Viale AM, Vallejos RH: Identification of an essential arginine residue
114:203-209,1993 in the beta subunit of the chloroplastATPase. J Bioi Chern 260: 4958-
36. Shoshan-Barmatz: Chemical modification of sarcoplasmic reticulum. 4962,1985
Biochem J 240: 509-517, 1986 43. Werber MM, Sokolovsky M: Proceedings: Carboxypeptidase B:
37. Mohan PM, Basu A, Basu S, Abraham KI, Modak MJ: DNA binding Modification offunctional arginyl residues. Israel J Med Sci II: 1169-
domain of Escherichia coli DNA polymerase I: Identification of 1170,1975
arginine-841 as an essential residue. Biochemistry 27: 226-233, 1988 44. Shoshan-Barmatz V: Stimulation of Ca 2+ efflux from sarcoplasmic
38. Bond MW, Chiu NY, Cooperman BS: Identification ofan arginine reticulum by preincubation with ATP and inorganic phosphate.
important for enzymatic activity within the covalent structure of yeast BiochemJ247:497-504,1987
inorganic pyrophosphatase. Biochemistry 19: 94-102, 1980 45. Campbell KP, MacLennan DH: Labeling of high affinity ATP binding
39. Konishi K, Fujioka M: Chemical modification ofa functional arginine sites on the 53,000- and 160,000-dalton glycoproteins of the sar-
residue ofrat liver glycine methyltransferase. Biochemistry 26: 8496-8502, coplasmic reticulum with the photoaffinity probe 8-N J -[a- 32 P]ATP. J
1987 Bioi Chern 258: 1391-1394,1983
Molecular and Cellular Biochemistry 190: 179-184, 1999.
© 1999 Kluwer Academic Publishers.

Intrinsic inhibitor of inositoll,4,5-trisphosphate

binding

Masato Hirata, Masako Yoshida, Takashi Kanematsu and Hiroshi

Takeuchi
Department ofBiochemistry, Faculty ofDentistry, Kyushu University, Fukuoka, Japan

Abstract
Rat brain cytosol was applied to a heparin column and eluted with 0.9 M-NaCI. The total binding activity of [3H]inositol
1,4,5-trisphosphate to the eluate was increased about 6-fold compared with the original cytosol. When the eluate was mixed with
a flow-through fraction from the heparin column, however, the activity retumed to the original level, suggesting that the flow-through
fraction contained an inhibitory factor(s) which prevented the binding. The factor(s) was purified by sequential column chro-
matography using gel permeation, a hydrophobic gel, and finally, a hydroxylapatite gel. Silver staining of sodium dedecyl sulfate
gel electrophoresis of the sample thus purified showed a broad band located between the authentic molecular weight markers of
580 and 390 k. A carbohydrate staining method showed that the factor is a glycoprotein. (Mol Cell Biochem 190: 179-184, 1998)

Key words: inositol 1,4,5-trisphosphate, receptor, pleckstnin homology domain

Abbreviations: Ins( I,4,5)P3 - inositol I,4,5-trisphosphate; PH domain - pleckstnin homology domain

Introduction Ins( I,4,5)P3 -immobilized column which was synthesized as

an extension of a series of Ins(l,4,5)P 3 analogues [9-12].
D-myo-Inositol 1,4,5-trisphosphate (lns( I,4,5)P 3)1, a product Determination of the partial amino acid sequence of both
of the receptor-activated hydrolysis of phosphatidylinositol molecules revealed that the 85 kDa protein is phospholipase
4,5-bisphosphate, plays an important role as an intracellular C-OI (PLC-O I), whereas the 130 kDa protein is as yet uniden-
second messenger by mobilizing Ca 2+ from non-mito- tified. We localized the region of PLC-Ol responsible for
chondrial store sites [1]. Ins( 1,4,5)P 3 is metabolized by two binding Ins( I,4,5)P3 by molecular biological, peptide synthetic
known routes; One is dephosphorylation, catalyzed by and immunological techniques [13, 14]. More recently, we
Ins( I,4,5)P 3 5-phosphatase present in both the cytosol and cloned the cDNA encoding the 130 kDa protein and, the
membrane fraction of cells, the result being formation of predicted amino acid sequence, deduced that the molecule is
Ins(l,4)P 2 which is subsequently degraded to free inositol similar to PLC-Ol, but has no PLC catalytic activity [15, 16].
by other phosphatase activities [2]. The other is phos- During the course of these experiments we noticed that
phorylation of the 3-hydroxyl group ofIns(l,4,5)P3 by an there is an intrinsic inhibitory factor(s) in the rat brain
ATP-dependent kinase present in the cell cytosol, producing cytosolic fraction which prevents the binding ofIns(1 ,4,5)P 3
Ins(l,3,4,5)P 4 [3]. Three types of proteins have been to both the 85 kDa (PLC-O I) and 130 kDa proteins, as
identified as Ins(l,4,5)P 3-interacting macromolecules: discussed previously [6]. Furthermore, the factor(s) was
Ins(l,4,5)P3 receptors on the endoplasmic reticulum involved verified to be effective on the purified Ins(1 ,4,5)P3 receptor
in Ca 2+ release [4, 5] and two types of enzymes related to isolated from rat cerebellum. Therefore, it is possible that the
Ins(l,4,5)P 3 metabolism. factor(s) plays a regulatory role in Ca 2+ release from the
In this laboratory, two novel Ins( 1,4,5)P3 binding proteins, endoplasmic reticulum mediated by Ins(l ,4,5)P 3" In the
with a molecular mass of 130 and 85 kDa respectively, have present study, one of the inhibitor(s) ofIns(l,4,5)P 3 binding
been isolated from rat brain [6-8], with the aid of an was purified from rat brain cytosol, and characterized.

Address for offprints: M. Hirata, Department of Biochemistry, Faculty of Dentistry, Kyushu University, Fukuoka 812-8582, Japan
180

Materials and methods sample was applied to a TSK-HA-IOO column (0.5 x 7 cm)
mounted in a high pressure liquid chromatography system,
Preparation ofrat brain cytosol and was eluted with a linear gradient ofphosphate from 1 mM
to 150 mM, and finally with a stepwise increase to 300 mM.
Using Wister rats ofboth sexes, brains without the cerebellum Flow rate was I ml/min.
were homogenized in a glass homogenizer with a teflon pestle
in 3 vol of a solution containing 50 mM NaCI, 10 mM Hepes
buffer (pH 8.0), I mM EDTA, 2 mM NaN 3 and 10 mM Assay of[JHjlns(l,4,5)P 3 binding
2-mercaptoethanol (buffer A) supplemented with a mixture
ofprotease inhibitors (1 ~g/ml aprotinin, 1.25 ~g/ml pepstatin The binding protein was incubated with 1.3 nM [3H]Ins(l ,4,5)P3
A, 0.1 mM p-amidinophenylmethanesulfonyl fluoride and in a solution (0.45 ml) of 50 mM Tris-HCI buffer (pH 8.3), 1
10 ~M leupeptin). The cytosol fraction was obtained by mM EDTA and 0.2% Triton X-100 on ice for 15 min. After
centrifuging the homogenates at 130,000 x g for 60 min. the addition of 50 ~I of 10 mg/ml bovine y-globulin and 0.5
ml 000% polyethyleneglycol6000, the tube was centrifuged
at 15,000 rpm for 5 min. The precipitate was dissolved in 0.5
Heparin-immobilized resin affinity chromatography ml of0.1 N NaOH and then counted for radioactivity in a liquid
scintillation counter. Non-specific binding was assayed in the
Rat brain cytosol (-50 ml) was applied to a heparin- presence of! ~M cold Ins(l ,4,5)P3and subtracted from the test
immobilized column (1.5 x 9 cm) equilibrated with buffer A. values obtained.
After washing the column first with buffer A and sub-
sequently with buffer A containing 0.3 M NaCl, absorbed
proteins were eluted with a solution containing 0.9 M NaCl. Polyacrylamide gel electrophoresis and silver staining or
Flow rate was 1 ml/min. carbohydrate staining

Sodium dodecyl sulfate gel electrophoresis was performed

Gel filtration column chromatography in 6% polyacrylamide according to the method ofLaemmli
[17]. Silver staining was carried out according to the
The flow-through fraction from a heparin column was maker's protocol. For a carbohydrate staining, the gel was
concentrated to about 5 ml by ultrafiltration using a membrane immersed in 12.5% trichloroacetic acid for 30 min and
with 20 k-cut-off, and the concentrates were applied to a washed once with distilled water. After immersion in 1%
HiLoad 26/60 superdex 200 column (2.6 x 60 cm) equilibrated sodium periodide for 50 min and extensive washing with
with buffer C (0.3 M NaCI, 10 mM Hepes buffer, pH 7.2, 1 distilled water, the gel was incubated in Schiff solution in
mM EDTA, 2 mM NaN 3 and 10 mM 2-mercaptoethanol). the dark for 10 min. Finally the gel was developed in 0.5%
Flow rate was 2 ml/min. sodium metabisulfite.

Phenyl sepharose column chromatography Materials

Fractions obtained from gel filtration chromatography were [3H]Ins(l,4,5)P 3(specific radioactivity, 777 GBq/mmol) was
supplemented with powdered KCI to give a final concentration purchased from Du Pont-New England Nuclear (Boston,
of 3 M and the mixture was applied to a phenyl Sepharose MA, USA). Heparin-immobilized resin, phenyl Sepharose
column (1 x6 cm)equilibrated with bufferD (3 MKCI, 10mM resin and a prepacked HiLoad 26/60 superdex 200 column
Hepes buffer, pH 7.2,1 mM EDTA, 2 mM NaN 3 and 10 mM (2.6 x 60 cm) were from Pharmacia (Sweden). A TSK-HA-
2-mercaptoethanol). The column was washed with a linear 100 column was from Tosoh (Japan). All other reagents were
gradient of a KCI concentration from 3-0 M. Flow rate was of the highest grade available.
1.2 ml/min.

Results and discussion

Hydroxylapatite column chromatography
Presence ofinhibitor(s) for Ins(1,4,5)P 3 binding in the
Fractions obtained from phenyl sepharose column chroma- cytosol
tography were dialyzed against buffer E (1 mM ~HP04' 50
mMNaCI, 10mMHepes buffer, pH 7.2, 1 mMEDTA,2mM The total binding activity of the cytosol fraction obtained
NaN 3and 10 mM 2-mercaptoethanol). Following dialysis, the from ten rat brains was 2.1 pmol at 1.3 nM [3H]lns( 1,4,5)P3'
 181

2000 results indicate that there is an inhibitory factor(s) in the

cytosol fraction which passes through the column without
retention. Similar results were also obtained when the cytosol
~
0
fraction was applied to a heparin column. As shown in Fig.
I, the binding observed to the fraction eluted from the column
E Q.
by O.9M-NaCI was about 6-fold that observed with the
~
Cl
c
cytosol. The eluate showed decreased binding if mixed with
'is
c 1000 the flow-through fraction, indicating that the flow-through
:c fraction contains a factor(s) which inhibits binding. The
M
a. possibility that the factor(s) is a phosphate compound such
iii'
o:r:
~
as ATP or pyrophosphate must be excluded because the
VI inhibitor(s) could not be dialyzed through a membrane with
.=
~ 15 k-cut-off. Furthermore, chymotrypsin treatment of the
M
flow-through fraction destroyed the inhibitory activity,
0 indicating that the factor(s) is of a protein-origin. Therefore,
2 3 4 we used the flow-through fraction of the brain cytosol on a
Fig. I. [JH]Ins( I,4,5)PJ binding to fractions obtained from heparin column
heparin column as a starting material for purifying the
chromatography. The cytosol fraction from 10 rat brains was applied to a factor(s). It was thought that the factor(s) would be mainly
heparin column. The flow-through fraction and the fractions eluted by 0.9M- present in the cytosol because the total binding activity of a
NaCI were assayed for [JH]Ins(1,4,5)PJ binding at 1.3 nM. Fractions of 2M-NaCl eluate of a membrane extract from an Ins(l ,4,5)P 3
cytosol, flow-through and fractions eluted by 0.9M-NaCI were all adjusted affinity column was just twice as much as that of the original
to 50 ml in volume, and then 10 J.l1 of each fraction was assayed for binding
or inhibition. Column 1--4 represents the cytosol, flow-through, 0.9M-NaCI
extract (see Table 1 of Ref. [8]).
eluate and flowthrough plus 0.9M-NaCI eluate, respectively. Each value
was the mean of triplicate determinations, and three other preparations gave
essentially the same results. Putification ofinhibitory factor(s)

whereas the fraction eluted from an Ins( I,4,5)P3-immobilized The flow-through fraction of a heparin column was con-
resin by 2M-NaCl increased the total binding activity more centrated and then applied to a gel filtration column (Fig. 2).
than 30-fold, up to 66.5 pmol (see Table 2 of Ref. [6]). These Inhibitory activity was eluted between the authentic samples,

2000K
01
ISOK 66K C
2.0
t t 50
'6
:.8
M
0..
~
V)

~
o
CD 40 ~

..:=..
N
«
VI

Q 1.0
..,I ~

20 0
c
o
:.0
E
.S
O.O __------r-------.---------r---~
0
50 70 80 90 "*
Retention lime (min)

Fig. 2. Gel filtration chromatography of the flow-through fraction. The flow-through fraction from a heparin column was applied to a HiLoad 26/60
superdex 200 column, and fractions were collected for I min (about 2 ml). A hundred J.l1 of each fraction was assayed for the inhibition of ['H]Ins(1 ,4,5)P J
binding to 10 J.l1 of the 0.9M-NaCI eluate from a heparin column. Each point represents the mean of duplicate determinations. The graph shown is a typical
and more than ten trials afforded a similar pattern of inhibition. Subsequent graphs with respect to column chromatography were the same.
182

120 the binding activity of [3H]lns(l ,4,5)P3because of the de-

C:
.~ creased concentration during incubation on ice. This is un-
Q.
CI> 100 likely, however, because neither phosphatase nor kinase
.c.
activity was detectable in the fractions obtained in Fig. 2
£ l (results not shown). Furthermore, inhibitory activity was
O'l
c c 80
'6
c
.2
(j
attenuated when the concentration of binding proteins was
15 doubled, as shown in Fig. 3.
M
~ 60
a.. u The fraction indicated in Fig. 2 were fractionated on a
if) Q)
.D
phenyl Sepharose column using a linear decrease of KCI
""
~
<5
<J)
.D
40 concentration from 3-0 M. Each fraction was assayed for the
<J) III
C
c inhibition of [3H]lns(l,4,5)P 3 binding. As shown in Fig. 4,
M
I E
:J 20 three peaks of inhibitory action were observed. The fractions
"0 in the first peak, becuase they contained the least protein
6 0

concentration as assessed by absorbance at 280 nm, were then

a
a 50 100 150 200 applied to a hydroxylapatite column, and a single peak of
inhibitory activity was obtained by fractionation using a linear
Fraction (fd)
gradient of phosphate concentration (Fig. 5).
The final preparation of the sample was analyzed by
Fig. 3. Dose-dependence of the inhibitor(s). The 0.9M-NaCI eluate at 5 sodium dodecyl sulfate-polyacrylamide gel electrophoresis,
(e) or 50 f.Il (0), was assayed for binding in the presence of various volumes
followed by silver staining. As shown in Fig. 6A, staining
of flow-through fraction from a heparin column. Each point represents the
mean of duplicate determinations, and other experiments showed similar revealed a broad band at the top of the 6% gel. When a 4%
results. polyacrylamide gel was used, the silver stained band migrated
into the gel and was located between authentic molecular
weight markers of580 and 390k (results not shown). Because
blue dextran (2,000 k) and ~-amylase (200 k), indicating that of the broadness of the band, the protein inhibitor for
the factor(s) has a relatively high molecular weight. Ins( I,4,5)P3 binding was suspected to be a glycoprotein.
Metabolizing enzymes such as Ins( I,4,5)P3-5-phosphatase Carbohydrate staining verified the protein as being a glyco-
and Ins(l,4,5)P3-3-kinase, both of which are present in the protein as shown in Fig. 6. The purification steps employed
brain cytosol, could contribute to the apparent reduction in here are summarized in Table I.

I 0>
c
0.6 1::::;:':::::::1 60 '6
c 3M
:is
M
a..
in c
""- .9
0 0.4 40 v:; <0
CD ~
c c
Ie
C\J

- -
4)
~ 0
M c
Q (5 8
c (3
0.2 20 .g :::.:::
:is
:.c
..
.S
0....
M

0.0 0
0 10 20 30
Fraction number

Fig. 4. Hydrophobic column chromatography of inhibitor(s). Fractions having inhibitory activity, as indicated in Fig. 2, were collected, adjusted with
powdered KCI to 3 M-KCl and then applied to a phenyl Sepharose column. Fractionation was carried out using a linear gradient of3--o M-KCl, and fractions
were collected for 1 min (about 1.2 ml). Fifty f.Il of each fraction was assayed for inhibition.
 183

3.0,----------------. Table I. Purification of an inhibitor protein of Ins(l ,4,5)P3 binding. The

A 80 cytosol fraction from 10 rat brains was used as the starting material. The
0.3M
g> protein concentration required for 50% inhibition (IC so ) of[lH)Ins( I,4,5)P1
'6
c bmdmg to type I-Ins(1 ,4,5)P 3 receptor purified from rat cerebellum (4 ng
:0
60 cr protein) is shown.
..,- 2.0 u;: c
o ". .9
Fraction protein IC so purification
E "§
40
.s'"
}:: • c
'"u
C
0
u Cytosol
mg

129
Ilg

150
-fold

1
'0
1.0 c <3 Heparin Sepharose 30 4.1 36
.g Cl.
J: 2.7 2.2 68
HiLoad 26/60
"
20 :0
:.c Phenyl Sepharose 0.21 1.7 88
.s
~ Hydroxylapatite 0.010 0.2 750
0 OM
20 30 40

Fraction number for binding in both cases [13, 14, 16]. Although the tertiary
structure of the PH domain of the 130kDa protein has not
Fig. 5. Hydroxylapatite column chromatography ofinhibitor(s). Fractions yet been clarified, it is likely to be similar to that ofPLC-8 I,
indicated in Fig. 4 were collected and applied to a TSK-HA-I 00 column in because the tertiary structures of the PH domains from many
a high pressure liquid chromatography system. Each fraction was collected kinds of proteins, as assessed by NMR spectroscopic and/
{forO.5 min (0.5 ml)} and 20 III of each fraction was assayed for inhibition.
or crystallographic analysis, are reported to be similar [18-
22]. However, the structure of the region responsible for
Ins( I,4,5)P3binding in the receptor molecule whose binding
In the present study, we have isolated one of the intrinsic
was also destroyed by the factor isolated here, is probably
inhibitors of Ins( I ,4,5)P3binding from rat brain cytosol. As
different from that of the PH domain of the binding proteins.
to the mechamism underlying the inhibition, the following
Furthermore, structures other than the binding region would
could be postulated; (I) A reduction of the actual concentra-
also be different one another. Such speculation makes the
tion of [3H]Ins(l,4,5)P3 due to metabolism or binding to the
remaining possibilities also unlikely. At present time we
inhibitor molecule itself. (2) Competition with [3H]Ins( 1,4,5)P
suggest that the factor would act partially on several basic
for the binding site. (3) Allosterical inhibition (binding to a sit~
amino acids which are essential for binding and present in
different from the binding site and thus causing inhibition).
the binding region of PLC-8 I and the 130 kDa, protein as
The inhibitor molecule was unable to metabolize nor bind
well as in the Ins(l,4,5)P 3 receptor ([13, 14] and personal
to [3H]Ins( I,4,5)P3' indicating that possibility (I) is unlikely.
communications from Dr. K. Mikoshiba), but not on the
As to the site ofIns( I,4,5)P3binding of both PLC-8 I and the
binding pockets as a whole molecule.
130 kDa protein, both of which are similar in the primary
In the case that the factor is present inside cells, it is likely
sequence and have a pleckstrin homology (PH) domain, we
to be a negative regulator of Ca 2+ mobilization mediated by
clarified that the N-terminal of the PH domain is responsible
Ins(l,4,5)P 3 inside cells. Further work is obviously required
for clarification of the localization of the factor in cells and
which part of the molecule, carbohydrate or protein moiety,
Silver PAS is essential for inhibition.
staining staining

Acknowledgements

200K~
The authors express sincere thanks to Professor Setsuro
Ebashi. One of the authors (MH), as a graduate student of
Kyushu University, worked under the supervision of Prof.
Setsuro Ebashi in The University ofTokyo for 2.5 years from
117K~ 1976. Although MH was neither a good student nor made a
valuable contribution to the achievement of Prof. Ebashi' s
laboratory, he himselfleamed much from the atmosphere of
scientific excellence in Prof. Ebashi's laboratory. He believes
Fig. 6. Sodium dodecyl sulfate polyacrylamide gel electrophoresis. Six % that the atmosphere has influenced greatly his works in the
polyacrylamide was used. field of inositol I ,4,5-trisphosphate/Ca2+. The work was
184

financially supported by a grant-in-aid for scientific research II. Hirata M, Yanaga F, Koga T, Ogasawara T, Watanabe Y, Ozaki S:
from the Ministry of Education, Science, Sports and Culture Stereospecific recognition of inositol 1,4,5-trisphosphate analogs by
the phosphatase, kinase, and binding proteins. J Bioi Chern 265: 8404-
of Japan. 8407, 1990
12. Hirata M, Watanabe Y, Ishimatsu T, Yanaga F, Koga T, Ozaki S: Inositol
1,4,5-trisphosphate affinity chromatography. Biochem Biophys Res
References Commun 168: 379-386,1990
13. Yagisawa H, Hirata M, Kanematsu T, Watanabe Y, Ozaki S, Sakuma
K, Tanaka H, Yabuta N, Kamata H, Hirata H, Nojima H: Expression
I. Berridge MJ, Irvine RF: Inositol trisphosphate, a novel second and characterization of an inositol I,4,5-trisphosphate binding domain
messenger in cellular signal transduction. Nature 312: 315-321, 1984 ofphosphatidyl inositol-specific phospholipase C-ol. J BioI Chern 269:
2. Storey DJ, Shears SB, Kirk CJ, Michell RH: Stepwise enzymatic 20179-20188,1994
dephosphorylation of inositol 1,4,5-trisphosphate to inositol in liver. 14. Hirata M, Kanematsu T, Sakuma K, Koga T, Watanabe Y, Ozaki S,
Nature 312: 374-376, 1984 Yagisawa H: D-myo-Inositoll ,4,5-trisphosphate binding domain ofphos-
3. Irvine RF, LetcherAJ, Heslop JP, Berridge MJ: The inositol tris/tetrakis pholipase C-ol. Biochem Biophys Res Commun 205: 1563-1571,1994
phosphate pathway-demonstration ofIns(1 ,4,5)P J 3-kinase activity in 15. Kanematsu T, Misumi Y, Watanabe Y, Ozaki S, Koga T, Iwanaga S,
animal tissues. Nature 320: 631--634, 1986 Ikehara Y, Hirata M:A new inositol I,4,5-trisphosphate binding protein
4. Supattapone S, Worley PF, Baraban JM, Snyder SH: Solubilization, similar to phospholipase C-ol. Biochem J 313: 319-325, 1996
purification and characterization of an inositol trisphosphate receptor 16. Taketuchi H, Kanematsu T, Misumi Y, Yaakob BH, Yagisawa H, Ikehara
binding site. J BioI Chern 263: 1530-1534, 1988 Y, Watanabe Y, Tan Z, Shears SB, Hirata M: Localization of a high
5. Furuichi T, Yoshikawa S, Miyawaki A, Wada K, Maeda N, affinity inositol 1,4,5-trisphosphate/inositol 1,4,5,6-tetrakisphosphate
Mikoshiba K: Primary structure and functional expression of the binding domain to the pleckstrin homology module of a new 130 kDa
inositol I,4,5-trisphosphate binding protein P400. Nature 342: 32- protein: Characterization of the determinants of structural specificity.
38, 1989 Biochem J 318: 561-568, 1996
6. Kanematsu T, Takeya H, Watanabe Y, Ozaki S, Yoshida M, Koga T, 17. Laemmli UK: Cleavage of structural proteins during assembly of the
Iwanaga S, Hirata M: Putative inositol 1,4,5-trisphosphate binding head of bacteriophage T4. Nature 227: 680--685, 1970
protein in rat brain cytosol. J Bioi Chern 267: 6518--6525, 1992 18. Yoon HS, Hajduk PJ, Petros AM, Olejrnczak ET, Meadows RP, Fesik
7. Hirata M, Kanematsu T: Inositol I,4,5-trisphosphate-binding proteins SW: Solution structure ofa pleckstrin-homology domain. Nature 369:
in rat brain cytosol. In: J. Fain (ed.) Methods in Neurosciences, Vol. 672--675, 1994
18. Academic Press, 1993, pp 298-311 19. Macias MJ, Musacchio A, Ponstingi H, Nilges M, Saraste M, Oschkinat
8. Yoshida M, Kanematsu T, Watanabe Y, Koga T, Ozaki S, Iwanaga S, H: Structure of the pleckstrin homology domain from p-spectrin.
Hirata M: D-myo-inositol 1,4,5-trisphosphate-binding proteins in rat Nature 369: 675--677, 1994
brain membranes. J Biochem 115: 973-980, 1994 20. Downing AK, Driscoll PC, Gout I, Salim K, Zvelebil MJ, Waterfield
9. Hirata M, Sasaguri T, Hamachi T, Hashimoto T, Kukita M, Koga T: MD: Three-dimensional solution structure of the pleckstrin homology
Irreversible inhibition ofCa'- release in saponin-treated macrophages domain from dynamin. Curr Bioi 10: 884-891,1994
by the photoaffinity derivative of inositol I,4,5-trisphosphate. Nature 21. Ferguson KM, Lemmon MA, Schlessinger J, Sigler PB: Crystal
317: 723-725,1985 structure at 2.2 A resolution of the pleckstrin homology domain from
10. Hirata M, Watanabe Y, Ishimatsu T, Ikebe T, Kimura Y, Yamaguchi human dynamin. Cell 79: 199-209, 1994
K, Ozaki S, Koga T: Synthetic inositol trisphosphate analogs and their 22. Hyvonen M, Macias M, Nilges M, Oschkinat H, Saraste M, Wilmanns
effects on phosphatase, kinase, and the release of Ca". J Bioi Chern M: Structure of the bindingsite for inositol phosphate in a PH domain.
264: 20303-20308, 1989 EMBO J 14: 4676-4685, 1995
Molecular and Cellular Biochemistry 190: 185-190, 1999.
© 1999 Kluwer Academic Publishers.

Dynamic regulation ofintraceUular calcium signals

through calcium release channels

Masamitsu Iino
Department of Pharmacology, Faculty of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo; CREST, Japan Science
and Technology Cooperation

Abstract
After the seminal work of Ebashi and coworkers which established the essential role of the intracellular Ca2+ concentration
([Ca2+]) in the regulation of skeletal muscle contraction, we have witnessed an explosive elongation of the list of cell functions
that are controlled by the [Ca 2+l. In numerous instances, release of intracellular Ca 2+ stores plays important roles in Ca 2+
signalling which displays significant variation in spatio-temporal pattern. There are two families of Ca2+ release channels,
ryanodine receptors and inositol 1,4,5-trisphosphate (IP 3) receptors. These Ca2+ release channels are structurally and
functionally similar. In particular, the activity of both types of channels is regulated by the [Ca 2+]j' The [Ca 2+l dependence of
the Ca2+ release channel activity provides both types of channels with properties of a Ca2+ signal amplifier. This function of
the ryanodine receptor is important in striated muscle excitation-contraction coupling, whereas that of the IP 3 receptor seems
to be the basis of the generation ofCa 2+ waves. Thus the wide variety ofCa 2+ signalling patterns seem to be critically dependent
on the [Ca2+]j dependence of the Ca2+ release channels. (Mol Cell Biochem 190: 185-190, 1999)

Key words: calcium, calcium wave, calcium oscillation, inositol 1,4,5-trisphosphate receptor, ryanodine receptor,
excitation-contraction coupling

Introduction such is the [Ca 2+l increase of the eggs of Medaka, small fish
commonly found in rural creeks in Japan, during fertilization.
Only with swift command offinger strokes, can a pianist fully Immediately after the penetration of a sperm into a Medaka
convey his or her artistic expression to the audience. During egg, a [Ca 2+]j increase starts in the immediate vicinity of
each of the key strokes, which may occur at a rate exceeding the penetration site. This [Ca 2+]j increase then propagates
10 times per second, the intracellular Ca 2+ concentration beneath the cell membrane at a constant speed and reaches
([Ca2+]) of the skeletal muscle cells of the fingers and arms the opposite pole of the egg, which has a diameter of about
changes by several-ten-fold simultaneously throughout the 1 mm, within a minute or two (Fig. 1B) [3]. This propagating
entire length of the muscle fibers, which may span several Ca 2+ signal is called a Ca 2+ wave, and in Medaka eggs it
millimeters to several centimeters (Fig. lA). Owing to the triggers the formation of a fertilization membrane, which
pioneering work of Ebashi and coworkers during 1950's and prevents polyspermy, and the initiation of embryogenesis.
1960's, it is known that [Ca2+) increases are the key regulatory Mammallan eggs are smaller, but they show similar Ca2+
factors of skeletal muscle c1ontraction [1, 2]. Thus [Ca 2+]j waves upon fertilization [4, 5].
changes of skeletal muscle cells represent swift and These examples highlight the wide spectrum of spatio-
homogeneous signal transmission throughout the cell for temporal patterns found in different Ca2+ signalling mech-
initiation of contraction. anisms [6]. However, recent results suggest that a common
There are, however, many different patterns of Ca 2+ fundamental mechanism is involved in many store-mediated
signalling among different cell types. A classic example of Ca 2+ signalling mechanisms having quite different spatio-

Addressfor offprints: M. lino, Department of Pharmacology, Faculty of Medicine, The University ofTokyo, Bunkyo-ku, Tokyo 113, Japan
186

-10. 1 S
A

I -1 mm

o
B

D • I
-1()2s

Fig. I. Schematic representation ofCa 2• signalling in mammalian skeletal muscle (A) and Medaka egg (8). The [Ca H ]; is indicated by the degree of shading
(the darker the higher).

temporal patterns. Here, I would like to discuss the basic tail artery smooth muscle cells, it is about 20 Ilm/sec or
properties of the Ca2• signalling mechanisms which may be higher [7]. The spread of the [Ca 2+) change as the Ca 2+wave
common to both skeletal muscle cells and eggs, and to many cannot be due to passive diffusio~ of Ca 2+, because such a
other cell types. process will not propagate at a constant speed. Further-
more, the peak [Ca 2+l attained at a point within the cell
would decrease with an increase in the diffusion distance,
Ca2+ waves but the peak [Ca 2+]; remains constant during propagation of
the Ca2+ wave. Therefore, there must be a mechanism via
Cal. waves observed within a tissue which Ca 2+ release at the wave front is regeneratively
enhanced.
Ca 2+ waves have been detected in many types of cells
including eggs as described above, but usually in isolated
cells or cultured cells. To determine the physiological Ca l + wave and IP3 receptor
significance of the Ca2+ waves, it is important to determine
whether Ca2+ waves occur in cells within tissue, where the There are two families of intracellular Ca 2+ release channels,
cells are under the influence of neighboring cells. We have ryanodine receptors (RyRs) and inositoll,4,5-trisphosphate
therefore carried out a [Ca2+l imaging study using a segment receptors (IP3Rs), both of which can support Ca 2+ waves.
of an artery, the wall of which is composed chiefly of Ca 2+ waves mediated by the RyR will be considered in
endothelial cells, smooth muscle cells and sympathetic 'Functional differences between RyR subtypes'. In many cell
nerve cells. We showed that Ca 2+ waves indeed occur in types, including oocytes and smooth muscle cells, agonist
vascular smooth muscle cells within the arterial wall during stimulation that causes generation ofCa2+ waves also induces
perivascular sympathetic nerve stimulation [7]. Ca2+ waves synthesis ofIP3. Binding ofIP3 to the IP 3R induces opening
were also observed when noradrenaline was applied to the of the IP 3R as shown in Fig. 2A. Although this simple
smooth muscle cells in the arterial wall [7]. mechanism may be sufficient for mediation of a simple
Since Ca 2+ waves can be observed in the absence of increase in [Ca 2+]j' it alone cannot possibly explain the
extracellular Ca2+, Ca2+ is considered to be mobilized from complex spatio-temporal patterns such as those of Ca 2+
the intracellular Ca2+ stores, which are very likely located waves in agonist-stimulated cells. There must be more
in the endoplasmic reticulum. The wave velocity in many elaborate mechanisms via which the complex spatio-
types of cells ranges between 10 and 100 Ilm/sec [4]; in rat temporal patterns are generated.
 187

Coagonists ofIPfi membrane spanning regions [13]. These two receptors

are therefore likely to share a common ancestral proto-
During the study of the kinetics ofIP3-induced Ca2+release, type Ca 2+release channel [14].
I found that the IP 3R activity is sensitive not only to the IP 3 b) Both Ca 2+release channels are tetramers.
concentration but also to the cytoplasmic Ca2+concentration c) The RyR is known to function as a Ca 2+-induced Ca 2+
[8]. Indeed, we later found that the IP 3-induced Ca2+ release release (CICR) channel [15, 16]. CICR was discovered
rate depends biphasically on the Ca 2+concentration [9, 10]. independently by M. Endo in Professor Ebashi's
A Ca2+concentration change around 100 nM enhances IP 3- laboratory [17] and RJ. Podolsky [18]. As described in
induced Ca2+release. However, with a larger increase in the Ca2+ the previous section, the IP3R activity is sensitive to the
concentration around 111M, this enhancement effect is reduced. [Ca2+]j and also functions as a CICR. While RyR activity
The Ca 2+sensitivity of the IP 3-induced Ca2+release clearly is insensitive to [IP 3], the IP 3R functions as a CICR
indicates that IP3is only a partial agonist of the IP3R, and for channel only in the presence of IPr
the full activation of the IP 3R both Ca2+and IP 3 are required
[9-12] (Fig. 2B). In other words, IP3R can be regarded as a In excitable cells, depolarization of the cell membrane during
Ca2+signal amplifier, which induces release ofCa2+only upon an action potential is converted to a Ca2+signal which mediates
receipt of a Ca2+ signal when the cytoplasmic IP3 concen- regulation ofvarious cell functions. The first step in this signal
tration remains constant. The function of the IP3R as a Ca2+ transduction is mediated by a voltage-sensitive Ca2+channel,
signal amplifier seems to be the underlying mechanism of the such as the dihydropyridine receptor (DHPR) (Fig. 3a). In
generation of Ca2+waves. As shown in Fig. 2C, a local Ca 2+ mammalian cardiac muscle cells, the membrane depolarization
increase due to opening oflocal IP 3 receptors would induce first induces Ca2+influx into the cells through the DHPR, which
activation of neighbouring sites. A cascade of such local in turn triggers release ofCa2+from the SR through activation
activation would result in the generation of a Ca 2+wave. ofCICR via the RyR (RyR-2) [19]. This process provides a
Ca 2+signal amplification mechanism, and the resulting [Ca2+l j
change seems to become several-fold greater than that which
Membrane depolarization and Ca 2+ results from the Ca 2+influx alone (Fig. 38).
release In skeletal muscle cells, the depolarization-induced Ca 2+
signalling also involves both the DHPR and the RyR (RyR-I),
RyR as a Ca 2+ signal amplifier although Ca 2+ influx through the DHPR is not required for
functional coupling between the two molecules to occur
In both skeletal muscle cells and cardiac muscle cells, [20]. Thus, direct protein-protein interaction seems to
depolarization of the cell membrane results in Ca2+ release mediate transmission of the signal from the DHPR to the
from the intracellular Ca 2+stores, the sarcoplasmic reticulum RyR [21] (Fig. 3C). These E-C coupling mechanisms in
(SR). This process is referred to as excitation-contraction striated muscle cells indicate that the RyRs function, in
(E-C) coupling. The Ca 2+ release channel in these cells is collaboration with the DHPR, as Ca 2+ signal amplifiers. In
the RyR, which is both functionally and structurally similar cardiac muscle cells the amplification ratio is finite, while
to the IP 3R . The similarities between the RyR and IP 3R the skeletal muscle E-C coupling can be regarded to have
include the following. an infinite amplification factor since no Ca 2+ influx is
required for Ca 2+ release from the SR (Fig. 3).
a) There are regions of amino acid sequence identity Interestingly, in crustacean skeletal muscle cells, Ca2+influx
between the two receptors, especially in the putative through the DHPR triggers Ca2+release through the RyR as

A B

Fig. 2. Ca'+ dependence ofCa'+ release through IP) receptors. (A) The rate of IP)-induced Ca'+ release in the absence ofCa'+ is low (indicated by the thin
arrow). (8) Simultaneous activation by Ca 2+and IP) enhances the Ca'+ release through the IP) receptor (indicated by the thick arrow). (C) The cascade of
Ca'+-mediated enhancement of Ca'+ release results in the formation of a Ca'+ wave.
188

A B c
Ca 2 +

RyR subtype RyR-2 RyR-1

Amplification factor -10

Fig. 3. Ca'· signal amplification by ryanodine receptors. (A) Depolarization of the cell membrane induces opening of voltage-dependent Ca2+ channels,
and thus Ca'· influx into the cell. (B) Ca'· influx through the voltage-dependent Ca'· channels induces activation of the RyR, leading to amplification of the
Ca'· signal as observed in mammalian cardiac muscle cells. C. Direct protein-protein interaction induces the release ofCa'· from the SR through the RyR
in skeletal muscle cells.

in the case of mammalian cardiac muscle E-C coupling [22, been shown that RyR- I is capable of substituting for RyR-2
23]. During evolution, the coupling mechanism between the in the functional coupling with the DHPR [3 I]. Is then RyR-2
DHPR and RyR seems to have become increasingly specia- capable of substituting for RyR- I in skeletal muscle E-C
lized to facilitate the very fast skeletal muscle movement. coupling? We have attempted to answer this question
through expression studies in cultured skeletal muscle cells
of RyR- I-deficient mice generated by a gene-targeting
Functional differences between RyR subtypes technique [32]. The skeletal muscle cells of these RyR-
I-deficient mice do not exhibit Ca 2+ release from the SR
Three different RyR subtypes, with 67-70% amino acid upon membrane depolarization; that is, they do not exhibit
sequence identity between them, have been identified [24]. E-C coupling [32, 33]. Transfection with RyR-I eDNA
All of the subtypes (RyR- I, RyR-2 and RyR-3) function as restores the E-C coupling in these cells, but transfection
Ca 2+release channels, but they are differentially expressed RyR-2 eDNA does not [34]. This indicates that while RyR-I
among different tissues [24, 25]. RyR-I is dominantly can substitute for RyR-2, RyR-2 cannot replace RyR-I in
expressed in mammalian skeletal muscle cells. Although in skeletal muscle E-C coupling.
skeletal muscle cells of non-mammalian vertebrates RyR-I The RyR-2-cDNA-transfected skeletal myocytes show
and RyR-3 are expressed at almost equal levels [26,27] the spontaneous and repetitive generation of Ca2+waves, which
expression level ofRyR-3 in mammalian skeletal muscle cells we have never observed in RyR-I- or RyR-3-expressing
is extremely low [24]. Cardiac muscle cells express RyR-2. skeletal myocytes. As is the case in IP 3R-mediated Ca2+waves
RyRs are expressed in smooth muscle cells, although we still (see 'Coagonists of IP 3R' section), local [Ca 2+l-dependent
do not know which subtype is dominantly expressed. RyRs regenerative Ca2+release is though to underlie RyR-mediated
are also expressed in the central nervous system, exhibiting Ca2+waves [35]. Therefore, RyR-2 seems to be most prone to
differences in distribution pattern within the brain [25, 28]. regenerative release ofCa2+stores among subtypes ofthe RyR.
However, the functions of RyRs in the brain are poorly Such a property and resulting Ca2+waves may have some rele-
understood. RyRs might be expressed in non-excitable cells vance to the delayed after depolarization which occurs in Ca2+-
[29, 30], but further work is required for elucidation of the overloaded cardiac muscle cells which express RyR-2 [36].
functional significance of RyRs in such cells. RyR-3 also functions as a CICR channel, although an
Are there any functional differences among the RyR about 10-fold higher concentration of Ca 2+ is required for
subtypes? In cardiac muscle cells Ca 2+ influx through the activation of Ca2+release through RyR-3 as compared with
DHPR triggers CICR through RyR-2 [19] (Fig. 3B). It has that through RyR-I [33]. The physiological roles of RyR-3
 189

is still unclear. RyR-3-deficient mice grow and reproduce functions remain unknown. Some examples of its proposed
normally, exhibiting no gross anatomical abnormalities [37]. significance are summarized below.
Although the Ca2+ concentration dependence of CICR is Depending on regulatory molecular mechanisms, cell
altered in RyR-3-deficient mice, the E-C coupling in these mice functions may have different activation and deactivation
appears normal. RyR-3-deficient mice are significantly kinetics. For a cell function in which deactivation after a
hyperactive, very likely due to central nervous system abnor- decrease in [Ca 2+]i is slow (smooth muscle contraction may
malities [37). The roles of the RyRs in the central nervous be one case), a fusion of responses corresponding to each
system are an interesting subject for future investigation. Ca 2+ oscillation into a sustained response may occur.
Therefore, Ca 2+oscillation can induce sustained activation
ofthe function even at a very low average increase in the Ca2+
Spatio-temporal patterns of Ca2+ signals concentration. However, for another cell function with either
rapid deactivation kinetics or with slow activation kinetics,
Ca 2+ sparks and Ca 2+ puffs then oscillatory changes in [Ca2+l may not be an effective
stimulus and a steady increase in [Ca2+l may be required for
Localized (diameter < several 11m) and transient « 0.1 sec) the regulation of the function in question. Thus, Ca 2+
[Ca 2+l changes have been detected in muscle cells or oocytes oscillations may be able to activate certain type of cell
injected with IP 3 [38, 39). These localized and transient functions selectively among many Ca 2+-dependent functions
[Ca 2+]j changes have been referred to as Ca2+sparks and Ca 2+ with different activation and deactivation kinetics.
puffs, which are assumed to result from Ca2+ release from
intracellular Ca2+stores due to stochastic opening of a single
or a few RyRs and IP 3Rs, respectively. Conclusions
Ca 2+ sparks seem to be the unitary events recruited in
cardiac E-C coupling [38], although smaller units may be Since the discovery of Ca2+as a regulator of skeletal muscle
present. Furthermore, they might be involved in the regulation contraction by Professor Setsuro Ebashi around 1960, the
of the activity of ion channels of smooth muscle cells. Ca2+ study ofCa 2+as a cell signal has advanced tremendously. One
sparks, which occur in the vicinity of the surface membrane, of the most spectacular advances was the explosion in the
may activate Ca 2+-activated K+ channels. Since the Ca 2+- number of cell functions that have been recognized to be
activated K+ channels have an extremely large conductance, regulated by changes in the [Ca 2+]j' In regulation of the
their activation even in a very small fraction of the cell surface multifarious cell functions, Ca 2+ signals exhibit diverse
area may induce a significant outward current leading to spatio-temporal patterns. Despite the apparent differences
hyperpolarization of the cell membrane. In some vascular in the spatio-temporal patterns, many Ca2+ store-mediated
smooth muscle cells, hyperpolarization induced by Ca2+sparks Ca 2+ signalling mechanisms depend critically on the Ca 2+
is assumed to inhibit voltage-dependent Ca2+channels and thus signal amplification mechanisms of the intracellular Ca 2+
decrease the [Ca2+l [40). Therefore, a generalized increase in release channels, RyR and IP 3 R. Therefore, the Ca 2 + sig-
the intracellular Ca2+concentration induces contraction, while nalling mechanisms in many types of cells may depend
a localized rise induces relaxation ofsmooth muscle cells. This critically on a very simple property of the Ca 2+ release
point highlights the importance of spatial patterns of Ca 2+ channels, i.e. Ca 2+sensitivity of the release of intracellular
signals in the regulation of cell functions. Ca 2+ stores.

Ca 2+ oscillation
Acknowledgements
Even at a constant level of agonist-induced stimulation, the
magnitude of [Ca2+l change may fluctuate at a frequency of This work was supported in part by grants from the Ministry
once in every several seconds to several minutes; such a of Education, Science, Sports and Culture of Japan, the
pattern is called Ca 2 + oscillation, which was initially Mitsubishi Foundation, the Uehara Memorial Foundation and
observed in hepatocytes [41]. Ca2+oscillations have also been the Naito Foundation.
detected in perivascular sympathetic nerve or noradrenaline
stimulated vascular smooth muscle cells. Each Ca2+oscillation
is accompanied by a Ca2+wave. With an increase in agonist References
concentration, the frequency of the Ca2+ oscillations rather
than the peak height is often increased. Details of the 1. Ebashi S, Endo M: Calcium ion and muscle contraction. Prog Biophys
significance of the frequency-modulated regulation of cell Mol Bioi 18: 123-183, 1969
190

2. Ebashi S, Endo M, Ohtsuki I: Control of muscle contraction. Q Rev 24. Sutko JL, Airey JA: Ryanodine receptor Ca2• release channels: Does
Biophys2:351-384,1969 diversity in form equal diversity in function? Physiol Rev 76: 1027-
3. Gilkey J, Jaffe L, Ridgway E, Reynolds G: A free calcium wave traverses 1071,1996
the activating egg of the medaka, Orizias taplis. J Cell Bioi 76: 448- 25. Hakamata Y, Nakai J, Takeshima H, Imoto K: Primary structure and
466,1978 distribution ofa novel ryanodine receptor/calcium release channel from
4. Jaffe LF: The path of calcium in cytosolic calcium oscillations: A rabbit brain. FEBS Lett 312: 229-235, 1992
unifying hypothesis. Proc Natl Acad Sci USA 88: 9883-9887,1991 26. Oyamada H, Murayama T, Takagi T, Iino M, Iwabe N, Miyata T, Ogawa
5. Miyazaki S, Yuzaki M, Nakada K, Shirakawa H, Nakanishi S, Nakade S, Y, Endo M: Primary structure and distribution of ryanodine-binding
Mikoshiba K: Block ofCa'· wave and Ca'· oscillation by antibody to protein isoforms ofthe bullfrog skeletal muscle. J BioI Chern 269: 17206-
the inositol 1,4,5-trisphosphate receptor in fertilized hamster eggs. 17214,1994
Science257: 251-255,1992 27. Ottini L, Marziali G, Conti A, Charlesworth A, Sorrentino V: Alpha and
6. Berridge MJ: Inositol trisphosphate and calcium signalling. Nature beta isoforms of ryanodine receptor from chicken skeletal muscle are
361:315-325,1993 the homologues of mammalian RyRI and RyR3. Biochem J 315: 207-
7. lino M, Kasai H, Yamazawa T: Visualization of neural control of 216,1996
intracellular Ca'· concentration in single vascular smooth muscle cells 28. Furuichi T, Furutama D, Hakamata Y, Nakai J, Takeshima H, Mikoshiba
in situ. EMBO J 13: 5026--5031, 1994 K: Multiple types of ryanodine receptor/Ca 2• release channels are
8. lino M: Calcium dependent inositol trisphosphate-induced calcium differentially expressed in rabbit brain. J Neurosci 14: 4794-4805,
release in the guinea-pig taenia caeci. Biochem Biophys Res Commun 1994
142:47-52,1987 29. Hakamata Y, Nishimura S, Nakai J, Nakashima Y, Kita T, Imoto K:
9. Iino M: Biphasic Ca'· dependence of inositol 1,4,5-trisphosphate- Involvement of the brain type of ryanodine receptor in T-cell
induced Ca release in smooth muscle cells of the guinea pig taenia proliferation. FEBS Lett 352: 206--210, 1994
caeci. J Gen Physiol95: 1103-1122, 1990 30. Giannini G, Conti A, Mammarella S, Scrobogna M, Sorrentino V: The
10. lino M, Endo M: Calcium-dependent immediate feedback control of ryanodine receptor/calcium channel genes are widely and differentially
inositol I,4,5-trisphosphate-induced Ca'· release. Nature 360: 76--78, expressed in murine brain and peripheral tissues. J Cell Bioi 128: 893-
1992 904,1995
II. Bezprozvanny I, Watras J, Ehrlich BE: Bell-shaped calcium-response 31. TanabeT, Beam K, Powell J, Numa S: Restoration ofexcitation-contraction
curve of Ins( I,4,5)P J -and calcium-gated channels from endoplasmic coupling and slow calcium current in dysgenic muscle by dihydro-
reticulum ofcerebellum. Nature 351: 751-754,1991 pyridine receptor complementary DNA. Nature 336: 134-139, 1988
12. Finch EA, TumerTJ, Goldin SM: Calcium as a coagonist of inositol 1,4,5- 32. Takeshima H, lino M, Takekura H, Nishi M, Kuno J, Minowa 0, Takano
trisphosphate-induced calcium release. Science 252: 443-252, 1991 H, Noda T: Excitation-contraction uncoupling and muscular de-
13. Furuichi T, Yoshikawa S, Miyawaki A, Wada K, Maeda N, Mikoshiba generation in mice lacking functional skeletal muscle ryanodine-receptor
K: Primary structure and functional expression of the inositol 1,4,5- gene. Nature 369: 556--559, 1994
trisphosphate-binding protein P41IO • Nature 342: 32-38, 1989 33. Takeshima H, YamazawaT, IkemotoT, Takekura H, Nishi M, NodaT,
14. Takeshima H, Nishi M, Iwabe N, Miyata T, Hosoya T, Masai I, Hotta Y: lino M: Ca'·-induced Ca'· release in myocytes from dyspedic mice
Isolation and characterization of a gene for a ryanodine receptor / lacking type-I ryanodine receptor. EMBO J 14: 2999-3006, 1995
calcium release channel in Drosophila metanogaster. FEBS Lett 337: 34. Yamazawa T, Takeshima H, Sakurai T, Endo M, lino M: Subtype
81--87,1994 specificity of the ryanodine receptor for Ca 2• signal amplification in
15. Endo M: Calcium release from the sarcoplasmic reticulum. Physiol Rev excitation-contraction coupling. EMBO J 15: 6172--6177, 1996
57: 71-108,1977 35. Takamatsu T, Wier WC: Calcium waves in mammalian heart: Quan-
16. Meissner G: Ryanodine receptor/Ca'· release channels and their tification of origin, magnitude, waveform, and velocity. FASEB J 4:
regulation by endogenous effectors. Annu Rev Physiol 56: 485-508, 1519-1525,1990
1994 36. Berlin JR, Cannell MB, Lederer WJ: Cellular origins of the transient
17. Endo M, Tanaka M, Ogawa Y: Calcium induced release of calcium from inward current in cardiac myocytes. Role of fluctuations and waves of
the sarcoplasmic reticulum of skinned skeletal muscle fibres. Nature elevated intracellular calcium. Circ Res 65: 115-126, 1989
228:34-36,1970 37. Takeshima H, IkemotoT, Nishi M, Nishiyama N, Shimuta M, Sugitani
18. Ford LE, Podolsky RJ: Regenerative calcium release within muscle cells. Y, Kuno J, Saito I, Saito H, Endo M, Iino M, Noda M: Generation and
Science 167: 58-59, 1970 characterization of mutant mice lacking ryanodine receptor type 3. J
19. Niibauer M, Callewaert G, Cleemann L, Morad M: Regulation ofcalcium Bioi Chern 271: 19649-19652,1996
release is gated by calcium current, not gating charge, in cardiac 38. Cheng H, Lederer WJ, Cannell MB: Calcium sparks: Elementary events
myocytes. Science 244: 800--803, 1989 underlying excitation - contraction coupling in heart muscle. Science
20. Armstrong C, Bezanilla F, Horowicz P: Twitches in the presence of 262:740-744,1993
ethylene glycolbis-(b-aminoethylether)-N,N'-tetraacetic acid. Biochim 39. Yao Y, Choi J, Parker L: Quantal puffs of intracellular Ca 2• evoked by
Biophys Acta 468: 31-50, 1972 inositol trisphosphate in Xenopus oocytes. J Physiol 482: 533-553,
21. Rios E, Pizarro G: Voltage sensor of excitation-contraction coupling in 1995
skeletal muscle. Physiol Rev 71: 849-908, 1991 40. Nelson MT, Cheng H, Rubart M, Santana LF, Bonev AD, Knot F,
22. Gyorke S, Palade P: Calcium-induced calcium release in crayfish skeletal Lederer WJ: Relaxation of arterial smooth muscle by calcium sparks.
muscle. J Physiol457: 195-210, 1992 Science 270: 633--637, 1995
23. Ushio H, Watabe S, lino M: Crayfish skeletal muscle requires both 41. Woods NM, Cuthbertson KS, Cobbold PH: Repetitive transient rises
influx of external Ca'· and Ca'· release from internal stores for in cytoplasmic free calcium in hormone-stimulated hepatocytes. Nature
contraction. J Exp Bioi 181 : 95-105,1993 319:600-602,1986
Molecular and Cellular Biochemistry 190: 191-201, 1999.
© 1999 Kluwer Academic Publishers.

Comparison of properties of Ca2+ release channels

between rabbit and frog skeletal muscles

Yasuo Ogawa, Takashi Murayama and Nagomi Kurebayashi

Department ofPharmacology, Juntendo University School ofMedicine, Tokyo, Japan

Abstract
Biochemical investigation of Ca2+release channel proteins has been carried out mainly with rabbit skeletal muscles, while frog
skeletal muscles have been preferentially used for physiological investigation ofCa2+release. In this review, we compared the
properties ofryanodine receptors (RyR), Ca2+release channel protein, in skeletal muscles between rabbit and frog. While the
Ryrl isoform is the main RyR of rabbit skeletal muscles, two isoforms, (X- and ~-RyR which are homologous to Ryrl and Ryr3
isoforms in mammals, respectively, coexist as a homotetramer in a similar amount in frog skeletal muscles. The two isoforms
in an isotonic medium show very similar property in [3H]ryanodine binding activity which is parallel to Ca2+-induced Ca2+release
(CICR) activity, and make independent contributions to the activities of the sarcoplasmic reticulum. CICR and pH]ryanodine
binding activities of rabbit and frog are qualitatively similar in stimulation by Ca 2+, adenine nucleotide and caffeine, however,
they showed the following quantitative differences. First, rabbit RyR showed higher Ca2+affinity than the frog. Second, rabbit
RyR showed higher activity in the presence of Ca 2+ alone with less stimulation by adenine nucleotide than the frog. Third,
rabbit RyR displayed less enhancement of pH]ryanodine binding by caffeine in spite of having a similar magnitude of Ca2+
sensitization than the frog, which may explain the occasional difficulty by researchers to demonstrate caffeine contracture with
mammalian skeletal muscles. Finally, but not least, rabbit RyR still showed marked inhibition of [3H]ryanodine binding in the
presence of high Ca2+concentrations in the 1 M NaCI medium, while frog RyR showed disinhibition. Other matters relevant to
Ca2+release were also discussed. (Mol Cell Biochem 190: 191-201,1999)

Key words: Ca2+release, frog, rabbit, ryanodine receptor, skeletal muscle

Introduction release channel or protein, i.e. ryanodine receptor RyR, have

been made with rabbit skeletal muscle, and it is generally
Professor Ebashi always stressed to his students that bio- assumed that the property ofRyR from rabbit might hold with
chemical investigation ofmuscle proteins was carried out with other species. We noticed that this was not the case in some
rabbit skeletal muscles, while frog skeletal muscle had been respects with frog RyR. We discuss in this article the similarity
preferred for physiological approaches to muscle contraction, and the difference ofCa 2+release channels between rabbit and
and that quantitative investigation of calcium ion balance was frog skeletal muscle. General aspects ofRyR can be found in
necessary to establish the regulatory role of calcium ion in the many reviews already published [1-9]. Some differences
muscle contraction. One of the authors (YO), when a post- in Ca 2+-uptake by SR between rabbit and frog have been
graduate student, had a chance to help Dr. S. Winegrad who discussed elsewhere [10].
was attempting to prepare an entire sarcoplasmic reticulum
(SR) from frog skeletal muscle in Dr. Ebashi's laboratory
during the latter's sabbatical leave. Frog skeletal muscle was, The principal component of Ca 2+-release
thereafter, the primary material used in our experiments. channels
There are few laboratories except ours where biochemical
investigation of SR from frog skeletal muscles has ever been The Ca2+-release channel in skeletal muscle is largely com-
routinely conducted. Many investigations of isolated Ca2+- posed oftetramer ofryanodine receptor (RyR), a protein of

Address for offprints: Y. Ogawa, Department ofPharmaco!ogy, Juntendo University School of Medicine, 2-1-1, Hongo, Bunkyo-ku, Tokyo 113, Japan
192

about 5000 amino acid residues with a molecular weight of brain, a well-known expression site of Ryr3, was less than
about 560 kDa. Three different genes express the corre- 2% of total RyR (mostly Ryr2) in that organ. They recently
sponding isoforms in mammals: Ryr 1 by ryr J, Ryr2 by ryr2 obtained results with rabbit diaphragm that the fraction of
and Ryr3 by ryrJ [6,9]. The protein-staining pattern of SR Ryr3 content in the muscle was much less than its value in
on SDS-PAGE showed a single band of RyR with rabbit the brain [26]. In the back muscles no Ryr3 was detected by
skeletal muscle in contrast to double bands with frog skeletal immunoprecipitation. This may not be consistent with the
muscle [11, 12]. The RyR in rabbit skeletal muscles turned hypothesis that rapid contraction requires the presence of
out to be Ryrl. Double bands of similar density with frog Ryrl alone. Rigorous re-examination will be necessary to
skeletal muscle are referred to as a- and /3-RyR isoforms. The determine whether nonmammalian skeletal muscles that
two isoforms occur as a homotetramer in a muscle fiber [12, showed a single band ofa-RyR on SOS-PAGE have, in fact,
13]; their coexistence occurs not only in frog skeletal muscles no /3-RyR or too little to detect it by conventional analysis.
but also in chicken, and fishes [11, 14, 15]. They are distinct We have mentioned above that a- and /3-RyRs form
proteins, and /3-RyR, which shows the greater mobility, is not homotetramers and function as Ca 2+-induced Ca2+ release
a degradation product of a-RyR [12,13]. /3-RyR is not a (CICR) channels. Ryr3 also functions as a homotetramer not
homologue of Ryr2, although this was originally the pre- only in rabbit brain [25] but also in skeletal muscles [26].
vailing expectation among many investigators [16, 17]: Sutko Preliminary results in our work showed that Ryr3 coexisted
and his colleagues [18] showed a third distinct isoform in with Ryrl in the same cells of skeletal muscle fibers. Forma-
avian cardiac muscle, and no positive band was observed on tion of a homotetramer even in as small a fraction as below
Western blot analysis with anti-/3-RyR antibody of micro- 2%, but with no evidence of a heterotetramer, is in marked
somes from bullfrog cardiac muscle [14]. Sequences of the contrast to the inositol trisphosphate receptor, another Ca2+
cloned cONAs showed that a- and /3-RyR of frog skeletal release channel protein. The potential formation of a hetero-
muscle are most homologous to mammalian Ryrl and Ryr3, tetramer of inositol trisphosphate receptor (at least 3 iso-
respectively [19]. Recently a full amino acid sequence for forms) may be an underlying mechanism accounting for its
/3-RyR of chicken skeletal muscle was deduced, while that diverse properties [27,28].
for a-RyR remains only partial [20]. We can conclude, The content of Ryr3 may be far less than 2% of Ryr I, at
however, that a- and /3-RyR of chicken skeletal muscle are least in rabbit skeletal muscle. The difference in [3H]ryanodine
most homologous to mammalian Ryrl and 3, respectively, as binding activity between Ryrl and Ryr3 was not marked
true offrog skeletal muscle. We can further assume that most enough to compensate for the difference in their contents [25,
skeletal muscles from non-mammalian vertebrates have two 26]. Putative Ryr3 from ryr I-targeted mouse was, in fact,
isoforms (a- and /3-RyRs, i.e., Ryrl and 3 homologues) in reported to be about 20 times lower in Ca 2+sensitivity in Ca 2+-
similar amounts in the same fiber. Not all skeletal muscles induced Ca 2+release (CICR) [22]. Therefore, the contribution
from non-mammalian vertebrates have two coexisting of Ryr3 to CICR from and FH]ryanodine binding to SR must
isoforms: lizards and snakes showed a single band ofa-RyR, be negligible in rabbit skeletal muscle. The involvement of
while turtles and crocodiles showed a- and /3-RyRs as did Ryr3 in excitation-contraction coupling is also probably
fishes, amphibians and birds [15]. Extraocular and swim- minor, if any. This conclusion is consistent with the finding
bladder muscles from fishes which contract very rapidly that the skeletal muscles from ryrJ-knocked out mouse
showed only a-RyR while their body skeletal muscle had showed contractions similar to those from a normal animal,
both a- and /3-RyR. The rattlesnake tail-shaker muscle which while those from ryr I-targeted mouse did not contract on
produces sound by rapid vibration as the swimbladder depolarization [21,29]. Correspondingly, a-RyR, the homo-
muscles do also has a single isoform, a-RyR [15]. logue to Ryrl, is likely to play an important role in the
Caffeine sensitive Ca2+-release was detected in skeletal excitation-contraction coupling in nonmammalian verte-
muscle cells from ryr I-gene targeted mouse, and this was brates, because skeletal muscles from a crooked-neck dwarf
ascribed to Ryr3 because mRNA from ryrJ was recognized mutant chick which lacks a-RyR failed to contract on
[21,22]. The detection of Ryr3 in normal skeletal muscles depolarization [30]. Because those mutant muscles showed
by Western blot analysis, however, has been unsuccessful to Ca 2+release on application of caffeine, /3-RyR is functional
date except by Sorrentino's laboratory. He estimated that Ryr3 as a CICR channel. In view of the characteristic disposition
accounted for one twentieth to one-fiftieth of Ryr1 in mam- of Ryrl to which putative voltage sensors are alternately
malian skeletal muscle [23, 24]. The contents of Ryr3 varied apposed in swimbladder muscles [31] and mammalian
among different muscles, e.g. relatively higher in diaphragm skeletal muscles [4], the disposition of a- and /3-RyR, i.e.,
and soleus, and lower levels in abdominal muscles and tibialis alternate or random arrays in nonmammalian vertebrates is
anterior. No detectable levels of Ryr3 were observed in the of interest. The biological significance of /3-RyR which is
extensor digitorum longus [24]. Murayama and Ogawa [25] the homologue to Ryr3 and occurs in a similar amount
estimated by immunoprecipitation that Ryr3 in the rabbit remains to be elucidated.
 193

Ca2+-induced Ca2+ release (CICR) and nucleotide, 4 mM AMP in this case (circles), enhanced these
rates as much as 100 times without significant change in their
RyR Ca 2 + dependency. Endo [38] reported that the factor of
enhancement could reach as high as 1000. The rate in-
General properties of CICR creased with increase in the Ca2+ concentration up to 0.1 mM
and at about 10 IlM Ca 2+ was half the maximum rate. Ca2+
Since the first demonstration ofCICR by Ford and Podolsky at concentrations over 0.1 mM, in contrast, were dose-
[32] and Endo et al. [33], CICR has been energetically dependently inhibitory to CICR. ATP, ADP, andAMPOPCP,
studied; it is a Ca2+ release mechanism not only in skeletal an unhydrolyzableATP analogue, are more potent than AMP
muscles but also in various kinds of cells including other in their stimulating effect [39]. Caffeine at 5 mM increased
types of muscle cells. However, there are two different both the Ca 2+ sensitivity in CICR and the maximum rate at
opinions about the biological role of CICR in the physio- the optimal Ca 2+ concentration (squares in Fig. 1). However,
logical contraction ofskeletal muscle: one is that CICR serves enhancement of the maximum rate by caffeine was less than
as an amplifier of Ca 2+ release triggered by conformation that by AMP (squares) or by other adenine nucleotides [38].
change of the voltage sensor (i.e., the dihydropyridine- The effects of caffeine and AMP are strongly potentiating to
derivative receptor), and the other is that CICR has no each other (filled circles). Procaine decreased the rate of
function in the physiological depolarization-induced con- CICR over the whole range of Ca2+ concentrations without
traction [34-37]. Because an understanding of CICR is significant change in Ca2+ dependency [38,40], while Mg 2+
helpful in clarifying the properties of RyR, we offer a decreased both the Ca2+ sensitivity and the maximum rate at
summary of the findings on CICR to date. the optimum Ca2+ concentration [38, 40]. The effect ofMg2+
Figure 1 shows CICR in skinned frog iliofibularis muscle may be explained by the combined effect of competitive
fibers. The rates of CICR triggered by Ca2+ alone were very antagonistic action in stimulatory Ca2+ sites and synergistic
small (triangles). However, the addition of an adenine action in an inhibitory Ca2+ concentration range. Procaine and

rI·h-',----.,.------,r-:----'"7T"----T"""""1
at 4'C
~. .1- 1
~•
1=0.16
3.0 30

o • •
<] 0
:g«I 2.0 20 :g«I
..
.!?
Q)

«I
..
Q)
Q)
t.Il
U U
'0 '0
~
a:
1.0 10 Q)

~
I
•
0 ~,., o
G 7.0

°
Fig. 1. Ca'+-induced Ca release in skinned frog iliofibularis muscle fibers. For details of the method, refer to Kurebayashi and Ogawa (1986) [40].li - Ca'+
alone (10 mM Ca-EGTA buffer); 0 - 5 mM caffeine added; -4 mM AMP added; .-4 mM AMP and 5 mM caffeine added. The medium was at an ionic
strength of 0.16, the temperature was 4°C, and sarcomere length was 2.8 !lm. Closed circles with arrows mean that the rate of Ca release was too fast to be
determined, G means that 10 mM EGTA was added in the absence of Ca, and the ordinates reflect the relative rate of Ca release. Note that the scale in the
ordinate was increased by a factor often in the presence of AMP (0, .). Reproduced from Crit Rev Biochem Mol BioI 29 (4): 229-274, 1994 [6] with
permission of CRC Press, Inc., Boca Raton, Florida.
194

MgZ+ differ in their inhibitory mechanisms, as do adenine might be lower than that for rabbit Ryrl ([12], but see also
nucleotide and caffeine in their stimulatory mechanisms. [45]). Modulation of the gating process, however, may be
Table 1 summarizes the Caz+concentrations which would different in some respects.
give half the maximum rate ofCICR «pCa)l/z) using various
sources of skinned skeletal muscle. Determinations with or
without an adenine nucleotide according to similar experi- Comparison between a-RyR and f3-RyR
mental protocol were chosen just because of easier com-
parison although numerous results were reported using a Before further discussion of rabbit and frog Ca z+ release
variety of experimental procedures. Although the results vary channels, it will be helpful to clarify the difference between
slightly depending on experimental conditions, we are able a- and P-RyR of frog skeletal muscle, because the two
to conclude that mammalian skeletal muscle has a higher Ca z+ isoforms coexist in closely similar amounts and CICR is
sensitivity to CICR by a factor of 2-3 than frog skeletal believed due to summation of the independent contribution
muscle. of each one. In an isotonic medium, [3H]ryanodine binding
activity of bullfrog P-RyR was very similar to a-RyR [46]:
biphasic Caz+ dependence with almost the same values of
[3H]Ryanodine binding to RyR (pCa)l/z' not only for activation but also for inactivation. They
are also similarly modulated by AMPOPCP and caffeine. To
Open CICR channel protein, i.e. RyR, can bind [3H]ryanodine inhibitory agents, e.g., Mgz+, procaine, and ruthenium red,
[41,42]. [3H]Ryanodine binding activity to SR vesicles from they are also similarly sensitive in 0.17 M NaCI medium. In
bullfrog skeletal muscle was modulated by Caz+, AMPOPCP the I M NaCI medium where both [3H]ryanodine binding and
and/or caffeine as in CICR [43]. Similarresults were obtained CICR channel activity were enhanced markedly, neither
with SR vesicles from rabbit skeletal muscle [44], allowing a-RyR nor P-RyR showed inhibition at high Caz+concentra-
us to conclude that [3H]ryanodine binding activity closely tions; their Caz+-dependences were monophasic (see Fig. 6,
parallels CICR channel activity. It should be noted, however, panel 'Frog'). Some differences between a- and P-RyR,
that [3H]ryanodine binding is an endothermic reaction and however, were noted in the I M NaCI medium: bullfrog
that at a lower temperature it shows a decreased affinity for P-RyR showed about 20 times higher Ca z+ sensitivity than
ryanodine and retarded association and dissociation [43]. a-RyR [12,46]; P-RyR is also more resistant to those
Under fully activated conditions, the heavy fractions, i.e., inhibitors than a-RyR in I M NaCI medium [46].
terminal cisternae-rich fractions of SR vesicles from rabbit Lai et al. [16] reached the conclusion that P-RyR would
and bullfrog skeletal muscles showed similar values of Kd be similar to mammalian Ryr2. Bull and Marengo [17]
(2-4 nM) and Bm3x (10--12 pmol/mg protein) (Murayama, consistently reported two types ofCaz+release channels with
unpublished results). This indicates that these fractions from distinct Caz+ dependences in SR vesicles isolated from frog
the muscles of both rabbit and bullfrog have similar density skeletal muscle, one of biphasic dependence with marked
in RyR, i.e. CICR channels. Furthermore, purified rabbit Ryrl inhibition at high Ca z+ concentrations and the other of
and bullfrog a- and P-RyR showed similar values for Kd (2- monophasic dependence with little inhibition which is similar
5 nM) and Bm3x (l mol [3H]ryanodine/mol tetramer) under to the property of Ryr2. Similar results were obtained from
fully activated conditions [3, 5, 6]. This may suggest that the SR vesicles of toadfish white swimming muscles by O'Brien
properties of CICR channels in the open state are similar et al. [47]. Percival et al. [48] investigated the ion channel
between frog and rabbit skeletal muscle, although published properties ofa- and P-RyR from chicken skeletal muscles and
results reported that the unit conductance for frog a-RyR reported that the two isoforms embodied Caz+ channels with
similar conductances but different gating properties on
activation by Caz+ andATP, and on inactivation by Ca z+. Thus
Table I. (pCa)", for CICR in skinned skeletal muscle fibers these investigators concluded that the two RyR isoforms in
non-mammalian skeletal muscles had distinct CICR pro-
Material (pCa)'/' Ref perties in contrast to our conclusion. It is possible that this
clawed toad (2°C, pH 7.1) 5.0 [35] may be due to differences of the species examined, in view
frog (4°C, pH 6.8) 4.8-5.5 [40] ofthe limited immunologic cross-reactivities among a-RyRs
frog WC, pH 6.8) 5.0 Fig. 1 from different species [14]. A key to the solution of this
frog (16°C, pH 6.8 or 7.0) 5.5 Kurebayashi, unpublished
discrepancy would be determinations ofCICR with skinned
rabbit (2o-22°C, pH 7.0) 5.7 [78]
chicken and fish skeletal muscle fibers.
human (20°C, pH 7.0) 5.5-5.3 [35,79] Here, we would like to further discuss the inhibitory effect
guinea pig (20 or 38°C, pH 7.0) 6-5.7 [35,80] of Ca z+ at high concentrations. As described above, P-RyR
mouse (25°C, pH 7.0) 6.1 [22] was similar to a-RyR in Ca z+ dependence in isotonic 0.17 M
 195

NaCI medium, and was also sensitive to the inactivating 200

action ofCa2+ at high concentrations. This was also the case
with Ryr3 from rabbit brain [25] and skeletal muscle (dia-
a. Rabbit
"0
1::_ 150
phragm) (unpublished results). This is in marked contrast to :::JI::
0'-
the case with Ryr2 which showed weak inhibition by high .0.$
Ca2+ in an isotonic medium [1-3,5]. we
1::0.
'60> 100
g.§
co-
>0,0
Comparison between Ca 2+ release channels ofrabbit and lIE
~o.

frog skeletal muscles :r:~

'2....
50

Adenine nucleotides
Adenine nucleotides stimulate [3H]ryanodine binding as true 0
of CICR. Rabbit Ryrl and bullfrog a-RyR showed similar 8 7 6 5 4 3 2
pCa
dependence on AMPOPCP as shown in Fig. 2; their EC 50s
were approximately 0.3 mM. This was also the case with
bullfrog P-RyR (see also [43]). Since the value in the absence 200
ofAMPOPCP was significantly lower with frog a-RyR than
b. Frog
with rabbit Ryrl, the enhancement by AMPOPCP looked "0
1::_
more marked with the former than with the latter (Fig. 2). This :::JI:: 150
0'-
is clearly seen in Fig. 3 which shows [3H]ryanodine binding .0.$
at various Ca 2+ concentrations with and without 1 mM
we
1::0.
AMPOPCP. Rabbit Ryrl showed moderate amounts of '60>
1:: __ 100
°E
biphasically Ca2+ dependent ryanodine binding in the absence co-
>0,0
of AMPOPCP (Fig. 3a), while frog a-RyR showed only lIE
-0.
marginal amounts (Fig. 3b). The enhancement by 1 mM
:r:~
'2....
50
AMPOPCP was about 10-fold at the optimum Ca2+ con-
centrations with frog a-RyR, whereas it was about 2.5-fold
with rabbit Ryrl. The Ca2+ concentrations which would give
OL-~~~~!!l=:ll:L....::::.........L_----I--=~_----.J

half the maximum value «pCa)1/2) in the stimulatory Ca2+

8 7 6 5 4 3 2
pCa
concentration range were 5.7-5.5 with rabbit Ryrl, and

Fig. 3. Ca 2+ dependent [3H]ryanodine binding to rabbit Ryrl (panel a) and

200 , . . . . . . . { I - - - - - - - - - - - - - - - , bullfrog a-RyR (panel b) and stimulation by I mM AMPOPCP. Experi-
mental conditions were similar to those in Fig. 2 except for varied Ca 2+
o concentrations with (filled symbols) and without (open symbols) I mM
"0
1::_ 150 AMPOPCP. Note that ryanodine binding to Ryrl was moderate without
:::JI::
0'- AMPOPCP, while that to a-RyR was marginal. With AMPOPCP, the
.0.$
we
1::0.
enhanced ryanodine bindings to Ryrl and a-RyR were comparable.

'60> 100
°E
1:: __
co-
>0,0 around 5.1 with frog a-RyR. This is consistent with the results
lIE for CICR in skinned fibers (Table 1).
50
~o.
:r:~
'2.... In skinned fibers from frog and rabbit skeletal muscles,
we can observe CICR, corresponding to the results with
o y'I-_~L...._ _ ____l. __'_ __J [3H]ryanodine binding. When the activator for CICR was
o 0.01 0.1 10 Ca2+ itself, the highest rate constant at the optimum Ca2+
concentration with frog skinned fiber was less than 0.1 min-I
AMPOPCP (mM)
[6,40]. With rabbit skinned fiber, in contrast, the rate constant
was about 4--Q min-I (Kurebayashi, unpublished results),
Fig. 2. AMPOPCP dependence of [3H]ryanodine binding. Rabbit Ryrl which was comparable to the value of frog in the presence
(circles) or bullfrog a-RyR (triangles) was incubated with 8.5 nM
of4 mM AMP [6,40]. Stimulation by adenine nucleotide, in
[3H]ryanodine at 25°C for 4 h in 0.17 M NaCI medium containing 40 11M
Ca2+ buffered by I mM EGTA, 10 mM MOPSOlNaOH, pH 6.8, I% CHAPS, tum, was less marked with the rabbit preparation. The factor
0.5% egg lecithin, 2 mM DTT and AMPOPCP indicated at the abscissa. ofenhancement by 4 mM AMP was several fold with a rabbit
196

specimen, while it was an order of several tens with frog 6r-----------------,

fibers (see also Fig. 3). The maximum rate ofCa2+ release with
rabbit SR appears to be larger than that with frog SR in the a. Rabbit
presence of an adenine nucleotide. With rabbit, we could
detect Ca2+ release from SR even in the presence of 10 mM
EGTA with no added Ca (pCa>8).
Donoso et al. [49] compared CICRs from triads isolated
from frog and rabbit skeletal muscle, and reported that the
two showed similar maximal rate constants (10-12 sec-I) for
CICRs induced by 10 J.!M Ca2+ and 2 mM ATP. Their results
were 60 or more times higher than ours. One reason for the
discrepancy is partly due to their usage of2 mMATP instead
of 4 mM AMP to activate CICR. Because determination of
the very rapid rate of Ca 2+ release is difficult in an experi-
mental system with skinned fiber, our experimental con-
ditions were set to be within release rates appropriate to make
this possible. Another reason is the difference in the pre-
parations used: theirs was a triad-rich fraction, while ours was
a skinned fiber. The former has relatively higher average
density in RyR than the latter.

Caffeine
Figure 4 shows activation by caffeine of [3H]ryanodine
binding to rabbit (circles) and frog (triangles) SR vesicles.
Common effects of caffeine are: (1) increase in the Ca2+
sensitivity in Ca2+ dependent ryanodine binding at a concen-
tration of 10 mM caffeine or less; and (2) adenine nucleotide-
like effect, i.e., increase in the maximum [3H]ryanodine
binding at the optimum Ca2+ concentration. The latter effect
is more remarkable at a concentration higher than 10 mM o
caffeine where the Ca2+-sensitizing effect is at the maximum 8 7 6 5 4 3 2
[43]. The enhancement by caffeine is, however, much smaller pea
than that by ATP as true with CICR [38, 39]; the maximum
[3H]ryanodine binding in the presence of 10 mM caffeine was
Fig. 4. Potentiation by caffeine of [3H]ryanodine binding to rabbit and
0.06 pmol/mg protein with rabbit SR vesicles and 0.45 pmol/ bullfrog SR vesicles. Experimental conditions were similar to those in Fig.
/mg protein with frog SR vesicles, whereas I mM AMPOPCP 3 in the presence of 1 mM AMPOPCP, except for the use of SR vesicles
gave rise to [3H]ryanodine binding of 0.8 and 1.44 pmol/mg instead of purified RyR isoforms, omission of CHAPS and phospholipids,
protein, respectively. No significant ryanodine binding was and addition of 10 mM caffeine (filled symbols) where indicated. While
Ca2> sensitization by 10 mM caffeine was similar in rabbit (~ pCa-LO)
detected above the background level with rabbit or frog SR
and frog (~ pCa-I.2), the amounts of maximum binding differed greatly:
vesicles when neitherAMPOPCP nor caffeine was added. Co- 2.0 pmol/mg protein and 4.5 pmol/mg protein, respectively. The values of
existence of 1 mM AMPOPCP and 10 mM caffeine was mu- 8 m" under fully activated conditions were similar in the two preparations
tually potentiating and remarkably enhanced the [3H]ryanodine as shown in Fig. 6.
binding as shown in Fig. 4, being consistent with CICR (Fig.
1, see also [50]).
Weber, A. and Herz [51], who were the first to show the Ca-uptake activity. The Ca-uptake activity of frog SR is not
Ca releasing action of caffeine, reported that several of their lower than that of rabbit SR. Less sensitivity of CICR
rabbit preparations showed no response to the addition of channels to caffeine may be a common property of mammals
caffeine, whereas in 11 out of 12 frog preparations caffeine in view of the difficulties reported with rat muscle [53]. The
did cause Ca release. Ogawa also confirmed difficulty in difference between frog and rabbit SR in stimulation by
observing Ca release by caffeine from rabbit SR [52]. These caffeine of [3H]ryanodine binding (Fig. 4) may be one
results were obtained in the presence ofATP which enabled explanation for the difference in caffeine sensitivity between
Ca-uptake activity, although the effect of caffeine on rabbit these animal species. The enhanced amount of ryanodine
SR vesicles was thereafter shown mostly in the absence of binding was much greater with frog SR than with rabbit SR.
 197

This suggests that the fraction of the channels activated by 6

caffeine may be larger with the former, because the densities
ofCICR channels are similar between the two species on the
"0
a
basis ofsimilar [3H]ryanodine binding activities under the full c_
~c
activating conditions as shown in Fig. 6 (filled symbols). 0'-
4
Although caffeine is a well-known Ca 2+ sensitizer in CICR me
.02

co.
and [3H]ryanodine binding, the effect of enhancement of the '50>
binding may be more important in its impact on the Ca 2+ °E
C .....
(lj-
releasing action. There was no difference between (X.- and >0- 0
a:E 2
~-RyR in the effect ofcaffeine on FH]ryanodine binding [46]. ~o.
I-
~
It will be interesting to examine the effect of caffeine on
mammalian Ryr3, especially its enhancement ofFH]ryanodine
binding, because Ryr3 was initially reported to be character- 0
istically insensitive to caffeine [54]. 8 7 6 5 4 3 2
pCa
Monovalent salts
1 M NaCl or KCI enhanced [3H]ryanodine binding to SR
vesicles so much with an increase in both (l/KD) and Bmax that
a massive amount of bound FH]ryanodine was observed in 100
the presence ofCa2+ alone [55]. Correspondingly, Ca2+ release "0
C
from SR vesicles was also stimulated [44, 56]. The extent of ~
0
.0
enhancement was salt-specific [46]. A portion of the salt was m
replaced by an osmotically equivalent amount of sucrose in c
'5
the stimulating effect [55]. Another interesting effect of 1 M 0
c 50
NaCI was the disinhibition of inactivating Ca 2+ as shown in (lj
>0-
Fig. 6 (see also [46, 55]). The notable difference between a:
rabbit and frog is that rabbit SR vesicles showed marked
I
~
inhibition at high Ca 2+ concentrations, while that of frog SR ~
vesicles was slight.
0
CHAPS and phospholipids 8 7 6 5 4 3 2 1
The effects of CHAPS and phospholipids on [3H]ryanodine pCa
binding activity are of interest. A marked increase in
FH]ryanodine binding by these reagents was observed in the Fig. 5. Effects of CHAPS and phospholipids on ['H)ryanodine binding
presence of I mM AMPOPCP in 0.17 M NaCl medium, not activity of rabbit SR vesicles and purified Ryrl. (a) Potentiating effect of
CHAPS and phospholipids on SR vesicles from rabbit skeletal muscle.
only with frog SR vesicles (3-fold) but also with rabbit SR ['H)ryanodine binding activity was determined as shown in Figs. 2--4 in
vesicles (5-fold) (Fig. 5a). Because there was no significantly the presence of 1 mM AMPOPCP with (filled circles) and without (open
detectable activity in the absence ofAMPOPCP, CHAPS and circles) 1% CHAPS and 0.5% egg lecithin. Similar results were obtained
phospholipids markedly potentiate the effect of AMPOPCP. with bullfrog SR vesicles. (b) Changes in [Ca2+) dependence by addition
Ca2+ sensitivity increased slightly in Ca2+ activation, and it also of I% CHAPS and 0.5% egg lecithin. For comparison, the results for
purified Ryrl (open squares) are also shown. The maximum value for each
decreased in inactivating Ca2+ concentration range with un- set of determinations is calibrated to be 100%.
changed optimum Ca2+ concentration (Fig. 5b). Whether these
changes could be due to dissociation of an RyR-binding
protein such as FKBP12 [57, 58] or triadin [59---Ql] remains served that rabbit SR showed a hyperbolic dependence on
to be determined. They might be also due to a change in inter- luminal Ca 2+, while that of frog SR was sigmoidal. They
actions among RyR monomers which form a homotetramer. claimed that calsequestrin-bound Ca 2+ was important for Ca2+
release. Their results, however, were curious in the respect
that the release rates were independent ofCa2+ concentrations
Some problems in Ca2+ release in situ outside the vesicles between pCa5 and pCa7, while they were
dependent on luminal Ca 2+ concentration. Our results showed
Donoso et al. [49] reported that Ca2+ release was dependent that the Ca 2+ release rates were not different between full-
on the extent ofCa 2+ loading, i.e., luminal Ca2+ concentration, loading and 1/3-10ading level. Shirokova et al. [66] con-
confirming results reported previously [62-65]. They ob- sistently concluded that Ca release flux rates from SR offrog
198

and rat skeletal muscle fibers were independent of the Ca Although there have been numerous in vitro experiments
content in SR. These conclusions, however, may not be of depolarization-induced Ca2+release with skinned skeletal
inconsistent with the results by Donoso et at., because it was muscle fiber [74] and triad-rich fraction [75--77], the results
in the lower loading level that the release rate was strongly are not yet sufficient to understand quantitatively the sequence
dependent. Volpe and Simon [67] prepared calsequestrin of events in intact and/or cut fiber. The effect of high Ca 2+
from SR isolated from frog skeletal muscle and compared concentrations on [3H]ryanodine binding activity, however,
those Ca2+ binding properties with those ofcalsequestrin from might give a clue to understanding the difference in the
rabbit skeletal muscle [68]. The two were similar in their Ca2+-release rate between frog and rat. As shown in Fig. 6,
properties of Ca 2+binding: ionic strength dependent affinity the inhibitory action ofCa2+at high concentrations seems to
and non-cooperative Ca 2+ binding. They discussed quan-
titative analysis of Ca2+ binding sites in the SR lumen and
concluded that calsequestrin is not the sole Ca 2+binding site
8
with K d of approximately I mM, and that the contribution of
low affinity binding sites ofCa 2+-ATPase protein [69] inside a. Rabbit
the lumen would be considerable. They further concluded that "0
c_
most released Ca 2+must come directly from free Ca 2+. This ;:]c 6
0·-
.0$
conclusion may be reasonable because most biological Q)e
co.
material not only of proteins but also of lipids would have
'60> 4
Ca 2+ binding sites of K d on the order of mM or larger. °E
c-.
ctl-
Although channel activity of heavy SR fused into the lipid >.0
bilayer was reported to be augmented enormously by the a::E
~o.

addition ofcalsequestrin into the trans compartment [65], many I- 2

~
reports showed vivid channel activity of purified RyR in-
corporated into this bilayer without the addition of cal-
sequestrin [1-3, 5]. Since Ca flux through the Ca 2+ release 0
channel was observed in the reversed direction into empty SR 8 7 6 5 4 3 2
([70]; Kurebayashi, unpublished results), conformation change pCa
of calsequestrin cannot be the direct trigger of Ca release.
Shirokova et al. [66] determined Ca2+release rate from SR 8
on depolarization in cut fibers from frog and rat skeletal

• •
muscle. They calculated Rp (early peak rate of Ca 2+ release b. Frog
"0
on depolarization) and Rs (steady rate ofCa 2+release during c_ 6
;:]c
depolarization) from observed Ca 2+transient. They reported 0·-
.0$
that Rp and Rs for frog are larger by 5 and 3 times, re- Q)e
co.
spectively, than those for rat. Since mammalian Ryr1s (among '60> 4
rabbit, pig and human) show more than 95% identity in the °E
c-.
ctl-
primary amino acid sequence, CICRs from rabbit and rat SR >.0
a::E
are likely to be very similar. Thus we may use the results with ~o.
I- 2
rabbit SR for interpretation of rat Rand p
R s . Their inter- ~

pretation is that R p is the peak rate of Ca 2+ release which is

amplified by the CICR mechanism, while R s is the steady rate OL-.......~.....~~~_....l...-_---C::t!=-6-_--.J
ofCa 2+ release caused by conformation change ofthe voltage 8 7 6 5 4 3 2 1
sensor during depolarization [37, 71]. Chandler and his pCa
colleagues [72, 73] also showed a similar time course ofCa 2+
release on depolarization of intact and/or cut fibers. Their
Fig. 6. Effect of I M NaCI (filled symbols) on [3H]ryanodine binding to
interpretation, however, was as follows: Rp is the peak rate SR vesicles. Experimental conditions were similar to those in Fig. 4 except
of Ca release triggered through conformation change of the for I M NaCI (filled symbols) instead of 0.17 M NaCI (open symbols)
voltage sensor and R s is caused by the inactivation mechanism where indicated. It is well known that the ryanodine binding is markedly
by Ca 2+ of the Ca 2+ release. It should also be noted that the enhanced in I M NaCI medium as shown here. Note that rabbit SR vesicles
value for R p for frog differed between the two groups: Rios' showed marked inhibition with high Ca2+ concentrations, while frog SR
vesicles showed very little inhibition in the I M NaCI medium. K o and
[66, 71] and Schneider's [37] laboratories reported it to be
8 m" for [3H]ryanodine were 3.6 nM and 11.5 pmollmg protein for rabbit
about 30 mM/s, while Chandler's laboratory [72, 73] reported SR vesicles, and 2.5 nM and 9.4 pmol/mg protein for bullfrog SR vesicles,
it as 100 mM/s or greater. respectively, under fully activated conditions in the I M NaCI medium.
 199

be stronger with rabbit SR than that with frog SR, because 15. O'Brien J, Meissner G, Block BA: The fastest contracting muscles of
intact SR vesicles from rabbit showed marked inhibition in 1 nonmammalian vertebrates express only one isoform of the ryanodine
receptor. Biophys J 65: 2418-2427, 1993
M NaCl medium, whereas with frog SR vesicles the inhibition 16. Lai FA, Liu QY, Xu L, EI-Hashem A, Kramarcy NR, Sealock R,
was very weak under the same conditions. We do not know Meissner G: Amphibian ryanodine receptor isoforms are related to
whether the inactivation mechanism by high Ca2+ concen- those of mammalian skeletal or cardiac muscle. Am J Physiol 263:
trations on CICR is still effective on the depolarization-induced C365-C372, 1992
Ca release if the latter release mechanism is distinct from the 17. Bull R, Marengo JJ: Sarcoplasmic reticulum release channels from
frog skeletal muscle display two types of calcium dependence. FEBS
mechanism of CICR. Further investigations are required. Lett 331: 223-227, 1993
18. Airey JA, Grinsell MM, Jones LR, Sutko JL, Witcher D: Three
ryanodine receptor isoforms exist in avian striated muscle. Bio-
Acknowledgements chemistry 32: 5739-5745, 1993
19. Oyamada H, MurayamaT, Takagi T, Iino M, Iwabe N, MiyataT, Ogawa
Y, Endo M: Primary structure and distribution of ryanodine-binding
Some ofthe results cited here were supported partly by grants protein isoforms of the bullfrog skeletal muscle. J BioI Chern 269:
from the Uehara Memorial Foundation and the Suzuken 17206-17214,1994
Memorial Foundation. We thank Ms. Naomi Ariji for her 20. Ottini L, Marziali G, Conti A, Charlesworth A, Sorrentino V: a. and P
skillful secretarial assistance. isoforms ofryanodine receptor from chicken skeletal muscle are the homo-
logues of mammalian RyRI and RyR3. BiochemJ 315: 207-216,1996
21. Takeshima H, Iino M, Takekura H, Nishi M, Kuno J, Minowa 0,
Takano H, Noda T: Excitation-contraction uncoupling and muscular
References degeneration in mice lacking functional skeletal muscle ryanodine-
receptor gene. Nature 369: 556-559, 1994
I. Fleischer S, Inui M: Biochemistry and biophysics of excitation- 22. Takeshima H, Yamazawa T, Ikemoto T, Takekura H, Nishi M, Noda T,
contraction coupling. Annu Rev Biophys Chern 18: 333-364, 1989 Iino M: Ca2+-induced Ca'· release in myocytes from dyspedic mice
2. McPherson PS, Campbell KP: The ryanodine receptor/Ca2+ release lacking the type-I ryanodine receptor. EMBO J 14: 2999-3006, 1995
channel. J Bioi Chern 268: 13766-13768, 1993 23. Giannini G, Conti A, Mammarella S, Scrobogna M, Sorrentino V: The
3. Coronado R, Morrissette J, Sukhareva M, Vaughan DM: Structure and ryanodine receptor/calcium channel genes are widely and differentially
function ofryanodine receptors. Am J Physiol 266: C 1485-C 1504, 1994 expressed in murine brain and peripheral tissues. J Cell BioI 128: 893-
4. Franzini-Armstrong C, Jorgensen AO: Structure and development of 904,1995
E-C coupling units in skeletal muscle. Annu Rev Physiol 56: 509- 24. Conti A, Gorza L, Sorrentino V: Differential distribution ofryanodine
534, 1994 receptor type 3 (RyR3) gene product in mammalian skeletal muscles.
5. Meissner G: Ryanodine receptor/Ca2+ release channels and their re- Biochem J 316: 19-23, 1996
gulation by endogenous efTectors.Annu Rev Physiol56: 485-508, 1994 25. Murayama T, Ogawa Y: Properties of Ryr3 ryanodine receptor isoform
6. Ogawa Y: Role of ryanodine receptors. Crit Rev Biochem Mol Bioi in mammalian brain. J BioI Chern 271: 5079-5084, 1996
29: 229-274, 1994 26. Murayama T, Ogawa Y: Properties of homotetrameric Ryr3 ryanodine
7. Sitsapesan R, Mcgarry SJ, Williams AJ: Cyclic ADP-ribose, the receptor in mammalian diaphragm muscle. Biophys J 72: A 169, 1997
ryanodine receptor and Ca'· release. Trends Pharmacol Sci 16: 386- 27. MonkawaT, MiyawakiA, Sugiyama T, Yoneshima H, Yamamoto-Hino
391,1995 M, FuruichiT, SarutaT, Hasegawa M, Mikoshiba K: Heterotetrameric
8. Wagenknecht T, Radermacher M: Three-dimensional architecture of complex formation of inositol 1,4,5-trisphosphate receptor subunits.
the skeletal muscle ryanodine receptor. FEBS Lett 369: 43-46, 1995 J BioI Chern 270: 1470G-14704, 1995
9. Sorrentino V: Ryanodine Receptors. CRC Press, Boca Raton, NY, 28. Nucifora FC Jr., Sharp AH, Milgram SL, Ross CA: Inositol 1,4,5-
London, Tokyo, 1996 trisphosphate receptors in endocrine cells: Localization and association
10. Ogawa Y: Comparative aspects of the mechanisms of energy trans- in hetero- and homotetramers. Mol Bioi Cell 7: 949-960, 1996
duction in sarcoplasmic reticulum between rabbit and frog skeletal 29. Takeshima H, Ikemoto T, Nishi M, Nishiyama N, Shimuta M, Sugitani
muscle. In: H. Kleinkauf, H. von Dohren, L. Jaenicke (eds). The Roots Y, Kuno J, Saito I, Saito H, Endo M, Iino M, Noda T: Generation and
of Modern Biochemistry. Walter de Gruyter & Co., Berlin, New York, characterization of mutant mice lacking ryanodine receptor type 3. J
1988, pp 747-758 Bioi Chern 271: 19649-19652,1996
II. Olivares EB, Tanksley SJ, Airey JA, Beck CF, Ouyang Y, Deerinck 30. Ivanenko A, McKemyt DD, Kenyon JL, Airey JA, Sutko JL: Em-
TJ, Ellisman MH, Sutko JL: Nonmammalian vertebrate skeletal bryonic chicken skeletal muscle cells fail to develop normal excitation
muscles express two triad junctional foot protein isoforms. Biophys J contraction coupling in the absence of the a. ryanodine receptor:
59: 1153-1163, 1991 Implications for a two-ryanodine receptor system. J Bioi Chern 270:
12. Murayama T, Ogawa Y: Purification and characterization of two 4220-4223,1995
ryanodine-binding protein isoforms from sarcoplasmic reticulum of 31. Block BA, Imagawa T, Campbell KP, Franzini-Armstrong C: Structural
bullfrog skeletal muscle. J Biochem 112: 514-522, 1992 evidence for direct interaction between the molecular components of
13. Airey JA, Beck CF, Murakami K, Tanksley SJ, DeerinckTJ, Ellisman the transverse tubule/sarcoplasmic reticulum junction in skeletal
MH, Sutko JL: Identification and localization of two triad junctional muscle. J Cell Bioi 107: 2587-2600, 1988
foot protein isoforms in mature avian fast twitch skeletal muscle. J 32. Ford LE, Podolsky RJ: Regenerative calcium release within muscle
BioI Chern 265: 14187-14194, 1990 cells. Science 167: 58-59, 1970
14. Murayama T, Ogawa Y: Relationships among ryanodine receptor 33. Endo M, Tanaka M, Ogawa Y: Calcium induced release of calcium
isoforrns expressed in vertebrate skeletal muscle based on immunologic from the sarcoplasmic reticulum of skinned skeletal muscle fibres.
cross-reactivities. J Biochem 116: 1117-1122, 1994 Nature 228: 34-36, 1970
200

34. Endo M: Calcium release from the sarcoplasmic reticulum. Physiol 55. Ogawa Y, Harafuji H: Osmolarity-dependent characteristics of
Rev 57: 71-108,1977 ['H)ryanodine binding to sarcoplasmic reticulum. J Biochem 107: 894-
35. Endo M: Calcium release from sarcoplasmic reticulum. Cur Top Memb 898, 1990
Trans 25: 181-230, 1985 56. Kasai M, Kawasaki T, Yamamoto K: Permeation of neutral molecules
36. Rios E, Pizarro G: Voltage sensor of excitation-contraction coupling through calcium channel in sarcoplasmic reticulum vesicles. J Biochem
in skeletal muscle. Physiol Rev 71: 849-908, 1991 112: 197-203, 1992
37. Schneider MF: Control of calcium release in functioning skeletal 57. Jayaraman T, BrillantesAM, TimermanAP, Fleischer S, Bromage HE,
muscle fibers. Annu Rev Physiol56: 463-484,1994 Tempst P, Marks AR: FK506 binding protein associated with the
38. Endo M: Mechanism of calcium-induced calcium release in the SR calcium release channel (ryanodine receptor). J Bioi Chern 267: 9474-
membrane. In: S.T. Ohnishi and M. Endo (eds). The Mechanism of 9477,1992
Gated Calcium Transport across Biological Membranes. Academic 58. Brillantes AMB, Ondrias K, Scott A, Kobrinsky E, Ondriasova E,
Press, New York, 1981, pp 257-264 Moschella MC, Jayaraman T, Landers M, Ehrlich BE, Marks AR:
39. Endo M, Kakuta Y, Kitazawa T: A further study of the Ca-induced Ca Stabilization of calcium release channel (ryanodine receptor) function
release mechanism. In: A.D. Grinnell, M.A. Brazier (eds). The by FK506-binding protein. Cell 77: 513-523, 1994
Regulation of Muscle Contraction: Excitation- Contraction Coupling. 59. Caswell AH, Brandt NR, Brunschwig J-P, Purkerson S: Localization
Academic Press, New York, 1981, pp 181-195 and partial characterization of the oligomeric disulfide-linked
40. Kurebayashi N, Ogawa Y: Characterization of increased Ca 2+ efflux molecular weight 95 000 protein (triadin) which binds the ryanodine
by quercetin from the sarcoplasmic reticulum in frog skinned skeletal and dihydropyridine receptors in skeletal muscle triadic vesicles.
muscle fibres. J Muscle Res Cell Motil 7: 142-150,1986 Biochemistry 30: 7507-7513, 1991
41. Pessah IN, WaterhouseAL, Casida JE: The calcium-ryanodine receptor 60. Knudson CM, Stang KK, Moomaw CR, Slaughter CA, Campbell KP:
complex of skeletal and cardiac muscle. Biochem Biophys Res Primary structure and topological analysis ofa skeletal muscle-specific
Commun 128: 449-456, 1985 junctional sarcoplasmic reticulum glycoprotein (Triadin). J Bioi Chern
42. Fleischer S, Ogunbunmi EM, Dixon MC, Fleer EAM: Localization of 268: 12646-12654, 1993
Ca2+ release channels with ryanodine in junctional terminal cisternae 61. Liu G, Pessah IN: Molecular interaction between ryanodine receptor
of sarcoplasmic reticulum of fast skeletal muscle. Proc Nat! Acad Sci and glycoprotein triadin involves redox cycling offunctionally important
USA 82: 7256-7259, 1985 hyperreactive sulfhydryls. J BioI Chern 269: 33028-33034, 1994
43. Ogawa Y, Harafuji H: Effect of temperature on ['H)ryanodine binding 62. Ikemoto N, Ronjat M, Meszaros LG, Koshita M: Postulated role of
to sarcoplasmic reticulum from bullfrog skeletal muscle. J Biochem calsequestrin in the regulation of calcium release from sarcoplasmic
107: 887-893, 1990 reticulum. Biochemistry 28: 6764-6771, 1989
44. Meissner G, EI-Hashem A: Ryanodine as a functional probe of the 63. Ikemoto N, Antoniu B, Kang J-J, Meszaros LG, Ronjat M: Intra-
skeletal muscle sarcoplasmic reticulum Ca 2+ release channel. Mol Cell vesicular calcium transient during calcium release from sarcoplasmic
Biochem 114: 119-123,1992 reticulum. Biochemistry 30: 5230-5237,1991
45. Bull R, Marengo JJ, Suarez-Isla BA, Oonoso P, Sutko JL, Hidalgo C: 64. Gilchrist JSC, Belcastro AN, Katz S: Intraluminal Ca 2+ dependence
Activation of calcium channels in sarcoplasmic reticulum from frog ofCal+ and ryanodine-mediated regulation of skeletal muscle sarco-
muscle by nanomolar concentrations of ryanodine. Biophys J 56: 749- plasmic reticulum Ca 2+ release. J Bioi Chern 267: 20850-20856,1992
756, 1989 65. Kawasaki T, Kasai M: Regulation of calcium channel in sarcoplasmic
46. Murayama T, Ogawa Y: Similar Ca 2+ dependences of ['H)ryanodine reticulum by calsequestrin. Biochem Biophys Res Commun 199: 1120-
binding tou- and ~-ryanodine receptors purified from bullfrog skeletal 1127,1994
muscle in an isotonic medium. FEBS Lett 380: 267-271, 1996 66. Shirokova N, Garcia J, Pizarro G, Rios E: Cal+ Release from the
47. O'Brien J, Valdivia HH, Block BA: Physiological differences between sarcoplasmic reticulum compared in amphibian and mammalian
the u and ~ ryanodine receptors offish skeletal muscle. Biophys J 68: skeletal muscle. J Gen Physiol 107: 1-18, 1996
471-482,1995 67. Volpe P, Simon BJ: The bulk ofCa 2+ released to the myoplasm is free
48. PercivalAL, WilliamsAJ, Kenyon JL, Grinsell MM,Airey JA, Sutko in the sarcoplasmic reticulum and does not unbind from calsequestrin.
JL: Chicken skeletal muscle ryanodine receptor isoforms: Ion channel FEBS Lett 278: 274-278, 1991
properties. Biophys J 67: 1834-1850, 1994 68. Ikemoto N, Nagy B, Bhatnagar GM, Gergely J: Studies on a metal
49. Oonoso P, Prieto H, Hidalgo C: Luminal calcium regulates calcium binding protein of the sarcoplasmic reticulum. J BioI Chern 249: 2357-
release in triads isolated from frog and rabbit skeletal muscle. Biophys 2365, 1974
J 68: 507-515, 1995 69. Ikemoto N: The calcium binding sites involved in the regulation of
50. Ogawa Y, Ebashi S: Ca-releasing action of beta, gamma-methylene the purified adenosine triphosphatase of the sarcoplasmic reticulum. J
adenosine triphosphate on fragmented sarcoplasmic reticulum. J Bioi Chern 249: 649-651,1974
Biochem 80: 1149-1157, 1976 70. Kitazawa T and Endo M: Increase in passive calcium influx into
51. Weber A, Herz R: The relationship between caffeine contracture of sarcoplasmic reticulum by 'depolarization' and caffeine. Proc Japan
intact muscle and the effect of caffeine on reticulum. J Gen Physiol Acad 52: 599-602, 1976
52: 750-759, 1968 71. Rios E, Pizarro G, Stefani E: Charge movement and the nature of
52. Ogawa Y: Some properties offragmented frog sarcoplasmic reticulum signal transduction in skeletal muscle excitation-contraction coupling.
with particular reference to its response to caffeine. J Biochem 67: Annu Rev Physiol54: 109-133,1992
667-683, 1970 72. Pape PC, Jong O-S, Chandler WK, Baylor SM: Effect of fura-2 on
53. Gutmann E, SandowA: Caffeine-induced contracture and potentiation action potential-stimulated calcium release in cut twitch fibers from
of contraction in normal and denervated rat muscle. Life Sci 4: 1149- frog muscle. J Gen Physiol 102: 295-332, 1993
1156, 1965 73. Jong D-S, Pape PC, Chandler WK, Baylor SM: Reduction of calcium
54. Giannini G, Clementi E, Ceci R, Marziali G, Sorrentino V: Expression inactivation of sarcoplasmic reticulum calcium release by fura-2 in
of a ryanodine receptor-Ca 2+ channel that is regulated by TGF beta. voltage-clamped cut twitch fibers from frog muscle. J Gen Physiol
Science 257: 91-94,1992 102: 333-370, 1993
 201

74. Lamb GD, Stephenson DG: Calcium release in skinned muscle fibres 19936-19942,1995
of the toad by transverse tubule depolarization or by direct stimulation. 78. Ikemoto T, lino M, Endo M: Enhancing effect of calmodulin on Ca 2+_
JPhysioI423:495-517,1990 induced Ca2+ release in the sarcoplasmic reticulum of rabbit skeletal
75. Corbett AM, Bian J, Wade JB, Schneider MF: Depolarization induced muscle fibres. J Physiol487: 573-582, 1995
calcium release from isolated triads measured with impermeant fura 79. Endo M, Vagi S, Ishizuka T, Horiuti K, Koga K, Amaha K: Changes
2. J Membrane Bioi 128: 165-179, 1992 in the Ca-induced Ca release mechanism in the sarcoplasmic reticulum
76. Ikemoto N, Yano M, EI-Hayek R, Antoniu B, Morii M: Chemical of the muscle from a patient with malignant hyperthermia. Biomed
depolarization-induced SR calcium release in triads isolated from rabbit Res 4: 83-92, 1983
skeletal muscle. Biochemistry 33: 10961-10968, 1994 80. Ohta T, Ito S, Ohga A: Inhibitory action of dantrolene on Ca-induced
77. Yano M, El-Hayek R, Ikemoto N: Role of calcium feedback in Ca2+release from sarcoplasmic reticulum in guinea pig skeletal muscle.
excitation-contraction coupling in isolated triads. J Bioi Chern 270: Eur J Pharmacol 178: 11-19, 1990
Molecular and Cellular Biochemistry 190: 203-204, 1998.

Index to Volume 190

Abe H, see Kusano K-I et al.

Endo M: Dedication 3-4

Fujita A, see Murahashi T et al.

Fujita K, Ye L-H, Sato M, Okagaki T, Nagamachi Y, Kohama K: Myosin light chain kinase from skeletal muscle
regulates an ATP-dependent interaction between actin and myosin by binding to actin 85-90
Fujita S, see Yoshida Y et al.

Gergely J: Professor Ebashi's impact on the study of the regulation of striated muscle contraction 5-8

Hagiwara Y, see Ozawa E et al.

Hartshorne DJ, Hirano K: Interactions of protein phosphatase type I, with a focus on myosin phosphatase 79-84
Hasegawa K, see Masuda T et al.
Hayashi K, see Sobue K et al.
Hidaka H, see Niki I
Hirano K, see Hartshorne DJ
Hirata M, Yoshida M, Kanematsu T, Takeuchi H: Intrinsic inhibitor of inositol 1,4,5-trisphosphate binding 179-184

lino M, see Masuda T et al.

lino M: Dynamic regulation of intracellular calcium signals through calcium release channels 185-190
Imai S, see Yoshida Y et al.
Islam MO, see Yoshida Y et al.

KalabokisVN, see Szent-Gyorgyi AG et al.

Kamidochi M, see Yazawa Y
Kanematsu T, see Hirata M et al.
Kawakita M, see Yamamoto H
Kawamura Y, see Soeno Y et al.
Kimura S, see Soeno Y et al.
Kitazawa T, see Murahashi T et al.
Koga T, see Yoshida Y et al.
Kohama K, see Fujita K et al.
Kurebayashi N, see Ogawa Y et al.
Kusano K-i, Abe H, Obinata T: Detection ofa sequence involved in actin-binding and phosphoinositide-binding
in the N-terminal side of cofilin 133-141

Maruyama K, see Soeno Y et al.

Masaki T, Ninomiya H, Sakamoto A, Okamoto Y: Structural basis of the function of endothelin receptor 153-156
Masuda T, Ohmi K, Yamaguchi H, Hasegawa K, Sugiyama T, Matsuda Y, lino M, Nonomura Y: Growing and
differentiating characterization of aortic smooth muscle cell line, p53LMACO I obtained from p53 knock out
mice 99-104
Matsuda Y, see Masuda T et al.
Murahashi T, FujitaA, Kitazawa T: Ca2+-induced Ca2+ desensitization of myosin light chain phosphorylation and
contraction in phasic smooth muscle 91-98
Murayama T, see Ogawa Y et al.
204

Nagamachi Y, see Fujita K et at.

Nakashima K-i, see Yazawa M et at.
Niki I, Hidaka H: Roles of intracellular Ca2+ receptors in the pancreatic ~-cell in insulin secretion 119-124
Ninomiya H, see Masaki T et at.
Nishida W, see Sobue K et at.
Nonomura Y, see Masuda T et at.

Obinata T, see Kusano K-I et at.

Obinata T, see Soeno Y et at.
Ogawa Y, Murayama T, Kurebayashi N: Comparison of properties ofCa 2+ release channels between rabbit and
frog skeletal muscles 191-201
Ohmi K, see Masuda T et at.
Ohtsuki I: Calcium ion regulation of muscle contraction: The regulatory role oftroponin T 33-38
Okagaki T, see Fujita K et at.
Okamoto Y, see Masaki T et at.
Ozawa E, Hagiwara Y, Yoshida M: Creatine kinase, cell membrane and Duchenne muscular dystrophy 143-151

Perreault-Micale CL, see Szent-Gyorgyi AG et at.

Perry SV: Troponin I: Inhibitor or facilitator 9-32

Sakamoto A, see Masaki T et at.

Sato M, see Fujita K et at.
Sobue K, Hayashi K, Nishida W: Expressional regulation of smooth muscle cell-specific genes in association
with phenotypic modulation 105-118
Soeno Y, Yajima H, Kawamura Y, Kimura S, Maruyama K, Obinata T: Organization of connectin/titin filaments
in sarcomeres of differentiating chicken skeletal muscle cells 125-131
Sugiyama T, see Masuda T et at.
Suzuki Y, see Tanokura M
Szent-Gyorgyi AG, KalabokisVN, Perreault-Micale CL: Regulation by molluscan myosins 55-62

Takeuchi H, see Hirata M et at.

Tanokura M, Suzuki Y: A phosphorus-31 nuclear magnetic resonance study on the complex of chicken gizzard
myosin subfragment I with adenosine diphosphate 75-78
Toyosato A, see Yoshida Y et at.

Weber A: Actin binding proteins that change extent and rate of actin monomer-polymer distribution by different
mechanisms 67-74

Yagi K, see Yazawa M et a/

Yajima H, see Soeno Y et at.
Yamada K: Thermodynamic analyses of calcium binding to troponin C, calmodulin and parvalbumins by using
microcalorimetry 39-45
Yamaguchi H, see Masuda T et at.
Yamamoto H, Kawakita M: Chemical modification of an arginine residue in the ATP-binding site of Ca2 +_
transporting ATPase of sarcoplasmic reticulum by phenylglyoxal 169-177
Yazawa M, Nakashima K-i, Yagi K: A strange calmodulin of yeast 47-54
Yazawa Y, Kamidochi M: The properties and function of invertebrate new muscle protein 63-66
Ye L-H, see Fujita K et at.
Yoshida M, see Hirata M et at.
Yoshida M, see Ozawa E et at.
YoshidaY, ToyosatoA, Islam MO, Koga T, Fujita S, Imai S: Stimulation of plasma membrane Ca2+ -pump ATPase 157-167
ofvascular smooth muscle by cGMP-dependent protein kinase: Functional reconstitution with purified proteins
Developments in Molecular and Cellular Biochemistry

Series Editor: Naranjan S. Dhalla, Ph.D., M.D. (Hon.), FACC

1. VA. Najjar (ed.): Biological Effects of Glutamic Acid and Its Derivatives. 1981 ISBN 90-6193-841-4
2. VA. Najjar (ed.): Immunologically Active Peptides. 1981 ISBN 90-6193-842-2
3. VA. Najjar (ed.): Enzyme Induction and Modulation. 1983 ISBN 0-89838-583-0
4. VA. Najjar and L. Lorand (eds.): Transglutaminase. 1984 ISBN 0-89838-593-8
5. GJ. van der Vusse (ed.): Lipid Metabolism in Normoxic and Ischemic Heart. 1989 ISBN 0-7923-0479-9
6. I.F.C. Glatz and GJ. van der Vusse (eds.): Cellular Fatty Acid-Binding Proteins. 1990
ISBN 0-7923-0896-4
7. H.E. Morgan (ed.): Molecular Mechanisms of Cellular Growth. 1991 ISBN 0-7923-1183-3
8. GJ. van der Vusse and H. Starn (eds.): Lipid Metabolism in the Healthy and Diseased Heart. 1992
ISBN 0-7923-1850-1
9. Y. Yazaki and S. Mochizuki (eds.): Cellular Function and Metabolism. 1993 ISBN 0-7923-2158-8
10. I.F.e. Glatz and GJ. van der Vusse (eds.): Cellular Fatty-Acid-Binding Proteins, II. 1993
ISBN 0-7923-2395-5
11. R.L. Khandelwal and J.H. Wang (eds.): Reversible Protein Phosphorylation in Cell Regulation. 1993
ISBN 0-7923-2637-7
12. I. Moss and P. Zahradka (eds.): ADP-Ribosylation: Metabolic Effects and Regulatory Functions. 1994
ISBN 0-7923-2951-1
13. VA. Saks and R. Ventura-Clapier (eds.): Cellular Bioenergetics: Role of Coupled Creatine Kinases. 1994
ISBN 0-7923-2952-X
14. I. Slezak and A. Ziegelh6ffer (eds.): Cellular Interactions in Cardiac Pathophysiology. 1995
ISBN 0-7923-3573-2
15. I.A. Barnes, H.G. Coore, A.H. Mohammed and R.K. Sharma (eds.): Signal Transduction Mechanisms. 1995
ISBN 0-7923-3663-1
16. A.K. Srivastava and J.-L. Chiasson (eds.): Vanadium Compounds: Biochemical and Therapeutic Applications.
1995 ISBN 0-7923-3763-8
17. I.MJ. Lamers and P.O. Verdouw (eds.): Biochemistry of Signal Transduction in Myocardium. 1996
ISBN 0-7923-4067-1
18. E.-G. Krause and R. Vetter (eds.): Biochemical Mechanisms in Heart Function. 1996
ISBN 0-7923-4118-X
19. R. Vetter and E.-G. Krause (eds.): Biochemical Regulation ofMyocardium. 1996 ISBN 0-7923-4259-3
20. G.N. Pierce and W.C. Claycomb (eds.): Novel Methods in Molecular and Cellular Biochemistry of Muscle.
1997 ISBN 0-7923-4387-5
21. F.N. Gellerich and S. Zierz (eds.): Detection ofMitochondrial Diseases. 1997 ISBN 0-7923-9925-0
22. P.K. Singal, V. Panagia and G.N. Pierce (eds.): The Cellular Basis of Cardiovascular Function in Health and
Disease. 1997 ISBN 0-7923-9974-9
23. S. Abdel-aleem and I.E. Lowe (eds.): Cardiac Metabolism in Health and Disease. 1998
ISBN 0-7923-8104-1
24. A.K. Srivastava and B. Posner (eds.): Insulin Action. 1998 ISBN 0-7923-8113-0
Developments in Molecular and Cellular Biochemistry

25. v.A. Saks, R. Ventura-Clapier, X. Leverve, A. Rossi and M. Rigoulet (eds.): Bioenergetics of the Cell:
Quantitative Aspects. 1998 ISBN 0-7923-8118-1
26. G.N. Pierce, H. Rupp, T. Izumi and A. Grynberg (eds.): Molecular and Cellular Effects of Nutrition on
Disease Processes. 1998 ISBN 0-7923-8171-8
27. K. Ahmed, E. Chambaz and O.G. Issinger (eds.): Molecular and Cellular View ofProtein Kinase CK2. 1998
ISBN 0-7923-8208-0
28. M.V. Cohen, J.M. Downey, R.I. Gelpi and J. Slezak (eds.): Myocardial Ischemia and Repeifusion. 1998
ISBN 0-7923-8173-4
29. D.A. Bernlohr and L. Banaszak (eds.): Lipid Binding Proteins within Molecular and Cellular Biochemistry.
1998 ISBN 0-7923-8223-4
30. R. Alvarez-Gonzalez (ed.): ADP-Ribosylation Reactions: From Bacterial Pathogenesis to Cancer. 1998
ISBN 0-7923-8235-8
31. S. Imai and M. Endo (eds.): Muscle Physiology and Biochemistry. 1998
ISBN 0-7923-8264-1

KLUWER ACADEMIC PUBLISHERS - DORDRECHT / BOSTON / LONDON