Molecular mechanisms of

drug action
Molecular mechanisms of
drug action
Second Edition

Christopher J. Coulson

Glaxo Group Research, UK

UK Taylor & Francis Ltd, 4 John St., London WCIN 2ET

USA Taylor & Francis Inc., 1900 Frost Road, Suite 101, Bristol, PA 19007

Copyright 0 Taylor & Francis Ltd 1994

All rights reserved. No part of this publication may be reproduced,

stored in a retrieval system, or transmitted, in any form
or by any means, electronic, electrostatic, magnetic tape,
mechanical, photocopying, recording or otherwise, without
the prior permission of the copyright owner.

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British
Library.

ISBN 0-7484-0068-O (cloth)

ISBN o-7484-0078-8 (paper)

Library of Congress Cataloging-in-Publication Data are available

Cover design by Russell Beach

Typeset by Photo.Graphics
Printed in Great Britain by Burgess Science Press, Basingstoke, on
paper which has a specified pH value on final manufacture of not
less than 7.5 and is therefore @acid free’.
 Contents

Preface

Preface to second edition

Glossary

1. General principles 1

1.1 Background 1
1.2 Do drugs have a specific mode of action? 1
1.3 Basic processes 2
1.4 Drug binding to enzymes 3
1.5 Drug binding to receptors 5
1.6 Further considerations 6
1.7 Partition coefficient 7
1.8 Drug nomenclature 8
1.9 Stereochemistry 9
1.10 The future 11
References 11

2. Nucleic acid biosynthesis and catabolism 13

2.1 Introduction 13
2.1.1 Overall scheme 15
2.2 Nucleotide biosyntheses - enzyme targets of drugs 18
2.2.1 Dihydroorotate dehydrogenase - actovaquone as
antimalarial 18
2.2.2 Dihydropteroate synthetase - sulphonamides as anti-
bacterials 19
2.2.3 Dihydrofolate reductase - trimethoprim and
pyrimethamine as anti-bacterials, methotrexate as
anti-cancer 21
2.2.4 Ribonucleotide reductase - hydroxyurea as anti-cancer 23
2.2.5 Thymidylate synthetase - 5-fluorouracil as anti-cancer,
5-fluorocytosine as antifungal 25

V
vi Molecular mechanisms of drug action

2.2.6 Inosine monophosphate dehydrogenase - ribavirin as

anti-viral 26
2.3 DNA biosynthesis 28
2.3.1 Herpesvirus DNA polymerase - acyclovir and
vidarabine as anti-virals 30
2.3.2 Mammalian DNA polymerase - cytarabine as anti-
leukaemic 32
2.3.3 DNA topoisomerases 33
2.3.4 Reverse transcriptase - azidothymidine ror AIDS 38
2.3.5 Bacterial RNA polymerase - rifampicin as
antimycobacterial 40
2.4 Nucleotide catabolism 40
2.4.1 Adenosine deaminase - 2deoxycoformycin as anti-
leukaemic 41
2.4.2 Xanthine oxidase - allopurinol for gout 43
2.4.3 Guanylate cyclase 45
Questions 46
References 47

3. Protein biosynthesis 51

3.1 Introduction 51
3.2 Aminoglycosides 54
3.3 Chloramphenicol 57
3.4 Tetracyclines 59
3.5 Erythromycin 60
3.6 Clindamycin 61
Questions 62
References 63

4. Carbohydrate metabolism 65

4.1 Introduction 65
4.2 Metronidazole action - anaerobic protozoa1 and bacterial
infections 66
4.3 Medical use of metronidazole 69
4.4 a-Glucosidase inhibitors - acarbose as antidiabetic 70
4.5 Sialidase inhibitor-4-guanidino-NeuSAcZen for influenza 71
Questions 72
References 72

5. Cell wall biosynthesis 75

5.1 Introduction 75
5.2 Plactams 77
5.2.1 Penicillin-binding proteins 78
 Contents vii

5.2.2 PIactamases 81
5.2.3 PLactam therapy 84
5.3 Vancomycin 86
5.4 Mycolic acid synthesis - isoniazid for tuberculosis 88
5.5 PGlucan synthetase 90
Questions 90
References 91

6. Steroid biosynthesis and action 95

6.1 Introduction 95
6.2 Sterol biosynthesis 100
6.2.1 PHydroxy-pmethylglutaryl CoA reductase - lovastatin
as hypocholesterolaemic 100
6.2.2 Squalene epoxidase - naftitine as antifungal 103
6.2.3 Lanosterol demethylation - ketoconazole as antifungal 104
6.3 Steroid biosynthesis 107
6.3.1 Steroid 17,20-lyase - ketoconazole for steroid-
dependent tumours 107
6.3.2 11 PSteroid hydroxylase - metyrapone for
hypercortisolism 108
6.3.3 Aromatase - aminoghttethimide for hormone-
dependent cancers 109
6.3.4 So-Reductase 110
6.4 Steroid receptor ligands 111
6.4.1 Oestrogen/progesterone receptor agonists - oral
contraceptives 111
6.4.2 Oestrogen antagonists - tamoxifen for oestrogen-
dependent cancer, clomiphene to stimulate
ovulation 114
6.43 Progesterone antagonist - mifepristone as abortifacient 116
6.4.4 Androgen antagonist 116
6.4.5 AIdosterone antagonist - spironolactone as diuretic 117
Questions 119
References 120

7. Prostaglandin and leukotriene biosynthesis and

ZWtiOll 123

7.1 Introduction 123

7.2 Phospholipase inhibition - glucocorticoids for asthma and
arthritis 124
7.3 Prostaglandin synthetase inhibitors - aspirin and other non-
steroidal anti-inflammatories 127
7.4 Rheumatoid arthritis 130
7.5 Thromboxane synthesis 132
 ..a
Vlll Molecular mechanisms of drug action

7.6 Leukotriene biosynthesis 132

Questions 133
References 133

8. Zinc metalloenzymes 135

8.1 Introduction 135
8.2 Carbonic anhydrase - methazolamide for glaucoma 136
8.3 Angiotensin-converting enzyme 141
8.3.1 Angiotensin-converting enzyme inhibitors - captopril
and enalapril for hypertension 145
8.3.2 Pharmacology of ACE inhibitors 146
8.4 Endopeptidase 24.11 - thiorphan as analgesic 148
Questions 149
References 150

9. Neurotransmitter action and metabolism 151

9.1 Introduction 151
9.1.1 G-Protein linked receptor structure 154
9.1.2 Signal transduction and receptor occupancy 155
9.1.3 Drug development 159
9.2 Adrenergic receptors 160
9.2.1 cu-Adrenergic receptors 161
9.2.2 PAdrenergic receptors 163
9.2.3 Mixed cr- and @unagonist 166
9.3 Dopamine receptors 166
9.3.1 Dopamine agonists - L-dopa and bromocryptine for
Parkinsonism, fenoldapam for hypertension 167
9.3‘2 Dopamine antagonists - phenothiazines,
butyrophenones and diphenylbutylpiperidines for
schizophrenia 169
9.4 Serotonin receptors 174
9.4.1 5HT,, receptors, anxiety and depression 175
9.4.2 5HT,, receptors and migraine 176
9.4.3 5HT, receptors 178
9.4.4 5HT, receptors 178
9.5 Serotonin and noradrenaline re-uptake mechanisms - tricyclic
anti-depressants 180
9.6 Monoamine oxidase - tranylcypromine and moclobemide for
depression, deprenyl for Parklnsonism 183
9.6.1 Monoamine oxidase and Parkinsonism 187
9.7 Acetylcholine action 188
9.7.1 Muscarinic receptor - pirenzepin for ulcers, atropine
pre-anaesthetic medication, pilocarpine for
glaucoma 189
 Contents ix
9.7.2 Acetylcholinesterase - pyridostigmine for myasthenia
gravis 192
9.8 4Aminobutyric acid receptor - benzodiazepines as hypnotics,
avermectin as anthehninthic, baclofen for spastic@ 194
9.8.1 Benzodiazepines 195
9.8.2 Avermectin 197
9.8.3 Baclofen 200
9.9 Opiate receptors - morphine for pain 201
9.9.1 Opioid dependence 204
9.92 Pentazocine as analgesic 204
9.9.3 Loperamide as anti-diarrhoea1 205
9.10 Histamine receptors - mepyramine as anti-allergic, cimetidine
as anti-ulcer 206
9.10.1 H, receptor antagonists 207
9.10.2 H, receptor antagonists 208
Questions 210
References 210

10. Membrane-active agents 215

10.1 Introduction 215
10.1.1 Membrane structure 215
10.1.2 Dynamics of the heart beat 217
10.2 The sodium channel 219
10.2.1 Sodium channel blockers - procaine as a local
anaesthetic 220
10.2.2 Anti-arrhythmic agents - lidocaine as an anti-
arrhythmic 222
10.2.3 Diuretics - amiloride and triamterene 223
10.3 The calcium channel 225
10.3.1 Calcium channel antagonists - verapamil, nifedipine
for hypertension and heart failure 227
10.4 Coupled sodium-chloride ion channels - furosemide and
ethacrynic acid as diuretics 230
10.5 Membrane-bound ATPases 231
10.5.1 Sodium-potassium-ATPase - cardiac glycosides for
heart failure 232
10.5.2 Potassium-hydrogen-ATPase - omeprazole for ulcers 236
10.6 CromogIycate - calcium antagonist or membrane stabilizer? 239
10.6.1 Cromoglycate action 239
10.7 Cyclosporin 241
10.8 Potassium channel opening 243
10.9 Sterol Iigands - polyene antibiotics 244
10.9.1 Amphotericin action 245
Questions 247
References 248
X Molecular mechanisms of drug action

11. Mlcrotubule assembly 251

11.1 Introduction 251
11.2 Colchicine for gout 252
11.3 Vinca alkaloids as anti-tumour agents 255
11.4 Griseofulvin as an antifungal agent 256
11.5 Benzimidazoles as anthehninthics 257
11.6 Tax01 as an anti-tumour agent 259
Questions 260
References 260

12. Hormonalmodulators 261

12.1 Introduction 261

12.2 Diabetes mellitus 261
12.2.1 The action of insulin 262
12.2.2 Insulin therapy 264
12.3 Sulphonylureas as hypoglycaemic agents 266
12.4 Gonadotrophin-hormone-releasing hormone (GnRH) analogues 268
Questions 273
References 273

Appendix: Quantifkationoflipd-ma~omole~ebhdhg 275

A.1 Enzyme kinetics: I,, or Ki? 275
A.2 Drug binding to receptors 278
A.3 Ligand-protein binding 280
References 282

IlRdeX 285
 Preface

As my experience is that of an industrial biochemist who has worked in the

pharmaceutical industry for over 20 years, this book is written from a practical
standpoint. It includes the mechanism of drug action, encompassing drugs for
infectious disease, as well as for ‘endogenous conditions’ such as cancer,
arthritis, heart disease, schizophrenia etc. It is written from the point of view
of targets - whether enzymes in pathways, receptors or ion channels in mem-
branes.
The alternative procedure, where disease is made the central focus, has been
used in a number of other publications and so I have approached the subject
in a different way which was better able to link the practical aspects of drug
targeting with the biochemical and pharmacological background. My aim was
to bring together our present state of knowledge with respect to drug mechan-
isms (information that is otherwise scattered throughout the literature) in a
readable and accessible form, which could assist both those teaching the sub
ject, and students who wish to study it. I felt that there was a gap between
academia and industry which such a book might help to close.
The introductory chapter is concerned with the basic principles that cover
enzyme inhibition and receptor binding by drugs. The rest of the book is div-
ided into two sections; the next seven chapters deal with drugs that modulate
biochemical pathways, both of synthesis and breakdown, while the last four
are concerned with organizational structures of the cell. Chapters 2 to 5 are
concerned with the biosynthesis of DNA, protein, carbohydrate and cell walls.
Much of the chemotherapy of cancer, viruses and bacteria is to be found
within these chapters. Next are two chapters on lipid biosynthesis; the first
one covering the pathways of sterol and steroid interconversions and the
second on the various pathways that lead from arachidonic acid. Although not
a pathway, I have next included a discussion on inhibitors of zinc metalloen-
zymes as they form a coherent group.
In the second section the agonists and antagonists at neurotransmitter recep-
tors are discussed first, followed by those agents that interfere directly with
membranes. In these two chapters I have deliberately covered membrane
enzymes alongside ion channels and receptors. Microtubule ligands form the
subject of Chapter 11 and hormone modulators are discussed in Chapter 12.
In the Appendix, I discuss the basis for the measurement of binding constants
for ligand/macromolecule binding.

xi
xii Molecular mechanisms of drug action

Some readers may feel that there are notable lacunae in the coverage of
pharmaceuticals. This has occurred because I have tried to cover those drugs
that show recognized principles in their mode of action; some have not been
included because their mode of action is not known and the work done on
them cannot be related in a coherent fashion in a textbook of this sort. No
attempt has been made to cover all drugs. Others are included, even though
their mode of action is not fully understood, if their development illustrates a
useful point. The use of cromoglycate in asthma is such a case. Other drugs
of interest have been discussed if they have been designed particularly for
a given condition and may be approaching the market - although not yet
fully launched.
I have not tried to cover drug delivery systems, metabolism, side-effects etc.,
except where these are germane to the mechanism. The quantity of material
available in these areas is so great that the length of the book would have
reached unmanageable proportions if it had been included. Furthermore, the
amount of coverage does not relate to market value or quantity of drug sold,
but rather to scientific interest, so that a widely used drug whose mode of
action is clearly defined may not warrant a long coverage, while an interesting
but less used drug may have a greater coverage.
The book is intended for third, and possibly second, year students studying
subjects or modules in microbiology, pharmacology, biochemistry, pharmacy
and medicine. It may also be useful for medicinal scientists in the drug industry
who are changing fields and need a quick entree into a fresh area. Comprehen-
sive reviews have been listed, if available, to help the reader to follow up
points of interest and to go more deeply into the subject. These are largely in
accessible journals and the occasional book.
I am deeply indebted to Charles Ashford, John Foreman, Ian Kitchen, Hugh
White, Alan Wiseman, Helen Wiseman and David Wiiams for reading all of
part of the manuscript and who made many useful suggestions for improve-
ment and corrections.
 Preface to second edition

A few new drugs (and some old ones) have been added if their mode of action
is sufficiently clear. Other drugs that have been withdrawn from the market
have been omitted. As a result of responses to the first edition, questions have
been added at the end of each chapter which it is hoped will help in studying
the subject. Some questions can be answered by reading the chapter alone,
while others will require some background reading from the review articles
mentioned.
A much greater emphasis has been placed on chirality as it is becoming a
major feature of drug development. If the information on drug isomers
developed in earlier years is not available, this point is usually noted in the
text. The techniques of molecular biology have permitted enormous advances
in our understanding. I have made references to these wherever appropriate.
Some drugs are known by different generic or non-proprietary names in
different parts of the world. A glossary of those that differ markedly between
Britain and the USA is included to clarify the situation for the reader. As a
general rule the USA adopted name includes the counterion of the salt if appro-
priate but the British approved name does not. The name used here is that
commonly accepted.

. ..
xm
 Glossary

British Approved Name @AN) United States Adopted Name (USAN)

Adrenaline Epinephrine
Azidothymidine Zidovudine
Cromoglycate Cromolyn
Frusemide Furosemide
Glibenclamide Glyburide
Isoprenaline Isoproterenol
Lignocaine Lidocaine
Mepyramine Pyrilamine
Noradrenaline Norepinephrine
FWampicin Rifampin
Salbutamol Albuterol
 Chapter 1

General principles

1.1 Background

It would have been impossible to write a book of any length detailing mechan-
isms of drug action more than 20 years ago, because very little was known at
that time. Penicillin, for example, had been on the market for 20 years, but
we were still far from detailing its precise mode of action. The philosophical
attitude, prevalent in those days and still common now, was that drugs were
just used because they worked: enquiry into their mode of action was deemed
unnecessary. The fact that, at that time, our detailed understanding of biologi-
cal processes was very limited, may not be unconnected. Many of us entering
the drug industry in the 1960s hoped, however, to be able to design drugs
more effectively on the basis of molecular structure, whether it be of the
enzyme or the receptor.
There have been a number of major advances in the intervening years. It
was, after all, only 23 years ago that the connection between aspirin and
prostaglandin biosynthesis was realized by Vane (1971) which led in its turn
to a much greater understanding of the biosynthesis of prostanoids. In fact,
drugs have frequently provided researchers in the medical sciences with tools
to unravel a ‘knotty’ scientific and/or medical problem.

1.2 Do drugs have a specific mode of action?

Perhaps one question that should be considered is whether drugs do have a

singular mechanism of action that can be reduced to molecular terminology.
The concept of a drug as a ‘magic bullet’ that can penetrate through a myriad
of biochemical systems, acting at only one specific site to have the desired
effect, is perhaps reductionist to the point of absurdity. Nevertheless, in practi-
cal terms the history of drug development suggests that, in the initial analysis,
it is possible to view the process in this fashion. The basis for this assertion
lies in the linking of enzyme inhibition, receptor binding or ion channel block-
ing with expected effects on substrate levels or events inside and outside the
cell. These must lead to organ effects or death of parasitic microorganisms,

1
2 Molecular mechanisms of drug action
and improvement of the patient’s condition. These factors must all be consist-
ent with the proposed mode of action.
Nevertheless, drugs frequently act at sites other than the intended one -
molecular, cellular and/or organ. As a consequence, they show undesirable
side-effects. Indeed, the process of developing a drug is intended to reduce
undesirable activities to an acceptable level. Unless the aspect of the structure
of the drug molecule that gives rise to the activity also gives rise to the side-
effect, it is usually possible to reduce the undesirable activity by making minor
changes to the structure.
In the last analysis the crucial factor is the therapeutic ratio. This is the ratio
between the dose required to treat the condition and the dose which gives
rise to unacceptable side-effects. There are occasions when the side-effect is
apparently random (rare but often lethal), e.g. the development of aplastic
anaemia associated with some drugs, notably chloramphenicol (see Chapter
3). Difficult decisions then have to be made, weighing up the possible risk to
the patient against the benefit.
Another factor that affects such a decision is the condition for which the
drug is prescribed. A drug for a condition that is frequently lethal (such as
cancer) can be tolerated with a lower therapeutic ratio than one intended
for reducing blood pressure. For viral diseases very few drugs are available;
permission might therefore be given to use a drug with a low safety margin
if there was no other drug available for a certain condition.
We must not forget that it is a patient who has to take the selected drug.
The side-effects, therefore, need to be minimal and not of such a nature as
not to be in any way distressing. At present, our need is for effective, non-
toxic drugs to treat a variety of conditions in a preventive or palliative manner.

1.3 Basic processes

Drugs can be used to help the body reject an invading pathogenic organism
(whether parasite, fungus, bacterium or virus) or to modify some aspect of
the metabolism of the body that is functioning abnormally. In the former case,
a drug is normally used which is toxic to the pathogenic organism but not to
the host, and stimulation of the body’s normal processes for combating
invaders may play a part. In the latter case, the approach is usually more subtle,
with a modulation of a process as the requirement.
In all cases, the drug can have biological activity only by interacting with
the molecules of the target organ or organism. These molecules are usually
proteins - enzymes that catalyse reactions essential for the functioning of the
organism, or proteins called receptors which transmit signals by interacting
with messenger molecules such as hormones and neurotransmitters. The drug
has its effect by binding to either an active site or to a secondary site that
influences the active site on the enzyme or receptor and thereby prevents
access by the normal substrate or ligand (inhibitor or antagonist) or it provokes
a signal where none was wanted (substrate or agonist).
 General principles 3

The interaction of the drug with the protein molecule can be measured in
terms of the strength of inhibition of an enzyme reaction or the strength of
drug binding to a receptor. The former may be quantified as I,,, or I&,, the
concentration of inhibitor required to reduce the rate of a reaction or the
binding of a ligand by one-half. This, however, varies with the amount of
substrate or ligand available to the enzyme or receptor, thereby making com-
parisons between data obtained under different conditions almost impossible.
With more data, a binding constant for ligand attaching to receptor or inhibi-
tor constant for enzyme reaction may be determined. These quantities are
usualIy expressed in terms of the dissociation constant of the ligand-receptor
complex (KJ or enzyme-inhibitor complex (Ki). For the ligand-receptor inter-
action, this constant is equal to the ligand concentration at which 50 per cent
of the receptors are occupied by ligand and it is assumed, although this is not
always correct, that the measured response has fallen to one-half of its
maximum value. For typical drug-receptor interactions, the dissociation con-
stants are of the order of lo-’ to 10-i’ M.

1.4 Drug binding to enzymes

Inhibitors can bind to enzymes in several ways. They may bind either reversi-
bly or irreversibly, and with either the active site or with another part of the
enzyme. In the latter case, the binding can cause a conformational change in
the enzyme, thus producing distortion of the active site. Different effects of
inhibitor concentration upon the kinetics of the enzyme in these various cases
allow evidence to be obtained on the type of binding involved in any particu-
lar situation.
Reversible inhibition can occur in various ways including competitive, non-
competitive and uncompetitive binding. If the inhibitor binds to the enzyme
alone, the inhibition is said to be competitive. In this case, the inhibitor con-
stant (KJ is defined as the dissociation constant of the enzyme/inhibitor com-
plex (see Appendix for a more detailed discussion of KJ. In principle, increas
ing the substrate concentration can eventually restore the reaction rate to what
it was in the absence of inhibitor. Frequently, the substrate competes with
the inhibitor for the substrate binding site, but there are cases in which com-
petitive inhibition is observed although substrate and inhibitor bind to dis-
tinct sites.
In non-competitive inhibition the inhibitor can bind to the enzyme-substrate
complex as well as to the enzyme itself, at a site separate from the substrate
binding site. In this case, increasing the substrate concentration may reduce
but cannot eliminate the inhibition. Uncompetitive inhibition is found when
the inhibitor will only bind to the enzyme-substrate complex.
The binding constants noted above are, in principle, independent of ligand
or substrate concentration and so results obtained in different laboratories can
be compared with some confidence. They do not, however, necessarily bear
4 Molecular mechanisms of drug action

any relation to the 1% values mentioned above. (See Appendix 1 for a

discussion.)
The links between drug and target are usually of a reversible nature and so
the drug will eventually be released from the complex. The complex is usually
formed instantaneously, although this can take many hours if the binding con-
stant is of the order of 10e9 M or smaller. In some cases, however, the drug and
target will react covalently, and this is often chamcterized by time-dependent
inhibition and by lack of enzyme recovery after dialysis. If the ligand is bound
reversibly, it dissociates during dialysis and enzyme activity is recovered. If
irreversible inhibition has occurred, the activity is not recoverable; instead
synthesis of fresh protein will be needed to restore the original level of the
target. Some reversible inhibitors whose binding is extremely tight may cycle
on and off the enzyme so slowly that for all practical purposes the enzyme is
inactivated by their presence.
Tightly binding inhibitors therefore need a considerable time for preincu-
bation for the inhibition to reach its maximum effect. They do not satisfy the
normal kinetic criteria of enzyme reactions in that the equilibrium is not set
up rapidly and the concentrations of inhibitor and substrate are not very much
higher than that of the enzyme. In fact, the inhibitor and enzyme concen-
trations are often very similar and in some cases stoichiometric or one-to-one.
Under these conditions the Lineweaver-Burk plots (double reciprocal plots of
reaction velocity against substrate concentration) will have both curved and
linear portions while the progress curves will be aptly named because, in
effect, a portion of the enzyme is being put gradually out of action (Morrison,
1969, 1982). An example is methotrexate binding to dihydrofolate reductase
with a Kd of lo-ii M (Chapter 2).
Irreversible inhibition is particularly useful if the step in a pathway chosen
as the target for drug action is not the rate-limiting one. Reversible inhibitors
are likely to be ineffective in these situations because substrate levels increase
and, even with noncompetitive inhibitors, the inhibition is reduced. Most
drugs are administered at ieast on a daily basis. Irreversible inhibitors, how-
ever, probably do not need to be administered daily, since it may take several
days for enough enzyme to be resynthesized to return to steady state levels.
A dosage regimen in which the drug is only given on alternate days is likely
to be more useful than one requiring daily dosage. Otherwise the effect of the
drug is likely to become magnified to the point of producing side-effects. A
case in point is the use of monoamine oxidase inhibitors (Chapter 9), which
were given on a daily basis and eventually lowered the level of the enzyme
in the stomach wall to such a level that tyramine and dopamine ingested from
food were not deaminated and caused hypertensive crises that were sometimes
lethal. This was known as the ‘cheese effect’ because cheese contains a con-
siderable amount of tyramine.
A special type of irreversible inhibition occurs when the drug is chemically
unreactive but the enzyme converts it to a highly reactive species that inacti-
vates the enzyme. This has been given the name of mechanism-based irrevers-
 General principles 5

ible inhibition and the drugs are called suicide substrates or Kc, inhibitors
because they require the catalytic activity of the enzyme. Examples of this
include allopurinol inhibiting xanthine oxidase and 5-fluorodeoxyuridylate
inhibiting thymidylate synthetase (both in Chapter 2) clavulanate inhibiting
plactamase (Chapter 5) and the monoamine oxidase inhibitors (Chapter 9).
For a fuller discussion see Walsh (1984).
Another class of enzyme inhibitors is described as transition state inhibitors.
Every reaction proceeds through an intermediate state in which the structure
attached to the enzyme is neither substrate nor product but an intermediate
form. The transition state is the structure of highest energy on the pathway
from substrate to product. As enzymes catalyse a specific reaction, compounds
that interfere with the transition state are likely to be highly selective in their
action. One example of a transition state inhibitor is 2deoxycoformycin which
inhibits adenosine deaminase (Chapter 2) (Lienhard, 1973).
An inhibitor does not always bind to the active site of an enzyme. It may
modulate the enzyme activity by binding to another part of the protein, and
is called an allosteric effector - allosteric is derived from Greek words meaning
‘other shape’. Allosteric effecters are thus distinguished from those competi-
tive inhibitors which resemble the substrate in shape, and are termed isosteric.
Inhibition of this type is also known as negative cooperativity, whereas acti-
vators that bind allosterically give rise to positive cooperativity. An enzyme
that behaves in this way is normally composed of more than one subunit.
Ribonucleotide reductase is a case in point where the effecters are the purine
and pyrimidine trinucleotides (Chapter 2).

1.5 Drug binding to receptors

It is important to be clear about what constitutes a receptor. For the binding

of a ligand to a receptor to be a genuine physiological phenomenon and not
just non-specific binding, the ligand binding should:

1. be saturable, in which case a plot of amount of drug bound against drug

concentration will level off and reach a plateau;
2. be high alfinity (binding constants less than lop6 M) and low capacity for
the binding to be regarded as specific;
3. obey the Law of Mass Action (Appendix); and
4. be linked to a pharmacological response characteristic of the particular
ligand.

If these criteria are not fulfilled, the data may indicate, fortuitously, the exist-
ence of a binding protein for a particular ligand but not a receptor with physio-
logical significance. It is also advisable, if possible, to have a series of agonists
and antagonists, preferably of different structural types, that are specific for a
given receptor and can thus define its role.
6 Molecular mechanisms of drug action

An agonist is an agent that interacts with the receptor to produce a clearly

defined change in the target cell. An antagonist also binds to the receptor, but
lacks the structural features essential to initiate any further changes and so can
block the action of an agonist. In a cell-free system, therefore, both agonist
and antagonist will bind, but it will be impossible to distinguish between them
without data on activity on a cellular or tissue basis. A partial agonist shows
activity at a receptor but not to the same extent as the natural agonist.
The measurement of the reversible binding of a radioactive ligand to a recep-
tor preparation (radioligand binding) has greatly increased our understanding
of receptors and, provided it is understood that the binding occasionally may
have to be discounted because it is an artefact of the preparation, the knowl-
edge gained has been of great value in identifying and, in some cases, isolating
the receptor. One of the greatest challenges in modem pharmacology is (a)
to link the binding data with pharmacological effect and (b) to understand
how in molecular terms the signal is transduced in a given cell to produce a
given response (Chapter 9).

1.6 Further considerations

The attachment of a drug to its target is only part of the process that leads to
an effective drug. Anti-bacterial drugs have to be able to reach the bacterium
in sufficient quantity to be able to kill it (bactericidal action). In some cases
the drug merely prevents the bacterium from growing and does not kill it
(bacteristatic action), but the host may still recover from the infection by the
anti-bacterial action of the immune system. Similarly, some antifungal drugs
merely stop fungal growth, but here the situation is more serious because
fungal infections frequently occur in individuals with impaired immune sys-
tems, for example in cancer patients or in patients who are receiving immune
suppressant drugs after an organ transplant. It is therefore most important to
have fungicidal drugs.
The activity of anti-bacterial and antifungal drugs is usually measured by the
minimum concentration at which the drug completely inhibits the growth of
the microorganism. This concentration is known as the minimum inhibitory
concentration (MIC).
Furthermore, absorption, serum binding and metabolism must all be taken
into account in developing an effective drug. The drug, if given by mouth,
must be able to cross the stomach wall, reach the bloodstream, and be trans
ported in sufficient quantity to the site of action. Serum proteins may play a
part in this process - notably albumin, which is particularly effective at carry-
ing drugs that exist in anionic form at pH 7.4, such as the non-steroidal anti-
inflammatory drugs (Chapter 7). Albumin binding can, however, if too strong,
prevent a compound that shows good activity in vitro from demonstrating
efficacy in vivo (e.g. clorobiocin activity against Gram-negative bacteria is a
case in point, Chapter 2). Other compounds may fail to show activity in vivo
because they are metabolized to inactive metabolites.
 General principles 7

I. 7 Partition coefficient

A partition coefficient is a measure of how a solute distributes itself between

two immiscible solvents. Often denoted P and expressed as the logarithm (log
P) it is particularly useful in describing the potential of a drug for reaching
the brain. The brain (and the spinal cord) is surrounded by a fatty sheath
and a considerable degree of lipid partitioning or lipophilicity (i.e. preferential
transfer into lipid, not just solubility) is required for the drug to cross into the
brain. Molecular weight and charge are also of importance. The concept of
blood-brain transport is of crucial importance if a drug is expected to act
within the brain, as with an anti-psychotic agent, or to handle infections in
the membrane that surrounds the brain and spinal cord (meninges), such as
meningitis. Conversely, it is essential to avoid side-effects of a mental nature,
such as drowsiness or hallucinations, when the drug is intended to act else-
where in the body.
The partition coefficient is measured by allowing the compound to come to
equilibrium, assisted by shaking, between an aqueous and an organic medium,
usually n-octanol or hexane. The quantity of material in each phase at equilib-
rium is determined by a physical method, such as ultraviolet absorption, and
the ratio between the concentration in the lipid phase and that in the aqueous
phase is known as the partition coefficient.
The partition coefficient, as measured above, views the compound as an
uncharged species. The distribution coefficient (D) is a measure of how a drug
distributes itself between aqueous and lipid phases at a given pH, usually close
to neutrality, which may be more relevant conceptually to biologists since the
blood plasma and intracellular pH are normally about pH 7.4. This is particu-
larly relevant to molecules which are capable of ionization since the charged
species is less likely to be able to enter a lipophilic membrane (e.g. the local
anaesthetics discussed in Chapter 10). The aqueous medium chosen for the
measurement is often phosphate buffer at pH 7.4.
For ionizable drugs, the tendency to ionize is measured in terms of a dis
sociation coefficient K, which, like hydrogen ion concentration, is often
expressed as its negative logarithm: --log K or PK. If the pK of the drug is
within one log unit of the pH at which the measurement is made, more than
5 per cent of the compound will be in the charged form. The pK is identical
to the pH at which equal numbers of the drug molecules are charged and
uncharged. pK, refers to the molecule ionizing as an acid (releasing a proton)
while pK, is the similar position for a base (accepting a proton). Usually only
the non-ionized form of the drug will move into the lipid phase.
It should be noted that this concerns passive transport across membranes.
Active transport is a completely different phenomenon that requires the
expenditure of energy, and can be saturated and inhibited in a similar fashion
to enzyme reactions. There is an active transport system into the brain for the
aromatic amino acids, for example. Note that there is often confusion in the
literature about which is intended.
8 Molecular mechanisms of drug action

An interesting use of the partition coefficient alone to rationalize binding of

ligands to macromolecules has been made in the case of various families of
drugs and other compounds binding to serum albumin. This transport protein
carries tryptophan, fatty acids and other naturally occurring materials around
the body, and also a number of foreign molecules (xenobiotics) including
drugs. The binding of members of a closely related group of compounds, such
as penicillins, is directly proportional to log P; no other parameter appears
to be necessary to explain the binding. This is because the same factors of
hydrophobic versus hydrophilic interaction govern the albumin binding as
determine the partition coefficient. Sulphonamide binding, on the other hand,
may be defined by a term representing the pK, of the sulphonamide side-chain
as well as lipophilicity (see Jusko and Gretch, 1976 for a review).
Other attempts to rationalize the activity of families of related structures on
biological systems (structure-activity analyses) have been made using various
physical constants including log P, together with molecular size, electronic
effects in aromatic rings and shape of substituents. Ligands may bind to macro-
molecules by ionic, hydrogen or covalent bonds or by hydrophobic or van der
Waals interactions. The intention of this analysis is to make predictions for
new, more active, structures and also to reduce toxicities. A detailed discussion
on this subject is beyond the scope of this book but is available in the review
by Hansch (1985).

1.8 Drug nomenclature

Drug nomenclature is often rather confusing so it may be helpful to explain
the different names that may attach to a drug and how they were derived.
1. The detailed chemical name (according to IUPAC or other systematic
nomenclature).
2. The name by which the drug is often known (the generic or non-pro-
prietary name), which is usually a shortening of the chemical name and
is usually the name approved by drug regulatory authorities. These can
often vary from country to country. This name is useful for scientists and
physicians to describe the drug.
3. The manufacturer’s patented trade name by which the drug is sold, which
is usually short and ‘punchy’ so that it is easily remembered by hard-pressed
general practitioners. There may be several of these if different drug compa-
nies are marketing the same product, possibly in differing formulations.
4. There is also a number given by a drug company to any compound, synthes-
ized by the organic chemists or extracted from a fermentation culture
medium, that is usually of the form of an initial letter or letters followed
by a number which may extend to six digits. Usually when the decision is
taken to progress the compound towards the market, it is given a generic
name. Earlier publications may have the number only, and it may be neces-
sary to correlate name and number from later publications.
 General principles 9

An example of these names is provided by the following structure of a drug

which is used for the treatment of infections caused by anaerobic microorgan-
isms (Chapter 4).
1. The detailed chemical name is l-(2’-hydroxyethyl)2-methyl-5-nitroimid-
azole.
2. The generic name is metronidazole. We can see how this name was arrived
at from the detailed chemical name with the relevant letters in capitals:
l-(2’-hydroxyethyl)2-METhyl-5-NitROimIDAZOLE
(there has been a slight change in the order to make a more harmonious
sound).
3. The trade name of metronidazole is Flagyl.
4. The original company number was 8823 RP (short for Rhone-Poulenc).

W ,CH2CH20H
Metronidazole /N
tr N’ CH3

Generic names are not always entirely based on the chemical name but in
some cases the use of the drug is taken into account. The anti-herpesvirus
drug, originally numbered BW 248U, was subsequently named acyclovir,
although it might have been given the generic name acycloguanosine since
guanosine is the purine base in the structure (Chapter 2). Zovirax is the
trade name.

Acyclovir

Throughout this book the generic name of the drug is used, except in one
or two instances when the best known trade name and generic name are
both quoted.

1.9 Stereochemistry

The interaction of biological molecules is frequently a question of asymmetry,

whereby for example L- and n-amino acids and sugars are respectively active
or inactive in the natural biochemical pathways, and L-nor-adrenaline is the
10 Molecular mechanisms of drug action

ligand for adrenergic receptors. The L- and o-forms of the amino acids are
known as stereoisomers (IUPAC, 1970).
If a carbon atom has four different subs&tents attached, it is impossible to
superimpose one form of the molecule on the other. One fotm rotates the
plane of polarized light in one direction and the other in the opposite direction
to an equal degree - hence the use of the terms + and - for isomers, known
as optical isomers. An equal mixture of the two stereoisomers does not rotate
the plane of polarized light and is known as a racemic mixture or racemate.
Use of the terms D and L, noted above, is known as the Fischer convention
and indicates the configuration in relation to glyceraldehyde. This convention
is used mainly for amino acids, sugars and other related molecules of import-
ance in biochemistry and pharmacology.
There is another system of terminology named after Cahn, Prelog and Ingold
which classifies the stereoisomers into R (rectus or right-handed) and S
(sinister or left-handed), derived from the order of precedence in atomic
weight of the substituent groups on the asymmetric carbon atom. The substitu-
ent atom with lowest atomic number is imagined as lying behind the asymmet-
ric centre. The other substituents project towards the viewer. If the sequence
of decreasing priority is clockwise then R is assigned, if it is anti-clockwise, S.
This system is in more general use in organic chemistry and is applied to drugs.
There is no particular relationship between the different systems. Spectre
scopic techniques, such as circular dichroism (CD) or optical rotatory disper-
sion (ORD) which depend on the spatial relationship of the substituent groups,
must be used to determine the absolute configuration of a molecule. Receptors
and enzymes will interact with the form for which they are selective and are
likely to ignore the other.
Furthermore, the presence of a double bond in the molecule may also give
rise to stereoisomerism (but not optical isomerism). The two forms which
arise from the substituents being on the same or opposite sides of the double
bond are known as cis and trans. An example is the carboxylic acids, fumaric
and maleic - only fumaric acid, the trans isomer, is recognized as a substrate
by the Krebs cycle enzyme, fumarase. Cis/trans isomerism can also occur
when two substituents project out of a saturated ring (e.g. pilocarpine, Chap
ter 9).
Many drugs are prepared as racemates or mixtures of isomers, frequently a
pair, sometimes four or more if there are several optically active centres. One
or more of the individual compounds may be active, may contribute to the
toxicity or may just be inactive ballast (Ariens et al., 1988). Chemical synthesis
normally produces an equal quantity of each individual isomer, and in the
past it has not seemed sufficiently cost-effective to attempt to separate them.
Recently, however, there has been more of an effort to prepare the isomers,
especially where one isomer is responsible for the toxicity and another for
the activity as in the case of lamivudine, the anti-HIV drug (Chapter 2).
The importance the medicinal scientific community attaches to stereoisom-
erism is shown by the launching of a journal (Cbirality) specifically devoted
 General principles 11

to the biological effects of chirality. Furthermore, the attitude of the regulatory

authorities such as the Federal Drug Agency in the USA is leaning towards the
regulation of optically pure drugs; data for each stereoisomer must be pro-
vided. Wherever possible, I have Included data on the stereoisomers, although
in many cases it is still not available.

I.10 The future

We may be standing on the threshold of an explosion of biological knowledge

which will allow us to treat the major diseases of today. Such a leap forward
is obviously to be desired. One of the most important messages that this book
is intended to convey is that most of the major diseases are still, despite the
advances of recent years, neither fully understood, nor curable.
Thus, most of the drugs for cancer are nearly as toxic to the host as to the
tumour. The inflammation of arthritis may be suppressed but the underlying
disease process still cannot be halted and is poorly understood. In the case of
mental illness, we treat schizophrenics with drugs that largely sedate the pati-
ent and do not cure, although they may partially arrest the course of the dis-
ease. The human irnmunodeficiency virus (HIV) now presents a major chal-
lenge. In spite of our hard-won success in dealing with bacterial disease, we
still have to be eternally vigilant (a) to cope with mutants that are resistant to
many antibiotics and (b) to devise new drugs to kill fungi colonizing the areas
rendered vacant by the destruction of bacteria. Nature does, indeed, abhor
a vacuum!

References
Ariens, E.J., 1986, Trends in Pbarmacol. Sci., 7, 200-5.
Ariens, EJ., Wuis, E.W. and Veringa, E.J., 1988, Biocbem. Pbarmacol., 37, 9-18.
Hansch, C., 1985, hg Metab. Rev., 15, 1279-94.
IUF’AC, 1970,J. Org. Cbem., 35, 2849-63.
Jusko, W.J. and Gretch, M., 1976, Drug Metab. Rev., 5, 43-140.
Lienhard, G.E., 1973, Science, 180, 149-54.
Morrison, J.F., 1969, Biocbim. Biopbys. Acta., 185, 269-86.
Morrison, J.F., 1982, Trends in Biocbem. Sci., 7, 102-5.
Vane, J.R., 1971, Nature New Biology, 231, 232-6.
Walsh, C., 1984, Ann. Rev. Biocbem., 53, 493-536.
Wolfenden, R., 1976, Adv. Biopby. Bioeng., 271-306.
 Chapter 2

Nucleic acid biosynthesis and

catabolism

2. I Introduction

As nucleic acids are a fundamental part of any organism and have to be synthes-
ized early on in the growth cycle, it is not surprising that many of the drugs
in clinical use for the chemotherapy of neoplastic disease and microbial infec-
tion have the enzymes of nucleic acid synthesis as their primary target. A num-
ber of enzymes involved in viral and bacterial replication are either unique to
that organism, or are sufficiently different from the host enzyme which cata-
lyses a similar reaction for that difference to be exploited.
Cancer cells are characterized by rapid, uncontrolled growth; there do not
seem to be any major differences between tumour enzymes and those from
normal cells. It is, therefore, the rate of synthesis of DNA that forms the target
for chemotherapeutic agents. Consequently, toxicity problems do arise by vir-
tue of inhibition of these enzymes in rapidly growing non-tumour tissue -
hence the practice of delivering anti-tumour drugs in ‘cocktails’ of three or
four drugs at a time, working on different pathways, so that the toxicity of
any one drug does not become overwhelming. An outline scheme of the drugs
and their targets discussed in this chapter is presented in Table 2.1.
Another difference between microbial infection and cancer lies in the effi-
cacy of the immune system. Usually with bacterial and viral infection we rely
on the immune system to assist in, and eventually take over, the healing pro-
cess. In cancer, and also in some fungal infections, the immune system is weak-
ened and, therefore, a complete kill of the tumour or fungal cells is required;
drugs that merely arrest the growth of these cells are frequently not sufficient
to cure the disease.
The type of interference with enzymes that we find in this area is not only
the straightforward competitive inhibition of the ratedetermining step in a
pathway - although that may occur. Other approaches include the concept
of false substrates whereby one enzyme accepts a foreign substrate and yields
a product that can react with another enzyme; in the cancer field this is known
as the anti-metabolite approach. Furthermore, the synthesis of a macro-molecu-
lar product allows the possibility of incorporation of false substrate into the

13
14 Molecular mechanisms of drug action
Table 2.1 Drugs discussed in Chapter 2

Target Disease

2.2 Nuukotide biosynthesis

Dihydropteroate synthase Sulphonamides Bacterial infection
Dihydrofolate reductase Trimethoprim Bacterial infection
Pyrimethamine Malaria
Methotrexate Cancer
Bibonucleotide reductase Hydroxyurea Cancer
Orotate phosphoribosyltransferase 5-Fluorouracil Cancer
(thymidylate synthetase)
Cytosine deaminase (thymidylate synthetase) 5-Fluorocytosine Candidosis
Inosine monophosphate dehydrogenase Bibavirin Influenza
Dihydroorotate dehy,drogenase Atovaquone Malaria

2.3 Nucleic acid biosynthesis

Acyclovir Herpesvirus infection
Thymidine kinase (viral DNA polymerase) Vidarabme Herpesvirus infection
Deoxycytidine kinase (mamm alian DNA Cytarabme Cancer
polymerase)
DNA gyrase Ciprofloxacin Bacterial infection
Nortloxacin Bacterial infection
Novobiocin Bacterial infection
DNA topoisomerase II’ Doxorubicin Cancer
Etoposide Cancer
DNA polymerase (RNA primed), reverse Azidothymidine Infection by human
transcriptase immunodeficiency
virus (AIDS)
RNA polymerase Rifampicin Tuberculosis

2.4 Nucleotide catabolism

Adenosine deaminase 2-Deoxycoformycin Cancer
Xanthine oxidase Allopurinol Gout
Hypoxanthine phosphoribosyl transferase Allopurhlol Leishmaniasis
Guanylate cyclase Isosorbide nitrate Angina pectoris

Where two enzymes are recorded as targets, the first enzyme uses the drug as a false substrate
while the enzyme in brackets is the ultimate target.

mis assignation remains controversial.

polymer chain, resulting in either chain termination or the production of a

faulty macromolecule.
We also find the occasional occurrence of a suicide substrate (allopurinol,
section 2.4.2) and transition state analogues (2deoxycoformycin, section
2.4.1) which have aroused considerable interest (see Chapter 1 for an outline
of these concepts). The recent concentration on the widespread screening of
secondary microbial metabolites (i.e. those compounds formed that are not on
the normal pathways required for cell growth) and of the products of organic
chemical synthesis, has greatly increased the scope of chemical structures that
are under investigation.
 Nucleic acid biosynthesis and catabolism 15

2.1.1 Overall scheme

Nucleic acids in mammalian organisms are constructed from nucleotides

which in turn are made up from a purine or pyrimidine base, a sugar (ribose
in ribonucleic acid or deoxyribose in deoxyribonucleic acid) and a phosphate
group attached to the 5’-position of the ribose ring (Fig. 2.1). The synthesis
of these moieties may be de nova (Fig. 2.2) - i.e. constructed from small
molecular building blocks - or via salvage pathways which effectively recycle
a pre-formed base. Purines are synthesized by attaching an amino group to a
ribose ring and then using this as the anchor from which to build up an imida-
zole ring to yield 5-aminoimidazole-4carboxamide ribotide (AICAR).
One more step leads to the formation of inosine monophosphate (the first
purine nucleotide) from which the pathways diverge to yield guanosine and
adenosme monophosphates. The monophosphates have a second and third
phosphate group added to give the triphosphates that are required for RNA
biosynthesis. The dinucleotides may be reduced to deoxynucleotides, which
after conversion to the trinucleotides, are used for DNA biosynthesis.

ii
O--P-o-H& 0 R
I
0-
72
6H i)H
Phosphate Ribose

PURINE BASES

NH2 0

R=

Adenine Guanine

PYRJMIDINE BASES

I I I
Uracil Cytosine Thymine
Figure 2.1 Nucleotide structure.
16 Molecular mechanisms of drug action
0
0

”
” Glutamine Glutamate +
pyrophosphate
Phosphoribosyl pyrophosphate Phosphoril bosylamine

H2NC0
N
7 steps
AICAR J
I ’
X) .
HzN “:

lo- Formyl THF

THF

FAICAR

lnosine monophosphate

2 ste/ \ps

Adenosine monophosphate Guanosine monophosphate

R, ribose 5’-phosphate; AICAR, 5-aminoimidazole-4-carboxamide ribotide; FAICAR, 5-

formamidoimidazole-4-carboxamide ribotide.
Figure 2.2 Outline of purine nucleotide biosynthesis.

Folk acid analogues are essential cofactors for the transfer of some one-
carbon units. The particular conversions of interest in the nucleotide biosyn-
thetic pathways are the addition of a formyl (-CHO) group to glycinamide
ribotide and subsequently to AICAR by lO-formyltetrahydrofolate in the purine
pathway, and the methylation of deoxyuridine monophosphate (UMP) to deox-
ythymidine monophosphate (dTMP) catalysed by thymidylate synthetase in
pyrimidine biosynthesis. The first step in the biosynthesis of folate is the forma-
tion of dihydropteroate from p-aminobenzoate and 2-amino4hydroxyG
methylpteridine (Fig. 2.4). This step is the target for the sulphonamide antibac-
 Nucleic acid biosynthesis and catabolism 17

terials (hence the build-up of AICAR in cells treated with sulphonamides -

see below).
The heterocyclic ring of pyrimidines, on the other hand, is constructed
before the ribose is attached (Fig. 2.3). Indeed, in mammals the first three
steps to yield dihydroorotate are all catalysed by the same protein molecule,
although the activities are on separate proteins in bacteria. Orotate is con-
verted to the first pyrimidine nucleotide, orotidine 5-phosphate, which loses
a carboxyl group to form uridine 5-monophosphate. The addition of two more

y2
c=o
A -0oc.
NH2 ‘CH,
o--6=0 + AH -COO- I_
L CH-COO-
’
I I 4 ‘NH’
0‘ +NH, 0

Carbamylphosphate Aspartate Carbamylaspartate

II
1

::

“N&H,
&
I I
//C,N,CH-COO-
’ H
Orotate Dihydroorotate

PRPP
IV

PP I

OH OH i)H i)H

Orotidylate Uridine monophosphate

1, Aspartate transcarbamylase; II, dihydroorotase; III, orotate dehydrogenase; IV, orotidylate pyrophos-
phorylase; PRPP, phosphoribosyl pyrophosphate; PP, pyrophosphate

Figure 2.3 Pyrimidine nucleotide biosynthesis.

18 Molecular mechanisms of drug action

phosphate groups forms uridine triphosphate from which cytidine triphos-

phate can be made. Thymidine monophosphate is formed by the methylation
of uridine monophosphate catalysed by thymidylate synthetase. Again, as with
the purine nucleotides, it is the ribodinucleotides which are reduced to the
deoxy form.
The action of DNA polymerases, using the four deoxyribotrinucleotides,
produces DNA. RNA is synthesized from a DNA template in a process
(transcription) requiring the four ribotrinucleotides, and messenger and trans-
fer RNA are employed to generate protein on the ribosome (translation, Chap-
ter 3). Exceptions to this sequence are found with viruses that contain RNA
as the genetic code. Two examples are the influenza virus, where the RNA
acts as its own template, and the virus that produces the disease known as
acquired immune deficiency syndrome (AIDS), where the RNA acts to form
DNA, catalysed by a viral enzyme known as reverse transcriptase (section
2.3.5).
Purine nucleotides are first broken down to nucleosides by removal of the
phosphate group and ribose groups to produce inosine. Subsequent changes
include oxidative steps to hypoxanthine, xanthine and uric acid (Fig. 2.10).
Pyrimidine nucleotides are also converted to the free bases, uracil and 5-
methyluracil), that eventually yield malonate (from cytosine) and 2-methyl
malonate (from thymine).

2.2 Nucleotide biosyntheses - enzyme targets of drugs

2.2.1 DShydroorotate dehydrogenase - atovaquone as

amtimalarial

Atovaquone is a naphthoquinone under development for the treatment of

protozoal infections in man, notably malaria caused by Plasmodium falcipa-
rum and toxoplasmosis caused by Toxoplasma gondii. Quinones are unusual
among marketed drugs, probably because of their innate reactivity. The drug
may also be effective against Pneumocystis carinii, a microorganism originally
classed as a protozoon but now believed to be phylogenetically closer to a
fungus. Infections by this agent have become more serious with the increasing
number of patients whose immune systems are suppressed, e.g. in AIDS.
The malarial parasite depends entirely on de nova biosynthesis of pyrimidine
nucleotides for nucleic acid biosynthesis, unlike man where salvage pathways
from preformed nucleosides play a significant part. Atovaquone acts by
inhibiting the membrane-bound complex dihydroorotate dehydrogenase - the
third enzyme in the pyrimidine biosynthetic pathway - see Fig. 2.3. Although
the enzyme is not the rate-limiting step, which in most pathways is the first
enzyme (in this case aspartate transcarbamoylase), the inhibition of the target
enzyme is irreversible; a build-up in the precursors such as carbamoylaspartate
 Nucleic acid biosynthesis and catabolism 19

and a reduction in the end-products such as UTP results. Furthermore, the

inhibition could not be reversed by adding uridine (Hammond et al., 1985).
Dihydroorotate dehydrogenase is found on the inner side of the mitochon-
drial membrane and passes electrons to ubiquinone in the electron transport
chain. The specific molecular target is a polypeptide between cytochromes b
and cl in mitochondria isolated from P. falciparum (Fry and Pudney, 1992).
Atovaquone is active as the trans isomer.

Atovaquone
Relative stereochemistry.

The IC,, for various strains of l? falciparum is between O-7 and

4.3 X 10m9 M, which is more potent than the other antimalarial drugs tested
(amodiaquin, mefloquin, pyrimethamine, chloroquin and quinine; Hudson et
al., 1991).

2.2.2 Dihydropteroate synthetase - sulphonamides as anti-

bacterials
One family of drugs which was recognized early to have anti-bacterial proper-
ties was the sulphonamides. These have been in medical use since the 1940s -
indeed Winston Churchill was given one of the earliest compounds, M &
B 693, in North Africa in February 1943 when he had an attack of pneumonia,
and it is likely that it played a major role in his recovery (Churchill, 1951).
These drugs interfere with the purine biosynthetic pathway; 5-aminoimidazole-
4carboxamide ribotide (AICAR) accumulates in bacterial cells treated with sul-
phonamides. The mechanism of interference is indirect in that AICAR accumu-
lates not because the next enzyme in the pathway is inhibited, but because
there is insufficient folate cofactor to service the enzyme. Mammals, however,
require preformed folate in the diet as an essential vitamin, and so inhibition
of this step is not likely to lead to mammalian toxicity.
The biosynthesis of folate requiresp-aminobenzoic acid as a starting material
and sulphonamides are sufficiently close in structure to act not only as inhibi-
tors of dihydropteroate synthetase but also as false substrates (reviewed in
Woods, 1962) (Fig. 2.4). Any agent that can penetrate a bacterium which needs
a functioning enzyme is very likely to have activity against that bacterium.
20 Molecular mechanisms of drug action

0.69 nm 0.67 nm
C
-o/ +.

v 1
e * * *
0.24 nm 0.23 nm

Sulphanilamide p-Aminobenzoate

+ W COzH

2-Amino-4-hydroxy-6- pAminobenzoic acid

methylpteridine

OH

Dihydropteroate Glutamic acid

\
Pteroylglutamic
(folic) acid

The conversion of 2-amino-4-hydroxy-6-methylpteridine to dihydropteroate and to folic

acid is shown.
A typical sulphonamide is compared with paminobenzoate and is shown to be
structurally similar.

Figure 2.4 Dihydropteroate synthetase reaction.

Sulphonamides are not now usually used alone to treat infections in man as
there are many other more effective families of anti-bacterials currently avail-
able, although sulphonamides may be used in conjunction with inhibitors of
dihydrofolate reductase (see below). An exception to this is dapsone, which
is sometimes used alone for the treatment of leprosy (Shepard et al., 1976).
 Nucleic acid biosynthesis and catabolism 21

2.2.3 Dihydrofolate reductase (DHFR) - trimethoprim and

pyrimethamine as anti-bacteria@ methotrexate as anti-cancer
The reaction carried out by this enzyme is crucial to the recycling of the folate
cofactors involved in methylation of the DNA bases:

Deoxyuridine monophosphate 4.+Deoxythymidine monophosphate

9,10-Methylene

I A/
Tetrahydrofolate

DHFR has turned out to be an effective target for both anti-microbial and anti-
cancer drugs, possibly because one folate molecule has to be reduced for every
deoxythymidine monophosphate molecule synthesized, and in situations of
rapid growth this requirement could be considerable. The mechanism of thym-
idylate synthetase requires that the one-carbon fragment is transferred as a
methylene group to the 5-position of the uracil ring of dUMP. This is then
reduced to a methyl group by the pteridine ring of the coenzyme to yield
dihydrofolate. DHFR is required to reform tetrahydrofolate.
Unlike the case of dihydropteroate synthase above, mammals carry out the
same interconversion and so selectivity becomes of prime importance. In this
instance trimethoprim, used as an anti-bacterial agent, and pyrimethamine, an
effective agent for killing the protozoa1 parasites (plasmodia) that cause
malaria, are far less inhibitory towards a mammalian liver DHFR than for the
enzyme from their respective microbial targets. This is shown in Table 2.2
(the effect of the drugs on the parasite, Plasmodium berghei, that infects mice
is shown for comparison) and see Matthews et al. (1985).
DHFR has been confirmed as the target for trimethoprim by the identifi-
cation of an altered, and largely uninhibited, DHFR in resistant organisms
(reviewed in Hitchings and Smith, 1980).
These agents are able to penetrate their target organisms by non-ionic dif-
fusion. Methotrexate, a diaminopyrimidine which is used for choriocarcinoma
and trophoblastic tumours (Hammond et al., 1981>, strongly inhibits DHFR in
tumour cells with a Ki of 1 X lop9 M. The drug forms a ternary complex with
enzyme and NADPH, as shown by X-ray crystallography (Matthews et al.,

Table 2.2 Concentration for 50 per cent inhibition of DHFR (X lo-” M)

Rat Nver E. coli P. berghei

Pyrimethamine 700 2500 0.5

Trimethoprim 260 000 5 70

(from Ferone et al., 1969)

22 Molecular mechanisms of drug action

1978). The action of the drug in vivo relies on the greater need of the tumour
cell for purine nucleotides compared with normal cells, and so gives rise to
considerable toxicity in rapidly dividing cells such as in the bone marrow.
Methotrexate is transported into mammalian cells by the active transport sys
tern for folate, to which it bears a marked resemblance. Microorganisms do
not have such a transport system and, although methotrexate is as effective
against the microbial enzymes as against the mammalian, it has little anti-
microbial activity (Ferone et al., 1969).

Pyrimethamine Trimethoprim

Methotrexate
0
il
HOOC-(CH2)2-CHNH-C

LOOH ‘o-, -cH2 fo””

As the targets for sulphonamides and diaminopyrimidines lie on different

pathways that intersect, it seemed appropriate to test combinations of these
drugs to establish whether they could be used in conjunction more effectively,
i.e. whether they showed synergy. Synergy was found to occur both in vitro
(Moody and Young, 1975) and in vivo (Ramachandran et al., 1978) when
trimethoprim was combined with sulphamethoxazole against bacterial infec-
tions. With pyrimethamine similar effects have been observed against human
malaria, but in this case the value of such combinations is neither to reduce
toxicity nor to obtain high potency, but rather to lower the dose of pyrimeth-
amine such that there is less likelihood of resistance developing. Resistance is
a serious problem with the malarial parasite at present.
Combinations of trimethoprim with other sulphonamides such as sulphadia-
zine (Hurly, 1959) or dapsone (Lucas et al., 1969) were shown to be very
effective in man. Sulphadoxine currently appears to be the sulphonamide of
choice for synergy with trimethoprim as it has a particularly long half-life for
a sulphonamide - seven to nine days in plasma. This is clearly of great value
for reducing the number of doses required for treatment.
 Nucleic acid biosynthesis and catabolism 23

Sulphadiazine Sulphamethoxazole

Dapsone Sulphadoxine

H2NaSO~oNH, H.N+OzNH 0

CH30 OCH3

2.2.4 Ribonucleotidereductase -hydroxyureaasanti-cancer

The reduction of the nucleoside diphosphates to their respective deoxy
counterparts is the first committed step in deoxyribonucleotide biosynthesis,
and is catalysed by ribonucleotide reductase. The activity of this enzyme has
been found to follow closely the regulation of DNA biosynthesis, although this
connection may not be any more than a correlation (Thelander and Reichard,
1979). The same enzyme reduces all four ribonucleotide diphosphates by
directly replacing the 2-hydroxyl group on the ribose ring by a hydrogen atom.
With the mammalian enzyme, magnesium and ATP are also required for activity
and a number of allosteric regulators have been identified (see Chapter 1 for
a discussion on allosterism). In order to carry out the reduction, the enzyme
requires the iron-sulphur protein thioredoxin, together with thioredoxin
reductase which uses NADPH to regenerate thioredoxin. Thioredoxin contains
two free sulphydryl groups positioned in such a way that a disulphide bond
can be formed between them:

Ribonucleotide reductase
Ribonucleoside Deoxyribonucleoside

Thioredoxin Thioredoxin

I I I I

SH SHLd-s
Thioredoxin reductase

NADP+ $NADPH + H+

The E. coli enzyme on which most of the work has been done is a hetero-
dimer composed of two subunits Bl and B2. The larger subunit contains two
equivalent polypeptides of 86 kDa carrying the substrate binding site and
24 Molecular mechanisms of drug action

allosteric effector sites. The smaller subunit composed of two equivalent poly-
peptides of 45 kDa carries a unique prosthetic group consisting of two cata-
lytic, but nonequivalent, high spin ferric iron atoms stabilizing a tyrosyl rad-
ical. The oxygen of the phenolic hydroxyl formally carries the radical in a one-
electron oxidized state. The tyrosine (Tyr-122) is buried in the protein about
1 nm from the surface and is surrounded by hydrophobic side-chains. Each
iron atom is associated with one polypeptide chain and the pair are anti-ferro-
magnetically coupled to each other (i.e. their spins are opposed or anti-
parallel) and are bridged by an oxygen atom and two carboxylate residues.
These groups are located in the active site at the interface between the two
subunits, together with two cysteine thiols which are reduced by the NADPH-
thioredoxin system (reviewed in Stubbe, 1989; Elgren et al., 1991; Fig. 2.5).
The mechanism of this unusual reaction involves removal of a hydrogen
atom from position 3 of the sugar by the tyrosyl radical, followed by acid-
catalysed cleavage of the 2’ C-OH bond to yield a radical cation. This inter-
mediate is then reduced by addition of two hydrogens from the two thiols.
Finally the same hydrogen atom that was originally abstracted from the 3’
position is returned to the product and the tyrosyl radical is regenerated
(Stubbe, 1989; Fig. 2.6).
Since the activity of ribonucleotide reductase correlates well with the level
of DNA biosynthesis it might be expected to be a logical target for chemo-
therapy (reviewed in Elford et al., 1981), but it is perhaps surprising that so
far only one drug on the market, hydroxyurea, owes its effectiveness to inhi-
bition of this enzyme. This drug is used for the treatment of chronic granulo-
cytic leukaemia and malignant melanoma (Kennedy and Yarbro, 1966; Ariel,
1970). The use of hydroxyurea suffers from the drawback that frequent large
doses are required to maintain effective concentrations in vivo. This is under-
standable in view of the high drug concentration, 5 X lo-* M, required to
produce 50 per cent inhibition of the enzyme (Elford et al., 1979). Hydroxy-
urea originally presented a challenge with regard to its mechanism of action.
Recent studies, however, show that the drug quenches the tyrosyl radical by

Postuloled Cofactor Centre of RDPR

His\Fe3+‘OA
His -
/ ‘\\Fe3+-His
/His
’ ‘0 0’ ’
His ‘,’ X

Figure 2.5 Proposed structure of the binuclear iron centre of ribonucleotide reductase
 Nucleic acid biosynthesis and catabolism 25

I b

Figure 2.6 Postulated mechanism of reduction of nucleotides to deoxynucleotides catalyzed by

ribonucleotide reductase.

the transfer of an electron, thus forming transient nitroxide-like free radicals

which have a very short half-Life. The enzyme is thereby inactivated. Hydroxy-
urea is poor as a general free radical scavenger, however, but is much more
specific for ribonucleotide reductase than are other free radicals, possibly
because it is small enough to pass along the 1 rnn cleft in the protein
(Lassmann et al., 1992).

2.2.5 Thymidylate synthetase - 5-fluorouracil as anti-cancer, 5-

fluorocytosine as anti-fimgal
The methylation of uridine monophosphate to yield thymidine monophos-
phate is a crucial step in the provision of pyrimidine bases for the synthesis
of DNA, and so is another possible target for cancer chemotherapy. 5-Fluorour-
acil is believed to exert its anti-tumour effect, after conversion to ribo- and
deoxyribonucleotide, partly on thymidylate synthetase and partly on RNA
biosynthesis (reviewed in Heidelberger et al., 1983).
5-Fluorouracil (5-FU) is converted to 5-fluorouridine monophosphate (5-
FUMP), probably in one step catalysed by orotate phosphoribosyltransferase
using phosphoribosyl-1-pyrophosphate as co-substrate. 5-FUMP is phosphoryl-
ated to the diphosphate (5-FUDP), reduced in the presence of ribonucleotide
reductase to 5-FdUDP and then hydrolysed to 5-FdUMP. This latter compound
26 Molecular mechanisms of drug action

forms a covalently bound ternary complex with the folate cofactor Ns,‘“-methy-
lene-tetrahydrofolate and thymidylate synthetase, which closely resembles the
transition state of the normal reacton. 5-FdUMP acts as a false substrate since
the folate methylene group is transferred to the 5-position of the uracil ring,
but further conversions cannot take place because of the presence of the fluor-
ine atom at position 5 of the uracil ring. As a consequence, a covalent link
remains between the methylene group and either positions 5 or 10 of the
pteridine ring of the cofactor (see Fig. 2.7). The inhibitor, therefore, behaves
as a suicide substrate and the synthesis of thymidylate is shut off.
Other studies, however, have tended to suggest that thymidylate synthetase
may not be the only target. The incorporation of 5-FU into RNA to yield a
faulty ribosomal RNA may also play a part in the drug’s efficacy and both
mechanisms may be important to varying extents in different situations
(Heidelberger et aE., 1983). 5-FU is used with good effect in certain types of
carcinoma of the breast and of the gastrointestinal tract (Bonadonna and Valo-
gussa, 1981; Kisner et al., 1981).
A closely related compound, 5-fluorocytosine (5-FC), has found favour as a
treatment for fungal infections, namely candidosis and cryptococcosis, particu-
larly when used in conjunction with amphotericin B (Cohen, 1982). 5-FC is
not recommended as a single medication in view of the rapid development to
resistance. Up to 30 per cent of the patients in one trial treated with 5-FC
alone developed resistance (Block et al., 1973) - hence the concomitant use
of amphotericin. 5-FC is converted to 5-FU by a fungal enzyme, cytosine deami-
nase; similar metabolic conversions to those for 5-FU in mammalian cells fol-
low. 5-FU is initially converted to 5-FUMP by UMP pyrophosphorylase (an
enzyme that normally acts to salvage uracil in Candida albicans). This appears
to be the point at which resistance is induced, since resistant C. albicans has
little or no detectable enzyme (Whelan and Kerridge, 1984). Cytosine deamin-
ase is believed to be largely absent from the host, although Diasia et al. (1978)
found some evidence of 5-FU production in man.

2.2.6 Inosine monophosphate dehydrogenase - riiavirin as anti-

viral
Kibavirin (virazole) has shown variable clinical effectiveness against a variety
of viruses, including those with RNA as the genetic material, such as influenza,
and those containing DNA, for example, herpes, measles and some adeno-
viruses (reviewed in Chang and Heel, 1981). There is some uncertainty about
the effectiveness of the drug, but, at least in the case of influenza, the delivery
of the drug by aerosol is more effective than by the oral route. The latter may
deliver insufficient drug to the lungs for activity to be shown (McClung et
al., 1983). Kibavirin sufficiently resembles adenosine to be phosphorylated
intracellularly by host adenosine kinase, and is subsequently converted to the
di- and trinucleotides by other host enzymes. The drug structure can also be
drawn to resemble guanosine (Gilbert and Knight, 1986).
 27
Nucleic acid biosynthesis and catabolism

0
0
II + E-Glutagi;

Ho--F;-o
OH

6H

CH3

0
0 II
II + ‘C - Glutamic
acid
Ho--I;-o
OH

REACTION INTERMEDIATE
(PROPOSED):

TERNARY COMPLEX
(PROPOSED):

OH

Figure 2.7 Thymidylate synthetase reaction.

28 Molecular mechanisms of drug action

5-FC
NHz

Intracellular pools of GTP are reduced by up to 45 per cent in influenza-

infected cells treated with ribavirin (Wray et aE., 1985). This results from inhi-
bition of the first committed enzyme in de nouo synthesis of GMP; the NAD-
linked inosine monophosphate dehydrogenase, by the monophosphate of riba-
virin (Streeter et al., 1973). This enzyme converts to IMP to xanthosine mono-
phosphate which is then transformed into GMP by GMP synthetase (see Figure
2.8). The inhibition is competitive with a K, of 2.5 X lo-’ M against IMP
dehydrogenase from Ehrlich ascites tumor cells (K,,, for IMP is 1.8 X 10e5 M).
This inhibition accounts for part of the antiviral effect, but the fact that
guanosine can only partly reverse the inhibition of proliferation of influenza
virus in canine kidney cells suggests that other mechanisms may be involved.
This is supported by the observation that the lowering of the GTP levels
reaches a minimum at 2.5 X lop5 M ribavirin but the anti-viral effect continues
to increase up to 1.4 X lop4 M drug (Wray et al., 1985). Other possible targets
include ribavirin triphosphate inhibition of (a) the capping reaction of viral
messenger RNA (GTP is added to the 5’ end of freshly synthesized viral mRNA
to render it effective), and (b> the viral RNA polymerase complex (see review
by Gilbert and Knight, 1986).
The nature of the virus under consideration may well play a part in the
mechanism of action. Patterson and Fernandez-Larson (1990) believe that inhi-
bition of the viral RNA polymerase complex is the mechanism by which riba-
vii-in inhibits the growth of two other RNA viruses: La Crosse and vesicular
stomatitis. Multiplicity of mechanisms of action may also explain why it is very
difficult if not impossible to obtain ribavirin-resistant virus strains.

2.3 DNA biosynthesis

So far we have been considering drugs that interfere with the process of
nucleotide biosynthesis. We now turn to the next step in the formation of
nucleic acid; the magnesium-dependent process in which the four nucleoside
triphosphates are utilized in the polymerization of DNA. In mammalian cells
more than one enzyme is required to do this conversion; two enzymes, (Yand
p, are responsible for the formation of chromosomal DNA, while y catalyses
the synthesis of mitochondrial DNA.
The viral genome, on the other hand, codes for a number of enzymes which
it needs to synthesize a fresh virus particle. Foremost among these is the RNA
 Nucleic acid biosynthesis and catabolism 29

lnosine monophosphate Xanthosine monophosphate

OOC- CH,-CH - COO-

I I IV
+NH3
-4 1

-OOC-CH,-YH-COO-

Adenylosuccinate Guanosine monophosphate

II
II
CH
I
coo-

I: Adenylosuccinate synthetase
II: Adenylosuccinate lyase
III: IMP dehydrogenase
IV: GMP synthetase
Adenosine monophosphate R: Ribose

Figure 2.8 The biosynthesis of GMP

P
“2N’cN*p
Ribavirin ‘N
HOCH2 o

‘6$’
OH OH
30 Molecular mechanisms of drug action

or DNA polymerase with which, after subversion of the host cell’s protein
synthesizing machinery, more genome is synthesized. Although all DNA poly-
merases require magnesium, the enzyme coded by herpesviruses 1 and 2 differ
from their mammalian counterparts in a number of ways, including a fivefold
activation by 6 X lop2 M ammonium sulphate while the host cell enzymes are
completely inhibited at this concentration (Mar and Huang, 1979).
A few drugs that interfere with DNA biosynthesis have been developed for
the treatment of cancer and for viral disease. Indeed, the main drugs used to
treat viral infection may be found in this section, which is interesting consider-
ing how many other activities of the virus could be targets, such as the binding
of the virus to the outside of the target cell; the removal of the protein outer
coating of the virus and the penetration of the nucleic acid into the cell; the
induction of protein as well as nucleic acid biosynthesis by the cellular appar-
atus; assembly of protein and nucleic acid into infectious virus particles fol-
lowed by their exit from the cell.
This process destroys the cell, whether bacterial or mammalian, and can
lead to a fresh round of virus replication if other cells are nearby. A zone of
lysis, known as a plaque, is then formed wherever replicating viral particles
are located. If host cells are grown on a suitable nutrient in a plate, plaque
formation can be used to quantify the amount of active virus present.

2.3.1 Herpesvirus DNA polymerase - acyclovir and vidarabine as

anti-virals
9-[(2-Hydroxyethoxy)methylguanine (generic name, acyclovir), an analogue of
a purine nucleoside, is the first of a new family of drugs for the treatment of
infections caused by the family of DNA viruses, known as herpesvirus, which
range in severity from the ubiquitous mouth ulcer, through serious cases of
blindness, to life-threatening disease (reviewed in Whitley and Gnann, 1992).
Herpesvirus infections in humans can be divided into four types: herpesvirus
type 1 which is usually responsible for mouth ulcers and eye damage; type 2
which mainly gives rise to serious genital infections; varicella zoster virus
which produces chicken pox and shingles; cytomegalovirus (large-cell-produc-
ing virus) which gives rise to a type of pneumonia, particularly in immuno-
compromised conditions such as AIDS.
These viruses have the ability to become latent, i.e. after an infection the
virus migrates to the nervous system where it remains dormant until reacti-
vated by stimuli such as stress. When latent, herpesvirus 1 resides in the tri-
geminal ganglion (a relay station for the fifth cranial nerve) near the inner ear.
There is a need not only to treat viral infections but also to block either latency
or reactivation of these viruses.
Acyclovir is converted to a trinucleotide which takes part in the reaction
catalysed by viral DNA polymerase by incorporating the false purine base into
the template primer. The addition of the next nucleotide (dCTP) produces a
dead-end ternary complex that brings DNA synthesis to a premature end, since
 Nucleic acid biosynthesis and catabolism 31

acyclovir does not have the 3’-hydroxyl group necessary for chain elongation
(Reardon and Spector, 1989).
Another virally coded enzyme, thymidine kinase, is responsible for con-
verting the drug to a nucleotide analogue. In practice, the enzyme is unfortu-
nately named because it is not specific for pyrimidines, and also shows thymid-
ylate kinase activity. Nevertheless, acyclovir is a poor substrate for the enzyme
from herpesvirus 1 with a K, of 4 X lo-* M and a V,, of 6.1 X lO-‘*
mol/min/~l of enzyme preparation. The respective figures for thymidine, in
contrast, are 8.5 X 10m6 M and 218 (Ashton et al., 1982). Acyclovir mono
phosphate is converted to the diphosphate by the action of viral thymidine
kinase or by cellular guanosine monophosphate kinase and a number of kin-
ases can add a third phosphate (Miller and Miller, 1982). Acyclovir is not toxic
to uninfected cells because they do not convert it to a triphosphate analogue.

Acyclovir

Acyclovir is active to a similar extent against types 1 and 2 virus with 50

per cent inhibition of viral plaque formation at concentrations of 0.046 to
1.8 X 10m6 M for type 1 and 0.65 to 1.8 X lop6 M for type 2 (Parris and Harring-
ton, 1982). Since there are two viral enzymes required to convert acyclovir
into an antiviral agent, resistance can and has occurred in all three possible
ways: an altered thymidine kinase that phosphorylates thymidine but not acy-
clovir through mutations in the thymidine binding region, a virus lacking thym-
idine kinase (which is not an essential enzyme) and single base mutations in
the ultimate target DNA polymerase (Chatis and Crumpacker, 1992).
Acyclovir is also used to treat varicelIa zoster infections, albeit at a higher
dose (Ljungmann et al., 1986). Cytomegalovirus is resistant as it lacks a thymid-
ine kinase. Ganciclovir, the drug of choice for cytomegalovirus infections, is
phosphorylated by host kinases to the triphosphate which inhibits the viral
DNA polymerase. Infections by these viruses are not normally life-threatening -
except in immunocompromised conditions such as AIDS (Chatis and Crum-
packer, 1992).
Vidarabine (arabinosyladenine or Am-A) has both anti-tumour and anti-viral
activity. The drug is active against keratinization of the cornea and encephalitis
caused by herpesvirus 1 in adults and neonates, and against herpes zoster
infections in immunocompromised patients (reviewed in Whitley et al., 1980).
The nucleoside is phosphorylated to the monophosphate by both host and
virus-induced thymidine kinase; further enzymes convert it to the triphosphate.
Ara-ATP is more inhibitory to herpesvirus than to host DNA polymerase; a Ki
32 Molecular mechanisms of drug action

of 1.4 X lo-’ M has been reported for herpes hominis DNA polymerase from
infected cells (K, for dATP was 1.37 X 10m5 M) while DNA polymerase cr from
rabbit kidney cells was inhibited with a Ki of 7.4 X low6 M (K, for dATP was
6.4 X 10e6 M) (Muller et al., 1977). Furthermore, Am-A is incorporated into
viral DNA and slows the rate of DNA elongation, but can be removed by the
exonuclease activity associated with viral replication (see Chatis and Crum-
packer, 1992).
All these inhibitions were competitive and were measured with activated
DNA as the template. Clearly, low doses of the drug will inhibit viral repli-
cation, while higher doses will inhibit the mammalian enzyme and may lead
to toxicity as the drug is not absolutely selective for viral replication (Muller
et al., 1977). Furthermore, Am-A is incorporated into viral DNA, and it is poss-
ible that this effect contributes to its anti-viral action.
The development of Am-A as an anti-tumour agent has been greatly limited
by its rapid deamination to the inactive metabolite arabinosylhypoxanthine by
adenosine deaminase, an enzyme that is widely distributed in body tissues and
fluids. The drug is, therefore, insufficiently active in the clinic for practical
use. Attempts have been made to circumvent this deamination by the use of
2-deoxycoformycin, a powerful adenosine deaminase inhibitor, which
enhances and prolongs the activity of Ara-A in man (Major et al., 1983) but
this practice has not yet found general acceptance.

2.3.2 M ammalian DNA polymerase - cytarabine as anti-leukaemic

Arabinosylcytosine (At-a-C) is a drug that is commonly used for the treatment

of acute leukaemia, both myelogenous and lymphocytic, and non-Hodgkin’s
lymphoma (Rubin, 1981). It is specific for the S-phase of the cell cycle. In
order to exert its anti-tumour effect, Am-C must first be converted to the mono-
phosphate. This is performed by deoxycytidine kinase; subsequently deoxycy-
tidylate kinase and nucleoside diphosphate kinase catalyse the conversion of
the nucleotide to Am-CTP. This metabolite interferes with DNA synthesis by
inhibiting DNA polymerase competitively with dCTP (Ki for Am-CTP is
8.7 X lop6 M and K,,, for dCTP is 9.0 X lop6 M (Graham and Whitmore, 1970).
Resistance to the drug can occur either by virtue of reduced deoxycytidine
kinase levels or by deamination to inactive uracil metabolites by cytidine deam-
inase. In addition, it has been shown that Am-C is incorporated into DNA not
so much at terminal, but rather at internucleotide linkages. This suggests that
chain elongation is not stopped by the presence of the false sugar (Graham
and Whitmore, 1970). More recent work with higher concentrations of drug
has shown that a greater proportion of Am-C may be found at the 3’ terminus,
which implies that chain elongation is greatly slowed by the drug (Major et
al., 1982). This is understandable since the 2’-hydroxyl of arabinose is trans
to the 3’-hydroxyl, in contrast to ribose, and will cause steric hindrance to
the rotation of the pyrimidine base about the nucleoside bond. The bases of
 Nucleic acid biosynthesis and catabolism 33

polyarabinonucleotides cannot, therefore, stack in the same way as do deoxyri-

bonucleotides, and slowing of DNA elongation could well be the consequence.

NHz NH2

OH OH
Ara-A Ara-C

2.3.3 DNA topoisomerases

Another recently discovered family of enzymes is closely involved in the repli-
cation of DNA. DNA as normally found in the cell, whether prokaryotic or
eukaryotic, is supercoiled, i.e. not only do the nucleoside bases form a double
helical coil but the helix itself is twisted on itself to form supercoils. The
topoisomers that result from a difference of two superhelical turns have differ-
ent mobility on SDS-polyacrylamide gels and can be visualized, after ethidium
bromide staining, in the form of a ‘staircase’.
In the bacterial cell, chromosomal and plasmid DNA is in the form of a
closed circle; supercoiling greatly reduces the overall volume of DNA allowing
it to be packaged more tightly in the cell. DNA as normally extracted from
bacterial cells has fewer supercoils than expected and is called negatively
supercoiled. The enzyme-catalysed unwinding of supercoils is referred to as
relaxation.
In the mammalian cell, the DNA is supercoiled on the histone proteins in
a long helical coil which restrains the supercoils, i.e. prevents them from
relaxing. In both prokaryotic and eukaryotic cells, supercoiled DNA is required
for the major processes in which DNA is involved; replication, transcription,
conjugation, etc.
A class of enzymes called topoisomerases has been recently discovered that
can catalyse the interconversion of these topoisomers; they do not catalyse
the net formation or cleavage of covalent bonds, but rather modify the three-
dimensional structure or topography of DNA. In the mammalian cell, the
arrangement of the DNA into a long helical coil requires an enzyme to remove
the twists that are produced when DNA is replicated. Otherwise there will be
a build-up of positive supercoils in front of the replication fork and negative
supercoils behind (see Wang, 1991).
In the presence of magnesium and ATP, topoisomerases break and reform
the strands of the DNA; topoisomerase I breaks one strand by forming a bond
with the 4-hydroxy group of a tyrosine to the 3’ terminus of the DNA, while
34 Molecular mechanisms of drug action

topoisomerase II breaks both strands but with the covalent link to the 5’ pos-
ition (Rowe et al., 1986). In both cases the other strand or strands is passed
through the break and the DNA chain is resealed, introducing two supercoils.

2.3.3.1 Bacterial topoisomerase II - quinolone carboxylic acids as anti-

bacterials
Often in the history of medicine a drug has been used for many years without
any precise knowledge of its mode of action, e.g. nalidixic acid and its ana-
logues. They first went on the market in the early 1960s for the treatment of
urinary tract infections caused by Gram-negative organisms such as Escberichia
coli and Proteus species. The drug has the drawback that resistance rapidly
appears in previously sensitive organisms.

CH2CH3

Nalidixic acid

Recently, the advent of analogues, the fluoroquinolones, with a greatly broad-

ened spectrum of action has occurred, notably ofloxacin, and ciprofloxacin;
the clinical use of these agents has been reviewed by Hooper and Wolfson
(1991).
It became clear early on that nalidixic acid was interfering with DNA
biosynthesis, but it required the discovery of DNA gyrase before the precise
enzyme target was identified (Gellert, 1980). Many of the activities that bac-
terial DNA undergoes, including recombination, conjugation, replication and
repair, require DNA to be supercoiled. The enzymes that largely control super-
coiling are topoisomerases I and II (Drlica, 1992) - the latter is known as DNA
gyrase. They play a crucial role in the bacterial cell.
Like the eukaryotic topoisomerase II, DNA gyrase catalyses the interconver-
sion of topoisomers differing by units of two. Gyrase is unique, however, in
that it can introduce negative supercoils into DNA in the presence of ATP and
magnesium. Without ATP, it can relax both negatively and positively super-
coiled DNA. DNA gyrase is composed of two subunits in the form of a
tetramer, A,B,. Like the mammalian enzyme, the linkage with DNA is made
through the 4-hydroxy group of a tyrosine (position 122; Horowitz and Wang,
1987); furthermore, nalidixic acid and the newer fluoroquinolones bind to the
A subunit in the presence of DNA, the cleavable complex, to form a ternary
complex. The genetic evidence strongly confirms this view since almost all
drug-resistant bacteria contain a point mutation in the A subunit. These drugs
may also bind direct to DNA (see Reece and Maxwell, 1991).
 Nucleic acid biosynthesis and catabolism 35

There is a paradox, however, in that the level of cellular supercoiling and

the inhibition of in vitro growth is inhibited by drug concentrations that do
not inhibit the enzyme. Normally one would expect a drug to inhibit a target
enzyme at a lower concentration than an organism because the drug has to
cross the cell wall and membrane to reach the target enzyme. To explain
this paradox, a ‘poison’ hypothesis has been proposed, whereby the enzyme
required to relax the DNA in front of the replication fork is poisoned by the
enzyme-DNA-drug ternary complex while the rest of the cellular gyrase is not
affected. Replication would thus be blocked but there would be no effect on
cellular supercoiling (see Reece and Maxwell, 1991).
The fluoroquinolones are broad spectrum anti-bacterials effective against a
wide range of Gram-negative and some Gram-positive bacteria. These drugs
are amongst the preferred agents for the treatment of infections of the gastro-
intestinal and urinary tracts, but lack activity against blood-borne infections of
Streptococcus and Enterococcus (Hooper and Wolfson, 1991).
Ofloxacin is a chiral agent, the S-isomer being 10 to lOO-times more active
than the R-isomer, indicating the asymmetry of the binding site. The s-isomer
happens to be laevorotatory and is under development as laevofloxacin both
for the treatment of bacterial infections and HIV (Wentland et al., 1988). The
mode of action is unclear in the latter case.
4Hydroxycoumarin antibiotics, of which novobiocin is the only one to have
reached the market, have been isolated from Streptomyces species and are
mainly active against Gram-positive species. Coumarins interfere with ATP-
induced supercoiling but not ATP-independent relaxation by binding to the B
subunit (Mizuuchi et al., 1978). Mutations in the B subunit block their effect.
These antibiotics also inhibit cellular levels of supercoiling at levels consistent
with inhibition of gyrase. Steady-state kinetic experiments suggest that cou-
mar-ins are competitive inhibitors of ATP hydrolysis and DNA supercoiling, but
this does not imply that they are necessarily binding at the same site as ATP
(Sugino and Cozzarelli, 1980). Structurally they are not particularly similar to
ATP, although they do have a negative charge at neutral pH from the 4-hydroxy
group and a sugar ring.

Novobiocin

2.3.3.2 Mammalian DNA topoisomerase II

No enzyme equivalent to DNA gyrase has been found in mammalian cells,
probably because mammalian DNA is negatively supercoiled by being wrapped
36 Molecular mechanisms of drug action

around histones. There is no need, therefore, to ‘wind up’ the nucleic acid,
merely to unwind it with the help of ATP. Topoisomerase II works down an
energy gradient using the Mg-ATP complex by removing positive or negative
supercoils (relaxation). Unlike DNA gyrase it does not introduce supercoils.
Topoisomerase II recognizes topological structure rather than base sequences,
unlike other DNA-interacting enzymes, and so acts at asymmetric sites
(Osheroff et al., 1991).
In addition, in mammalian cells topoisomerase II is an integral part of the
chromosome scaffold. This structure is seen most clearly in mitosis when the
highly compacted DNA forms loops anchored at intervals to the chromatid
axis. If the histones are removed, the overall structural form of the chromo-
some still remains and one of the major proteins at the base of the chromatid
axis is topoisomerase II. The function of the enzyme is presumably, amongst
other things, to untangle loops in the sister chromosomes at anaphase.
Mutations to produce a non-functional enzyme result in a lethal event, as the
cells cannot separate their chromosomes at anaphase (Earnshaw, 1988).
Any interference with this structure is likely to be crucial to the functioning
of the cell and at least two classes of anti-tumour drugs owe their action to
a binding to the cleavable complex of enzyme and DNA; the epipodophyllo-
toxins and the anthracyclines.
Doxorubicin (adriamycin), produced by Streptomyces peucetius var. caes-
ius, is an example of the anthracyclines which were originally believed to act
by slotting in (intercalating) between nucleic acid bases like actinomycin and
thereby interfering with DNA and RNA biosynthesis. Doxorubicin, however,
also forms a complex with topoisomerase II and DNA in the same fashion as
the antibacterial drugs and DNA gyrase, and this is an essential part of its
mechanism of action. Studies with closely related structures have shown that
cleavage of DNA is related to cytotoxicity. Although the precise molecular
detail of the interaction has not been worked out, a covalently linked protein
is attached to the 5’ end of a DNA strand. The four (planar) rings of the drug
intercalate with DNA while the hydroxyl group on the side-chain interacts
with topoisomerase II (Bodley et al., 1989).

t
CCH20H

“OH
Q(&

OCHJ 0 OH
0
Doxorubicin CH3
HO 0
NHz

Doxorubicin is effective in the treatment of acute leukaemias and malignant

lymphomas and also in a variety of solid tumours. The drug is part of a cocktail
 Nucleic acid biosynthesis and catabolism 37

for the treatment of carcinoma of the ovary, breast and small cell carcinoma
of the lung, and a wide range of sarcomas (Calabresi and Chabner, 1990).
Etoposide is the most important member of the epipodophyllotoxins and
its derivation from the plant product podophyllotoxin makes an interesting
story (Stahelin and Von Wartburg, 1991). It is an example of how the process
of drug discovery actually takes place, rather than the sanitized version that
is often presented. The process took at least 20 years and involved the classic
situation of a very active impurity present in low quantity.
OH

OMe

Me

Etoposide
Podophyllotoxin inhibits the growth of cancer cells by acting as a spindle
poison and arresting the cells in metaphase with clumped chromosomes. In
an attempt to improve on the cytotoxic activity of podophyllotoxin crude
Podophyhm glucoside fraction was treated with benzaldehyde to give the
benzylidene derivatives. The mixture had activity against the 112 10 leukaemia
cell that could not be explained by the known major constituents. After con-
siderable effort at purification, a material was discovered that turned out to
have a different mechanism of action. It did not interact with microtubules,
but prevented cells from entering mitosis by arresting them in the G2 phase.
Using this information a number of analogues were synthesized, one of which
was etoposide (Stahelin and Von Wartburg, 1991).
Etoposide does not intercalate with DNA and owes its anti-tumour action
entirely to formation of ternary cleavage complexes with DNA and topoisomer-
ase II. Both single and double-stranded cleavage reactions appear to be stimu-
lated while the religation reaction is inhibited (Osheroff, 1989). The 4’-
hydroxyl is essential for both activity against the enzyme and cytotoxicity to
tumour cells (Sinha et al., 1990). Etoposide has been used in combination with
other agents for small cell carcinoma of the lung and is being tried in other
solid tumours, lymphomas and leukaemias (Henwood and Brogden, 1990). Sar-
coma is the term given to tumours arising in bone, connective tissue or muscle,
while carcinomas arise in covering or lining membranes. The latter are more
commonly encountered than sarcomas.
38 Molecular mechanisms of drug action

2.3.4 Reverse transcriptase - azidothymidine for AIDS

A number of RNA viruses that can infect man and animals are known as retro-
viruses because they use their genomic RNA to prime an enzyme to synthesize
complementary DNA. This enzyme is called reverse transcriptase and it has
both RNA and DNA-dependent DNA polymerase activity. The RNA of the viral
genome is replicated into a RNA/DNA duplex, the RNA is stripped off by
RNAase activity also found in the enzyme and the DNA acts as a template to
form double-stranded DNA (see Hostomsky et al., 1992; Darby, 1992). Reverse
transcriptase is a heterodimer composed of two subunits (66 and 51 kDa); a
polypeptide cleaved from the carboxy terminus of the larger produces the
smaller. Both subunits show nucleic acid polymerizing activity, derived from
the N-terminal domain, but only the larger has RNAase activity, based in the
C-terminus.
The viral genome encodes an enzyme known as an integrase to allow the
DNA to be inserted into the host DNA and thus become latent, subsequently
being transcribed by the host cell to produce virus particles.
If ever the medicinal scientist fraternity needed to be reminded that micro-
organisms are capable of changing in such a way as to present new threats to
the health of man, the development of acquired immune deficiency syndrome
(AIDS) as a result of virus infection has provided such a reminder. The disease
is caused by the virus known as human immunodeficiency virus (HIV) (earlier
names were human T-cell lymphotropic virus III, HTLV-III, and lymphadeno-
pathy-associated virus, LAV, which appears to be a recent import into humans
possibly from the African green monkey.
The virus infects the T cells of the immune system and thus the host
defences are seriously weakened. Consequently, the AIDS patient frequently
dies of other diseases that are normally controlled by the T cells, for example
pneumonia, caused by Pneumocystis carinii or cytomegalovirus, or an other-
wise rare form of cancer, Kaposi’s sarcoma (Shaw et al., 1985). Furthermore,
there may be degenerative changes in nervous tissue as a consequence of
infection, but it is probably too early to be sure how serious these changes are.
A family of nucleosides, based on pyrimidine and purine bases, have been
shown to prolong the life of AIDS sufferers, but not to induce a cure
(Sandstrom and Oberg, 1993). Azidothymidine (thymidine) was the frost to be
used, while didanosine (hypoxanthine), zalcitabine and lamivudine (cytidine)
followed. These agents are all converted intracellularly by host kinases to the
triphosphate which then interferes with viral nucleic acid synthesis in two
ways: the triphosphate at concentrations below lo-’ M (Mitsuya et al., 1985)
itself binds to the nucleotide binding site preventing the natural nucleotide
from binding. The enzyme can also treat the nucleotide as a false substrate by
introducing it to the growing nucleotide chain, thereby causing the chain to
terminate prematurely - there is no 3’-hydroxyl group to bind the next phos-
phate group to form a 5’-3’ diester (Yarchoan et al., 1989; Hart et al., 1992).
 Nucleic acid biosynthesis and catabolism 39

CHz-OH

Didanosine

Lamivudine

CH2 -OH

Azidothymidine

N3

In the case of lamivudine and its 5-fluoro analogue, the more active agent is
the ‘unnatural’ - or L-isomer while the b-isomer shows less activity but greater
toxicity. The suggestion is that cellular enzymes, such as cytidine deaminase,
recognize the ‘natural’ isomer thus inducing toxicity, whereas the ‘unnatural’
isomer is not so recognized (Hoong et al., 1992).
Mammalian DNA polymerases differ in their sensitivity to these drugs, with
type (Y much less (Ki l-2 X lo-*M), but types p and y Ki 2.6-70 X 10-6~
and less than 4 X lo-'M, respectively; (Yarchoan et al., 1989). These inhibi-
40 Molecular mechanisms of drug action

tory effects may be responsible for drug toxicity. The mammalian enzymes do
not use these nucleotides as false substrates.

2.3.5 BacterialRNA polymerase-rifampicinas

adimycobacterial
Rifampicin, which is derived by chemical synthesis from the rifamycins, natural
products of Nocardia mediterranae, inhibits bacterial, but not eukaryotic,
DNA-primed RNA polymerase. The drug is more effective (i.e. has a lower
MIC) against Gram-positive bacteria than Gram-negative, but this is probably
because of permeability barriers in the latter. The major use of the drug is
to treat infections produced by the mycobacteria (rod-shaped Gram-positive
bacteria), so-called because approximately 40 per cent of their cell walls con-
sists of mycolic acid - complexed to the peptidoglycan (see section 5.4).
Mycolic acid is a hydroxy fatty acid with long hydrocarbon chain branches,
the presence of which causes the bacterium to stain in a characteristic way
known as acid-alcohol fastness. Important pathogenic mycobacteria are M.
leprae, the cause of leprosy, and M. tuberculosis causing tuberculosis.
The target of rifampicin is the p subunit of RNA polymerase. The p subunit
must be complexed in the trimer ozp for specific binding to occur. With the
complete pentameric enzyme <a&‘@) rifampicin binds with a binding con-
stant of lo-9111 at 37°C (see Wehrli, 1983). If the p subunit alone is treated
with a rifampicin quinone derivative it is non-specifically modified, whereas
in the c@ complex the binding site only is modified. This suggests that the
binding site develops as a result of formation of the enzyme complex, presum-
ably by allosteric interactions (Lowder and Johnson, 1987). Mutations in the
p subunit alone, however, will generate resistance to rifampicin.

Me’

0 HO OH

Rifampicin

Binding of the drug produces abortive chain initiation through the formation
of the dinucleotide pppApU and thereby prevents further elongation of the
RNA chain. This is probably because rifampicin interferes with the binding
site for the growing RNA chain, thereby destabilizing the binding of the inter-
mediate oligonucleotides to the DNA-enzyme complex (Wehrli, 1983).
 Nucleic acid biosynthesis and catabolism 41

Rifampicin has to be given as part of a cocktail of drugs because a small

proportion of the naturally occurring enzyme is resistant. The need for drugs
active against M. tuberculosis has become more urgent in recent years as the
occurrence of tuberculosis is increasing.

2.4 Nucleo tide catabolism

2.4.1 Adenosine deaminase - 2-deoxycoformycinas

antileukaemic

Adenosine deaminase catalyses the deamination of adenosine and deoxyadeno-

sine to inosine and deoxyinosine respectively. This enzyme is important in
lymphocytic function and its absence through genetic defect leads to severe
immune deficiency (Thompson and Seegmillar, 1980). 2-Deoxycoformycin is
the most potent known inhibitor of the enzyme and acute lymphoblastic leu-
kaemia of T-cell origin responds to such inhibition, probably because T cells
are particularly rich in adenosine deaminase and the enzyme levels are raised
even further in various forms of acute lymphocytic leukaemia (Dearden et al.,
1991). The enzyme plays an important role in the salvage pathway for purines
since inosine may be phosphorylated to IMP which can, as we have seen, be
converted into GMP. Inhibition of this salvage pathway in times of high purine
requirement is likely to be critical to the cell.
The action of 2deoxycoformycin as a powerful adenosine deaminase inhibi-
tor may also be indirect. A build-up of dATP, derived by phosphorylation from
adenosine and adenylate kinases, can inhibit ribonucleotide reductase which
is particularly sensitive to the triphosphate balance. Secondly, accumulated
adenosine can drive Sadenosylhomocysteine hydrolase into reverse to form S-
adenosylhomocysteine (Her&field et al., 1983). This agent is a powerful inhibi-
tor of the methylation of RNA and DNA, possibly by antagonizing the methyl
transfer activity of S-adenosylmethionine.
2-Deoxycoformycin is a fermentation product of Streptomyces antibioticus.
The mode of action is reviewed in Agarwhal(l982). The drug is slightly more
tightly bound than its ribose analogue, coformycin (Ki of 2.5 X 10-“~ com-
pared with 1 X 10-l’ M). Since a similar relationship occurs with the substrates
with adenosine (K,,, 2.5 X 1O-5 M) and deoxyadenosine (7 X lop6 M), it is likely
that the enzyme has a greater affinity for deoxyribose than ribose. The drug
is a close structural analogue of the transition state of the enzyme reaction
(see Chapter 1 for discussion on transition states). The six-membered ring of
the purine nucleus has been expanded to a seven-membered ring by the
addition of a carbon atom. There are two double bonds in that ring so the
aromaticity is lost and the ring becomes puckered; the hydroxyl group is a
secondary alcohol, and so the keto-enol tautomerism is also lost. This leads
to a structure similar to the transition state intermediate in which a hydroxyl
42 Molecular mechanisms of drug action

group has been added at position 6 to give a tetrahedral carbon atom

(Wolfenden et al., 1977). The reaction mechanism is shown in Fig. 2.9.
This type of inhibition leads to some rather unusual kinetic results. The
binding of the inhibitor is to the form of the enzyme present in the transition
not to the ground state, and so the inhibition takes time to develop. The initial
velocity plots show only relatively moderate inhibition if the inhibitor is not
pre-incubated with the enzyme. A double reciprocal plot demonstrates charac-
teristics of competitive inhibition with a Ki of 1 X lo-’ M. With pre-incubation,
however, the plot obtained indicated a noncompetitive inhibitor with a Ki of
2.5 X 10P”~. The situation is similar to what is obtained with irreversible
time-dependent binding to an enzyme, namely that a proportion of the enzyme
is put out of action by the inhibitor because the enzyme-inhibitor complex
takes so long to dissociate (the half-life for dissociation is 25 to 30 hours). The
activity observed is that of the remaining uncomplexed enzyme, and so the

OH OH OH OH

Adenosine Intermediate

OH

2-Deoxycoformycin lnosine

Figure 2.9 Adenosine deaminase reaction.

 Nucleic acid biosynthesis and catabolism 43

K,,, remains the same while V,,,, is reduced - a classical example of non-com-
petitive behaviour. The rates of association and dissociation of this complex
are slow, possibly as a result of a slow conformational change in the enzyme
as it is transformed from the ground state to the activated transition state
(Agarwhal, 1982).

2.4.2 Xanthine oxidase - allopurinolfor gout

Xanthine oxidase is the last enzyme on the breakdown pathway of purine

bases in primates (Fig. 2.4) and it catalyses the conversion of hypoxanthine to
xanthine and of xanthine to uric acid. The latter is normally excreted, although
quantities of’the other purines may also find their way into the urine. In some
diseases, notably gout, the production of purines can be increased as a primary
cause of the disease (Scott, 1980). Enzyme deficiencies with a genetic origin
may play a part. One such case is deficiency of the salvage enzyme hypoxan-
thine phosphoribosyltransferase (HPRT) which leads to an elevated level of
hypoxanthine phosphoribosylpyrophosphate. The latter stimulates de HOUO
purine biosynthesis at the initial rate-limiting step of the formation of phos-
phoribosylamine (Fig. 2.2).
The consequence of increased purine synthesis is an increased throughput
down the catabolic pathway to uric acid. When levels of the latter rise above
saturation, crystals of monosodium urate form in the synovial fluid. The charac-
teristic symptoms of gout derive from an inflammatory response to these crys-
tals and thus closely resemble the painful joint swellings in rheumatoid
arthritis. This may occur in one joint only or in several (Kelley and Wyngarden,
1974). In advanced gout, deposits (tophi) of sodium urate form on or near
joints or tendon sheaths, which are soft initially but eventually harden
(Scott, 1980).
For therapy the major need is to lower serum uric acid levels, although anti-
inflammatory drugs will relieve the symptoms on a short-term basis. One of
the most useful drugs in effecting a long-term cure is allopurinol. Xanthine
oxidase is the target of the drug, and so serum and urine hypoxanthine and
xanthine levels are raised while, more importantly, uric acid levels are lowered.
In addition, the drug is useful when given in combination with anti-leukaemic
drugs since serum urate levels can rise sharply as the leukaemia cells die. This
is an example of secondary gout, secondary in that uric acid formation is
increased as a consequence of other changes. In this case, the danger is not
only that acute episodes of gout may occur, but also that sodium mate crystals
may form in the distal tubule of the kidney (Scott, 1980).
Clearly, if a drug is metabolized by xanthine oxidase, its action is likely to
be potentiated by allopurinol. For example, Gmercaptopurine (a drug used for
the treatment of leukaemia) is metabolized by xanthine oxidase to Gthiouric
acid, an inactive metabolite (Elion, 1967). The dose of mercaptopurine
required when given in conjunction with allopurinol must therefore be
44 Molecular mechanisms of drug action

reduced to avoid widespread toxicity which would otherwise occur if higher

mercaptopurine levels were sustained for longer periods of time.
Xanthine oxidase is a complex enzyme containing, in effect, a transport
system involving molybdenum, flavin nucleotide and two iron-sulphur centres
which convey electrons to oxygen to yield superoxide anion (0;) (Hille et
al., 1993). Allopurinol inhibits the enzyme in a complex fashion, and may be
regarded as one of the earliest examples of a suicide substrate (Spector and
Johns, 1970; Massey et al., 1970 - see Chapter 1 for a discussion on suicide
substrates). If the inhibition is studied without pre-incubation of enzyme and
inhibitor, allopurinol behaves as though it were a competitive inhibitor with
a Ki of 7 X lo-‘M. With pre-incubation in the presence of air, the inhibition
increases and is no longer competitive with substrate. Allopurinol is also a
substrate for xanthine oxidase and the product of the reaction, oxipurinol
(alloxanthine), is also an inhibitor. In the presence of xanthine as substrate
and oxygen, or anaerobically without substrate, the enzyme is inactivated by
oxipurinol. If the oxidation of xanthine, which requires the enzyme to cycle
between reduced and oxidized forms, and for the enzyme to be in an anaerobic
environment, both result in enzyme inactivation by oxipurinol, it is likely that
the reduced form of the enzyme is sensitive to oxipurinol (Massey et al., 1970).
The dissociation constant of the oxipurinol-enzyme complex is 5.4 X 10-‘“~
(Spector and Johns, 1970). Inhibition can be reversed by prolonged dialysis
or by allowing the complex to be reoxidized in the presence of air, thus con-
firming that it is the partly reduced form of the enzyme that is receptive to
oxipurinol inhibition.

Allopurinol Oxipurinol

The inactivation of reduced xanthine oxidase by oxipurinol follows first-

order kinetics by appearing to be dependent on the concentration of reduced
enzyme. This may be the result of an internal rearrangement of the enzyme-
inhibitor complex in a time-dependent fashion (Spector and Johns, 1970). The
similarity between the tight or stoichiometric binding of oxipurinol to xan-
thine oxidase, and of coformycin to adenosine deaminase was noted by Cha
et al. (1975).
Allopurinol has been found to be effective in the treatment of kala-azar
(leishmaniasis) (Berman, 1988). In this instance the drug is acting as a false
substrate for the parasite’s hypoxanthine phosphoribosyltransferase (Fig.
2.10) - much more efficiently than for the human erythrocyte enzyme (Tuttle
and Krenitsky, 1980). Subsequent enzymes convert the ribonucleotide into
 Nucleic acid biosynthesis and catabolism 45

AMP
adenylate
deaminase
1

Phosphomonoesterase

HPRT HPRT

Purine nucleoside
phosphorylase

Xanthine

Hypoxanthine

Uric acid

HPRT: Hypoxanthine phosphoribosylamine transferase

Figure 2.10 Purine nucleotide catabolism and salvage

an analogue of ATP which is then incorporated into a faulty RNA (Nelson et

al., 1979).

2.4.3 Guanylate cyclase

Angina pectoris is characterized by sudden sharp pains in the chest which

may spread out into the left arm, and is brought on by exercise, stress, emotion
or eating. A reduction in the volume of the coronary vessels (atherosclerosis)
is the cause of the usual angina, while spasm of the coronary vessels underlies
variant (Prinzmetal’s) angina. Both of these increase the demand for oxygen;
lack of oxygen (ischaemia) brings on the pain which can, therefore, result
from insufficient blood flow or increased requirement for oxygen or both.
46 Molecular mechanisms of drug action

A family of organic nitrates has been used for over a century in the treatment
of angina pectoris and congestive heart failure amongst other cardiac con-
ditions (Abrams, 1987). Glyceryl trinitrate was the first to be used in this way,
while isosorbide din&rate has also been favoured. These drugs dilate the per-
ipheral blood vessels, reducing the strain and oxygen demand on the heart,
and also dilate the coronary arteries in an area short of oxygen, thus increasing
the oxygen supply (Ignarro, 1989; Berlin, 1987).
Despite this long use it is only recently that the molecular mechanism of
their action has been identified. Nitric oxide is the key, and it may be formed
in two ways: to make compounds more water-soluble, hepatic glutathione
organic nitrate reductase catalyses the formation of nitrite ion which breaks
down to nitric oxide and denitrated compound; alternatively, nitrates react
with cysteine to produce S-nitrosocysteine which breaks down to nitric oxide.
Nitric oxide activates soluble guanylate cyclase by binding to the central iron
of the prosthetic heme Ognarro, 1989).
Cyclic GMP levels activate a cyclic-GMPdependent protein kinase which
relaxes smooth muscle by reducing the levels of free calcium inside the cell,
probably by binding to the endoplasmic reticulum (Feelisch and Noack, 1987;
Kukovetz et al., 1991). A direct correlation exists between the rate of forma-
tion of nitric oxide and the inhibition of guanylate cyclase and the elevation
of cyclic GMP (Greenberg et al., 1991).
Another more recent introduction, nicorandil which is a nitrate ester of
nicotinic acid, also works through cyclic GMP but, in addition, seems to have
another mode of action which is independent of cyclic GMP because a very
high concentration of drug is required to activate guanylate cyclase. The
activity of nicorandil varied greatly according to the tissue, unlike the other
organic nitrates - a result which is not likely to be due to a potassium channel
blockade because potassium chloride inhibits smooth muscle relaxation by
organic nitrates and the differences between nicorandil and the other esters
remained in KCl-contracted blood vessels (Greenberg et al., 1991).

Questions

1. What four bases are present in DNA? How does RNA differ from DNA?
2. What does the term synergy mean? What classes of antibacterial drugs
show synergy and why?
3. What part does the tyrosyl radical play in ribonucleotide reductase?
4. Why is the nucleic acid polymerase from retroviruses called reverse tran-
scriptase?
5. Why does rifampicin not kill E. coli?
6. What is meant by supercoiling? What advantages to a bacterial cell does
supercoiling confer?
7. How does topoisomerase I differ in the mechanism of action from topoiso-
merase II?
 Nucleic acid biosynthesis and catabolism 47

8. How do the fluoroquinolones interfere with bacterial topoisomerase II?

9. What mechanisms of action can be put forward to explain the action of
ribavirin?
10. What is meant by a transition state complex? Name an enzyme that can
be inhibited by a drug resembling such a complex.

References
Abrams, J., 1987, Dncgs, 34, 391-403.
Agarwhal, R.P., 1982, Pbarmacol. The?-., 17, 399-429.
Ariel, I., 1970, Cancer, 25, 705-14.
Ashton, W.T., Karkas, J.D., Field, A.K. and Tolman, R.L., 1982, Biocbim. Biopbys. Res.
Commun., 108, 1716-21.
Berlin, R., 1987, Dncgs, 33 (Suppl 4), l-4.
Berman, J.D., 1988, Revs. Infect. Dis., 10, 560-79.
Block, E.R., Jennings, A.E. and Bennett, J.E., 1973, Antimicrob. Ag. Cbemotber., 3,
649-56.
Blum, R.H. and Carter, SK., 1974, Ann. Intern. Med., 80, 249-59.
Bodley, A., Liu, L.F., Israel, M., Seshadri, R., Koseki, Y., Guiliani, F.C., Kirschenbaum,
S., Silber, R. and Potmesil, M., 1989, Cancer Res., 49, 5989-93.
Bonadonna, G. and Valagussa, P., 1981, New Engl. J. Med., 304, 10-15.
Calabresi, P. and Chabner, B.A., 1990, p. 1243 in the Pharmacological Basis of Tbera-
peutics, 8th edn, (eds A.G. Gilman, T.W. Rall, A.S. Nies and P. Taylor), Pergamon,
New York.
Cha, S., Agarwhal, R.P. and Parks, R.E., 1975, Biocbem. Pbarmacol., 24, 2187-97.
Chang, T.-W. and Heel, R.C., 1981, Drugs, 22, 111-28.
Chatis, P.A. and Crumpacker, C.S., 1982, Antimicrob. Ag. and Cbemotber., 36,
1589-95.
Chen, G.L., Yang, L., Rowe, T.C., Halligan, B.D., Tewey, K.M. and Liu, L.F., 1984, J.
Biol. Cbem., 259, 13560-6.
Churchill, W.S., 1951, In: The Hinge of Fate, Vol 4 in the Second World War (London:
Cassell), p. 582.
Cohen, J., 1982, Lancet, ii, 532-7.
Cozzarelli, N.R., 1980, Science, 207, 532-7.
Darby, G., 1992, Biocbem. Sot. Trans., 20, 505-B.
Dearden, C., Matutes, C. and Gatovsky, D., 1991, Brit. J. Cancer, 64, 903-6.
De Clerq, E. and Walker, R.T., 1984, Pbarmacol. Tber., 26, l-44.
Diasio, R.B., Bennett, J.E. and Myers, C.E., 1978, Biocbem. Pbarmacol., 27, 703-7.
Di Marco, A., 1975, Cancer Cbemotber. Rep., 6, 91-106.
Drlica, K., 1992, Molec. Microbial., 6, 425-33.
Earnshaw, W.C., 1988, Bioessays, 9, 147-51.
Elford, H.L., Wamplen, G.L. and Van’t Riet, B., 1979, Cancer Res., 39, 844-51.
Elford, H.L., Van’t, Riet B., Wamplen, G.L., Lin, A.L. and Elford, R.M., 1981, Adv.
Enzyme Regul., 19, 151-68.
Elgren, T.E., Lynch, J.B., Juarez-Garcia, C., Munck, E., Sjoberg, B.-M. and Que, L., 1991,
J. Biol. Cbem., 266, 19265-8.
Elion, G.B., 1967, Fed. Proc., 26, 898-904.
Feelisch, M. and Noack, E.A., 1987, Eur. J. Pbarmacol., 139, 19-30.
Ferone, R., Burchall, J.S. and Hitchings, G.H., 1969, Molec. Pbarmacol., 5, 49-59.
Fry, M. and Pudney, M., 1992, Biocbem. Pbarmacol., 43, 1545-53.
Gellert. M., 1980, Annu. Rev. Biocbem., 50, 879-910.
48 Molecular mechanisms of drug action

Gilbert, B.E. and Knight, V., 1986, Antimicrob. Ag. Cbemotber., 30, 201-5.
Graham, F.L. and Whitmore, G.F., 1970, Cancer Res., 30, 2636-44.
Greenberg, S.S., Cantor, E., Ho, E. and Walega, M., 1991, J. Pharmacol. Exptl. Tber.,
258, 1061-71.
Hammond, C.B., Weed, J.C., Barnard, D.E. and Tyray, L., 1981, Cancer, 31, 322-32.
Hammond, D.J., ButchelI, J.R. and Pudney, M., 1985, Molec. Biocbem. Parasitol., 14,
97.
Hart, G., Orr, D.C., Penn, C.R., Figueiredo, H., Gray, N.M., Boehme, R.E. and Cameron,
J.M., 1992, Antimicrob. Ag. Cbemotber., 36, 1688-94.
Heidelberger, C., Danenberg, P.V. and Moran, R.G., 1983, Adv. Enzymol., 54, 57-119.
Henwood, J.M. and Brogden, R.N., 1990, Drugs, 39, 438-90.
HershBeld, MS., Kredich, N.M., Keller, C.A., Mitchell, B.S., Kurtzberg, J., Kinney, T.R.
and Falletta, J.M., 1983, Cancer Res., 43, 3451-6.
Hille, R., Kim, J.H. and Hemann, C., 1993, Biochemistry, 32, 3973-80.
Hitchings, G.H. and Smith, S.L., 1980, Adv. Enzyme Regul., 18, 349-71.
Hooper, D.C. and Wolfson, J.S., 1991, N. Engl. J. Med., 324, 384-94.
Hoong, L.K., Strange, L.E., Liotta, D.C., Koszalka, G.W., Burns, C.L. and Schinazi, R.F.,
1992,J. Org Cbem., 57, 5563-5.
Horowitz, D.S. and Wang, J.C., 1987, J Biol. Cbem., 262, 5669.
Hostomsky, Z., Hostomska, Z., Fu, T.-B. and Taylor, J., 1992,J. ViroL, 66, 3179-82.
Hsiang, Y.-H. and Liu, I.F., 1988, Cancer Res., 48, 1722-6.
Hudson, A.T., Dickins, M., Ginger, C.D., Gutteridge, W.E., Holdich, T., Hutchinson,
D.B.A., Pudney, M., Randall, A.W. and Latter, VS., 1991, Drugs Exptl. Clin. Res.,
17, 427-35.
Hurly, M.G.D., 1959, Trans. R. Sot. Trop. Med. Hyg., 53, 412-13.
Ignarro, LJ., 1989, Pbarm. Res., 6, 651-9.
Kager, P.A., Rees, P.N., Wellde, B.T., Hockmeyer, W.T. and Lyerley, W.T., 1981, Trans.
R. Sot. Trop. Med. Hyg., 75 556-9.
Kelley, W.N. and Wyngaarden, J.B., 1974, Adv. EnzymoL, 41, l-33.
Kennedy, B.J. and Yarbro, J.W., 1966,J. Am. Med. Assoc., 195, 1038-43.
Kisner, D.L., Schein, P.S. and Macdonald, J.S., 1981, Rec. Results Cancer Res,, 79,
28-40.
Kukovetz, W.R.D., Holzmann, S. and Schmidt, K., 1991, Eur. HeatiJ, 12 (Suppl E),
16-24.
Lassmann, G., Thelander, L. and Graslund, A., 1992, Biocbem. Biopbys. Res. Commun.,
188, 879-87.
Ljungmann, P., Lonngvist, B., Gahrton, G., Ringden, O., Sundgvist, V.A. and Wahren,
B., 1986,J. Znfect. Dis. (Suppl B), 840-7.
Lowder, J.F. and Johnson, R.S., 1987, Biocbem. Biopbys. Res. Commun., 147, 1129-36.
Lucas, A.O., Hendrickse, R.G., Okubadse, O.A., Richards, W.H.G., Neal, R.A. and Kofie,
B.A.K., 1969, Trans. R. Sot. Trop. Med. Hyg., 63, 216-29.
Major, P.P., Agarwhal, R.P. and Kufe, D.W., 1983, Cancer Cbemotber. Pbarmacol., 10,
128-38.
Major, P.P., Egan, E.M., Herrick, D.J. and Kufe, D.W., 1982, Biocbem. Pbarmacol., 31,
2937-40.
Mar, E.-C. and Huang, E.-S., 1979, Znteruirol., 12, 73-83.
Massey, V., Komai, H., Palmer, G. and Elion, G.B., 1970,J. Biol. Cbem., 245, 2837-44.
Matthews, D.A., Alden, R.A., Bolim, J.T., Filman, D.J., Freer, ST., Hamlin, R., Wim,
G.J.H., Kislink, R.L., Pastore, EJ., Plante, L.T., Xuong, N. and Krant, J., 1978, J.
Biol. Cbem., 253, 6946-54.
Matthews, D.A., Bolin, J.T., Burridge, J., Filman, D.J., Volz, K.W. and Kraut, J., 1983,
J. Biol. Cbem., 260, 392-9.
McClung, H.W., Knight, V., Gilbert, B.E., Wilson, S.Z., Quarles, J.M. and Divine, G.W.,
1983, J Am. Med. Assoc., 249, 2671-4.
 Nucleic acid biosynthesis and catabolism 49

Miller, W.H. and Miller, R.L., 1982, Biochem. Pbarmacol., 31, 3879-84.
Mitsuya, H., Weinhold, K.J., Furman, P.J., St. Clair, M.H., Nusinoff-Lehrman, S., Gallo,
R.C., Bolognese, D., Barry, D.W. and Broder, S., 1985, Proc. Nat. Acad. Sci. (U.S.A.),
82, 7096-100.
Mizuuchi, K., O’Dea, M.H. and Gellert, M., 1978, Proc. Nat. Acad. Sci. (LLS.A.), 75,
5998.
Moody, M.R. and Young, V.M., 1975, Antimicrob. Ag. Cbemotber., 7, 836-9.
Muller, W.E.G., Zahn, R.K., Brittlingmaier, K. and Felke, D., 1977, Ann. N.Y Acad. Sci.,
28e, 34-48.
Nelson, D.J., LaFon, S.W., Tuttle, J.V., Miller, W.H., Miller, R.L., Krenitsky, T.A., Elion,
G.B., Berens, R.L. and Marr, J.J., 1979,J. Biol. Cbem., 254, 11544-9.
Osheroff, N., 1989, Biochemistry, 28, 6157-60.
Osheroff, N., Zechiedrich, E.L. and Gale, K.C., 1991, Bioessays, 13, 269-75.
Part-is, D.S. and Harrington, J.E., 1982, Antimicrob. Ag. Cbemotber., 22, 71-7.
Patterson, J.L. and Fernandez-Larsson, R., 1990, Rev. Infect. Dk, 12, 1139-46.
Ramachandran, S., Godfrey, JJ. and Lionel, N.D.W., 1978,J. Trop. Med. Hyg., 81, 36-9.
Reardon, J.E. and Spector, T., 1989, J Biol. Cbem., 264, 7405-l 1.
Reece, RJ. and Maxwell, A., 1991, Crit. Rev. Biocbem. Mol. Biol., 26, 335.
Rowe, T.C., Chen, G.L., Hsiang, Y.-H. and Liu, L.F., 1986, Cancer Res., 46, 2021-6.
Rubin, R.N., 1981, Clin. Ther., 4, 74-94.
Sandstrom, E. and Oberg, B., 1993, Drugs, 45, 488-508.
Scott, J.T., 1980, Brit. Med.=/., 281, 1164-6.
Shaw, G.M., Harper, M.E., Hahn, B.H., Epstein, L.G., Gadjusek, D.C., Price, R.W., Navia,
B.A., Petito, C.K., O’Hara, C.J., Groopman, J.E., Cho, E.-S., Oleske, J.M., Wong-staal,
F. and Gallo, R.C., 1985, Science, 227, 177-82.
Shepard, CC., Ellard, G.A., Levy, L., Upromolla, V., Pattyn, S.R., Peters, J.H., Rees,
R.J.W. and Waters, M.F.R., 1976, Bull. WHO, 53, 425-33.
Sinha, B.K., Politi, P.M., Eliot, H.M., Kerrigan, 0. and Pommler, Y., 1990, Eur. J Cancer,
28, 590-3.
Spector, T. and Johns, D.G., 1970, J Biol. Cbem., 245, 5079-86.
Stahelin, H.F. and Von Wartburg, A., 1991, Cancer Res., 51, 5-15.
Streeter, D.G., Witkowski, J.T., Khare, G.D., Sidwell, R.W., Bauer, R.J., Robins, R.K.
and Simon, L.N., 1973, Proc. Nat. Acad. Sci. (U.S.A.), 70, 1174-8.
Stubbe, J.A., 1989, Annu. Rev. Biocbem., 58, 257-85.
Sugino, A. and Cozzarelli, N.R., 1980, J. Biol. Cbem., 255, 6599-604.
Thelander, L. and Reichard, P., 1979, Annu. Rev. Biocbem., 48, 133-58.
Thompson, L.F. and Seegmiller, J.E., 1980, Adv. Enzymol., 51, 167-210.
Tuttle, J.V. and Krenitsky, T.A., 1980,J. Biol. Cbem., 255, 909-16.
Wang, J.C., 1991, J. Biol. Cbem., 266, 6659-62.
Wehrli, W., 1983, Rev. Infect. Dis., 5, Suppl. 3, S407-11.
Wentland, M.P., Pemi, R.B., Dorff, P.H. and Rake, J.B., 1988,J. Med. Cbem., 31, 1694-7.
Whelan, W.L. and Kerridge, D., 1984, Antimicrob. Ag. Cbemotber., 26, 570-4.
Whitley, R.J. and Gnann, J.W., 1992, New Engl. J. Med., 327, 782-9.
Whitley, R.J., Alford, C., Hess, F. and Buchanan, B., 1980, Drugs, 20, 267-82.
Wolfenden, R.L., Wentworth, D.F. and Mitchell, G.N., 1977, Biochemistry, 16, 5071-7.
Woods, D.D., 1962, J. Gen. Microbial., 29, 687-702.
Wray, S.K., Gilbert, B.E., Noall, M.W. and Knight, V., 1985, Antiviral Res., 5, 29-37.
Yarchoan, R., Mitsuya, H., Myers, C. and Broder, S., 1989, New Engl. J. Med., 321,
726-38.
 Chapter 3

Protein biosynthesis

3. I Introduction
The synthesis of protein from individual amino acids is an essential feature of
living organisms. Major differences exist between the process in eukaryotes
and prokaryotes, and these have been exploited serendipitously to produce
anti-bacterial agents. Subtle differences are found between higher and lower
eukaryotes (for example mammalian and yeast cells), but these have not so
far indicated a compound that could be used to kill pathogenic fungi.
This chapter is, therefore, concerned solely with anti-bacterial
agents.
Protein synthesis in eukaryotes and prokaryotes takes place in ribosomes,
both in the cytoplasm and in mitochondria, but as no drugs are believed to
owe their mechanism of action to interference with the mitochondrial process
(although chloramphenicol does interfere with this process in a few individ-
uals giving rise to side-effects), only the cytoplasmic system will be considered
in this chapter. Protein synthesis can be divided into three stages: chain
initiation, elongation and termination. Although almost all the medically useful
inhibitors of protein synthesis act on the elongation stage, initiation will also
be discussed in order that the whole process may be understood (reviewed
in Kozak, 1983; Maitra et al., 1982). Eukaryotic protein synthesis is reviewed
by Merrick (1992).
An amino acid first has to be activated by combination with a transfer RNA
(tRNA) that is specific for that amino acid. This is done in two stages, both
catalysed by the same specific aminoacyl-tRNA ligase or synthetase: the amino
acid first reacts with ATP and Mg*+ to form an enzyme-bound amino acid
adenylate and then the amino acid is transferred to the appropriate tRNA, with
the concomitant release of AMP. The triplet anti-codon on the tRNA recognizes
the codon for that amino acid on the messenger RNA, this complex formation
takes place on the ribosome in the presence of GTP (Pig. 3.1).
Protein synthesis occurs on the ribosome, a complex particle containing
various proteins and RNA. In bacteria, the overall sedimentation coefficient of
the ribosome is 7OS, and it readily dissociates into unequal subunits of 30s
(containing 21 proteins and 16s RNA) and 50s (34 proteins, and both 23s and
5S RNA), respectively. Ribosomes are normally linked together by a messenger

51
52 Molecular mechanisms of drug action

Table 3.1 Antibacterial drugs discussed in Chapter 3

Section hug Target

3.2 Aminoglycosides 30s ribosomal subunit

3.3 Chloramphenicol 50s ribosomal subunit
3.4 Tetracycline 30s ribosomal subunit
3.5 Erythromycin; Azithromycin 50s ribosomal subunit
3.6 Clindamycin SOS ribosomal subunit

RNA (mRNA) molecule like beads on a wire (polysomes), with the growing
peptide chain at increasing lengths according to the ribosomal position on the
mRNA. The ribosome directs the addition of amino acyl groups to a growing
peptidyl tRNA by a condensation reaction that forms the peptide bond, and
then moves with its growing chain along the messenger RNA to the next
codon to be read.
In prokaryotes, protein synthesis is always initiated by formylmethionyl-
tRNA @Met-tRNA), while in eukaryotes this role is carried out by unformylated
methionine. Formylation of methionine takes place after the methionine is
linked to a form of tRNA specific for chain initiation. Also required are three
proteins or initiation factors called IF 1 to 3 and also GTP. IF 1 and 2 are
needed to position the mRNA and fMet - tRNA on the ribosome, while the
role of IF 3 is to recognize the mRNA. GTP is hydrolysed to GDP in the process.
In order to assemble the active complex, the three initiation factors bind to
the 30s subunit of the ribosome. FMet-tRNA and mRNA bind to the complex
with a specific start codon, AUG, as part of the binding site on the mRNA.
GTP also binds at this stage. A 50s subunit joins the complex to form the
functional 70s ribosome, and GTP is hydrolysed to GDP and phosphate by
ribosome-bound IF 2. IF 1 and 2 are then released from the complex so that
fMet-tRNA is unblocked ready for peptide bound formation.
Chain elongation requires two sites on the ribosome (Pig. 3.1): one called
the A-site (acceptor) where the fMet-tRNA is initially located before it transfers
to the P-site (peptide nearer the 5’ end of the nucleic acid). The A-site is
located on the 3’ side of the P-site and is largely on the 30s subunit. The 50s
subunit carries the P-site, although there are clearly interactions between the
two subunits as the substrates bind. The incoming aminoacyl-tRNA binds to
the A-site, and then the fMet is condensed with the second amino-acid residue
to make a dipeptidyl-tRNA; a reaction which requires a protein known as
elongation factor 1 together with GTP which is hydrolysed in the process. The
reaction is catalysed by a specific protein on the ribosome itself, and the spent
tRNA is released from the ribosome.
The deacylated tRNA vacates the P-site and the dipeptidyl-tRNA is transferred
back to the P-site still hydrogen-bonded to its mRNA codon - a process which
requires the intervention of another molecule of GTP and elongation factor 2.
 Protein biosynthesis 53

ATT-- --

Binding of fMet-tRNA Activation

I Thr

Release of spent fMet-tRNA Peptide bond formation

Figure 3.1 Chain initiation and formation of first peptide link.

This allows a third aminoacyl-tRNA defined by the mRNA codon to take up

position in the A-site and the condensation process is repeated etc. (Pig. 3.1).
Protein release factors (RF1 and RP2) are needed in order to terminate the
process. They recognize ‘stop’ codons: UAA, UAG and UGA. Termination is
stimulated by GTP and probably requires the hydrolysis of GTP by one of the
release factors.
54 Molecular mechanisms of drug action

The drugs that are discussed in this chapter, namely the aminoglycosides,
tetracyclines, chloramphenicol, erythromycin and clindamycin, have their
major effect on the chain elongation procedure. They are specific for protein
synthesis in bacteria; apart from the tetracycline family which also inhibit euka-
ryotic protein synthesis. Nevertheless, despite this specificity, without excep-
tion they have undesirable side-effects and, in addition, their use over several
decades has induced a considerable level of resistance among the bacterial
population. For both these reasons, their use has been superseded, except in
particular instances, by the much safer plactams (penicillins and
cephalosporins) discussed in Chapter 5.
Bacteria may be divided into two main classes as to whether they take up
the Gram stain; the cells are treated with crystal violet and then with iodine
to fur the stain, followed by decolourization with ethanol or acetone. Finally
the cells are exposed to safranine, a counterstain. Gram-positive microorgan-
isms stain violet, whilst Gram-negative stain red. This difference in response
to a histological stain is derived from major differences in the cell wall of the
two types of bacteria; Gram-positive cells contain teichoic acids, polymers of
glycerol or ribitol phosphate, unlike Gram-negative cells, which may explain
the affinity of the Gram-positive cell wall for a basic dye such as crystal violet.
It is, however, not only a matter of chemistry but also of integrity as to
whether a cell will take up the Gram stain because old, ruptured or dead Gram-
positive cells may stain as though they were Gram-negative. Furthermore,
Gram-positive cells contain very little lipid whereas Gram-negative are rich in
lipid. In sharp contrast, one portion of the wall that both types have in com-
mon, but contributes more to the Gram-positive cell, is peptidoglycan (also
known as murein or glycopeptide; Chapter 5).

3.2 Aminoglycosides

Aminoglycosides are a family of antibiotics produced mainly by streptomycetes

(names ending in -mycin) or micromonospora (ending in -micin). Structurally,
they consist of aminosugars connected to a central hexose/aminocyclitol -
streptidine in streptomycin or L-deoxystreptamine in the other drugs men-
tioned in this chapter. Streptomycin was the first to be discovered but now
is much less widely used than previously, while tobramycin is frequently
regarded as the drug of choice when an aminoglycoside is indicated. Amikacin,
a semisynthetic agent, is useful for the treatment of infections caused by amino-
glycoside-resistant bacteria (Bailey, 1981).
Aminoglycosides are most useful therapeutically for infections caused princi-
pally by Gram-negative organisms such as the Enterobacteriaceae and pseudo-
monads, and are largely ineffective against Gram-positive agents such as strep-
tococci (DeTorres, 1981). In aerobic environments, these drugs are
bactericidal, i.e. they kill the bacterium (in certain instances a drug might only
prevent the bacterium from growing and multiplying and would be termed
 Protein biosynthesis 55

bacteriostatic). Aminoglycosides, unlike other antibacterials, are taken up by

the bacterium in three phases. The first phase is the binding of the drug to
the outer surface of the cell wall and is rapid and passive. In the second phase,
the drug passes across the cytoplasmic membrane by a fairly slow process,
dependent on the energy derived from electron transport which is rate-limit-
ing. The membrane potential is negative inside and drives the uptake (the
aminoglycosides are cationic and thus attracted inwards). The third phase is
fast and is probably the uptake of drug into the ribosome (see Amyes and
Gemmell, 1992).

Aminoglycosides

CHzNHz

Rl R2 R3 R4

Amikacin OH OH COCH(OH)(CH2)2NH2 H

Tobramycin H NH2 H NH2

The primary intracellular site of action of the aminoglycosides is the 30s

ribosomal subunit which consists of 21 proteins and a single 16s molecule of
RNA. A single substitution of asparagine for lysine in protein S12 (i.e. protein
12 of the smaller ribosomal subunit as numbered by position of the protein
on two-dimensional gel electrophoresis (Kaltschmidt and Wittman, 1970) of
Escherichiu coli prevents streptomycin binding to the ribosome (Birge and
Kurland, 1969) but it is doubtful whether this is the only protein involved
(Land0 et al., 1976; Davies and Courvalin, 1977).
Dihydrostreptomycin binds to the 30s subunit with a binding constant of
2.7 X ~O-‘M (Iando et al., 1976). Competition is observed between strepto-
mycin and dihydrostreptomycin, not surprisingly, but not with gentamicin;
indeed gentamicin appears to bind to the 50s subunit since protein ~16
(protein 16 of the large ribosomal subunit) is altered in a gentamicin-resistant
mutant (Buckel et al., 1977). Tobramycin appears to bind to both subunits of
the ribosome and at several different sites, since no saturation curve kinetics
can be obtained when increasing concentrations of [3H]tobramycin are used.
Nevertheless tobramycin antagonism of acetyltobramycin binding suggests a
56 Molecular mechanisms of drug action

binding constant of 1.42 X 10e6~ for the most tightly bound site (LeGoffic et
al., 1979).
One major effect of streptomycin binding is to cause misreading of the mess-
age so that the wrong amino acids are inserted into the growing polypeptide
chain (Tai et al., 1978). The aminoglycosides vary in their ability to cause
misreading of the genetic code, presumably as a result of their varying affinity
for different ribosomal proteins. This effect is found both in cell-free systems
and in whole bacteria. Streptomycin can also interfere with chain elongation
by inhibiting, in a partial and reversible manner, protein synthesis by the full
complex of ribosomes attached to messenger RNA (polysomes).

Streptomycin Gentamicin Cl

The drug does not, however, affect the process of initiation. It merely
ensures that the initiation complex formed in its presence is less stable and
less likely to engage in appreciable protein synthesis. The presence of the full
complex would appear to mask the major streptomycin binding site, however,
as streptomycin will additionally bind irreversibly to uncomplexed ribosomes,
and thus cause the ribosomes to fall off the messenger RNA prematurely. The
drug-bound ribosomes are impaired in IF-3-dependent dissociation and become
engaged in abortive initiation and elongation (Wallace et al., 1973). These two
types of effect may cause the observed phenomena of partial inhibition of
synthesis at lower concentration and total inhibition at higher concentration
leading to cell death, in that at low concentrations streptomycin will not be
able to block all the initiation sites in a cell but merely interfere transiently
with polysomal complexes (Wallace et al., 1973).
Drug resistance can occur, as one might expect, by alteration of the ribo-
some. Furthermore, in an anaerobic environment where the energy of oxidat-
ive phosphorylation is not available to facilitate the active transport of the drug
across the outer membrane into the periplasmic space, otherwise sensitive
bacteria become resistant. Clinically, however, the most important mechanism
of resistance is by the development of metabolizing enzymes, secreted into
 Protein biosynthesis 57

the periplasmic space, that inactivate the antibiotic. Acetylation of amino

groups and phosphorylation or adenylation of hydroxyl groups can all occur.
Plasmids carry the genes for no fewer than 20 such enzymes (Amyes and Gem-
mell, 1992; Bryan, 1988).
The use of aminoglycosides is restricted by a narrow therapeutic margin
since they are in varying degrees toxic to the ear and kidney (Phillips, 1982).
Streptomycin, in particular, damages the auditory and vestibular branches of
the eighth cranial nerve. This drug is still in use for the treatment of tubercu-
losis, although isoniazid and rifampicin are sometimes preferred.

3.3 Chloramphenicol

Chloramphenicol is a natural product isolated from the culture filtrate of Strep-

tomyces venezuelue. It is a molecule of low molecular weight, characterized
by the presence of a nitrobenzene group, unusual both for a natural product
and also for a drug. The drug has a broad spectrum of activity against both
Gram-positive and -negative organisms, and is able to penetrate freely into host
tissues. In general, chloramphenicol is believed to be bacteriostatic, although
it can be bactericidal to some species - notably Haemophilus injluenzae
which causes meningitis (Turk, 1977).
The drug binds primarily to the 50s subunit of the bacterial ribosome and,
in vitro, protein Ll6 appears to be the target (Nierhaus and Nierhaus, 1973).
When monoiodoamphenicol is incubated with whole cells, however, proteins
~6 and L24 are labelled in addition to ~16, suggesting that all three proteins
are involved with the receptor site (Pangs and Messer, 1976). The binding
constant for chloramphenicol binding to free ribosomes is 6 X ~O-‘M
(Fernandez-Munos et al., 1971). L-Chloramphenicol is the active form.

0
II
NHC - CHCI,

R
Chloramphenicol NO2
Thiamphenicol SO&H3

Chloramphenicol also interferes with a rather interesting reaction called the

puromycin fragmentation reaction. Puromycin is a natural product that
resembles the terminal adenosine of an aminoacyl- (or peptidyl-) tRNA (Pig.
3.2). Puromycin can block chain elongation both by substituting for an
incoming aminoacyl-tRNA at the A-site, and by accepting the peptidyl-tRNA
bound at the P-site. The peptidyl-puromycin thus formed is incapable of further
reaction as it contains an amide link attached to the ribose ring rather than
58 Molecular mechanisms of drug action

HOCHz 0 R,-O-CH2 0

0 d
b t)H

O-;-CH--R2

AH,
Puromycin Terminal adenosine of an
aminoacyl (or peptidy&tRNA
R,represents transfer RNA
R2 is the side-chain of an amino acid or a growing peptide.
Puromycin resembles methyltyrosine sufficiently closely to accept the growing peptide chain. No
futher change can take place because of the amide link at the 3’ position of the sugar ring in
puromycin instead of the expected, and more easily broken, ester linkage.
Figure 3.2 The resemblance of puromycin to an aminoacyl adenosine.

the expected ester; the peptide chain consequently falls off the ribosome. AU
of the peptides so formed have the correct amino acid at the N terminus but
puromycin at the C terminus.
The overall process is known as the puromycin fragmentation reaction and
requires 50s subunits, a peptidyl- or aminoacyl-tRNA fragment, puromycin and
K+, Mg*+ and the presence of 33 per cent ethanol. This reaction is inhibited
by other antibiotics that bind at or near the A-site of the ribosome, notably
chloramphenicol with a Ki of approximately lo-5111 (Pestka, 1970).
Chloramphenicol was extremely useful in earlier decades for combatting a
wide range of infections, but it soon became clear that the drug was respon-
sible for haematological side-effects, notably bone marrow suppression and,
more seriously, aplastic anaemia. One in 25 000 patients who take the drug
succumb to the latter condition, which is frequently fatal (Kucers, 1982). Bone
marrow suppression is a consequence of mitochondrial injury which appears
to result from the inhibition of mitochondrial protein synthesis. Aplastic anae-
mia, however, may result from reduction of the nitro group of the drug to
less chemically stable intermediates since thiamphenicol, which has an SO&H3
group instead of the nitro, does not show the same toxicity (Yunis et al.,
1980).
The recognition, however, that various serious infections, notably with
anaerobic bacteria such as Bacteroides fragilis, are sensitive to chloramphen-
icol has to some extent brought back the use of the drug - particularly when,
as in this case, the safer @lactams are rendered less effective by the bacterial
production of plactamases. Nevertheless, chloramphenicol therapy should be
 Protein biosynthesis 59

limited to those infections for which the benefit of the drug outweighs the
risks of the side-effects. If other drugs are available that are equally effective
but are less toxic, they should be used instead (see Kucers, 1982).
One condition where chloramphenicol appears to be the antibiotic of
choice is for the treatment of meningitis and brain abscesses caused by
Haemophilus influenzue type b (Kucer, 1982) although, in the USA, chloram-
phenicol is recommended only for those strains of organism that are ampicillin
resistant (AMA Drug Evaluation, 1983). B. frugilis infection of the central ner-
vous system and systemic infections by Salmonella species (e.g. typhoid fever)
are often treated by chloramphenicol (AMA Drug Evaluation, 1983).

3.4 Tetracyclines

Various Streptomycetes elaborate members of the tetracycline family of anti-

biotics, notably oxytetracycline from S. rimosus and chlortetracycline from S.
aureofuciens. Later additions to the family, such as minocycline and doxy-
cycline, have been synthesized chemically and have clinical advantages in
being better absorbed by host tissues.
For drugs which have been used for so many years, there is still a surprising
amount of disagreement as to which protein is the target of their action. Tetra-
cycline binds to the 30s ribosomal subunit in a 1: 1 fashion. Specific targets
are primarily proteins S4 and S18 with S7, S13 and S14 as minor targets as
shown by photoincorporation studies with 70s ribosomes (Goldman et al.,
1980). There is only one strong binding site for tetracyclines on the ribosome
and it is likely that the occupancy of this site is sufficient to block protein
synthesis. The binding constant is 2 X 10e5~ (T&ton, 1977). Tetracycline
binding stabilizes the ribosome thermally and raises the melting temperature
of the ribosomal RNA. This effect correlates moderately well with the inhi-
bition of polyphenylalanine synthesis for a series of tetracyclines (T&ton,
1977).

Chlortetracycline
Oxytetracycline
Doxycycline
Minocycline

As a consequence of this binding, aminoacyl-tRNA is prevented from binding

to the A-site, and so no peptide bond can be formed. Elongation factor, EF,,
60 Molecular mechanisms of drag action

dependent hydrolysis of GTP is not affected by tetracycline (Modolell et al.,

1971). The interaction between the ribosome and EF, may be interrupted
directly, or indirectly by interfering with the process whereby EF, recognizes
the anti-codon on the tRNA (Smythies et al., 1972).
Some interesting studies have been carried out on the process by which
tetracyclines gain access to the cell (reviewed in Chopra and Howe, 1978). In
Gram-negative species, passive diffusion of the non-ionized form of the drug
takes place through hydrophilic pores in the outer cell membrane, specifically
through protein lA, one of three proteins situated there. In contrast, mino-
cycline and doxycycline, which are more lipophilic than the original natural
products, pass more readily through the lipid bilayer rather than the pore. The
result is better uptake, as shown by studies on E. coli in which the uptake of
minocycline was 10 to 20 times more rapid than that of tetracycline (McMurray
et al., 1982). The second process involves an energy-dependent active trans-
port system that pumps tetracycline through the cytoplasmic membrane. Less
is known about transport into Gram-positive bacteria but it also requires an
energy-dependent process. Plasmids which confer resistance to tetracyclines
in E. coli, and possibly other species, code for proteins which are located in
the cell envelope and probably interfere with the uptake of the drug (Chopra
and Howe, 1978).
The tetracyclines are active against a wide range of Gram-positive and -nega-
tive bacteria but a number of organisms such as Pseudomonas aeruginosa,
and staphylococcal and streptococcal species have acquired resistance poss-
ibly as a result of indiscriminate use of these drugs. Tetracyclines are still used
for the treatment of chlamydial infections, notably urethritis, but resistance
limits their use in other infections (reviewed in Speer et al., 1992).

3.5 Erythromycin
Erythromycin is a larger molecule than the other antibiotics considered in this
chapter, with a molecular weight of 734. It consists of a 14-membered lactone
ring with two sugar molecules attached, and was first detected as a product
of Streptomyces elythreus. Erythromycin can be either bacteriostatic or bac-
teriocidal, depending on both the organism and the concentration of the drug.
In general it is not active against most aerobic Gram-negative bacilli. It is active
against some Gram-positive cocci such as Streptococcus pneumoniae, and is
particularly useful in patients who show an allergy to penicillins. Erythromycin
can also be helpful against staphylococcal strains that are resistant to penicil-
lins. It is still the drug of choice in Legionnaire’s disease caused by Legionella
pneumophila and for the treatment of diphtheria (Kucers, 1982; Blagg and
Gleckman, 1981). Unfortunately, erythromycin usage is not free from side-
effects, notably liver toxicity, reversible hearing loss and pseudomembraneous
colitis (Blagg and Gleckman, 1981) - see under clindamycin.
The binding site for erythromycin lies on the 50s subunit of the 70s ribo-
 Protein biosynthesis 61

Erythromycin

some. The precise position remains uncertain, however, as both protein ~16
and the ribosomal RNA itself have been implicated (B&son-Noel et al., 1!988).
The loss of protein ~16 by washing with ammonium chloride parallels the loss
of binding to the ribosome in one study (Bernabeu et al., 1977) whereas
methylation of adenine base in position 2058 of E. coli 23s RNA prevents the
binding of various macrolides to the ribosome (Cundliffe, 1987). It is possible
that the macrolide binding site is composed of both protein and RNA, particu-
larly as free ~16 or any other protein will not bind erythromycin.
The action of erythromycin is to cause premature release of the growing
peptidyl-tRNA chain from the ribosome. This may occur either because the
drug weakens the binding between ribosome and peptidyl-RNA or, alterna-
tively, because it interferes with the translocation step when the newly syn-
thesized peptidyl-tRNA moves from acceptor to donor site.
A recent addition to the antibacterial armoury is a close relative of erythro-
mycin, azithromycin. Structurally, it contains a nitrogen atom in the macrolide
ring as its name implies (azote is Latin for nitrogen) and acts by a similar
mechanism. This drug is slightly less active against Gram-positive bacteria but
rather more active against Gram-negative than its parent. It is also apparently
less toxic (Drew and Gallis, 1992).

3.6 Clindamycin

Clindamycin is the 7deoxy,7-chloro analogue of lincomycin, which was

initially obtained from Weptomyces lincolnensis. The former is less toxic,
however, and its use has superseded that of lincomycin. Broadly speaking, the
spectrum of activity of clindamycin resembles that of erythromycin, and the
former is of considerable use in the treatment of infections caused by anaer-
obic bacteria, notably B. frugilis (Dhawan and Thadepalli, 1982). Central ner-
vous system infections are best managed by chloramphenicol which crosses
the blood-brain barrier more effectively (Fass et al., 1973). More recent work
suggests that metronidazole and cefoxitin are at least as effective as clindamy-
tin in vivo (Dubreil et al., 1984).
62 Molecular mechanisms of drug action

Antibiotics of the lincomycin family inhibit protein synthesis in whole bac-

teria, and in cell-free systems, by interacting with the 50s subunit at a site very
close to that occupied by erythromycin and chloramphenicol. This is shown
by a parallel loss of binding for all three antibiotics when protein ~16 is washed
out of the ribosome by ammonium chloride and ethanol treatment (Bernabeu
et al., 1977). Furthermore, lincomycin inhibits the binding of chloramphenicol
and erythromycin and vice versa (Fernandez-Munos et al., 1971). One mol-
ecule of lincomycin binds per ribosome, preventing peptide bond formation
by blocking binding of the 3’ terminal end of the substrate to the A-site on
the ribosome (Pestka, 1970). The binding constant for lincomycin binding to
E. coli ribosomes was 3.4 X lop5 M as measured by direct binding of radiolab-
elled drug by equilibrium dialysis. Under conditions of the puromycin fragment
assay with 33 per cent ethanol the binding was increased to give a figure of
1.7 X 10e6~, but this is clearly not a physiological situation.

R,-c 4,
I

6 &
Lincomycin OH H
Clindamycin H Cl

Clindamycin suffers from the drawback that diarrhoea frequently

accompanies treatment, which can occasionally degenerate into pseudomem-
braneous colitis, a sometimes lethal disease which is caused by Clostridium
difficile and characterized by diarrhoea, fever and abdominal pain (Dhawan
and Thadepalli, 1982).

Questions

1. How does bacterial protein synthesis differ from mammalian?

2. What mechanisms of resistance occur against the aminoglycosides? Which
is the most important clinically?
3. Which antibacterials block peptide bond formation? How do they do this?
4. Which antibacterial can be used for the treatment of infections caused by
anaerobic bacteria? Why is this possible?
 Protein biosynthesis 63

References
AMA Drug Evaluation, 1983, Prepared by A.M.A. Drug Division 5th edition
(Philadelphia, U.S.A.: W.B. Saunders Co), p. 1255.
Amyes, S.G.B. and Gemmell, C.G., 1992, J. Med. Microbial, 36, 4-29.
Bailey, R.R., 1981, Dncgs, 22, 321-7.
Bernabeu, C., Vazquez, D. and Ballesta, J.P.G., 1977, Eur. J. Biocbem., 79, 469-77.
Birge, E.A. and Kurland, C.G., 1969, Science, 166, 1282-4.
Blagg, N.A. and Gleckman, R.A., 1981, Postgrad. Med., 69, 59-67.
B&son-Noel, A., Trieu-Coot, P. and Courvalin, P., 1988, J. Antimicrob. Cbemotber.,
22, Suppl. B, 13-23.
Bryan, L.E., 1988, J. Antimicrob. Cbemotber., 22, Suppl. A, 1-15.
Buckel, P., Buchberger, A., Bock, A. and Wittman, H.G., 1977, Molec. Gen. Genetics,
158, 47-54.
Chopra, I. and Howe, T.G.B., 1978, Microbial. Rev., 42, 707-24.
Cuchural, J.G. et al., 1988, Antimicrob. Ag. Cbemotber., 32, 717-22.
CundLiBe, E., 1987, Biocbimie, 69, 863-9.
Davies, B.D., 1988,J. Antimicrob. Cbemotber., 22, l-3.
Davies, J. and Courvalin, P., 1977, Am. J. Med., 62, 868-72.
DeTorres, O.H., 1981, Clin. Tber., 3, 399-412.
Dhawan, V.K. and Thadepalli, H., 1982, Rev. Infect. Dis., 4, 1133-53.
Drew, R.H. and Gallis, H.A., 1992, Pbarmacotber., 12, 161-73.
Dubreil, L., Devos, J., Neut, C. and Romond, C., 1984, Antimicrob. Ag. Cbemotber.,
25,764-6.
Fass, RJ., Scholand, J.F., Hodges, G.R. and Saslaw, S., 1973, Ann. Intern. Med., 78,
853-9.
Fernandez-Munos, R., Memo, R.E., Torres-Pinedo, R. and Vazquez, D., 1971, Eur. J
Biocbem., 23, 185-93.
Goldman, R.A., Cooperman, B.S., Strycharz, W.A., Williams, B.A. and T&ton, T.R., 1980,
FEBS Letters, 118, 113-18.
Kaltschmidt, E. and Wittman, H.G., 1970, Proc. Nat. Acad. Sci. (U.S.A.), 67, 1276-82.
Kozak, M., 1983, Microbial. Rev., 47, l-45.
Kucers, A., 1982, Lancet, ii, 425-8.
Lando, D., Cousin, M.A., Ojasoo, T. and Raynaud, J.P., 1976, Eur. J Biocbem., 66,
597-606.
LeGoffic, F., Capmau, M.L., Tangy, F. and Baillarge, M., 1979, Eur. J. Biocbem., 102,
73-81.
McMurray, L.M., Cullinane, J.C. and Levy, S.B., 1982, Antimicrob. Ag. Cbemotber., 22,
791-9.
Maitra, U., Stringer, E.A. and Chaudhuri, A., 1982, Annu. Rev. Biocbem., 51,869-900.
Merrick, W.C., 1992, Microbial. Rev., 56, 291-315.
Modolell, J., Cabrer, B., Parmeggiani, A. and Vazquez, D., 1971, Proc. Natl. Acad. SCE’
(USA.), 68, 1796-800.
Nierhaus, D. and Nierhaus, K.H., 1973, Proc. Natl. Acad. Sci. (U.S.A.), 70, 2224-8.
Pestka, S., 1970, Arch. Biocbem. Biopbys., 136, 80-S.
Pongs, 0. and Messer, W., 1976,J. Mol. Biol., 101, 171-84.
Smythies, J.R., Benington, F. and Morin, R.D., 1972, Experientia, 28, 1253-4.
Speer, B.S., Shoemaker, N.B. and Salyers, A.A., 1992, Clin. Microbial. Rev., 5, 387-99.
Tai, P.-C., Wallace, B.J. and Davis, B.D., 1978, Proc. Natl. Acad. Sci. (U.S.A.), 75, 275-9.
T&ton, T.R., 1977, Biochemistry, 16, 4133-8.
Turk, D.C., 1977,J. Med. Microbial., 10, 127-31.
Wallace, BJ., Tai, P.-C., Herzog, E.L. and Davis, B.D., 1973, Proc. Natl. Acad. Sci.
(U.S.A.), 70, 1234-7.
Yunis, A.A., Miller, A.M., Salem, Z. and Arimura, G.K., 1980, Clin. Toxicol., 17, 359-73.
 Chapter 4

Carbohydrate metabolism

4. I Introduction

Metronidazole is a drug which is used widely for the treatment of infections

caused by anaerobic organisms, either (a) protozoa such as i?ric&omonas
vaginal&, which provokes vaginal discharges, and Entamoeba histolytica,
causing amoebic dysentery, or (b) bacteria, e.g. in abscesses or tissue necroses
caused by Bacteroides fragilis and antibiotic-induced pseudomembraneous
colitis (Clostridium difficile). For a general review on metronidazole, see
Muller (1981). Although there is agreement that the drug has to be reduced
to show activity (metronidazole is in fact a pro-drug) the ultimate target may
be DNA in bacteria and, possibly, carbohydrate metabolism in protozoa. In
view of this uncertainty, therefore, it has been placed in a chapter on its own.
Carbohydrate metabolism in anaerobes appears to be similar to that in
aerobes - at least as far as the conversion of glucose to pyruvate is concerned.
Although anaerobes subsequently transform pyruvate to acetyl-coenzyme A,
they use a very different type of enzyme from the mammalian or bacterial
pyruvate dehydrogenase. Pyruvate decarboxylation in anaerobes is linked to
the reduction of hydrogen ion by ferredoxin (Fd), a protein which contains
four iron-sulphur centres and is a very powerful reducing agent. This reaction
is catalysed by hydrogenase, to give hydrogen gas, and the reduction of mol-
ecular oxygen to give hydrogen peroxide. Pyruvate dehydrogenase
(pyruvate:ferredoxin oxidoreductase) appears to have been a very early pro-
duct of evolution as it is also present in Archaebacteria which are thought to
have existed initially in a highly reduced environment (Kerscher and Oes-
terhelt, 1982). The overall reaction is shown below:

Fd oxidized Fd reduced

H202
\ \
02

65
66 Molecular mechanisms of drug action

In trichomonads, acetyl-coenzyme A is converted to acetate directly by an

enzyme which simultaneously transfers the coenzyme A moiety to succinate.
Succinyl-coenzyme A is then hydrolysed to succinate and free coenzyme A
coupled with phosphorylation of ADP to ATP (this is known as substrate level
phosphorylation in contrast to oxidative phosphorylation in which an electron
transport chain is required). In prokaryotes, acetate is still the end-product
but an intervening step of acetylphosphate formation is included (sometimes
referred to as the phosphoroclastic reaction). In trichomonads, but not in bac-
teria, pyruvate decarboxylase and the other enzymes involved in the later
stages of carbohydrate metabolism are enclosed in a cytoplasmic organelle
called the hydrogenosome (anaerobic protozoa do not have mitochondria and
peroxisomes) (Lindmark et al., 1975).

4.2 Metronidazole action - anaerobic protozoa1 and

bacterial infections
Metronidazole is the original and most widely used of a family of alkyl-substi-
tuted 5nitroimidazoles. It is inactive against organisms that grow aerobically
and have reducing activity equivalent to a redox potential of -350 mV
whereas anaerobic organisms (whether Gram-positive or -negative) possess a
reducing system sufliciently powerful to reduce the nitro group
(E, = -470 mV; O’Brien and Morris, 1972). These products are likely to be
the nitro group radical-ion for one-electron reduction, nitroso for two-electron
reduction, reduction to a hydroxylamino group for four while a Gelectron
reduction would give rise to products at the oxidation level of the amino
group. Whether all or only some of these products occur in the intact cell is
not established but the instability of some of them makes their detection
extremely difficult. The chemical instability of one product at least allows it
to react with DNA, causing strand breakage and cell death (reviewed in
Edwards, 1993). Intracellular reduction of the drug causes more to be trans-
ported into the cell to maintain equilibrium. The net effect is that accumulation
of the drug occurs in sensitive cells (Muller, 1981).

OzN ,CH2CH20H

Metronidazole
t’ N’,,,

The powerful reducing agent required to carry out this reduction is probably
the iron-sulphur protein ferredoxin (E, = -640 mV; O’Brien and Morris,
1972) acting as a cofactor to the enzyme pyruvate:ferredoxin oxidoreductase.
This enzyme catalyses the decarboxylation of pyruvate to acetylphosphate and
carbon dioxide. There is a correlation in anaerobes between enzyme level,
degree of susceptibility of the organism and rate of drug uptake (Narikawa,
1986). Ferredoxins contain equal numbers of iron and sulphur atoms at the
 Carbohydrate metabolism 67

active site (i.e. 4 Fe:4 S in mammalian systems), capable of acting as redox

agents in a respiratory chain because of the ability of the iron to change val-
ency. They are characterized by the presence of acid-labile sulphur atoms -
so-called because treatment with acid liberates H,S. Ferredoxin-linked
reduction of metronidazole has been demonstrated in cell-free extracts of Clos-
tridium pasteurianum, and it appears that the drug acts as an electron sink,
siphoning reducing equivalents away from other cellular reductive processes.
Methylviologen is less effective, while the pyridine nucleotide coenzymes are
totally ineffective. In principle, aromatic nitro groups can be reduced in
four stages:
Ze 2e
R-N02- le R - No), - -R-NO-
le R-NHOH- R - NH,

Although it is difficult to-detect the intermediates in the metronidazole path-

way because of instability, the antimicrobial action may involve a four-election
reduction taking place to the oxidation level of a hydroxylamino derivative
(Lockerby et al., 1984; Kedderis et al., 1988). Alternatively, the radical anion
may take up a proton to give R-NO,H, which removes electrons from
(oxidizes) DNA causing strand breaks. In the process, the active agent is
reduced to the nitroso form R-No (Edwards, 1993). In essence this is oxidation
by the hydroxide radical OH.
Experimentally, the first effects detectable after drug treatment are the cess-
ation of first hydrogen emission and then carbon dioxide, as seen with C.
welchii for example (Edwards et al., 1973). The radical ion produced by
metronidazole acting as an electron acceptor in place of hydrogen ion has
been detected by electron spin resonance (e.s.r.) measurements both in hydro-
genosomes from 2: vaginalis (Lloyd and Kristensen, 1985) and in the intact
organism (Chapman et al., 1985). The radical ion is then reduced further to
the active species that attacks crucial cell constituents. If the reduced
product(s) were chemically stable it is likely that the effect of the drug would
be reversible, in the sense that when all the drug had been reduced, hydrogen
ion would continue as the electron acceptor and hydrogen evolution would
restart.
In aerobic situations, oxygen can accept the electron from the radical ion
because the redox potential of the 0,/O; is greater than that of RNOJRNO;.
Superoxide ion and the uncharged drug molecule are produced - a futile meta-
bolic cycle. Such a cycle has been observed in rat liver microsomes in the
presence of NADPH (Perez-Reyes et al., 1979), and in hydrogenosomes from
Tritricbomonas foetus, the cattle pathogen, in the presence of oxygen
(Moreno et al., 1984). In the latter case, pyruvate and addition of coenzyme
A stimulated the production of the radical ion, but NADPH was totally ineffec-
tive, as one would expect. Oxygen thereby acts to detoxify the drug and so
protects aerobic organisms.
The precise nature of the protozoa1 target for the reduced active species is
not known. Chapman et al. (1985) showed that progressive damage occurs
68 Molecular mechanisms of drug action

to the radical generating system as a consequence of metronidazole reduction

and considered that the radical ion is likely to be reactive enough not to travel
very far in the whole cell. Furthermore, Lloyd and Kristensen (1985) proposed
that a metabolite of the drug may irreversibly interfere with part of the system
for hydrogen production, since the level of gas release falls progressively with
time in the presence of metronidazole. The likelihood, therefore, of a reactive
metabolite crossing both the hydrogenosomal and nuclear membranes to
attack the DNA is not very high.
In bacteria, where there are neither nuclear nor hydrogenosomal mem-
branes, DNA synthesis has been proposed as the target for the active species.
Although protein and RNA synthesis in B. fragilis are unaffected by treatment
with 10 @ml metronidazole, DNA synthesis, as measured by [3H]thymidine
incorporation, halted 20 minutes after treatment. The DNA remained structur-
ally intact, as shown by subsequent extraction and gel electrophoresis, and it
is unlikely that the level of supercoiling was affected, thus ruling out an effect
on DNA gyrase (Chapter 2). The precise target may be in a multi-enzyme repli-
cation step since DNA polymerase was also not inhibited by drug in sonicated
cells and DNA was clearly capable of serving as a template for RNA synthesis
(Sigeti et al., 1983). Earlier studies had shown that if metronidazole is reduced
in the presence of DNA, the reduced products will bind covalently to the
nucleic acid but without interfering with the sedimentation velocity or melting
temperature (LaRusso et al., 1978).
The development of resistance to 5-nitromidazoles does not occur fre-
quently. Two possibilities that, in theory, can occur with an anti-infective agent
are: (a) reduced transport into the cell, which would be very unlikely with
such a small molecule as metronidazole, or (b) metabolism by enzymes which
have not yet been discovered. Resistance modes normally give a clue as to the
mode of action of a given drug, but in this case there is little evidence. Oxygen
detoxification of metronidazole has been suggested as a possible method of
inducing resistance (Morent et al., 1984), since oxygen can react with the
radical ion to reform uncharged drug. Indeed, if sensitivity tests to 5-nitroimida-
zoles are carried out under aerobic conditions, the microorganisms often
appear more or less resistant. In studies with 7: vaginalis isolates that mark-
edly differed in their susceptibility to the drug in vivo, the availability of oxy-
gen was shown to be an important factor in that NADH oxidase, which oper-
ates to keep oxygen tensions low, was lowered in one strain (Clackson and
Coombs, 1982).
Mutations in pyruvate:ferredoxin oxidoreductase could be a possible mode
of resistance, as shown for Clostridium perfringens by mutagenesis with N-
methyl-N’-nitro-N-nitrosoguanidine (Sindar et al., 1982). In these mutants,
pyruvate and lactate levels rose whereas acetate and carbon dioxide levels fall
during growth, and so the enzyme activity was not totally abolished. Muller
and Gorrell(1983) however, discounted the importance of this enzyme since
activity can be quite low even in susceptible strains. Furthermore, a number
of clinically resistant isolates defied identification of their locus of resistance
 Carbohydrate metabolism 69

(Muller and Gorrell, 1983). The overall difficulty in assigning the resistance to
a given site is that the organism, by virtue of its ferredoxin or other powerful
reducing agent, is almost certain to be able to reduce metronidazole to chemi-
cally unstable intermediate(s). The nature of the ultimate target could vary
from species to species depending on the nature of the nearest sensitive mol-
ecule.

4.3 Medical use of metronidazole

Metronidazole was originally introduced on to the market in 1959 for the treat-
ment of protozoa1 infections. The development of the drug for the treatment
of infections caused by anaerobic bacteria coincided with, and played a major
part in, the recognition of the importance of these organisms in infectious
disease (Finegold, 1981; Bartlett, 1982). Metronidazole is regarded as the drug
of choice in the treatment of these conditions, particularly those likely to be
caused by B. fragilis. In one study on the suceptibility of anaerobic bacteria
isolated from patients in several French hospitals, metronidazole and cefoxitin
were found to be equally effective while clindamycin was less so. Metronida-
zole was considered to be the more suitable drug, partly because fewer resist-
ant strains have been recorded and, in addition, a smaller dose is normally
required than for cefoxitin (Dubreil et al., 1984).
Metronidazole is also being used in conjunction with bismuth salts, hista-
mine H, receptor blockers and other antibiotics, such as tetracycline, for the
treatment of Helicobacter pylori associated with duodenal ulcers (Rauws,
1992).
The role of H. pylori in ulcer formation is not fully understood. Almost all
patients with ulcers proven by endoscopy are colonized with the bacterium.
In contrast, some 20 per cent of asymptomatic volunteers are also colonized,
which implies that the aetiology of ulcers is multi-factorial and, indeed, acid
secretion and pepsin are also strongly implicated.
Ulcers take longer to heal with an H, blocker such as ranitidine, alone than
with a combination of ranitidine, bismuth and an antibiotic. Ulcers are also
more likely to relapse if H. pylori is not eradicated. H. pylori possesses urease
which converts urea to ammonia, thus raising the pH in the immediate vicinity
and, in addition, various hydrolytic enzymes such as lipase, phospholipase and
proteases which could cause damage to the mucosal lining (Rauws, 1992).
H. pylori can grow if the oxygen content of the atmosphere is low, i.e. 5
per cent, and is known as microaerophilic. The bacterium can survive, but
probably not grow, in anaerobic conditions. Resistance to metronidazole, like
the early reports of B. fragilis resistance, is associated with microaerophilic
conditions where presumably the detoxicant effects of oxygen come into play.
If H. pylon’ is incubated under anaerobic conditions before testing, it is suscep-
tible to the drug (Cederbrand et al., 1992).
Metronidazole is particularly effective in z&o, being readily absorbed from
70 Molecular mechanisms of drug action

the gastrointestinal tract and widely disseminated in body fluids and tissues.
Furthermore, in the considerable period of the use of the drug, very few genu-
inely resistant strains have been reported. Another potential problem with the
use of the drug has been the possibility of mutagenicity as indicated by the
Ames test (Ames et al., 1975). This test measures the ability of a compound
to reverse a mutation in Salmonella typhimuria that lacks an enzyme in the
histidine biosynthetic pathway (phosphoribosyl adenosine triphosphate
synthetase). The bacterium is unable to grow in histidine-deficient medium
unless a reverse mutation occurs. Since the microsomal mixed-function oxi-
dase system which involves cytochrome P-450 is often responsible for the
metabolism of inactive precursors to rank carcinogens, liver microsomes and
a NADPH generating system are also included. Metronidazole is positive in
this test.
It should be noted that the Ames test is not the only test used for potential
carcinogens and that some of its results are controversial. In practice over 30
years, the use of metronidazole apparently has not given rise to an abnormal
incidence of tumours (Friedman and Selby, 1989). The drug is also very cheap,
at least in oral tablet form.

4.4 a-Glucosidase inhibitors - acarbose as antidiabetic

Various approaches have been made to the treatment of non-insulin-dependent
diabetes mellitus; diet to control carbohydrate uptake, the sulphonylureas
which stimulate insulin secretion and sensitize the target tissues to insulin
action, and the injection of insulin itself. Recently, a different approach has
become possible by the development of a-glucosidase inhibitors, which delay
the hydrolysis of complex polysaccharides to glucose, slow glucose absorption
and reduce its plasma levels (Lebovitz, 1992).
The carbohydrate content of the diet in Western countries is composed of
the polysaccharide starch (- 60 per cent) and the disaccharide sucrose (= 30
per cent). Salivary and pancreatic cy-amylaseshydrolyse starch to disaccharides
(maltose), trisaccharides (maltotriose) and polysaccharides (dextrins). The
brush border of the intestinal mucosa contains the cY-glucosidases. The most
important enzymes are maltase hydrolyzing maltose, sucrase hydrolyzing
sucrose and isomaltase that catalyses the breakdown of maltotriose. In
addition, glucoamylase removes one glucose residue at a time from the non-
reducing end of a dextrin. With the exception of sucrase, which produces
fructose as well as glucose, the product of all these enzymes is glucose (see
WiRiam-Olsson, 1985; Clissold and Edwards, 1992).
Acarbose is one of the leading a-glucosidase inhibitors and is a modified
tetrasaccharide with two o-o-glucose residues linked to acarviosine - a 46
dideoxyhexose attached to cyclohexitol through a secondary amino group.
The Ki for sucrase is 1.5 X 10p6~, the K, for the substrate sucrose is
1.9 X 10e3 M so the drug has a 1500-fold greater affinity for the enzyme than
 Carbohydrate metabolism 71

does the substrate. The secondary amino group of acarbose, instead of the
usual glycosidic oxygen bond of the substrate, accepts a proton from a car-
boxylate anion on the enzyme thus preventing the first step in hydrolysis
(William-Olson, 1985).
Acarbose has lowered glucose levels in diabetics after a meal, but not neces-
sarily insulin levels, irrespective of whether the patients were receiving sul-
phonylureas (see Clissold and Edwards, 1992). In patients receiving insulin,
daily requirements were sometimes reduced. Acarbose may also be useful in
preventing the diabetic complications that result from higher plasma glucose,
but the evidence requires long-term trials which have not yet been carried out
(Zimmerman, 1992).

Acarbose

4.5 Sialidase inhibitor - 4-guanidino-Neu5Ac2en for

in$uenza

In earlier years it was a dream of the medicinal scientist to design a drug on

the basis of the X-ray structure and mechanism of an enzyme. This dream may
be realized in a potential drug for influenza through inhibition of the viral
enzyme sialidase (neuraminidase, N-acetylneuraminic acid hydrolase).
Sialidase has been known for a long time and used to be sold, moderately
purified, as ‘receptor-destroying enzyme’. It cleaves cY-keto-linked sialic acids
from terminal positions on oligosaccharide side-chains. Haemagglutinin, the
other protein on the virus surface binds to sialic acid residues on target and
other cells. Sialidase cuts off these false receptors and allows the virus to reach
its target cell. In addition, after the virus has replicated inside the cell, siahdase
cleaves the sialic acid residues from the debris of the cell and allows the virus
to escape.
The enzyme reaction passes through a sialosyl cation intermediate with a
proton donated by a water molecule. The sugar transition state contains a
double bond in which the ring is flattened (Chung et al., 1992). An unsaturated
compound NeuSAc2en that mimics this state was a moderate inhibitor of the
enzyme (Ki near 10m6~). The inlhtenza virus, however, unlike the bacterial
72 Molecular mechanisms of drug action

and mammalian analogues, has a pocket near the oxygen at the 4position of
the sugar ring. A glutamate residue is sited in this pocket and the suggestion
was made that amino or guanidino side-chains would form salt-bridges with
this residue producing very powerful inhibitors. This turned out to be the case
with the amino compound Ki at 5 X 10m8~ and the guanidino at 2 X lo-‘OM
respectively (von Itzstein et al., 1993). The latter may be a slow-binding inhibi-
tor.
Both compounds block the replication of both types of influenza virus, A
and B, in tissue culture and are active in uivo intranasally. Although these
compounds are good leads there is much that can go wrong in the develop-
ment process. Nevertheless, it is hoped that a compound will be available for
use before the next worldwide influenza epidemic occurs.

-R NAME

OH Neu-S AC 2en

4-Amino-Neu-S AC 2en
NH*
//NH
NHC 4-Guanidino-Neu-S AC 2en
‘NH,

Questions

1. How many electrons are required to reduce metronidazole to (a) the radical
ion, (b) the hydroxylamino derivative?
2. Why is metronidazole not active against B. fragilis grown in the presence
of oxygen?
3. What part do anaerobic organisms play in (a) bacterial and (b) protozoa1
infection?
4. What enzymes hydrolyze oligosaccharides in the intestine?
5. How does acarbose antagonize the symptoms of diabetes?

References
Ames, B.N., McCann, J. and Yamasaki, E., 1975, Mutut. Res., 31, 347-64.
Bartlett, J.G., 1982, Lancet, ii, 478-81.
Cederbrand, G., Kahlmeter, G. and Ljungh, A., 1992,J. Antimicrob. Cbemotber., 29,
115-20.
Chapman, A., Cammack, R., Linstead, D. and Lloyd, D., 1985,J. Gen. MicrobioZ., 131,
2141-4.
 Carbohydrate metabolism 73

Chung, KJ., Pegg, MS., Taylor, N.R. and von Itzstein, M., 1992, Eur.J Biochem., 207,
333-43.
Clackson, T.E. and Coombs, G.G., 1982, J. Protozool., 29, 636.
Clissold, S.P. and Edwards, C., 1988, Drugs, 35, 214-43.
Dubreil, L., Devos, J., Neut, C. and Romond, C., 1984, Antimicrob. Ag. Chemotber.,
25, 764-6.
Edwards, D.I., 1993,J. Antimicrob. Chemotber., 31, 9-20.
Edwards, D.I., Dye, M. and Carne, H., 1973,J. Gen. Microbial., 76, 135-45.
Finegold, SM., 1981, Stand. J. Infect. Dis., Suppl. 26, 9- 13.
Friedman, G.D. and Selby, J.V., 1989,J. Am. Med. Assoc., 261, 866.
Kedderis, G.L., Argenbright, L.S. and Miwa, G.I., 1988, Arch Biocbem. Biopbys., 262,
40-S.
Kerscher, L. and Oesterhelt, D., 1982, Trends in Biochem. Sci., 3, 371-4.
LaRusso, N.F., Tomasz, M., Kaplan, D. and Muller, M., 1978, Antimicrob. Ag. Cbem-
other., 13, 19-24.
Lebovitz, H.E., 1992, Drugs, 44, Suppl. 3, 21-8.
Lindmark, D.G., MuIIer, M. and Shio, H., 1975,J. ParasitoL, 61, 552-4.
Lindmark, D.G. and Muller, M., 1976, Antimicrob. Ag. Cbemotber., 10, 476-82.
Lloyd, D. and Kristensen, B., 1985,J. Gen. Microbial., 131, 849-53.
Lockerby, D.L., Rabin, H.R., Bryan, L.E. and Iaishley, F.J., 1984, Antimicrob. Ag. Cbem-
other., 26, 665-9.
Moreno, S.N.J., Mason, R.P. and Decampo, R., 1984,J Biol. Cbem., 259, 8252-9.
Muller, M., 1981, Scand.J Znfect. Dis., Suppl. 26, 31-41.
MuIIer, M. and GorreII, T.E., 1983, Antimicrob. Ag. Cbemotber., 24, 667-73.
Narikawa, S., 1986,J. Antimicrob. Cbemotber., 18, 565-74.
O’Brien, R.W. and Morris, J.G., 1972, Arch. Mikrobiol., 84, 225-33.
Perez-Reyes, E., Kalyanaraman, B. and Mason, R.P., 1979, Molec. Pbarmacol., 17,
239-44.
Rauws, E.A.J., 1992, Drugs, 44, 921-77.
Sigeti, J.S., Guiney, D.G. and Davis, C.E., 1983,J. Infect. Dis., 148, 1083-9.
Sindar, P., Britz, M.L. and Wilkinson, R.G., 1982, J. Med. Microbial., 15, 503-9.
von Itzstein et al., 1993, Nature, 363, 418-23.
William-Olsson, T., 1985, Acta Med. Scand., 706, Suppl. l-39.
Zimmerman, B.R., 1992, Drags, 44, Suppl. 3, 54-9.
 Chapter 5

Cell wall biosynthesis

5.1 Introduction

Fungi and bacteria both require a structure external to their cell membrane
to maintain the integrity of the organism - preventing a high internal osmotic
pressure from rupturing the cell in a hypotonic external environment. Drugs
that interfere with the biosynthesis of the cell wall might therefore be
expected to be of value in combating infection by these organisms, particularly
since mammalian cells do not require such a structure. The plactams
(penicillins and cephalosporins) and vancomycin fall into this category. PLac-
tams, indeed, have been on the market for many decades; the frost observation
that led to the identification of the precursor penicillin was made in 1929,
although it took nearly 20 years for the drug to reach the market. In addition,
cilofungin is known to interfere with the biosynthesis of &lucan in the fungal
cell wall and is effective in the treatment of human mycoses.
Peptidoglycan is a polymeric molecule which consists of parallel polysacch-
aride chains covalently joined by peptide cross-links. The whole structure can
be regarded as a single molecule encompassing the entire cell in a sort of rigid
bag. The glycan chain is a polymer consisting of alternating pyranose residues
of N-acetylglucosamine and N-acetylmuramic acid connected in a @(l-4) link-
age (similar to the fungal cell wall polysaccharide chitin) with o-lactyl groups
attached to alternate sugar residues. Attached via amide links through the car-
boxy1 groups of the o-lactate are tetrapeptide chains (Fig. 5.1). The sequence
of this peptide is L-Ala-y-o-Glu-L-Lys (or mesodiaminopimelic acid)-D-Ala (meso-

Table 5.1 Drugs discussed in Chapter 5

Section Target Drug Use

5.2 Peptidoglycan synthesis PLactams Bacterial infections

5.3 Peptidoglycan synthesis Vancomycin Bacterial infections
5.4 Mycolic acid synthesis Isoniazid Tuberculosis
5.5 @-Glucan synthesis Cilofungin Candida infections

75
76 Molecular mechanisms of drug action

CH,OH CH,OH

\OQ:QIL3 k!-Acetylmuramic acid (NAM)

3 0

H,C-AH-C=0

ALAcetylglucosamine I
(NAG) L-Ala
I
o-Glu

L-Lys

I
o-Ala
The y-carboxyl group of ,,-Glu is used for peptide bond formation

Figure 5.1 Peptidoglycan repeating unit (S. aureus)

diaminopimelic acid is found in Gram-negative organisms, lysine in the Gram-

positive Staphylococcus aureus).
The bifunctional amino acid in the chain (e.g. lysine) acts as a cross-link to
accept the terminal carboxyl group of a o-Ala residue from an adjacent chain.
This linkage may be either direct as in Escherichia coli, or through a short
connecting peptide such as the pentaglycine found in Staphylococcus aureus
(Fig. 5.2). This structure is resistant to peptidases which cannot attack peptides
containing n-amino acids. The enzyme lysozyme, however (found in tears and
egg-white) is able to split the polysaccharide back-bone at the /3-(l-4) link, to
yield disaccharides of the basic repeating unit to which peptide chains are still
attached. This breaks the strength of the peptidoglycan so that it can no longer
contain the osmotic pressure inside the cell and so the membrane ruptures
with loss of cell contents.
The biosynthesis of the peptidoglycan involves the action of about 30
enzymes, starting with the formation of active precurors by soluble enzymes
in the cytoplasm. The first major precursor or monomer is the pentapeptide
attached to uridine diphosphate: UDP-MurNAc-L-Ala-y-o-Glu-L-Lys-D-Ala-D-Ala
(MurNAc is shorthand for N-acetylmuramic acid). Synthesis of the D-Ala-D-Ala
fragment, which is added last, occurs by conversion of L-alanine to n-alanine
(epimerization) followed by condensation of two u-alanine molecules. The
monomer is then transferred to a C,,-terpene alcohol with a terminal phos-
phate group, known as bactoprenol, which acts as a membrane-bound carrier
instead of UDP, and is linked to a phospholipid in the cell membrane via
a pyrophosphate bridge. Condensation of UDP-N-acetylmuramyl-pentapeptide
with UDP-N-acetylglucosamine is effected. In the case of the Gram-positive
organism Staphylococcus aureus, five glycine residues are transferred from
 Cell wall biosynthesis 77

NAM
/NAG
NAG’ 1

KAG. N-acctyl~luco~am~ne: S4.M. N-acetylmuramlc acid.

The rcarboxyl group of D-Glu is used for peptide bond formation.
Figure 5.2 Mode of peptidoglycan cross-linking (S. aweus)

glycinyl-tRNA to form a chain attached at one end to the e-amino group of the
L-lysine residue in the side-chain. The unit is then inserted into the cell wall
and simultaneously the bactoprenol is released as a pyrophosphate. The pyro-
phosphate then has to be converted to a monophosphate before being used
again.
The third and final stage involves the formation of the cross-link. The ter-
minal glycine is linked to the penultimate D-alanine residue of a neighbouring
chain (a reaction catalysed by a transpeptidase) and at the same time releasing
the terminal D-alanine. It is this last step in peptidoglycan biosynthesis that
was originally thought to be the only target for the plactam antibiotics, prob-
ably because of their structural resemblance to the D-Ala-D-Ala dipeptide
(Tipper and Strominger, 1965). The overall process is reviewed in more detail
in Gale et al. (1981).

5.2 ELactams

PLactams comprise a number of different classes of antibiotic whose salient

feature is a four-membered cyclic amide ring. This may be fused to a six-mem-
bered ring with or without sulphur, as in the cephalosporins and carbace-
78 Molecular mechanisms of drug action

phems, a five-membered ring in the penicillins and penems or no ring at all

in the monobactams. The activity of penicillin was first observed in 1929 by
Alexander Fleming. The antibiotic is elaborated by a fungus, Penicillium nota-
turn, and the extraction and identification of the antibiotic from broth cultures
was carried out by Chain, Florey and Abraham starting in 1939. The drug
became available for human use during the Second World War. A multitude
of semi-synthetic penicillins have been produced with a variety of substituents
on the Gamin0 group.
Cephalosporium acremonium was the first source of the cephalosporin
family of antibiotics, and was isolated in 1948 from a sewer outlet off the
Sardinian coast. One of the earliest cephalosporins identified was cephalospo-
rin C which can be hydrolysed to yield 7-aminocephalosporanic acid; this has
allowed the preparation of a vast range of semi-synthetic agents. The cephamy-
tins are a closely related family of antibacterials which are derived from Strep-
tomyces species, in particular Streptomyces lactamdurans. The latter prod-
uces cephamycin C, which has a methoxyl group at position 7 of the plactam
ring of the 7-aminocephalosporanic acid nucleus. Recently, the monobactam
series of structures has been developed of which one of the main compounds
of interest is aztreonam (Neu, 1986).
Originally the mode of action of the penicillins was attributed to inhibition
of the transpeptidase that catalyses the fmal cross-linking of the peptide side-
chain of the nascent peptidoglycan (Tipper and Strominger, 1965). In Staphy-
Zococcus aureus treated with penicillin these authors detected large amounts
of non-cross-linked peptidoglycan units, confirmed in later studies by the
detection of soluble linear peptidoglycan in penicillin-treated cells (Waxman
et al., 1980).
More recent studies have shown, however, that the situation is not so clear-
cut (reviewed in Spratt, 1980, 1983). In addition to the transpeptidase, both
a o-alanine carboxypeptidase and a peptidoglycan endopeptidase have been
found that are sensitive to penicillin, and it has become clear that bacteria
possess a number of enzymes that are plactam sensitive. It appears, however,
that the carboxypeptidase reaction at least is not crucial for the organism, as
Gaminopenicillanic acid inactivates at least 95 per cent of the enzyme at non-
lethal concentrations in Bacillus subtilis, whereas cephalothin kills the bac-
terium at concentrations that do not affect the enzyme (Blumberg and Strom-
inger, 1971).

5.2.1 Penicillin-binding proteins

In contrast, more recent approaches have been to study the pattern of binding
when radiolabelled benzylpenicillin is incubated with bacterial cells, the mem-
brane portion is isolated and solubilized, and the proteins are separated by gel
electrophoresis (Curtis et al., 1979a; Spratt and Pardee, 1975). A number of
proteins to which the j?-lactams will bind have been identified by this process
and are known as penicillin-binding proteins (PBP) - numbered l-6 in decreas-
 Cell wall biosynthesis 79

Penicillins

Benzylpenicillin : R =

Amoxycillin : R = HO

Methicillin : R =

Ampicillin : R =

Cephalosporins

R2
k02H

f32 R3

Cefoxitin -CH20CONH2 -0CH2

NHS\
Cephalosporin C CH(CH2)3- -CH20COCH2 -H
H02C’

Cefuroxime -CH20CONH2 -H

+-
Ceftazidime -CH2-N, , -H
3
80 Molecular mechanisms of drug action

Aztreonam

ing order of apparent molecular weight. The enzymic reactions associated with
these proteins are not fully understood, although they must carry out transpep-
tidation and endopeptidase reactions.
Antibiotics that bind to these proteins produce a variety of effects: in E.
coli, blockade of PBPla and lb leads to cell lysis although it appears that the
loss of either one alone is not lethal, presumably because the other protein
can compensate; PBP2 inhibition leads to cessation of growth and change of
shape from rod to sphere and eventual lysis; PBP3 appears to be responsible
for the formation of the cross-wall to divide the two daughter cells at cell
division (septation), and blockade of this protein results in the formation of
long filaments and eventual cell death. Binding to the three other penicillin-
binding proteins does not lead to any obvious growth defects (reviewed in
Spratt, 1983).
Other distinctions that correlate with the separation of the penicillin-binding
proteins into two groups depending on their importance to the bacterium are
that PBPs l-3 have molecular weights between 60 and 92 kDa in E. coli, for
example, and exhibit transglycosylase and transpeptidase activity, while PBPs
4-6 have molecular weights 40-49 kDa and show o-alanine carboxypeptidase
activity. This work has been carried out both with agents that bind specifically
to one or other penicillin-binding protein, and also by the use of mutants that
lack, or produce thermolabile varieties of, these proteins (Spratt, 1980, 1983).
It is clearly of importance to correlate the binding of an antibiotic to a given
penicillin-binding protein with the effect on the whole cell. Cefoxitin, for
example, exhibits an I,,, of O-1 pg/ml for PBPla and 3.9 pg/ml for PBPlb from
E. coli K12, values which are close to the minimum inhibitory concentration
(MIC, the minimum concentration that totally inhibits growth; Chapter 1)
(Curtis et al., 1979b), and so it is likely that the penicillin-binding proteins are
the targets, particularly as the drug causes the bacteria to lyse. Mecillinam is
an example of Plactam that binds particularly effectively to PBP2
(15” < 0.25 pg/ml), kills E. coli with an MIC of 0.05 pg/ml, producing in the
process spherical forms of the organism as we would expect from a binding
of PBP2 (see above). In sharp contrast, ceftazidime has a potent affinity to
PBP3 in both E. coli and Pseudomonas aeruginosa, with an 15, of 0.06 I.Lg/ml
for PBP3 of the former bacterium and MIC of 0.2 pg/rnl (Hayes and Orr, 1983).
As a consequence, filamentation of the organism results.
Resistance to ~lactams among Gram-negative bacteria may result from a
reduction in drug afftnity by the penicillin-binding protein (Malouin and Bryan,
 Cell wall biosynthesis 81

Mecillinam

1985). In Gram-positive bacteria, this can also happen and, in addition, the
acquisition of a differen: PBP.
Staphylococcus am-em normally has five PBPs (1, 2, 3, 3’, 4) of which three
(I, 2, 3) are essential for cell survival and growth. Some strains have become
resistant to methicillin and other plactams by acquiring an altered PBP2
(PBP2A, molecular weight 78 kDa) which has a low affinity for these drugs.
The PBP gene is chromosomal and seems to have been derived from the fusion
of the genes for a plactamase and a PBP from another source. The gene is
highly conserved in that limited proteolysis of the PBP from unrelated strains
produces similar peptides. Although almost all cells within a population carry
the gene, only a very few express it, so there must be other factors involved
in methicillin resistance (Hackbarth and Chambers, 1989). The altered PBP
does not reduce growth or virulence as the mutant strains are as virulent as
the wild type at causing life-threatening infections, for example pneumonia or
meningitis. The crucial site of action is probably the septum (the division
formed from plasma membrane and cell wall that forms across the middle of
the original cell, splitting it into two daughter cells); methicillin-sensitive cells
lyse when growth at this point is blocked, whereas resistant cells do not.
Infection is often transmitted within an institution (nosocomial), often a hospi-
tal and may reach epidemic proportions. The problem is that the resistant
strains may not be recognized in time to be treated with an appropriate drug
such as vancomycin (Chambers, 1988).

5.2.2 ~Lactamases
Bacterial resistance to the plactams may be derived, as with other antibiotics,
by preventing the ingress of the antibiotic into the cell or by altering the
enzyme target. Although the former possibility does not appear to play a major
part, it is clear that the latter does (Malouin and Bryan, 1986). In addition,
another major method for the development of bacterial resistance to the p
lactams is detoxification by enzymes which destroy the plactam ring, namely
8lactamases (see Amyes and Gemmell, 1992). The genetic coding for the
enzyme may be found both on a chromosome or a plasmid. In Gram-positive
bacterial Plactamase production is the major method of resistance in that the
bacteria produce a large quantity of the enzyme and secrete it extracellularly.
In Gram-negative bacteria the situation is more complex. The organism prod-
uces less plactamase and secretes it into the periplasmic space, i.e. the space
between the peptidoglycan band around the cytoplasmic membrane and the
bilayer outer membrane, which is very hydrophobic (construction of the
82 Molecular mechanisms of drug action

Gram-negative cell envelope is reviewed in Costerton and Cheng, 1975). The

enzyme is thereby positioned at the site where the peptidoglycan is formed
on the outer side of the cytoplasmic membrane, and it can therefore protect
the organism to the greatest extent. Furthermore, the outer membrane pre-
sents a much more serious obstacle to the transport of the antibiotic in Gram-
negative bacteria than it does in Gram-positive ones. As a consequence, the
possession of a plactamase correlates better with bacterial resistance for a
Gram-positive than for a Gram-negative bacterium (Sykes and Matthew, 1976).
PLactamases have been classified in various ways but probably the most
widely accepted classification depends on their amino acid sequence and cata-
lytic properties. Two groups of enzymes with serine at the active site have a
preference for penicillins on the one hand (class A), and cephalosporins on
the other (class C). The latter are usually chromosomally encoded enzymes
derived from Gram-negative Enterobacteria (Jaurin and Grundstroem, 1981).
Class B includes the Zn*+-containing enzymes (see Frere and Joris, 1985). The
enzymes are frequently extremely active, although the most useful way to
describe the activity is by the ratio of catalytic rate constant to Michaelis con-
stant WC,, to K,,,). For benzylpenicillin this figure is 9.8 X lo5 m/s, which is
close to the diffusion limit, i.e. the substrate is hydrolysed almost as fast as it
can diffuse to the active site of the enzyme. The enzyme splits most penicillins
by breaking open the plactam ring to yield penicilloic acids (Fig. 5.3).

CH3
CH3

COOH

6-Aminopenicillanic acid Penicilloic acid

I. Penicillin amidase breaks the amide link in the side-chain of the drug. This enzyme is
not used by the microorganism for protection against the drug. Penicillin amidase has
been used to prepare 6-aminopenicillanic acid from which new semi-synthetic deriva-
tives can be made by linking the amino group to new acyl side-chains.
I. /?-Lactam hydrolysis of the /I-lactam ring proceeds through attack by a nucleophilic
serine hydroxvl on the /3-lactam carbonyl, yielding a transient acyl-enzyme inter-
mediate. It is this enzyme that confers drug resistance upon the bacterium, and organic
chemists have been at pains to synthesize less susceptible new compounds.

Figure 5.3 Pathways of plactam metabolism.

 Cell wall biosynthesis 83

PLactamases vary markedly from each other in primary sequences having

in some cases less than 40 per cent homology, and so the overall shape of the
molecule is the key, rather like the G-protein family (see section 9.1.1). Indeed,
some PBPs are closer in structure to alactamases than to other PBPs. The
common feature both classes share is a serine at the active site, which forms
an acyl-enzyme intermediate (reviewed in Ghuysen, 1991).
In kinetic terms, the anti-bacterial binds reversibly to the enzyme. The com-
plex then isomerizes irreversibly into a complex with a covalent link between
enzyme and anti-bacterial which breaks down into enzyme and product:

K kz b3
E+I++EI+EI*+E+P

where EI* represents inactivated enzyme, K is an equilibrium constant and kz,

k, are rate constants. Provided that k, remains small and kz/K is high the
enzyme remains inactivated as in the PBPs, whereas if k, is high we have a
recipe for drug detoxification as in the plactamases.
Translated into molecular terms, the active site serine makes a nucleophilic
attack on the carbon atom of the carbonyl group in the lactam ring. The leaving
group is too bulky to diffuse away and remains attached to the serine. Only
a water molecule can attack the complex and release the leaving group. In
the PBPs the access is extremely slow or negligible, while in the alactamases
it occurs rapidly (Ghuysen, 1991). This may be because the plactamases have
a glutamate moiety able to act as a general base in the active site by abstracting
a proton from the water molecule to allow hydroxyl attack on the carbonyl
group. PBPs have a phenylalanine at the equivalent position which cannot do
this. P-Lactamases may not need the type of charge relay system seen in the
serine proteases, partly because the lactam as a cyclic amide is intrinsically
more reactive than an open chain amide. Furthermore, the lactam carbonyl is
likely to be further polarized by binding to two backbone amides (Escobar et
al., 1991; Strynadka et al., 1992).
A number of suicide substrates have been described, i.e. drugs which the
enzyme converts into products that destroy the enzyme’s activity (Chapter 1).
These are frequently derived from fermentation cultures, as in the case of
clavulanic acid from Streptomyces clavuligerus. Clavulanic acid interacts with
plactamases in a complex fashion but, in general, the first step involves the
formation of a complex that is non-covalent and readily dissociable. Secondly,
the ligand acts as a substrate and the Plactam ring is opened with the active
site serine forming an ester link. A true substrate would now dissociate from
the enzyme, but the clavulanic acid moiety rearranges to produce a conjugated
series of double bonds giving rise to an ultraviolet absorbance spectrum. Sub-
sequently, in some cases, the enzyme-inhibitor complex is reactivated and the
enzyme recovers its full activity. In others the complex is stable and the
enzyme remains irreversibly inhibited (Labia et al., 1985).
84 Molecular mechanisms of drug action

5.2.3 PLactam therapy

It has been of great importance to synthesize drugs which are resistant to p

lactamase detoxification. The introduction of a 2,6dimethoxyphenacyl group
to form an amide with the amino group at position 6 of the /3-lactam ring
sterically hindered @actam hydrolysis as in methicillin. Replacing the CH,
between the ring and the carbonyl group with a methoxyimine side-chain, as
in cefuroxime, also conferred stability to plactamases. Another approach is to
insert a 7cz-methoxy group in the plactam ring as in the family of cephamycins
of which cefoxitin is an example, although larger substituents in this position
produced a reduction in activity. Further substitution of the imino grouping, as
in ceftazidime, was designed to increase intrinsic activity against Gram-negative
organisms of clinical importance (such as E. coli and Klebsiella pneumoniae)
by virtue of the aminothiazolyl side-chain, whilst retaining resistance to plac-
tamases. Unfortunately, however, the series of plasmid-coded /!l-lactamases
have mutated to produce enzymes that hydrolyse the drug (see Amyes and
Gemmell, 1992). The introduction, however, of a positively charged quatern-
ary nitrogen at the 3 position has produced drugs that are more active against
enteric bacteria and Pseudomonas aeruginosa, largely through increased
resistance to beta-lactamases (Hancock and Bellido, 1992).
Another approach has been to design structures that have a carbocyclic
five- or six-membered ring attached to the lactam, known as carbapenems and
carbacephems. The most important examples of these classes are imipenem
and loracarbef. They have the great advantage that they are resistant to class
A and C plactamases (imipenem has a trans side-chain) albeit sensitive to the
class B zinc-containing enzymes. They also have a very wide spectrum of
activity against a range of organisms.
Imipenem binds strongly to PBP2 of E. coli and Pseudomonas aeruginosa
and to all four PBPs of Staphylococcus aureus. Imipenem appears to damage
Ps. aeruginosa more severely than other plactams without allowing regrowth.
Imipenem resistance results from a change in the permeability of the outer
membrane, specifically in the porins that form channels to allow the inward
passage of nutrients (reviewed in Buckley et al., 1992). Imipenem resembles
a substrate for the kidney dipeptidase, dehydropeptidase I, and is rapidly
degraded in viva by this enzyme. The drug is administered in a 1: 1 mixture
with cilastatin, a compound which inhibits dehydropeptidase (Farrell et al.,
1987). Loracarbef is given as a single drug and has the advantage of being
orally active (Cooper, 1992).
An alternative approach is to use plactams that are potent plactamase
inhibitors, but not necessarily anti-bacterial, to act in synergy with plactam
antibiotics. Various compounds with these properties have been discovered
in culture filtrates of Streptomycetes. In particular, clavulanic acid has been
developed for this purpose. A combination of clavulanic acid with amoxicillin
(trade name Augmentin) is effective in treating infections produced by organ-
isms that produce plactamase in large quantities, notably Bacteroides fragilis.
 Cell wall biosynthesis 85

It can be shown in vitro that clavulanic acid will enhance the activity of p
lactams against B. fragilis by lowering the MIC by two orders of magnitude
(Bansal et al., 1985).

CHCH20H

C02H

Clavulanic acid

The Plactams are probably the most widely used anti-bacterials at the pre-
sent time, partly because of their broad spectrum of action and partly because
of their very low toxicity. They must be about the least toxic drugs available
on the market.
Oral activity is clearly a great advantage for the use of drugs in the com-
munity, while injectable drugs are more easily given in a hospital setting.
Although the original penicillin (G) required injection, some a-amino-~lactams
are orally active. They are structural analogues of tripeptides that are taken up
by an oligopeptide transport system, used for small nutrient peptides, that
operates across the brush border membrane of the enterocyte (Kramer et al.,
1992). FLactams are themselves synthesized from a tripeptide, aminoadipyl-
cysteinyl-valine. The transport system is chirally sensitive in that n-cephalexin
is taken up while the L-isomer is not (Kramer et al., 1992).
A goal of plactam synthesis has been to make orally active drugs, either by
synthesis of a structure that is orally active, such as ampicillin or cephalexin,
or by modification of a structure that is only active by injection, i.e. by forming
a pro-drug. Cefuroxime axetil is an example of a pro-drug where the orally
inactive cefuroxime is converted at position 4 of the carboxyl to the acetoxye-
thy1 ester. The pro-drug is absorbed because it is more lipophilic not because
it is taken up through the peptide transport system. The pro-drug is absorbed
intact and then hydrolyzed in the intestinal mucosa or portal blood (Williams
and Harding, 1984).
The esterification of the carboxyl group introduces another chiral centre
into the molecule (positions 6 and 7 are already chiral) and the hydrolysis of
the isomeric esters in vivo may depend on the chirality of the 1’ position;
1 ‘R,bR,7R is rapidly hydrolyzed by an esterase in rat stomach while 1 ‘S,GR,7R
is almost completely resistant (Mosher et al., 1992).
The absorption of cefuroxime axetil was not straightforward, however, as
the first available tablets dissolved at variable rates and so the dose was vari-
able; a new test for measuring the rate of disintegration of a tablet had to be
developed (Emmerson, 1988). Furthermore, 50 per cent of the drug is
absorbed if taken after food but only 30 per cent after overnight fasting
86 Molecular mechanisms of drug action

(Williams and Harding, 1984). This is one practical aspect of the pharmaceut-
ical industry that can be overlooked in the quest for new medicines.

5.3 Vancomycin

Vancomycin is a large glycopeptide with a molecular weight in the region of

1500, obtained from the culture filtrates of Streptomyces orientalis. It has a
complex molecular structure (see Sheldrick et al., 1978) containing an amino
sugar called vancosamine linked to three aromatic rings; several amino acid
residues and a bis-resorcinol system. Vancomycin is active primarily against
Gram-positive bacteria. It is not a new antibiotic, having been discovered in
the mid-1950s but while it demonstrated undesirable side-effects it had no
advantage over the plactams, especially with the advent of the plactamase-
stable @actams.

Vancomycin

OH

743

OH +NH2

NHCO - dH

HC -CH3
dH,

------______

The three-dimensional molecular structure is very compact and excludes water mol-
ecules. A hydrophobic cleft is present on the side of the molecule on which the chlorine
atoms reside. The D-Ala-o-Ala portion of the muramyl pentapeptide binds in this cleft.
The groups postulated to take part in the interaction are surrounded by a dotted line.
See Fig. 5.4 for the detailed interaction proposed by Sheldrick et al. (1978).
 Cell wall biosynthesis 87

Latterly, however, the importance of methicillin-resistant Staphylococcus

aureus infections has indicated the need for a tried and tested antibiotic such
as vancomycin. In addition, pseudo-membraneous colitis responds well to van-
corny&. This condition is induced by the action of other antibiotics which
‘scour’ the gut leaving it almost empty of bacteria and thus allowing Clostrid-
ium difficile to colonize. Furthermore, more modern preparations appear to
be ‘cleaner’ and, therefore, the occurrence of side-effects is markedly reduced
compared with impure preparations of earlier years (Kucers, 1984; Watanaku-
nakorn, 1984).
The mode of action of vancomycin involves interference with the biosynth-
esis of cell wall peptidoglycan, but in a different way from the plactams. Van-
comycin binds tightly to UDP-pentapeptides containing o-alanine-u-alanine at
the free carboxyl end, thus causing UDP-Nacetylmuramyl-peptide precursors
to accumulate. The tightest binding was observed with UDP-N-acetylmuramyl-
glycyl-D-glutamylhomoseryl-u-alanyl-u-alanine, with a binding constant of
6 X lo-‘M, as measured by ultraviolet difference spectroscopy. The smallest
fragment to which binding was detected was acetyl-D-alanine-n-alanine
(Perkins, 1969). The molecule carries a marked hydrophobic cleft on one side
to which the peptide is bound as shown by nuclear magnetic resonance stud-
ies. The binding is initiated by the interaction between the peptide carboxylate
group and a protonated amine on the antibiotic (Williamson and Williams,
1984; Fig. 5.4).
As the target is a structural component of the cell wall, rather than an
enzyme, resistance to vancomycin was expected to be very unlikely. Recently,
however, resistant clinical isolates have turned up, where the pentapeptide
terminates in u-lactate with greatly reduced affinity for vancomycin

CH3

Interaction with acetyl-o-Ala-o-Ala is shown above. Ar, the trihydroxylated benzene ring
in the middle of the structure. Dotted lines indicate hydrogen bonds which are postulated
to be the major force in holding the complex together. The portion of the molecule that
interacts with the dipeptide moiety is outlined in the structure on previous page.

Figure 5.4 Vancomycin interaction with D-ala-D-ala moiety of pentapeptide

88 Molecular mechanisms of drug action

(Handwerger et al., 1992). Figure 5.4 shows three of the five hydrogen bonds
holding the drug-pentapeptide complex together - the other two arise from
the carboxylate oxygen not involved in binding in the figure (Barna and Willi-
ams, 1984). Resistance arises from the substitution of o-lactate for o-alanine
and so the NH in the amide link of the dipeptide is replaced by the oxygen
of an ester Link, thus removing one of the hydrogen bonds; a major contri-
bution to the binding is thereby lost. The pentapeptide altered from amide to
ester may act more efficiently as a substrate with the cross-linking enzymes to
form the wild-type peptidoglycan.
The primary mode of action of vancomycin is inhibition of cell wall
biosynthesis, and consequently results in cell lysis. Gram-positive organisms
are primarily affected because vancomycin is too large to cross the outer mem-
brane of Gram-negative organisms. The drug may also alter the permeability
of the cell membrane and, in addition, interfere with ribonucleic acid synthesis
(Watanakunakorn, 1984).

5.4 Mycolic acid synthesis - isoniazid for tuberculosis

Tuberculosis, known as consumption in earlier years, is the leading cause of

death from infectious diseases (Collins, 1993). The disease has recently
become much more frequent in developing countries and is also a serious
complication of AIDS, where the patient is suffering from a weakened immune
system. Tuberculosis is mainly caused by an infection with Mycobacterium
tuberculosis and can be lethal. The lung is most often the site of infection,
although the brain or skin can also be affected. The symptoms of pulmonary
tuberculosis are cough, fever and, in the later stages, extreme loss of weight,
severe night sweats and often haemorrhage. The disease gets its name from a
granule composed of leukocytes, epithelial cells, a few giant cells and myco-
bacteria, called the tubercle. As the disease progresses, the tubercles fuse to
form a soft mass and invade the healthy tissue. Eventually cavities form and
the lung collapses.
Most front-line anti-bacterial agents are inactive, or only weakly active,
against mycobacteria. Isoniazid (isonicotinic acid hydrazide) and rifampicin are

lsoniazid
 Cell wall biosynthesis 89

still the mainstays of anti-tubercular chemotherapy. They are normally given

in conjunction with a third drug, in case the microorganism is resistant to one
or both. Rifampicin binds to bacterial RNA polymerase (see section 2.3.5) but
isoniazid’s mode of action is still in doubt, despite having been used clinically
for over 40 years. Various proposals include reaction of the hydrazine with
the carbonyl group of pyridoxal phosphate enzymes, which can happen even
in vitro (Shoeb et al., 1985) or conversion of isoniazid to the isonicotinic acid
analogue of NAD. The evidence for these effects is, however, weak. The reason
for the inclusion of isoniazid in this chapter is that one important hypothesis
is inhibition of mycolic acid formation, a major component (- 40 per cent)
of mycobacterial cell walls (Quemard et al., 1991). This assignment is contro-
versial as there is also good evidence that activation of isoniazid by peroxidase
is involved. It may be, however, that these two mechanisms are not mutually
exclusive, in that peroxidase may activate isoniazid to a product that may
attack mycolic acid synthesis.
Mycolic acids, a family of closely related lipids found in mycobacteria and
a few other bacterial genera, are o-alkyl-phydroxy long-chain fatty acids with
between 60 and 90 carbon atoms. They contain branched methyl groups, dou-
ble bonds and cyclopropane rings. The general formula of the major mycolic
acid, cY-mycolic acid, present in M. tuberculosis is shown below (from Qureshi
et al., 1978):

OH COOH
CH&CH,). CH - CH (CH,), CH - CH (CH,). AH - AH (CH,), CH,
where a = 17, 18, 19; b = 10; c = 15, 17, 19, 21 and d = 21, 23.
The fatty acid synthetase which produces the Cz4 and Cz6 acids required for
the straight CYchain part of the mycolic acid has been purified from M. tubercu-
losis and is composed of two 500 kDa monomers. The enzyme is unusual in
being a soluble cytoplasmic source of fatty acids with carbon chains longer
than C,, or CzO.The enzyme also resembles the mammalian fatty acid synthet-
ase complex in that both acetyl and malonyl-coenzyme are substrates and all
the enzymic steps are sited on a single polypeptide. Fatty acids are released
from the synthetase as coenzyme esters (Kikuchi et al., 1992).
Mycolic acid synthesis has long been recognized as a possible target for
isoniazid. Sensitivity to isoniazid paralleled inhibition of mycolic acid synthesis
in two closely related strains of M. aurum. The concentration for mycolic
acid inhibition in a cell-free system was five times higher than the MIC (&, of
3.5 X lop5 M compared with MIC of 7.3 X lop6 M) but this relationship may
be explained by the ability of the mycobacterium to concentrate isoniazid.
Peroxidase activity was similar in the two strains (Quemard et al., 1991).
There is also evidence that oxidation by peroxidase is the key target. Mycotu-
bercular resistance to isoniazid is associated clinically with gene deletion of
an enzyme which has peroxidase and low catalase activity. Hydrogen peroxide
seems important for the sensitivity, because if the catalase activity is higher,
90 Molecular mechanisms of drug action

as in some mycobacterial species, hydrogen peroxide is broken down to water

and oxygen and the microorganism is resistant (Shoeb et al., 1985). Hydrogen
peroxide oxidizes isoniazid to a radical with peroxidase as the catalyst. Inser-
tion of the gene into resistant M. tuberculosis or E. coli confers sensitivity to
isoniazid. The suggestion is that the radical, whether drugderived or super-
oxide, is responsible for damaging the cellular machinery (Zhang et al., 1992).
It is likely that this uncertainty about the mode of action of isoniazid will
soon be resolved now that tuberculosis is becoming a much more serious
therapeutic problem.

5.5 PGlucan synthetase

The fungal cell wall consists of three polysaccharides in varying proportions:

chitin, mannan and pglucan, held together by cross-links. Chitin is a homo-
polymer of Nacetylglucosamine, mannan of mannose complexed to a small
protein, and Pglucan of glucose. PGlucan provides the main structural poly-
mer and is largely made of glucose 1,3-linked with a small proportion of 1 ,G
linked material to provide the cross-links (Cabib et al., 1982).
PGlucan, like most polysaccharides, is synthesized from a nucleotide sugar,
UDP-glucose, by the cell membrane enzyme pglucan synthetase, and is
inhibited non-competitively by the semisynthetic lipopeptide, cilofungin, with
an apparent Ki of 2.5 X 10e6~. Both the peptide nucleus and the lipophilic
side-chain are essential for the inhibition (Tang and Parr, 1991); the $hydroxy-
4-methylproline residue enhances activity, but the t-homotyrosine moiety is
crucial for both antifungal activity and @glucan synthetase activity (Zambias
et al., 1992).
Cilofungin inhibits the growth of the pathogenic Candida albicans and C.
tropicalis in vitro with minimum inhibitory concentrations (MIC)
CO.31 pg/ml, but other Candida species less powerfully. The growing organ-
ism forms protoplasts, a form without a cell wall in the presence of the drug,
and may burst depending on the external osmotic pressure. To growing cells,
therefore, cilofungin is fungicidal and to stationary cells it is fungistatic in
vitro. Althoug pglucan is widespread in the fungal kingdom, other fungi are
resistant to the action of the drug, possibly because they contain less &&can
(Hall et al., 1988).
Cilofungin is under development for the treatment of systemic or topical
infections by Candida albicans, particularly in immunosuppressed patients.
The cell wall of Pneumocystis carinii also contains 1,3-pglucan, and echino-
candin analogues may also be useful for the treatment of this microorganism
in AIDS patients.
 Cell wall biosynthesis 91

OH

CH-OH
OH
CH-OH I

CH-Me 1
I c=o
OH I

Cilofungin

Questions

1. What are the effects on E. coli of drug binding to penicillin-binding proteins

(a) 2 and (b) 3?
2. What families of plactams are available to the clinician? How do they dif-
fer structurally?
3. What is the therapeutic advantage of an orally active plactam compared
with one that has to be given intravenously?
4. How do plactamases and penicillin-binding proteins differ kinetically?
5. What methods do bacteria use to detoxify plactams? What is their relative
importance clinically?
6. Vancomycin is particularly useful in the treatment of infections caused by
which bacteria? Why is it effective?
7. How have bacteria developed vancomycin resistance?
8. Isoniazid blocks the synthesis of which component of the mycobacterial
cell wall?
9. What are the main constituents of the fungal cell wall? With the synthesis
of which component does cilofungin interfere?

References
Amyes, S.G.B. and GemmelI, C.G., 1992, J. Med. MicrobioZ., 36, 4-29.
Bansal, M.B., Church, SK., Onjema-Lalobo, M. and Thadepalli, H., 1985, Chemotherapy
(Basel), 31, 173-7.
Barna, J.C.J. and Williams, D.H., 1984, Annu. Rev. Microbial., 38, 339-57.
92 Molecular mechanisms of drug action

Blumberg, P.M. and Strominger, J.L., 1971, Proc. Natl. Acad. Sci. (U.S.A.), 68, 2814-17.
Buckley, M.M., Brogden, R.N., Barradell, L.B. and Goa, K., 1992, Drugs, 44, 408-44.
Cabib, E., Roberts, R. and Bowers, B., 1982, Ann. Rev. Biocbem., 51, 763-93.
Chambers, H.F., 1988, Clin. Microbial. Rev., 1, 173-86.
Collins, F.M., 1993, Crit. Rev. Microbial., 19, l-16.
Cooper, R.D.G., 1992, Amer. J. Med., 92, Suppl 6A, 2-6.
Costerton, J.W. and Cheng, KJ., 1975, Bact. Rev., 38, 87-110.
Curtis, N.A.C., Brown, C., Boxall, M. and Boulton, M.G., 1979a, Antimicrob. Ag. Cbem-
other., 15, 332-6.
Curtis, N.A.C., Orr, D., Ross, G.W. and Boulton, M.G., 1979b, Antimicrob. Ag. Cbem
other., 16, 533-9.
Emmerson, A.M., 1988,J: Antimicrob. Cbemotber., 22, 101-4.
Escobar, W.A., Tan, A.K. and Fink, A.L., 1991, Biocbemisty, 30, 10783-7.
Farrell, O.A., Allegretto, N.J. and Hitchcock, M.F., 1987, Arch. Biocbem. Biopbys., 256,
253-9.
Frere, J.M. and Joris, B., 1985, CRC Crit. Rev. MicrobioZ., 11, 299-396.
Gale, E.F., Cundliffe, E., Reynolds, P.E., Richmond, M.H. and Waring, M.J., 1981, Eds.
The Molecular Basis of Antibiotic Action, 2nd Ed. (London: John Wiley) Ch. 3.
Ghuysen, J.-M., 1991, Annu. Rev. Microbial., 45, 37-67.
Gooday, G.W., 1979, J Gen. Microbial., 99, 1- 11.
Hackbarth, C.J. and Chambers, H.F., 1989, Antimicrob. Ag. Cbemotber., 33, 991-4.
Hall, G.S., Myles, C., Pratt, K.J. and Washington, J.A., 1988, Antimicrob. Ag. Cbem-
other., 32, 1331-5.
Hancock, R.E.W. and Bellido, F., 1992, J Antimicrob. Cbemotber., 29, Suppl. A, l-6.
Handwerger, S., Pucci, M.J., Volk, K.J., Liu, J. and Lee, M.S., 1992,J. Bacterial., 174,
5982-4.
Hayes, M.V. and Orr, D.C., 1983,J. Antimicrob. Cbemotber., 12, 119-26.
Jaurin, B. and Grundstroem, T., 1981, Revs. Znfect Dis., 8, Suppl. 3, S237-59.
Kikuchi, S,, Rainwater, D.L. and Kolattukudy, P.E., 1992, Arch. Biocbem. Biopbys.,
295, 318-26.
Kramer, W., Girbig, F., Gutjahr, U., Kowalewski, S., Adam, F. and Schiebler, W., 1992,
Eur. J Biocbem., 204, 923-30.
Kucers, A., 1984,J. Antimicrob. Cbemotber., 14, 564-7.
Labia, H., Barthelemy, M. and Peduzzi, J., 1985, Drugs Exptl. Clin. Res., 11, 765-70.
Malouin, F. and Bryan, L.E., 1986, Antimicrob. Ag. Cbemotber., 30, 1-6.
Mosher, G.L., McBee, J. and Shaw, D.B., 1992, Pbarm. Res., 9, 687-90.
Neu, H.C., 1986, Revs. Infect. Dis., 8, Suppl. 3, S237-59.
Perkins, H.R., 1969, Biocbem.J, 111, 195-205.
Quemard, A., Lacave, C. and Laneelle, G., 1991, Antimicrob. Ag. Cbemotber., 35,
1035-9.
Qureshi, N., Takayama, K., Jordi, H.C. and Schnoes, H.K., 1978, J. Biol. Cbem., 253,
5411-17.
Samraoui, B., Sutton, B.J., Todd, R.J., Artymiuk, P.J., Waley, S.G. and Phillips, D.C.,
1986, Nature, 320, 378-80.
Sheldrick, G.M., Jones, P.G., Kennard, O., Williams, D.H. and Smith, G.A., 1978, Nature,
271,223-5.
Shoeb, H.A., Bowman, B.U., Ottolenghi, A.C. and Merola, A.J., 1985, Antimicrob. Ag.
Cbemotber., 27, 399-403.
Spratt, B.G. and Pardee, A.B., 1975, Nature, 254, 516-17.
Spratt, B.G., 1980, Phil. Trans. R. Sot. Lond. B., 289, 273-83.
Spratt, B.G., 1983, J. Gen. Microbial., 129, 1247-60.
Strynadka, N.C.J., Adachi, H., Jensen, S.E., Johns, K., Sielecki, A., Betzel, C., Sutoh, K.
and James, M.N.G., 1992, Nature, 359, 700-705.
Sykes, R.B. and Matthew, M., 1976, J Antimicrob. Cbemotber., 2, 115-57.
 Cell wall biosynthesis 93

Tang, J. and Parr, T.R., 1991, Antimicrob. Ag. Cbemother., 35, 99-103.
Tipper, D,J, and Strominger, J.L., 1965, Proc. Natl. Acad. Sci. (U.S.A.), 54, 1133-41.
Watanakunakom, C., 1984, J. Antimicrob. Chemother., 14, Suppl. D, 7-18.
Waxman, D.J., Yu, W. and Strominger, J.L., 1980,J. Biol. Cbem., 255, 11577-87.
Williams, P.E.O. and Harding, S., 1984, J. Antimicrob. Cbemotber., 13, 191-6.
Williamson, M.P. and Williams, D.G., 1984, Eur. J. Biocbem., 138, 345-8.
Zambias, R.A., Hammond, M.L., Heck, J.V., Bar&al, K., Trainor, C., Abruzzo, G.,
Schmatz, D.M. and Nollstadt, KM., 1992,J. Med. Cbem., 35, 2843-55.
Zhang, Y., Heym, B., ALlen, B., Young, D. and Cole, S., 1992, Nature, 358, 591-3.
 Chapter 6

Steroid biosynthesis and action

6 I Introduction
The major sterols to be discussed in this chapter are ergosterol and cholesterol,
which are essential constituents of cell membranes of yeasts and fungi, and
of mammals respectively. Both sterols are synthesized in a similar fashion start-
Table 6.1 Drugs discussed in Chapter 6

Section Target Drug Use

6.2 Sterol biosynthesis

6.2.1 FHydroxy-pmethyl glutaryl Hypercholesterolaemia
CoA reductase Lovastatin Hypercholesterolaemia
6.2.2 Squalene epoxidase Naftitine Dermatophyte infections
Terbinatine Dermatophyte infections
Tolnaftate Dermatophyte infections
6.2.3 Lanosterol 14odemethylase Ketoconazole Fungal infection
Itraconazole
Fluconazole

6.3 Steroid biosynthesis

6.3.1 Steroid 17,2@lyase Ketoconazole Hormone-dependent
prostate tumours
6.3.2 Steroid 1 l/?-hydroxylase Metyrapone Hypercortisolaemia
6.3.3 Aromatase Aminoglutethimide Hormone-dependent breast
cancer
6.3.4 5wreductase Finasteride Prostate hyperplasia

6.4 Steroid receptor ligands

6.4.1 Oestrogen receptor agonists Ethinyloestradiol Oral contraceptive
Mestranol Oral contraceptive
Progesterone receptor agonists Norethisterone Oral contraceptive
Norgestrel Oral contraceptive
6.4.2 Oestrogen receptor antagonist Tamoxifen Hormonedependent breast
cancer
Oestrogen receptor antagonist Clomiphene Fertility stimulant
6.4.3 Progesterone receptor Mifepristone Pregnancy terminator
antagonist
6.4.4 Androgen antagonist Flutamide Prostate cancer
Nilutamide
6.4.5 Aldosterone receptor Spironolactone Diuretic
antagonist

95
96 Molecular mechanisms of drug action

ing from acetyl-coenzyme A (Fig. 6.1). The synthesis takes place in different
parts of the cell: acetate to mevalonate is carried out in the microsomal fraction
while mevalonate to squalene is a cytoplasmic function; squalene conversion
to ergosterol and cholesterol is performed in the microsomes and this is the
first part of the pathway to require oxygen. The pathways to ergosterol and
cholesterol are common up to lanosterol. Furthermore, the removal of the
methyl group attached to the carbon atom at position 14 of lanosterol is also
common, although in yeasts and fungi it does not necessarily occur next in
sequence (Fig. 6.3). Not all the enzymes have been purified, but it appears
that, although yeast and mammalian enzymes catalyse the same interconver-
sions, they are not necessarily identical, and thus lend themselves to selec-
tive inhibition.

Acetyl-CoA

1
Acetoacetyl-CoA

1
P-Hydroxy-P-methylglutaryl-CoA

Mevaionete
1
/ \
lsopentenylpyrophosphate --Dimethylallylpyrophosphate

Farnesylpyrophosphate

Squakne

Squalbe epoxide

1
Lanosterol

/ \
Cholesterol Ergosterol
Figure 6.1 Sterol biosynthetic pathway.
 Steroid biosynthesis and action 97

Cholesterol can undergo a variety of reactions. Part of the side-chain is

removed as the first step in the formation of the steroid hormones (Pig. 6.2).
This leads to (a) glucocorticoids, such as cortisol which, amongst other
actions, increase gluconeogenesis and stimulate lipolysis, (b) mineralocorti-
coids, such as aldosterone that cause sodium to be reabsorbed in the kidneys
thus ensuring that sufficient water is retained to maintain the body’s osmotic
balance and (c) the final removal of the side-chain yields the sex steroids, such
as testosterone in the male and oestradiol in the female. Although the through-
put down these pathways is fairly small, the effects of the hormones produced
is very marked.
The greatest proportion of cholesterol is converted into bile acids which
are released into the gut to emulsify triglycerides so that they can be trans-
ported into the circulation. Cholesterol can also be esterified at the 3-hydroxy
position with long-chain fatty acids donated from either triacyl glycerols such
as lecithin (catalysed by the plasma enzyme lecithin cholesterol acyl-transfer-
ase: LCAT) or their respective coenzyme-A esters (tissue-bound acyl-CoA acyl-
transferase: ACAT). The esters so formed act as a storage depot for cholesterol,
and also in the former case to keep cholesterol within the circulating lipopro-
tein and prevent it from transferring into the tissues.
The cholesterol nucleus cannot be broken down directly to carbon dioxide
and water, and so if cholesterol levels rise there is no simple metabolic path-
way that can remove it. Elevated cholesterol, when attached to very low den-
sity lipoproteins, gives rise to one of the major diseases facing the Western
world today (section 6.2.1).
Ergosterol, strangely enough, is only known to undergo (a) esteribcation
and (b) conversion to ergocalciferol (vitamin D,) under the influence of ultra-
violet light. At present there are no other metabolic transformations described,
although by analogy with cholesterol we might expect some conversions, per-
haps to a yeast hormone. On the other hand, like cholesterol, the ergosterol
nucleus cannot be broken down directly to carbon dioxide and water.
The sterol biosynthetic pathway has not, until recently, attracted much
attention from the pharmaceutical point of view, except perhaps from those
seeking to lower cholesterol levels. Much effort went into this direction in the
1960s, after the important connection between cholesterol and coronary heart
disease was found. There was considerable success in reducing the flow of
metabolites through this pathway, but the agents led instead to the formation
of toxic intermediates and were of no value as drugs. Recently, however, the
discovery of non-toxic agents that interfere with the rate-limiting step in the
early part of the pathway, phydroxy-fimethylglutaryl-CoA reduction to meva-
lonate, has greatly improved the prospects for the treatment of hypercholester-
olaemia and, hopefully, coronary heart disease.
More recently, the part of the pathway leading from lanosterol to ergosterol
has been spotlighted as the target of various antifungal agents, some useful
as agricultural fungicides and others in human medicine. This interest is likely
to turn what was previously rather a backwater, inhabited only by organic
 98 Molecular mechanisms of drug action

Cholesterol

26

27

Cholesterol
(1) Hydroxylation at C,, and cleavage of
C,-C,, bond
1
Pregnenolone
(2) 3-OH + 3-keto
Double bond at C,,, migrates to Cd,,
I
Progesterone
21-H + 21-OH 1 17-H --) 17x-OH

(3) 11 -Deox&ticosterone 171.OH-Ptogesterone (4)

(6) 11-H + 116-OH (3) 21-H + 2l-OH 1-1 ;;;;aK of C,,,,

1
Corticosterone 11 -Deoxycortisol Androstenedione
1 17-keto --, 17-OH
(6) C,s+ aldehyde (6) 11-H+ ll6-OH Testosterone
3-keto--r 3f3-OH (5)
I I A ring becomes aromatic
Aldosterone Cortisol 176Oestradioi

Aldosterone ~H20H Oestradiol

- CH OH
I 2

(1) Cholesterol 20.22.lyase, mitochondrial.

(2) 3-P-Hydroxysteroid dehydrogenase. mitochondrial.
(3) 21Steroid hydroxylase, microsomal.
(4) I7-&Steroid hydroxylase and I7,20-lyase. microsomal.
(5) Aromatase, microsomal.
(6) 1lPSteroid hydroxylase, mitochondrial.
Ail the enzymes except 3 are cytochrome P-4505, similar to the liver mixed-function oxidases.
and thus require NADPH and molecular oxygen to function. Unlike the latter, however, they are
selective for their particular substrates.
Figure 6.2 Main pathways of steroid hormone biosynthesis.
 Steroid biosynthesis and action 99

Lanosterol
HO

HO

HO

The 14a-methyl group of lanosterol is sequentially hydrcxylated in the presence of a cytochrome

P-450-linked enzyme, the tlavoprotein cytochrome P-450 reductaae. NADPH and molecular oxygen.
The final oxidation yields formic acid, together with a sterol double-bonded at position 14-15.

Figure 6.3 Lanosterol 14cydemethylation.

chemists interested in the mechanism of squalene epoxide cyclization, for

example, into one of the most fruitful areas in biochemistry.
The conversion of cholesterol into various steroid hormones has been of
major pharmaceutical importance for many years because of the development
of oral contraceptives. Although the use of oestrogens and progestins to pre-
vent conception is not to combat disease, it is nonetheless one of the major
uses of drugs. Possibly because the use of drugs has greatly increased the life
expectancy of the population of a large part of the world, the use of drugs to
restrain population growth may be regarded as a moral imperative for the
industry. Alternatively, it may be argued that the industry is merely responding
to a need, in that women increasingly desire to regulate their reproductive
function. Whichever view is taken, there is still considerable effort being
expended to find other means of achieving the same end, particularly in view
of the uncertainty as to whether the pill can cause breast cancer (Anon, 1986).
100 Molecular mechanisms of drug action

Another pharmaceutical outlet for inhibitors of steroid biosynthesis, which

are often hydroxylations carried out by cytochrome P-450 dedicated to a par-
ticular reaction, has been to lower the levels of sex hormones, glucocorticoids
etc., in conditions where these pathways are hyperactive, and in situations
where tumours are particularly dependent on sex steroids. Antagonists of min-
eralocorticoid action are also useful as diuretics.

6 2 Sterol biosynthesis

6.2.1 P-Hydroxy-f3-methylglutaryl-CoA reductase - lovastatin as

hypocholesterolaemic
The lowering of plasma cholesterol levels, whether by inhibition of biosynth-
esis or reduction in gastrointestinal absorption, has been a major pharmaceut-
ical target for many years, since the establishment of a strong causative connec-
tion between plasma lipid, the deposition of fat and other material in the
arteries, or atherosclerosis and coronary heart disease. In outline, atheroscler-
osis results in the narrowing of the arteries in the following way. Cholesterol
is transported via the plasma in three types of lipoprotein complexes charac-
terized by their densities: very low density (VLDL) which has the highest lipid
content, low density (LDL) which is intermediate and high density (HDL) with
the least lipid. These complexes are of high molecular weight, with a core of
cholesteryl esters, triglycerides and free cholesterol surrounded by phospho-
lipid and lipoprotein and they provide the transport mechanism for hydro-
phobic lipids in an aqueous environment.
At the surface of the cells in the lining of the arteries (e.g. the aortic intimal
cells) are receptors for the protein component of LDL, which is the lipoprotein
whose levels correlate most closely with the development of atherosclerosis.
LDL binds and the complex is taken up by the cell. There, any remaining
free cholesterol is converted to esters by acyl-CoA cholesterol acyltransferase
(ACAT). These esters can be laid down intracellularly to produce what is
known as a ‘fatty streak’, and at this stage the deposition can be reversed.
Subsequently, lipids begin to accumulate extracellularly, particularly at points
where previous injury to the lining has occurred. Fibrin and collagen deposit
around the growing plaque. In addition, the smooth muscle cells proliferate
at that point to form at least 50 per cent of the advanced plaque, and the
artery begins to be blocked. At this point, the lesion is no longer reversible.
High plasma cholesterol levels, particularly associated with LDL, are correlated
with plaque formation, and with the development of coronary heart disease
(Wissler, 1991).
The rate-limiting step in the earlier part of the pathway is phydroxy-p
methylglutaryl-CoA reductase, which is subject to feedback inhibition by diet-
ary cholesterol. A powerful inhibitor of this enzyme, isolated from broths of
the fungus Aspergillus terreus, is lovastatin (originally known as mevinolin -
 Steroid biosynthesis and action 101

CH, CH, CH,

HO : HO :
COO- NADPHTH’ coo- NADPH+H+ H”~coo-

0 OH OH
L
T -7
SCoA SCoA

HMG-&A Half-reduced Mevolonate

IntermedIate

OH
0

Active form of
lovastatin
HMG-CoA is reduced in two stages by two molecules of reduced nicotinamide adenine dinucleotide
phosphate (NADPH) to form mevalonate. The half-reduced intermediate is a structural analogue of the
active form of lovastatin.

Figure 6.4

Alberts et al., 1980). The drug is isolated as the lactone form but the active
principle is the open chain hydroxy acid. Lovastatin inhibits the enzyme com-
petitively with a & of 6 X lo-” M; the upper part of the drug in Fig. 6.4 closely
resembles the half-reduced reaction intermediate (Grundy, 1988).

HO 0

Lovastatin
Lovastatin does not work, however, by sharply lowering the rate of choles-
terol synthesis, as plasma cholesterol levels are only slightly reduced in man.
It is the intracellular levels of cholesterol that are important. Cholesterol sup-
presses the synthesis of hydroxymethylglutaryl-CoA (HMG-CoA) synthase,
102 Molecular mechanisms of drug action

HMG-CoA reductase and the LDL receptor by binding to a DNA sequence in

the promoter of the genes for all three proteins (Smith et al., 1988). When
lovastatin reduces cholesterol levels in the cell, the synthesis of all three pro-
teins increases. Although raised HMG-CoA synthase and reductase can partially
offset lovastatin inhibition of cholesterol synthesis, the net effect is still to
increase the concentration of LDL receptors, thereby stimulating the removal
of LDL from the circulation. In people who have a genetic deficiency of LDL
receptors, this mechanism cannot work and there is no change in LDL levels
or rates of production or breakdown (Uauy et al., 1988).
Clearly, it is necessary to show that lowering LDL levels does in fact reduce
heart attacks, and a number of trials have been carried out with cholesterol-
lowering drugs. There is considerable controversy about the results of these
trials. There appears to be no doubt that the drugs reduce the frequency of
what is euphemistically called ‘coronary events’ (i.e. heart attacks). There does
not, however, appear to be an overall reduction in mortality from all causes
(Ravnskov, 1992; Smith and Pekkanen, 1992). Even the largest trial to date
with lovastatin (EXCEL) seems to be giving equivocal results, although it is
perhaps too early to tell (Bradford et al., 1991). Improving the diet may be at
least as effective as drug treatment (Smith and Pekkanen, 1992).

6.2.2 Squaleneepoxidase - naftifineasanti--gal

The intermediates immediately before squalene in the sterol biosynthetic path-
way are rendered soluble by the presence of pyrophosphate groups. Squalene
is the first substrate that is insoluble in aqueous media - indeed it must be
one of the most, if not the most, insoluble compounds in the cell. It appears
that a sterol carrier protein has evolved to handle this situation. In addition,
squalene epoxidase is of considerable interest because it is responsible for the
first oxidative step in the pathway.
The enzyme is present in the microsomes of eukaryotic cells, and requires
the presence of cytoplasmic fraction to show activity, presumably because the
latter contains the sterol carrier protein. The cofactors are NADPH or NADH
and FAD, but the enzyme is not a cytochrome P-450. Although the enzyme
catalyses the placement of an oxygen atom across positions 2 and 3 of the
squalene ring to form an epoxide, it may also be able to put a second one
across positions 22 and 23, because conditions that induce an accumulation of
the mono-epoxide, such as inhibition of the next step in the pathway (squalene
epoxide cyclase) by chloroquine (Chen and Leonhardt, 1984) yield a pro-
portion of the di-epoxide as well. The enzyme has been purified from rat liver
(Ono et al., 1982) but not so far from any fungal or yeast source, although
the Candida albicans enzyme has requirements for activity similar to the rat
liver enzyme (Ryder and DuPont, 1984).
Recently a new class of antifungal agents, the allylamines, has been found
to exert its antifungal effect by virtue of powerful inhibition at this step
(reviewed by Balfour and Faulds, 1992). Naftifme was the first member of this
 Steroid biosynthesis and action 103

series to be synthesized but a more recent one, terbinafine, is more potent at

inhibiting both enzyme activity and fungal growth. Terbinafine shows non-
competitive inhibition of the enzyme from Candida albicans with a Ki of
3 X lo-’ M (Petranyi et al., 1984). The enzyme from rat liver is some 2000-
times less sensitive to the drug, and this selectivity appears to be reflected in
viva (Petranyi et al., 1984). The inhibition is apparently reversible, despite the
system of conjugated double and triple bonds which might be expected to
interact covalently with the enzyme.
Squalene epoxidase inhibition will lead to rising levels of squalene as the
substrate and falling levels of ergosterol, the final end-product (Fig. 6.1).
Against Tricbophyton species, that cause skin infections such as athlete’s foot,
inhibition of fungal growth parallels rising levels of squalene. Terbinahne is
fungicidal for these fungi. Fungistatic action against Candida species appar-
ently correlates with falling levels of ergosterol (Balfour and Faulds, 1992).
Candida albicans is a dimorphic fungus, i.e. it can exist either as a yeast
or as a mycelial form; the latter is believed to be the pathogenic form as it
can force its way through tissues. The change from one form to the other,
known as the dimorphic switch, has been the subject of much research. Terbi-
naGme is almost inactive against the yeast form while very active against the
mycelia, and thus blocks the switch. This could also explain fungistatic activity
against C. albicans as the fungus could revert to yeast form in the presence
of the drug and resume growth in its absence.
Terbinatine is orally and topically active in humans against a wide variety
of fungal skin infections and may become a first-line therapy in the future
(Balfour and Faulds, 1992).

Terbinafine

Another family of compounds which have been in use for some time against
topical infections is exemplified by tolnaftate. Recent studies have shown that
it also inhibits squalene epoxidation. Although active against the Candida
enzyme, the drug apparently cannot reach its site of action in these yeasts
Barrett-Bee et al., 1986). At concentrations that inhibit ergosterol biosynth-
104 Molecular mechanisms of drug action

esis, both families of drugs appear to sensitize the C albicans membrane to

further damage (Georgopapadakou and Bertasso, 1992).

6.2.3 Lanosterol demethylation - ketoconazole as anti-hngal

One of the crucial rate-limiting steps in the later part of the pathway of choles-
terol biosynthesis is the removal of the CS2methyl group from the 14a position
of lanosterol with a concomitant introduction of a double bond at the 14-15
position (Pig. 6.3). These transformations appear to be carried out by a single
microsomal enzyme that is cytochrome P-450 dependent, and requires
NADPH, molecular oxygen and cytochrome P-450 reductase, a flavoprotein,
for activity (reviewed by Co&on et al., 1984). The methyl group is sequen-
tially hydroxylated to yield -CH,OH and -CHO, and then removed as formic
acid as the double bond is inserted at the 14-15 position. Carbon monoxide
which binds to the haem iron of cytochrome P-450 and interferes with a num-
ber of P-450 linked reactions, inhibits the first step but not the latter two
(Gibbons et al., 1979) but that may merely reflect tighter binding of the oxid-
ized intermediates to the enzyme, i.e. carbon monoxide may be able to dis-
place lanosterol, but not the oxidized intermediates, at the binding site. More
recent work has confirmed that the same cytochrome P-450 will catalyze all
three steps in the process (Aoyama et al., 1987).
In fungi, the demethylation of lanosterol is not rate-limiting as it is in liver,
but it is nevertheless essential for the formation of ergosterol and other sterols
that are acceptable materials for the fungal cell membrane. Lanosterol and its
analogues which retain the methyl group do not fit properly into the mem-
brane, causing it to become permeable to protons and eventually burst (Nes
et al., 1978; Thomas et al., 1983). The 14cy-methyl group of lanosterol is axial,
protruding from the otherwise planar face. Van der Waals interactions of the
sterol underside with the fatty acyl groups in the phospholipid bilayer of the
membrane are therefore very much less favourable (Bloch, 1979).
A number of medical and agrochemical antifungal agents in use today owe
their activity to potent inhibition at this step. The majority are substituted
nitrogen heterocycles. The presence of a lone pair of electrons which can be
donated from the nitrogen to the iron of the haem appears to be mandatory.
The binding of ligand to enzyme gives rise to a shift in the ultraviolet spectrum
whereby the absorbance of the haem is shifted from 410 nm to 420-430 nm
(Fig. 6.5; Van den Bossche et al., 1984). This results in what is known as a
Type II difference spectrum characterized by a maximum in the region of 420-
430 nm and a minimum around 390-400 nm. Substrates of cytochrome P-450-
linked reactions such as lanosterol give a totally reversed type of difference
spectrum, called Type 1, where a peak is found at 380-390 nm and a trough
at 410-420 nm.
Inhibitors of lanosterol demethylation in medical use include substituted
imidazoles, such as ketoconazole, and the newer triazoles, fluconazole and
itraconazole. The inhibitory effect of the ‘azoles’ is usually measured by the
 Steroid biosynthesis and action 105

AAbsorbance
(i.e. difference 0
spectrum)

A difference spectrum is used because the changes in ultraviolet spectrum during inter-
action of cytochrome P-450 with a ligand are often too small to be detected by reading the
direct spectrum. The spectrum of cytochrome and ligand uncomplexed is subtracted from
the spectrum of the complex. This can be done electronically by subtracting spectra
stored in memory or, experimentally, by the use of split compartment cells.
A type I difference spectrum is characteristic of a substrate binding to a cytochrome
f-450 whereas type II is given by a poorer substrate or an inhibitor. The absorbance
maximum of the cytochrome is shifted to longer wavelength from about 410 nm by such
inhibitors to give the type II spectrum, while a substrate causes the wavelength maximum
to move to shorter wavelength. Ketoconazole gives a type II difference spectra.
This subject is also discussed in Chapter 7, section 7.5.

Figure 6.5 Cytochrome P-450 difference spectra (diagrammatic representation)

Fluconazole

inhibition of incorporation of ‘*C-acetate or -mevalonate into ergosterol, which

is more convenient to use than a specific assay for lanosterol demethylation
and is more flexible in that inhibition of various steps in the pathway may be
identified. It is quite clear in this case from the pattern of inhibition obtained,
namely a build-up of lanosterol and a decrease in ergosterol, where the
inhibited step lies. Ketoconazole and fluconazole give 15o values in the region
of 10P8~ to 10P9~ in this system (Grant and Clissold, 1990).
Ketoconazole has two asymmetric carbon atoms and thus has four stereoiso-
mers, all of which show some activity against lanosterol demethylase. The
most active is the 2S,4R isomer (Rotstein et al., 1992). There is no information
available about the stereoisomers of itraconazole, and fluconazole is not
optically active as the central carbon atom has symmetrical substituents.
106 Molecular mechanisms of drug action

Ketoconazole

The main action of this class of drug is to disrupt the supply of ergosterol
for membrane synthesis. This effect is likely to be fungistatic not fungicidal
because the inhibitor will be effective only when the organism is growing
and dividing.
The advent of the azoles has greatly improved the treatment of a variety of
fungal infections. They are orally active and can be used for infections that
are carried throughout the body (systemic) and dermatophyte infections that
are confined to the surface of the skin (reviewed in Grant and Clissold, 1990;
Hay, 1991).
Ketoconazole is relatively specific in inhibiting the fungal lanosterol deme-
thylase as opposed to the same enzyme from rat liver (Van den Bossche et al.,
1980, 1984). Nevertheless, a number of other cytochrome P-450 conversions,
particularly those in the interconversion of steroid hormones, are sensitive to
inhibition by the drug. Two of the most obvious of these are (a) the rate-
limiting step in the production of glucocorticoids and sex steroids, i.e. choles-
terol 22,24 side-chain cleavage (Loose et al., 1983) and (b) 17,20-lyase (Sikka
et al., 1985; reaction 4 in Fig. 6.2). These effects are probably responsible for
the occurrence of undesirable side-effects such as the growing of breast tissue
in men (gynaecomastia), probably through lowered testosterone levels (Pant
et al., 1982). Other problems include teratogenicity and some incidence of
serious, indeed fatal, liver toxicity (Lewis et al., 1984).
These inhibitions of cytochrome P-450 linked enzymes may be put to good
effect, however, in conditions in which there is overproduction of ACTH lead-
 Steroid biosynthesis and action 107

ing to overstimulation of the adrenals (Cushing’s syndrome); ketoconazole has

been found effective in this instance (Angeli and Frairia, 1985). Furthermore,
some tumours are dependent on sex steroids and respond to the drug (Amery
et al., 1986).In this case, the observation of side-effects has led to the develop-
ment of the drug for another purpose.

6.3 Steroid biosynthesis

6.3.1 Steroid 17,201yase - ketoconazole for steroiddependent

tumours
The conversion of progesterone into 17a-hydroxyprogesterone followed by
the removal of the acetyl side-chain to yield androstenedione (reaction 4 in
Fig. 6.2) is believed to be carried out by one enzyme - but there is still a
considerable lack of knowledge about some of the finer details of steroid
interconversions (see Takemori and Kominami, 1984, for review). The enzyme
is sited in the endoplasmic reticulum of adrenals and testes. It is a cytochrome
P-450 and, therefore, requires molecular oxygen and NADPH for activity. The
molecular weight of the purified enzyme is 52 000.
The inhibition is presumably effected by the donation of the imidazole nitro-
gen lone pair of electrons to the haem iron (as in lanosterol 14a-demethylase,
Section 6.2.3). Sikka et al. (1985) find an Z,,, of 1 X lo-“M for ketoconazole
inhibition of the 17,20-lyase reaction. Ketoconazole appears to be relatively
unselective in its inhibition of dedicated sterol and steroid P-450s. Other drugs
are being developed to be selective for this enzyme system.
As mentioned in the previous section, the antifungal drug ketoconazole
produces an occasional side-effect in a few men taking the drug in that it
causes them to grow breast tissue (gynaecomastia). Testosterone levels are
reduced reversibly by the drug in a dose-related fashion, while at the same
time progesterone and 17a-hydroxyprogesterone accumulate. Since andro-
stenedione is reduced to form testosterone, these results can be explained by
an inhibition of the steroid 17,20-lyase (reviewed in Amery et al., 1986).It
appears that the gynaecomastia occurs in men whose 17/3-oestradiol levels
are high.
This unfortunate side-effect was, however, noted and put to good use in a
male condition where the reduction of androgen levels is the goal of therapy,
namely prostate cancer. The hormone stimulates the growth of the tumour.
Any reduction in androgen level, and possibly antagonism by increasing levels
of progesterone, is an advance on one of the previously used treatments -
castration!

6.3.2 ll@Steroid hydroxylase - metyrapone for

hypercortisolism
This enzyme catalyses the conversion of deoxycorticosterone to corticos-
terone, and deoxycortisol to cortisol (reaction 6 in Fig. 6.2). Like cholesterol
108 Molecular mechanisms of drug action

20,22-lyase, the enzyme is found in adrenal mitochondria and requires a power-

ful reducing agent containing iron-sulphur clusters like ferredoxin (Chapter
4) known as adrenodoxin, an enzyme to reduce it called adrenodoxin
reductase, NADPH and molecular oxygen to form an electron transport chain.
Metyrapone inhibits the enzyme competitively with a Ki of 1 X lo-‘M, and
shows a type II ultraviolet difference spectrum with adrenal mitochondria
(Williamson and O’Donnell, 1969). 1 I-P-Deoxycorticosterone shows a type I
difference spectrum characteristic of a substrate. Both substrate and metyra-
pone stabilize the enzyme. Furthermore, metyrapone induces a shift in the
electron paramagnetic resonance spectrum suggesting that the haem is
involved in the interaction (Wilson et al., 1969).

Metyrapone
Metyrapone has been used to treat the hypersecretion of cortisol by adrenal
tumours (reviewed by Temple and Liddell, 1970), but on long-term adminis-
tration the level of ACTH may rise to counteract the effect of the drug. The
use of metyrapone alone for this indication is no longer recommended, and
aminoglutethimide may be given in conjunction (Gold, 1979).

6.3.3 Aromatase - aminoglutethimide forhormone-dependent

cancers
Another major area of interest concerning the synthesis of steroid hormones is
the dependence of certain tumours on steroids for the maintenance of growth.
Testosterone appears to be essential for the growth of some tumours of the
prostate gland, while certain breast cancers are dependent on oestradiol. At
least one third of human breast cancers are hormone-dependent, and regress
when the sources of oestrogen are eliminated (Kiang et al., 1978). Inhibition
of steroid-producing enzymes is one way of achieving this.
Aminoglutethimide is an inhibitor of both cholesterol 20,22-lyase and aroma-
tase - the latter so-called because it introduces an aromatic ring into the steroid
nucleus (ring A, reaction 5 in Fig. 6.2) in catalysing the transformation of testos-
terone and androstenedione to 17-P-oestradiol and oestrone respectively.

NHz

Aminoglutethimide
 Steroid biosynthesis and action 109

The mechanism of removal of the C, methyl group shows a formal similarity

in chemical terms to the demethylation of lanosterol in that the reaction takes
place in three stages; first to -CH,OH, next -CHO, followed by total removal
of the fragment as formic acid and the introduction of a double bond into the
sterol skeleton. Conversion of the 3-keto to hydroxy produces a third double
bond to give aromatic@. This has prompted the synthesis of some mechanism-
based irreversible (suicide) inhibitors (Chapter 1; Covey et al., 1981).
It is not clear whether aromatase or cholesterol side-chain cleavage is the
key target for aminoglutethimide, although it inhibits aromatase at a lower
concentration than cholesterol 20,22-lyase (Ki for aromatase 6 X lo-’ M; testos-
terone concentration 1.5 X 1O-6 M; Ki for 20,22-lyase 4.5 X lo-” M, cholesterol
concentration as normally found in bovine adrenal mitochondria). Aromatase
may be the enzyme that is the more clinically relevant since the fall in oestro-
gens after dosing is immediate, unlike progesterone levels which initially rise
(Foster et al., 1983).
These inhibitory effects have promoted the study of aminoglutethimide as
a treatment for metastatic breast cancer, against which it is moderately effec-
tive (Smith et al., 1978). Aminoglutethimide shows a number of undesirable
side-effects and attempts to improve on the original structure have led to a
number of compounds being tested in the clinic. These structures contain
imidazole and are competitive inhibitors by virtue of binding to the haem iron.
One leading compound is fadrozole (CGS16949A) which has a Ki for aromatase
of 1.9 X 10Pio~ (Santen, 1991). Non-steroidal aromatase inhibitors are very
unlikely to have oestrogen agonist activity unlike tamoxifen and the other
oestrogen antagonists, and thus can increase the options for treatment for
breast cancer.

CN

Fadrozole

6.3.4 5a-Reductase
Androgens control the development of male sexual characteristics. Testoster-
one is the main hormone responsible for these changes but inside the cell it
is reduced to dihydrotestosterone which acts as the intracellular mediator. The
prostate gland in older men can increase in tissue size and interfere with the
110 Molecular mechanisms of drug action

flow of urine (benign hyperplasia - the cells are normal not cancerous hence
benign). This condition is hormone-dependent. Blockade of the reducing
enzyme, 5a-reductase, has been attempted to reduce the size of the prostate.

0
! - NHBu-t

Me

Me
/

0 LY+
!

Finasteride

Finasteride is an 4-azasteroid which is a potent competitive inhibitor of the

rat isoenzyme 1 with a Ki of 4 X 10p9~. It inhibits the human isoenzyme 1
much less powerfully with a Ki of 3 X ~O-‘M. Cloning and expression of the
two enzymes in human embryonic kidney cells has shown that there is a tetra-
peptide sequence (Val-Ser-Ile-Val) in the rat enzyme binding site (residues 22-
25 at the amino terminus) that is responsible for drug sensitivity. The anal-
ogous sequence in man is Ala-Val-Phe-Ala (residues 26-29) which is partially
resistant (Thigpen and Russell, 1992). Clearly there will be further attempts
to design a more potent human enzyme inhibitor.
Despite the weaker inhibition in man, finasteride reduces the size of the
prostate and increases the flow of urine. Levels of dihydrotestosterone in the
serum were greatly reduced but testosterone levels were unchanged
(testosterone can be metabolized to other products such as oestradiol)
(McConnell et al., 1992). There were side-effects such as decreased libido and
impotence which could be attributed to hormone interference but testoster-
one itself seems to be able to substitute for dihydrotestosterone in many other
tissues, thus limiting the side-effects experienced (Gormley et al., 1992).

6.4 Steroid receptor ligands

6.4.1 Oestrogen/progesterone receptor agonists - oral

contraceptives
Oestrogen and progesterone receptors are present in the nucleus - unlike the
G-protein-linked receptors found in the cell membrane (Chapter 9). Corticos-
teroid receptors may be cytoplasmic with a connection to the nucleus
(Chapter 7). Ligands cross the cell membrane by passive diffusion and bind
to the steroid receptors; the complex acts as a transcription factor by activating
 Steroid biosynthesis and action 111

particular sections of the genome to produce specific mRNA. General RNA

synthesis occurs later followed by DNA biosynthesis (Murad and Kuret, 1990).
Recent cloning studies have greatly increased our understanding of the struc-
ture of these receptors but we still do not know entirely how these processes
function (reviewed in Gronemeyer, 1991; Bailieu et al., 1990).
There are seven regions in steroid receptors, labelled A to F except the
progesterone receptor which does not have F. Regions C and E are highly
conserved and contain the DNA and hormone-binding domains respectively.
At rest these receptors are bound by a 90 kDa protein called a heat-shock
protein, so-called because its synthesis is greatly stimulated when the cell’s
temperature is raised to about 42°C. The heat-shock protein binds to the DNA
binding domain and produces an oligomer with sedimentation coefficient of
8s. Another protein of molecular weight 59 kDa is also found in the 8’S recep-
tor-heat-shock protein complex.
When the ligand binds, the oligomer dissociates as a result of a confor-
mational change and the drug-receptor complex is now 4s. There are two
molecules of oestrogen receptor but probably only one of progesterone recep-
tor complexed with the heat-shock protein. The heat-shock protein thus acts
to repress the steroid signal and probably also to stab&e the receptor protein
while not needed (Bailieu et al., 1990).
The oestrogen and progesterone receptors bind to DNA as dimers. In the
C region of the receptor the protein folds into two zinc fingers, so-called
because each is produced by a zinc coordinated to four cysteine residues. The
zinc fmgers are essential for DNA binding, while the hormone binding domain
and other areas are required to stabilize the binding to DNA. The promoter
region of the hormone-inducible gene contains a sequence some 13 or 15
nucleotides long depending on the specific receptor; the sequence is palin-
dromic i.e. the five bases at each end are complementary. The sequence of
the oestrogen-responsive element is 5’-GGTCAXXXTGACC, where X can be
any base. Transcription also requires the presence of two regions in the recep-
tor known as transcription activation factors, one of which is in the N-terminal
region, but the precise site of these areas is not known (Gronemeyer, 1991).
The progesterone receptor exists in two isoforms, named A and B, which
result from the initiation of transcription from the same gene but two different
sites. The two forms differ by some 130 amino acids in length at the N-ter-
minus, with the A form being the smaller (Gronemeyer, 1991). There is
another short eight amino acid stretch with more than four cationic residues
in the centre of the receptor which appears to control nuclear siting.
One of the major uses of pharmaceutical products is of a prophylactic nat-
ure - the avoidance of the condition of pregnancy. Both oestrogen and pro-
gesterone receptor agonists are used for this, usually in conjunction, although
progestins can be used alone if the patient is believed likely to suffer side-
effects from oestrogen treatment. The rationale for the use of such agents is
that they inhibit ovulation, with oestrogen mainly responsible for suppressing
follicle-stimulating hormone (ESH) secretion, while the progestin suppresses
 112 Molecular mechanisms of drug action

luteinizing hormone (LH) secretion (Briggs et al., 1970) O;SH and LH are
known as gonadotrophins). Even if ovulation and subsequent fertilization did
occur, it is very unlikely that the fertilized egg could implant into the lining
of the womb. Oestrogen produces a cornitied or keratinized layer on the sur-
face lining of the womb to promote adhesion of the egg. Progesterone antagon-
izes this and, in addition, produces a very viscous mucus layer through which
the egg would fmd it difficult to penetrate.
The oral contraceptives inhibit the production of FSH and LH by binding to
receptors in the cytosol of hypothalamus and pituitary (Fig. 6.6). The effect
on the former site suppresses the release of gonadotrophin releasing hormone
(GnRH) while the effect on the latter is to reduce gonadotrophin release. This
is a negative feedback loop that the endogeneous hormone controls, and is a
necessary part of pregnancy (Schally, 1978). It is interesting in this context
that the oral contraceptives do not trigger the positive feedback at the level
of the pituitary whereby oestrogen together with GnRH stimulates the pre-
ovulatory surge of LH (Asch et al., 1983 - see Fig. 6.6). The negative feedback
closes the cycle since FSH stimulates the granulosa cells of the follicle to prod-
uce increasing amounts of oestrogen as the follicle develops. Following ovu-
lation the granulosa cells differentiate into luteal cells which secrete progester-
one under the influence of LH in the later part of the cycle (known as the
luteal phase) until the corpus luteum begins to degenerate (Mm-ad and Kuret,
1990). Both peptide hormones, LH and FSH, act through adenylate cyclase,

Hypothalamus

luteum

Oestrogen Progesterone

+ indicates a positive feedback

- indicates a negative feedback

The hypothalamus releases GnRH which acts on the pituitary to produce both FSH and LH.
FSH stimulates follicles to grow and produce oestrogen which has a negative feedback on
LH and FSH release. Eventually, however, an oestrogen level is reached that is sufficient,
acting in concert with GnRH, to release LH and FSH in a surge from the pituitary. This
converts one follicle to a corpus luteum, synthesizing progesterone that inhibits LH
production. Only the negative feedback to the pituitary is shown for the sake of clarity.

Figure 6.6 Hypothalamus - Pituitary - Gonad axis.

 Steroid biosynthesis and action 113

after binding to their specific receptors, in a similar fashion to ACTH to

stimulate cholesterol 20,22lyase, cyclic-AMP levels and steroidogenesis
(Punkenstein et al., 1983; and references therein).
Two of the most frequently used oestrogens are ethinyloestradiol and its O-
methyl analogue, mestranol. The former binds to rat anterior pituitary and
hypothalamic cytosol with a potency similar to that of oestradiol, with a Kd
from 1 to 1.3 X 10pl”~ compared with 2.2 to 3.2 X lo-“‘M for oestradiol.
Mestranol, on the other hand, is bound much less tightly, with a Kd of 2.9 to
5.5 X 10p8~, as shown by competitive binding with oestradiol. It has been
suggested that mestranol is dealkylated to the more active ethinyloestradiol by
liver mixed-function oxidases - as occurs in the rat (Eisenfeld, 1974).

R = CH,, Mestranol
R = H, Ethinyloestradiol

After the drug-receptor complex is formed it is converted to a different

species by a conformational change and translocated into the nucleus. Specific
mRNA and certain proteins are synthesized initially, followed by a general
increase in the synthesis of RNA. Specific DNA biosynthesis is a much later
event (Murad and Kuret, 1990). How this process actually interferes with the
secretion of GnRH is not known at present.
Norethisterone and norgestrel are typical examples of the most widely used
progestins. They bind to receptors in the cytosol of human uterus cells with
binding constants of 6.8 X 10e9~ and 2.4 X 10-9~ respectively (Kasid et al.,
1978; Srivastava, 1978). Both compounds compete with progesterone for these
binding sites.

Diethylstilboestrol

R = CH3, Norethisterone
R = CzH5, Norgestrel

If an oral contraceptive is required after intercourse, diethylstilboestrol, a

non-steroid oestrogen is often used. Oestrogens have uses other than oral con-
114 Molecular mechanisms of drug action

traceptives including hormone replacement therapy during the menopause

(Murad and Kuret, 1990).

6.4.2 Oestrogen antagonists - tamoxifen for oestrogen-

dependent cancer, clomiphene to stimulate ovulation
Breast cancers may be divided into two distinct types: oestrogen-receptor-rich
and receptor-poor. The receptor-rich group (about 50 per cent of pre-meno-
pausal women and 75 per cent of post-menopausal women) depends on the
hormone for growth. Oestrogen antagonism has been found to be very effec-
tive in a large proportion, particularly post-menopausal, of this group. Surgery
is the treatment of choice for the pre-menopausal group. In the case of the
receptor-poor group, hormone therapy alone is unlikely to be of much value,
but the use of oestrogen antagonists as part of a cocktail of drugs including
cytotoxic agents has recently been developed (Jordan, 1992).
The oestrogen antagonist of choice is tamoxifen, which is a potent ligand
of oestrogen receptors, with a Kd of cytosolic receptors from human mammary
tumours of 3.7 X 10P9~ compared with 1.2 X lo-‘OM for 17poestradiol by
direct measurement (Nicholson et al., 1979). A figure of one to two orders of
magnitude less activity is obtained in competitive binding studies, possibly due
to a breakdown of the drug-receptor complex. Using a cell line of human
breast cancer cells (MCF 7) Horwitz et al. (1978) found that the action of
tamoxifen was biphasic in that at low extracellular levels (lo-‘M) the drug
showed agonist properties in stimulating growth and inducing progesterone
receptors. Higher doses (10P6~) were antagonistic. Interference with oestra-
diol may take the form of blocking the binding of oestradiol to its receptors,
and of reducing the number of receptors available for the natural hormone.
Consequently, the natural hormone is unable to maintain tumour growth.

Tamoxifen Clomiphene

A close analogue of tamoxifen, clomiphene, has found favour as an agent

for the induction of fertility in women who either have no menstrual cycle
or cycles in which they do not ovulate (Marshall, 1978). The most obvious
physiological effect of this drug is the enlargement of the ovaries, and this
hyperstimulation does on occasion give rise to multiple births - approximately
6 to 8 per cent compared with 25 per cent of births when gonadotrophins
 Steroid biosynthesis and action 115

are given to induce fertility. Clomiphene at high dose is effective for the palli-
ation of breast cancer but tamoxifen is preferred for this indication because
of lower toxicity (Marshall, 1978).
Clomiphene acts as an anti-oestrogen showing side-effects consistent with
this action, including hot flushes. The major pharmacological effect is to pre-
vent the normal feedback inhibition of oestrogen on the release of GnRH from
the hypothalamus (Fig. 6.6). Higher GnRH levels then release more luteinizing
hormone and follicle-stimulating hormone from the pituitary and ovulation
is stimulated.
The binding constant of clomiphene to various rat brain and uterine recep-
tors is in the region of lo-s111 compared with 10P’OM for 17/3-oestradiol under
the same conditions. Unlike the studies with tamoxifen, direct binding and
competitive experiments gave similar results (Ginsberg et al., 1977). This fam-
ily of anti-oestrogens show activity as antagonists in the tram isomer and as
agonists in the cis. Tamoxifen is marketed as the tram isomer while clomi-
phene is a racemic mixture.

6.4.3 Progesterone antagonist - mifepristone as abortifacient

In addition to inhibition of progesterone biosynthesis as an approach to contra-
ception and abortion, receptor antagonism has been found to be effective.
Mifepristone (RU-486 or RU-38486) has been developed as a ligand of pro-
gesterone receptors and is apparently considerably more potent than pro-
gesterone itself (Schreiber et al., 1983). Mifepristone also binds to glucocort-
icoid receptors but at markedly higher concentration. The drug also binds to

OH

Mifepristone

receptors in the hypothalamus which leads to a fall in GnRH secretion and

subsequently to luteolysis through a fall in LH levels. A marked fall in progester-
one secretion is the result.
Mifepristone appears to antagonize progesterone in two ways. The drug
binds very tightly to the B isoform of the progesterone receptor in the 8s
complex and, by slowing the release of the heat-shock protein, mifepristone
prevents progesterone binding. In addition, the drug receptor complex will
still bind to DNA but fails to activate transcription (Gronemeyer et al., 1992).
Information about the mifepristone binding site has come from a study of
116 Molecular mechanisms of drug action

chicken progesterone receptors which cannot bind mifepristone because

there is a cysteine instead of a glycine at the critical position 575. The region
containing the glycine lies in a so-called 1 lppocket of the receptor, i.e. it lies
near progesterone’s 11 p position. The bulky 11 Pdimethylaminophenyl group
of mifepristone fits very well into this pocket provided position 575 does not
carry a blocking side-chain.
Mifepristone can be used either as a contraceptive if taken 72 hours after
intercourse, although it may also work effectively after a longer time. In
addition, the drug is licensed for use in conjunction with a prostaglandin to
achieve abortion in up to nine weeks of pregnancy and is proving a realistic
alternative to surgery, particularly in developing countries where surgical abor-
tion carries a much greater risk (Henshaw and Templeton, 1992).

6.4.4 Androgen antagonist

Like the majority of breast cancers, prostate cancer is dependent on hormones,

viz. androgens such as testosterone or androsterone, to maintain its growth.
As noted earlier for oestrogen antagonism, the antagonism of androgens might
be expected to inhibit the growth of prostate cancer.

Nilutamide

NH -- ! Pr-i

NOz
Flutamide

Flutamide and nilutamide have been developed for this indication. The for-
mer is a pro-drug for the 2-hydroxylated metabolite which has binding con-
stants in the range of 0.8 to 2.05 X 10m8~ to androgen receptors from various
 Steroid biosynthesis and action 117

genital organs including human prostate cancer (Simard et al., 1986). Fluta-
mide is often used in conjunction with luteinizing hormone releasing hormone
agonists such as leuprolide, because flutamide blocks the transient increase of
androgen release by leuprolide before down-regulation desensitizes the pitu-
itary (Chapter 12). The efficacy of such combinations over single agent therapy
is not yet proven (Brogden and Clissold, 1989).
It is interesting that both drugs contain an aromatic nitro group which in
nilutamide may be reduced by NADPH and microsomal fraction, first by one-
electron reduction to a radical ion, and then by further reduction to reactive
products which become covalently attached to cellular protein (Berger et al.,
1992). The nitroimidazole, metronidazole, reacts in a similar fashion
(Chapter 4).

6.4.5 Aldosterone antagonist- spironolactone as diuretic

Aldosterone is the steroid (known as a mineralocorticoid because it helps to
control the levels of minerals in the blood plasma) which controls reabsorption
of sodium ion in the distal portion of the kidney tubule (Fig. 6.7). Up to 90

Bowman’s
capsule
Blood
out

Proximal

Descending
limb

Thin

limb

Loop of Henle

Shaded area represents the site of action of

aldosterone antagonists.

Figure 6.7 Site of aldosterone and spironolactone action

118 Molecular mechanisms of drug action

per cent of the sodium is reabsorbed with chloride as the counter-ion, but a
portion appears to be linked to the excretion of potassium or hydrogen ion
thereby maintaining electrical neutrality. Water is reabsorbed with the sodium
ion; the net effect is an increase in plasma volume and fall in plasma potassium.
Aldosterone is synthesized in the zona glomerulosa of the adrenal cortex.
Like the other steroids mentioned in this chapter, the hormone functions by
crossing the cell membrane and binding to a cytoplasmic receptor with a Kd
of 2 X 10P9~ (Funder et al., 1974). Two sets of binding sites in the cortex
have been proposed by Farman et al. (1982) who obtained a biphasic Scatch-
ard plot (see Appendix) for the binding of hormone to preparations from rab-
bit kidney cortex. Although such a plot does not always imply two sets of
sites, and insufficient data points were obtained to reach a definite conclusion
about the absolute values of the binding constant, these authors proposed that
the higher affinity binding sites were for mineralocorticoids and the lower
for glucocorticoids.

CH2 OH
OHC ;=O

Aldosterone

The steroid-receptor complex undergoes a conformational change before

it is translocated into the cell nucleus, where it binds to the DNA thus causing
specific mRNAs to be synthesized (Corvol et al., 1981). Several proteins in the
cytosol and plasma membrane of the kidney cell are synthesized in response
to aldosterone action - a process which is blocked by cycloheximide - but
the precise role each plays in the facilitation of sodium reabsorption is not yet
known (Scott et al., 1981).
Steroid antagonists of aldosterone, of which the best known is spironolac-
tone, are valuable diuretics in that not only do they antagonize the anti-diuretic
action of the steroid by preventing sodium reabsorption, but also they do not
cause a loss of potassium (this is referred to as potassium sparing action), an
action not shared by the other types of diuretic. This is because aldosterone
induces a low but measurable exchange of potassium for sodium noted above.
Spironolactone and its analogues bind to the aldosterone receptor; the com-
plex so formed does not rearrange into the active conformation and therefore
cannot translocate into the nucleus (Funder et al., 1974; Sakauye and Feldman,
1976). These drugs thus act in a fashion different from tamoxifen and its com-
plex with the oestrogen receptor (section 6.4.3).
The binding constant for spironolactone is approximately 4 X lo-” M when
 Steroid biosynthesis and action 119

Spironolactone

measured competitively against 3H-aldosterone binding to rat kidney (Farman

et al., 1974). The correlation of binding to the receptor and physiological
effect has been studied by Rossier et al. (1983) who showed that receptor
binding in toad bladder correlated with inhibition of sodium transport across
the membrane induced by aldosterone. Aldosterone has to be present for spi-
ronolactone to be active, i.e. the latter is not a partial agonist, and the two
compounds compete for the receptor (Rossier et al., 1983; Horisberger and
Giebisch, 1987).
Spirolactones, as a group, are also able to inhibit the biosynthesis of aldos-
terone but at rather higher concentrations than are needed to interfere with
aldosterone receptor binding. The conclusion is, therefore, that the inhibition
of biosynthesis plays little or no part in the antidiuretic action of the drugs
(Corvol et al., 1981).
Spironolactone is not, however, selective for aldosterone receptors, and is
well able to block androgen receptors, albeit at higher doses. This is clearly
a disadvantage when a diuretic effect only is required, because it can lead to
the growth of breasts in men (gynaecomastia) and to menstrual disturbances
in women. Nevertheless, in some conditions, notably hirsutism in women
caused by high androgen levels, an anti-androgen effect is a positive advantage.
The drug has been of clinical value in this condition (Shapiro and Evron, 1980).

Questions

1. Which protein is complexed with the oestrogen receptor in its inactive

state?
2. Name three enzymes described in this chapter that use cytochrome P-450
as a prosthetic group.
3. Which enzymes are rate-limiting in the synthesis of cholesterol?
4. The plasma level of which lipoprotein is linked with heart disease? How
does lovastatin interfere with its action?
5. Name one tumour that is hormone-dependent. Name a drug that can be
used to treat this condition. How does the drug prevent the tumour growth?
6. Which enzyme does fluconazole block? What condition is fluconazole used
to treat?
120 Molecular mechanisms of drug action

7. Why is terbintine more effective against Trichophyton than Candida spec-

ies?

References
Alberts, A.W., Chen, J., Kuron, G., Hunt, V., Huff, V., Hoffman, C., Rothbrook, J., Lopez,
M., Joshua, H., Harris, E., Pitchett, A., Monaghan, R., Currie, S., Stapley, E., Albers-
Schonberg, G., Hersens, O., Hi&field, J., Hoogsteen, K., Liesch, J. and Springer,
J., 1980., Proc. Nat. Acad. Sci. (lIS.A.), 77, 3957-61.
Amery, W.K., De Coster, R. and Caers, I., 1986, Drug. Dev. Res., 8, 299-307.
Angeli, A. and Frairia, R., 1985, Lancet, i, 821.
Anon, 1986, Lancet, ii, 665-6.
Aoyama, Y., Yoshida, Y., Sonoda, Y. and Sato, R., 1987,J. Biol. Cbem., 262, 1239-43.
Asch, R.H., Balmaceda, J.P., Borghi, K.R., Niesvisky, R., Coy, D.H. and Schally, A.V.,
1983, J. Clin. Endocrinol. Metab., 57, 367-72.
Bailieu, E.E., Binart, N., Cadepond, F., Catelli, M.G., Chambraud, B., Gamier, J., Gasc,
J.M., Groyer-Schweizer, G., Oblin, M.E., Radanyi, C., RedeuiIh, G., Renoir, J.M. and
Sabbah, M., 1990, Ann. N. Y. Acad. Sci., 595, 300-15.
Balfour, J.A. and Faulds, D., 1992, Drugs, 43, 259-84.
Barrett-Bee, KJ., Lane, A.C. and Turner, R.W., 1986, J. Med. Vet. MycoZ., 24, 155-60.
Berger, V., Barson, A., Wolf, C., Chachaty, C., Fan, D., Fromenty, B. and Pessayre, D.,
1992, Biochem. Pharmacol., 43, 654-7.
Bloch, K.E., 1979, C.R.C. Crit. Rev. Biochem., 7, l-5.
Bradford, R.H., et al., Amer.J Med., 91, Suppl lB, 18S-23s.
Briggs, M.H., Pitchford, A.G., Stamford, M., Barker, H.M. and Taylor, D., 1970, Adv.
Steroid. Biochem. Pbarmacol., 2, 11 I-222.
Brogden, R.N. and Clissold, S.P., 1989, Drugs, 38, 185-203.
Chen, H.W. and Leonhardt, D.A., 1984,J. Biol. Cbem., 259, 8156-62.
Corvol, P., Claire, M., Oblin, M.E., Geering, K. and Rossier, B., 1981, Kidney Znt., 20,
1-6.
Coulson, C.J., King, D.J. and Wiseman, A., 1984, Trends in Biocbem. Sci., 9, 446-9.
Covey, D.F., Hood, W.F. and Parikh, V.D., 1981, J. Biol. Cbem., 256, 1076-9.
Eisenfeld, A., 1974, Endocrinology, 94, 803-7.
Farman, N., Vandewalle, A. and Bonvallet, J.P., 1982, AmJ Pbysiol., 242, F69-77.
Foster, A.B., Jarman, M., Leung, C.-S., Rowlands, M.G. and Taylor, G.N., 1983, J. Med.
Cbem., 26, 50-4.
Funder, J.W., Feldman, D., Highland, E. and Edelman, T.S., 1974, Biocbem. PbarmacoL,
23, 1493-501.
Funkenstein, B., Waterman, M.R., Masters, B.S.S. and Simpson, E.R., 1983, J. Biol.
Cbem., 243, 10187-9.
Garnier, J., Gasc, J.M., Groyer-Schweizer, G., Oblin, M.E., Radanyi, C., Redeuilh, G.,
Renoir, J.M. and Sabbah, M., 1990, Ann N. Y Acad. Sci., 595, 300- 15.
Georgopapadakou, N.H. and Bertasso, A., 1992, Antimicrob. Ag. Cbemotber., 36,
1779-81.
Gibbons, G.F., Pullinger, C.R. and Mitropoulos, K.A., 1979, Biocbem. J., 183, 309- 15.
Ginsburg, M., Maclusky, N.J., Morris, I.D. and Thomas, P.J., 1977, Brit. J. Pbarmacol.,
59, 397-402.
Gold, E.M., 1979, Ann. Zntern. Med., 90, 829-44.
Gormley, G.J., Stoner, E. and Bruskewitz, R.C., et al., 1992, N. Engl. J. Med., 327,
1185-91.
Grant, S.M. and Clissold, S.P., 1990, Drugs, 39, 877-916.
Gronemeyer, H., 1991, Annu. Rev. Genet., 25, 89-123.
 Steroid biosynthesis and action 121

Gronemeyer, H., Benhamou, B., Berry, M., Bocquel, M.T., Gofflo, D., Garcia, T., Ler-
ouge, T., Metzger, D., Meyer, M.E., Tora, L., Vergezac, A. and Chambon, P., 1992,
J. Steroid Biocbem. Molec. Biol., 3-8, 217-21.
Grundy, S.M., 1988, New Engl. J. Med., 313, 24-33.
Hay, R.J., 1991,J. Antimicrob. Cbemotber., 28, Suppl A, 35-46.
Henshaw, R.C. and Templeton, A.A., 1992, Drugs, 44, 531-6.
Horisberger, J.-D. and Giebisch, G., 1987, Renal Pbysiol., 10, 198-220.
Horwitz, K.B., Koseki, Y. and McGuire, W.L., 1978, Endocrinology, 103, 1742-51.
Jordan, V.C., 1992, Cancer, 70, 977-82.
Kasid, A., Buckshee, K., Hingorami, V. and Laumas, K.R., 1978, Biocbem. J, 176,
531-9.
Kiang, D.T., Frenning, D.H., Goldman, AI., Ascensao, V.F. and Kennedy, B.J., 1978,
New Eng1.J Med., 277, 1330-4.
Lewis, J.H., Zimmerman, H.J., Benson, G.D. and Ishak, K.G., 1984, Gastroenterology,
86, 503-13.
Loose, D., Kan, P.B., Hirst, A., Marcus, R.A. and Feldman, D., 1983,J. Clin. Invest., 71,
1495-9.
Marshall, J.R., 1978, Clin. Obstet. Gynaecol., 21, 147-62.
McConnell, J.D., Wilson, J.D., George, F.W., GeIIer, J., Pappas, R. and Stoner, E., 1992,
J. Clin. Endocrinol. Metab., 74, 505-8.
Murad, F. and Kuret, J.A., 1990, Ch. 58 in the Pharmacological Basis of Iberapeutics,
(eds A.G. GiIman, T.W. Rall, A.S. Nies and P. Taylor), Pergamon, New York, 8th
edn.
Nes, W.R., Sekula, B.C., Nes, W.D. and Adler, J.H., 1978,J. Biol. Cbem., 253, 6218-25.
Nicholson, RI., Syne, J.S., Daniel, C.P. and Griffiths, K., 1979, Eur. J Cancer, 15,
317-29.
Ono, T., Nakazone, K. and Kosaka, H., 1982, Biocbim. Biopbys. Acta., 709, 84-90.
Petranyi, G., Ryder, N.S. and Stutz, A., 1984, Science, 224, 1239-41.
Pont, A., Williams, P.L., Azher, S., Reitz, R.E., Rochra, C., Smith, E.R. and Stevens, D.A.,
1982, Arch. Intern. Med., 142, 2137.
Ravnskov, U., 1992, Brit. Med.J, 305, 15-19.
Rossier, B.C., Claire, M., Rafestin-Oblin, M.E., Geering, K., Gaggeler, H.P. and Corvol,
P., 1983, Am.J Pbysiol., 244, C24-31.
Rotstein, D.M., Kertesz, D.J., Walker, A.M. and Swinney, D.C., 1992, J. Med. Cbem.,
35, 2818-25.
Ryder, N.S. and DuPont, M.-C., 1984, Biocbim. Biopbys. Acta, 774, 466-71.
Sakauye, C. and Feldman, D., 1976, Am. J. Pbysiol., 231, 93-7.
Santen, R.J., 1991,J Steroid Biocbem. Molec. Biol., 40, 247-53.
SchaIly, A.V., 1978, Science, 202, 18-28.
Schreiber, J.R., Hsueh, A.J.W. and Barlieu, E.B., 1983, Contraception, 28, 77-85.
Scott, W.N., Yang, C.P., Skipski, LA., Cobb, M.H., Reich, I.M. and Terry, P.M., 1981,
Ann. N. Y Acad. Sci., 372, 15-28.
Shapiro, G. and Evron, S., 1980,J. Clin. Endocrinol. Metab., 51, 429-32.
Sikka, S.C., Swerdloff, R.S. and Rajfer, J., 1985, Endocrinology, 116, 1920-5.
Smith, G.D. and Pekkanen, J., 1992, Brit. Med. J., 304, 431-4.
Smith, I.E., Fitzharris, B.M., McKinna, J.A., Fahmy, D.R., Nash, A.G., NeviUe, A.M., Gazet,
J.-C., Ford, H.T. and Powles, TJ., 1978, Lancet, ii, 646-8.
Smith, J.R., Osborne, T.F., Brown, MS., Goldstein, J.L. and GiI, G., 1988,J. Biol. Cbem.,
263, 18480-7.
Srivastava, A.K., 1978,J. Steroid Biocbem., 7, 1241-4.
Takemori, S. and Kominami, S., 1984, Trends in Biocbem. Sci., 7, 393-6.
Temple, T.E. and Liddle, G.W., 1970, Annu. Rev. Pbarmacol. Toxicol., 10, 199-218.
Thigpen, A.E. and Russell, D.W., 1992, J. Biol. Cbem., 267, 8577-83.
Thomas, P.G., Haslam, J.M. and Baldwin, B.C., 1983, Biocbem. Sot. Trans., 11, 713.
122 Molecular mechanisms of drug action

Uauy, R., Vega, G.L., Grundy, S.M. and Bilheimer, D.M., 1988,J. Pediatr., 113, 387-92.
Van den Bossche, H., Willemsens, G., Cools, W., Cornelissen, F., Lauwers, W.F. and
Van Cutsem, J.M., 1980, Antimicrob. Ag. Cbemotber., 17, 922-8.
Van den Bossche, H., Lauwers, W., Willemsens, G., Marischal, P., Cornelissen, P. and
Cools, W., 1984, Pesticide Sci., 15, 188-98.
Williamson, D.G. and O’Donnell, V.J., 1969, Biochemistry, 8, 1311-16.
Wilson, L.D., Oldham, S.B. and Harding, B.W., 1969, Biochemistry, 8, 2975-80.
Wissler, R.W., 1991, Amer.J U. Med., 91, Suppl lB, 3S-9s.
 Chapter 7

Prostaglandin and leukotriene

biosynthesis and action

7. I Introduction

The transformation of arachidonic acid into a multiplicity of lipid intermedi-

ates, many of which are extremely active biologically, has been an area of
great interest in recent years. The prostaglandins themselves were first isolated
and identified by Samuelson and co-workers in the 1960s but the other agents
with quaint, albeit accurate, names like ‘rabbit aorta contracting substance’
(thromboxane AZ) and ‘slow reacting substance of anaphylaxis’ (leukotrienes
C4 and DJ defied identification until the early 1980s largely because of their
instability and low concentrations in biological fluids. The development of
sensitive mass spectroscopic and gas chromatographic methods for their iso-
lation was crucial in their identification.
The details of arachidonic acid metabolism have been elucidated at a remark-
able rate over recent years. The present outline of the main pathways is shown
in Fig. 7.1. The metabolism of arachidonic acid can give rise to four distinct
pathways: the cyclooxygenase reaction leads to the prostaglandins D,, E,, F,,
G2 and H,, the thromboxanes and prostacyclin; 5-lipoxygenase leads to the
leukotrienes (Fig. 7.2); a cytochrome P-450 linked reaction gives rise to the
eicosatrienoic acids, and 12lipoxygenase yields another family of hydroxyeico-
satrienoic acids (Taylor and Ritter, 1986). It is the first two pathways with
which this chapter is primarily concerned.

Table 7.1 Drugs discussed in Chapter 7

Target Drug Major uses

PhosphoIipase A, Corticosteroids Asthma, arthritis

Prostaglandin synthetase Aspirin, Pain, inflammation in arthritis
Indomethacin

123
124 Molecular mechanisms of drug action

Prostaglandin biosynthesis

Phospholipid

Arachidonic acid s

Cyclooxygenaselliydroperoxidase
\

%:;GH
Prostaglandin GZ

OOH

l;;GOH
Prostaglandin Hz

OH

Thromboxane synthase

Prostacyclin Thromboxane B2

Figure 7.1 Prostaglandin biosynthesis.

7.2 Phospholipase inhibition - glucocorticoids for asthma

and arthritis

Corticosteroids have long been the main therapy for the treatment of severe
asthma and other allergic conditions. With the advent of aerosol systems that
Prostaglandin and leukotriene biosynthesis and action

Leukotriene biosynthesis

COOH

Arachidonic acid

5-S-Hydroperoxyeicosatetraenoic acid

COOH

Leukotriene A,
\ Leukotriene A, hydrolase

Leukotriene C,
synthase i

COOH

&-Glutamvl transferase

COOH

S
H
f-W
i,
0 NYCooH
Leukotriene Cl4 &wDtidase
- -
b OH
-
- COOH

H2N ‘:OOH
Leukotriene E4

Figure 7.2 Leukotriene biosynthesis.

126 Molecular mechanisms of drug action

deliver the drug direct to the lung, however, steroids can also be used for less
severe asthma without the risk of the side-effects, sometimes severe, that occur
with oral treatment. Control of symptoms can often be more effective than
with cromoglycate or the pagonists.
The initial step in prostaglandin and leukotriene biosynthesis is the release
of arachidonic acid from phospholipids catalysed mainly by phospholipase A,.
The enzyme hydrolyses the fatty acyl group from position 2 of the phospho-
lipid. Phospholipase A, is located on the inside of the plasma membrane and
requires calcium ion for activity - but not calmodulin (Withnall et al., 1984).
Corticosteroids inhibit phospholipase A, and thus block the synthesis of
pro-inflammatory prostaglandin metabolites deriving from the prostaglandin
synthetase and lipoxygenase pathways. Steroids switch on the gene that con-
trols the synthesis of a protein called lipocortin, a protein of molecular mass
15 to 40 kDa depending on the cell type, that completely inhibits phospho-
lipase A,. Accordingly, the anti-inflammatory effects of steroids can be blocked
by inhibitors of nucleic acid synthesis (see Sautebin et al., 1992).

CH2 OH
I

Cortisol

Lipocortin is now recognized as a member of a larger family of calcium

binding proteins known as annexins. These proteins when activated by cal-
cium can bind lipids and, consequently, draw two membranes together, sug-
gesting that they may be involved in exocytosis (reviewed in Creutz, 1992).
The N-terminal side is on the cytoplasmic face which contains tyrosine and
serine residues as potential phosphorylation sites, the tyrosine being the target
for protein kinase C. Phosphorylation reduces the ability of lipocortin to aggre-
gate chromaffin granules (Wang and Creutz, 1992). In lipocortin there are four
stretches or folds of 70 repeated amino acids, 17 of which are highly conserved
between lipocortins. The calcium binding sites, three of high and two of low
affinity, are on one side of the approximately planar molecule, which is also
the membrane binding face. Each high affinity site involves the first few amino
acids of the fold and an acidic residue in a second loop downstream.
The human lipocortin gene has been cloned and sequenced, and a glucocort-
icoid response element detected in the first intron similar to the hormone
response elements in other steroid-induced genes. These hormone response
elements respond to the transactivating function of the steroid receptor and
initiate transcription of the gene (Chapter 6; Kovacic et al., 1991).
 Prostaglandin and leukotriene biosynthesis and action 127

The glucocorticoid receptor is a multimeric complex located in the cyto-

plasm containing heat-shock protein. In unstressed cells heat-shock proteins
exist as monomers with no DNA binding ability. When the cell is stressed by
heat or other stress, they assemble in the nucleus in the form of trimers that
activate responsive genes (reviewed in Morimoto 1993). The receptor mol-
ecule is bound to heat-shock proteins of molecular weight 90, 70 and 56 kDa
in a complex that does not bind to DNA. When the steroid binds, a dimer of
the 90 kDa protein dissociates from the receptor and the receptor has a great
affinity for DNA in that the DNA binding site is now unmasked. The activated
receptor dime&es either before or at the same time as it binds to DNA. The
receptor molecule can be viewed as having three sections; an N-terminal
region which is involved in transactivation and dimerization, a DNA binding
section and a third domain carrying the steroid and 90 kDa protein binding
sites and also involved in transactivation (reviewed in Dahlman-Wright et
al., 1992).
The receptor initiates glucocorticoid responses in vivo by turning on a num-
ber of genes through binding at the response element:
AGAACAnnnTGTTCT
where A, G, T, C are the respective nucleotide bases and n can be any of the
four. This sequence is sometimes referred to as a partial palindrome, i.e. the
two stretches of six amino acids produce an imperfect mirror-image in DNA
base-pairing terms.
General DNA binding is mediated by the middle domain of the receptor
containing a section 66 to 68 amino acids long, highly conserved between
tissues. There are nine conserved cysteines, of which eight are involved in the
tetrahedral coordination of the two zinc atoms consistent with cysteine ligands
for structural zinc (see Chapter 8). The zinc atoms are located at the base of
loops or ‘zinc fingers’ that protrude from the main structure of the protein.
The zincs are apparently essential for structural integrity and DNA binding
capacity.

7.3 Prostaglandin synthetase inhibitors - aspirin and other

non-steroidal anti-infammatories

One of the oldest medications still in use is aspirin or acetylsalicylic acid. The
stimulus for the derivation of aspirin was the finding that the bark of the wil-
low tree had a soothing effect on headaches and fevers. The active ingredient
turned out to be a glycoside of salicyl alcohol, and subsequently salicylic acid
was also found to be effective in reducing temperature (antipyretic action).
Acetylation of the phenolic hydroxyl was the next step to yield aspirin (Flower
et al., 1985).
A very interesting advance in our understanding of pharmacology arose from
the realization that aspirin and other similar drugs, classified as the non-ster-
128 Molecular mechanisms of drug action

oidal anti-inflammatory agents, inhibited the synthesis of prostaglandins at the

step catalysed by prostaglandin endoperoxide synthetase (Vane, 1971). There
are a considerable number of these drugs which can be broadly grouped into
structural families: salicylates such as aspirin, substituted propionic acids like
naproxen and flurbiprofen, pyrazolones such as phenylbutazone, fenamates
like meclofenamic acid, together with indomethacin which does not readily
fall into any of the above structural groupings. Nevertheless, they all show
similar pharmacological activity: anti-pyretic, analgesic and anti-intlammatory
properties.
Prostaglandin synthetase is a homodimer with a molecular weight of

Me

Ph
F

Flurbiprofen

lndomethacin Acetylsalicylic acid

J-4
CH&Hz)3 0

Phenylbutazone
 Prostaglandin and leukotriene biosynthesis and action 129

144 kDa; it is a dioxygenase containing haem as an essential cofactor, mem-

brane-bound in the microsomes and particularly rich in the vesicular gland. It
exhibits two activities: cyclooxygenase and hydroperoxidase. Cyclooxygenase
adds both atoms of a molecule of oxygen to the carbon atoms at positions
11 and 15 of the arachidonic acid molecule to form a 15-hydroperoxy-9,1 l-
endoperoxide (prostaglandin G2) followed by reduction of the hydroperoxy
group to hydroxy by a hydroperoxidase activity yielding prostaglandin H, (Fig.
7.1). The two activities appear to operate separately, although the haem is
shared between the two active sites. Aspirin, for example, inactivated cycloox-
ygenase activity but not hydroperoxidase by acetylating serine-530 at the
amino terminus of the protein (Smith and Marnett, 1991; Smith et al., 1992).
In addition, iron can be replaced with manganese to give an enzyme which
retains the cyclooxygenase but no hydroperoxidase activity.
The reaction mechanism proceeds via a protein free radical, although the
precise details of the mechanism may differ from another iron enzyme, ribonu-
cleotide reductase. Although the details are not fully understood, the reaction
requires oxidization of the iron by a peroxide from the four- to five-valent state
and the abstraction of a hydrogen atom from the 13-pro-S position of arachi-
donic acid by a radical. A tyrosyl radical (probably Tyr-385) is formed when
the enzyme is incubated with arachidonate but may not be involved in the
cycloxygenase reaction. The native enzyme loses activity as a result of cataly-
sis - a sort of suicide inactivation (this is obviously an advantage for the cell
to close down the production of such reactive intermediates but it bedevils
efforts to unravel the mechanism). It is possible that the tyrosyl radical is some-
how involved in the inactivation (Smith et al., 1992). The anti-inflammatory
agents first inhibit the cyclooxygenase activity reversibly and then a time-
dependent inactivation follows with a stoichiometry of 1 mol drug:1 mol syn-
thetase dimer. The enzyme retains some 4- 10 per cent of activity. Dissociation
constants calculated for the initial binding are 1.7 X 10m5M for indomethacin,
2-O X lo-’ M for flurbiprofen and 8 X lo-’ M for meclofenamic acid. There is
a stereochemical basis to this action in that only the S(+) (dextrorotatory)
isomer of flurbiprofen is active (Kulmacz and Lands, 1985) and the o-isomer
of naproxen is 70 times more active than the L-isomer (Flower, 1974). Since
the absorbance at 410 nm due to the haem group did not change, it is not likely
to be involved in anti-inflammatory agent binding (Kulmacz and Lands, 1985).
The time-dependent change is likely to be conformational in the synthetase

lvie

Meclofenamic acid
130 Molecular mechanisms of drug action

because indomethacin can be recovered intact after inhibiting the enzyme, so

covalent reaction has not taken place. Esters of these agents, however, do not
show irreversible inhibition, so the charge on the carboxyl is important, per-
haps for binding to the arachidonic acid site.
The involvement of prostaglandin endoperoxide synthetase in the pathogen-
esis of inflammation rests on two types of evidence. The first is that the anti-
inflammatory agents inhibit the enzyme in a rank order which correlates with
their anti-inflammatory activity (Vane and Botting, 1987). The second lies in
the discovery that prostaglandins play a major part in the oedema, inflam-
mation, pain and fever that accompany rheumatoid arthritis. Other important
mediators include leukotriene B4 which attracts lymphocytes into the joint,
platelet-activating factor and interleukin 1 (Harris, 1986).
The use of these aspirin-like drugs can result in the development of stomach
ulcers. This side-effect has led to a greater understanding of the role of prosta-
glandins in protecting the stomach wall against acid by activating the secretion
of mucus in the intestine. In addition, the anti-inflammatory drugs can suppress
the release of gastric acid in the stomach. The drugs themselves may also cause
tissue damage by diffusing into the cells in the neutral form. They are weak
acids with pK, in the region of 5, so they will be uncharged at the pH of the
stomach. Inside the cells they will ionize at neutral pH and thus become
trapped (Price and Fletcher, 1990).

7.4 Rheumatoid arthritis

Rheumatoid arthritis is characterized by inflammation of the synovium of par-

ticular joints associated with destructive effects, leading to the dissolution of
cartilage and bone. Rheumatoid arthritis is not to be confused with osteoar-
thritis which contains less of an inflammatory component, in addition to carti-
lage breakdown. Both types of arthritis include an element of proliferative
growth in different tissues. In osteoarthritis, this can reach the point of excru-
ciating pain as the nodules grind in the joint and require removal, as in the
operation for hip replacement for example. As well as the hip, the knee is
commonly affected in osteoarthritis. It is a condition more common in later
life, although not necessarily a part of the ageing process. Both types of
arthritis respond to non-steroidal anti-inflammatory agents.
Rheumatoid arthritis is classified as an auto-immune disease in that anti-
bodies of the immunoglobulin M class (&Ms) found are against the body’s own
constituents, notably the Fc fragment of immunoglobulin G (IgG). These are
called rheumatoid factors which may be found in the serum as well as the
joint, although the level does not necessarily parallel the severity of the disease.
Indeed, the relevance of possible damaging effects of the autoantibodies to
disease progression is not at all clear. The disease is complex involving both
T and B cells of the immune system (see Insel, 1990).
The process seems to be initiated by an antigen that is attracted to the joint.
 Prostaglandin and leukotriene biosynthesis and action 131

Antibodies made within the joint space complex with the antigen and activate
the complement system. Antigen-presenting cells present the complex to T
cells which produce cytokines, including interleukins, that recruit more T
cells. The T cells differentiate and produce factors that induce both tissue
destruction and inflammation. More lymphocytes are recruited, the cells of the
lining of the synovium proliferate and synthesize connective tissue (pannus)
eroding cartilage, bone, tendon and ligaments. At some stage the process
becomes perpetuated, but the reasons for the persistence and fluctuation of
inflammation are poorly understood. The initiating antigen may reappear or
there may be a cycle of positive and negative immune responses.
Some view the T cell as driving the reaction, whereas others see the T cell
as being held in check as the body tries to control the inflammation. Which-
ever view is correct, the T cell clearly plays a major role in the process. T
cells need to have antigens presented to them in a way that they can be recog-
nized - a function carried out by antigen-presenting cells. The latter express
molecules on their surface encoded by the major histocompatibility complex
(MHC). These molecules bind antigen which, in the case of the macrophage,
may be derived from intracelhtlar degradation of pathogens. The antigen recep-
tor of T cells will only bind to antigen if it is complexed to an MHC molecule.
MHC la is expressed on the lymphocytes in the synovium probably as a result
of activation by y-interferon. Lymphocytes associated with MHC la rich cells
can produce lymphokines that promote synovial growth and B-cell differen-
tiation to antibody secreting cells - both events occurring in the arthritic joint
(Tinsel, 1990; Zvaifler, 1988). The persistent observation of an immune
response in rheumatoid arthritis has led to frequent searches for an infective
agent. Mycoplasma were suspected to be the culprits in earlier years, indeed
one species of mycoplasma, M. arthritidis, can cause a transient condition
which resembles arthritis, as its name suggests. More recently, however, atten-
tion has switched to viruses for the aetiology of the disease, but it is not clear
whether the virus is the causative agent or merely a passenger in a damaged
synovium.
The role of prostaglandins in this process is confined to pain sensitization,
oedema formation, increase in blood flow and fever. Chemotactic agents that
recruit more T cells are likely to be products of the lipoxygenase pathway
such as leukotriene B4. The non-steroidal anti-inflammatory agents can reduce
or eliminate the pain and inflammation in the short term, but the progressive
deterioration of the synovium and pannus formation continues, resulting in
osteoarthritis at least in the formation of painful and irregular nodules (Baker
and Rabinovitz, 1986).

7.5 Tbromboxane synthesis

The next step in the prostaglandin metabolic pathway is conversion of prostag-

landin H2 to thromboxane A, (Fig. 7.1). This material is extremely unstable
132 Molecular mechanisms of drug action

with a half-life of approximately 30 seconds in tissue fluids, being converted

by non-enzymic hydrolysis of the epoxide bridge to thromboxane B,, which
is almost biologically inactive. Much interest has centred on thromboxane A2
in view of its vasoconstrictor activity and its irreversible induction of platelet
aggregation which can lead to blockage of blood vessels particularly in the
heart. Thromboxane A2 is generated in vascular smooth muscle and cardiac
cells as well as platelets, among other tissues, thus being implicated in angina
and acute myocardial infarction.
One major therapeutic objective has been to blockade the action of throm-
boxane A,, either by inhibiting its synthesis or blocking its receptors. Throm-
boxane A, synthesis is rate-limited in the platelet by the availability of throm-
boxane synthetase. Prostaglandin H, diffuses out of the platelet and is taken
up by the epithelial cells in the vessel wall (aortic intima). Microsomal prosta-
cyclin synthetase catalyzes the formation of prostaglandin I, (prostacyclin), a
vasodilator and extremely powerful inhibitor of platelet aggregation at
between 1 and 10 X lo-“M. Most of the prostacyclin is apparently synthesized
this way. This diversionary process may be an important regulatory mechanism
for limiting clot (thrombus) formation in damaged vessels (reviewed in Gresele
et al., 1991).
Thromboxane synthetase inhibitors were synthesized and tested in the clinic
for the treatment of cardiovascular conditions including angina and hyperten-
sion. These drugs had, however, little consistent therapeutic effect (Fiddler
and Lumley, 1990), probably because they caused a build-up of prostaglandin
H, which binds more tightly to thromboxane receptors than thromboxane A2
itself, and is a more powerful aggregating agent. Attempts are now being made
to synthesize compounds which both inhibit thromboxane synthetase and are
antagonists at thromboxane receptors. Receptor antagonists on their own are
not expected to be effective, and also cannot redirect prostaglandin equiva-
lents from platelet to vessel wall (Gresele et al., 1991).

7.6 Leukotriene biosynthesis

Although no drug has yet reached the market that owes its activity to inhibition
of leukotriene biosynthesis, a great deal of effort is going into the search for
such a compound. This is largely because of the expectation that such a com-
pound would be useful in the treatment of asthma, since the ‘slow reacting
substance of anaphylaxis’, that is released in asthma and is responsible for
much of the bronchoconstriction of the condition, is now recognized to be a
mixture of leukotrienes. In addition, a leukotriene biosynthesis inhibitor might
be useful in arthritis as leukotriene B4 is a potent chemotactic agent for poly-
morphonuclear leukocytes.
The first committed step is peroxidation of arachidonic acid catalyzed by 5-
lipoxygenase (Fig. 7.2) which contains a non-haem iron at the active site, pro-
ducing 5-hydroperoxy-7,9,11,14-eicosatetraenoic acid which can then be con-
 Prostaglandin and leukotriene biosynthesis and action 133
verted into leukotriene A+ This is the branch-point from where leukotriene
B4 may be obtained by hydrolase action or C4 by the action of a glutathione
S-transferase. Leukotrienes D4 and E, result from the sequential action of y-
glutamyl peptidase and a dipeptidase.
Inhibition of the first enzyme in the pathway, 5-lipoxygenase, has turned
out to be the most fruitful. Inhibitors, that either bind to the active site iron
or compete with arachidonic acid, have shown promise in the clinic for
allergic and inllammatory disorders (reviewed in McMillan and Walker, 1992).

Questions

1. What separate catalytic activities are carried out by prostaglandin endo-

peroxide synthetase?
2. Are the non-steroidal anti-infIammatory agents irreversible inhibitors of
prostaglandin endoperoxide synthetase?
3. What therapeutic advantage would an inhibitor of thromboxane synthetase
be expected to have over a thromboxane A, receptor antagonist?
4. What features of the lipocortin/annexin molecule are unusual? Can you list
any cellular processes in which they might be involved?

References
Baker, D.G. and Rabinowitz, J.L., 1986,~ Clin. Pharmacol., 26, 2-21.
Creutz, C.E., 1992, Science, 258, 924-30.
Dahlman-Wright, K., Wright, A., Carlstedt-Duke, J. and Gustaffson, J.-A., 1992,J. Steroid
Biochem. Molec. Biol., 41, 249-72.
Fiddler, G.I. and LumIey, F., 1990, Circulation, 81, Suppl 1, 69-78.
Flower, R.J., 1974, Pharmacol. Rev., 26, 33-67.
Flower, R.J., Moncada, S. and Vane, J.R., 1985, In: The Pharmacological Bash ofThera-
peutics, 7th Edn., Eds. A.G. GiIman, L.S. Goodman, T.W. RaII and F. Murad (New
York: Macmillan), Ch. 29.
Gresele, P., Deckmyn, H., Nenci, G.G. and Vermylen, J., 1991, Trends in Pharmacol.
Sci., 12, 158-63.
Harris, E.D., 1986, Amer. J Med., 80, Suppl. 4B, 4-10.
Insel, P., 1990, Ch. 26 in the Pharmacological Basis of Therapeutics, 8th edn (eds
A.G. Goodman, T.W. Rall, A. Nies and P. Taylor) Pergamon, New York.
Kovacic, R.T., Tizard, R., Cate, R.L., Frey, A.Z. and WaIlner, B.P., 1991, Biochemistry,
30,9015-21.
Kulmacz, R.J. and Lands, W.E.M., 1985,J. Biol. Cbem., 260, 12572-8.
McMillan, R.M. and Walker, E.R.H., 1992, Trends in Pbarmacol. Sci., 13, 323-30.
Morimoto, RI., 1993, Science, 257, 1409-10.
Price, A.H. and Fletcher, M., 1990, Drugs, 40, Suppl 5, l-l 1.
Sautebin, L., Camuccio, R., Ialenti, A. and Di Rosa, M., 1992, Pharmacol. Res., 25, l- 12.
Skorodin, M.S., 1993, Arch. Intern. Med., 153, 814-28.
Smith, W.L. and Marnett, L.J., 1991, Biocbem. Biopbys. Acta, 1083, 1-17.
Smith, W.L., Eling. T.E., Kulmacz, R.J., Mamett, L.J. and Tsai, A.-L., 1992, Biochemistry,
31, 3-7.
134 Molecular mechanisms of drug action

Taylor, G.W. and Ritter, J., 1986, Trends in Pharmacol. Sci., 7, March centrepage.
Vane, J.R., 1971, Nature New Bioloa, 231, 232-6.
Vane, J.R. and Botting, R., 1987, FASEBJ, 1, 89-96.
Wang, W. and Creutz, C.E., 1992, Biochemistry 31, 9934-9.
Withnall, M.T., Brown, T.J. and Diocee, B.K., 1984, Biochem. Biophys. Res. Commun.
121, 507-13.
Zvaifler, N.T., 1988, Amer. J Med., 85(4A), 12- 17.
 Chapter 8

Zinc metalloenzymes

8.1 Introduction
Approximately 200 enzymes have so far been discovered that contain at least
one zinc atom as part of the integral functioning protein. Vallee and Galdes
(1984) have classified the roles that zinc can play in metalloproteins.
1. Catalytic, in which the metal plays an active part in the catalysis. Notable
examples of this are carbonic anhydrase, angiotensin-converting enzyme
and alcohol dehydrogenase from liver.
2. Structural, where the zinc is required to maintain the structural integrity of
the protein as in pamylase and the glucocorticoid receptor (see Chapter 7).
3. Regulatory, either inhibitory or activatory, as in leucine aminopeptidase.
4. Other unspecified role(s) which may be regarded as non-catalytic for want
of better knowledge.
The complete complement of 10 electrons in the 3d shell of the zinc atom
has a screening effect on the positive charge of the nucleus, compared with
calcium which lacks this screen. Consequently, zinc is able to form covalent
coordinate complexes in addition to the ionic complexes characteristic of cal-
cium. Ligands can, therefore, easily enter its coordination sphere and donate
a pair of electrons to form a coordinate or dative bond. Zinc may have either
four or six such ligands in simple inorganic compounds.
Furthermore, the external shell of zinc’s electrons is polarizable and there-
fore distorts in an asymmetric environment. Characteristic features of the cata-
lytic type of zinc protein are (a) a highly asymmetric environment of the zinc
such that the atom can be regarded as either 4- or 5coordinated and (b) three
ligands to the protein amino acid side-chains, while water occupies the fourth
position. Non-catalytic and structural zinc is normally bound to the protein by
four ligands and is not as distorted as catalytic zinc. The polarizability plays a
practical part in the catalysis as it has a higher energy than a non-distorted
state and thus the activation energy for the reaction is correspondingly low-
ered (Vallee and Auld, 1990).
Covalent coordinate interactions require a closer inter-atomic approach than
is seen in ionic compounds. Furthermore, zinc will accept a variety of ligands
based on oxygen, nitrogen or sulphur, as exemplified by the formation of

135
136 Molecular mechanisms of drug action

sodium zincate (Na,ZnO,) and zinc diamine chloride [Zn(NH3>JC12. In this

way a variety of reaction types can be catalyzed, with cysteine, histidine or
glutamic acid side-chains acting as anchors for the binding of zinc to the pro-
tein.
Zinc, as an electrophilic cation, can activate the substrate chemically by
behaving as a Lewis acid (meaning that it can readily accept a share in a pair
of electrons from a ligand, with the concomitant formation of a coordinate
covalent bond). This has the effect of enhancing ligand nucleophilicity, as for
example by ionizing a water molecule and binding the hydroxide ion formed,
which can then participate in hydrolysis reactions that the initial water mol-
ecule itself could not achieve. In addition, zinc can facilitate catalysis by orien-
tating reactants through coordination on the metal as a template (Vallee and
Galdes, 1984).
Another feature of catalytic zinc proteins is the occurrence of acyl-enzyme
intermediates in the catalytic process, as exemplified by acyl-phosphates in
alkaline phosphatase. The nature of the nucleophile which breaks the acyl-
enzyme intermediate is not certain, but in a number of hydrolytic metallo-
enzymes the proposition of a zinc attached hydroxide ion has been made
(Coleman, 1984). If the variation of enzyme activity with pH is measured (pH
profile) there is a requirement for catalysis of a group with a pK, of 7 to
7.5 which nuclear magnetic resonance (n.m.r.) studies have shown is not an
imidazole. In effect the zinc is lowering the pK, of the hydroxide moiety of
the water molecule so that it can act as a nucleophile (see Vallee and
Galdes, 1984).
It is the catalytic type of zinc atom with which we are concerned in this
chapter, particularly as the inhibitors of the enzyme actually form complexes
which involve direct attachment of a ligand to the zinc. Inhibitors of two
enzymes containing catalytic zinc have found pharmaceutical use. Carbonic
anhydrase inhibitors are useful both in situations where there is an excess of
fluid (oedema) in various tissues and, therefore, in the treatment of glaucoma,
and also when urine needs to be ‘alkalinized’ because acids like uric acid might
otherwise precipitate. Secondly, angiotensin-converting enzyme (ACE) inhibi-
tors are now playing a major part in the reduction of blood pressure. Poten-
tially, a third application lies in inhibiting an enzyme that can degrade enke-
phalin. Such inhibitors may be approaching clinical trial as painkillers.

8.2 Carbonic anhydrase - methazolamide for glaucoma

An enzyme exists in a large number of tissues including kidney, erythrocyte
and pancreas catalyzing the hydration of carbon dioxide to form carbonic acid,
which immediately ionizes to form a bicarbonate ion:

CO, + H,O = H&O, + HCO, + H+

This enzyme (carbonic anhydrase) can also catalyse the hydration of aldehydes
 Zinc metalloenzymes 137

and ketones, such as pyruvate, but it is unlikely that any of these reactions
have physiological significance. It may seem unnecessary for an enzyme to
carry out this reaction - estimates of the ration of enzyme-catalysed to uncata-
lysed rate (catalytic rate enhancement) range from 8000 (Maren, 1984) to lo5
(Coleman, 1984), depending on the source of the enzyme. Nevertheless, the
ubiquity and high level of the enzyme in, for example, the kidney suggest
that the rate enhancement of this reaction plays a very important role in the
metabolism of the cell.
Carbonic anhydrase normally exists in at least three forms (or isoenzymes)
designated CAl, CA2 and CA3, which are distinguished mainly by a consider-
able difference in catalytic rate with CA2 being the most active. The isoen-
zymes also differ to some extent in primary sequence but not in overall length
of some 260 amino acid residues, resulting in a molecular weight of approxi-
mately 30 kDa. There are 105 differences between human erythrocyte CA1
and CA2 which represents 59.6 per cent homology - the similarity is particu-
larly close about the active site as would be expected. Other species show
greater homology between the two forms (Deutsch, 1984).
As shown by X-ray crystallographic data, carbonic anhydrase contains one
zinc atom per active site coordinated to three nitrogen atoms from imidazole
side-chains of histidines residues 94, 96 and 119, and to one oxygen atom from
a water molecule in an approximately tetrahedral conformation. The reaction
mechanism has been the subject of much discussion over the years (reviewed
in Silverman and Lindskog, 1988).
The zinc atom based in a hydrophobic pocket acts to lower the pK, of the
water molecule and binds a coordinated hydroxide ion although the pH of the
surrounding medium may be neutral. Another enzyme residue, probably His
64, accepts the proton before transferring it to buffer. The hydroxide ion
behaves as a nucleophile, by attacking carbon dioxide to form bicarbonate
ion. A sigmoid pH profile is obtained for both hydration and dehydration reac-
tions but in opposing senses, which may be a consequence of the requirement
for a base such as the hydroxide ion or His 64 in the active site with a pK,
in the region of 7 as an intermediate for both reactions (Silverman and Lind-
skog, 1988). A ii-coordinated zinc atom (with a second water molecule
attached via its oxygen atom to the zinc) has been shown to have theoretical
advantages for the catalysis in that the second water molecule is ready to slip
into the place of the first water molecule as soon as reaction has occurred,
but it is at present difficult to be sure how many water molecules are at the
active site (Cook and Allen, 1984).
The detailed interaction of the sulphonamides, used originally as diuretics,
with carbonic anhydrase has been a subject of considerable interest. The sul-
phonamide group has to be unsubstituted for activity, with the amide nitrogen
atom binding to the zinc atom at the fourth coordination site displacing the
water molecule. One of the two sulphonamide oxygen atoms links with the
f&h coordination site of the zinc (see Vedani and Meyer, 1984). Kinetic studies
have shown that the sulphonamides are competitive inhibitors of the dehy-
138 Molecular mechanisms of drug action

dration of carbonic acid as well as the hydration of carbon dioxide, as one

would expect from the principle of microscopic reversibility. In this situation,
competitive kinetics have shown that the inhibitor binds at the substrate bind-
ing site (Cook and Allen, 1984).
Carbonic anhydrase is normally present in tissue in great excess, as indeed
are a number of other enzymes, and over 99 per cent of the enzyme has to
be inhibited before physiological effects become apparent. The importance of
this high concentration relates to bicarbonate ion transport across epithelial
cell membranes where sodium is the counter-ion, in that a large excess of
enzyme is required to act on the substrate.
One example of this occurs in the proximal tubule of the kidney where the
majority of bicarbonate ion that has previously been filtered out in the glom-
erulus is reabsorbed (Kokko, 1984) (Fig. 8.1). In a coupled transport process,
sodium ion is reabsorbed while hydrogen ion is secreted into the filtrate pass-
ing through the lumen of the proximal tubule (Fig. 82a). Hydrogen and bicar-
bonate ions form carbonic acid which disproportionates to carbon dioxide
and water, catalyzed by carbonic anhydrase in the luminal membrane (Lucci
et al., 1983). Carbon dioxide diffuses across the luminal membrane of the
proximal tubule cell and the cytoplasmic enzyme reconverts in into bicarbon-
ate ion which is transported across the basement membrane of the cell into
the peritubular fluid that drains into the blood. The proton that is also formed
in the reaction is recirculated across the luminal membrane in exchange for
sodium (DuBose et al., 1981; Lucci et al., 1983). The overall result is the net
reabsorption of sodium and bicarbonate ion.
If insufficient hydrogen ion is secreted to react with all the tubular bicarbon-
ate, then the urine will contain larger amounts of bicarbonate than usual and
consequently will have a higher pH. The content of sodium ion will also
increase (Kokko, 1984) and in order to maintain the osmotic pressure water
will be secreted together with the ions (Fig. 8.2a). This situation occurs as the
consequence of carbonic anhydrase inhibition because the supply of protons
is greatly reduced. It appears that at least two-thirds of the carbon dioxide
reabsorption is mediated by carbonic anhydrase because, when the enzyme is
completely inhibited, the proportion of bicarbonate reabsorbed falls to about
37 per cent of its control value (Maren, 1984). Carbonic anhydrase inhibitors
were originally developed for use as diuretics but their action is only weak
because their major effect is in the proximal tubule (Kokko, 1984). As a conse-
quence the greater volume of fluid reaching the distal tubule produces a com-
pensatory reabsorptive effect there which markedly reduces the effect of
the drug.
Another family of diuretics, the widely-used thiazides, were originally
developed out of studies on the sulphonamides. Although the thiazides are
weak inhibitors of carbonic anhydrase, they act primarily at the early portion
of the distal tubule and, since carbonic anhydrase is mainly located in the
proximal tubule, inhibition of this enzyme is unlikely, therefore, to play much
 Zinc metalloenzymes 139

Bowman’s
capsule

Descending

Thin

Loop of Henle

Shaded area represents the site of action of

carbonic anhydrase inhibitors.

Low molecular weight substances are ultrafiltered from the blood through a capillary
system in the glomerulus. The filtrate passes into Bowman’s capsule and then through the
proximal tubule, the loop of Henle and the distal tubule. Ii finally reaches a collecting duct
that drains several nephrons.

Figure 8.1 Site of action of carbonic anhydrase inhibitors

of a part in the mode of action of the thiazide diuretics. Inhibition of sodium

chloride cotransport may be the key (Gesek and Friedman, 1992).
Although carbonic anhydrase inhibitors are now very rarely used as
diuretics, they are sometimes used to increase the pH of the urine. This
becomes necessary when, for example, drugs that cause tumours to diminish
in size (oncolytic) release large amounts of the breakdown product of purines,
namely uric acid. At normal pH this could begin to precipitate and so a car-
bonic anhydrase inhibitor is used, often in conjunction with sodium bicarbon-
ate, to keep the acid in the form of the more soluble sodium salt.
Carbonic anhydrase inhibitors are also used for the treatment of glaucoma.
The object is to reduce the volume of liquid in the aqueous humor of the eye
where this has become excessive and is putting a great deal of pressure on
the eyeball. The situation in the eye differs from that in the kidney in that the
140 Molecular mechanisms of drag action

(a) Proximal tubule of the kidney

Glomerular filtrate Proximal tubular cell Peritubular fluid

(leading to blood)

cl- . Cl-

Luminal membrane Peritubular membrane

0 represents direct ion exchange

represents ion pump linked to ATPase

0

probably occurs in the brush border membrane

*

(b) Ciliary epithelium in the eye

Aqueous humour Cilian/ epithelial cell Leading to blood

HCOJ-

Na+ 4 Na++ Na+

cl-
Cl- 4 cl-

&O 4

/
Luminal membrane Basolateral membrane

Figure 8.2 Ion transport in kidney and eye.

 Zinc metalloenzymes 141

protons are secreted across the basolateral membrane into the blood and not
into the lumen or aqueous humor. Bicarbonate ion is secreted into the aqueous
humor, however, and inhibition of carbonic anhydrase reduces the level of
the ion in the humor by about 50 per cent thus reducing the concomitant
fluid outflow (Maren, 1984). The details of ion transport in the eye are less
well understood, however, than in the proximal tubule cell of the kidney
(Fig. 8.2b).
Carbonic anhydrase inhibitors are particularly useful for the treatment of
glaucoma since they reduce the pressure behind the iris, and assist in opening
the angle between the iris and the cornea thus allowing the aqueous humor
to leak into the anterior chamber and be reabsorbed. Drugs are used in the
acute condition in order to manage the attack while the patient is prepared
for surgery - normally a necessity because a physical obstruction has usually
induced the condition in the first place. The main intention of acute treatment
is to reduce liquid volume, and thus intraocular pressure, so preventing dam-
age to the optic nerve. The drugs are less useful in treating chronic primary
open angle glaucoma, which slowly progresses as a consequence of a gradual
closing of the channel through which the aqueous humor circulates until it
may result in irreversible loss of vision.
Acetazolamide was originally used to treat glaucoma but its use has largely
been superseded by the more lipophilic methazolamide. The major problem
with carbonic anhydrase inhibitors given orally has been the occurrence of
side-effects. Topically administered agents are now under development for
glaucoma which should have less potential for causing adverse reactions
(Hurvitz et al., 1991).

Acetazolamide Methazolamide

8.3 Angiotensin-converting enzyme

The second zinc metalloenzyme to be discussed in this chapter is the metallo-

peptidase, angiotensin-converting enzyme (ACE), the search for inhibitors of
which has turned out to be both a very interesting intellectual study and also
of great importance medically. The search was initiated because ACE converts
the biologically inactive decapeptide, angiotensin I, to the highly active pressor
(blood pressure raising) octapeptide, angiotensin II, by removing the C-ter-
minal dipeptide His-Leu (Fig. 8.3). The enzyme will also cleave the nonapep-
tide, bradykinin, which is a depressor agent, to inactive metabolites. ACE is
still sometimes known as kininase 2, and it plays a pivotal role in the control
of blood pressure.
142 Molecular mechanisms of drug action
Hydrolysis of angiotensin and bradykinin
Angiotensinogen: Bradykininogen

Asp-Arg-Val-Tyr-lle-His-Pro-Phe-His-Leu-Leu-Val....
m Kallikrein

Angiotensin I: Bradykinin:
I I
Asp-Arg-Val-Tyr-lie-His-Pro-Phe-His-Leu Arg-Pro-Pro-Gly-Phe-Ser- Pro-Phe-Arg

I
Angiotensin-converting enzyme = Kininase II
Angiotensin ll:
1 I
Asp-Arg-Val-Tyr-lie-His-Pro-Phe Arg-Pro-Pro-Gly-Phe + SW-Pro + Phe-Arg

Angiotensinogen, the precursor protein, is cleaved at one specific site by renin, secreted
by the kidney, to yield biologically inactive angiotensin I. This is converted by angiotensin-
converting enzyme to the active octapeptide angiotensin Il.
Precursor bradykininogen is converted to the active nonapeptide bradykinin by kallikrein
and thence to the inactive pentapeptide by angiotensin-converting enzyme.

Figure 8.3 Hydrolysis of angiotensin and bradykinin

In general, the enzyme prefers substrates with hydrophobic amino acids in

the ante-penultimate position. It will cleave a variety of substrates, both pep-
tide and ester, provided that the C-terminal amino acid has a free carboxyl
group, and proline is not present in the penultimate position. Thus angiotensin
II is not hydrolyzed further by ACE, whereas bradykinin is attacked in three
places (Fig. 8.3) and so the enzyme may be regarded as not particularly spec-
ific (Ondetti and Cushman, 1984).
Angiotensin II is one of the most active biological agents known. It raises
blood pressure by constricting blood vessels and, in addition, it causes aldos-
terone to be released from the adrenal cortex. The latter acts on the kidney
to retain sodium ion, and therefore also water, thereby raising blood volume
and pressure. The complex system for maintaining blood volume and pressure
initially depends on the action of the endopeptidase renin. When blood vol-
ume or sodium ion concentration in the plasma falls below a certain critical
value, renin is released from the kidney and acts on an cY-globulin known as
angiotensinogen - a glycoprotein of molecular weight 60 kDa - to yield angio-
tensin I. Fig. 8.4 outlines the interactions of the renin-angiotensin-aldos-
terone system.
It was clear in the 1970s from this knowledge that an inhibitor of ACE might
be expected to have anti-hypertensive properties, and a study began to design
agents that would bind to the active site. Although ACE had not been isolated
at all at that time, advantage was taken of the much greater understanding of
the active site of another zinc metallopeptidase, namely carboxypeptidase, an
enzyme which is elaborated by the pancreas and cleaves the C-terminal amino
acid from peptides and proteins.
The structures of enzymes where zinc has a catalytic role show three ligands
from the protein attached to the zinc: in carboxypeptidase A they are His-69,
 Zinc metalloenzymes 143

Interactions of the renin-angiotensin-aldosterone axis

Angiotensinogen

Kidney

Angiotensin I

Vasoconstriction Aldosterone

Sodium,
4
water retention
Blood pressure

@ Denotes activation or elevation

@ Denotes inhibition

Figure 8.4 Interactions of the renin-angiotensin-aldoaerone axis.

Glu-72 and His-196. The fourth ligand is the oxygen of a water molecule
(reviewed in Vallee and Auld, 1990). This is replaced in the enzyme-substrate
complex by the oxygen of the carbonyl group of the peptide link to be broken,
and the latter is thereby polarized by the action of the zinc atom acting as a
Lewis acid. Glu-270 then functions as a base or nucleophile to attack the car-
bon of the carbonyl group to yield a transient acyl-enzyme intermediate. Tyr-
248 donates a proton to the departing amino group. The free carboxylate of
the substrate is linked by attraction between a negative and positive ion (salt
bridge) to Arg-145.
The nature of the group that, in effect, breaks up the enzyme-substrate
complex by attacking the carbonyl carbon is likely to be a hydroxide ion,
which may be activated by Glu-248 or possibly the zinc (Lipscomb, 1980). It
would be intellectually satisfying if the zinc atom played a role in carboxypepti-
dase mechanism similar to that in carbonic anhydrase, in lowering the pK, of
the water molecule and allowing a hydroxide ion to attack a carbon of a car-
bony1 group. The presence of three histidine ligands (positions 94, 96 and
118) around the zinc suggests that this is the case. For a diagrammatic rep-
resentation of a possible enzyme mechanism see Fig. 8.5.
Although the information regarding the active site of angiotensin-converting
enzyme is still fairly sparse, it is clear that similar groups are involved in the
catalysis, as shown by group-specific inactivating agents backed up by the use
of ligands to protect against inactivation (Brunning, 1983; reviewed in Ondetti
and Cushman, 1984). One difference between the two enzymes is the require-
ment for chloride ion demonstrated by ACE. Chloride ion has no effect on the
144 Molecular mechanisms of drug action

R2

LJ
B- LO

AH zn 4-.
A-H

(4>
1. The base group B (Glu-270 in carboxypeptidase) attacks the carbon of the carbonyl
group already distorted by activation in the zinc.
2. An acyl-enzyme intermediate is formed, and then rearranges to release the amine
F&NH, taking the proton from the acid group A (Tyr-248).
3. The hydroxyl of the water molecule activated by the zinc attacks the carbonyl group to
yield the acid R,COOH, while acid group A receives the proton.
4. The acid R&OOH is released and the active site resumes its original format.

Figure 8.5 Putative Zn-metallopeptidase mechanism.

rate of modification of arginyl, carboxyl and tyrosyl residues by these reagents,

but it does affect the rate of reaction of pyridoxal phosphate with the enzyme,
which suggests that lysine, which forms a Schiff s base with the aldehyde
group of pyridoxal, may bind the halide ion. Activation by chloride ion is seen
with most substrates and at all pH values, and is characterized by a lowering
of K,,, but no change in KC,, (Ondetti and Cushman, 1984). Recent studies have
defined the active site glutamic acid residue in bovine lung ACE in the
sequence: H,N-Phe-Thr-Glu-Leu-Ala-Ser-Glu. This sequence shows considerable
homology with bovine carboxypeptidase A, but little with the bacterial zinc
endopeptidase, thermolysin (Harris and Wilson, 1985).

8.3.1 Angiotensin-converting enzyme inhibitors - captopril and

enalaprD for hypertension
The story of the development of the ACE inhibitors that are marketed today
is a triumph of logical and imaginative planning, combining features of both
substrates and inhibitors. As mentioned above, the information about the
active site of ACE was strictly limited. Nevertheless, the potential similarities
between ACE and carboxypeptidase A were exploited to the full. Early studies
 Zinc metalloenzymes 145

with inhibitory peptides showed that having proline or an aromatic amino

acid in the C-terminal position gave the most effective inhibitors, but proline
inhibitors gave the better results in vivo; Val-Trp is some 150 times more active
in vitro than Ala-Pro but no more active in vivo (Petrillo and Ondetti, 1982).

(q?OH
;H3 *

HSCH,CH-C-N
3
cl ii

Captopril

Enalapril

Succinylproline could be regarded as a good starting point for inhibitor

development with four sites interacting with the protein; the carboxylate
group of the proline, the proline ring itself, the carbonyl of the amide link
and the carboxylate of the succinate all binding to the zinc atom (Z50 of
3.3 X lo-* M). A methyl group placed on the carbon atom next to the carbonyl
group of the amide improved the binding markedly. Replacement of the car-
boxylate by a sulphydryl gave a further enormous improvement, to produce
captopril with an Z5, of 2.3 X 10p8~. Enalaprilic acid was derived by three
further changes; the addition of further binding affinity by a phenethyl group
(possibly interacting with the binding site for the phenylalanine side-chain
characteristic of angiotensin I), together with a restoration of the carboxyl
group in place of sulphydryl, and the introduction of a secondary amino group.
These changes lowered the Z50to 1 X lo-9111 (reviewed in Ondetti and Cush-
man, 1984; Mackaness, 1985) (Fig. 8.6). Captopril has two, and enalaprll three,
optically active carbon atoms, and interestingly all five centres are in the S
configuration (Patchett et al., 1980: Brittain and Kadin, 1990).
Detailed kinetic studies with different substrates and inhibitors, excluding
captopril and enalaprilic acid, have revealed a complex pattern of interactions.
It appears that an inhibitor can bind to more than one form of the enzyme
and, in addition, there are parallel reaction pathways. Moreover, both captopril
and enalaprilic acid are slow and tight binding inhibitors (Morrison, 1982). As
a consequence, sufficient time must be allowed for complete binding to place,
and, in addition, allowance must be made for the depletion of active enzyme
by drug binding as both drug and enzyme concentrations are in the nanomolar
146 Molecular mechanisms of drug action

Inhibitors 0’ 0 0- M/M
I II I
1 O=C-CH2--Hz--C- N c=o 333.0
Succinyl Proline

0- CH3 f 0-
I 1 I
2 O=C--Hz-CH -C-N c=o 22.0

0
II ?-
3 HS-CH2 - CH2-C-N CL c=o 0.20

0.023

Enalaprilic acid 0.001

Figure 8.6 The development of ACE inhibitors.

range. Ki values of 3.3 X 10PiO~ and 5-O X lo-“M have been calculated for
captopril and enalaprilic acid respectively (Shapiro and Riordan, 1984). A two-
step inhibition is identified characterized by a fast step followed by a slow
second one which is probably an isomerization of the complex. The overall
type of inhibition is judged to be competitive.

8.3.2 Pharmacology of ACE inhiiitors

Captopril and enalapril have made a major contribution to our understanding
of the mechanism for maintaining blood pressure as well as to the treatment
of hypertension. These drugs lower angiotensin II levels in the peripheral net-
work of blood vessels, causing them to dilate, and consequently lower blood
pressure. There is no particular effect on the force or rate of contraction of
the heart (Todd and Goa, 1992). Enalaprilic acid is given as the ethyl ester,
enalapril, as the acid is poorly absorbed from the gastrointestinal tract. Sub-
sequent hydrolysis by esterases yields the active agent.
In the clinic, captopril and analapril are extremely effective at lowering the
blood pressure in hypertensive patients, particularly those in whom the reno-
vascular system is the originator of the condition and the renin-angiotensin-
aldosterone axis is involved. Elevated aldosterone and angiotensin II levels in
the plasma lead to increased sodium and water retention and so to increased
 Zinc metalloenzymes 147

blood pressure. Treatment with ACE inhibitors lowers these levels to normal.
The efficacy of the drugs extends to essential hypertension, i.e. of unknown
origin, whether mild, moderate or severe, but in these cases a diuretic is usu-
ally prescribed, in addition, to lower sodium levels and blood volume and
thereby assist the action of the ACE inhibitor (Brunner et al., 1983; Ferguson
et al., 1984).
Since these drugs were designed to combat angiotensin II action, we might
expect a correlation between initial plasma renin concentration and the blood
pressure lowering effects, and this is indeed found to be the case for captopril
(see Ferguson et al., 1984) and enalapril (see Brunner et al., 1983). Further-
more, the levels of angiotensin II prior to treatment correlate with falls in
blood pressure as a consequence of treatment (MacGregor et al., 1983). Plasma
renin rises as a consequence of captopril treatment - not unexpectedly in view
of the positive feedback effect of lowered sodium levels and blood volume on
the secretion of renin. Plasma renin may not be the only target, however,
because both renin and ACE have been found in the blood vessel walls, where
renin levels are elevated as a consequence of captopril treatment (Unger et
al., 1983). In addition, the renin-angiotensin system has been found in brain
and it is possible that ACE inhibitors that can cross the blood-brain barrier
will have a further antihypertensive effect (Unger et al., 1983).
In principle, we would expect that plasma renin would be an indicator of
blood pressure, since elevated blood pressure usually acts to lower plasma
renin levels. This does not always seem to occur, however, since a number
of patients with high blood pressure have normal, instead of lowered, renin
levels (Lindpainter et al., 1988). ACE inhibitors are found to work even when
renin levels are normal. As noted above, the renin-angiotensin system is found
outside the kidney and a major contribution may be made by the heart. Angio-
tensin II produced in the heart may assist in contraction, but on the negative
side produces vasoconstriction thus lowering the supply of oxygen to the heart
with the risk of producing ischaemia and arrhythmias (Grinstead and
Young, 1992).
The ACE inhibitors produce relatively few side-effects but dry cough and
proinflammatory properties may result from inhibition of bradykinin hydroly-
sis, also carried out by ACE. Angiotensin II antagonists are being developed as
more specific antihypertensives (Smith et al., 1992).
There are two other effects of these drugs that have emerged as a conse-
quence of their use. One is the interaction of angiotensin II with the trans-
mission of adrenergic impulses, as the peptide stimulates the release of norad-
renaline from presynaptic vesicles and inhibits the re-uptake of the
neurotransmitter. Consequently, angiotensin II potentiates the vasoconstrictor
response to sympathetic nerve stimulation as well as to exogenously applied
noradrenaline. Clearly, the inhibition of angiotensin II production will limit
this effect, and supplement the action of ACE inhibitors as vasodilators
(reviewed in Unger et al., 1983).
Another very interesting effect that seems to be associated with ACE inhibi-
148 Molecular mechanisms of drug action

tor treatment is euphoria (Zubenko and Nixon, 1984). The reason for this is
not entirely certain, but it may be due to the inhibition of the hydrolysis of
the enkephalin precursor, [Met]-enkephalin-Arg6-Phe’ which is probably car-
ried out by ACE in the brain (Normal et al., 1985). It is already clear that
enkephalinase, which cleaves [Metl-enkephalin at the Gly3-Phe* bond, is not
inhibited by captopril, and so it is intriguing to speculate that other opioid
peptides are responsible for the euphoria; particularly as [Metl-enkephalin-
Arg”Phe’ is derived physiologically from the proenkephalin A precursor pro-
tein. Opioid peptides are discussed in section 9.7 in Chapter 9.

8.4 Endopeptidase 24. I1 - thiorphan as analgesic

The treatment of pain in an effective but non-addictive manner remains one

of the most important pharmacological goals of the future. The discovery of
the opioid peptides as the endogenous ligands for the morphine receptor sug-
gested that prolongation of their action could lead to a new treatment of pain,
as an alternative to the use of morphine (Chapter 9). This approach has led
to two structures under clinical trial: thiorphan for pain and acetorphan for
diarrhoea. Acetorphan is the S-acetylphenyl ester of thiorphan and could be
regarded as a pro-drug.
One enzyme that can hydrolyze enkephalins is endopeptidase 24.11 (it has
been given this curious shorthand as an abbreviation of the Enzyme Com-
mission number 3.4.24.11). The enzyme has been isolated from kidney
microvilli and shown to be identical with an enzyme from pig brain. The
enzyme should not, strictly speaking, be referred to as enkephalinase because
it will hydrolyze various peptide substrates (Hersch, 1986). Nevertheless, in
the central nervous system its main role does appear to be termination of the
action of enkephalins by cleaving [Met]- and [Leu]-enkephalin to yield Tyr-Gly-
Gly and either Phe-Leu or Phe-Met - acting, in fact, as a dipeptidylcarboxypep-
tidase (reviewed in Roques et al., 1993).
Endopeptidase 24.11 is a zinc-containing enzyme and bears a considerable
resemblance to the family of zinc metallopeptidases. Approaches to inhibition
based on the strategy that originally led to the angiotensin-converting enzyme
inhibitors has yielded a compound that is analogous to captopril but which
shows considerable selectivity for endopeptidase 24.11 (Mumford et al., 1982).
As in captopril binding to ACE, the sulphur atom of thiorphan binds to the
zinc and the carboxylate to an arginine on the enzyme surface. Not all activity
is lost, however, with these two sites blocked because acetorphan has a Ki of
5 X 1Op6M. The proposal is that the heteroatoms in the amide bond also make
an important contribution to the binding by forming hydrogen bonds to the
enzyme (Monteil et al., 1992).
Enkephalinase also inactivates atria1 natriuretic factor, a hormone produced
by the heart which has sodium excreting (natriuretic) and diuretic activity
amongst other attributes. Enkephalinase inhibitors may also be useful for
 Zinc metalloenzymes 149
potentiation of the effects of atria1 natriuretic factor (Schwartz et al., 1990;
Roques and Beaumont, 1990).

f-342
I
HS-CH2-CH-NHCOCH2-COOH

Questions

1. What features of the zinc atom allow it to form covalent complexes?

2. Why do cells have carbonic anhydrase?
3. What is the present therapeutic use of carbonic anhydrase inhibitors?
4. What chiral centres does captopril have?
5. Why do ACE inhibitors lower blood pressure in patients whose renin levels
are normal?
6. What biological effects does angiotensin show?
7. What therapeutic effects would an enkephalinase inhibitor be expected
to show?

References
Brittain, H.G. and Kadin, H., 1990, Pbarm. Res., 7, 1082-5.
Brunner, H.R., Turini, G.A., Waeber, B., Nussberger, J. and Biollaz, J., 1983, CZin. Exper.
Hype&ens., A5, 1355-66.
Brunning, P., 1983, Gin. Exper. Hypertens., A5, 1263-75.
Coleman, J.E., 1984, Ann. N. Y. Acad. Sci., 429, 26-48.
Cook, C.M. and Allen, L.C., 1984, Ann. N. Y. Acad. Sci., 429, 84-S.
Deutsch, H.F., 1984, Ann. N. Y. Acud. Sci., 429, 183-94.
DuBose, T.D., Pucacco, L.R. and Carter, N.W., 1981, Am. J. Pbysiol., 240, F138-46.
Ferguson, R.K., Vlasses, P.H. and Rotmensch, H.H., 1984, Am. J. Med., 77, 690-S.
Gesek, F.A. and Friedman, P.A., 1992,J. Clin. Invest., 90, 429-38.
Grinstead, W.C. and Young, J.B., 1992, Amer. HeartJ, 123, 1039-45.
Harris, R.B. and Wilson, I.B., 1985, J Biol. Chem., 260, 2208-l 1.
Hurvitz, I.M., Kaufman, P.L., Robin, A.L., Weinreb, R.N., Crawford, K. and Shaw, B.,
1991, Drugs, 41, 514-32.
Kokko, J.P., 1984, Am.J Med., 77, (5A), 11-17.
Lindpainter, K., Jin, M., Wilhelm, MJ., Suzuki, F., Linz, W., SchoeIkens, B.A. and Ganten,
D., 1988, Circulation, 77, Suppl. 1, 18-23.
Lipscomb, W.N., 1980, Proc. Nat. Acud. Sci. (US.A.), 77, 3875-B.
Lucci, M.S., Tinker, J.P., Weiner, I.M. and DuBose, T.D., 1983, Am. J Pbysiol., 237,
F443-9.
150 Molecular mechanisms of drug action

MacGregor, G.A., Markandu, N.D., Smith, S.J., SagneIla, G.A. and Morton, J.J., 1983,
Clin. Exper. Hypertens., AS. 1367-80.
Mackaness, G.B., 1985, J. Cardiovasc. Pbarmacol., 7, Suppl. 1, S30-4.
Maren, H., 1984, Ann. N. Y. Acad. Sci., 429, 568-79.
Monte& T., Kotera, M., Duhamel, I., Duhamel, P., Gros, C., Noel, N., Schwartz, J.C.
and Lecomte, J.M., 1992, Bioorg. Med. Chem. Lett., 2, 949-54.
Morrison, J.F., 1982, Trends in Biocbem. Sci., 7, 102-5.
Mumford, R.A., Zimmerman, M., ten Broeke, J., Joshua, H. and Bothrock, J.W., 1982,
Clin. Exper. Hypertens., A5, 1362-80.
Norman, J.A., Autry, W.L. and Barbaz, B.S., 1985, Molec. Pharmacol., 28, 521-6.
Ondetti, M.A. and Cushman, D.W., 1984, CRC Crit. Rev. Biochem., 16, 381-411.
Patchett, A.A., et al., 1980, Nature, 288, 280-3.
PetriIIo, E.W. and Ondetti, M.A., 1982, Medicinal Res. Rev., 2, 1-41.
Roques, B.P. and Beaumont, A., 1990, Trends in Pharmacol. Sci., 11, 245-9.
Roques, B.P., Noble, F., Dange, V., Fournie-Zaluski, M-C. and Beaumont, A., 1993, Phar-
macol. Rev., 45, 87-146.
Schwartz, J.C., Gros, C., Lecomte, J.M. and Bralet, J., 1990, Life Sci., 47, 1279-97.
Shapiro, R. and Riordan, J.F., 1984, Biochemistry, 23, 5225-33.
Silverman, D.N. and Lindskog, 1988, Act. Chem. Res., 21, 30-36.
Smith, R.D., Chiu, A.T., Wong, P.C., Herblin, W.F. and Timmermans, P.B.M.W.M., 1992,
Annu. Rev. Pbarmacol. Toxicol., 32, 135-65.
Todd, P.A. and Goa, K.L., 1992, Drugs, 43, 346-81.
Unger, T., Ganten, D. and Lang, R.E., 1983, Clin. Exper. Hypertens., A5, 1333-54.
ValIee, B.L. and AuId, D.S., 1990, Proc. Nat. Acad. Sci. (U.S.A.), 87, 220-4.
Vallee, B.L. and Galdes, A., 1984, Adv. Enzymol., 56, 281-430.
Vedani, A. and Meyer, E.F., 1984,J. Pbarm. Sci., 73, 352-8.
Zubenko, G.S. and Nixon, R.A., 1984, Am. J. Psycbiat., 141, 110-l 1.
 Chapter 9

Neurotransmitter action and

metabolism

9.1 Introduction

In this chapter we consider the drugs that have been developed deliberately,
or found in hindsight (serendipitously), to modulate neurotransmitter action
or metabolism. In order to understand more about the way in which drugs
can interfere with the transmission of nerve impulses we need to know how
impulses are transmitted along a nerve.
The resting nerve cell or neurone maintains a potential difference between
the cytoplasm and extracellular fluid of approximately 75 mV (cytoplasm
negative). This is achieved largely by the functioning of a membrane-bound
sodium pump acting to export positive sodium ions from the cell using the
energy of hydrolysis of ATP. The membrane is practically impermeable to
sodium ions at this potential, thus a tenfold concentration gradient is main-
tained between the nerve cell cytoplasm and the extracellular fluid.
An impulse may be generated by a variety of means such as heat, mechanical
deformation, specific chemicals or neurotransmitters etc., depending on the
type of neurone in question, and it produces a reduction in the membrane
potential difference to approximately 60 mV. When this happens, ‘gates’ or
channels in the membrane open at one end of the neurone, which are specific
for the active transport of sodium ions, causing the sodium cation to rush in
and thereby neutralize the negative charge inside the membrane
(depolarization). The neurone remains in this state of depolarization, usually
for 1 to 3 milliseconds in man, until sufficient potassium ion has left the cell
through specific channels to repolarlze the membrane. This condition is also
known as refractory because no further impulses can pass until the resting
potential of 75 mV is re-established. The intracellular concentrations of potass-
ium and sodium ion are then restored by the action of the Na+/K+-adenosine
triphosphatase which uses the energy of hydrolysis of ATP to drive an
exchange of sodium inside for potassium outside the cell. This enzyme is dis-
cussed further in Chapter 10.
The impulse or action potential passes down the nerve cell until it reaches
the end of the cell body opposite another nerve cell at a junction called a

151
152 Molecular mechanisms of drug action
Table 9.1 Drugs and their targets discussed in Chapter 9

Section Receptor/Enzyme Type Drug Use in

9.2 Adrenergic Clonidine Hypertension

Prazosin Hypertension
Propranolol Arrhythmia
Metoprolol Hypertension
Atenolol Hypertension
Salbutamol Asthma
Salmeterol
Labetalol Hypertension

9.2 Dopamine D, Bromocryptine Parkinsonism

D, L-Dopa + carbidopa Parkinson&m
D, Fenoldapam Hypertension
D2 Haloperidol Schizophrenia
D2 Pimozide Schizophrenia

9.4 Serotonin 5-HT,, Buspirone Anti-anxiety

5-HT,, Sumatriptan Migraine
5-HT, Ketanserin Hypertension
5-HT, Mianserin Depression
5-HT, Ondansetron Migraine
Granisetron

9.5 Serotonin/ Imipramine Depression

Noradrenaline Iprindole Depression
uptake Maprotiline Depression
Fluoxetin Depression

9.6 Monoamine oxidase MAOA Clorgyline Depression

MAOA Moclobemide Depression

9.7 Cholinergic Ml Pilocarpine Glaucoma

Muscarinic Ml Pirenzepin Ulcers
M I.2 Atropine Pre-anaesthetic medication
Acetylcholinesterase Pyridostigmine Myasthenia gravis
Neostigmine Myasthenia gravis

9.8 4Aminobutyrate GABA, Chlordiazepoxide Hypnotic/sedative/muscle

relaxant
GABA, Diazepam HeIrninth infection
GABA, Avermectin Spasticity
GABA, Baclofen

9.9 Opiate Pow> Morphine Pain

PL(K.4 Pentazocine Pain
PL(Kd Loperamide Diarrhoea

9.10 Histamine H, Mepyramine Allergic reactions

HI Chlorpheniramine Hay fever
HZ Cimetidine Ulcers
H* Ranitidine Ulcers

For some drugs there are additional receptors to which the drug may bind and which may contrib-
ute either to its mode of action and/or to its toxicity. Chlorpromazine, for example, is known to
bind to several other types of receptor, possibly because of its affinity for membranes, and these
effects are likely to contribute to its toxicity.
Furthermore, some of these drugs have other uses, as for example, neostigmine for the reversal
of neuromuscular blockade in surgery.
 Neurotransmitter action and metabolism 153

synapse. The cell from which the action potential arrives at the synapse is
known as the presynaptic cell. Communication with the second (post-
synaptic) cell is performed by a chemical agent known as a neurotransmitter,
specific for that synapse. The chemical is stored in vesicles immedately before
the synapse and the release of these is triggered by the arrival of the action
potential. The contents of the vesicles are extruded into the gap between the
two nerve cells, known as the synaptic cleft, and the neurotransmitter passes
across to bind to specific protein targets on the post-synaptic cell membrane,
known as receptors. This complex formation may elicit a response of either
an excitatory nature, i.e. a continuing impulse (accompanied by
depolarization) or an inhibitory response, i.e. to prevent an impulse arising in
the responding cell. The latter condition is accompanied by hyper-polarization
up to - 100 to - 120 mV, usually by the inward transport of anions such as
chloride ion, and this renders the cell insensitive to stimulation.
Clearly, the action of the neurotransmitter needs to be terminated as soon
as it is completed, and hydrolytic enzymes accomplish this action; acetylcho-
line is hydrolysed by acetylcholinesterase in the synaptic cleft, while the trans-
mitters that contain a primary amine group (aminergic) such as noradrenalin,
dopamine and serotonin are taken up into the presynaptic cell and oxidatively
deaminated to aldehydes by monoamine oxidase. Catecholamine transmitters
may also be taken up outside the neurone where catechol Omethyltransferase
renders the amine inactive by catalysing the transfer of a methyl group to the
meta-hydroxyl on the catechol ring.
For a substance to be confirmed as a neurotransmitter, two major actions
have to be seen: firstly the agent must be released from the appropriate nerve
ending by an impulse, and secondly the action of the compound when applied
to post-synaptic membranes must mimic the action that occurs when the nerve
is stimulated in the normal way.
Furthermore, a number of other factors may be taken into account. The
compound has to be present in the appropriate nerve terminals in sufficient
quantity to act as a transmitter. This means that the enzymes for its biosyn-
thetic pathway must occur in those terminals. Substances that interfere with
the binding of the natural ligand to its receptor in subcellular preparations
should have a similar action on the natural transmitter released in the normal
way. In addition, there is usually a re-uptake mechanism and/or a metabolic
enzyme for removal of the transmitter from the synaptic cleft. These conditions
have been met for all of the transmitters covered in this chapter, with the
possible exception of histamine. The similarities of drugs binding to histamine
receptor compared with ligand binding to the other receptors discussed in
this chapter suggest its inclusion here.
It should be noted that receptor binding requires the correct stereochemical
form of an agonist. In the case of the catecholamines it is the naturally occur-
ring L-form that is active while the o-isomers are at least ten times less active.
Mere binding of an agent to a membrane preparation that contains a receptor
does not guarantee that a pharmacological response will ensue. Parallel studies
154 Molecular mechanisms of drug action

on isolated tissues or whole animals should also be carried out to confirm the
receptor occupancy as a valid pharmacological result. The cloning of the requi-
site gene will confirm the existence of the receptor.
Membrane-bound receptors are discussed in this chapter (for intracellular
receptors see Chapters 6 and 7). There are various receptor families now
known: one is linked directly to an ion channel. Another very widespread
theme is the receptor linked via a guanine nucleotide binding protein (G-
protein) to an intracellular signalling system.

9.1.1 GProtein linked receptor structure

Bacteriorhodopsin is the model protein for the G-protein-linked receptors
because it is the only structure subjected to experiment. Bacteriorhodopsin is
not a receptor at all but a visual pigment, acting as a proton pump, found in
the membranes of the bacterium Halobacterium halobium. A characteristic
feature of bacteriorhodopsin is seven cY-helicesarranged in a bundle perpen-
dicular to the plane of the lipid bilayer (see Savarese and Fraser, 1992).
The structure of G-protein-linked receptors has been analysed by analogy
with bacteriorhodopsin through the hydrophobic&y of the amino acids in a
given region of the protein. The hydrophobic regions would be expected to
lie in the hydrophobic environment of the membrane, while the more hydro-
philic or polar amino acids would be more favourable sited in the aqueous
environment either inside or outside the cell. The receptors have seven
domains, comprised mainly of hydrophobic amino acids which are modelled
to span the membrane with intervening cytoplasmic and extracellular loops.
The receptor structure can be divided into three domains: extracellular
which includes the amino terminus where there is a major site of N-glycosyl-
ation and the loops between transmembrane regions 2 and 3, 4 and 5, 6 and
7; the membrane domain which consists of the seven transmembrane regions
and the intracellular area with the carboxyl terminus and the loops between
transmembrane regions 1 and 2,3 and 4, 5 and 6. They are often referred to as
7-transmembrane (TM) receptors and are common to a wide variety of ligands
discussed in this chapter including noradrenaline, dopamine, serotonin and his-
tamine.
From mutagenesis studies, the ligand-binding site is believed to lie in the
transmembrane region. The structure of the aminergic hormones, noradrena-
line, dopamine and serotonin all have a cationic amino function and an aro-
matic ring, while acetylcholine also has a positive charge based on the quatern-
ary nitrogen. The anionic carboxylate of an aspartate is probably the binding
site for this charge (Asp-113 in TM3 in Padrenergic receptors). A serine (Ser-
204 in transmembrane-5 TM5) is likely to form a hydrogen bond with the
catechol moiety in noradrenaline and dopamine and the 5-hydroxy group in
serotonin. A phenylalanine in TM6 probably stacks hydrophobically with the
aromatic ring. The different receptors are similar but sufhciently different to
bind different agonists. One of the most interesting aspects of these studies is
 Neurotransmitter action and metabolism 155

to define the molecular difference between agonists and antagonists, i.e. which
amino acid is involved in linking binding to G-protein activation. It appears
that Asp-79 in TM2 may be the crucial residue in the &receptor because the
other interactions noted above are essential for agonists and antagonists alike,
while alteration of Asp-79 only reduces agonist binding, not antagonist binding.
The third cytoplasmic loop also plays a part in the activation of the G-protein
leading to adenylate cyclase or phospholipase activation (Shih et al., 1991;
Saverese and Fraser, 1992).

9.1.2 Sigmltransduction and receptor occupancy

The chain of events that connect receptor occupancy with cell response is
not fully understood. Nevertheless it is clear that there are at least three princi-
pal methods of converting the extracellular signal to activate intracellular path-
ways. Two methods are enzymic. One involves the activation of the mem-
brane-bound enzyme adenylate cyclase which catalyzes the conversion of ATP
to cyclic adenosine 3’,5’-monophosphate (CAMP) - this in turn activates a spec-
itic protein kinase which is crucial in regulating the activity of many enzymes
by controlling their phosphorylation state. The second pathway involves
hydrolysis of phospholipids in the membrane to yield prostaglandins and leu-
kotrienes, as well as release of calcium from intracellular stores, leading to a
wide variety of physiological effects. The third method is direct stimulation of
a cation channel, whether sodium, potassium or calcium, which allows the
cation to pass into the cell (see Chapter 10 for a discussion on ion channels).
Usually a particular type of receptor will only activate one or other of these
pathways directly, although there may be interaction between receptors which
can lead to a more complex picture of pathway stimulation.
G-proteins play a crucial role in signal transduction; they act as intermedi-
aries between receptor occupancy and either enzymatic pathways or the acti-
vation of some ion channels. There are six types of G-protein - three of which
are involved in drug action: G,, the Gi proteins, and G,. G, activates adenylate
cyclase and calcium channels, the Gi family inhibit adenylate cyclase and acti-
vate phospholipase A, and potassium channels while Go regulate voltage sensi-
tive calcium channels (Birnbaumer et al., 1990). A recently discovered family
of G, stimulate phospholipase C (Sternweis and Smrcka, 1992).
The G-protein is a heterotrimer consisting of an (Y, /3 and y subunit. At rest,
G-protein is linked to the receptor and GDP is bound to the a-subunit; when
an agonist binds to the receptor, GTP is exchanged for GDP and the a-subunit
dissociates from the rest of the trimer and activates the enzyme. GTP is hydro-
lysed to GDP after one catalytic turnover of the enzyme, thus rendering the
complex inactive until GTP again replaces GDP (Savarese and Francis, 1992).
The complex of /3r is also able to regulate some forms of adenylate cyclase
(Taussig et al., 1993) and phospholipase C (Sternweis and Smrcka, 1992).
When adenylate cyclase is activated, the rate of breakdown of adenosine
triphosphate to adenosine 3’,5’-cyclic monophosphate and pyrophosphate is
 156 Molecular mechanisms of drug action

raised from a very low, almost undetectable level in some cases, and the cyclic
nucleotide is able to carry out various reactions - the most important of which
is the activation of protein kinases. These exist as inactive tetramers composed
of two catalytic and two regulatory subunits. Cyclic AMP binds to the regulat-
ory subunits causing them to dissociate from the catalytic dimer, leaving the
latter free to express its activity in phosphorylating a number of crucial pro-
teins. These include triglyceride lipase - that catalyses the initial and rate-limit-
ing step in lipolysis - and the phosphorylase kinase which activates phos-
phorylase, again by phosphorylation, to hydrolyse glycogen to glucose
(Gilman, 1984).
The process as a’type of cascade allows a considerable amplification of the
original hormone signal through the catalytic activity of adenylate cyclase and
the protein kinases. The adenylate cyclase system is outlined in Fig. 9.1, while

Cell membrane

ATP cAMP + PPf

CAMP& + 2C

R and X, receptors for hormones and neurotransmitters

G,, the inhibitory guanine nucleotide binding protein
G,, the stimulatory guanine nucleotide binding protein
AC, adenylate cyclase catalytic subunit
CAMP, cyclic AMP; PP,, pyrophosphate
C,R,, the tetramer of protein kinase A; C is the catalytic, R the regulatory, subunit

The binding of a hormone or neurotransmrtter to receptors R or X causes release of G, or

G,, thus allowing it to interact with adenylate cyclase. For every turnover of the enzyme,
one molecule of GTP is hydrolysed to GDP. The enzyme catalyses the hydrolysis of ATP to
CAMP and PP,. CAMP binds to the regulatory subunit of protein kinase thus allowing the
catalytic subunit to phosphorylate various proteins with crucial rate-limiting roles.

Figure 9.1 Adenylate cyclase modulation by receptor occupancy.

 Neurotransmitter action and metabolism 157

the receptors discussed in this chapter are listed by the signal transduction
pathway that they activate in Table 9.2.
The second major method of signal amplification that the target cell uses is
via hydrolysis of the phospholipid phosphatidylinositol that is found primarily
in the plasma membrane of the cell. When membrane receptors are activated
through ligand binding, a chain of events is set in motion whereby phospho-
lipase C cleaves the phospholipid to release, amongst other products, arachi-
donic acid which is further metabolized to prostaglandins, thromboxanes and
leukotrienes (Chapter 7).
The other products of phospholipid hydrolysis are diacylglycerol and either
inositol l-phosphate, the 1,4-diphosphate or the 1,4,5-triphosphate. Diacylgly-
cerol activates protein kinase C in the presence of calcium ion and phospho-
lipid to phosphorylate a number of proteins (Fig. 9.2). The triphosphate acts
intracellularly to induce the release of calcium from the endoplasmic reticulum
in order to produce a transient rise in the cytoplasmic concentration of the
cation.
This transient rise in intracellular calcium ion level is sufficient, for example,
to sustain the release of aldosterone by angiotensin II (reviewed in Rasmussen,
1986), but a more prolonged charge is required for smooth muscle contraction
and other cellular processes (Berridge, 1993). This is obtained by a cycling

Table 9.2 Association of ligands with signal transduction pathways

Receptor sub&pe GTP binding protein

1. Adenylate cyclase
Serotonin 5-HT,,, 5-HT,,,
Muscarinic Ma M4
GABA B
Opiate CL, 6
Histamine HZ
Dopamine D,
D2
Adrenergic
L

2. Inositol phospholipid
Serotonin 5-HT,,, 5-HT, G‘l
Muscarinic MI, M,> M, %
Histamine HI %
Adrenergic aI %

G, represents the inhibitory guanine nucleotide regulatory protein leading to inhibition of adenyl-
ate cyclase.
G, represents the stimulatory GTP binding protein.
G, activates phospholipase C.
AU the various receptors discussed in this chapter are listed, in order of the discussion, with
respect to the signal transduction pathway that they activate. Exceptions are GABA, which does
not appear to operate through either of these pathways and, moreover, unlike the receptors above,
contains a chloride channel as part of the receptor; opiate (K) and 5-HT, operates a certain channel.
158 Molecular mechanisms of drug action

Cell membrane

DG PIP*
Protein Protein

Ip3 _+ Ca” store

CT7
+
CNP e Ca”

I
+ indicates activation +
Other actions
R, receptor
PIP,, phosphatidylinositol 4,5-diphosphate
IP3, inositol 1,4,5-triphosphate
DG, diacylglycerol
PKC, protein kinase C
CNP, cyclic nucleotide phosphodiesterase
G,, GTP binding protein
PLC, phospholipase C
Binding of a ligand to the receptor activates Gq which eliminates phospholipase C to
hydrolyse PIP2 to IP, and DG. IP, causes calcium to be released from intracellular stores
such as mitochondrion and, in muscle, sarcoplasmic reticulum. DG activates PKC to phos-
phorylate various proteins and possibly to open a voltage-dependent channel to allow
calcium influx into the cell. Calcium ion may activate certain processes directly or via its
complex with calmodulin, e.g. CNP, one form of which is calcium-dependent, and often
used to assay calmodulin.

Figure 9.2 Activation of inositol phospholipid pathway by receptor occupancy

process, the details of which are not fully understood, but it is proposed that
protein kinase C is involved in the latter case because polymyxin B, an inhibi-
tor, also inhibits the release of catecholamine while phorbol diester, an acti-
vator, reverses this inhibition (Wakade et al., 1986). Indeed, modulation by
phorbol diester is usually a clear indication that protein kinase C and phosphat-
idylinositol hydrolysis are involved in a secretory process.
In the case of excitatory cells, such as those neurones from which acetylcho-
line is released, and also cardiac cells, the depolarization of the membrane is
sufficient to open voltage-sensitive channels, so that as the voltage across the
membrane falls calcium is drawn into the cell (reviewed in Rasmussen, 1986).
Calcium ion activates a number of proteins either directly or as a complex
 Neurotransmitter action and metabolism 159

with the specific binding protein, calmodulin. The latter complex frequently
acts through protein kinases but may activate other proteins directly (such
as cyclic nucleotide phosphodiesterase), thereby interacting with the cyclic
nucleotide system.

9.1.3 Drug development

Drugs can interfere with the process of synaptic transmission in various ways.
Occasionally, a false substrate can be provided for the biosynthetic pathway
of the neurotransmitter, resulting in a false neurotransmitter with altered
properties which competes with the natural agent. The mechanisms for
removal of used neurotransmitter (both re-uptake and metabolism) have been
found useful as targets for the rational development of drugs. Probably the
major effort in drug development, however, has gone into the development
of agents which act like the neurotransmitter (agonists) and those which anta-
go&e its action.
It is a feature of pharmacology in particular, and possibly science in general,
that initial concepts are simple but become progressively more complex as
more information is obtained. Neuropharmacology is no exception - for every
neurotransmitter discussed in this chapter, at least two kinds of receptor, with
different effects, have been identified, although drugs for use as agonists or
antagonists at some of the subtypes of receptor (see below) have not yet
been devised!
For many of the receptors discussed in this chapter the endogenous ligand
has been known for some time, as in the case of the receptors acted on by
the catecholamines, but for others, e.g. the opiate receptors, drug binding
studies suggested that there were tight binding sites before a natural ligand
was discovered. The opiate peptides, such as enkephalin and endorphin, were
subsequently found to fit the bill. At the present time, the diazepam binding
site, although part of, and modulatory of, the 4-(or y-)aminobutyric acid
(GABA) receptor is clearly separate from the latter, and may also have an
unknown endogenous ligand.

9.2 Adrenergic receptors

Adrenergic receptors occupy a privileged place in receptor pharmacology, in

that the first separation of hormonal action into distinct activities was pro-
posed on the basis of excitatory and inhibitory action of the catecholamines
(named because of the 1,2dihydroxybenzene or catechol moiety that they
contain) on smooth muscle (Ahlquist, 1948). These activities were termed (Y
and /3, and from this arose the concept of a specific receptor for each type
of physiological activity.
The three catecholamines [noradrenaline, adrenaline and isopropylnoradren-
160 Molecular mechanisms of drug action

aline (isoprenaline); Fig. 9.3 shows the biosynthesis of the two former agents]
differ markedly in their activities on the two receptor types. Noradrenaline
and adrenaline are the most active on (Y receptors, while isoprenaline is the
more potent agonist on p receptors. In cardiac muscle the predominant recep-
tor is the p type, which in this tissue governs the increase in rate and force
of contraction of the heart, known respectively as the positive chronotropic
and inotropic effects.
Gene cloning has revolutionized our understanding of the adrenergic recep-
tor - indeed the p receptor has often been the model. Until the middle of the
1980s there were only three known subtypes: (pi, (Y* and /3. Recently disco
vered receptors have increased this number to nine by dividing each of the
three subtypes into three, with a possible tenth in the (Y,, (Bylund, 1992;
Savarese and Fraser, 1992).
The receptor is one of the superfamily of receptors that couple to G-proteins
and is composed of seven a-helical stretches of hydrophobic amino acids that
cross the membrane separated by six loops, three extracellular and three cyto-
plasmic (for reviews see Kobilka, 1992; Savarese and Fraser, 1992). The C-
terminus lies in the cytoplasm, and the N-terminal segment with glycosylation
sites is outside the cell. The ligand binding domain lies in the transmembrane

Ho
NHz
C02H

Tyrosine

HO CH#ZHzNHz

HO Noradrenaline HO
Dopamine
I

Ho CHCH2NHCH3

HO’ Adrenaline

I: Tyrosine hydroxylase, tetrahydrobiopterin

II: Aromatic amino acid decarboxylase, pyridoxal phosphate
Ill: Dopamine /5hydroxylase, ascorbate
IV: Phenylethanolamine N-methyltransferase, S-adenosylmethionine
(Enzymes are noted together with their prosthetic groups)

Figure 9.3 Pathway of catecholamine biosynthesis.

 Neurotransmitter action and metabolism 161

region. The carboxylate group of an aspartate residue in the third transmem-

brane (TM3 domain binds the cationic amino group. Two serines in TM5 (Ser-
204 and Ser-207 in the hamster & receptor) form hydrogen bonds with the
catechol hydroxyls; TM7 appears to differentiate between pz and cyzantagon-
ists.
The coupling of receptors to G-proteins and the subsequent chain of events
leading to signal transmission is obviously of great interest. A second aspartate
in TM2 (Asp-79 in the /3* receptor) appears to be part of the agonist but not
the antagonist binding site. Furthermore, mutation of this residue greatly
reduces the ability of the receptor to activate (p2) or inhibit (a& adenylate
cyclase and therefore seems to be an integral part of this mechanism. The
third cytoplasmic loop and C-terminus has been implicated in the coupling of
the (Y, receptor to phospholipase C.
Repeated stimulation of a receptor can lead to a reduction in the signal
sensitivity (a) by removal of the receptor from the membrane (sequestration),
(b) by destruction of the receptor (down-regulation), or (c) by tachyphylaxis
through phosphorylation. The pz receptor is phosphorylated at a site in the
third cytoplasmic loop, thus impairing coupling to the G-protein.
At present we do not know the function of many of these recently disco-
vered receptor subtypes - that is a challenge for the immediate future. We
have, however, a range of relatively non-specific agents that are in therapeutic
use for high blood pressure, asthma and other disorders.

9.2.1 a-Adrenergic receptors

The detailed study of the molecular pharmacology of the (Yreceptor has lagged
behind that of the /3 receptor, possibly because selective pharmacological
modulators of the CYreceptors have only been discovered more recently. The
(or receptor mediates most of the responses to sympathetic nervous system
activation including smooth muscle contraction (Kendall, 1992). CQReceptors
are also involved in sympathetic transmission; agonists reduce the passage of
impulses to peripheral tissue. Autoreceptors at sympathetic nerve terminals
regulate neurotransmitter release. Post-synaptic CY*receptors modulate smooth
muscle tone and control sodium ion excretion in the kidney (Kobilka, 1992).
With regard to intracellular signal transduction, the or.,, receptor is unusual in
that it activates calcium channels while the aIB stimulates phospholipase C.
All three (Y* subtypes inhibit adenylate cyclase (Bylund, 1992).

q-Agonists
The catecholamines by definition are agonists at the a-receptor and do not
distinguish between the receptor subtypes. Clonidine, however, is specific for
the (Yereceptor but again does not distinguish between the subtypes.
The binding affinity of clonidine at the LY*receptor is high. Whether the
receptor is in the intact membrane of the human platelet or solubilized in the
162 Molecular mechanisms of drug action

Structures of a-receptor; agonists and antagonists

Prazosin

Cl

4
H
-
N
\ / NH I
cl N
Cl

Clonidine

Prazosin antagonists. Clonidine is an agonist.

presence of digitonin, the binding constant lies between 4 and 9 X lop9111

(Limbird et al., 1982). A similar binding constant was obtained for the rat brain
CY*receptor in membrane-bound and detergent-solubilized form (Matsui et
al., 1984).
Clonidine is used to reduce blood pressure often in conjunction with
another drug such as a thiazide diuretic (Rudd and Blaschke, 1980) or a p
adrenergic receptor blocker (Vauholder et al., 1985). The advantage of the
latter combination is that clonidine tends to induce a rapid fall in blood press-
ure while the reduction is much slower with @-blockers, so that an immediate
but also longer lasting effect can be obtained with the drug combination.

cY-Adrenergic antagonists
Blockade of the (Yereceptor antagonizes smooth muscle contraction; veins and
other small blood vessels may dilate leading to a fall in vascular resistance and
concomitant drop in blood pressure. Prazosin, an antagonist which does not
distinguish between the (Y, receptor subtypes, is used for hypertension
because of its action as an LY,blocker particularly by post-synaptic receptors.
Prazosin thereby inhibits the vasoconstriction produced by noradrenaline that
is released from the sympathetic nerve endings, while the drug’s lack of
activity on (Y* receptors allows the neurohormone to exert negative feedback
control of its own release, thus reducing the cardiac stimulation that follows
non-selective cY-adrenergic blockade (Rudd and Blaschke, 1985).
 Neurotransmitter action and metabolism 163

Prazosin binds very tightly to or receptors with a binding constant of

between 1 and 6 X 10-l’ M to membrane-bound receptors, and appears to be
some four orders of magnitude more selective for cyl than c+ receptors (see
McPherson and Summers, 1982 and references therein).

9.2.2 fiA&energic receptors

PAdrenergic receptors mediate many of the actions of the catecholamines;
the action of PI receptors includes governing the force and rate of contraction
of the heart beat, increasing secretion of renin from the kidney and antidiuretic
hormone from the anterior pituitary. pz Receptors control relaxation in lung
airways, but contraction in skeletal muscle. The & receptor may mediate lip-
olysis in fat cells. AU three receptor subtypes stimulate adenylate cyclase.

PAdrenergic antagonists
The structure of isoprenaline underwent a number of modifications in order
to devise a compound that would block the p activity of the catecholamines
and be pharmacologically valuable as an anti-hypertensive. These modifications
initially concentrated on extending the side-chain so that first three and then
four atoms were interposed between the nitrogen atom and the benzene ring.
Thus /3antagonists were derived from pagonists. The p receptor is, therefore,
the target for a number of drugs that are either agonists or antagonists. Pro-
pranolol, the first effective antagonist or P-blocker, is still used to lower blood
pressure and in the management of cardiac arrhythmias (see Chapter 10 for a
discussion on arrhythmia). The drug is, however, non-selective in that it blocks
the p response in heart, lung and a variety of other tissues non-specifically
(Lefkovitz, 1975) and this can lead to severe bronchoconstriction in the lung -
highly undesirable in an asthma patient, where propranolol should not be
used.
A search for more selective Pblockers has resulted in the development of,
for example, metoprolol and atenolol which are more active on cardiac recep-
tors (termed PI> although they still inhibit lung p2 receptors at higher doses.
For example, metoprolol is 10 to 20 times mores selective for p, while atenolol
is about three times more potent. Propranolol, on the other hand, shows no
such selectivity. There is a problem in that the concentrations of drug required
to have an effect on membrane preparations are much higher than those in
intact tissue. This has been rationalized as being a consequence of some distor-
tion of the normal receptor-enzyme coupling mechanism in the broken mem-
brane fragments (Minneman et al., 1979).

PAdrenergic agonists
Isoprenaline is the prototype of p receptor agonists and found a use in earlier
years as a treatment for asthma. It suffered, however, from a lack of specificity
164 Molecular mechanisms of drug action

PH / CH3
OCH2CHCH2NHCH
‘CH3
?H /CH3
CH30CH2CH2 OCH2CHCH2NHCH
’ CH,

5: YH,
Propranolol Metoprolol

WzCCHz OCH2CHCH2NHCH
’
CH3

CH3

since it activated receptors in the cardiovascular system and gave rise to

unpleasant side-effects, notably palpitations and occasionally cardiac
arrhythmias (uneven beating of the heart) through its positive inotropic effect.
In addition, the drug is rapidly inactivated by catechol Omethyltransferase and
thus had a very short half-life. A longer-acting, more selective agent was
required.
Accordingly, salbutamol was synthesized specifically for this purpose
(Cullum et al., 1969) and was shown to be much more specific for lung (&)
than for cardiac (&) receptors. As a consequence, in vivo the drug showed
lower activity in raising heart rate than in raising bronchial muscle tone in cats
(Apperley et al., 1976). In vitro the dose for half-maximal effect (ED,,) in
contracting the guinea pig trachea was in the region of 3 X lop9 M while ED,,
for contraction on the right atrium was ~O-‘M (Buckner and Abel, 1974).

HO
OH
! YH 73
CHCH2NHC(CH3)3 CHCHz NHCH
Ho ‘CH3

Salbutamol lsoprenaline

Salbutamol was only active for up to about six hours and a longer acting
agent was needed for night-time asthma attacks. Eventually salmeterol, a much
more lipophilic compound, was synthesized (see review by Jack, 1991) and
was found to have a much greater duration of action in man; twice daily doses
gave 24 hours coverage.
Salmeterol acts at p2 receptors in the lung, activating adenylate cyclase and
eventually producing bronchodilatation. Estimates of the ratio of efficacies of
salbutamol and salmeterol vary from O-1 to 3 according to the method of
measurement (Dougall et al., 1991; Jack, 1991). Salmeterol binds less tightly
to p1 receptors, however, than salbutamol and so is more selective. If (--)
 Neurotransmitter action and metabolism 165

Ph

HO

\
OH

CN Salmeterol

isoprenaline is taken as being non-specific and has a ratio of 1 between agon-

ism in airway smooth muscle (&) and cardiac atria &), salbutamol has a ratio
of 650, and salmeterol 50 000 (see Brogden and Faulds, 1991). With regard to
binding to & receptors, salbutamol is 50 times less tightly bound to rat lung
membranes as shown by competition with ‘251-pindolol (2.5 X lo-“M versus
5.3 X 10m8M; Coleman et al., 1990).
The extremely potent binding of salmeterol to the /3 receptor is shown by
attempts to wash the receptors free of the drug. 1251-pindolol can remove iso-
prenaline and salbutamol from membranes but has little effect on salmeterol
binding. Another unusual effect is shown by the pblocker, sotalol. Relaxation
of guinea-pig trachea by salbutamol and salmeterol is reversed by sotalol. When
the sotalol is removed, salbutamol action does not reappear, whereas that of
salmeterol does. This has been explained by suggesting that the polar head-
group of salmeterol remains bound to a site outside the ligand site, called an
exo-site, while the @blocker binds to the ligand site, and when the latter is
washed off, salmeterol reasserts its binding at the ligand site (Brogden and
Faulds, 1991; Jack, 1991).
If binding to the exo-site plays a large part in the overall binding and binding
to the ‘active’ ligand site controls the efficacy, we can see how the membrane
binding of salmeterol is so much greater than salbutamol, whereas they are
much closer in efficacy (Dougall et al., 1991).
In addition, the drug has an anti-inflammatory effect, like other /3-agonists,
through inhibiting the release of inflammatory mediators, including histamine
and leukotrienes (C,, D4 and E4) from rat mast cells and human lung fragments.
This effect was antagonized by propranolol, suggesting that p receptors were
involved (Butchers et al., 1991). Whether salmeterol has any clinically
important effect on the underlying inflammatory processes in asthma remains
to be seen.
Asthma is characterized by early and late responses to challenge by an aller-
gen, together with an increase in hyperresponsiveness of the bronchi to an
allergen; corticosteroids and cromoglycate (Chapter 10) are both able to
inhibit these effects, but this is not usually so for a short-acting Pago&t.
Salmeterol is unusual among @agonists in inhibiting these effects. Both salbuta-
mol and salmeterol are optically active but the drugs are usually in a racemic
166 Molecular mechanisms of drug action

mixture. They are normally prescribed in conjunction with steroids, as the

latter are intended to stop the underlying inflammatory process, while the
bronchodilatation produced by the @agonists gives immediate relief and a
feeling of well-being.

7.2.3 Mixed a- and j?-antagonist

Labetalol has two asymmetric centres and thus has four stereoisomers; it is
normally given as a mixture of all four isomers. They vary markedly in receptor
activity blocking both (pi, /3i and pz receptors. The R,R isomer is largely respon-
sible for the Pblocking activity and is only weakly effective at (Y receptors.
The S,R isomer is the most potent at czl receptors. The S,S isomer is relatively
weak at all three subtypes of receptor (Brittain et al., 1982). Furthermore, the
R,R isomer has agonist activity at the p2 receptor which can be blocked by
propranolol (Prichard, 1992). This is not surprising, since the structure bears
a resemblance to both noradrenaline and salbutamol, with the nitrogen atom
only two carbon atoms away from the ring and a hydroxyl group on the side-
chain o-carbon.
The advantage of this combination of multiple actions is that p blockade
reduces blood pressure by slowing the heart rate while czl blockade results
in relaxation of arterial smooth muscle. Furthermore, vasodilatation results
from p2 agonism (Prichard, 1992).

?” YH3
Ho CHCH2NHCHCH2CH2

Labetalol

9.3 Dopamine receptors

In earlier years, radioligand binding studies found only two dopamine recep-
tors: D, and D,. D, was linked via the stimulatory G-protein, G,, to adenylate
cyclase whereas D, not only inhibited adenylate cyclase via Gi, but also acti-
vated Kt (hence the hyperpolarization sometimes noted for this receptor) and
inhibited calcium channels. As with most of the receptors noted in this chap
ter, however, molecular biological methods have since discovered three more
receptors: D,, D4 and D,. D, is similar to D, in stimulating, whereas D, and
D4 relate to D, in inhibiting, adenylate cyclase. The challenge the molecular
biologists have set the pharmacologists is to find a role for these new receptors
(Sibley and Monsma, 1992).
 Neurotransmitter action and metabolism 167

The D, receptor is one of the 7-transmembrane variety with the N-terminus

lying outside the cell and the C-terminus in the cytoplasm. The N-terminus
carries three potential glycosylation sites. The C-terminus is short and the third
cytoplasmic loop between transmembrane regions (TM) 5 and 6 is large and
carries a potential phosphorylation site for cyclic-AMP-dependent kinase - a
characteristic of many receptors that inhibit adenylate cyclase. D, on the other
hand, has a small third cytoplasmic loop and a long C-terminal section in the
cytoplasm which is characteristic of receptors that activate adenylate cyclase.
As with the Padrenergic receptor, the 7-transmembrane domain carries the
ligand binding site with a conserved aspartate residue in TM2 probably respon-
sible for maintaining the active conformation, another aspartate in TM3 to bind
the protonated amino group and two serines in TM5 to interact with the two
catechol hydroxy groups.
The behavioural correlates of dopamine action are believed to be exem-
plified by the effects of dosage with apomorphine, a dopamine agonist
(dopamine does not pass the blood-brain barrier), namely a stereotypic type
of behaviour in rats that involves persistent biting, gnawing and licking, and
hypothermia in mice etc. These activities can all be antagonized by ligands at
both types of receptor. A further action of dopamine is to inhibit the pro-
duction of prolactin which appears to be under the control of D, receptors
since it is only antagonized by the neuroleptic drugs and not by SCH 23390
(Iorlo et al., 1983). Agonists and antagonists at D, and D, subtypes are used
clinically for a number of conditions, including Parkinsonism, hypertension,
and schizophrenia.

3.3.1 Dopamineagonists - L-dopaandbromocryptinefor

Parkinsonism, fenoldopam forhypertension
One of the most widely used dopamine agonists is bromocryptine, one of the
ergot alkaloids derived from the growth of the fungus Clavicepspurpurea on
damp rye. The fungus forms a purple growth - the ergot - around the gram
from which the ergot alkaloids are obtained. In the Middle Ages, ingestion of
infected rye gave rise to a number of severe reactions including gangrene,
abortion and, after high doses, death. A variety of pharmacologically active
agents have been isolated and identified from the ergot, including the related
structures, lysergic acid diethylamide and bromocryptine. The latter binds to
D, receptors in the brain and pituitary, with affinities in the region of ~O-‘M
as measured by competition with 3H-spiroperidol, a specific D, antagonist
(Creese et al., 1977).
Dopamine and acetylcholine exert balancing inhibitory and excitatory con-
trol over voluntary movements. When the dopamine neurones in a certain
portion of the brain, known as the substantia nigra, serving the basal ganglia
fail to inhibit these movements, Parkinsonism is the consequence. The brain
can suffer the loss of many of these dopamine neurones without any apparent
changes in behaviour, but when the loss reaches about 80 per cent, the symp-
168 Molecular mechanisms of drug action

tams of Parkinsonism appear, namely involuntary movements or tremors, tics,

rigidity, a gradual loss of motor function and eventually, in the late stages of
the disease, the inability to stand upright. Furthermore, the main psychic corre-
late is apathy (Bianchine, 1985). The substantia nigra governs the transmission
of motor impulses that arise elsewhere in the brain, and contains a large num-
ber of pigmented neurones with several synaptic junctions with other neu-
rones. Dopamine acts here as an inhibitory neurotransmitter.
The reason for this loss of neurones and the dopamine deficiency is not
known but a toxin present as an impurity in some forms of heroin (I-methyl-4-
phenyl-1,2,3,Gtetrahydropyridine: MPTP) can give rise to symptoms in animals
indistinguishable from Parkinsonism in man. A similar destruction of dopami-
nergic neurones in the substantia nigra seems to occur (see section 9.6 where
monoamine oxidase inhibitors are discussed). Some dopamine antagonists,
specifically the drugs used to treat schizophrenia, can also give rise to symp-
toms of Parkinsonism by blocking the post-synaptic dopamine receptors in the
basal ganglia. This condition is reversible, however, since it disappears when
the drug is withdrawn (Borison et al., 1973; Baldessarini, 1979).
The dopamine deficiency in Parkinsonism can be treated in a number of
ways. Dopamine can be supplied to the remaining neurones or dopamine agon-
ists can be used which bypass the damaged neurones and act on the post-
synaptic receptors. Alternatively, the excitatory transmission governed by ace-
tylcholine can be suppressed in order to achieve a more appropriate balance
of inhibition and excitation (Pearce, 1984; Calne, 1984). The latter approach
is discussed below (section 9.7).
Dopamine cannot be supplied directly to the brain since systemically admin-
istered catecholamines do not cross the blood-brain barrier. The precursor of
dopamine, Ldopa (L-dihydroxyphenylalanine), can however be administered
as a pro-drug (see Fig. 9.3 for the biosynthetic pathway). L-Dopa is decar-
boxylated by aromatic amino acid decarboxylase to yield dopamine. Pyridoxal
phosphate is the cofactor. Unfortunately, much of the amino acid is decar-
boxylated in the liver and only about 5 per cent is available to cross the blood-
brain barrier. Accordingly, a decarboxylase inhibitor, which itself cannot cross
into the brain is usually co-administered with L-dopa. The most useful is carbi-
dopa, which inhibits dopa decarboxylase by forming a Schiff s base between
the terminal amino moiety of its phenylhydrazine group and the aldehyde of
pyridoxal. Although daily dosage with Ldopa is often effective for a number of
years, it gradually loses its efficacy and must be supplemented in various ways.
Bromocryptine can be used in conjunction with L-dopa in order that the
dose of the latter may be reduced (see Lieberman and Goldstein, 1985, for a
review on the use of bromocryptine in Parkinsonism). The use of an additional
drug is also helpful in reducing the side-effects of L-dopa treatment, namely
psychosis, hallucinations, nausea and respiratory distress.
Another condition where bromocryptine has found a medical role is in the
treatment of infertility due to the hypersecretion of prolactin (Weinstein et
al., 1981). Prolactin secretion is suppressed by dopamine, and bromocryptine
 Neurotransmitter action and metabolism 169

Bromocryptine

can also inhibit this secretion at the level of the pituitary; this is consistent
with the view noted above that bromocryptine acts at D, receptors. High levels
of prolactin will inhibit the release of gonadotrophins in humans and, conse-
quently, will suppress ovulation. This effect is clearly of great biological value
when nursing mothers are feeding their children because it blocks the possi-
bility of another pregnancy, but in cases of inappropriate prolactin release
infertility is the result. Bromocryptine can also be used to treat the symptoms
of excessive secretion of prolactin by a pituitary tumour. In general, therefore,
bromocryptine is widely used in situations where an agent that acts like dopa-
mine is required.
Fenoldapam, an agonist at D1 receptors in the periphery (often referred to
as DAl) is under development for the treatment of malignant hypertension.
Fenoldapam dilates blood vessels in the periphery, especially in the kidney,
to induce secretion of water (diuresis) and sodium ion (natriuresis) (Hegde et
al., 1989) thus lowering the blood volume and reducing pressure. The drug
is optically active and most of the activity resides in the R isomer (Holcslaw
and Beck, 1990).

9.3.2 Dopamine antagonists - phenothiazines,butyrophenones

anddiphenylbutylpiperidines for schizophrenia
A large number of drugs that are used for the treatment of psychosis, notably
schizophrenia, are antagonists of dopamine at the D, receptor. Schizophrenia
is a complex disorder over the diagnosis of which there has been considerable
confusion in the past, and so an outline is presented here. Recent attempts to
categorize the diagnosis more accurately have apparently led to greater treat-
ment success (Manschrek, 1981). The Diagnostic and Statistical Manual of the
American Psychiatric Association (DMS-III) has laid down a number of criteria
which must be satisfied before schizophrenia can be diagnosed. These include:
(a) a period of at least 6 months in which social withdrawal, impaired self-
170 Molecular mechanisms of drug action

Ligand structures

SKF 38393

Cl

These structures contain the framework of the dopamine molecule drawn in heavy type.

awareness and markedly peculiar behaviour predominate - this is

known as the ‘prodromal phase’;
(b) an active phase with delusions and hallucinations, possibly associated
with incoherent and fragmented thinking, together with flatness of
emotional expression;
(c) a phase similar to the prodromal phase in which the symptoms subside
(residual phase);
(d) at least 6 months continuous signs of the condition in addition to (a),
and
(e) the onset must occur under 45 years of age and not be due to any
organic disorder.
The course of the disease varies widely but the active phase may be inter-
spersed with residual phases over a number of years. The active phase can be
treated and the signs greatly ameliorated by antipsychotic drugs, and in this
respect the treatment has been greatly improved. The residual phase, social
withdrawal and emotional aspects of the condition are much more difficult to
treat. The anti-psychotic agents ameliorate the condition but do not effect a
cure, although remission can occur (Manschrek, 1981).
A large number of anti-psychotic agents have been developed for the treat-
ment of schizophrenia. The phenothiazine group of drugs, of which the first
was chlorpromazine, interact with both types of dopamine receptor (and a
number of other receptors as well). In fact, these drugs are extremely non-
selective in their widespread biological actions. The more selective agents such
as the butyrophenones (e.g. haloperidol) and diphenylbutylpiperidines (e.g.
 Neurotransmitter action and metabolism 171

pimozide) interact with D, receptors more specifically, although not totally

selectively, and the relative binding affinities of the ligands correlate with their
anti-psychotic potency in man (Snyder, 1981). The term neuroleptic is often
used to describe these drugs, mainly to distinguish them from the other types
of psychoactive drugs that depress the central nervous system such as general
anaesthetics, sedatives, etc.

Phenothiazines

a;n,, QcllQF
I
dH2CH2CH2NKH31 CH2CH2CH~-N h-CH CH OH
w 22
Chlorpromazine Fluphenazine

Diphenylbutylpiperidine

‘a----2~~2~~~~ fg

Butyrophenones

Other characteristic behaviour of the neuroleptics stems from their binding

to dopamine receptors in that they reduce (a) the firing of dopaminergic neu-
rones (Bunney et al., 1973) and (b) the rates of both biosynthesis of dopamine
from tyrosine and of its metabolism to $methoxytyramine, homovanillic and
dihydroxyphenylacetic acids (Fig. 9.4) (Anden and Stock, 1973).
It is probably incorrect, however, to assume that dopaminergic systems are
overactive in schizophrenia - on the other hand a reduction in dopaminergic
activity is certainly beneficial to the patient. The effect is reversed on with-
drawal of the drug and so they are at best a palliative for the condition. An
associated feature of the action of these drugs is the slow onset of clinical
172 Molecular mechanisms of drug action

HO Dopamine

HO CHpCHzNHz
I
/
HO

3-Methoxytyramine Dihydroxyphenylacetaldehyde

CH2CH0
Ho CH2CozH

-
Homovanillaldehyde Dihydroxyphenylacetic acid

Ho
e\/ CH2C02H
III
HO 1

Homovanillic acid

I: Catechol 0methyltransferase, S-adenosylmethionine

II: Monoamine oxidase, FAD prosthetic group
Ill: Aldehyde dehydrogenase, NAD
Figure 9.4 Dopamine metabolism

effect in man compared with the rapid onset of their pharmacological action
in vitro. This suggests that, perhaps like the tricyclic antidepressants (section
9.5) secondary, indirect changes produced by the drug administration may
make a major contribution to their clinical efficacy. In addition, the use of
these drugs increases the number of D, receptors in the brain, as shown by
measurement of dopamine binding to post-mortem material (Snyder, 1981) -
an interesting result of the brain trying to nullify the effect of the drug.
Although in general the discussion of drug side-effects is not intended to be
a major feature of this book, the neuroleptic drugs show at least three types
of side-effect that are a consequence of dopamine receptor blockade in other
parts of the brain. On the one hand, the therapeutic effects of the neuroleptic
drugs result from their blockade of receptors in the projections to the limbic
system that regulates emotion in the brain. Parkinsonism, however, can result
as a consequence of post-synaptic blockade of the basal ganglia in the corpus
striatum which are served by nerve fibres from the substantia nigra (Snyder,
1986). Neuroleptic treatment for schizophrenia can, therefore, cause some
similar effects. Tardive dyskinesia, which is characteristic of Parkinsonism, and
 Neurotransmitter action and metabolism 173

prolactin release may also result as side-effects of neuroleptic treatment. The

former is characterized by a variety of movement disorders such as:
1. constant chewing movements with the tongue intermittently darting out of
the mouth (fly-catcher tongue):
2. body rocking while sitting; and
3. marching on the spot.
These symptoms take time to develop, hence the term tardive. They are
induced by receptor supersensitivity caused by prolonged drug treatment, in
other words the receptors that are not blocked by the drug are supersensitive
to dopamine @orison et al., 1983).
Prolactin release from the pituitary is inhibited by dopamine acting at D,
receptors. Dopamine blockade has the effect, therefore, of increasing prolactin
levels in the plasma leading to breast engorgement and a persistent or recur-
rent discharge of milk from the breast (galactorrhoea).
The hope that clarification of the mode of action of the neuroleptic drugs
might lead to a greater understanding of the aetiology of schizophrenia has
not yet been realized. Hyperactivity of the dopaminergic receptor is not a
feature of the condition. Various attempts to demonstrate the formation in
vivo of hallucinogenic by-products, e.g. by methylation of the neurotransmitter
amines to form such compounds as bufotenin or mescaline, have failed. Never-
theless, we desperately need a much greater understanding of a disease which
is very complex and difficult to treat.
Recent so-called ‘atypical’ neuroleptics have been developed of which one
of the leading examples is clozapin - atypical because it does not produce the
side-effects characteristic of D, blockade such as tardive dyskinesia (reviewed
in Fitton and Heel, 1990). It binds only weakly to D, receptors compared with
a variety of other receptors including D,, histamine, serotonin (5HT,) and oz.
Various suggestions have been made as to the basis of clozapin action; the D,
(Ellenbroek et al., 1991) the newly discovered D4 receptor (see Sibley and
Monsma, 1992) 5HT, (Meltzer, 1992) and the CY~adrenergic receptor (Pickar
et al., 1992) but at present the mode of action is not clear.

Me

0 N

N-
Cl

1; I)
do N
H

Clozapin
174 Molecular mechanisms of drug action

Although clozapin only shows extrapyramidal side-effects at a low level it

does, however, have the serious side-effect of agranulocytosis (i.e. it can
destroy the granulocytes - white blood cells such as neutrophils and basophils)
and under treatment the white blood cell count has to be monitored weekly.

9.4 Serotonin receptors

Serotonin (5-hydroxytrytamine, 5H’I’) is an indolealkylamine which is mainly

located in the chromaffrn cells of the gut (so-called because they readily stain
brownish-yellow with chromium salts); there are smaller concentrations in
platelets and the brain. The physiological effects of serotonin are widespread
including regulation of gastric motility and effects on platelet aggregation. The
central effects include the perception of pain, sleep, control of behaviour and
affective disorders and release of other hormones.
Recently, our understanding of serotonin receptors has greatly increased,
partly by the discovery of specific ligands for the receptor subtypes and partly

CH*
CH ‘C02H

Ttyptophan

5-Hydroxytryptophan

CH2CH2 NH2

Serotonin (5-hydroxytryptamine)

I: Tryptophan 5-hydroxylase
II: Aromatic amino acid decarboxylase

Figure 9.5 Serotonin biosynthesis.

 Neurotransmitter action and metabolism 175

by the techniques of gene cloning (see Harrington et al., 1992, for a review
on the molecular biology of serotonin receptors). There are now known to
be at least four major subtypes, 5HT, to 5HT4. The 5HT, subtype is further
subdivided into six (A to F), 5HT2 into three (A to C) and whether 5HT, and
5HT4 are homogeneous is still under study. The ligands which interact with
these receptors are numerous, but only those that are in use as drugs will be
considered in this chapter.
The 5HT receptors are coupled to G-protein-mediated pathways with the
exception of 5HT3 which is linked to a ‘fast’ monovalent cation channel - fast
because the effects occur more swiftly than with enzyme-linked pathways.
Cloning of receptors has confirmed the existence of almost all the receptors
defined by pharmacology. For example, the 5HY, receptor has been cloned
and expressed in Xenopus oocytes (large egg cells from the South African
frog). The expected cellular response to 5HT was obtained, namely influx of
cations, which could be blocked by specific antagonists (Maricq et al., 1991).
The relationship of receptor to signal pathway is listed in Table 9.3.
5HT,, and 5HT, show 49 per cent homology overall - not surprising as the
receptors couple to the same intracellular pathway - rising to over 80 per
cent in the transmembrane regions. These receptors govern different func-
tions, however, because, they are found in different sites and furthermore,
serotonin has two orders of magnitude greater affinity for the 5HT,, receptor
than for the 5HT, (Julius et al., 1990).

9.4.1 5HT,, receptors,anxiety and depression

Buspirone - the most developed member of a group of compounds known as
the azapirones - is a partial agonist at 5HT,, receptors with an Ki of
2.0 X IO-’ M with some activity at D, receptors (see Hamik et al., 1990). The
drug is used for the treatment of anxiety (anxiolytic) and is under study for
the treatment of depression (Lucki, 1991; Napoliello and Domantay, 1991).
The azapirones apparently do not act by the same mechanism as the anxiolytic
benzodiazepines, although both types of drug may reduce serotonin trans-

Table 9.3 Serotonin receptor classification

Receptor Signal Pathway Cloned (AA) 7.Transmembrane

5HT,, AC inhibited Yes (421) Ye.5

5HT,, AC inhibited Yes (??) Yes
5HT,, PC stimulated Yes (460) Yes
5HT,,> AC inhibited Yes (377) Yes
5HT, PC stimulated Yes (471) Yes
5HT, Cation channel Yes (487) No
5HT, AC stimulated No ?

Key: AC, adenylate cyclase; PC, phospholipase C; AA, number of amino acids in the cloned recep
tor.
176 Molecular mechanisms of drug action

mission. The azapirones also do not show addiction which has been recently
recognized as a serious side-effect of the benzodiazepines.

5HT,, receptors are found both pre- and post-synaptically in the brain. The
presynaptic receptors appear to inhibit the firing of 5HT neurones, i.e. they
are known as auto-receptors. It is an intriguing possibility that agonism at the
presynaptic receptors is responsible for the anxiolytic effects while partial
agonism at the post-synaptic receptors controls anti-depressive effects (Lucki,
1991). Section 9.5 discusses the relative importance of serotonin and noradren-
aline in the aetiology of depression.

7.4.2 SHT,, receptors andmigraine

Migraine is a condition of a unilateral and pulsating headache often
accompanied by nausea and vomiting and occasionally by characteristic visual
effects. There are two theories for the origin of migraine; nervous or vascular.
The vascular theory (Lance et al., 1989) proposes that blood vessels, inside
the skull but outside the brain, become distended; protein is lost across the
vessel membrane (extravasation) putting further pressure on the surrounding
tissue. These sensory stimuli activate the trigeminal (ftih) cranial nerve which
takes impulses to the pain centres in the cortex. This nerve also has connec-
tions with (a) the chemoreceptor trigger zone which is linked to the part of
the brain that controls vomiting (see below) and (b) the hypothalamus which
controls the visual stimuli (Saxena and Ferrari, 1989; Humphrey and Feniuk,
1991).
The involvement of serotonin in migraine attacks was suggested by the low-
ering of serotonin levels in the brain at the beginning of an attack. Further-
more, levels of the metabolite of serotonin, 5-hydroxyindoleacetic acid,
increase in the urine after an attack, suggesting that the hormone has been
released in the brain. Serotonin is able to constrict cranial arteries and shunts
between arteries and veins. Vasoconstrictors, such as methysergide, known to
be a (non-specific) antagonist at 5HT receptors, have been used to treat mig-
raine. Although an infusion of serotonin does alleviate migraine it produces
too many side-effects to be acceptable. The need was for a specific vasocon-
strictor to operate only on the cranial blood vessels (Lance et aE., 1989).
Detailed studies with ligands showed that the 5HT,, subtype is the most
common subtype in human brain. Sumatriptan is a potent and specific agonist
 Neurotransmitter action and metabolism 177

OH
CHzCHzNHz

Serotonin

0
Mianserin
Ketanserin

at the 5HT,, receptor subtype with a pK, of 7.54 - only five times less active
than serotonin itself. Sumatriptan can also inhibit forskolin-activated adenylate
cyclase activity in membranes from the bovine substantia nigra (Schoeffter and
Hoyer, 1989) which is to be expected as 5HT,, receptors interact with the
adenylate cyclase pathway. Furthermore, sumatriptan does not cross the
blood-brain barrier, which supports the peripheral aetiology of migraine
(Humphrey and Feniuk, 1991) and is effective even after an attack has started.

Tandospirone

MeNH

Sumatriptan

Sumatriptan is believed to act by activating the 5HT,, receptors in blood

vessels in the dura mater (the membrane surrounding the brain). The vessels
178 Molecular mechanisms of drug action

become constricted, thereby easing the pressure on the vessel wall, preventing
extravasation and stimulation of the sensory fibres of the f&h cranial nerve
(Humphrey and Feniuk, 1991). This may inhibit release of neuropeptides,
which play a role unknown as yet, blocking neurogenic inflammation.

7.4.3 5HT, receptors

5HT, receptors mediate the contraction of gastrointestinal smooth muscle, as
well as platelet aggregation and other inflammatory responses (Leysen et al.,
1984) by raising intracellular levels of calcium through the phospholipase C
pathway. Ketanserin is a powerful antagonist of serotonin binding at this sub
type (& 3 X 10-iO~), and stimulates the phospholipase C pathway. Ketan-
serin is also a weak antagonist of czl adrenergic receptors.
Serotonin causes platelets to change shape and aggregate to produce clots,
and also augments the effects of other agents such as ADP and collagen. In
addition, serotonin causes contraction of smooth muscle cells and amplifies
the contraction produced by other agents. Ketanserin can prevent both actions
(see Brogden and So&in, 1990).
Ketanserin is used to treat hypertensive conditions in which vasoconstric-
tion and platelet aggregation are due to serotonin. The relative importance of
adrenergic blockade to the action of ketanserin is still under investigation. A
close analogue of ketanserin, ritanserin, which is devoid of a,-blocking activity
is not effective in lowering blood pressure, which suggests that binding at
both receptors may be required.
Central effects of 5HT, receptor antagonists are shown by mianserin and
ritanserin. Mianserin, a clinically effective antidepressant, is also a potent
antagonist at 5HT, receptors with the +-isomer being the more effective
(Alexander and Wood, 1987). Ritanserin has been developed for depression
and as an anxiolytic.

7.4.4 5HT, receptors

The nausea and vomiting associated with a number of anti-cancer drugs, princi-
pally cisplatin, causes a great deal of distress to the patient and places much
strain on their compliance with the treatment. For some years attempts were
made to prevent emesis by co-administration of dopamine antagonists such as
metoclopramide because some forms of vomiting responded to this treatment.
This drug is only effective at high doses against cisplatin-induced emesis, how-
ever, suggesting that another receptor was involved. Eventually this was ident-
ified as the 5HT, receptor.
The nausea was believed to stem from the damage done to the epithelial
lining of the gut by the anti-cancer drug, resulting in release of serotonin from
the chromaffin cells. Serotonin activates the afferent fibres of the vagal nerve
and allows the influx of sodium and potassium, thus depolarizing the nerve
by neutralizing the normal negative charge. The vagal nerve leads to a sector
 Neurotransmitter action and metabolism 179

of the brain known as the nucleus tractus solitarius where the complex motor
activity required for vomiting appears to be controlled. Alternatively, the signal
could be relayed to this sector via a section of the area postrema, called the
chemoreceptor trigger zone, which lies outside the blood-brain barrier and
responds to toxic materials in blood or cerebrospinal fluid (Barnes et al., 1991;
Hesketh and Gandara, 1991).
The involvement of serotonin in the emetic response was suggested by the
following data:
Administration of 5-hydroxytryptophan, the precursor of serotonin (Fig.
9.5) to man resulted in nausea and vomiting.
Depletion of the serotonin stores in the gut, either by the action of reser-
pine or inhibition of synthesis, prevented the emetic action of cisplatin.
Levels of 5-hydroxyindoleacetic acid in the urine of cisplatin-treated sub
jects rose sharply during emesis (see Barnes et al., 1991).
A number of chemical synthetic programmes were undertaken to find a
serotonin antagonist which would block the action of the transmitter. Several
compounds of different structural types were eventually identified, including
granisetron, ondansetron, tropisetron and zacopride, which were specific
antagonists at 5HT, receptors and inhibited cisplatin-induced emesis (Hesketh
and Gandara, 1991; Aapro, 1991). These antagonists also control radiation-
induced emesis.

N
‘N
&’ 1\r H
N..*..
0

Granisetron

(yeyJ N

0
CH2-N
A
Me

\--/
‘N

Ondansetron

Radiolabelled versions of these agents showed the high density of 5HT,

receptors in the area postrema and nucleus tractus solitarius (Barnes et al.,
1991). These regions receive the highest density of vagal nerves which orig-
180 Molecular mechanisms of drug action

inate from the gut and so it is possible to reconcile the peripheral and central
theories of emetic action.
As shown by molecular modelling techniques, the structural requirements
for antagonism included an aromatic ring, a carbonyl group and a basic nitro-
gen at the correct distances from each other (Evans et al., 1991). An indole
ring is not necessary.
5HT, receptor antagonists are also being studied for potential use as anti-
psychotic agents, as it appears that 5HT, receptors modulate dopamine D2
receptors by reducing activity without causing sedation - unlike the neurolep-
tic agents that are D, receptor antagonists (Barnes et al., 1992). In addition,
5HT, antagonists show anxiolytic activity that differs from the benzodiazepines
in not inducing withdrawal symptoms when the drug is withdrawn.

Tropisetron

9.5 Serotonin and noradrenaline re-uptake mechanisms -

tricyclic anti-depressants
As noted in the introduction, there are uptake systems that efficiently remove
the amine neurotransmitters from the synaptic cleft in order to terminate the
signal. These mechanisms are believed to be related in the case of serotonin
to re-uptake mechanisms in the blood platelet and the latter are often used as
models of the neuronal system, even though platelets and synapses differ mark-
edly in other ways.
Depression is a condition that can be treated by different types of drug,
notably the monoamine oxidase inhibitors and the tricyclics in addition to
electroconvulsive therapy. The aetiology of the condition, in spite of intensive
research over the last 20 years, still defies a full understanding. A reduction
in activity in either serotonin or noradrenaline transmission has been variously
canvassed and the question is stiIl open at the present time.
The inhibition of monoamine oxidase A, the isoenzyme that primarily cata-
lyses the oxidation of noradrenaline and serotonin, in principle raises the brain
levels of both monoamines. It does not, however, answer the question as to
which, if either, monoamine is functionally inadequate in depression. Indeed,
there is a further question which needs to be answered; although depression
can be treated by drugs that promote monoaminergic transmission, is this a
necessary and sufficient explanation of their anti-depressive activity, particu-
larly in view of the long time interval between potentiation of monoamine
 Neurotransmitter action and metabolism 181

activity and the onset of clinical relief of symptoms? This interval may be of
the order of days or even weeks, and has led to a shift in research emphasis
from acute effects to a study of slower adaptive changes induced by chronic
anti-depressive therapy.
The discussion was initiated in 1967 by Coppen who suggested that norad-
renaline levels were associated with drive and energy, and serotonin with
mood. This view may owe a lot to the age-old connection between the cat-
echolamines and ‘fight or flight’ that is a classic feature of school biology text-
books. There are difficulties here, in that the effect of monoamine oxidase
inhibitors in animal tests that measure activity (e.g. tetrabenazine-induced
sedation) have been shown to correlate with raised levels of serotonin rather
than with noradrenaline (Christmas et al., 1972). Nevertheless, there are
always problems in extrapolating from the results of animal tests to effects in
man in any aspect of drug development, particularly in the area of depression
because of the difi?culty in designing tests that measure mood.
The debate has been thrown into sharp relief by studies on a group of drugs
known as the tricyclic antidepressants, of which the best known is probably
imipramine. These compounds bear a structural resemblance to the phenothia-
zincs, since they have three aromatic rings fused in a linear fashion as well as
a side-chain containing an amine group, but they do not affect dopamine recep-
tors. They do, however, interfere with the re-uptake of serotonin and noradren-
aline into the presynaptic nerve terminal and in some cases have a variety of
effects on receptors for other neurotransmitters. One view is that tertiary
amines tend to favour the inhibition of serotonin uptake while secondary
amines affect noradrenaline re-uptake (Carlsson, 1984). A newer group of
agents that specifically inhibit serotonin uptake but are not tricyclics are now
available, of which one of the best known is fluoxetine. Although optically
active the + and - forms are almost equiactive as inhibitors of serotonin
uptake at 2.1 X 10e8~ and 3.3 X 10m8~ respectively (Schmidt et ul., 1988).
These drugs are less toxic than the tricyclics but are just as effective
(Leonard, 1992).

Fluoxetine

On the other hand, there are other agents which are effective antidepress
ants and inhibit the re-uptake of noradrenaline selectively. One such com-
pound is maprotiline which has a secondary amine side-chain attached to a
182 Molecular mechanisms of drug action

tetracyclic structure, not greatly different from the classical tricyclics (Pinder
et aZ., 1977).

lmipramine

Maprotiline lprindole

Another approach has been to investigate the brain levels of serotonin, and
its metabolite 5-hydroxyindoleacetic acid (to indicate turnover), in the brains
of patients who suffered from depression, suicide etc. (reviewed in Goodwin
and Post, 1983). Clear indications have been found of a lowered level of sero-
tonin activity at least, and in some cases a lowered level of amine. The concept
of neuronal activity has been extended by Blier et al. (1990) who measured
the effect of all the anti-depressive treatments on the firing of brain serotonin
neurones. Both tricyclic drugs and electroconvulsive therapy sensitized post-
synaptic neurones to serotonin (the post-synaptic receptor participates in the
neurotransmission of the nerve impulse, whereas the presynaptic receptor or
autoreceptor regulates neurotransmitter release and re-uptake). Furthermore,
serotonin re-uptake blockers allowed serotonin to desensitize the autorecep-
tors more efficiently, while 5HT,, agonists activated the post-synaptic recep-
tors. In all these cases the therapy increased the neuronal activity of serotonin.
In sharp contrast Lipinski et al. (1987), have proposed that the post-synaptic
czl receptor is the common pathway for anti-depressive action as noradrena-
line-containing neurones are able to reach most parts of the brain. Chronic
stimulation of this pathway, whether directly or by interaction with other
neurotransmitter receptors, eventually leads to supersensitivity, while the
other adrenoceptors appear to be desensitized by anti-depressive treatments.
The net effect of these changes is enhanced neurotransmission, and a correc-
tion of a subnormally functioning receptor-transmitter relationship. Whether
this involves serotonin or noradrenaline depends on the drug involved, but it
is possible that there are interactions between the two systems which may
lead to a similar result after long-term administration. To discover how these
 Neurotransmitter action and metabolism 183

neurotransmitters interact is clearly of paramount importance in understanding

the aetiology of the condition and in the development of new and more effec-
tive drugs.

9.6 Monoamine oxidase - tranylcyromine and moclobemide

for depression, deprenyl for Parkinsonism

The importance of monoamine oxidase (MAO) lies in its function of destroying

excess neurotransmitter after action. Clearly, any interference with MAO
activity will prolong the action of the transmitter and such interference has
been found of therapeutic value in the treatment of depression.
The initial discovery came after two drugs, isoniazid and iproniazid, were
put on the market in 1951 for the treatment of tuberculosis (iproniazid is
the isopropyl analogue of isoniazid; the latter is still used for the therapy of
tuberculosis). Eventually, it was found that the tubercular patients became
euphoric (i.e. highly elated) on iproniazid. Depressed patients subsequently
were also found to respond. Iproniazid was eventually withdrawn from the
market in the USA because of liver toxicity, although it is still available in the
UK, and a number of other compounds were developed in its wake. One
characteristic they all have in common is the irreversible inhibition of MAO.

OH

HO CH2CH2NH2

CHCH*NH*
HO

Noradrenaline
Serotonin

, CH3
CONHNHCH

lproniazid Tranylcypromine

MAO is a flavoprotein, present in the outer mitochondrial membrane, which

catalyses the oxidation of the primary amine group of neurotransmitters (e.g.
noradrenaline, dopamine and serotonin) to an aldehyde; hydrogen peroxide
is formed in the process (Figs 9.6 and 9.7). Serotonin is thereby oxidized to
5-hydroxyindoleacetaldehyde and not-adrenaline to 3,4dihydroxymandelal-
dehyde. The mechanism requires a flavin to become reduced by abstracting
184 Molecular mechanisms of drug action

OH

OH OH
CH30 :H- CH2NH2 HO LH-CH0

HO

Ho

CH,0 LH- CH0

HO

MOPEG VMA
I: Monoamine oxidase
II: Catechol Qmethyltransferase
III: Aldehyde dehydrogenase
IV: Aldehyde reductase
VMA: Vanillylmandelic acid
MOPEG: 3-Methoxy-4-hydroxyphenylethyleneglycol

Figure 9.6 Noradrenaline metabolism.

an electron from the amine to give a radical cation; subsequent loss of a hydro-
gen from the carbon produces a carbon radical and rearrangement of the car-
bon produces an imine which hydrolyses to the aldehyde (Silverman, 1991).
Subsequent conversions yield either acid or alcohol, and in the case of the
catecholamines, methylation of one of the catechol hydroxyl groups takes
place primarily in the liver by catechol O-methyltransferase (COMT). COMT
will also catalyse the methylation of the original monoamines (the methyl
group is obtained from Sadenosylmethionine).
A detailed study of MAO was carried out using a number of acetylenic inhibi-
tors. Two isoenzymes were identified (called A and B by Johnston, 1968).
MAO-A was primarily responsible for the oxidation of serotonin and noradrena-
line and was inhibited specifically by clorgyline, whereas MAO-B was respon-
 Neurotransmitter action and metabolism 185

CH&HO

Hol&f I
N
H
II III
\

CH2 CH2 OH HO\O\NSC”&O,H

-0;;l” I

H H

5-HTP SHIAA

I: Monoamine oxidase
II: Aldehyde reductase
Ill: AIdehyde dehydrogenase
5-HTP: 5-Hydroxytryptophol
5-HIAA: 5-Hydroxyindoleacetic acid
Figure 9.7 Serotonin metabolism.

sible for the oxidation of phenylethylamine (benzylamine, although not occur-

ring physiologically, is often used as the substrate in vitro). Deprenyl was
subsequently found to be specllic for MAO-B (Knoll and Magyar, 1972). Other
naturally occurring monoamines, such as dopamine, tryptamine and tyramine
are substrates for both isoenzymes. Most tissues contain both isoenzymes but
there are cases where only one isoenzyme is present that is able to catalyse,
slowly, the oxidation of substrates of the other isoenzyme.
The mechanism underlying this specificity has been addressed by Fowler et
al. (1982) who showed that, since the structures formed with the selective
acetylenic inhibitors are similar, the selectivity must derive from a kinetic dif-
ference. The inhibitors are K,, or suicide inhibitors (Chapter 1) and the reac-
tion can be represented as follows:
k+l
E + I - EI 2 EI’ (Ki = B+,/k,)
k-1
where E and I represent the free enzyme and inhibitor respectively, EI rep-
resents the non-covalently bound inhibitor/enzyme complex and EI’ is the
covalent complex. In the case of clorgyline the Ki for MAO-A form is
5-4 X lo-’ M and 5.8 X 10m5M for MAO-B. This difference in the rate of forma-
tion of the reversible complex is sufficiently large to account for almost all
186 Molecular mechanisms of drug action

the specificity, although & for the inactivation of MAO-A is greater by one
order of magnitude than k?, for MAO-B. Deprenyl, on the other hand, owes
much of its selectivity for MAO-B to a much higher kz for that form, i.e. the
rate of formation of the irreversible adduct is much faster.

Cl
, CH3
CH?-
yH3743

CH NCH2-C=CH
OKHM’J
‘CH2C aCH Deprenyl

Clorgyline

A detailed study of the interaction of acetylenic irreversible inhibitors with

MAO has shown that the flavin portion is the target for the inhibitor. The
inhibitor initially may be oxidized to an allenic grouping that subsequently
reacts irreversibly with the N5 position of the isoalloxazine ring (Maycock et
al., 1976) to give a structure of the type:

(the flavin is linked covalently to the enzyme through a sulphur atom). In

contrast, the cyclopropylamine type of inhibitor, e.g. tranylcypromine,
although also a suicide substrate, yields a product that can be attacked by
a nucleophilic sulphhydryl group and no binding at all occurs to the flavin
(Silverman, 1983):

This adduct can be broken down more easily at neutral pH after the protein
has been denatured.
The two human isoenzymes show 70 per cent homology and derive from
two distinct genes (see Shih, 1991). Each polypeptide contains a FAD (flavin
adenine dinucleotide) residue but there is some doubt as to whether the
enzyme is a monomer or homodimer. A stretch of about 20 amino acids near
the cysteine at the C-terminus that binds the flavin is conserved, while other
conserved residues near the N-terminus may bind the adenine of FAD. The
flavin is linked to the cysteine via a thioester bond. Regions of the protein
that anchor it in the membrane have also been proposed (Powell, 1991).
 Neurotransmitter action and metabolism 187

Monoamine oxidase is also found in many other tissues and one of its princi-
pal tasks in the gut is to detoxify monoamines absorbed in the diet as, for
example, tyramine from cheese, pickled herrings, chianti etc., or dopamine in
broad beans. These amines will otherwise reach the circulation and are likely
to produce a hypertensive crisis by releasing noradrenaline from nerve endings
(section 9.2). Clearly, any drug that irreversibly inhibits MAO non-specifically,
and is given on a daily basis, will result in a greatly reduced level of the enzyme,
since it takes almost three weeks to regain its original level by re-synthesis
after a single dose. This was found to be the case with the earlier non-specific
irreversible inhibitors such as tranylcypromine, and use of these agents was
eventually restricted to a hospital environment where close supervision of the
patient could be maintained. Even the use of selective but irreversible inhibi-
tors could cause hypertensive crises in patients brought on by eating cheese.
Recent re-evaluation suggests that concerns over the ‘cheese effect’ may have
been too great. Irreversible MAO inhibitors may be of considerable value in
‘atypical’, as opposed to ‘endogenous’, depression where electroconvulsive
therapy and the tricyclics may not be effective (Bass and Kerwin, 1989).
More recently, a new class of reversible inhibitor has been described that
shows specificity for MAO-A, but does not apparently potentiate tyramine’s
ability to raise blood pressure. Moclobemide, for example, shows an IC,, for
MAO-A of 6.1 X 10p6~ and >10m3 M for MAO-B (reviewed in Haefely et al.,
1992). The inhibition for MAO-A is initially competitive but gradually changes
over time to a more tightly bound complex. Moclobemide is more effective
in vivo than these figures would predict, possibly because of the increase in
inhibition or alternatively through MAO metabolism to a more active agent.
Moclobemide is equally effective in endogenous and atypical depression but
shows only a very slight ‘cheese’ effect (Fitton et al., 1992).

0
II
o~-cH.-cHI--NH-c , >

-Q Cl

$6.1 Monoamine oxidase and Parkinsonism

An interesting recent development involving monoamine oxidase is the dis-

covery that it may be involved in the generation of some types of Parkinsonism.
Various heroin addicts injected synthetic heroin contaminated with l-methyl-
4-phenyltetrahydropyridine (MPTP). They subsequently developed irreversible
symptoms characteristic of Parkinsonism (slowness of movement, tremor and
rigidity) normally confined to those over 55 years of age. Parkinsonism
develops as a consequence of the destruction of the dopamine neurones in
188 Molecular mechanisms of drug action

the nigro-striatal region of the brain and at least 80 per cent of these must be
inactivated before the disease manifests - hence the connection with old age.
There is now considerable concern that other chemicals in the environment
may also be responsible for the development of Parkinsonism.
MFI’P itself does not produce the effect. The compound is oxidized by
MAOB, surprisingly fast for a tertiary amine, via a two electron oxidation to
the dihydropyridine (MPDP) which subsequently disproportionates into MPTP
and the aromatic iV-methyL4-phenylpyridinium ion (MPP+; Fig. 9.8). This MPP+
is taken up by the catecholamine uptake system into the nigro-striatal region
and probably reacts with the surrounding tissue. This transformation also takes
place in other parts of the brain and it is not clear why the nigro-striatal area
should either be particularly sensitive to the toxin, or should concentrate it
so effectively.
L-(-)-Deprenyl is an irreversible inhibitor of MAO-B in a similar fashion to
clorgyline for MAO-A (p. 185). L-Deprenyl first forms a non-covalent complex
with the enzyme; the flavin is then reduced while the drug is oxidized and
the product reacts covalently with the N5 position of the isoalloxazine ring
(Gerlach et al., 1992).
For much of its life, Ldeprenyl had been a drug in search of a condition. The
discovery that Ldeprenyl inhibited the conversion of MPTP to MPPf triggered
clinical trials in Parkinsonism, notably the DATATOP study (Deprenyl and
Tocopherol Antioxidative Therapy of Parkinsonism). The drug greatly pro-
longed the initial phase of the disease before Ldopa had to be given, presum-
ably by protecting the dopamine neurones against damage (LeWitt, 1991;
Tetrud and Langston, 1989).

MAO-B ?
b

MPTP MPOP MPP+

MPTP: l-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine
MPDP: l-Methy@phenyl-2,3-dihydropyridine
MPP+: 1-Methyl-4-phenylpyridinium ion

Figure 9.8 The oxidation of MPTPby MAO-B.

 Neurotransmitter action and metabolism 189

9.7 Acetylcholine action

Acetylcholine is probably the most widely found neurotransmitter in the
human body as it serves large areas of the central and peripheral nervous
systems, and mediates a variety of actions both inhibitory and excitatory. In
the autonomic nervous system, i.e. the system that governs involuntary func-
tions such as heart rate, digestion etc., acetylcholine is the predominant neuro-
transmitter. Acetylcholine is, therefore, responsible for carrying messages to a
variety of organs including heart, blood vessels, glands and smooth muscles.
CH&O,CH&H;,(CH,),
Acetylcholine
Two major subdivisions of acetylcholine receptors have been defined on
the basis of agonist and antagonist action, namely nicotinic and muscarinic.
The former are characterized by the action of nicotine as an agonist and of
tubocurarine as an antagonist, and are located in autonomic ganglia (i.e. where
two neurones or sets of neurones interact), skeletal muscle and some of the
synapses in the central nervous system. The nicotinic receptors on ganglia and
skeletal neuromuscular junctions differ, however, in some respects; tubocurar-
ine blocks both but its action is much more marked on the latter, and there
are other more selective antagonists.
Muscarinic receptors, on the other hand, are found in smooth muscle, car-
diac muscle and glands, and are the predominant form of the acetylcholine
receptor in the central nervous system. Muscarine and pilocarpine are agonists;
atropine and hyoscine antagonists.

7.7.1 Muscarini c receptor - pirenzepin for ulcers, atropine for

pre-anaesthetic medication, pilocarpine for glaucoma
Gene cloning has emphasized the heterogeneity of muscarinic receptors, as
with other receptors discussed in this chapter. Five genes have been found to
encode muscarinic receptors where two had originally been defined by ligand
affinities. As in other cases, not all the proteins expressed have pharmacologi-
cal functions yet assigned to them. The muscarinic receptors mediate a wide
variety of functions including phosphoinositol hydrolysis, adenylate cyclase
inhibition, mobilization of calcium from intracellular stores and modulation of
potassium channels (reviewed in Hulme et al., 1990; Lambert et al., 1992).
Muscarinic receptors are one of the many groups of receptors that are
coupled to G-proteins and structurally have seven transmembrane (TM)
regions. As with the other receptors, the N-terminus is extracellular and the
C-terminus cytoplasmic thus forming six loops, three inside and three outside
the cell. Agonists bind primarily to TM3, TM6 and TM7, although there is a
suggestion that other transmembrane domains may help to form a hydrophilic
ligand binding pocket OIulme et al., 1990). The third intracellular loop and
possibly the C-terminus are involved in G-protein binding.
190 Molecular mechanisms of drug action

Agonists bind to an aspartate in TM3 and a tyrosine in TM7 which are

believed to hydrogen bond in the ground state. When the amino group of the
agonists binds to the aspartate carboxyl, the hydrogen bond is broken and a
conformational change may result, allowing the G-protein to bind (Hulme et
al., 1990). Another aspartate residue is conserved in TM2 and is involved in
signal transduction, as mutation greatly reduces the ability of the Ml receptor
to activate phospholipase C. Two cysteines in the first and second extracellular
loops form a disulphide bridge which may act to stab&e the receptor
(Saverese and Fraser, 1992).
G-Protein Gi couples muscarinic receptors M, and M4 to adenylate cyclase in
an inhibitory fashion, while (probably) G, links M,, M, and M, to phospholipid
hydrolysis. The results for agonists binding to receptor have given a wide range
of results, probably due to varying levels of guanine nucleotides because the
presence of guanine nucleotides can reduce the affinity of the agonist for the
receptor markedly. The G-protein-receptor complex has a much higher aftin-
ity for the agonist than the receptor alone. This may reflect the energy of
another hydrogen bond between, possibly, the ester function of acetylcholine
and an amino acid residue. Antagonists do not show this range of effect (Hulme
et al., 1990).
The M,, M,, M, and M, receptors are normally found in neural tissue and
the M, and M, in exocrine glands as well. M, is a cardiac and smooth muscle
receptor. Receptors are found both pre- and post-synaptically, the presynaptic
receptors or autoreceptors (usually M,) inhibit the release of acetylcholine.

Muscarinic antagonists
Pirenzepin as a relatively selective ligand compared with atropine (see below)
has been used for the treatment of peptic ulcers (Carmine and Brogden, 1985).

Pirenzepin

Some side-effects were noted, characteristic of anti-muscarinic drugs and due

to blockade of the parasympathetic nervous system, notably dry mouth and
blurred vision, but they were of low incidence and severity. The lack of central
nervous system effects suggests that although the drug has a great affinity for
neuronal receptors, it does not cross the blood-brain barrier because it is
too hydrophilic.
As well as stimulating peristalsis in the gut, acetylcholine releases gastric
 Neurotransmitter action and metabolism 191

juices from the glands by activating M, receptors, both on postganglionic neu-

Tones and the parietal cells of the gastric mucosa. Since the latter have a low
aflkity for pirenzepin, the drug target is believed to be the postganglionic
neurone receptor (Goyal, 1988). Pirenzepin shows some selectivity for the M,
receptor with a binding constant of 4.6 X lo-* M to M, and -3 X lo-’ M for
M, (Lambert et al., 1992). It is interesting that efficacy and binding studies give
results that are very close unlike the p2 agonists noted earlier in this chapter.
DL-Atropine, an alkaloid derived from Atropa belladonna or deadly night-
shade, is non-selective but a very tight binder nevertheless, having a binding
constant in the region of 10-‘“~; the levorotatory form is the more active.
Atropine is used to counteract the effects that stem from the stimulation of
the vagus nerve during anaesthesia, and so is usually part of operation premedi-
cation.

Atropine

Muscarinic agonists
One type of glaucoma is characterized by a build-up of water in the aqueous
humor of the eye, a condition which is particularly prevalent in the elderly.
The iris and the cornea are normally separated by a channel through which
the aqueous humor can filter and be dispersed by absorption into the nearby
blood vessels. In later life the iris can be pushed forwards, possibly because
of inflammation or increasing size of the lens, thus closing the channel. The
aqueous humor can no longer filter out and pressure on the eyeball increases,
with particular danger of destruction of the optic nerve fibre leading to blind-
ness.
Pilocarpine, an alkaloid derived from the leaves of the South American shrub
Pilocarpus, is one of the drugs recommended for this condition (Shapiro and
Enz, 1992). The drug acts as a miotic, i.e. it narrows the pupils by causing
them to contract, thus drawing the iris away from the cornea and allowing
the fluid to dram away. The active form of pilocarpine is the isomer where
the side-chains at C-2 and C-3 of the imidazole ring are in the cis position.
Pilocarpine does not distinguish between the species of muscarinic receptor,
with binding and efficacy measured on ganglion depolarization (M,) and
guinea-pig ileum (M3) in the region of lo-6111 (Shapiro and Enz, 1992).
192 Molecular mechanisms of drug action

+
CH~N(CHB)J

Muscarine Me

Pilocarpine

9.7.2 Acetylcholinesterase -pyridostigmineformyasthenia

glWiS

Closely associated with cholinergic synapses is the enzyme acetylcholinester-

ase, which rapidly hydrolyzes the neurotransmitter and is essential to ensure
that the signal is terminated as soon as it has passed. The enzyme is attached
to the lipid environment of the membrane via a link with inositol phosphate
and diacylglycerol because phospholipase C will liberate the enzyme from its
normal membrane-bound state (reviewed in Low et al., 1986).
Acetylcholinesterase hydrolyses the ester link of the neurotransmitter to
yield acetate and choline with a very high turnover number. The enzyme has
two welldefmed parts in its binding site for substrates; one anionic which
binds the cationic head of the substrate (and of inhibitors), and an esteratic
site at which hydrolysis takes place with the formation of an acyl-enzyme inter-
mediate. There is apparently another anionic binding site which can be occu-
pied by bis-quaternary ligands (see the review by Soreq et al., 1992). A charge-
relay system similar to that defined for trypsin, chymotrypsin and elastase with
glutamate carboxyl, histidine imidazole and serine hydroxyl is proposed at the
active site of acetylcholinesterase.

Glutamate Histidine Serine

As with the proteases, the serine hydroxyl is activated by the charge relay
system sufficiently to be a strong nucleophile, thus forming a transient acyl-
enzyme intermediate, which is easier to detect with acetylcholinesterase as
the complex dissociates more slowly.
A family of compounds with structures related to acetylcholine and includ-
ing a cationic head group and carbamate ester, inhibit acetylcholinesterase by
acting as false substrates, but with very low turnover numbers, so that the
true substrate is unable to be hydrolysed. One of the best known is pyridostig-
 Neurotransmitter action and metabolism 193

mine which inhibits acetylcholinesterase by binding to the anionic site with

its cationic pyridine nitrogen, while the carbamate ester is available to be
hydrolysed by the esteratic site. The tetrahedral complex formed on hydrolysis,
a dimethylcarbamoyl-enzyme, resembles the transition state of the normal reac-
tion but is extremely slow to hydrolyse and prevents the hydrolysis of the
normal substrate. The drug is therefore a type of suicide inhibitor. The IC,,
for the reaction is 1.6 X lo-“M (Wilson et al., 1961) which is effectively k,k,
in the following scheme:

CH3

CH3

f
OCN(CH312 + HO-E -
::
OH +(CH3)2NC-O-E

f
(CH&NCOH
+HO-E

Inhibitors of acetylcholinesterase would be expected to potentiate the

action of acetylcholine and this has been made use of in the development of
the organophosphate insecticides. In medicine, less toxic inhibitors have been
used in conditions where the receptors are damaged or diminished in number,
as in myasthenia gravis. This condition is a genuine auto-immune disease. The
damage is caused by auto-antibodies to the nicotinic acetylcholine receptors
at the motor end-plate where the nerve connects with striated muscle. The
receptor has been cloned and consists of five subunits, CQ, p, E and S. The
auto-antibodies bind to the (Y subunit as a type of antagonist preventing the
binding of acetylcholine; the transmission of stimulatory signal from nerve to
muscle fails resulting in muscle weakness (reviewed in Vincent, 1991). In
addition, a degree of membrane lysis occurs, and a considerable amount of
debris is found in the synaptic cleft. The net result is a widening of the cleft
and elongation of the synaptic membrane, further reducing neuromuscular
transmission @adding and Havard, 1981).
Myasthenia gravis is most obviously recognized by muscle weakness which
characteristically starts around the eyes and face and then works its way down
via the limb girdle, outer limbs and trunk in that order. There is a marked
fatigability of skeletal muscle, in that if a motor nerve of a patient is stimulated,
194 Molecular mechanisms of drug action

synaptic transmission rapidly falls away, unlike the results obtained with a
normal subject. Pyridostigmine and neostigmine are currently widely used for
this condition (Havard and Fonseca, 1990). They compensate for the deficit
of acetylcholine receptors by slowing down the rate of degradation of acetyl-
choline and so increase the duration of its activity.

Pyridostigmine Neostigmine

9.8 4-Aminobutyric acid receptor - benzodiazepines as

hypnotics, avermectin as anthelminthic, baclofen for
spasticity

4Aminobutyrate (y-aminobutyrate; GABA) is a neurotransmitter that acts at

inhibitory synapses, often presynaptically, i.e. the inhibitory nerve terminal
secretes GABA on to a neighbouring excitatory nerve terminal and thereby
reduces the output of the excitatory transmitter (often acetylcholine). By this
means, specific neuronal pathways converging on another neurone can be
eliminated without influencing the efficacy of others. The relevant biosynthetic
pathway is found in GABA neurones, as found with other neurotransmitters;
GABA neurones can synthesize GABA from glutamate with the help of gluta-
mate decarboxylase (Fig. 9.9).

4-Aminobutyrate

Bicuculline

GABA receptors are divided into two major types. GABA, receptors are
unusual in that they contain a chloride ion channel in the receptor itself. The
binding of agonists thus allows negative ions to pass into the cell and increases
the negative membrane potential (hyperpolarization) so rendering it unable to
transmit the nerve impulse or action potential. The consequence of this action
is to induce muscle relaxation, to reduce anxiety and to block convulsions.
 Neurotransmitter action and metabolism 195

NHzNHCONHs NH20CHzCOOH
Semicarbazide Aminooxyacetic acid

FOOH -
H2NCHCH2CH2COOH H2NKH2)&OOH OHCCHsCH,COOH

Glutamic acid I. 1 {. 2
Succinic semialdehyde
4-Aminobutyric acid

3
1, Glutamate decarboxylase
2, 4-Aminobutyrate aminotransferase
3, Aldehyde dehydrogenase
1
-, inhibition
HOOCCH2CH2COOH

Succinic acid
Figure 9.9 The pathways of GABA synthesis and breakdown.

Sedation may also be a consequence @anger, 1985). GABA* receptors are dis-
cussed under baclofen (see section 9.8.3).

Baclofen Muscimol

7.8.1 Benzodiuepines

A well known family of psychoactive drugs used as hypnotics, sedatives and

tranquillizers, the benzodiazepines, that were originally marketed in the late
1960s mainly for the treatment of anxiety, were subsequently discovered to
bind very tightly to specific sites in the brain (Speth et al., 1978). A consider-
able body of evidence linked this site with the GABA,, receptor and, not for
the first time, the experimental use of a drug advanced the state of scientific
knowledge, having initially been of therapeutic value (Sanger, 1985). The best
known benzodiazepines are probably the anxiolytic (tranquillizer) drugs
diazepam and chlordiazepoxide (these are generic names; the drugs may be
more familiar under the trade names of Valium and Librium).
Benzodiazepines act in such a way as to simulate the effects of exogenously
administered GABA, but if GABA levels are depleted there is no such effect.
This occurs, for example, when semicarbazide is co-administered; glutamate
decarboxylase, an enzyme with pyridoxal phosphate as a prosthetic group, is
inhibited through the formation of a Schiff s base and GABA levels are lowered.
196 Molecular mechanisms of drug action

Conversely, benzodiazepine effects are potentiated by aminooxyacetic acid,

an inhibitor of GABA oxidation to succinic semialdehyde by GABA amino
transferase.
The best characterized benzodiazepine ligand is flunitrazepam which binds
to homogenates of human brain with a binding constant in the range l-
3 X 10-9~ (Speth et al., 1978). This was measured both by kinetic analysis
of the rates of association and dissociation, and also by Scatchard analysis.
Inhibition of flunitrazepam binding by a number of benzodiazepines correlated
well with a number of pharmacological tests (notably inhibition of convulsions
induced by pentylenetetrazole in mice and muscle relaxation in cats) and also
with the dose used to treat anxiety in man. Braestrup et al. (1977) obtained
similar results using 3Hdiazepam binding to a membrane preparation from
human brain. They also showed that GABA receptors varied in density between
different parts of the brain and this density paralleled the distribution of benzo-
diazepine receptors.

’0
CH3

7
3 -N
Rl

75
RI

Rl R2

Flunitrazepam -NO2 F
Chlordiazepoxide Diazepam -Cl H

Conversely, the binding of GABA and GABA agonists, such as muscimol,

potentiates the binding of benzodiazepines to their receptors, an action which
can be reversed by the GABA* antagonist bicuculline (Blaestrup et al., 1980).
The receptor complex has been cloned and expressed in Xenopus oocytes to
give receptors that control chloride channels; there are three families of subun-
its ((.u, p, 7). There are six versions of an (Ysubunit, two of the j3 subunit and
two y subunits, and possibly others. Benzodiazepine binding occurs on a
trimer ((~i, a2, a3 or as) with p2 and y2. The o! subunit carries the drug site,
although the y subunit is also required for maximum effect. The GABA site
resides mainly on the /3 subunit, although there is some overlap between them
(Sieghart, 1989; Vi&i, 1991). The variation in the (Y subunits gives rise to
GABA* receptors with differing GABA sensitivity which may relate to pre-
viously defined benzodiazepine receptors, although the receptors all control
GABA-gated chloride channels (Sieghart, 1989).
A model has been proposed for the receptor that contains four transmem-
brane regions and bears a resemblance to another gated ion channel, the nic-
 Neurotransmitter action and metabolism 197

otinic receptor (Vi&i, 1991). A histamine residue appears to be crucial for

the binding of benzodiazepines (Wieland et al., 1992).
Other studies, however, have cast doubt on this behavioural correlation
(Sanger, 1985). If rats are trained to respond to food and drink but responding,
in addition, is linked to an aversive stimulus such as an electric shock, benzodi-
azepines greatly increase what would otherwise be a low rate of response.
Furthermore, the magnitude of the drug effect correlates highly with the clini-
cal potency in treating anxiety. Not all agonists increase the rate of response
when given systemically, but do if injected into the appropriate part of the
brain. These differences may arise as a consequence of non-uniform distri-
bution of agents in the brain, or it may be that some of the anxiolytic actions
of benzodiazepines are not involved with GABA - unlike the muscle relaxant
and anti-convulsant actions (Sanger, 1985).

9.8.2 Avermectin

GABA acts as a neurotransmitter in a number of other organisms besides mam-

mals, including arthropods and nematodes, which are themselves pathogenic
for man. Conspicuous among these infections is one caused by a nematode
worm, Onchocerca VOZVUZUS, a native of West and Central Africa, which infects
man to produce the disease known as river blindness or onchocerciasis. The
worm in the larval form (microfilaria) is ingested from humans by a bite from
the blackfly of the genus Simulium, which breeds in fast-flowing streams.
After about a two-week incubation, the larvae can be injected by bite back
into man in an infective form. The microfilaria migrate in the surface tissue
where they are found in nodules, particularly around the joints, and to the
eye where they usually cause blindness by attacking the optic nerve. The larvae
mature over 1 to 3 years into adult worms (macrofilaria), the female of which
can release thousands of microfdariae daily. These migrate to the surface tis-
sues and the eye and either degenerate or are ingested by the blackfly bite.
The blackfly is the vector without which the worm could not attain such high
numbers, as neither the larvae or adult worms multiply in humans (Goa et al.,
1991). The only drugs that have been used are suramin, which is extremely
toxic, and diethylcarbamazine, which is less toxic but is liable to kill the adult
worms swiftly and precipitate an extreme allergic reaction by virtue of the
large amount of nematode protein rapidly entering the circulation. This reac-
tion can cause more damage than the nematode itself, particularly in the eye
(Mazzotti reaction).
Recently, however, a major improvement has been effected by the use of
ivermectin. This is a member of the avermectin family of compounds, first
obtained from the fermentation of Streptomyces avermitilis (Egerton et al.,
1979). The avermectins which possess a 16-membered lactone ring with a
spiroacetal system involving C-17 to C-25 and a disaccharide substituent at
position C-13 (see structures for details). The naturally occurring avermectins
198 Molecular mechanisms of drug action

fall into four major (A,,, Aza, Bi, and B,J and four minor (Alb, Azb, Bib and
B2,,) series (Campbell et al., 1983; Fisher and Mrozik, 1992). Ivermectin is

OCH,

A 0'
R, = CH3 CH3

a b
R2= W-b CH3

1 2
x= -CH=CH- -CH-CH-

OH
Avermectins

/OH
R= -T CH3 : Picrotin
CH3

/CH3

R= -C : Picrotoxinin
\\
CH2

Picrotoxin (1 mol picrotin:l mol picrotoxinin)

synthesized from avermectin B, by reduction of the 22,23 double bond (see

structure). Ivermectin is now regarded as the front-line treatment for river
blindness. Annual doses of 150 pg/kg are sufficient to reduce the microfilarial
load by SO-99 per cent in the surface tissues, but more slowly than diethylcar-
bamazine, so the Mazzotti reaction is much less marked. The drug does not
kill the adult worms, but it prevents the female from releasing the larvae and
they gradually degenerate in the womb. As a result, the blacktIy transmits the
parasite in far smaller numbers. Another advantage is that the infrequent dose
is much easier for health workers to organize in the African bush (Goa et
al., 1991).
 Neurotransmitter action and metabolism 199

Ivermectin was originally introduced to the veterinary market for the treat-
ment of internal parasites (endoparasitic nematodes) of mammals of economic
importance such as sheep, cattle and horses and of external parasites
(ectoparasites), namely arthropods such as ticks and mites. Lobsters are also
sensitive to the drug! The common link between these organisms is their pos-
session of neuronal synapses at which GABA acts as a transmitter (see
Campbell et al., 1983; Wright, 1987 for reviews on avermectin and its biologi-
cal action).
Most of the experimental studies have been carried out with avermectin Blar
referred to as avermectin below. In Ascaris suum, a relatively large nematode
that infects pigs, there are at least two sites of drug action: one is a block
between a nerve cell (known as an interneurone) that forms a synapse with
an excitatory motor neurone on the dorsal side of the nematode and leads to
depolarization of the muscle membrane. The second site is at the neuromuscu-
lar junction of an inhibitory motor neurone on the ventral side and muscle,
the activation of which leads to hyperpolarization. At the former site the action
of 6 X lo-6111 avermectin B,, in reducing the depolarization can be mimicked
by the agonists muscimol and piperazine, and reversed by the antagonist picro-
toxin, which indicates that the interaction is with the GABA* receptor (Kass
et al., 1984). On the other hand, avermectin’s action in reducing hyperpolariz-
ation at the second site does not respond to picrotoxin reversal and is presum-
ably nothing to do with GABA* receptors.
In sharp contrast, at 1 X lo-’ M concentration, the drug appears to antagon-
ize the action of GABA in reducing the length of time that ion channels are
open, and the probability of their opening - measured by using microelec-
trodes implanted into the nematode (Martin, 1987). The apparent contradic-
tion between these findings remains to be resolved, although the use of a low
physiological level of the drug may be crucial.
The arthropod target for avermectin is also a neuromuscular junction as
shown by experiments in lobsters where the action of the drug was synergistic
with GABA. The neuromuscular junction is known to be controlled by one
excitatory axon, with glutamate as the transmitter, and one inhibitory axon,
with GABA as the transmitter. Avermectin reduced both excitatory and inhibi-
tory post-synaptic potentials, probably by opening chloride channels and hyp-
erpolarizing the membrane. The response was, restored by picrotoxin. It
appears likely that avermectin acts as a GABA agonist either by potentiating
the binding of GABA to its receptor (Campbell et al., 1983) and/or by releasing
GABA presynaptically. The net effect of drug on both nematode and arthropod
is muscular paralysis.
In vitro studies indicate that avermectin B,, can stimulate the high affimity
binding of GABA to rat brain membranes by increasing the number of available
binding sites, rather than the binding constant, at an IC,, of 7X lO-6 M. This
effect is chloride-iondependent and is antagonized by picrotoxin and bicucul-
line (Pong and Wang, 1982). Avermectin cannot reach these sites in vivo as
it does not cross the blood-brain barrier. Avermectin also enhances the bind-
200 Molecular mechanbns of drug action

ing of diazepam to rat brain membranes at micromolar concentrations, and

potentiates the pharmacological activity of the tranquillizer, notably a decrease
in motor activity and muscle power (Williams and Yarbrough, 1979).
More recent studies in locust leg muscle (the exterior tibiae) have shown
that there are at least two other chloride channels which can be modulated
by avermectin besides that connected to the GABA receptor: one is probably
linked to glutamate receptors and the other is not linked to any known recep-
tor (Duce and Scott, 1985). Other avermectin binding sites have also been
found on chloride channels in non-pathogenic arthropods. In crayfish stomach
muscle there are sites that are extremely sensitive to the drug at concentrations
below ~O-‘*M. The channels can be opened directly and reversibly at these
levels and irreversibly above 10-l’ M. This effect can be blocked by picrotoxin
but is not sensitive to GABA. Avermectin is thus acting as an inhibitory neuro
transmitter without the need for the mediation of a second messenger system
(see Fisher and Mrozik, 1992).
In conclusion, the relatively simple picture of avermectin acting to potenti-
ate GABA requires some qualification. In some cases this view is correct, but
in other situations avermectin may act as an antagonist, and in yet others may
have actions which do not relate to GABA action but may still be connected
with chloride channel opening.

9.8.3 Baclofen

GABA, receptors differ from GABA* in that (a) no chloride channel is involved,
(b) they are insensitive to bicuculline and muscimol (Allan and Harris, 1986)
and (c) the benzodiazepines do not bind. The best known ligand at the GABA,
receptor is the anti-spastic agent baclofen. Baclofen is 3-$-chlorophenyl-GABA
and was synthesized as a more lipophilic analogue to pass the blood-brain
barrier and mimic the actions of GABA. t.-Baclofen is more potent as an agonist
at GABA, receptors than the o-isomer by a factor of about 1000 (reviewed in
Wojcik and Holopainen, 1992).
Like GABA* receptors, GABAt, are found presynaptically and can thus inhibit
the release of several neurotransmitters. The anti-spastic action of baclofen
probably derives from inhibition of many of the sensory fibres bringing
impulses into the dorsal side of the spinal cord at vertebrae I to IV (Allerton
et al., 1989). Normally the nerve impulse originating from the sensory neurone
is passed either directly or via interneurones to the motoneurone which then
activates the muscle reflex. Spastic neurones are hyperactive. The drug
reduces the hyperactivity by lowering the excitatory post-synaptic potential
on the motoneurone through two possible mechanisms: acting presynaptically
to reduce the release of the excitatory transmitter from the sensory nerve
fibres, or post-synaptically by hyperpolarizing the neurone.
The detailed mechanism of baclofen is not yet fully understood, but in some
cells the drug causes hyperpolarization (and thus prevents transmission) by
stimulation of an (outward) K+ current; in other cells baclofen inhibits an
 Neurotransmitter action and metabolism 201

inward calcium channel (Bowery, 1989). These actions are both mediated by
G-proteins and could explain the inhibition of nerve transmission. Inhibition
of adenylate cyclase and activation of inositol phospholipid hydrolysis, how-
ever, have also been described (Wojcik and Holopainen, 1992) but the rel-
evance of these to the anti-spastic action of baclofen is not clear.

9.9 Opiate receptors - morphine for pain

The opiates are a class of naturally occurring opium alkaloids of which the
best known is probably morphine. Opioids are chemically synthesized agents
such as meperidine, methadone and naloxone which mimic or antagonize the
pharmacological actions of morphine. Both opiates and opioids bind to recep-
tors designated as opioid.

N- CH&H =CH,

Morphine Neloxone

.Drugs of this type are powerful pain killers (analgesics); morphine, in par-
ticular, derived from the opium poppy, has been in use for this purpose for
a very long time. The sort of pain that these agents are used to treat is the
chronic nerve pain that often accompanies terminal cancer and recovery from
operations. Less severe pain is countered by the use of non-steroidal anti-
inflammatory drugs such as aspirin, which inhibit the biosynthesis of prosta-
glandins (Chapter 7).
A large number of analogues have been synthesized in order to improve
on the properties of morphine by reducing undesirable sideeffects such as
drowsiness, nausea, vomiting, constipation and, most serious of all, respiratory
depression that can be lethal if a sufficiently high dose is given. In addition,
there is also the grave problem of physical dependence on the opiate family
of drugs.
202 Molecular mechanisms of drug action

There are at least three and possibly four major types of opioid receptor:
CL, K, 6 and possibly U. Stimulation at all four receptors appears to relate to
relief of pain. Morphine and its analogues are primarily Al.agonists although
they often have appreciable activity for K and 6 receptors. The /.J receptor
controls respiratory depression (slowing of the rate of breathing), reduced
intestinal motility, nausea and vomiting and also drug dependence. K receptors
are primarily concerned with diuresis by inhibiting the release of anti-diuretic
hormone and sedation. The 6 receptor, which responds more markedly to
opioid peptides such as enkephalin than opioid alkaloids, appears to exist in
the guinea-pig ileum but not in the mouse vas deferens.
Some opioids also bind to other receptors known as o, which appear to
mediate disorientated and depersonalized feelings; the o receptor may be
regarded as non-opioid, as morphine has a very low affinity and naloxone, the
classical opioid antagonist, does not block morphine’s action. There appears
to be some cross-over between u and dopamine D,, however, because a num-
ber of atypical neuroleptics, including haloperidol, bind to both receptors in
the region of lo-’ M (Bowen et al., 1990).
Opioid receptors couple to G-proteins (mainly the Gi family - see Childers,
1991) although p may also couple through G,. II, K and 6 receptors inhibit
adenylate cyclase. In common with other G-protein coupled receptors GTP
reduces agonist, but not antagonist, binding.
p and 8 agonists stimulate the opening of inward potassium currents, thus
hyperpolarizing the membrane and reducing the rate at which signals are trans-
mitted, while K agonists close calcium channels. These actions seem to be
mediated directly by G-proteins and not through adenylate cyclase inhibition.
The opioids are transmitters at inhibitory synapses, usually by binding to recep-
tors located presynaptically and suppressing the release of excitatory transmit-
ters. p appears to inhibit noradrenaline release, K that of dopamine and both
p and 6 that of acetycholine (Schoffelmeer et al., 1992).
The finding that there were saturable, stereospecitic binding sites for opioid
drugs in the central nervous system prompted a search for endogenous ligands
to bind at these receptors. Very soon two naturally occurring peptides (the
enkephalins) were discovered in pig brain. Structurally they are very alike and
are known as Met-enkephalin (Tyr-Gly-Gly-Phe-Met) and Leu-enkephalin (Tyr-
Gly-Gly-Phe-Leu), differing only at the carboxyl terminus (Hughes et al., 1975).
Subsequently other, larger opiate peptides were isolated from the pituitary
gland and were called endorphins to distinguish them from the enkephalins
(Li et al., 1976; Chretien et al., 1976). Figure 9.10 shows opioid structural
relationships. It is interesting that Met-enkephalin forms the amino-terminus
of the endorphins, which in turn represent residues 61-91 at the carboxyl
terminus of the pituitary hormone, /%lipotrophin, although Plipotrophin does
not possess any opiate properties.
A further 17-amino-acid, peptide was isolated by Goldstein et al. (1977) and
named dynorphin. Dynorphin is Leu-enkephalin with an extension of 12 amino
acid residues at the carboxyl terminus. There are at least four peptides with
 Neurotransmitter action and metabolism 203

Aminoacid sequence of the opioid peptides.

Precursor Peptides Sequence

Pro-opiomelanocortin x-Endorphin Tyr-Gly-Gly-Phe-Met-Thr-Ser-GIu-Lys-

Ser-Gln-Thr-Pro-Leu-Val-Thr-Leu:
17 residues

H -Endorphin Tyr-Gly-Gly-Phe-Met-Thr-Ser-GIu-Lys-
Ser-Gln-Thr-Pro-Leu-Val-Thr-Leu:
17 residues

p-Endorphin Tyr-Gly-Gly-Phe-Met-Thr-Ser-GIu-Lys-
Ser-Gln-thr-Pro-Leu-Val-Thr-Leu-Phe-
Lys-Asp-Ala-lle-lle-Lys-Asn-Ala-His-
Lys-Lys-Gly-Gln: 31 residues

Proenkephalin Met-enkephalin Tyr-Gly-Gly-Phe-Met: 5 residues

Leu-enkephalin Tyr-Gly-Gly-Phe-Leu: 5 residues
Met-enkephalin Tyr-Gly-Gly-Phe-Met-Arg-Phe: 7 residues
Met-enkephalin Tyr-Gly-Gly-Phe-Met-Arg-Gly-Leu: 8 residues

Prodynorphin Dynorphin A Tyr-Gly-Gly-Phe-Leu-Arg-Arg-IIe-Arg-

Pro-Lys-Leu-Lys-Trp-Asp-Asn-Gln:
17 residues

Dynorphin 5 Tyr-Gly-Phe-Leu-Arg-Arg-lle:
8 residues

I-Neoendorphin Tyr-Gly-Gly-Phe-Leu-Arg-Lys-Tyr-Pro-
Lys: 10 residues

5-Neoendorphin Tyr-Gly-Gly-Phe-Leu-Arg-Lys-Tyr-Pro:
9 residues

Tyrosine is always at the amino-terminal end of the sequence. The classic sequence of the
first five amino acids (Tyr-Gly-Gly-Phe-Leu-) can be seen running through the entire group
of peptides. Further congruence can be seen within each of the three groups as we move
towards the carboxyl terminus.
Melanocyte stimulating hormones (MSH), both a. b and *f, are also derived from
pro-opiomelanocortin, and so is adrenocorticotrophic hormone (ACTH), u-MSH, indeed,
is residues 1 to 13 of ACTH.

Figure 9.10 Amino acid sequence of the opioid peptides

opioid activity coded for by the cDNA for the dynorphin precursor (Kakidani
et al., 1982). That these peptides are endogenous ligands for the opiate recep-
tor was shown early on since they closely mimic the action of morphine and
could be antagonized by the morphine antagonist naloxone (Frederickson,
1977; Fig. 9.10).
The endogenous ligands are not absolutely specific for the different recep-
tors, for example Pendorphin appears to bind equally well to p and K recep-
tors but has no affinity for 6. Leu-enkephalin, on the other hand, binds prefer-
entially to the &site but binds less well to the p-site. Dynorphin A favours the
K site, while still giving measurable binding at p. Morphine is also not entirely
204 Molecular mechanhns of drug action

selective in that it favours the /..Lsite but still has measurable affmity for both
K and 6. Naloxone, likewise, is more potent at the p site but can also bind at
the S and K sites. This lack of selectivity allows the opioid peptides to appear
similar to morphine in pharmacological terms, although they do not necessarily
favour the same receptor. Other agonists are being syntbesized that are more
specific (Kosterlitz, 1985).

9.9.1 Opioiddependence
Narcotic dependence is thought to relate in some way to the chronic inhibition
of adenylate cyclase activity, since levels of this enzyme are lowered in mem-
branes from morphmedependent rats and, as noted above, GTPase levels from
these animals are lowered. Naloxone increases these levels towards the normal
values (Barchfeld and Medzihradsky, 1984). It is interesting to note that metha-
done, although addictive itself, is often used to treat withdrawal symptoms of
heroin users gaffe and Martin, 1985).
A possible mechanism for the derivation of withdrawal symptoms is sug-
gested by the interaction of opioids with gonadotrophins. A long-acting ana-
logue of Met-enkephalin was shown to inhibit the release of follicle-stimulating
hormone and luteinizing hormone whilst naloxone produced a significant rise
in levels of gonadotrophins in both male and female subjects (Grossman et al.,
1981). The symptoms of withdrawal and naloxone treatment are very similar to
the symptoms of premenstrual syndrome (Reid and Yen, 1983) and opioid
agonists and antagonists have their greatest effect on serum luteinizing hor-
mone levels in the premenstruum. These symptoms may be a consequence of
high gonadotrophin levels, resulting partly from inadequate feedback inhi-
bition by ovarian steroids and also from continuous stimulation from the hype
thalamus by release of luteinizing-hormone-releasing hormone. The sudden
withdrawal of opiate narcotics, therefore, may sharply raise gonadotrophin
levels that can give rise to the withdrawal symptoms (Coulson, 1986).

9.9.2 Pentazocine as analgesic

In order to develop pain-killing drugs that do not cause dependence and pain-
ful withdrawal symptoms when their use is discontinued, considerable syn-
thetic effort has been expended. This eventually led to the development of
pentazocine among other drugs - structurally a derivative of benzomorphan.
When given orally this drug does not cause dependence, but when extracted
from the tablets and injected intravenously with the anti-histamine tripelenna-
mine, it may do so. Nevertheless, pentazocine is still one of the least addictive
morphine-like drugs (reviewed in Brogden et al., 1973).
Pentazocine shows a mixed type of action at opiate receptors, being a mod-
erately weak antagonist at p and a relatively powerful agonist at K and u. The
antagonism of p receptors helps to explain why the compound is less likely
to cause dependence, while the agonism at K receptors is probably responsible
 Neurotransmitter action and metabolism 205

for its analgesic effects. This type of analgesia differs from that induced by
morphine, since it only affects the spinal cord whereas morphine is able to act
at supra-spinal loci, presumably as a consequence of agonism at p receptors. At
high doses (60-90 mg in man), actions characteristic of stimulation of the (+
receptors occur, i.e. hallucinations and dysphoria.

HO

/CH,
N-CH,CH=C,
CH,
Pentazocine

9.9.3 Loperamideas anti-diaxrhoeal

One of the most marked side-effects of opiate agonist usage, constipation,
probably arises through action at the opiate receptor in the gastrointestinal
tract. This has been put to good use in the development of drugs for the
treatment of diarrhoea. Morphine has the ability to increase muscle tone, dim-
inish the amplitude of contractions and markedly reduce the propulsive
activity in the intestine. Consequently, over-the-counter preparations of mor-
phine adsorbed to kaolin are often effective in treating diarrhoea. Efforts have
been made, however, to fmd an opiate that does not suffer from the central
nervous system side-effects of morphine such as drowsiness etc.
Loperamide was developed as an answer to this problem and has proven
to be a safe and effective remedy for diarrhoea as well as being relatively free
of side-effects (Ericcson, 1990). Loperamide does not cross the blood-brain
barrier and thus does not give rise to sedation, although it will bind to opioid
receptors in brain homogenates. The drug slows gastrointestinal motility
(peristalsis) and is of particular value in cases where the bowel action is excess
ive, e.g. overactive bowel syndrome (Hughes et al., 1982).
Diarrhoea may, however, arise not only through excessive peristalsis but
also through reduced absorption or increased secretion of water and ions at
the mucosal surface. Loperamide antagonizes the secretion induced by prosta-
glandin El, theophylline and cholera toxin (Awouters et al., 1983) but appar-
ently not that induced by vasoactive intestinal polypeptide (Schiller et al.,
1984). Both cholera toxin and prostaglandin El induce the activation of adenyl-
ate cyclase, thus raising cyclic AMP levels, and the evidence suggests that
cyclic AMP is the mediator in the release of chloride ion, followed by water,
to maintain osmotic balance.
The mechanism of action of loper-amide is probably via the opioid receptor
since naloxone antagonizes its action. On the other hand, a direct inhibitory
effect of the opioid receptor on adenylate cyclase is unlikely because loperam-
206 Molecular mechanismsof drug action

Meperidine --CH, - COCH,CH, -H

bl

Loperamide
-OH -Cl

ide does not antagonize the rise in cyclic AMP induced by both prostaglandin
E, and cholera toxin. The suggestion has been made that calmoduhn may be
involved (Awouters et al., 1983) because:
(a) The secretory process is calcium-dependent and calmodulin mediates
many of the intracellular actions of calcium in secretion.
(b) Loperamide binds to other moderately high affinity sites in the gut wall,
where calmodulin is highly concentrated.
(c) Loperamide binds to calmodulin in the presence of calcium in vitro
with a binding constant of 1.2 X 10m5M (Zavecz et al., 1982).
This situation remains unresolved at the present time. What is clear, how-
ever, is that loperamide suppresses the secretion of chloride ion from the
mucosal cell into the gut lumen (Hughes et al., 1982).

9.10 Histamine receptors - mepyramine as anti-allergic,

cimetidine as anti-ulcer

The response of mammals to histamine has been studied extensively ever since
the work of Sir Henry Dale, one of the great pioneers of pharmacology. There
are three subtypes of histamine receptor: H,, H, and H,. The H, receptors
govern the contraction of smooth muscle in lung bronchi and gut, the increase
in capillary permeability and inflammatory response. H, receptors control the
action of histamine in inducing the secretion of gastric acid and positive chron-
otropic and inotropic effects in the heart. H, receptors are located, amongst
other places, presynaptically (autoreceptors) in the brain and govern the syn-
thesis and release of histamine and other neurotransmitters. Both H, and H,
play a part in the dilation of finer blood vessels and are also found in the brain
where histamine functions as a neurotransmitter (Garrison, 1990).
 Neurotransmitter action and metabolism 207

AU three receptors are linked to G-proteins which activate different cell-

signalling systems. Agonism at H, receptors activates the phospholipase C
pathway, while H, receptors, through G,, are linked to adenylate cyclase in a
positive fashion. H, may be linked negatively to adenylate cyclase (Arrang et
al., 1987) or possibly to an ion channel (Hill, 1990).
The H, and H, receptors have been cloned and expressed in various tissues.
Biogenic amines require an anionic ligand to bind the cationic amino group
which, in every case investigated so far, turns out to be an aspartate residue
in the third transmembrane segment. The histamine receptors are no excep-
tion. The f&h segment in the catecholamine receptors is important for cat-
echo1 binding. Histamine also requires an aspartate to bind the monocationic
nitrogen in the imidazole ring. For the H,, that appears to be sufficient to bind
the imidazole. The H2 receptor, however, recognizes both the nitrogen atoms -
the (neutral) second forms a hydrogen bond with the threonine residue only
present in TM5 of the H, receptor. Another aspartate appears necessary for
antagonist, but not agonist, binding in the H, receptor (Gantz et al,, 1992).
Both receptor subtypes are glycosylated with the carbohydrate residues being
sited on asparagine residues at the N-terminal segment. The large third intra-
cellular loop and the cytoplasmic C-terminus probably carries the G-protein
recognition site. The homology between the two receptor subtypes is, how-
ever, only 40.7 per cent which is less than between H, and the muscarinic
M, (44.3 per cent) (Gantz et al., 1992; Yamashita et al., 1991).

9.10.1 H, receptor antagonists

H, antagonists, such as mepyramine and chlorpheniramine, have been avail-
able for several decades and are used to block the action of histamine released
endogenously in allergic reactions, particularly in the upper respiratory tract,
as in hay fever. Skin allergies also respond favourably in some cases. Histamine
is one of the mediators (autacoids) of the allergic response generating itching
and some aspects of the inflammation that accompany this reaction. The lower
part of the respiratory tract is involved in the genesis of asthma which does
not respond to H, antagonists in man, since leukotrienes are the major cause
of allergic bronchoconstriction (see section 10.6 for a discussion on asthma).
In guinea-pigs, however, histamine is the major effector and H, antagonists
are protective (Garrison, 1990).

-N
\ / CH*
Cr AJCH~CH~N(CH~)~
CH30

Chlorpheniramine Mepyramine
208 Molecular mechanisms of drug action

Mepyramine has a low binding constant with H1 receptors as measured by

binding to homogenates of, for example, guinea-pig cortex (1.5 X 10e9~;
Daum et al., 1982) and rat cerebral cortex (2.1 X 10P9~; Hall and Ogren,
1984). Chlorpheniramine has a chiral centre: the (+)-isomer is more tightly
bound to the isolated bovine receptor by about lOO-fold more effectively than
the (-)-isomer (4.3 X lop9 M and 4.6 X lo-’ M, respectively; Yamashita et al.,
1991). (+)Chlorpheniramine was the most potent isomer, by three orders of
magnitude, in antagonizing contractions produced by histamine in the hamster
smooth muscle from the vas deferens (White et al., 1993). The drug was both
effective and bound at similar concentrations. The H, receptor activates turn-
over of phosphatidylinositol, as measured in rabbit aorta, and the effect is
inhibited most by H, receptor antagonists such as mepyramine. Histamine
stimulates smooth muscle contraction in this organ. The phosphatidylinositol
pathway (section 9.1) activates calcium uptake into the cell so that muscle
contraction may be stimulated (Villalobos-Molina and Garcia-Sainz, 1983; White
et al., 1993).
The antagonism of histamine is competitive and reversible. The HE antagon-
ists are generally lipophilic and demonstrate variations on the following struc-
ture:

Ah, /
Ar2/X-CH-CH-N\
where Ar is usually an aryl group and X is a nitrogen or carbon atom of a C-
O-ether group (Garrison, 1990).

9.10.2 Ha receptor antagonists

The rational development of the drug cimetidine, including the importance of
physicochemical factors such as partition coefhcient and ionization constant,
makes an interesting story (Ganellin, 1978). The structure of histamine was
taken as the starting point, and it was shown that 2-methylhistamine was a
more selective agonist for the HI receptor, while 4methylhistamine favoured
the H, receptor. After separation of agonist and antagonist activities, burimam-
ide was derived as the first true H, receptor blocker. Optimization of the H2
receptor blocking activity resulted in the synthesis of metiamide. Various struc-
tural changes to eliminate unwanted side-effects fmally led to the development
of cimetidine.
Histamine activates adenylate cyclase in brain and this has been shown to

CH2SCH2CH2N’CNHCH3
AHC=N

Cimetidine
 Neurotransmitter action and metabolism 209

be mediated by Hz receptors. Cimetidine antagonizes this activation in homo-

genates of the dorsal hippocampus of guinea-pig brain with a Ki of
S-9 X ~O-‘M (Kanof and Greengard, 1979). This compares well with a figure
of 7.9 X IO-’ M against histamine stimulation of the rate of contraction of the
right atrium of the guinea-pig in vitro, i.e. the positive chronotropic effect
(Ganellin, 1978). The values for other H, antagonists, burimamide and metiam-
ide, also agree, thus confirming the connection between adenylate cyclase
activation and binding to the H2 receptor. It should be noted that cimetidine
is a compound that is much more hydrophilic than an H, receptor antagonist
and thus does not cross the blood-brain barrier to react with brain receptors,
and so the correlation obtained above is all the more convincing.

CH2CH2CH2CH2NHCNHCH~ CH2SCH2CH2NHCNHCH3

HN/-\
dN
i
CH3)I=(
HN
I//”
:

Burimamide Metiamide

Tiotidine binds to H, receptors in homogenates of guinea-pig cerebral cortex

with a binding constant of 1.7 X lo-*M. That the H, receptor is involved in
the binding is suggested by the correlation of binding with the antagonism of
the chronotropic effect in guinea-pig heart. In addition, antagonism of hista-
mine-induced adenylate cyclase activity in guinea-pig gastric mucosa also corre-
lates very well with receptor binding for a number of antagonists including
tiotidine (Gajtkowski et al., 1983).
Cimetidine was put on the market in the mid-1970s for the treatment of
gastric acid secretion in peptic ulcer patients, and effected a revolutionary
improvement in the chemotherapy of ulcers. Most patients suffering from
ulcers hypersecrete gastric acid, possibly as a result of stress or for some other
reason. Gastric acid secretion is stimulated by food intake and by a variety of
agents, known as secretagogues, e.g. histamine or caffeine, and by stimulation
of the vagus nerve. Cimetidine will also block secretion induced by muscarinic
agonists and by gastrin non-competitively. This is because gastrin activates the
release of histamine from parietal cells (Sachs and Wallmark, 1989). Cimetidine
thereby inhibits both basal and stimulated secretion. The drug both relieves
the pain and allows the ulcer to heal.
Relapses are, however, quite common after treatment has stopped. In some
cases the ulcer may not have healed completely, although the patient may
have no symptoms. It has been remarked that the problem is not how to heal
an ulcer but how to keep it healed (Thomas and Misiewicz, 1984). If the ulcer
is healed initially but the predisposing factor is not addressed, then it is not
surprising that the ulcer may recur. Cimetidine may be used prophylactically
as a maintenance therapy to prevent relapse. Pepsin secretion is also reduced
by the drug.
After further synthetic optimization of H, receptor blocking activity, another
more potent receptor blocker, ranitidine, joined cimetidine on the market.
 210 Molecular mechanisms of drug action

CHN02 CH&.X2CH2NHCNHCH3

CH2SCH&H2NH:NHCHI l&=N
-

\o
~-NH*
c
CHMCH& Ranitidine NH’ Tiotidine

Structurally the need for a small heterocyclic ring has been met by a furan,
thus indicating that an imidazole, reminiscent of histamine, is not essential
(moreover, tiotidine has a thiazole ring). Ranitidine is now regarded as the
frontline therapy for suppression of gastric acid secretion (Grant et al., 1989).

Questions

1. Name three receptors that activate (a) adenylate cyclase or (b) phospholi-
pase C.
2. What special structural features can you identify for G-protein-coupled
receptors?
3. Name two receptors that are linked to ion channels. How do they differ
structurally from G-protein coupled receptors?
4. In the Padrenergic receptors, which amino acid residues are believed to
bind the agonists and where in the receptor are they sited? How do cat-
echolamine and serotonin receptors differ in the transmembrane domain?
5. What protein is used as the model template for assigning structure to the
7-transmembrane receptors? How valid is this extrapolation?
6. What role do guanine nucleotides play in signal transduction?
7. What is meant by the terms (a) affinity (b) efficacy (c) partial agonists?
8. Which receptors mediate the action of the benzodiazepines? How does this
interaction take place?
9. How does agonist binding to the G-protein-coupled receptors differ from
that of antagonists?

References

Aapro, MS., 1991, Drugs, 42, 551-68.

Ahlquist, R.P., 1948, Am. J Pbysiol., 153, 586-600.
Alexander, B.S. and Wood, M.D., 1987, J. Pharm. Pharmacol., 39, 664-6.
Allan, A.M. and Harris, R.A., 1986, Molec. Pbarmacol., 29, 497-506.
AIlerton, C.A., Boden, P.R. and Hill, R.G., 1989, Brit. J. Pbarmacol., 96, 29-38.
Anden, N.E. and Stock, G., 1973, J. Pbarm. Pbarmacol., 25, 346-8.
Anon, 1991, Trends in Pbarmacol. Sci., 12, 18.
Apperley, G.H., Daly, M.J. and Levy, G.P., 1976, Brit. J. Pharmacol., 57, 235-46.
Arrang, J.-W., Garbarg, M., Lancelot, J.-C., Lecomte, J.-M., Pollard, H., Robba, M., Schun-
ack, W. and Schwartz, J.-C., 1987, Nature, 327, 117-23.
 Neurotransmitter action and metabolism 211

Awouters, F., Niemegeers, C.J.E. and Janssen, P.A.J., 1983, Annu. Rev. Pbarmacol. Tox-
icol., 23, 279-301.
Baldessarini, R.J., 1979, Postgrad. Med., 65, 123-8.
Barchfeld, C.C. and Medzihradsky, F., 1984, Biocbem. Biophys. Res. Commun., 121,
641-8.
Barnes, J.M., Barnes, N.M., Costall, B. and Naylor, RJ., 1991, Pbarmaceut. J., 246,
112-14.
Barnes, J.M., Barnes, N.M. and Cooper, S.J., 1992, Neuroscience and Biobebavioural
Reviews, 16, 107-13.
Bass, C. and Kerwin, R., 1989, Brit. Med. J., 298, 345-6.
Bet-ridge, M.J., 1993, Nature, 361, 315-25.
Bianchine, J., 1985, In: The Pharmacological Basis of merapeutics, 7th Ed., A.G. Gil-
man, L.S. Goodman, T.W. Rail and F. Murad, Eds (New York: Macmillan) Ch. 21.
Bimbaumer, L., Abramowtiz, J. and Brown, A.M., 1990, Biocbem. Biopbys. Acta, 1031,
163-224.
Biier, P., de Montigny, C. and Chaput, Y., 1990,J Clin. Psycbiat., 51, 4, Suppl, 14-21.
Borison, R.L., Hitri, A., Blowers, AJ. and Diamond, B.I., 1983, Clin. Neuropbarmacol.,
6, 137-50.
Bowen, W.D., Moses, E.L., Tolentino, P.J. and Walker, J.M., 1990, Eur. J. Pbarmacol.,
177, 111-18.
Bowery, N., 1989, Trends in Pbarmacol. Sci., 10, 401-7.
Braestrup, C., Albrechtsen, R. and Squires, R.F., 1977, Nature, 269, 702-704.
Braestrup, C., Nielsen, M., Krogsgaard-Larsen, P. and Falch, E., 1980, Nature, 280,
331-3.
Brittain, R.T., Drew, G.M. and Levy, G.P., 1982, Brit.J. Pbarmacol., 77, 105-14.
Brogden, R.N. and Sorkin, E.M., 1990, Drugs, 40, 903-49.
Brogden, R.N. and Faulds, D., 1991, Drugs, 42, 895-912.
Brogden, R.N., Speight, T.M. and Avery, G.S., 1973, Drugs, 5, 6-91.
Brown, J.R. and Arbuthnott, G.W., 1983, Neuroscience, 10, 349-55.
Buckner, C.B. and Abel, P., 1974,J. Pbarmacol. Exp. Tier., 189, 616-25.
Bunney, B.S., Walters, J.R., Roth, R.H. and Aghajanian, G.K., 1973, J. Pbarmacol. Exp.
Tber., 185, 560-71.
Butchers, P.R., Vardey, C.J. and Johnson, M., 1991, Brit. J Pbarmacol., lO4, 672-6.
Bylund, D., 1992, FASEBJ, 6, 832-9.
Calne D.B., 1984, New Eng1.J Med., 310, 523-4.
Campbell, W.C., Fisher, M.H., Stepley, E.D., Albers-Schonbetg, G. and Jacob, T.A., 1983,
Science, 221, 823-8.
Carlsson, A., 1984, Adv. Biocbem. Psycbopbarmacol., 39, 213-21.
Carmine, A.A. and Brogden, R.N., 1985, Drugs, 30, 85- 126.
Caulfield, M.P. and Straughan, D.W., 1983, Trends in Neurosci., 6, 73-5.
Childers, S.R., 1991, Life Sci., 48, 1991-2003.
Chretien, M., Benjannet, S., Dragon, N., Seidah, N.G. and Lis, M., 1976, Biocbem.
Biopbys. Res. Commun., 72, 472-8.
Christmas, A.J., Coulson, C.J., Maxwell, D.R. and Riddell, D., 1972, Brit. J. Pbarmacol.,
45, 490-503.
Coleman, R.A., Johnson, M., Nials, A.T. and Stunner, MJ., 1990, Brit. J. Pbarmacol.,
99, ~Suppl), 121P.
Coppen, A., 1967, Brit. J. Psycbiat., 113, 1237-64.
Coulson, CJ., 1986, Med. Hypotbes., 19, 243-56.
Creese, I., Schneider, R. and Snyder, S.H., 1977, Eur. J. Pbarmacol., 46, 377-81.
Cullum, V.A., Farmer, J.B., Jack, D. and Levy, G.P., 1969, Brit. J Pbarmacol., 35,
141-51.
Daum, P.R., Hill, S.J. and Young, J.M., 1982, Brit. J. Pbarmacol., 77, 347-57.
212 Molecular mechanisms of drug action

DougaIl, I.G., Harper, D., Jackson, D.M. and I&f, P., 1991, Brit. J Pbarmacol., 104,
1057-61.
Duce, I.R. and Scott, R.H., 1985, Brit. J. Pbarmacol., 85, 395-401.
Egerton, J.R., Osthnd, D.A., Blair, L.S., F&y, C.H., Suhayda, D., CifeIIi, S., Riek, R.F. and
Campbell, W.F., 1979, Antimicrob. Ag. Cbemotber., 15, 372-8.
EIlenbroek, B.A., Artz, M.T. and Cools, A.R., 1991, Eur.J. Pbarmacol., 196, 103-108.
Ericsson, C.D., 1990, Amer. J. Med., 88<SbA), 105-45.
Evans, S.M., GaIdes, A. and Gall, M., 1991, Pharmacology, Biochemistry and Bebav-
iour, 40, 1033-40.
Fisher, M.H. and Mrozik, H., 1992, Annu. Rev. Pbarmacol. Toxicol., 32, 537-53.
Fitton, A. and Heel, R.C., 1990, Drugs 40, 722-47.
Fitton, A., Faulds, D. and Goa, K.L., 1992, DtrLgs, 43, 561-96.
Fowler, C.J., Mantle, TJ. and Tipton, K.F., 1982, Biocbem. Pbarmacol., 31, 3555-61.
Frederickson, RCA., 1977, Life Sci., 21, 23-42.
Gajtkowski, G.A., Norris, D.B., Rising, T.J. and Wood, T.P., 1983, Nature, 304, 65-7.
Ganellin, G.R., 1978, J. Appl. Cbem. Bbtecbnol., 28, 182-200.
Gantz, I., DeIVaIIe, J., Wang, L.-D., TashIro, T., Munzert, G., Guo, Y.-G., Konda, Y. and
Yamada, T., 1992,J. Biol. Cbem., 267, 20840-3.
Garrison, J.C., 1990, Ch. 23 in The Pharmacological Basis of Therapeutics, 8th Ed.
Eds, A.G. Giiman, T.W. RaII, A.S. Nies and P. Taylor. New York: Pergamon.
Gerlach, M., Riederer, P. and Youdim, M.B., 1992, Eur. J. Pbarmacol., 226, 97-108.
Gilman, A.G., 1984, Cell, 36, 577-9.
Goa, K.L., McTavish, D. and Clissold, S.P., 1991, Drugs, 43, 640-58.
Goldstein, A., Fischli, W., Lowney, L.I., HunkapilIa, M. and Hood, L., 1977, Proc. Nut.
Acad. Sci. (US.A.), 78, 7219-23.
Goodwin, F.K. and Post, R.M., 1983, Brit. J Clin. Pbarmacol., 15, 393S-405s.
GoyaI, R.K., 1988, Life Sci., 43, 2209-20.
Grant, S.M., Langtry, H.D. and Brogden, R.N., 1989, Drugs, 37, 801-70.
Grossman, A., Moult, P.J.A., Gaillard, R.C., DelitaIia, G., Toff, W.D., Rees, L.H. and
Besser, G.M., 1981, Clin. Endocrinol., 14, 41-7.
Haefely, W., Burkard, W.P., Cesura, A.M., KettIer, R., Lorez, H.P., Martin, J.R., Richards,
J.G., Scherschhcht, R. and de Prada, M., 1992, Psycbopbarmacol., 106, S6-14.
Hall, H. and Ogren, S.-O., 1984, Life Sci., 34, 597-605.
Han&, A., Oksenberg, D., Fischette, C. and Peroutka, S.J., 1990, Biol. Psycbiat., 28,
99-109.
Harrington, M.A., Zhong, P., Garlow, S.J. and Ciaranello, R.D., 1992, J Clin. Psycbiat.,
53,10, Suppi, 8-27.
Havard, C.W.H. and Fonseca, V., 1990, Dncgs, 39, 66-73.
Hegde, S.S., Ricci, A., Amenta, F. and Lokhandwala, F., 1989,J. Pbarmacol. Exp. Tber.,
251, 1237-45.
Hesketh, P.J. and Gandara, D.R., 1991,J. Natl. Cancer Inst., 83, 613-20.
Hi.& S.J., 1990, Pbarmacol. Rev., 42, 45-83.
Holcdaw, T.L. and Beck, T.R., 1990, Am. J. Hypertens., 3, 12OS-5s.
Hughes, J., Smith, T.W., Kosterlitz, H.W., FothergilI, L.A., Morgan, B.A. and Morris, H.R.,
1975, Nature, 258, 577-9.
Hughes, S., Higgs, N.B. and Turnberg, L.A., 1982, Gut, 23, 974-9.
Hulme, E.C., BirdsaIl, N. J.M. and Buckley, NJ., 1990, Annu. Rev. Pbarmacol. Toxicol.,
30, 633-73.
Humphrey, P.P.A. and Feniuk, W., 1991, Trends in Pbarmacol. Sci., 12, 444-6.
Jack, D., 1991, Brit.J Clin. Pbarmacol., 31, 501-14.
Jaffe, J.H. and Martin, W.R., 1985, in The Pharmacological Basis of Fberapeutics, 7th
Ed., A.G. Gihnan, L.S. Goodman, T.W. RaR and F. Murad, Eds. (New York:
Macmillan) Ch. 22.
Johnston, J.P., 1968, Biocbem. Pbarmacol., 17, 1285-97.
 Neurotransmitter action and metabolism 213
JuIins, D., Huang, K.N., Livelli,TJ., Axel, R. and Jesse1,T.M.. 1990, Proc. Nut. Acud. Scz:,
U.S.A., 87, 928-32.
Kakidani, H., Fumitani, Y., Takahashi, H., Noda, M., Morimoto, Y., Hirose, T., Asai, M.,
Inayama, S., Nakanishi, S. and Numa, S., 1982, Nature, 298, 245-9.
Kanof, P.D. and Greengard, P., 1979,J. Pbarmacol. Exp. Therapeut., 209, 87-96.
Kass, IS., Stretton, A.O.W. and Wang, C.C., 1984, Molec. Btocbem. Parasitol., 13,
213-25.
Kendall, 1992, Biocbem. Sot. Trans., 20, 109-13.
Kendall, D.A. and Nahorski, S.R., 1985, J. Pbarmacol. Exp. T&her., 233, 473-9.
Knoll, J. and Magyar, K., 1972, Adv. Biocbem. Psycbopbarmacol., 5, 393-408.
Kobilka, B., 1992, Annu. Rev. Neurosci., 15, 87-114.
Kosterlitz, H.W., 1985, Proc. Roy. Sot. Land., Series B, 225, 27-40.
Lambert, D.G., But-ford, N.T. and Nahorski, S.R., 1992, Biocbem. Sot. Trans., 20,130-5.
Lance, J.W., Lambert, G.A., Goadsby, PJ. and Zagami, A.S., 1989, Cepbalalgia, 9 (S9),
7-13.
Lefkovitz, R.J., 1975, Biocbem. Pbarmacol., 24, 583-90.
Leonard, B.E., 1992, Drzsgs, 43, Suppl 2, 3-10.
LeWitt, P.A., 1991, Acta Neural. Stand. Suppl., 136, 79-86.
Leysen, J.E., de Chaffoy de Courcelles, D., De Clerck, F., Niemegeers, C.J.E. and Van
Neuten, J.M., 1984, Neuropbarmacology 23, 1493-1501.
Li, C.., Chung, D. and Doneen, B.A., 1976, Biocbem. Biopbys. Res. Commun., 72,
1542-7.
Lieberman, A.N. and Goldstein, M., 1985, Pbarmacol. Rev., 37, 217-27.
Limbiid, L.E., Speck, J.L. and Smith, S.K., 1982, Molec. Pbarmacol., 2 1, 609- 17.
Lipinski, J.F., Cohen, B.M., Zubenko, G.S. and Watemaux, C.M., 1987, Life Sci., 40,
1947-63.
Low, M.G., Ferguson, M.A.J., Futerman, A.H. and Silman, I., 1986, Trends in Biocbem.
Sci., 11, 212-15.
Lucki, I., 1991,J Clin. Psycbiat., 52, S12, 24-31.
McPherson, G.A. and Summers, R.J., 1982, Brit. J. Pbarmacol., 77, 177-84.
Manschrek, T.C., 1981, New Eng1.J Med., 305, 1628-31.
Maricq, A.N., Peterson, A.S., Brake, AJ., Myers, R.M. and Julius, D., 1991, Science, 254,
432-7.
Martin, R.J., 1987, Biocbem. Sot. Trans., 15, 61-5.
Matsui, H., Imafuki, J., Asakura, M., Tsukamoto, T., Ino, M., Saitoh, N., Miyamura, S.
and Hasegawa, K., 1984, Biocbem. Pbarmacol., 33, 3311-14.
Maycock, A.L., Abeles, R.H., Satach, H.I. and Singer, T.P., 1976, Biochemistry, 15,
114-25.
Meltzer, H.Y., 1992, Brit. J. PJych.., Suppl, 22-9.
Minneman, K.P., Hegstrand, L.R. and Molinoff, P.B., 1979, Molec. Pbarmacol., 15,
21-33.
Napoliello, M.J. and Domantay, A.G., 1991, Brit. J. Psycbiat., 159, S12, 40-44.
Pearce, J.M.S., 1984, Brit. Med.J, 288, 1777-8.
Pickar, D., Owen, R.R., Litman, R.E., Konicki, P.E., Gutierrez, R. and Rapaport, M.H.,
1992, Arch. Gen. Psych.., 49, 345-53.
Pinder, R.M., Brogden, R.N., Speight, T.M. et al., 1977, Dncgs, 1, 321-52.
Pong, S.S. and Wang, C.C., 1982, J. Neurocbem., 38, 375-9.
Powell, J.F., 1991, Biocbem. Sot. Trans., 19, 199-201.
Prichard, B.N.C., 1992,J. Cardiovasc. Pbarmacol., 19, Suppl 1, Sl-S4.
Rasmussen, H., 1986, New Eng1.J Med., 314, 1094-101.
Reid, R.L. and Yen, S.S.C., 1983, Clin. Obstet. Gynaecol., 26, 710-18.
Rodbell, P., 1980, Nature, 284, 17-22.
Rosenberg, T.L., 1975, Adv. Enzymol., 43, 103-218.
Rudd, P. and Blaschke, T.F., 1985, In: The Pharmacological Basis of Therapeutics, 7th
214 Molecular mechanisms of drug action

Ed., Eds A.G. Gilman, L.S. Goodman, T.W. Rall and F. Murad (New York:
Macmillan) Ch. 32.
Sachs, G. and WaIlmark, B., 1989, Stand. J. Gastroenterol., 24, Suppl 166, 3-11.
Sanger, D.J., 1985, Life Sci., 36, 1503-13.
Saverese, T.M. and Fraser, C.M., 1992, Biochem. J., 283, 1-19.
Saxena, P.R. and Ferrari, M., 1989, Trends in Pharmacol. Sci., 10, 200-4.
Scadding, G.K. and Havard, C.W.H., 1981, Brit. Med.J, 283, 100812.
SchiIIer, L.R., Santa Ana, C.A., Morawski, S.G. and Fordtran, J.S., 1984, Gastroenterol.,
86, 1475-80.
Schmidt, MJ., Fuller, R.W. and Wong, D.T., 1988, Brit. J. Psycbiat., 153, Suppl. 3,
40-6.
Schoeffter, P. and Hoyer, D., 1989, Naunyn-Scbmiedeberg’s Arch Pbarmacol., 340,
135-8.
Schoffelmeer, A.N.M., Van VIiet, B.J., De Vries, T.J., Heijna, M.H. and Mulder, A.H.,
1992, Biocbem. Sot. Trans., 20, 449-53.
Shapiro, G. and Em, A., 1992, Drugs of the Future, 17, 489-501.
Shih, J.C., 1991, Neuropsycbopbarmacol., 4, l-7.
Shih, J.C., Yang, W., Chen, K. and Gallaher, T., 1991, Pbarmacol. Biocbem. and Bebav-
iour, 40, 1053-8.
Sibley, D.R. and Monsma, F.J., 1992, Trench in Pbarmacol. Sci., 13, 61-9.
Sieghart, W., 1989, Trenak in Pbarmacol. Sci., 10, 407- 11.
Silverman, R.B., 1983,J. Biol. Cbem., 258, 14766-9.
Silverman, R.B., 1991, Biocbem. Sot. Trans., 17, 201-6.
Snyder, S.H., 1981, Am.J Psycbiat., 138, 460-3.
Snyder, S.H., 1986, Nature, 323, 292-3.
Soreq, H., Gnatt, A., Lowenstein, Y. and Neville, L.F., 1992, Trenak in Biocbem. Sci.,
17, 353-8.
Speight, T.M. and Avery, G.S., 1972, Drugs, 3, 159-203.
Speth, RI., Wastek, G.J., Johnson, P.C. and Yamamura, HI., 1978, Life Sci., 22,859-66.
Sternweis, P.C. and Smrcka, A.V., 1992, Trends in Biocbem. Sci., 17, 502-6.
Taussig, R., Quarmby, L.M. and Gilman, A.G., 1993, J Biol. Cbem., 268, 9- 12.
Tetrud, J.W. and Langston, J.W., 1989, Science, 249, 519-20.
Thomas, J.M. and Misiewicz, G., 1984, Clin. Gastroenterol., 13, 501-41.
Vauholden, R., Lameire, N. and Ringoir, S., 1985, Eur.J. Clin. Pbarmacol., 28, 125-30.
Vicini, S., 1991, Neuropsycbopbarmacol., 4, 9-15.
Villalobos-Molina, R. and Garcia-Sainz, J.A., 1983, Eur. J. Pbarmacol., 90, 457-9.
Vincent, A., 1991, Biocbem. Sot. Trans., 19, 180-3.
Wakade, A.R., Malhotra, R.K. and Wakade, T.D., 1986, Nature, 321, 698-700.
Weinstein, D., Schenker, J.G., Gloger, I., Slonim, J.H., DeGroot, N., Hochberg, A.A. and
Folman, R., 1981, Fed. Eur. Biocbem. Sot. Lett., 126, 29-32.
White, T.E., Dickenson, J.M. and HiII, SJ., 1993, Brit. J Pbarmacol., 108, 196-203.
Wieland, H.A., Luddens, H. and Seeburg, P.H., 1992, J. Biol. Cbem., 267, 1426-30.
Williams, M. and Yarbrough, G.G., 1979, Eur. J. Pbarmacol., 56, 273-6.
Wilson, I.B., Harrison, M.A. and Ginsburg, S., 1961,J. Biol. Cbem., 236, 1498500.
Wojcik, W.J. and Holopainen, I., 1992, Neuropsycbopbarmacol., 6, 201-14.
Wright, D.J., 1987, Biocbem. Sot. Trans., 15, 65-7.
Yamashita, M., Fukui, H., Sugama, K., Horio, Y., Ito, S., Mizuguchi, H. and Wada, H.,
1991, Proc. Nut. Acad. Sci., (U.S.A.), 88, 11515-19.
Zavecz, J.H., Jackson, T.E., Limp, G.L. and Yellin, T.O., 1982, Eur. J. Pbarmacol., 78,
375-7.
 Chapter 10
Membrane-active agents

IO. 1 Introduction
This chapter considers those examples of drugs which act directly on the cell
membrane, whether mammalian or fungal. The processes concerned govern
the transport of ions across the cell membrane in either direction, and are
either channels through which ions pass, or require the hydrolysis of ATP to
drive various ions across the membrane. Alternatively, the drug may disrupt
the membrane of a pathogenic organism in such a way that it becomes highly
permeable to small molecules and consequently loses its effectiveness as a
semi-permeable membrane. These effects may all be regarded as direct;
indirect effects, whereby a drug binds to a receptor which subsequently modu-
lates membrane permeability, are discussed in Chapter 9.
In order to understand more about how drugs can interact with membranes,
it is important to know how present theories of membrane structure relate to
transport of ligands and ions across membranes, whether carried out by
enzyme or by ion channel. Furthermore, some understanding of the action
potential (potential difference) that exists across the membrane is also
important for an understanding of the mechanism of action of local anaes-
thetics and anti-arrhythmic agents that act via blockade of sodium channels,
and those anti-hypertensive drugs that interfere with calcium ion channels.

10.1.1 Membrane structure

Our present view of membrane structure in eukaryotic cells has been outlined
by Singer and Nicolson (1972) and is depicted in Fig. 10.1. The major matrix
of the membrane comprises a double leaflet of phospholipid molecules. The
polar headgroups of the two layers face out towards the aqueous environment
of the cytoplasm and the extracellular medium respectively, while the non-
polar fatty acyl hydrocarbon chains (two per molecule) are orientated towards
the middle of the membrane. Cholesterol in mammalian cell membranes and
ergosterol in fungal cell membranes are inserted into the lipid bilayer between
phospholipid molecules. The 3-hydroxyl group common to both of the sterols
is orientated towards the aqueous environments and interacts with the polar
headgroups of the phospholipids, while the non-polar sterol skeleton and

215
216 Molecular mechanisms of drug action

Table 10.1 Drugs and their targets discussed in Chapter 10

Section Target Drug l%erapeutk Use

Lidocaine
Lidocaine Local anaesthetic
Amiloride, Anti-arrhythmic
10.2 Sodium channel Triamterene Diuretic

Verapamil
10.3 Calcium channel Nifedipine Angina

Coupled sodium/ Ethacrynic acid

10.4 chloride channels Frusemide Diuretics

Lemakalin
10.5 Potassium channels Nicorandil Coronary vasodilator

10.6 Na/K ATPase Digoxin Cardiac failure

H/K ATPase Omeprazole Anti-ulcer

10.7 Membrane stabilizer Cromoglycate Anti-allergic

10.8 Calcineurin cyc10sp0rin Immune suppressant

0
10.9 Ergosterol Amphotericin Antifungal

i: Inside

In the lipid bilayer, phospholipid molecules are shown with their polar head groups as
small circles, and the non-polar tails as wavy lines. The leaflet is held together by the
non-polar interaction between the fatty acyl side-chains. A is an integral protein bound
in the extracellular surface of the membrane, while 6 is a transmembrane protein. C is an
integral protein facing the cytoplasm. Both A and C penetrate only one-half of the bilayer.
Figure 10.1 Membrane structure

hydrocarbon tail are positioned so that they interact with the hydrocarbon
chains of the phospholipid fatty acyl groups.
Proteins are also present in the membrane; the receptors for some hormones
span the membrane, unlike adenylate cyclase which is positioned on the
inside. Some transmembrane proteins, e.g. Na+,K+-ATPase, are the transport
carriers for ions, while there are other protein carriers for small molecules
such as glucose. These proteins have been shown to operate on both sides of
 Membrane-active agents 217

the membrane. Thus for Na+,K+-ATPase, the sodium and ATP binding sites
are on the cytoplasm& side of the membrane while the potassium and ouabain
binding sites are situated on the extracellular side. As noted in Chapter 9, the
GABA receptor contains a channel for chloride ions.
Membrane fluidity has a marked effect on the transport of small molecules
across the membranes - the more fluid the membrane the more permeable it
is. Fluidity is controlled to some extent by the type of fatty acid present in
the phospholipids: saturated fatty acids decrease and unsaturated fatty acids
increase the fluidity. Cholesterol also has the effect of reducing fluidity in areas
that are rich in unsaturated fatty acids and vice versa in areas high in saturated
fatty acids. The net effect is to produce regions of markedly differing per-
meability within the same membrane.
With regard to the mechanism of transport, Scarborough (1985) has sug-
gested that when membrane-bound proteins bind ligands, they undergo a con-
formational change rather like the operation of a hinge so that the ligand is
embedded within the protein. Scarborough views the open or ligand-free state
as having a cleft (state 1 in Fig. 10.2) within which the ligand binds (state 2).
This binding induces a conformational change which opens an aqueous dif-
fusion pathway from the binding site towards the far side of the membrane
(state 3). The ligand is then free to escape to the opposite side (state 4). The
conformation then changes back to the original to allow the process to
begin again.

10.1.2 Dynamics ofthe heartbeat

From a biochemical point of view the above description might be considered
sufhcient for an understanding of membrane events. The cell membrane has,
however, been shown to support a considerable potential difference; 70-
90 mV in cardiac cells and 200 mV in cells of the filamentous fungus Neuros-

P, the transmembrane protein that acts as a carrier.

M, the ligand to be transported; in this case a cation, and the carboxylate ion the binding
site - probably the aspartate side-chain.
1-4, the different states of the transport system.

Figure 10.2 Membrane transport

 218 Molecular mechanisms of drug action

pora crassa (the inside of the membrane is negative with respect to the
outside). The consequences that flow from changes in the electrophysiological
state of the membrane are crucial for an understanding of the mechanism of
action of three of the types of drug that are discussed in this chapter which
modulate the transport of cations: local anaesthetics and type 1 anti-arrhythmic
agents, both of which block the sodium channel, and calcium channel blockers
that are also used for arrhythmias and for high blood pressure. In order to
understand more about how these drugs exert their effects, an outline is pre-
sented here of the generation of the action potential in a pacemaker cell that
controls the beating of the heart, specifically in the Purkinje fibres (Kumana
and Hamer, 1979, described in Muhiddin and Turner, 1985; Fig. 10.3).
The sodium channel initiates the process in phase 0 by allowing a rapid
transport of sodium ion into the cell to occur, thus causing the membrane
potential to depolarize from a potential of -70 mV to approximately f25 mV.
These channels are then shut off or inactivated and cannot open again until
the membrane has been repolarized. Next in phase 1 there is a sharp fall of
about 20 mV probably due to chloride ion entry. In phase 2, calcium begins
to enter through the slow calcium channels (called slow because the time
constant for inactivation is 50 ms compared with O-5 ms for the ‘fast’ sodium
channel). Eventually chloride and calcium channels close and the major repola-
rization of the membrane is effected by a large outflow of potassium in phase
3. The potassium channels are inactivated when the membrane potential has
reached about -8O- -90 mV, the resting potential.
A second outward potassium channel, known as the pacemaker current,
continues to operate for a short time. Eventually, the point is reached at which
the inward sodium and calcium background current, which do not change
with time, produce a net inflow of current. The overall effect in phase 4,

-I- 50
1 1
& enters cell
0 4
500 lrns
mV

- 50

-100 J K leaks

The phases of the normal action potential are shown; 0, depolarization, 1,2,3, repolariz-
ation and the diastolic phase (4).
Figure 10.3 Diagram of myocardial action potential
 Membrane-active agents 219

therefore, is that of gradual depolarization as the background repolarizing cur-

rent gradually falls away. When the membrane potential has dropped to
-70 mV, the reactivated sodium channel is triggered and the process repeats
itself. It usually takes about 400 ms from start to finish in a pacemaker cell,
which acts, as its name suggests, to initiate the process by spontaneously gen-
erating an action potential - a property known as automatic&y. A group of
pacemaker cells act together to initiate an impulse which spreads throughout
the rest of the heart. Although the details of ionic involvement may differ in
other cells in the heart, the requirement for fast sodium and slow calcium
channels is constant.
Arrhythmias, or abnormal beating of the heart, can be caused either by faults
in the mechanism for initiating the electrical impulse or by disturbances in the
conduction of an impulse. Abnormal automatic&y occurs when the membrane
depolarizes too soon before phases 3 and 4 have folly run their course, and
an extra impulse is generated that is propagated throughout the heart. The
result can be almost random beating of the heart. This is often referred to as
afterdepolarization and, if it occurs during phase 3 (early afterdepolarization),
is due to the increased influx of a cation, either sodium or calcium, or reduced
efflux of potassium. The membrane potential has not fallen low enough to
activate the fast sodium channels. Afterdepolarizations may also occur that are
delayed until full repolarization has been reached (Bigger and Hoffman, 1990).
Alternatively, other cells that are not normally pacemaker cells may take it
upon themselves to initiate an impulse - a form of automatic&y that arises in
the wrong place. This may be a consequence of disease affecting the respon-
siveness in that the sodium channel is oversensitive to stimulation. In arrhyth-
mia caused by either situation, a drug that can partially block the initial sodium
current is of value in steadying the heart beat and returning the process to
normal. A partial blockade of the calcium channel is also found to be of value
in some cases, particularly after depolarization.
The conduction of an impulse can also lead to arrhythmias if it reenters a
part of the heart that has already been excited. This normally requires that
the rate of conduction be slowed, usually as a consequence of disease, and
that a block in conduction occurs which can divert the impulse back into a
region of the heart that it has already traversed (Wit, 1985).

10.2 The sodium channel

The mechanism of transmission of nerve impulses, and the propagation of

electrical impulses in the heart that accompany beating, both require a sharp
and rapid change in the trans-membrane potential difference. This is achieved
by means of a swift transport of sodium cations into the cell through what is
known as the sodium channel. Estimates of 10’ ions per second or an electrical
conductance of 2-8 pS (the unit of conductance is Siemens, S) per site have
been made (reviewed in Catterall, 1980). This rate is too fast to be associated
220 Molecular mechanisms of drug action

with a mobile carrier and the cation must pass independently through the
selective pore. It is not, however, open at all times and requires repolarization
of the membrane to take place before it is reactivated. After a rapid transient
influx of cations has taken place, the channel is closed by a mechanism that
is not yet understood.
At least two types of drug interact with the electrically excitable sodium
channel to block it; local anaesthetics, where the transmission of nerve
impulses must be blocked, and anti-arrhythmic drugs that are needed to restore
the beating of the heart to normal. Our understanding of the structure of the
sodium channel has increased greatly in recent years (reviewed in Catterall,
1988).
The sodium channel from mammalian brain consists of three subunits: (Y
(260 kDa), /3i (36 kDa) and p2 (33 kDa). The p2 subunit is linked to the (Y
subunit by disulphide bonds; all three subunits are heavily glycosylated at the
surface which faces the extracellular space, while the (Y subunit contains a
site for phosphorylation by cyclic-AMP-dependent kinase on the cytoplasmic
face. The (Y subunit contains the voltage-gated channel, while the other sub
units stab&e the channel in the functional state.
The (Y subunit consists of three cytoplasmic and four extracellular loops
separated by four homologous transmembrane domains - each with six seg-
ments that traverse the membrane in the form of cr-helices. The amino and
carboxy termini lie within the cell. The ion channel is believed to lie in an
almost symmetrical square array 0.3 to 0.5 nm in size, formed by the four
transmembrane domains.
The voltage dependence of channel opening is linked to conformational
changes that result from the movement of charges (voltage sensors) when the
transmembrane potential changes. Positively charged voltage sensors have
been found in the fourth segment (S,) of each transmembrane domain. At rest
all the positive charges in S4 are paired with negative ones and the transmem-
brane segment is held in that position by the negative membrane potential.
Depolarization reduces the forces holding the charges together. The S4 helix
slides and rotates in such a way as to produce an unpaired negative charge
on the inner surface and a positive one on the outside surface to give a net
charge transfer of +1 (Catterall, 1988; Grant and Wendt, 1992).

10.2.1 Sodium channelblockers -procaine as alocal anaesthetic

Local anaesthetics block both the generation and the transmission of sensory
nerve impulses by binding to sodium channels in the nerve cell membrane.
Normally, a slight depolarization of the membrane opens the sodium channel
to allow a rapid but transient inward current. Local anaesthetics raise the thres-
hold for electrical excitation of the membrane to such an extent that the per-
meability of the membrane becomes so low as to result in a block of nerve
conductance. Small nerve fibres are blocked more rapidly than thick ones,
possibly because the former are demyelinated, unlike the latter. All the com-
 Membrane-active agents 221

monly used local anaesthetics contain a secondary or tertiary amine with a

pK, between 8 and 9, so that at pH 7.4 between 5 and 20 per cent of the
drug will be in the form of a neutral amine (pK, is the pH at which 50 per
cent ionization occurs. As pH is a logarithmic scale, at one unit either side of
the pK,, only 5 per cent of one component is present.) The activity of the
drug increases with increasing lipophilicity, whereas the cationic form binds
to the sodium channel. Increasing lipophilicity probably helps the drug to
diffuse into the cell in the neutral form (Postma and Catterall, 1984). The drugs
act more rapidly when added to the cytoplasmic side of the axonal mem-
branes, rather than the extracellular side, suggesting that the binding site is
nearer the cytoplasm. Affinity of the drugs for the channel depends on the
voltage, and the channels are blocked more rapidly and completely when the
channels are repeatedly activated (use-dependent Inactivation or block). The
cationic form binds more rapidly to the open form of the channel because it
can gain access more readily but more tightly to the inactivated channel. In
the resting state the channel is not sensitive to the drug.
The drugs have a strictly limited life in the circulation since they are all
esters and are hydrolysed by plasma esterases. This ensures that one of the
major requirements of a local anaesthetic is met, namely a transient and revers-
ible anaesthesia. In principle, however, it is possible to see that such drugs
could be of use as anti-arrhythmics if their duration of action were sufficient;
in procainamide, an amide link replaces the ester group in procaine and the
reduced hydrolysis by esterases allowed the drug to be used as an anti-
arrhythmic. Lidocaine is an exception in being used both as a local anaesthetic
and anti-arrhythmic, but in the latter case it is usually during acute myocardial
infarction or cardiac surgery where a long duration of action is not required.

, CzH5 , (2%
H2N CO - NH(CH2)2N NH-CO-CH*-N
’ Cd% ’ WS

Procainamide Lidocaine

Local anaesthetics find a use in dentistry and a variety of surgical procedures

to reduce the pain suffered by the patients, without causing unconsciousness.
Examples of this procedure include the injection of a local anaesthetic to block
a particular nerve such as the median or ulnar at the elbow to allow surgery
at the wrist (nerve-block anaesthesia); infiltration anaesthesia where a particu-
lar tissue is to be removed and surface anaesthesia where an operation is to
be carried out on the mucous membranes of, for example, nose, mouth and
throat (Ritchie and Green, 1985).
222 Molecular mechanisms of drug action

10.2.2 Anti-arrhythmic agents - lidocaineasanadi-smhythmic

Anti-arrhythmic drugs may be classified into five groups (Muhiddin and Turner,
1985); sodium channel blockers such as lidocaine and procainamide belong
to class 1; sympatholytic agents such as Pblockers (Chapter 9) form class 2;
class 3 is composed of agents such as amiodarone that prolong the repolariz-
ation phase; calcium channel antagonists form class 4 (section 10.3.1) and
chloride channel antagonists form class 5 (section 10.4). It is the agents of
class 1 with which we are concerned in this section.
Class 1 anti-arrhythmic drugs block sodium channels in both nerve and car-
diac cells, but their efficacy is greater on the latter by a factor of between 10
and 200, and so local anaesthetic action is likely to make only a very modest
contribution while the drugs are being used for heart conditions. Three states
of the cardiac sodium channel are postulated: resting, activated (open) or inac-
tivated (closed) after use. These different states of the channel have different
affinities for the drugs. Furthermore, the idea is confirmed that drug binding
to the channel shuts off conductance. Repolarization of the membrane closely
parallels the activation of the channel in that when the membrane potential
has returned to the resting state, the channels are reactivated. The class 1 anti-
arrhythmics have, as their main effect, the reduction of the maximum rate of
depolarization. In addition, they may raise the threshold of excitability, making
the cardiac cell less sensitive to being triggered. These drugs may also prolong
the period during which no action potential can be initiated, known as the
refractory period. They produce no change in the resting potential (Muhiddin
and Turner, 1985).
Sodium channel blockers (group 1) vary somewhat in their mode of action
and are classified into three groups based on the effect they have on the dur-
ation of the action potential. Those that prolong the action potential, e.g. diso-
pyramide, are grouped in 1A. Those that shorten it, such as lidocaine, are in
group lB, while group 1C have no effect. Lidocaine binds to the inactivated
channel and so the degree of block increases as the membrane remains depola-
rized; the drug reduces the chance of the channel opening during depolariz-
ation (Grant, 1992).
The concept of a use-dependent block, noted above for local anaesthetics,
also holds good here - if the rate of heart beat is sufficiently fast the effect of
the drug increases with every beat. This has been termed the ‘modulated
receptor hypothesis’ (Hondeghem and Katzung, 1984). As with local anaes-
thetics (section 10.2.1) the neutral form of the drug can penetrate the channel
via the lipid membrane and ionize in the channel. This is particularly favoured
if the channel is closed (Hille, 1977).
Disopyramide binds rapidly to the open channel and little further block
occurs after the channel has closed. This drug is useful for both atria1 and
ventricular arrhythmias. In contrast, lidocaine is only useful for ventricular
arrhythmias where there is time for block to develop.
 Membrane-active agents 223

The sodium channel blockers also vary in the speed with which they dis
sociate from the channel, and this fortuitously follows their classification. Lido-
Caine dissociates rapidly with a rate constant of less than 1 second, whereas
disopyramide is between 1 and 10 seconds (type 1C drugs have a rate constant
greater than 10 seconds). The slower the rate of dissociation, the greater the
prolongation of slower conduction (Grant, 1992). This may have caused the
threefold Increase in sudden death noted for the 1C drugs on the Cardiac
Arrhythmia Suppression Trial (CAST, 1989). If the conduction is slowed too
much the impulse may re-enter the circuit and produce arrhythmias. This
effect may lie behind the increase in mortality. Furthermore, patients were
selected who were at a moderately low risk of dying so the drug treatment
was more dangerous than placebo (Grant and Wend& 1992).

10.2.3 Diuretics - amiloride andtriamterene

The type of sodium channel discussed in the previous two sections is that
present in tissue which is electrically excitable. In epithelial tissue that is not
electrically excitable, such as that lining the surface of the kidney or cornea,
there are three other types of sodium channel which are characterized by
insensitivity to toxins, such as tetrodotoxin even at high concentrations, and
by a slower rate of ion transport than the fast sodium channels of the heart.
One group of these channels is characterized by a high trans-epithelial poten-
tial difference and resistance; these are known as ‘tight’ epithelia and occur
in the toad urinary bladder. Another group, which includes the proximal
tubule of the kidney (Pig. 10.4) have a very low or negligible trans-epithelial
potential and are able to transport large volumes of isotonic fluid (leaky
epithelia). Intermediate between these extremes are the channels with moder-
ate trans-epithelial potential difference such as those in the cells of the distal
portion of the kidney tubule (reviewed in Cuthbert, 1977) and occur in the
hatched area for amiloride and triamterene action (see below) in Fig. 10.4.
The role of these sodium channels in the kidney is to reabsorb sodium ion
which, together with chloride ion and other solutes, has been filtered from
the blood in the glomerular portion of the kidney (Fig. 10.4). At the same
time, in order to maintain equal osmotic pressures either side of the kidney
cell membrane, fluid is absorbed.
The transport of sodium across the tight epithelia is reasonably well under-
stood in that sodium passes across the membrane from the bladder into the
cell, passively down an electrochemical gradient. Sodium pumps located on
the other side of the cell (known as the basolateral) remove the sodium by
an active transport process across the membrane coupled to an enzyme
hydrolysing ATP. The sodium ion flow may help to carry chloride ions in
addition, usually against the electrochemical gradient of the latter. Transport
systems in leaky and intermediate epithelia are less well understood at present
although it is probable that an electroneutral sodium exchange for protons
may take place as in the renal proximal tubule.
224 Molecular mechanisms of drug action

Bowman’s
capsule
Blood /

in Filt&e //

Proximal -
tubule

Descending
limb - -

Thin
ascending
:’ limb

Loop of Henle

The area hatched thus, represents the site of action of

;Eil;;$,um channel blockers, triamterene and

The area hatched thus, represents the site of action of

the chloride channel blockers, furosemide and
ethacrynic acid.

Figure 10.4 Site of action of sodium and chloride channel blocking diuretics.

Amiloride and triamterene are both commonly used diaminopyrimidine

diuretics which block the sodium channel in the distal portion of the kidney
tubule (Fig. 10.4). Amiloride has been shown to bind to the exterior surface of
the cell membrane instantaneously and reversibly. The drug thereby prevents
sodium transport and increases the electrical resistance of the toad bladder
that is used as a model of a tight epithelium. The reduction in the negative
charge of the membrane (depolarization) is blocked, and consequently the
secretion of potassium ion into the lumen of the tubule, which normally serves
to maintain the membrane potential, does not occur (reviewed in Kokko,
1984; Iant, 1985a).
Amiloride blocks sodium transport in tight epithelia with an I& less than
10m6~ (see the comprehensive review by Benos et al., 1992). The predomi-
nant form of the drug at neutral pH is cationic with the positive charge resonat-
 Membrane-active agents 225

ing in the amidinium moiety (pK, 8.7). As noted for local anaesthetics, how-
ever, the proportion of the uncharged form will be appreciable and this may
be the form that crosses lipophilic membranes.
The binding site for amiloride is not known at present, nor is the kinetics
of its interaction with sodium defined as competitive, non-competitive and
mixed inhibition have all been noted in different experiments. The evidence
suggests that amiloride acts as a molecular cork by physically plugging the
channel, rather than by binding outside and causing it to block by a confor-
mational change. The molecular weight of the channel is 730 kDa and it is
composed of six non-identical subunits ranging from 40 to 3 15 kDa in molecu-
lar weight. The smallest subunit is apparently the G-protein ale3 which renders
the channel sensitive to inhibitory control by atrial natriuretic hormone. The
membrane-spanning 150 kDa subunit binds amiloride on its extracellular sur-
face and may form the channel by itself (Benos et al., 1992).
Amiloride exerts its effects on the sodium channel and not on the outward
potassium channel, which indicates that the two alkali metal cations are trans-
ported by different channels - not surprising in view of their different sizes.
Amiloride and triamterene do not give rise to excessive potassium secretion
(spironolactone acts similarly; Chapter 6) - thus preventing hypokalaemia. The
drugs’ major action is to prevent sodium (and concomitantly chloride ion)
reabsorption together with fluid. The urine is thereby increased in volume,
sodium and bicarbonate ion excretion is increased (the pH of the urine rises)
and there.is a fall in potassium excretion.
There are some conditions where there is an excess of water (oedema) in
certain peripheral tissues of the body and also the lung, e.g. after heart failure,
when it is therapeutically advantageous to block the reabsorption of fluid. This
can be carried out by blocking the re-uptake in the kidney of sodium ion
through specific channels, as with amiloride and triamterene, or chloride ion
in a similar fashion by furosemide and ethacrynic acid (section 10.4), although
the specific agents operate at different sites in the kidney tubule. These drugs
all help to reduce excess fluid in the tissues, and thereby reduce the workload
of the heart - clearly an important contribution to the therapy of heart disease.

NH
4
Cl ,N CO-NH-C
\
NH2
\ ’
HzN xx N NH,

Triamterene Amiloride

10.3 The calcium channel

Calcium is indispensable for the initiation of a number of essential cellular
processes, notably muscle contraction (e.g. in the heart), secretion of neuro-
226 Molecular mechanism.s of drug action

transmitters and hormones from neurone and endocrine cells, and rhythmic
firing of heart and nerve cells. Some of the calcium required for this activation
may be found intracellularly, bound to endoplasmic reticulum, mitochondria
and to the calcium binding proteins such as calmodulin. The quantity of cal-
cium available in these sites is sufficient to support only a short burst of
activity, however, and for longer sustained activity calcium has to be drawn
into the cell from outside through specific ion channels (reviewed in Droog-
mans et al., 1985) except for the platelet which can sustain the release of
ADP from intracellular stores of calcium.
In the heart, the depolarizing effect of calcium entry has two effects: it
controls the conductance in the sinoatrial and atrioventricular (AV) nodes (a
small mass of conducting tissue that links atrium and ventricle via the bundle
of His and forms an important part of the conducting system of the heart). In
other parts of the heart, where the sodium current is responsible for the action
potential, the major role of calcium is to mediate excitation-contraction coup-
ling of the heart muscle. Calcium binds to troponin, thus relieving the inhi-
bition of contraction, and releasing actin and myosin to cause contraction.
Calcium channels are in a continuous state of flux, opening and closing
repeatedly, typically remaining open for 1 ms and then closing for between 1
and 100 ms. The channels may be opened by the binding of a neurotransmitter
or hormone to the cell surface, in which case they are termed receptordepen-
dent. Neurotransmitters activate the channel not by increasing the number of
ions flowing through per unit time (conductance), but rather by increasing
the length of time that the channel remains open (and reducing the time that
it is shut) (reviewed in Schramm and Towart, 1985). Catecholamines, such as
noradrenaline, usually act through cyclic AhIP and its attendant protein kinases
to phosphorylate a membrane protein to open the channel while, in contrast,
phosphatases dephosphorylate the protein during inactivation (see Glossman
et al., 1982).
Alternatively, other calcium channels may be influenced by changes in the
membrane potential and these are known as voltage-dependent channels.
Using the patch-clamp technique, whereby a single cell can be maintained at
a given potential difference across its membrane, the kinetics of opening and
closing (known as ‘gating’), conductance and sensitivity to various drugs can
be determined.
Recent studies indicate that there are three types of voltage-operated cal-
cium channels in sensory neurones and two types in cardiac cells (reviewed
in Spedding and Paoletti, 1992). They differ primarily in that different
reductions in the membrane potential are required to effect their gating. As
the membrane is depolarized a channel is opened (called T for transient) that
shows a moderate current and rapid decay. Stronger depolarization opens a
channel termed L (for longer lasting) as it decays only slowly, while another
channel termed N (neither T nor L) is also opened by strong depolarization
and grows more strongly with increasing depolarization. The most prominent
calcium channel types in the heart are T and L. The L type, which is the target
 Membrane-active agents 227

for the calcium channel blockers is pentameric with subunits of o1 that con-
tains the pore, LY*, p, y and 6. Although the calcium channel has two extra
subunits, it otherwise resembles the sodium channel, in that two of the subun-
its are linked by disulphide bonds to an (Ysubunit and the subunits are exten-
sively glycosylated. The respective (Y subunits show an extensive homology,
possibly caused by gene duplication. There are phosphorylation sites on the
(pi and p subunits which respond to Padrenergic stimulation and mediate
channel opening via cyclic-AMP-dependent kinases. The sodium channels,
however, operate a faster conduction close to the diffusion limit and generate
a more rapidly rising action potential than the calcium channels, originally
known as the slow channels (Catterall, 1988; Katz, 1992).
The calcium charmel subunits are made up of four potential membrane span-
ning domains separated by cytoplasmic and extracellular loops. As with the
sodium channel, there are six homologous transmembrane stretches of a-helix
in each domain. The fourth (S,) has been proposed as the voltage sensor
responding to depolarization as it contains a string of positively charged argi-
nine and lysine residues.
As with all biologically active molecules, mechanisms must exist for termin-
ating their action. In the case of calcium it clearly cannot be metabolized and
so it has to be removed from the cell cytoplasm. One mechanism is via the
calcium pump - a calcium-cahnodulin-activated membrane enzyme which
uses the energy from the hydrolysis of ATP to drive calcium ions out of the
cell. This enzyme is activated as soon as the cellular level of free calcium ion
rises much above the resting level of approximately 2 X lo-’ M. The maximum
activation of the contractile proteins in smooth muscle cells, for example, is
achieved at lop5 M, and this level also induces secretion in hormone-producing
cells that are activated by calcium ions.
Another cellular mechanism for removing calcium from the cytoplasm is by
the sodium-calcium exchange, whereby calcium is pumped out of the cell in
return for sodium entry. Cardiac glycosides exert their pharmacological effects
on this exchange (section 10.5.1). The extracellular level of calcium ion is
approximately 10e3~ and the gradient is maintained to a large extent by the
very low permeability of the cell membrane to the doubly charged alkaline
earth metal.

10.3.1 Calciumchannelantagonists -verapamil,nifedipinefor

hypertensionandheartfailure
In addition to the patch-clamp technique noted above, pharmacological inter-
vention has led to great advances in our understanding of the calcium channel.
There are three main families of calcium channel blockers (or calcium
antagonists) in clinical use: phenylalkylamines exemplified by verapamil, 1,4
dihydropyridines of which nifedipine is a major contender and benzothiazep-
ines such as diltiazem. These ligands include a number of drugs used for the
control of angina pectoris, arrhythmias and high blood pressure.
228 Molecular mechanisms of drug action

Rl R2 R3 R4 R5

Nifedipine H NO2 CH3 C02CH3 CH3

Nimodipine NO2 H CH (CH,), CO,KH,),OCH, H
Bay k 8644 H CF, CH3 NO2 CH3

WW’JCH~)z
Verapamil
Diltiazem

The earlier concept that the channel could be either open, closed and
resting, or closed and inactivated has been modified in the light of measure-
ments on a single channel The channel can remain open for either short or
long periods. The effect of the calcium channel blockers is, however, not to
act as molecular plugs but rather to reduce the probability that the channel
will be in the open state long. In addition, the ligands bind more effectively
to the channel when the membrane is depolarized.
These three classes of drug vary, however, in their effects on the (slow)
calcium channel. Although nifedipine reduces the inward calcium current, it
does not affect the rate of recovery of the calcium channel. Verapamil, on the
other hand, as well as reducing the inward flow of calcium also reduces the
rate of recovery of the channel. Furthermore, verapamil and diltiazem also
block more effectively as the frequency of stimulation increases (usedepen-
dent block).
Although there is a much greater understanding of the molecular architec-
ture of the calcium channel than in earlier years, we still do not know much
about the specific binding sites of the classes of antagonists. Recently, how-
ever, a potential verapamil-binding site on channels from skeletal muscle has
been identified on the cytoplasmic face of the channel. Verapamil and ana-
logues behave in a fashion similar to the local anaesthetic blockers of the
sodium channel in that they are weak bases with pK, in the region of 8.5.
There is sufficient drug in the uncharged form at neutral pH to enter the cell
and block the open state of the channel by approaching from the cytoplasmic
side. A putative binding site composed of aspartic acid residues has been pro-
posed on segment 6 near the mouth of the channel (Striessnig et al., 1991).
Chirality plays a part in the action of these drugs in that the S-(-)-form of
verapamil is six to ten times more active than the R-(+)-form; diltiazem is most
 Membrane-active agents 229

active as the 2S,3S form. In fact, ion channels have been described as chiral
receptors (Kwon and Triggle, 1991).
The effects of calcium entry blockade are more readily seen on the whole
heart. Verapamil was first shown by Fleckenstein’s group to mimic the effects
of calcium withdrawal, as the drug reduced (a) calcium-dependent contraction
without affecting sodium-dependent aspects of the action potential and (b)
calcium-dependent myofibrillar ATPase activation and oxygen consumption of
the beating heart (Fleckenstein-Grun et al., 1984).
The situation in which these effects may be particularly useful is when the
oxygen available to part of the heart is greatly reduced as, for example, after
a myocardial infarction. In this condition a blockage occurs in a branch of an
artery supplying the heart, cutting off part of its blood supply. The area
involved rapidly uses up the available oxygen and becomes hypoxic. It is
believed that catecholamines released as a result will greatly increase calcium
uptake into the myocardial cell, sending the muscular apparatus into a contrac-
tile spasm and using up ATP in the proces. Mitochondria try to take up calcium
from the cytoplasm and use up more ATP. Eventually the calcium load damages
the structure of the organelle and uncouples oxidative phosphorylation. The
loss of ATP ultimately shuts off the plasma membrane calcium pump, thus
allowing calcium to accumulate further and increase the damage. Calcium
channel blockers, although less effective against neurotransmitter-operated
channels may still be of use in a cardioprotective fashion (Kendall and Horton,
1985) and indeed are already used before cardiac surgery (Fleckenstein-Grun
et al., 1984).
At present, however, the main uses of calcium channel antagonists are for
the treatment of angina pectoris, hypertension and arrhythmias (Schramm and
Towart, 1985). Angina pectoris is characterized by dull chest pain brought on
by conditions of stress such as over-exercise, over-indulgence, smoking etc.
and is the result of insufficient oxygen for the oxidation of substrates in the
heart muscle. Calcium antagonists are able to reduce the oxygen demand of
the heart muscle by reducing the amount of muscular activity and are particu-
larly useful for the treatment of angina pectoris.
Another type of angina, known as variant angina, is characterized by a
reduction in blood flow in the heart rather than lack of oxygen. Calcium chan-
nel blockers are of use in this condition because they are particularly effective
at relaxing cardiac smooth muscle in the coronary arteries, thereby increasing
the diameter of the artery. As a consequence, the effort the heart has to make
to drive the blood through the vessels is reduced and the blood pressure falls.
In addition, calcium channel antagonists relax peripheral blood vessels, there-
by reducing blood pressure, and can be used to treat both acute hypertensive
crises and chronic hypertension.
The type of arrhythmia in which there is a change in the normal beating of
the heart by a group of cardiac cells beating in a disorganized fashion (when
pacemaker groups of cells arise in the ‘wrong’ part of the heart) responds to
verapamil, but not nifedipine (Spedding, 1985). In some cases these arrhyth-
230 Molecular mechanisms of drug action

mias can arise from a partial depolarization of the membrane causing the exci-
tation process to depend on the slow calcium channel rather than a fast
sodium channel (Fleckenstein-Grun et al., 1984).

10.4 Coupled sodium-chloride ion channels - furosemide

and ethacrynic acid as diuretics
As noted in the section on sodium ion channels (section 10.2.3) chloride is
reabsorbed in the kidney by active transport systems of which that in the
ascending limb of Henle’s loop is particularly active (Fig. 10.4). The transport
system has been proposed as a sodium-chloride co-transport with sodium
entering the cell down its electrochemical potential and chloride moving
against its electrochemical gradient. The chloride ion exits across the opposite
border of the cell down a favourable potential difference (Frizzell et al., 1979).
In earlier years it was thought that chloride uptake alone was the target of the
drugs discussed in this section and so much of the work has been directed to
the transport of chloride but not sodium (hence the emphasis on chloride in
the studies reported below). Nevertheless, it is now clear that the co-transport
of the ions is the major target.
Two structurally unrelated diuretics, furosemide and ethacrynic acid, among
others, act to inhibit this coupled transport and are known as ‘loop’ diuretics.
Up to 25 per cent inhibition of chloride and sodium reabsorption is possible
with these agents which is a far higher proportion than that obtainable with
other diuretics - hence the term ‘high ceiling’ is used to describe these
diuretics. Perhaps not surprisingly, such a powerful effect is characterized by
swift onset in about 15 minutes and rapid cessation in 4-6 hours (Kokko,
1984; Lant, 1985a).
Furosemide acts to inhibit active chloride transport from within the tubular
lumen, i.e. the fluid side not the intracellular side (Odlind, 1984). The drug
reaches the lumen by being transported from the proximal tubule cell through
a non-specific channel that secretes organic acids, thus using the normal kid-
ney transport pathways to reach its site of action (Kokko, 1984; Lant, 1985a).
Furosemide demonstrates competitive kinetics with chloride ion uptake and
noncompetitive with sodium ion which suggests, but does not prove, that the
drug may bind at or near the chloride uptake sites (Ludens, 1982). Coupled
transport exhibits saturation kinetics as would be expected of an active trans-
port system. Furthermore, ouabain blocks active transport of chloride which
suggests that the process is linked to a Na+,K+-ATPase as this drug is a specific
inhibitor (section 10.5). It is impossible, at present, to be more precise because
no definitive receptor has yet been identified for furosemide.
Frog cornea, as another epithelial system, is believed to contain similar
coupled transport systems to the kidney, although the cornea secretes chloride
whereas absorption occurs in the ascending limb of Henle. Electrochemical
studies have confirmed the selective inhibitory effect of furosemide on chlor-
 Membrane-active agents 231

ide uptake and have shown that the drug increases the electrochemical gradi-
ent of chloride across the apical membrane of frog cornea. It prevents equaliz-
ation of chloride concentrations either side of the membrane, leading to
hyperpolarlzation of the apical membrane and depolarization of the basolateral
membrane (Patarca et al., 1983). The net effect is to reduce the permeability
of the membrane to chloride transport.

Ethacrynic acid Furosemide

The receptor for ethacrynic acid has also not yet been identified. It is likely,
however, bearing in mind the differences in structure between furosemide
and ethacrynic acid, that the receptors for the two drugs differ. Ethacrynic
acid is effective at blocking chloride transport across the epithelial membrane
of the ascending loop of Henle with a concomitant decrease in potential differ-
ence. A decrease is found in this instance because the potential difference
across the luminal membrane is maintained by the chloride transport (Burg
and Green, 1973). Interestingly, the cysteine adduct of ethacrynic acid, which
appears to be a normal excretion product, is far more effective than the drug
itself; this adduct may therefore be the active form of the drug in vivo.
The contention that the loop of Henle is the sole site of action of these
drugs has been challenged by Caffruny and Itskovitz (1982) who demonstrated
that both drugs can act to block chloride transport in the proximal tubule,
but the importance of this observation is not entirely clear. Obviously, there
is much that we do not yet understand in the precise molecular mode of action
of the sodium-chloride channel blocking diuretics.

10.5 Membrane-bound ATPases

There are a variety of enzymes in eukaryotic and prokaryotic systems that use
the energy of hydrolysis of ATP to drive the transport of protons and other
cations across membranes that are otherwise impermeable to these ions. These
enzymes are all dependent upon magnesium, presumably because the true
substrate is an ATP-magnesium complex. One group that is part of the process
of oxidative phosphorylation is characterized by the reverse reaction of ATP
synthesis coupled to oxidation and is found in the membranes of mitochondria,
bacteria, chloroplasts etc. The mitochondrial enzyme is crucial to the chemios-
motic principle of Mitchell which states that phosphorylation of ADP is driven
by the energy of a proton gradient across the mitochondrial membrane. These
enzymes are often referred to as F,JF,-ATPases because they consist of two
232 Molecular mechanisms of drug action

portions: one is responsible both for anchoring the enzyme in the membrane
(F,) and for the translocation of hydrogen ion, while F, is the binding site for
the interconversion of adenine nucleotides. General inhibitors of these
enzymes are oligomycin which acts on F0 and blocks proton transport and
dicyclohexylcarbodiimide (DCCD), which reacts with a carboxyl group crucial
for the enzyme reaction (reviewed in Pedersen, 1982).
Another group of ATPases, bound to the plasma membrane of the cell, is
concerned with the pumping of cations against a concentration gradient out
of the cell; notable examples are: (a) the enzyme that catalyses the exchange
of sodium for potassium in the cardiac cell membrane particularly, although
it is also found in other tissues, (b) the calcium pump which drives calcium
out of secretory cells, among others and (c) the proton pump which
exchanges protons for potassium ion in the parietal cells of the stomach and
thereby produces the low gastric pH. In general, the cation pumps consist of
at least one polypeptide with molecular weight in the region of 100 kDa which
usually acts as a channel for the cation transport.
Oligomycin does not inhibit these enzymes but DCCD does, again because
there is a carboxyl group in a glutamate side-chain which is essential for
activity. A phosphate-enzyme intermediate is an essential step in the mechan-
ism, the phosphate being attached to the active site aspartate. Vanadate is a
characteristic inhibitor of these enzymes, probably because it can replace the
phosphate in the intermediate, forming a vanadyl-enzyme intermediate. A con-
siderable degree of homology between cation pumps has been discovered,
which suggests that they may have evolved from a common ancestor (Serrano
et aE., 1986).
Mitchell has suggested a charge relay type mechanism in which the transport
of a proton in one direction can be balanced by the transport of an ion such as
potassium in the opposite direction. The aspartyl-phosphate conveys a proton
outwards with the phosphate group, while potassium is translocated inwards
together with another phosphate on the same aspartyl group. The aspartyl-
phosphate group is assumed to rotate and go from a high energy to a low
energy conformation (reviewed in Pedersen, 1982). Not all the cation pumps
are electroneutral, however, since the Naf,Kf-ATPase pumps three sodium
ions out of the cell in exchange for the entry of two potassium ions for every
ATP molecule hydrolysed, resulting in a change of potential difference across
the membrane.

10.51 Sodium-potassium-ATPase - cardiac glycosides for heart

failure
Heart cells, in common with most eukaryotic cells, need to maintain gradients
of the alkali metal cations - potassium levels must be kept high inside the cell
and sodium low, in order to maintain the cell’s integrity. To do this, the cell
requires a pump to eject sodium ions from the cell in exchange for potassium
ions. This will counterbalance the effect of sodium intlux and potassium efflux
 Membrane-active agents 233

through the respective ion channels during the early part of the action poten-
tial which accompanies a heart beat. The transmembrane ATPase that effects
this exchange has been purified from a number of sources, and, when inserted
into lipid vesicles, can catalyse the exchange transport of alkali metal cations
at a rate approaching that seen in vivo (Jorgensen, 1982).
The enzyme is made up of two subunits: the (Ysubunit of molecular weight
about 100 kDa and the p subunit of about 40 kDa, although minor variations
in the cxsubunit may be found between tissues to a sufticient extent to warrant
the designation of isozymes (Lytton, 1985). The enzyme spans the membrane
with the catalytic and sodium activator sites facing the cytoplasm and the
potassium binding site on the extracellular surface. All these sites are located
on the (Ysubunit: no functional role has yet been described for the p subunit.
Both the amino and carboxyl termini of the (Ysubunit face the cytoplasm and
the polypeptide chain traverses the lipid bilayer several times in between. The
mode of transport of the cations is not known, although it has been suggested
that the ionic channel may pass either through the (Y subunit or between ad-
jacent (Ysubunits (Jorgensen, 1982).
In the presence of sodium ions and Mg-ATP on the intracellular side of the
enzyme, the carboxyl side-chain of an essential aspartate group is phosphorylated
to yield a high-energy phosphate. A conformational change takes place as ADP
is released (E,P - E,P; where E, and E, represent different conformational
forms of the enzyme). Potassium ions enter the (Y subunit from the other side
of the membrane and bind to E,P causing it to lose the phosphate group. K+E,
breaks down to free potassium and enzyme to complete the catalytic cycle.
The conformational change induced by ATP phosphorylation is potentiated
by sodium and dephosphorylation is activated by potassium. In separated and
purified enzyme preparations, sodium ion stab&es El and potassium E,. ATP
binds with high affinity to E,Na in a hydrophobic pocket and only weakly to
E2K in an open structure. Magnesium ion is required for activity as, with most
3enzymes that use ATP, Mg-ATP is the true substrate. This rather complicated
reaction scheme may be envisaged as follows (after Jorgensen, 1982):
Nati, K+i represent cytoplasmic cations; Nafe, K+, extracellular.

KEl ATP 4
234 Molecular mechanisms of drag action

Cardiac glycosides are used in the treatment of congestive heart failure, and
appear to act by inhibiting the sodium pump (Hansen, 1984). They are com-
posed of a large steroid-like structure with an unsaturated lactone ring attached
to position 17 and a sugar, or chain of sugars, linked to position 3. All these
parts of the molecule are essential for activity. These drugs were discovered
as part of the extract of foxgloves, digitalis, which was recommended for the
treatment of heart failure as long ago as 1785. Digitoxin is the major constitu-
ent of digitalis, while digoxin is also present to a lesser extent. The original
preparations from foxgloves have now been replaced by chemical preparations
because the dose is much easier to control and is not subject to variable com-
position. The cardiac glycosides are stilI very useful drugs for the treatment
of cardiac failure despite occasionally serious incidents of kidney toxicity.
Digoxin is probably the most widely used agent (Crazier and &ram, 1992).

Experimentally, ouabain is probably the most studied of the cardiac glyco-

sides and it has been shown to bind tightly but reversibly to the cr subunit of
the enzyme on the extracellular face with an IcsOof lo-’ M. The aftinity of the
drug for the isozyme (Y’ of brain and kidney is much less with an Icso of lo-* M
(Sweadner, 1979). The enzyme undergoes a large conformational change on
ouabain binding which can be partially antagonized in a non-competitive
fashion by potassium, suggesting that the latter binds at a different site. Oua-
bain binds preferentially to the E,P conformation of the enzyme but the affiity
of that form is reduced for the drug if it is complexed with potassium. Oua-
bain-enzyme complexes will also bind ATP but with greatly reduced afhnity
(Hansen, 1984).
Ouabain is thus able to block the translocation of alkali metal cations in and
out of the cell at concentrations of about 10-6~, and the rate of ouabain
binding parallels inhibition of rate of pumping. If the pump is inhibited, intra-
 Membrane-active agents 235

cellular sodium levels rise. Consequently, calcium levels also rise as a result
of a transmembrane sodium-calcium exchange. The raised calcium level acti-
vates the heart to contract more forcefully (the positive inotropic effect).
There have been suggestions, however, that this positive inotropic effect can
be dissociated from pump inhibition, since after the cardiac glycoside has been
washed out of the heart the enzyme was still significantly inhibited whereas
the positive inotropic effect was no longer present. One explanation may be
that another unknown mechanism operates in addition to pump blockade
(reviewed in Godfraind, 1984). On the other hand, the degree of enzyme inhi-
bition required to show an observable effect on the pumping of the heart may
be very high.

Ouabain
HO OH

Other studies by Brown and Erdmann (1984) however, have cast doubt on
this hypothesis. They obtained biphasic Scatchard plots and suggested that
there were two sets of binding sites for ouabain on rat and guinea-pig cardiac
membranes - a high affinity, low capacity site, the occupancy of which was
associated with the positive inotropic effect, and a low affinity, high capacity
site which was apparently the sodium pump. The high affinity site was not
identified. With cat and human cardiac membranes the position was much less
clear-cut with only a very slight curvature of the Scatchard plots. Bearing in
mind that biphasic Scatchard plots can arise for reasons other than multiple
sets of binding sites, the detailed mode of action of the cardiac glycosides is
still a subject for debate.
The action of the cardiac glycosides is particularly valuable in congestive
heart failure which is characterized by, among other things, an enlarged heart
and a high venous pressure as a consequence of blood flow congestion. The
increased contraction forces the blood out of the heart and it reduces in size.
As a consequence, the sympathetic activity is reduced and the venous pressure
falls back towards normal. In addition, the drugs slow the ventricular rate of
beating by rendering this part of the heart less sensitive to electrical stimuli
known as a refractory state. Digoxin may not be as effective as the angiotensin-
converting enzyme inhibitors captopril and enalapril, however, nor as free
236 Molecular mechanisms of drug action

from serious side-effects and the latter are preferred for mild to moderate con-
gestive heart failure (Crozier and &ram, 1992).

10.5.2 Potassium-hydroge - omeprazole for ulcers

Recently, an enzyme has been discovered which pumps protons out of the
parietal cells that release acid into the stomach and is believed to be the ter-
minal link in the chain of acid secretion. This proton pump is the target of a
new type of anti-ulcer drug that is exemplified by omeprazole, a substituted
benzirnidazole.
The enzyme appears to be unique to the parietal cell of the gastric mucosa,
and is the last link in the chain of the transport of hydrogen ions into the
small channels (canaliculi) that eventually lead to the surface of the cell leading
into the gut. These canaliculi form a highly acidic space in the cell and exhibit
a convoluted membraneous structure with short microvilli projections. The
pump is found in a particular type of vesicle which is isolated from the
microsomal fraction of gastric parietal cells. This fraction is probably derived
from the plasma membrane and vesicles which have a smooth surface. Nor-
mally, there are a large number of these vesicles (about 0.2 pm in diameter)
in the cytoplasm of the resting cell. When acid secretion is stimulated by hista-
mine, for example, these vesicles coalesce with the canaliculi and allow pro-
tons to be transported to the cell exterior (see Sachs, I986 for review).
Potassium ion and ATP are required for the action of the proton pump, in
fact the enzyme is a proton-stimulated ATPase which uses the energy of ATP
hydrolysis to drive the exchange of protons for potassium across the vesicle
membrane. ATP can thus generate a pH gradient across the canalicular mem-
brane in parietal cells. The following scheme has been proposed for the overall
reaction (Wallmark, 1989):

Mg-ADP Mg*+
1
ATP
E-E&P
KC

1
E2P ‘T+EZPK+ i4Tb EIK+,
ATP

K', P,

Kfi, internal potassium; K+,, external potassium

Studies on the cloned enzyme show that it is a single 1033-amino-acid
polypeptide with a molecular weight of 110 kDa, sharing 60 per cent homo-
 Membrane-active agents 237

logy with the Na’,K+-ATPase on the cardiac cell membrane, discussed in the
previous section. This similarity extends to the catalytic cycle where a phos-
phorylated form is the obligatory intermediate in the reaction, indeed the
phosphorylated forms of both enzymes are active whereas the dephosphoryl-
ated forms are not. Magnesium ion is also required for activity, as is common
for enzymes where ATP is a co-substrate and it is likely that the Mg-ATP com-
plex is the true substrate. The following scheme has been proposed for the
overall reaction (Wallmark, 1989).
The mechanism of protein pumping is outlined by Sachs and Wallmark
(1989). Mg-ATP binds to the cytosolic face of the enzyme and a hydrated
hydrogen ion (H,O+) binds to the ion-binding region. Phosphate is transferred
from the ATP to the carboxyl group of an aspartate residue to give the E,
form. The enzyme changes conformation to E,P to allow the hydronium ion
to face the outer luminal side; the affinity for the hydronium ion is sharply
reduced and it exchanges for the potassium ion. The phosphoenzyme breaks
down rapidly to give the E,K form which converts to the E,K thus undergoing
the reverse conformational change to present K+ to the cytosolic face. Potass-
ium dissociates, Mg-ATP binds and the whole process can recommence. Pot-
assium and chloride ion are co-transported out through the cell membrane
and thus recycling occurs. The pump maintains a pH gradient with the acid
outside, cytosol neutral, and is electroneutral.
Omeprazole is taken up from the circulation into the acidic canaliculus of
the parietal cell where it becomes protonated. Omeprazole is a pro-drug
because it is stable at neutral pH, but breaks down in the presence of acid to
form a tetracyclic sulphenamide (Fig. 10.5). The acidic lumen has a sufficiently
low pH for this to happen. Sulphydryl groups on the luminal side of the
enzyme make a nucleophilic attack on the sulphur atom, opening the ring to
give a covalent disulphide bond. Complete inhibition occurs when two moles
of inhibitor are bound per mole of active site. This reaction can be inhibited
by pH neutralization and also by sulphydryl reagents such as mercaptoethanol
(Walhnark, 1989).
Gastric mucosal ATPase is the final step in the induction of acid secretion by
various secretagogues; histamine is the most potent and acts through adenylate
cyclase and the formation of cyclic AMP; acetylcholine stimulates the uptake
of calcium ion into the cytoplasm and has no effect on cyclic nucleotide levels,
while the peptide gastrin, which is the physiological mediator of gastric acid
secretion in response to feeding, appears to release calcium from intracellular
stores. In addition, both acetylcholine and gastrin induce the release of hista-
mine from neighbouring paracrine cells (Sachs and Wallmark, 1989).
Clearly, it is possible to inhibit the secretion of acid by (a) blocking the
histamine receptor, as demonstrated by the anti-ulcer drugs cimetidine and
ranitidine and (b) the acetylcholine (muscarinic) receptor, e.g. the drug piren-
zepine. Furthermore, the histamine H,, receptor blockers inhibit gastrin and
acetylcholine induced secretion because these secretagogues also release
histamine (section 9.8.2). The proton pump, however, is the last link in the
238 Molecular mechanisms of drug action

OCF. ?CH, OCH,

L^H, &CH, CH, .&/CH,

OCH,

‘OCH, &CH,

Omeprczzk Sulphenic acid Svlphenamide

\ * Enzyme-SH J

W-4
CH,

OCH,

Enzyme-inhibitor complex
Figure 10.5 Chemical reactions of omeprazole that lead to inhibition of H+,K’-ATPase

chain of proton secretion and blockers of this may be expected to be at least

as effective as blockers of either of the two receptors described above. In
addition, the expected blockade of both histamine, dibutyryl cyclic AMP and
pentagastrin induced secretion in human gastric glands in vitro and in vivo
was found by Olbe et al. (1986). Pentagastrin is a synthetic pentapeptide con-
taining the C-terminal tetrapeptide amide moiety of gastrin and is often used
in tests to measure inhibition of acid secretion as it has appreciable
secretory activity.
In clinical trials, omeprazole has been shown to heal both gastric and duo
denal ulcers more rapidly and completely than cimetidine or ranitidine; in
some cases which were resistant to prolonged treatment with H, receptor
blockers, omeprazole was effective. As we might expect from its mode of
action producing a type of irreversible inhibition, less drug and less frequent
administration is required for omeprazole. Other gastric-acid-linked disorders
such as Zollinger-Ellison syndrome and reflux oesophagitis are also responsive
to omeprazole but not to Hz receptor blockers (McTavish et al., 1991). Omep-
 Membrane-active agents 239

razole is another drug that has made a pharmacological breakthrough and has
also taught us much about the workings of a previously little-known enzyme.

?CH3

Omeprarole

IO. G Cromoglycate - calcium antagonist or membrane

stabilizer?

Asthma is characterized by episodes of difficulty in breathing, varying from

the mild to extremely severe due to widespread narrowing of the airways in
the lung which can be due to a number of factors: the lining of the airway
may swell, there may be mucus in the airway, a spasm in the muscle in the
walls of the airway (Kauffman, 1984). In man, there is an initial acute phase of
the condition followed by two late phases, ranging from 17 to 72 hours later.
Many cells are involved in the pathogenesis of asthma. It seems likely that
the response of the mast cell to an allergen or other stimulus, such as exercise
or cold air, controls the acute phase by releasing chemical mediators such as
histamine and leukotrienes. The late phases may result from activation and
subsequent invasion of the lining of the airways by macrophages, eosinophils
and T lymphocytes. The airways are abnormally sensitive to stimuli which may
be a result of the presence of a large number of eosinophils, a type of white
blood cell that contains granules that readily stain with the red fluorescent
dye eosin (Barnes, 1992).
As a result, asthma is now regarded as a special type of inflammatory disease,
such that the anti-inflammatory steroids are now given much earlier in the
treatment instead of being retained for the really serious cases (Barnes, 1992).
Cromoglycate is still widely used 25 years after it was launched, although its
mode of action is, even now, not understood (Kuzemko, 1989).

10.6.1 Cromoglycate action

The story of the research project that eventually led to cromoglycate makes
very interesting reading, not only with regard to the science, but also the
politics of the pharmaceutical industry (Kauffman, 1984). Khellin, a chromone
(benzopyrone) obtained from the plant Ammi visnagi, was the model for syn-
thesis of chromone 2carboxylic acids because it possessed smooth muscle
relaxant properties. This was believed to be an advantage in relieving the bron-
240 Molecular mechanisms of drug action

choconstriction of asthma. A doctor, Roger Altounyan, who worked for the

company and who was himself a chronic asthmatic, tested the compounds on
himself by breathing in a soup of guinea-pig hair, producing an attack of
asthma by antigenic challenge.
Eventually, by chance, an agent was found that did not possess bronchodil-
ator activity but still protected against antigen challenge. The following syn-
thetic efforts resulted in the classic situation where the first batch of a com-
pound showed great activity but succeeding batches did not. A highly active
impurity was suspected and, on this occasion, the agent was isolated and found
to be a bis (or doubled-up) chromone carboxylic acid, a breakthrough which
led directly to cromoglycate. During the course of this work the project was
officially closed down but work still carried on clandestinely!

Cromoglycate

Cromoglycate is very hydrophilic, being a dicarboxylic acid with pK,s both

in the region of 2. It is therefore a dianion at neutral pH and would not be
expected to pass through the cell membrane easily if at all. Indeed, the drug
may not need to enter the mast cell to inhibit degranulation since it is active
when linked covalently to polyacrylamide beads (Mazurek et al., 1980).
Cromoglycate is only active when given by inhaled aerosol, and the rate of
passage across the lung membranes governs its rate of disappearance from
the circulation.
Mast cells obtained from different tissues vary greatly in sensitivity to cromo-
glycate inhibition of the anti-allergic reaction for which the K’,, for liberation
of histamine can vary from the high concentrations of lo-*-lop3 M for the
human lung mast cell (Holgate, 1989) to 5.5 X ~O-‘M for the mast cell from
rat peritoneum (Moqbel et al., 1988). It is difficult to understand how such a
high figure for the ‘correct’ tissue in humans can relate to the disease process
but, nevertheless, the rise in circulating levels of histamine in an asthma attack
are inhibited (Howarth et al., 1985).
Another mechanism must be at work, however, because the drug blocks
not only allergen-induced, ‘extrinsic’, asthma but also asthma in non-allergic
models, i.e. induced by exercise or cold air (‘intrinsic’). Furthermore, cromo-
glycate does not relax bronchial or smooth muscle in vitro, or acutely in
vivo, whereas after long-term administration, bronchoconstriction induced by
stimulants such as histamine or exercise is greatly reduced (Colgate, 1989).
The late reaction may be governed by the recruitment of neutrophils, eosino-
phils and possibly macrophages which can exert an inflammatory effect. Here
 Membrane-active agents 241

cromoglycate can be shown to be very active, i.e. at levels below lo-'M, in

inhibiting the stimulation of eosinophils by formyl-methionyl-let@-alanine, a
bacterial peptide which activates macrophages to kill bacteria (Kay et al.,
1987).
It is interesting to speculate that cromoglycate may owe its action in some
way to potentiation of the nitric oxide pathway. Nitric oxide has been impll-
cated as one possible neurotransmitter in the nerve system in the lung which
inhibits allergic bronchoconstriction and reduces arterial plasma histamine
(Iammers et al., 1992). Macrophages are activated by cytokines to produce
nitric oxide, with the help of cahnodulin, which stimulates cyclic GMP pro-
duction (Moncada et al., 1991) and the most efficient inhibitors of the allergic
reaction are compounds which potentiate cyclic GMP by inhibiting its hydroly-
sis rather than cyclic AMP hydrolysis (Coulson et al., 1977).

Et
n-Pr
I I
HO&,/$&.&O’”

0 0

Nedocromil

Nedocromil, an analogue of cromoglycate, is the only agent to reach the

market subsequently, despite the very great effort expended in recent years
trying to fmd new anti-allergic agents based on the in vitro mast cell degranu-
lation test. Many new compounds were discovered that were more active than
cromoglycate in vitro but show little or no advantage in vivo; indeed many
were inactive. Although in some cases this has been attributed, amongst other
things, to serum binding and different pharmacokinetic profiles, the lack of
correlation in other cases is likely to be due to an inappropriate in vitro test.
As Roger Altounyan observed, there is no substitute for testing in man
(Kauffman, 1984).

IO. 7 Cyclosporin
There have been a number of examples in the field of drug discovery where
a condition has been treated by a drug without its mechanism of action being
known. Cyclosporin is one such drug. Although its mechanism of action has
not been worked out, the involvement of the calcium-binding phosphatase
calcineurin, that can regulate calcium channels in the brain, suggested cyclo-
sporin’s inclusion in this chapter. Cyclosporin has played a major part in the
recent increase in transplant surgery for organ and bone marrow by sup
pressing the immune response at the level of the T lymphocyte - sufficiently
242 Molecular mechanisms of drug action

well to allow a great improvement in the survival rate for kidney transplants
(Schreiber, 1991; Walsh et al., 1992).
Cyclosporin is an 11-membered cyclic peptide with the methyl amide
between residues 9 and 10 in the cis position; all the other methyl amides are
trans. It was originally discovered as a metabolite of the fungus Tolypocladium
injlatun Cams and had weak antifungal activity. Cyclosporin interferes with

C. /CH3 H
CH CH3
I C&CH3 HO, ,Cq I
CH2 7H3 7” CH3 CH CH3 CH2 CH3
I I I I I
CH3 -N-CH-CO-N- CH-;-N-CH-CO-N-CH-;-N-CH,
I L L L
CH3 co 0 L L 0 I
\ I co
CH-CH2-CH,
I
N-CH3
CL3 CH,4 0
‘i’ II L 7 LI
OL :H -N-CO-:H-N-CO-kH-N-C-CH-N-CO-CH
-1 I I I
CH3 A ;H, CH2 CH3 CH fH2
I C&‘CH,
CH CH
C;;, ‘CH, C;;J ‘CH,

Cyclosporin

the activation of T lymphocytes. T cells need activation by an antigen or cyto-

kines before they can destroy virally infected cells and stimulate macrophages
to kill bacteria and aberrant cells. Activation is a complex series of changes
including the synthesis of interleukins, principally interleukin 2, and other
cytokines that induce proliferation and differentiation. Cyclosporin blocks
interleukin 2 gene expression in a T cell after antigen presentation (Kronke
et al., 1984; reviewed in Walsh et al., 1992).
Cyclosporin binds to a cytoplasmic receptor known as a cyclophilin, found
in many tissues and eukaryotic cells, catalysing the folding of proteins. Cyclo-
philin has the enzyme activity of peptidyl-proline cis-trans isomerase. Cyclo-
philins may help newly formed proteins to fold, particularly as proline cannot
participate in an a-helix and is often found at a turning point in protein struc-
ture. Cyclosporin initially binds in the cis form found in solution
(KD = 4 X 10e6~) but isomerizes in the complex into the trans form
(& = 4 X ~O-‘M) thus mimicking the normal substrates of cyclophilin. This
activity, however, is not directly responsible for the immunosuppressive action
because (a) most of the cyclophilin in the cell is not bound by cyclosporin
when T cell activation is inhibited, and (b) analogues of cyclosporin have been
 Membrane-active agents 243

synthesized which have cyclophilin binding activity but no immunosuppress-

ive action (Liu et al., 1992).
The active agent appears to be the drug-protein complex. The intracellular
target is even less obvious, being the calcium-activated protein, calcineurin
(Clipstone and Crabtree, 1992). Calcineurin is a phosphatase that removes
phosphate groups from serine and threonine. It is a heterodimer with subunits
61 and 19 kDa; the smaller subunit binds calcium with similar calcium-binding
sites to calmodulin. Calcineurin was originally discovered in mammalian brain,
hence the name, where it regulates voltage-activated calcium channels by
dephosphorylating a membrane-associated protein and inactivating the chan-
nel (Armstrong, 1989). It is not yet clear how calcineurin interferes with T
cell activation, but cyclosporin has already greatly increased our molecular
understanding of the immune process.

10.8 Potassium channel opening

The discovery of new information by novel drugs has shown up again in the
field of potassium channels. Nicorandil and cromokalin (or its optically active
isomer, lemakalin) open potassium channels and induce smooth muscle relax-
ation. This has led to potential use in hypertension, asthma and incontinence.
Relaxation of arterial smooth muscle leads to anti-hypertensive effect and ben-
eficial reduction in plasma lipids and effects on coronary heart disease. The
activity resides in the (-)-isomers of these drugs, disparate though they are
in structure (L.ongman and Hamilton, 1992).

C-NH-CH*-CH*-O-NO2
II

Nicorandil

Lemakalim

Although there is little mechanistic information available at the molecular

level, the inhibition of potassium efflux from the cell results in hyperpolariz-
244 Molecular mechanisms of drug action

ation. Calcium does not enter, is not released from intracellular stores, and the
tissue relaxes. The hyperpolarization can reach as high as -90 mV which is
the equilibrium level of potassium. Once this level is reached no further loss
of potassium can occur.
ATP regulates the calcium channel; at higher levels of ATP the channel is
more difficult to open. As the ATP level falls, the channel opens more easily
and consequently the drugs become more effective. Glibenclamide, the anti-
diabetic drug (Chapter 10) which is known to block ATP-sensitive K+ chan-
nels, antagonizes these effects.
Nicorandil is unusual by having two mechanisms of action. As well as potass-
ium channel activation, nicorandil is a nitrate and induces relaxation by stimul-
ating cyclic GMP formation (Kukovetz et al., 1992).

IO.9 Sterol ligands - polyene antibiotics

Fungal infections have become progressively a more serious problem in recent

years. This has been attributed partly to our mastery of bacterial infections,
leaving a vacuum that pathogenic fungi take the opportunity to fill, and partly
to the fact that a large number of hospital patients - particularly those with
cancer and transplant patients - have a partially suppressed immune system
as a consequence of drug treatment. Thus there has been a great increase in
the number of systemic infections by fungi, particularly opportunist infections
such as thrush, caused by Candida albicans, and they now present a serious
medical problem. It has been known for a long time, however, that Candida
infections can be lethal. Gooday (1977) notes that a paper of 1665 records at
the height of the plague: ‘The diseases and casualties this week, London, Sept.
12 to the 19; Consumption, 129 . . . Fever, 332 . . Plague, 6544 . . Thrush, 6’.
For nearly 30 years, systemic fungal infections have been treated with
amphotericin B as the drug of choice (indeed for much of that period the only
drug) because it is very active against a wide range of pathogenic organisms,
is fungicidal not fungistatic and does not induce resistance (reviewed in Medoff
et al., 1983). The drug’s major drawback, and it is a serious one, is kidney
toxicity; in about 80 per cent of the patients treated, the glomerular filtration
rate falls to about 40 per cent of normal. The effect is usually reversible when
treatment is withdrawn, but in some cases there is some residual impairment
of kidney function. Amphotericin B, however, was for many years the only
polyene antibiotic sufficiently safe to be used systemically in man. Nystatin, a
close analogue, may be given orally for Candida infections of the mouth or
intestinal tract, but since it is almost completely non-absorbed it is not used
for systemic infections.
Polyene antibiotics are natural products produced by the Actinomycetes
(notably the Streptomycetes) and are characterized by a macrolide ring closed
by an internal ester or lactone. As their name implies, polyenes have a system
of conjugated double bonds (normally between 4 and 7) in the tram contigur-
 Membrane-active agents 245

ation, which gives rise to a characteristic ultraviolet absorption spectrum with

a number of peaks in the visible and near ultraviolet. Amphotericin B is a
heptene (seven double bonds) and, additionally, contains an amino sugar
(mycosamine) attached to the ring. The ring size can range from 12 to 37
carbon atoms and is substituted by a number of hydroxyl groups. The combi-
nation of several polar hydroxyl groups and a carboxyl, on the one hand, and
several hydrophobic double bonds on the other, gives a dual character to the
molecule, particularly as these groups reside on opposite sides of the ring
(Norman et al., 1976).

10.9.1 Amphotericin action

The specific target of the polyenes is the sterol in the plasma membrane of
the eukaryotic cell; ergosterol in fungi and cholesterol in protozoal and mam-
malian cells. The evidence for this may be summarized as follows:
1. Bacteria do not contain sterol and are resistant to amphotericin. Protozoa,
fungi and mammalian cells are all sensitive. The mycoplasma, Acholeplasma
laidzawii, can grow whether sterol is present in the culture medium or not
and incorporates it into the membrane if it is present. The microorganism
is sensitive to the polyene only when sterol is present in the membrane.
2. Free sterols in the culture medium antagonize the effect of the drug on
microorganisms, such as Saccharomyces cerevisiae and Candida albicans.
3. A complex between amphotericin and sterols which can be demonstrated
by various techniques, such as ultraviolet difference spectroscopy, e.s.r. and
circular dichroism, is suft%ciently tightly bound to explain the antimi-
crobial action.
4. Fungal cells die as a result of damage to the membrane where the drug-
sterol complexes are found. The membrane begins to leak small ions,
notably ammonium and potassium.
The binding of polyenes to sterol can be demonstrated in model membrane
systems such as vesicles composed of sterol and phosphatidylcholine. The
binding constant for the amphotericin B-vesicle complex is 9.7 X ~O-‘M
when cholesterol is used and 1.52 X lop6111 with ergosterol. Although this
suggests that the drug might interfere more with animal cell membranes, the
number of binding sites is important - ergosterol containing vesicles have a
larger number of binding sites for the drug than those made with cholesterol,
giving a tenfold factor in favour of the former in terms of product of binding
sites and binding constant (Witzke and Bittman, 1984). Furthermore, circular
dichroism studies suggest that there is more than one conformer of the drug
present in membrane vesicles at higher drug concentrations but one confor-
mer predominates at lower concentrations whichever sterol is present (Vertut-
Croquin et al., 1983).
The fact that cholesterol binds polyenes tightly (indeed amphotericin is very
unusual in binding more effectively to ergosterol) probably explains why these
246 Molecular mechanisms of drug action

compounds are toxic to mammals. There is no direct correlation between anti-

Candida activity and binding affinity in model vesicles for a small group of
four polyenes including amphotericin B, but the group may be too small to
expect correlations.

X h
~~ ‘*OH

CH3 (0) 3
Amphotericin CH3 H H
HO

Alternatively, the nature of the phospholipid in the vesicles may be important

as phosphatidylcholine faces the extracellular side of the plasma membrane
in that its polar head groups are pointing in that direction while phosphatidyl-
ethanolamine and phosphatidylinositol face the cytoplasm, and so a different
composition in the vesicles might give a better correlation. It is also possible
that the leakiness induced by binding may not be directly related to the extent
of binding, but rather to the particular complexes made by the polyenes and
how they disrupt the membrane architecture (Witzke and Bittman, 1984).
A number of features on the sterol nucleus are essential for the interaction
to take place, namely a 3Phydroxyl group and a A-22 double bond together
with a cholestane skeleton. An intact polyene is also required - if the ring is
broken no interaction will take place (Norman et al., 1976).
Membrane fluidity plays a key role in the action of the polyenes. They can
induce permeability in the ‘solid’ or crystalline form of the membrane in the
absence of cholesterol and above the transition temperature when the mem-
brane is in the liquid form, amphotericin can still have a small effect, but the
addition of cholesterol markedly increases it (reviewed in Medoff et al., 1983).
The precise molecular nature of the interaction of polyenes with membrane
components, and how this induces permeability, is still obscure even for model
membranes. Electron microscope pictures of erythrocyte membranes treated
with amphotericin show a series of pits and protuberances but no obvious
pores, and so it is difficult to accept the original view that the polyene-sterol
complex forms aqueous channels or pores in the membrane through which
the ions pass. An alternative hypothesis is that the complex may alter the local
membrane fluidity by removing the sterol from its interaction with membrane
phospholipids, and thereby induce leakage of cell constituents. Particularly
 Membrane-active agents 247

sensitive points may occur where there is a change in the membrane from
solid to liquid phase (reviewed in Medoff et al., 1983). At present the situation
is still unresolved.
Amphotericin at low concentrations in vivo can stimulate cell growth -
indeed macrophages are stimulated to kill microorganisms, and the drug can
markedly increase the number of antibody-forming cells in the mouse spleen
and lymph node. Not surprisingly, therefore, the drug has a stimulatory effect
on the immune system in vivo, but the clinical relevance of this fmding is as
yet unknown. In view of the immune suppression of many of the patients
with fungal infections this property could be very useful.
In conclusion, therefore, amphotericin B is still the drug of choice for life-
threatening fungal infections by primary pathogens such as Histoplasma cap-
sulatum and Paracoccidioides brasiliensis as well as opportunist pathogens
such as Candida albicans. The problem of the toxicity of amphotericin may
be countered by the concomitant use of flucytosine which can allow the dose
of amphotericin to be reduced (Lyman and Walsh, 1992). Therapy with ampho-
tericin, however, is not always straightforward because, in addition to toxicity,
there can be a lack of clinical response even when the organism is susceptible
to the drug in vitro. In some cases this is because the drug is unable to pen-
etrate in sufficient quantity the particular part of the body which harbours the
pathogen. This is the case in infections of the central nervous system with
Coccidioides species which cause meningitis (Medoff et al., 1983).

Cholesterol Ergosterol

Questions

1. Name two ions that pass through channels in the mammalian cell mem-
brane without requiring the energy of ATP. Name two drugs that interfere
with the passage of these ions. What are the drugs used for?
2. Why is the energy of ATP required to drive ion pumps? Name two classes
of drug that block ion pumps. How do they do this?
3. Explain these terms (a) depolarization (b) action potential (c) arrhythmia.
4. How do local anaesthetics interfere with the sodium channel?
5. How do the sodium and calcium channels differ in structure?
6. What classes of calcium channel blockers are available? How does their
mechanism of action differ?
248 Molecular mechanisms of drug action

7. For what cellular processes is calcium required?

8. Name one major difference between the fungal and mammalian cell mem-
branes. How is this exploited for the treatment of fungal infection?
9. How might amphotericin be packaged to make it less toxic?
10. On what test does cromoglycate show the greatest activity? How does this
relate to the pathogenesis of asthma?

References
Armstrong, D.L., 1989, Trenak in Neurosciences, 12, 117-21.
Barnes, PJ., 1992,J. Znnt.Med., 231, 453-61.
Benos, D.J., Cunnin gham, S., Baker, R.R., Beason, K.B., Oh, Y. and Smith, P.R., 1992,
Rev. Pbysiol. Biochem. Pbarmacol., 120, 31- 113.
Bigger, J.T. and Hoffman, B.F., 1990, Ch. 35 in the Pharmacological Basis of Thera-
peutics, 8th edn (eds A.G. Gilman, T.W. Rail, AS. Nies and P. Taylor). New York:
Pergamon Press.
Brown, L. and Erdmann, E., 1984, Basic Res. Cardiol., (Suppl.), 50-5.
Burg, M. and Green, N., 1973, Kidney Znt., 4, 301-308.
Caffruny, EJ. and Itskovitz, H.D., 1982,J. Pbarmacol. Exp. Ther., 223, 105-109.
CAST investigators, 1989, N. Eng1.J Med., 321, 406-12.
CatteraIl, W.A., 1980, Annu. Rev. Pbarmacol. Toxicol., 20, 14-43.
Catterall, W.A., 1988, Science, 242, 50-61.
Clipstone, N.A. and Crabtree, G.R., 1992, Nature, 357, 695-7.
Coulson, C.J., Ford, R.E., Marshall, S., Walker, J.L., Wooldridge, K.R.H., Bowden, K. and
Coombs, T.J., 1977, Nature, 265, 545-7.
Crazier, I. and Ikram, H., 1992, Dncgs, 43, 637-50.
Cuthbert, A.W., 1977, Progr. Med. Cbem., 14, l-50.
Droogmans, G., Himpens, B. and Casteels, R., 1985, Experientia, 41, 895-900.
Fleckenstein-Grun, G., Frey, M. and Fleckenstein, A., 1984, Trend-s Pbarmacol. Sci., 5,
283-6.
Frizzell, R.A., Field, M. and Schultz, S.G., 1979, Am.J Pbysiol., 236, Fl-8.
Glossman, H., Ferry, B.R., Lubbecke, F., Mewes, R. and Hoffman, F., 1982, Trends Pbar-
macol. Sci., 3, 431-7.
Gooday, G.W., 1977, J. Gen. Microbial., 93, l- 11.
Godfraind, T., 1984, Eur. HeartJ, 5, Suppl. F, 303-308.
Grant, A.O., 1992, Amer. HeartJ, 123, 1130-6.
Grant, A.O. and Wendt, D.J., 1992, Trends in Pbarmacol. Sci., 13, 352-8.
Hansen, O., 1984, Pbarmacol. Rev., 36, 143-63.
HiRe, B., 1977, J. Gen. Pbysiol., 69, 497-515.
Holgate, S.T., 1989, Resp. Med., 83 (Suppl), 25-31.
Hondeghem, L.M. and Katzung, B.G., 1984, Annu. Rev. Pbarmacol. Toxicol., 24,
387-423.
Howarth, P.H., Durham, S.R., Lee, T.H., Kay, A.B., Church, M.K. and Holgate, S.T.,
1985, Am. Rev. Respir. Dis., 132, 986.
Jorgensen, P.L., 1982, Biocbem. Biopbys. Acta, 694, 27-68.
Katz, A.M., 1992, Amer. J. Cardiol., 69, 17E-22E.
Kauffman, G.B., 1984, Cbem. Educ., 21, 42-5.
Kay, A.B., Walsh, G.M., Moqbel, R., McDonald, AJ., Nagakura, A., Carroll, M.P. and
Richerson, H.B., 1987,J. Allergy Clin. Zmmunol., SO, l-8.
Kendall, MJ. and Horton, R.C., 1985,J. Royal ColC. Pbys. Land., 19, 85-9.
Kokko, J.P., 1984, Am.J Med., 77, 5A, 11-17.
 Membrane-active agents 249

Kronke, M., Leonard, WJ., Depper, J.M., Arya, SK., Wong-Staal, F., Gallo, R.C., Wald-
mann, T.A. and Greene, W.C., 1984, Proc. Nat. Acad. Sci. (USA.), 81, 5214-18.
Kukovetz, W.R., Holzmann, S. and Poch, G., 1992, J Cardiovasc. Pbarmacol., 20,
Suppl 3, s3-s7.
Kumana, C. and Hamer, J., 1979, In: Drugs for Heart Disease, Ed. J. Hamer (London:
Chapman and Hall), p. 61.
Kuzemko, J.A., 1989, Respir. Med., 83, (Suppl), 1 l- 16.
Kwon, Y.W. and Triggle, D.J., 1991, Cbirality, 3, 393-404.
Lammers, J.-W.J., Barnes, P.J. and Chung, K.F., 1992, Eur. J. Respir., 5, 239-46.
Lant, A., 1985a, Drugs, 29, 57-87.
Lant, A., 1985b, Drugs, 29, 162-88.
Liu, J., Albers, M.W., Wandless, T.J., Luan, S., Alberg, D.G., Belshaw, PJ., Cohen, P.,
Mackintosh, C., Klee, C.B. and Schreiber, S.L., 1992, Biocbemistly, 31, 3896-901.
Longman, S. and Hamilton, T.C., 1992, Med. Res. Rev., 12, 73-148.
Ludens, J.W., 1982,J Pharmacol. Exp. i%er., 223, 25-9.
Lyman, C.A. and Walsh, TJ., 1992, Drugs, 44, 9-35.
Lytton, J., 1985, Biochem. Biophys. Res. Commun., 132, 764-9.
Mazurek, N., Berger, G. and Pecht, I., 1980, Nature, 286, 722-3.
McTavish, D., Buckley, M.M. and Heel, R.C., 1991, Drugs, 42, 138-70.
Medoff, G., Brajtburg, J., Kobayashi, G.S. and Bolard, J., 1983, Annu. Rev. Pharmacol.
Toxicol., 23, 303-30.
Moncada, S., Palmer, R.M.J. and Higgs, E.A., 1991, Pharmacol. Revs., 43, 109-42.
Moqbel, R., Cromwell, 0.) Walsh, G.M., Wardlaw, A.J., Kurlak, A.K. and Kay, A.B., 1988,
Allergy, 43, 268-76.
Muhiddin, K.A. and Turner, P., 1985, Postgrad. Med.., 61, 665-78.
Norman, A.W., Spielvogel, A.M. and Wong, R.G., 1976, Adv. Lipid. Res., 14, 127-70.
Odlind, B., 1984, Acta Pbarmacol. Toxicol., 54, Suppl. 1, 5- 15.
Olbe, L., Lind, T., Cederberg, C. and Ekenved, G., 1986, Scand. J. Gastroenterol., 21,
S118, 105-7.
Patarca, R., Candia, O.A. and Reinach, P.S., 1983, Am. J. Pbysiol., 245, F660-9.
Pedersen, P., 1982, Ann. N.Y. Acad. Sci., Ml, l-20.
Postma, S.W. and Catterall, W.A., 1984, Molec. Pharmacol., 25, 219-27..
Ritchie, J.M. and Green, N.M., 1985, In: The Pharmacological Basis of Therapeutics,
7th Ed, eds A.G. Gilman, L.S. Goodman, T.W. Rail and F. Murad (New York:
Macmillan) Ch. 15.
Sachs, G., 1986, Scand. J Gastroenterol, 21, l-10.
Sachs, G. and Wallmark, B., 1989, Stand. J Gastroenterol., 24, (Suppl 166), 3-11.
Scarborough, G.A., 1985, Microbial. Rev., 49, 214-31.
Schramm, M. and Towart, R., 1985, Life Sci., 37, 1843-60.
Schreiber, S.L., 1991, Science, 251, 283-7.
Serrano, R., Kielland-Brandt, M.C. and Fink, G.R., 1986, Nature, 319, 689-93.
Singer, S.J. and Nicolson, G.I., 1972, Science, 175, 720-31.
Spedding, M., 1985, Trends Pbarmacol. Sci., 6, 109- 14.
Spedding, M. and Pauletti, R., 1992, Pbarmacol. Rev., 44, 363-76.
Striessnig, J., Glossmann, H. and Catterall, W.A., 1991, Proc. Natl. Acad Sci. (US.A.),
87, 9108-102.
Sweadner, K.J., 1979,J. Biol. Cbem., 254, 6060-67.
Vertut-Croquin, A., Bolard, J., Chabbert, M. and Gary-Bobo, C., 1983, J Biol. Cbem.,
22, 2939-44.
Wallmark, B., 1989, Stand. j. Gastroenterol., 24, (Suppl 166) 12-18.
Walsh, C.T., Zydowsky, L.D. and McKeon, F.D., 1992,J. Biol. Cbem., 267, 13115-18.
Wit, A.L., 1985, Clin. Pbysiol. Biocbem., 3, 127-34.
Witzke, N.M. and Bittman, R., 1984, Biochemistry, 23, 1668-74.
 Chapter 11

Microtubule assembly

Il. I Introduction
Microtubules are cytoplasmic organelles which play a crucial role in the pro-
cess of cell division (mitosis) in eukaryotic cells as part of the mitotic spindle.
Furthermore, various other activities including intracellular transport and
secretion are mediated by these organelles. Indeed, the secretion of cytoplas-
mic granules of hormonal mediators of the asthmatic process (see section 10.6)
also requires the involvement of microtubules (Kaliner, 1977). Interestingly,
the cilia of the protozoon Tetruhymena pyri~ormis, that the organism uses
for swimming, are mainly composed of microtubules.
During the stage of mitosis known as prophase the chromosome doublets
appear, linked by a body known as a centromere. The two centrioles, normally
located near the nucleus, migrate to opposite ends of the cell. Around each
centriole a system of radiating fibres, known as the aster, develops. In the next
phase, metaphase, the centrioles organize microtubules into the mitotic
spindle which links the centrioles. The centromeres become attached to a
spindle fibre and migrate to the median position between the poles. The loose
ends of the chromosomes are orientated in a random fashion but the centro-
meres of each chromosome lie exactly in a plane called the equatorial plate.
In anaphase which follows, the centromeres duplicate themselves. The two
resulting centromeres move apart towards the poles, still attached to the
spindle fibre, with their attached chromosome ‘dragging behind’. Sub-
sequently, in telophase a nuclear membrane begins to encircle the uncoiling
chromosomes, while a furrow appears at the equator, eventually deepening
and splitting the daughter cells apart.
The spindle is composed of a number of microtubules which have the gen-
eral structure of long hollow cylinders with inside and outside diameters of
15 and 25 nm, respectively. Microtubules are composed of the protein tubulin
which itself is constructed of two closely related subunits (40 per cent
sequence homology) with molecular weights in the region of 50 kDa. The
functional form of tubulin is an (Y/P heterodimer of sedimentation coefficient
(szO,J of 5.8s (reviewed in Correia, 1991).
Tubulin from brain, the most prolific source, can be assembled in vitro into
microtubules by raising the temperature from 4 to 37°C. The polymerization

251
252 Molecular mechanisms of drug action

may be measured by an increase in turbidity. Usual requirements are: (a) the

presence of a family of associated proteins, known not surprisingly as micro-
tubule-associated proteins (or MAPS) which assist the process in a way that is
not understood, (b) GTP which is hydrolysed to GDP and (c) magnesium ion.
Tubulin will assemble in the absence of MAPS but usually requires a high mag-
nesium concentration or the presence of a polycation. Calcium, on the other
hand, causes the microtubules to depolymerize, probably by binding to the
wall of the microtubule and inducing an increase in the rate of subunit dis-
sociation from the ends of the microtubules (Weisenberg and Decry, 1981).
Reduction of the temperature to 10°C also causes microtubules to disassemble,
and repeated cycles of assembly and disassembly are usually used to purify
tubulin.
Tubulin contains two GTP binding sites per dimer. One of these sites, the
E site, will rapidly exchange GTP for GDP in the presence of magnesium ion,
undergoing a conformational change in the process. GTP and GDP stab&e
different conformations of tubulin. GTP bound at the E site is hydrolysed when
microtubules are formed and is linked to the addition of a tubulin dimer to
the growing microtubule. The other site (N site) is not involved in catalysis
and is always occupied by non-exchangeable GTP.
In the absence of MAPS, the process of assembly requires a sufficiently large
polymeric aggregate before assembly will begin. A threshold concentration of
tubulin is required for this stage. Once the reaction is initiated, 543s dimers add
sequentially until the polymer reaches its fmal length within lo-20 minutes at
37°C. The process terminates if the tubulin dimer level falls below the critical
threshold value. Nevertheless, the process is dynamic with tubulin dimers join-
ing and leaving the polymer - provided there is sufficient GTP available. In
vivo, the polymer is in the form of a cylinder 30 nm in diameter composed
of 13 rods lying parallel to each other to form a cylindrical sheet (Correia,
1991). Growth or elongation occurs at both ends of the cylinder albeit at
different rates. The faster rate is at the plus end while the minus end is
anchored to a microtubule-organizing centre such as the centrosome. Subunit
disassembly occurs through loss of dimers when GDP bind at the E site (Pig. 11.1).
Various drugs interfere with the process of assembly. In general, they all
have similar actions in vitro in that they block mitosis. Their uses, however,
differ widely in a medical sense from antifungal agents like griseofulvin, anthel-
minthic drugs such as the benzimidazole family, to the vinca alkaloids, vincris-
tine and vinblastine, which are used as anti-tumour agents. Colchicine, one of
the best known of the microtubule ligands is used for the treatment of gout,
but there is some doubt as to whether its therapeutic efficacy relies only, if
at all, on inhibition of microtubule polymerization.

11.2 Colchicine for gout

Colchicine is an alkaloid derived from the autumn crocus (Colcbicum
autumnale) and other Colchicum species that grow in Colchis in Asia Minor
 Microtubule assembly 253

t-1or (+I (4)or t-1

end end

- /;,TP-fvlg
GTP-Mg [I
0
HPO,Mg

Figure 11.1 An idealized mechanism of microtubule assembly/disassembly. GTP-tubulin hetero-

dimers with GTP at the E-site are represented by dark ovals. GDP-tubulin heterodimers are rep
resented by light ovals. Microtubule growth occurs at both ends in vitro. In viva the slow growing
end, the minus end, may be anchored at MTOCs. GTP hydrolysis occurs upon assembly and, after
the release of HPO,Mg, generates a microtubule that is stabilized by a GTP cap, a layer of GTP-
tubulin subunits that prevent catastrophic disassembly. Disassembly may occur by the loss of
oligomers. Subunits may also form nonmicrotubule polymers, typically double walled rings.
(Correia, 1991)

and it is probably one of the most studied of the microtubule ligands. The
precise site at which the drug binds has yet to be determined and both sub-
units have been implicated by genetic data. This could either mean a site at
the interface between the (Y and /3 subunits or that changes in one subunit
can alter the conformation of the other (Correia, 1991; Levy et al., 1991).
The drug binds to the protein in a time-dependent fashion which resembles
an irreversible process in that a conformational change takes place in the pro-
tein. The colchicine bound to the tubulin is, consequently, not in equilibrium
with free drug, i.e. it is non-exchangeable. The process is not irreversible in
practice because colchicine gradually dissociates from the complex with a half-
life at 37°C of approximately 12 hours (Wilson, 1970). Estimates of the binding
constant vary between 0.1 to 3.0 X 10P6~, depending on how much time has
been allowed for equilibration. The reason for this apparent irreversibility is
the large activation energy required for binding (E = 100.4 kJ/mol) because of
a major conformational change - hence the dissociation is also very slow.
Colchicine has three rings denoted A, B and C, and all three play a part
in binding. The trimethoxybenzyl A ring interacts hydrophobically and with
hydrogen bonding, while the tropolone C ring provides stacking interactions.
The slow kinetics of the irreversible colchicine binding requires distortion of
the B ring. Removing the B ring increases the reaction rate and lowers the
activation barrier (Correia, 1991).
The drug inhibits the assembly of microtubules by binding to soluble tubu-
lin, forming a complex that adds to microtubule ends and sharply reduces the
affinity of the microtubule for fresh tubulin dimers. The tubulin addition rate
254 Molecular mechanisms of drug action

to the growing microtubule is thereby greatly reduced (Margolis et al., 1980).

Colchicine acts, in fact, as a type of chain terminator.
Furthermore, colchicine stimulates the GTPase activity of tubulin, perhaps
as a result of drug binding to a site which is responsible for inducing poly-
merization-dependent GTPase activity. Colchicine, therefore, induces a confor-
mational change in the absence of microtubule assembly that mimics assembly
by leading to increased GTPase activity. Vinblastine, an indole alkaloid, inhibits
this activation of GTPase activity by colchicine (David-Pfeuty et al., 1978).
Colchicine is used in the treatment of acute gouty arthritis to reduce pain
and inflammation. The swellings on joints and tendon sheaths composed
mainly of sodium m-ate crystals (tophi), characteristic of gout, are reduced in
size and eventually disappear. Polymorphonuclear leukocytes ingest the crys-
tals and produce a glycoprotein which acts as a chemotactic factor and may
be the cause of gouty arthritis by recruiting granulocytes into the inflamed
area which then release hydrolytic enzymes. Colchicine suppresses the release
of this factor. The role of microtubules in this is not clear, although colchicine
interferes with the microtubules in the neutrophil and disassociates microtu-
bules and the mitotic spindle. One problem is the activity of trimethylcolchi-
cinic acid, which is almost as active as colchicine itself against gout but is no
inhibitor of microtubule assembly (Wallace, 1975).
Trimethylcolchicinic acid could be converted in uiuo to colchicine (or an
analogue which can bind to microtubules). This could give rise to low levels
of colchicine in the circulation, which are as high as is needed for colchicine
action in uiuo (namely 5 X ~O-‘M) but still difficult to detect (Wallace, 1975).
Although this concentration is below the apparent binding constant for tubu-
lin-colchicine complex formation, the existence of a pseudo-irreversible com-
plex suggests that time-dependent inactivation could occur. Low levels of the
drug would then have the desired effect. This biotransformation has not been
confirmed, however, and the question remains unresolved.

OCH3 OH

Colchicine Trimethylcolchicinic acid

A second difficulty arises from the fact that colchicine is only a weak anti-
inflammatory agent in conditions induced by substances other than urate crys-
tals, and so it is possible that there is a particular step in crystal-induced inflam-
mation which is more important for that type of inflammation than it is for
any other. The release of a neutrophilderived chemotactic factor has been
proposed in this context (Spilberg et al., 1979). Whether microtubules can be
 Microtubule assembly 255

implicated in this secretory process is not entirely clear, although they are
known to be involved in a number of other secretory processes such as the
release of histamine from human lung tissue after antigen challenge (Kaliner,
1977). In conclusion, inhibition of microtubule assembly remains the only
plausible explanation for the mode of action of colchicine in the absence of
any other convincing hypothesis (Levy et al., 1991).

11.3 Vinca alkaloids as anti-tumour agents

Vincristine and vinblastine are two members of the family of indole alkaloids
derived from the periwinkle (Catharanthus or Vinca rosea - hence the name
vinca alkaloids). These drugs bind to tubulin in a different fashion and at a
different site for colchicine. The considerable difficulties, however, in measur-
ing binding constants accurately in this system have been noted by Correia
(1991).
There are two binding sites for vinblastine per tubulin dimer, instead of one
for colchicine. Moreover, vinblastine will bind instantaneously and reversibly,
unlike colchicine, to binding sites on the microtubule, in the absence of sol-
uble dimer, with a binding constant of 1.9 X 10e6~. The number of high-
affinity binding sites that is available on the microtubule, however, is far less
than the total number possible if one site on each individual subunit were
freely accessible (Wilson et al., 1982).
Inhibition of de novo polymerization has yielded ZC,, of 4.3, and 3.2 X lo-’ M
for vinblastine and vincristine respectively (Owellen et al., 1976). If the exper-
iment is carried out by measuring the inhibition of addition of tubulin dimers
to microtubules at steady state, known as ‘freewheeling’, in contrast to poly-
merization de rzovo, a figure of 1.38 X lo-‘M for half-maximal inhibition is
obtained for vinblastine (Wilson et al., 1982). Under these ‘freewheeling’ con-
ditions, 1.16 molecules of vinblastine are detected per microtubule, implying
that the addition of one vinblastine molecule at the assembly end of a microtu-
bule is enough to have a very marked effect on the polymerlzation process.
This is referred to as ‘substoichiometric poisoning’ since the concentration of
drug is well below that of the tubulin (although greater than that of the
microtubule ends). No effect on depolymerization is noted at these concen-
trations.
At higher concentrations (i.e. > lop6 M), the drug binds to a greater number
of sites on the microtubule and appears to loosen its structure. As a conse-
quence, splaying and peeling of protofilaments at microtubule ends occurs
together with active depolymerization (Wilson et al., 1982).
Since freewheeling is more sensitive to drug inhibition than is de nova poly-
merization, the ability of four vinca alkaloids to inhibit freewheeling has been
compared with their ability to inhibit cell proliferation in tumour cell cultures
in vitro. Although all four drugs were very potent in both tests, no correlation
was found between orders of potency on the two tests (Jordan et al., 1985).
256 Molecular mechanisms of drug action

It was suggested that this may be because the relative potency of the alkaloids
may vary with respect to tumour type studied. Alternatively, it may be sug-
gested that four compounds are hardly enough on which to base firm con-
clusions. In the cell, vinblastine blocks mitosis and causes metaphase arrest.

COOCHB

Vinblastine: R = CH3
Vincristine: R = CH0

The drug binds to microtubules causing disassociation and disruption of the

mitotic apparatus. Chromosomes cannot be segregated correctly, indeed they
aggregate in unusual groupings such as balls or stars, and the cell dies
(Calabresi and Chabner, 1990).
Vinblastine and vincristine are used normally in combination with other anti-
tumour agents for the treatment of Hodgkin’s and non-Hodgkin’s lymphomas,
breast carcinoma, cancers of the head and neck and others. One curious fea-
ture is that cross-resistance between the vinca alkaloids is not observed (see
Calabresi and Chabner, 1990).

11.4 Griseofulvin as an antifungal agent

Griseofulvin is a secondary metabolite elaborated by the fungus, Penicillium

griseofidvium dierckx. The action of griseofulvin is to arrest the target fungal
cell in metaphase (Gull and Trinci, 1973). As a consequence, multinucleate
hyphal cells are formed, frequently together with a Y-shaped metaphase equa-
torial plate. The microtubules maintain their normal morphological appearance
but are disorientated within the spindle (Grisham et al., 1973). The overall
effect is thus antimitotic, although the site of action differs from that of colchi-
tine. Griseofulvin does not prevent the binding of colchicine to tubulin nor
does it affect the ability of the vinca alkaloids to stabilize the binding of colchi-
tine to tubulin (Wilson, 1970).
Originally it was thought that the drug bound to the MAPS (Roobol et al.,
1977), thereby inhibiting microtubule assembly both in rate and extent
(Roobol et al., 1976). More recently, however, radioactive griseofulvin has
been found to bind directly to tubulin dimer (0433 mol/mol). Initiation of poly-
merization requires microtubule-associated proteins but subsequent elongation
 Microtubule assembly 257

does not, and griseofulvin blocked this later stage. Moreover, griseofulvin also
induces depolymerization of preformed microtubules at 37°C (reviewed in
Kerridge, 1986).

Cl CH3
Griseofulvin

Keates (1981) has drawn attention to the need to carry out studies in vitro
with griseofulvin at low concentrations since the solubility limit is 3 X 10e5 M in
aqueous buffers. Attempts to introduce higher levels into solution may induce
artifacts by precipitating proteins out of solution. Keates finds that griseofulvin
inhibits both polymerization and depolymerization with an ZC,, of
1.25 X 10e5M, but the equilibrium position is unchanged.
Griseofulvin is fungistatic in vitro for a number of fungi that cause skin
infections (dermatophytes, such as Tricbophyton, Micromonospora species)
and has been used for many years to treat infections such as athlete’s foot
(caused by Trichophyton mentagropbytes). The drug owes its efficacy partly
to the fact that it is concentrated in the skin in keratin precursor cells, binding
particularly to keratin so that the fungus is unable to gain a foothold on new
hair and nails (Gull and Trinci, 1973). Secondly, many dermatophytes concen-
trate the drug particularly if they are sensitive to it (El-Nakeeb and Lampen,
1965).

11.5 Benzimidazoles as antbelminthics

The benzimidazole family of drugs has long been used to treat helminthic
infections of sheep, cattle, goats etc. Various nematodes infect man and meb-
endazole is considered to be front-line treatment for several infections includ-
ing hookworm (Ancylostoma duodenale), roundworm (Ascaris lumbricoides)
and whipworm (Tricburis trichiura). Mebendazole is reasonably effective and
non-toxic - a very important consideration in anti-parasitic chemotherapy,
where many of the drugs available are toxic (Webster, 1990). Most of the
experimental work has been done with nematodes that do not infect man.
Several benzimidazoles inhibit the assembly of microtubules from a nema-
tode, Ascaridia galli, which infects chickens, with ZC,, of 5 X lo-“M and
6 X 1O-6~ for mebendazole and oxfendazole respectively. Some but not all
of the drugs bind to mammalian tubulin at similar concentrations, although
thiabendazole and oxfendazole had ZC,,, values in excess of 10m4~. Electron
microscopy of microtubules, polymerized under benzimidazole inhibition,
258 Molecular mechanisms of drug action

showed a reduction in both the number and the length of microtubules formed
(Dawson et al., 1984). Correlation of in vitro with in vivo results is not poss-
ible for all the compounds, perhaps because of pharmacokinetic and/or meta-
bolic considerations. Nevertheless, evidence from benzimidazole-resistant hel-

Rl R2

Mebendazole: - NHCOOCHJ

Oxfendazole: - NHCOOCH,

Thiabendazole: Ii

0
0IS N’

Flubendazole: F ;- - NHCOOCHJ

minths, such as Tricbostrongylus colubrirormis, suggests that tubulin has been

altered to a form which shows less drug affinity. Furthermore, resistant organ-
isms show no change in the appearance of their cytoplasmic microtubules
when treated with thiabendazole, unlike intestinal cells from the drug-sensitive
organisms, that lose almost all their organelles (Sangster et al., 1985).
Antagonism of colchicine binding can also be used as an index of microtu-
bule ligand affinity for the benzimidazole drugs because they bind at the colchi-
tine binding site, despite their obvious difference in structure. Mebendazole
has been shown to interfere with colchicine for the binding site on tubulin
from bovine brain with an inhibitor constant of 3.6 X lo-” M. One binding site
per tubulin dimer was found by Scatchard analysis (Laclette et al., 1980) -
see Appendix for discussion on Scatchard analysis. Furthermore, an KY,, of
1 X ~O-‘M was found for albendazole by inhibition of [3H]colchicine binding
to tubulin from the intestinal cells of Ascaris suum, a nematode that infects
pigs (Barrowman et al., 1984). In confirmation of these findings, benzimida-
zole resistance results in lack of colchicine binding to microtubules (Sangster
et al., 1985).
The importance of microtubules to the organism appears to be in the
secretion of acetylcholinesterase - regulating peristalsis in the host intestine.
This allows the nematode to hang on to the stomach wall more efficiently and
also delays the approach of lymphocytes. Mebendazole is a powerful inhibitor
 Microtubule assembly 259

of cholinesterase release at submicromolar concentrations and causes the

enzyme to accumulate inside the parasite. Other microtubule inhibitors such
as colchicine are moderately effective at inhibiting enzyme release but vinblas-
tine is ineffective, probably because it does not bind to nematode tubulin
(Tekwani, 1992).

11.6 Tax01 as an anti-tumour agent

The needles and bark of the common yew, Taxus baccata, yield an anti-
tumour agent named, not surprisingly, taxol. This compound has been known
for some time but only recently has it been developed for the treatment of
lung, breast and ovarian cancers.

0-C-Ph
I

Ph OH 0
I I II
Ph-C-NH-CH-CH-C-O

0 OH

Taxol

Taxol is unusual amongst anti-mitotic agents in causing tubulin to associate

into microtubules in the absence of GTP and MAPS. Tax01 also instigates
microtubule hydrolysis of GTP by assembled tubulin. There are more centres
for growth, the microtubules are shorter and ribbon-like structures are formed
(Schiff et al., 1979; Ringel and Horwitz, 1991). The drug binds to microtubules
with a binding constant near 1 X lO-(j~. The purification of brain microtu-
bules by taxol is a sufficiently reliable process to form the basis of a publication
(Wabbee and Collins, 1988).
Taxol binds at a different site from the other anti-mitotic drugs, colchicine,
vinblastine and vincristine (Correia, 1991) and there is no cross-resistance
between them. Taxol and its semi-synthetic analogue, taxotere, induce
microtubule bundle formation in cells and thereby prevent cell division (Ringel
and Horwitz, 1991).
The wide variety of structures that will bind to tubulin even at the same
site argues for considerable heterogeneity on the protein surface, and would
make a very interesting X-ray study. Furthermore, the variety of the uses of
microtubule ligands is intriguing and suggests that other structural types may
be found that bind to tubulin and show pharmacological activity.
 Chapter 12

Hormonal modulators

12. I Introduction

This chapter concerns the drugs that modulate the action of two types of
hormone: insulin and analogues of gonadotrophin hormone releasing hormone
(GnRH). Although we do not know the exact mechanism for the drugs that
lower blood glucose or the detailed mechanism of action of insulin itself (e.g.
the nature of the second messenger, if any), it is necessary to include the
important area of anti-diabetic drugs. It is interesting to compare our limited
knowledge of the mode of action of insulin, which has been investigated for
several decades, with the way in which GnRH analogues, that have only been
known for just over a decade, have broadened our knowledge of the relation-
ship between hypothalamus, pituitary and gonads.

12.2 Diabetes mellitus

The condition of diabetes is still a very serious medical problem which can
only partially be alleviated by pharmacological intervention. There are two
broad categories of diabetes mellitus: one in which the patient is totally depen-
dent on insulin injections because the fiells of the pancreas which normally
secrete insulin under the influence of glucose have been destroyed; and the
other in which the patient has circulating levels of insulin but is resistant to
the action of the hormone. The former, insulindependent, type is less frequent
than the latter but is more serious and usually occurs early in life (hence the
term ‘juvenile onset’ is used to describe it). It may result from viral infection -
Coxsackie B virus has been implicated in at least one case (Yoon et al., 1979).
Non-insulin dependent diabetes of which several sub-types exist, is more often
a disease of later life (‘maturity onset’) and is frequently associated with
obesity.
The basic metabolic defect lies in the fact that glucose cannot enter cells,
either because there is no insulin in the plasma to induce transport or because
the insulin that is present (and its level may be higher than normal) is rendered
ineffective, possibly by inadequate or insufficient receptors or by a post-recep-

261
262 Molecular mechanisms of drug action

tar lesion. Consequently, excess glucose is found in the plasma

(hyperglycaemia) and is excreted into the urine together with copious
amounts of water (polyuria). The patient becomes dehydrated and drinks to
assuage the thirst (polydipsia). Glucose cannot enter the appetite-regulating
cells of the hypothalamus, so the patient always feels hungry and eats fre-
quently (polyphagia).
The overall consequence is an increase in weight, leading to frank obesity
in some cases; the production of vast quantities of sweet-tasting urine (me1 is
the Latin for honey and diabetes is related to the Greek work for syphon); and
fatigue through the lack of available energy (Maurer, 1979).
The degree of severity of the condition may be studied by the administration
of a glucose ‘meal’ (50 or 100 grams), otherwise known as a glucose tolerance
test. In non-diabetics, the levels of blood glucose rise to a peak in 45 minutes
and return to the starting level in about lb hours. In those with mild disease,
the glucose level rises much higher and takes up to 4 hours to return to the
baseline, followed by an overshoot before returning to normal. Others with
more serious disease show elevated glucose levels in the absence of food and
very high levels of glucose for a long time after a glucose meal (Montgomery
et al., 1983).
If the condition is uncontrolled, the patient can relapse into a coma and
die. This is partially a consequence of the dehydration and partly from meta-
bolic acidosis. Because the cells cannot utilize glucose for the production of
energy, they break down triglyceride fat instead. Acetyl-CoA levels are raised,
leading to the production of acetone and acetoacetate and phydroxybutyrate,
which are secreted into the blood by the liver. The latter two can be used by
the brain, heart and skeletal muscle as a source of energy to a limited extent
but then they begin to accumulate in the plasma. Acetoacetate and phydroxy-
butyrate are relatively strong acids, and eventually the buffering capacity of
the blood is overcome, causing acidosis. The kidneys excrete these anions
together with cations, such as sodium, and water, thus contributing to the
dehydration, until eventually the patient goes into a coma as a result of the
combination of acidosis and dehydration (Montgomery et al., 1983).

12.2.1 The action ofinsulin

The action of insulin in cellular terms is responsible for the stimulation of
glycogen and lipid synthesis and glucose oxidation concomitantly with the
inhibition of glycogenolysis, lipolysis and gluconeogenesis. Any lesion in insu-
lin action, whether it occurs at the receptor on the cell surface or sub
sequently, will cause increases in lipolysis and glycogenolysis, with elevation
of glucose and acetyl-CoA levels leading to the formation of acetone etc. Extra-
cellular levels of glucose will rise since the major route of glucose removal
from the blood is into the peripheral muscle tissue (Pallet, 1983).
The intracellular effects of insulin are possibly initiated by means of an, as
yet, unidentified second messenger (Malchoff et al., 1987). Molecules of glu-
 Hormonal modulators 263

case carrier protein are transferred from the Golgi apparatus to the cell surface
ready to transport glucose into the cell. The major series of metabolic changes
noted above requires the phosphorylation status of the key enzymes of glyco-
gen breakdown, lipolysis and lipogenesis to be altered:
Dephosphorylation Result tnsulin
Glycogen synthase Glycogen synthesis +
activated
Glycogen phosphorylase Glycogen breakdown +
inhibited
Pyruvate dehydrogenase TCA cycle activity +
increased
HMG-CoA reductase Cholesterol synthesis +
activated
Phosphorylation Result
Glycogen synthase Glycogen synthesis
inhibited
Glycogen phosphorylase Glycogen breakdown
activated
AI-P-citrate lyase Lipogenesis activated +
Acetyl-CoA carboxylase Lipogenesis activated +

HMG = hydroxmethylglutaryl
TCA = tricarboxylic acid
These events follow from insulin binding to its receptor. The insulin receptor
is a glycoprotein heterodimer (c&) of molecular weight 430 kDa, composed
of two 125 kDa a subunits and two p subunits of 90 kDa. The receptor con-
tains two functional binding sites and is stabilised by disulphide bonds
between the subunits. The a subunits are extracellular, while the p subunits
are transmembmne proteins that can autophosphorylate (ATP-linked) the
hydroxyl of several tyrosine residues. Insulin binds to the a subunit and causes
the expression of tyrosine kinase activity, together with receptor aggregation.
The receptor can react with several cellular proteins to generate a cascade
of phosphorylation and dephosphorylation on serine or threonine residues as
noted above.
Not all of the actions of insulin may be mediated through autophosphoryl-
ation, as insulin receptor mutants without tyrosine kinase activity can still
mediate some of the actions of insulin (Sung, 1992). For example, a specnic
guanine regulatory protein activates a specihc phospholipase C, catalysing the
formation of phosphatidylinositol-glycan and diacylglycerol @ornero et aZ.,
1988; Yip, 1992).
The binding of hormone to receptors in liver and muscle also causes the
receptors to cluster in small groups followed by endocytosis of a receptor
cluster: subsequent fusion with lysosomes leads to the degradation of insulin
and the final event is recycling of the receptors.
264 Molecular mechanisms of drug action

12.2.2 Insulin therapy

From a therapeutic point of view, diabetes can be approached in a number

of ways. Insulin-dependent diabetics, by defnition, require insulin which in
earlier years was extracted from animal pancreas (usually that of pigs or cows).
In 1979, Goeddel et aZ. reported the expression of the human insulin gene in
E. coli. They chemically synthesised the genes for the A chain (21 amino acids)
and the B chain (30 amino acids) and inserted them into the plasmid pBR 322.
The insulin genes were attached to the carboxy terminus of the gene for p
galactosidase so that the human insulin would be efficiently transcribed and
translated as part of one long protein chain. The separate insulin chains were
subsequently cleaved from the @galactosidase using cyanogen bromide. The
sulphydryl groups were converted to S-sulphonyl groups, the chains were
mixed and were allowed to combine by first reducing the mixture followed
by air oxidation. The presence of insulin was confIrmed by radioimmunoassay.
The production of sufgcient quantities of msulin for human treatment is
described by Johnson (1983) - an interesting story of the problems faced by
the development of the first genetically engineered health-care agent to be
produced on a commercial scale.
Long-acting insulin preparations are produced by the reaction of insulin with
the basic protein protamine in the presence of zinc. When the preparation is
injected subcutaneously in aqueous suspension, it dissolves slowly at the injec-
tion site and is thus absorbed slowly.
Periodic injections of insulin, however, tend to cause the blood level of
glucose to oscillate wildly (Bressler, 1978) because the action of insulin is
normally balanced by glucagon, whose role is to raise blood glucose levels
by increasing liver glucose production and opposing liver glucose storage (as
glycogen). The net result in the normal subject is to maintain levels of glucose
close to the norm. The peaks in glucose level in the diabetic treated with
periodic insulin injections are believed to contribute to the complications
associated with diabetes. Recently, studies have been carried out with insulin
infusion pumps taped on to the wall of the abdomen, such that the point of
needle lies in the subcutaneous tissue. Insulin is then pumped through the
syringe from a reservoir at a steady basal level, with additional doses just before
meals (see Nathan, 1992).
In patients with non-insulin-dependent diabetes that is relatively mild, the
patients are likely to be insulin-resistant in the sense that they have a reduced
number of receptors. In this case the maximum response to the hormone is
merely attained at higher concentrations. Treatment with a drug that lowers
serum glucose levels (a hypoglycaemic agent) causes the number of receptors
usually to return to normal.
With patients more seriously affected by the disease, down-regulation of the
receptors is less likely to be of importance since the circulating insulin levels
range from slightly elevated to below normal. A marked post-receptor defect
is almost certainly present in these patients since an increased insulin level
 Hormonal modulators 265

does not return glucose levels to normal. The nature of this defect is not
known at present (Lockwood and Amatruda, 1983).
Glucagon is another peptide secreted by the pancreas, in this case by the
a cells. It has been proposed that too much glucagon as well as too little
insulin may be necessary for insulindependent diabetes to manifest (Unger and
Orci, l!%lb), since the actions ?f glucagon are antagonistic to those of insulin.
These actions are inhibition of cholesterol synthesis and stimulation of lip-
olysis, glycogenolysis and gluconeogenesis, with eventual hyperglycaemia.
Also produced in the pancreas is the 14-amino acid peptide somatostatin
which, amongst other actions, inhibits the release of both glucagon and insu-
lin. Indeed, since the a cells of the pancreas produce glucagon, the /3 cells
insulin and the 6 cells somatostatin, an intrapancreal control mechanism has
been proposed (Unger and Orci, 198la; Fig. 12.1). Consistent with this sugges-
tion, there is at least one report of combined treatment with insulin and soma-
tostatin which has given better results than insulin alone @askin and Unger,
1978). Somatostatin probably works by suppressing glucagon release (Fig.
12.1).
Although this is a book about the uses of pharmaceuticals to combat disease,
it is not the intention of the author to imply that drugs are the ultimate panacea
for all illness. In no disease is this more relevant than in non-insulindependent
diabetes. In many cases, dietary control with concomitant weight loss, coupled
with exercise, will often control the condition. The rationale for this approach
lies in the fact that many patients, particularly those who are obese, have
elevated levels of insulin together with a lowered number of receptors, as has
been seen in some diabetic animals (Roth et al, 1975). A reduction in weight
causes the number of receptors to rise and the level of circulating insulin to
fall as is likely to be the consequence if the intake of mono- and disaccharides
is restricted. Intake of polysaccharides apparently need not be restricted, pre-
sumably because these give rise to less available mono- and disaccharides. Satu-
rated fat intake should also be lowered, in view of the greatly increased risk
of heart disease with diabetes (American Diabetic Association, 1979). Weight
reduction, even to a modest extent, greatly improves the responsiveness of
obese subjects to insulin (Lockwood and Amatruda, 1983). Exercise is also

a
Somatostatin
#ii\ 4

4 +
Somatostatin 06
Arrows, an interaction between the hormone secreted from one cell type and the product
of another. +, stimulation, -, inhibition.

Figure 12.1 The interactions between pancreatic hormones.

 Hormonal modulators 267

pared with respect to both inhibition of radioactive ligand binding and hypo-

R,
1\
glycaemic activity in rabbits. A good correlation was observed, although a num-

-o-
-
S02NHCONH&

Toibutamide

Chlorpropamide

RCONHEHA+=& S02NHCONH 0

Glibenclamide

Glipizide CH3

ber of compounds with binding constants between lo-’ and lo-’ M had to be
omitted from the analysis because they failed to show suffkient hypoglycaemic
activity - probably because they were not absorbed sufllciently or were too
heavily serum-bound to be effective.
The second action of the hypoglycaemic agents is potentiation of the action
of insulin. They increase the absorption of glucose engendered by a given dose
of insulin; whether this is by increasing the number of insulin receptors
through preventing down-regulation or by a post-receptor effect is not known
(Lebovitz, 1984).
Chronic administration of these drugs eventually results in only normal (or
even reduced) insulin secretion (with the possible exception of glipizide). In
contrast, acute dosing raises insulin levels. It is likely, therefore, that the extra-
pancreatic effects are more important in the long term, although in some cases
the drugs may gradually lose their efficacy.
268 Molecular mechanisms of drug action

There is no clear indication that the second generation drugs are any more
effective therapeutically than the tirst, although the former are more active at
low concentrations. The number of failures with each type is almost identical,
and so there seems to be little to choose between them (Kreisberg, 1985).
They are nevertheless very useful for the treatment of the more severely hyper-
glycaemic patients for whom insulin is not yet necessary (Lebovitz, 1992).

12.4 Gonadotrophin-hormone-releasing hormone (GnRfQ

analogues

A growing area of pharmacological intervention has been opened up for a

number of cases of hormonal aberrations and cancers that depend on sex
steroid hormones for growth, by a greater understanding of the axis that links
the hypothalamus to the gonads via the pituitary gland (Fig. 12.2). The hor-
mones that carry the information from pituitary to gonads are collectively
known as the gonadotrophins: luteinizing hormone (LH) and follicle-stimulat-
ing hormone (FSH) in females - LH is referred to as interstitial-cell-stimulating
hormone (ICSH) in males.
Gonadotrophin-hormone-releasing hormone (G&H or gonadorelin) is
released from the hypothalamus and acts on the anterior portion of the pitu-
itary gland to release gonadotrophins, although there are suggestions that FSH
may also be released by another as yet undiscovered hormone. GnRH secretion
is under the control of higher centres in a way that we do not yet understand,
although it is clear that in women some form of ‘biological clock’ must operate
in order to maintain the functioning of the menstrual cycle.
GnRH is released in short pulses of millisecond duration. This can be reco-
gnized by following the pulsation in plasma levels of LH, since LH release is

Hypothalamus

Oestrogen Progesterone
+ indicates a positive feedback
- indicates a negative feedback
Figure 12.2 Hypothalamus-pituitary-gonad axis.
 Hormonal modulators 269

very sensitive to GnRH pulsation and the half-life of LH in plasma is shorter

than the time between pulses. FSH release is much less sensitive to GnRH
levels and does not pulse because the response is much slower and takes
longer to die away (reviewed in Lincom et al, 1985).
In the female, FSH stimulates the growth of follicles in the ovary which are
responsible for the production of the main oestrogen, the steroid hormone
17/3-oestradiol (oestrone and oestriol are derived from oestradiol in the liver
and are less active as oestrogens). By the time that ovulation is about to occur,
various follicles are almost mature. There is then a surge in the level of LH
which initiates ovulation in one or more follicles while in some way causing
the others to regress. The follicle bursts and discharges the egg. The follicle
then undergoes a re-organisation of its cells developing into a yellow body
known as the corpus luteum (a process therefore called luteinization) that
secretes progesterone still under the influence of LH.
Progesterone prepares the female for pregnancy by stimulating the develop-
ment of the lining of the uterus and inhibiting contraction of the smooth mus-
cle of the uterine wall. It also prevents the development of a new follicle.
Both progesterone and oestrogen act to close down FSH and LH release from
the pituitary and possibly GnRH release from the hypothalamus in a negative
feedback loop. Furthermore, oestrogen exerts a secondary, but positive, feed-
back effect on the pituitary in conjunction with GnRH which is believed to
be responsible for the pre-ovulatory surge in LH production (Fig. 12.2). The
use of GnRH analogues played a part in the discovery of these interactions
(Schally, 1978; Asch et al., 1983).
FSH and LH bind to membrane fractions from ovary tissue and exert their
biochemical action through the stimulation of adenylate cyclase. Cleavage of
the side-chain of cholesterol is induced, the rate-limiting step in steroid
biosynthesis, in a fashion similar to the events in the adrenal cortex
(Funkenstein et aZ., 1983; Fig. 6.2). Oestrogen and progesterone production
are thereby greatly increased (Fig. 12.3).
In the male, LH stimulates the Leydig cells of the testis to secrete testoster-
one. As these cells are also known as interstitial cells, LH is sometimes referred
to as interstitial-cell-stimulating hormone or ICSH. FSH, however, is required
only for the maturation of the spermatids into spermatozoa in some way that
is not fully understood. Nevertheless, there is a similarity with the female in
that testosterone exerts a negative feedback inhibition on LH secretion from
the pituitary.
LH and FSH are glycoproteins of molecular weight about 16 kDa and are
dimers of non-identical subunits cr and p. The o subunits are nearly identical
in the two hormones (and in thyroid-stimulating hormone), whereas the p
subunits show some differences - although homology is still 80 per cent
(Pierce and Parsons, 1981). GnRH is a decapeptide (see structures) that binds
to pituitary cell membrane receptors belonging to the 7-transmembrane group
with a binding constant of 3.0 X 10p9~ (Corm et aZ., 1985). Occupation of
the receptors causes them to aggregate, probably into dimers, and open a
270 Molecular mechanisms of drug action

Cholesterol
HO
Cholesterol
20,22 lyase
I
Pregnenolone
HO
I
3-Ghydroxysteroid
Dehydrogenase
I
Progesterone

Progesterone
17,20 lyase
I
?7=-OH Progesterone

Progesterone
17,20 lyase
I
Androstenedione

Steroid
17-ketoreductase
I
Testosterone

Aromatase

I
Oestradiol
HO
Figure 12.3 Sex hormone production from cholesterol.
 Hormonal modulators 271

calcium channel in the cell membrane. Calcium is also released from intracellu-
lar stores as normal for the phospholipase C pathway and acts as the second
messenger via calmodulin (section 9.1). Protein kinase C is also activated but
its role is not entirely clear, although it may be required for the expression of
the LH /3 subunit gene. The gonadotrophins LH and FSH are stored in secretory
granules which move to the cell surface, fuse with the cell membrane and
discharge their contents (reviewed in Corm and Crawley, 1991).
The GnRH receptor has been cloned and expressed in Xenopus oocytes
using as a detector the calcium-binding protein aequorin, which emits light
when calcium is bound. The receptor showed little homology with other G-
protein receptors, although there are structural similarities such as N-linked
glycosylation sites near the extracellular N-terminus. There were two potential
sites for phosphorylation by cyclic-AMPdependent kinase in the first intracellu-
lar loop and a protein kinase C site in the third intracellular loop. There was
no aspartate in the third transmembrane domain and, most unusual of all, there
was no cytoplasmic C-terminal domain. Normally this area controls G-protein
signalling and desensitisation. In view of the importance of desensitisation for
this receptor (see below) it will be very interesting to discover which residues
are involved (Reinhart et al., 1992).
The part of the molecule that is responsible for binding to the receptor (the
R site) is made up of residues from the C and N terminal ends, namely PyroGlu’
and Gly’“, and modification of these residues destroys binding. On the other
hand, His* and Tip3 appear to be essential to stimulate secretion of LH (M
site) which occurs as a consequence of receptor occupancy. Agonists of GnRH
such as the peptide leuprolide contain both M and R sites. Gly” is substituted
by an ethylamide group and a o-amino acid is in position 6 which reduces
proteolysis and increases receptor and plasma protein affinity. Recently, com-
pounds have been developed with a bulky hydrophobic group at position 6
to increase the affinity still further. Half-lives in uiuo are thus increased as well
as potency. In addition, calcium is required for agonists to show full activity
(Conn et aZ., 1985) as one would expect for a secretory process.
Furthermore, agonist action induces another effect of importance to the
therapeutic use of GnRH and its analogues, namely a surge of gonadotrophin
release followed by increasing insensitivity to GnRH and its analogues. Desensi-
tization can occur as a result of several mechanisms, including a reduction in
the number of available receptors on the cell surface because of internaliz-
ation, uncoupling of receptors from the signal pathways and an exhaustion of
gonadotrophin stores inside the cell. Desensitisation occurs in cultured pitu-
itary cells (Catt et al., 1985) and also in uiuo and is calcium-independent -
unlike LH release. Some of the internalised receptors may eventually be
returned to the cell surface (Hazum and Corm, 1988).
Antagonists, however, merely block GnRH binding to the receptor and do
not cause down-regulation. In contrast to agonist structure, GnRH antagonists
contain only the R site. In addition to the substitutions for agonists, hydro-
phobic amino acids are substituted in positions 1 to 3. The antagonists
272 Molecular mechanisms of drug action

Gonadorelin (GnRH):

pyroGlu - His - Trp - Ser - Tyr - Gly - Leu - Arg - Pro - GlyNH2

Leuprolide

pyroGlu - His - Trp - Ser - Tyr - D leu - Leu - Arg - Pro CONHC2H5

developed to date have shown unacceptable side-effects such as histamine

release but other more specmc compounds are under investigation (Come and
Crawley, 1991).
The conditions for which GnRH or its analogues are finding, or are likely
to fmd, therapeutic usage are numerous (Corm and Crawley, 1991). The major
use of the native hormone is to induce ovulation in women who suffer from
infertility because of insufficient gonadotrophin. This has given rise, on
occasion, to multiple pregnancies because the high level of hormones released
from the pituitary causes a large number of follicles to mature at the same
time. Other hormone abnormalities in men, such as delayed puberty
(hypogonadotrophic hypogonadism) and delayed descent of the testicles
(cryptorchidism) have also been shown to respond to treatment. These latter
treatments will probably need to be given for life, however, as the therapeutic
effects are reversible if drug administration is stopped.
Low doses of drug administered in pulses, mimicking the release of the
natural hormone, produce an elevation in gonadotrophins. If, however, the
drug is given systemically in higher doses, for example as a single daily subcut-
aneous injection, after a transient increase the pituitary is desensitised as a
result of receptor down-regulation, and the levels of FSH and LH fall sharply.
As a consequence, leuprolide and other GnRH agonists are, paradoxically,
being used to treat conditions where elevated GnRH levels are the cause of
the condition, notably the condition where GnRH levels rise very early in life
and induce sexual maturation (centrally derived precocious puberty).
Furthermore, there are some tumours that are dependent on sex steroids:
some breast cancers in women are dependent on oestrogen and 80 per cent
of prostate cancers require testosterone (Vance and Smith, 1984) and these
have responded to agonist therapy. A particular advantage that leuprolide ther-
apy has for prostate cancer over the usual treatment with diethylstilboestrol,
is the lack of feminising and cardiovascular side-effects seen with the latter
drug. Castration can also be used but is irreversible, and thereby unacceptable
to many patients. Santen (1992) indicates that depot leuprolide (acetate form)
given monthly is primary therapy for prostate cancer and may be combined
with an androgen antagonist to combat the initial disease ‘flare’ (an agonist
will initially produce a surge of gonadotrophins before desensitisation occurs).
 Hormonal modulators 273

It has also been proposed that premenstrual syndrome results from a cyclic
condition of elevated release of (or abnormal sensitivity to) LH and FSH
(Cot&on, 1986). The cyclic condition is characterized, amongst other symp-
toms, by weight gain, mood swings, irritability and inability to concentrate.
Trials have shown that GnRH analogues were effective in relieving this con-
dition (Muse, 1992) but further trials are required before the results can be
confirmed, since the placebo effect is very marked in this condition.

Questions

1. What enzymes do insulin action (a) stimulate (b) inhibit?

2. How does the insulin receptor transmit the signal to the cell interior?
3. How do sulphonylurea drugs combat diabetes?
4. What is structurally unusual about the gonadorelin receptor compared with
other 7-transmembrane receptors?
5. Why are gonadorelin agonists used to antagonise conditions which depend
on high gonadorelin levels?

References
American Diabetic Association, 1979, Diabetes, 28, 1027-30.
Asch, R.H., Balmaceda, R., Borghi, M.R., Niesvisky, R., Coy, D.H. and Schally, A.V.,
1983, J Clin. Endocrinol. Metab., 57, 367-72.
Bressler, R., 1978, Drugs, 17, 461-70.
Catt, J.K., Loumaye, E., Wynn, P.C., Iwashita, M., Hirota, K., Morgan, R.O. and Chang,
J.P., 1985,J Steroid Biocbem., 25(5B), 677-89.
Conn, P.M. and Crawley, W.F., 1991, New Eng1.J. Med., 324, 93-103.
Conn, P.M., Staley, D., Jinnah, H. and Bates, M., 1985, J. Steroid Biocbem., 23(5B),
703-10.
Coulson, C.J., 1986, Med. Hypotbes., 19, 243-56.
Funkenstein, B., Waterman, M.R., Masters, B.S.S. and Simpson, E.R., 1983, J. Biol.
Cbem., 258, 10187-91.
Garrino, M.G., Meissner, H.P. and Henquin, J.C., 1986, Eur. J. Pbarmacol, 124,
309-16.
Geisen, K., Hitzel, V., Okomonopoulos, R., Punter, J., Weyer, R. and Summ, H.-D., 1985,
Arzneim. Forscb./Drug 707-12.
Res., 35,
Goeddel, D.V., Kleid, D.G., Bolivar, F., Heyneker, H.L., Yansura, D.G., Crea, R., Hirose,
T., Kraszewski, A., Itakura, K. and Riggs, A.D., 1979, Proc. Nat. Acad. Sci. (lIS.A.),
76, 1106-10.
Hazum, E. and Corm, P.M., 1988, Endow Rev., 9, 379-87.
Jackson, J.E. and Bressler, R., 1981, mgs, 22, 21 l-45.
Johnson, IS., 1983, Science, 219, 632-7.
Kreisberg, R.A., 1985, Ann. Intern. Med., 102, 125-6.
Lebovitz, H.E., 1984, Diabetes Care, 7 (Suppl. l), 67-71.
Lebovitz, H.E., 1992, Drugs, wU, Suppl. 3, 21-8.
Lincoln, D.W., Fraser, H.M., Lincoln, G.A., Martin, G.B. and McNeilly, H.S., 1985, Rec.
Progr. Hormone Res., 41, 369-411.
274 Molecular mechanisms of drug action

Lockwood, D.H. and Amatruda, J.M., 1983, Amer. J. Med., 75, Suppl. 5B, 23-31.
Longman, SD. and Hamilton, T.C., 1992, Med. Res. Rev., 12, 73-148.
Malchoff, C.D., Huang, L., Gillespie, N., Palasi, C.V., Schwartz, C.F.W., Cheng, K., Hew
lett, E.L. and Lamer, J., 1987, Endocrinology, 120, 1327-37.
Maurer, A.C., 1979, Amer. Sci., 67, 422-31.
Montgomery, R., Dryer, R.L., Conway, T.W. and Spector, A.A., 1983, BiocbemzWy: a
case-orientated approach (St. Louis: C.V. Mosby) Ch. 7.
Muse, K.N., 1992, Clin Obstet. Gynecol., 35, 658-66.
Nathan, D.M., 1992, Drugs, e (Suppl 3), 39-46.
Pierce, J.G. and Parsons, T.F., 1981, Annu. Rev. Biocbem., SO, 465-95.
Pallet, R.J., 1983, Amer.J Med., 75, Suppl. 5B, 15-22.
Raskin, P. and Unger, R.H., 1978, New EngL J Med., 209, 433-6.
Reinhart, J., Mertz, L.M. and Catt, K.J., 1992, J. BioZ. Cbem., 267, 21281-4.
Romero, G., Luttrell, L., Rogol, A., Keller, K., Hewlett, E. and Iarner, J., 1988, Science,
240, 509-11.
Roth, J., Kahn, C.R., Lesnick, M.A., Gorden, P., De Meyts, P., Megyesi, K., Neville, D.M.,
Gavin, J.R., Soil, A.H., Freycher, P., Goldf?me, I.D., Bar, R.S. and Archer, J.A., 1975,
Rec. Progr. Hormone Res., 31, 95-128.
Santen, R.J., 1992,J. Clin. Endocrinol. Metab., 75, 685-9.
Schally, A.V., 1978, Science, 202, 18-28.
Sung, C.K., 1992,J Cell Biocbem., 48, 26-32.
Unger, R.H. and Orci, L., 198la, New Eng1.J Med., 3@4, 1518-24.
Unger, R.H. and Orci, L., 198lb, New EngLJ. Med., 304, 1575-80.
Vance, M.A. and Smith, J.A., 1984, Clin. PbarmacoL Zber., 36, 350-4.
Yip, C.C., 1992,J. Cell Biocbem., 48, 19-25.
Yoon, J.-W., Austin, M., Onodera T. and No&ins, A.L., 1979, New Engl. J. Med., 300,
1173-9.
 Appendix

Quantification of ligand-
macromolecule binding

The binding of a ligand to a macromolecule, whether it be receptor, enzyme

or just a binding protein, is governed by the Law of Mass Action, which states
that the rate of a reaction is proportional to the concentrations of the reactants.
A variety of plotting techniques have been used to elucidate the binding of
ligands to macromolecules. It is useful to show how ligand binding to macro-
molecules, enzyme kinetics and drug-receptor binding interrelate and, in
addition, to describe some other methods of plotting such as the Hill and
Scatchard plots.

A.1 Enzyme kinetics: I5o or Kb?

Much of the study of enzyme inhibition depends on measurement of Ki (the

dissociation constant of the enzyme-inhibitor complex). Another measure of
inhibition is I&,; the concentration of inhibitor required to reduce the reaction
velocity to SO per cent of its value in the absence of inhibitor. Ki and I&, are
not the same, except in very unusual circumstances, and furthermore, the Z&,
values will vary if different substrate conditions are used (see Cheng and Pru-
soff, 1973, for a detailed discussion).
Considering the simplest case of an enzyme reaction with one substrate, S;
if the concentration of substrate is very much higher than that of enzyme, the
initial velocity V. is given by:

where V,,,= = maximum velocity; Km = Michaelis constant of the substrate

which equals the concentration of substrate that produces the half-maximal
rate.
(a) In the most common form of competitive inhibition, described by the
following equation, the substrate and inhibitor compete for the same active
site, but with different affinities for it:

275
276 Molecular naecbanisms of drug action

E+S+ ES+E+P
11
EI

Here EI = enzyme-inhibitor complex and P is product. The equation govern-

ing the reaction rate is:

(3)

where Vi is the initial velocity in the presence of inhibitor at concentration i

and Ki is the dissociation constant of the enzyme-inhibitor complex. If the
initial velocity is one-half V& the inhibitor concentration i by definition equals
the KS0 with Vi = $VO, so after rearrangement:

(4)

Clearly, I&, is dependent on substrate concentration. Since different labora-

tories are unlikely to use the same substrate concentrations, their KS,, values
are not a sound basis for comparison of different inhibitors. Only if the sub
strate concentration is low compared with the Km does ZC50 approximate to
Ki. A low concentration (less than one-tenth of the K,,, will also render the
assay more sensitive because the effect of an inhibitor will thereby be magni-
fied, but this might lead to problems in detecting product formation (Bush,
1986).
(b) In the case of noncompetitive inhibition

E+S* ESeE+P
1 4I’ -Kis 1 Jr& (5)
EI ES1

where Kis is the affinity of the inhibitor for full enzyme and KiI for the enzyme-
substrate complex.
It can be shown that:

~c = UC,, + PI)
50

If the inhibitor binds equally well to the enzyme as to the enzyme-substrate

complex, Kis = KiI = Ki and the equation simplilies to:

IC5o = Ki m
Alternatively, if as is often the case [Sl B K,,, and KJ[SJ-G KiJKiI, the equation
reduces to:
 Quantzjiication of ligand-macromolecule binding 277

ZC50 = Ki

In the case of uncompetitive inhibition, provided that [SJ + Km, ZC&, is inde-
pendent of substrate concentration and equal to Ki. It is not always convenient
to maintain a substrate concentration that is very much higher than the Km
because this may mask the effect of weaker inhibitors. Further derivations for
two-substrate reactions may be found in Cheng and Prusoff (1973).
In practical terms there is not always enough time to measure sufticient data
points in order to define a Ki. If a large number of compounds have to be
tested, as is normally the case, it is often expedient to establish Z&,, and,
provided that the compounds all inhibit the enzyme in a similar fashion, the
Z&, values can be used for ranking purposes. Subsequently, when a few com-
pounds are chosen for development more work can be carried out in order
to estimate their Ki values and define the type of inhibition.
A further consideration in carrying out enzyme reactions with inhibitors in
vitro is that the enzyme concentration is set at a rate-limiting level (very much
less than the substrate concentration) in order that Michaelis-Menten kinetics
should be applicable. This usually requires enzyme concentmtions below ~O-‘M
and sometimes as low as 10-“~. This can be quite unlike the normal physio-
logical condition where reaction rates may be regulated by substrate avail-
ability rather than by enzyme concentration. Indeed, in the glycolytic pathway
of the yeast Saccbaromyces carlsbergensis, the molarities of only two
enzymes, phosphofructokinase and adenylate kinase, lie below 10p5~ (Hess
et al., 1969). If we consider the ‘concentration’ of enzyme active sites, the
range for all of the enzymes measured is between 2.5 X lO-5 and 2.0 X 10e4 M.
The substrate concentrations vary in the same range.
Although enzymes in the glycolytic pathway may be exceptionally highly
concentrated, there is no reason to suppose that other pathways may not be
similar and, in addition, may show molecular organization to facilitate the pass-
age of substrate along the pathway. It is therefore a matter of some surprise
that inhibitors often do show effects in vivo that can be related to data
obtained in vitro. That they do so makes life easier for the medicinal scientist.
One of the important features of enzyme inhibition is whether the inhibition
is reversible or not. Irreversible inhibition is often characterised by inhibition
that increases over time because it is a consequence of a chemical reaction
that requires covalent bonds to be broken. Consequently, the inhibited enzyme
will not recover during dialysis, unlike a reversibly-inhibited enzyme. Irrevers
ible inhibition is frequently missed if considerable pre-incubation is not a nor-
mal part of the assay protocol, and often appears to be noncompetitive inhi-
bition because a portion of the enzyme is effectively put out of action by
covalent moditication (Bush, 1986). The order of addition of substrate, inhibi-
278 Molecular mechanisms of drug action

tor and enzyme may be crucial but is frequently unreported. In fact, a plot of
log(remaining activity) against time should be a straight line for this type of
inhibition. Kinetically, the reactions are normally exemplified as follows (e.g.
Coulson and Smith, 1979):
E + I + EI - EI* (10)
where EI* represents the inactivated enzyme. Irreversible inhibitors include
amino-acid-modifying agents such as iodoacetamide, active-site-directed inhibi-
tors and suicide inactivators such as the monoamine oxidase inhibitors.

A.2 Drug binding to receptors

Just as enzymologists use a plot of substrate concentration against reaction
rate to quantify an enzyme reaction, pharmacologists also need to quantify the
response of an organ preparation to a drug. For drug-receptor binding:
D+Re DR - Effect (111
where D and R represent drug and receptor, respectively. The equation gov-
erning the reaction is then:
~~l-%,ax
W)
’ = ([D] + &)
where E represents the effect produced by a given drug concentration [D],
E is the maximum effect possible under the conditions and K,, is the dis-
szation constant of the drug-receptor complex. This clearly bears a formal
similarity to the Michaelis-Menten equation of enzyme kinetics (equation 1).
The method of plotting is different from enzyme kinetics in that log(drug
concentration) is plotted against response which normally gives a sigmoid type
of curve (Fig. A.1). If two drugs are compared by this approach and produce
parallel lines it is likely, but not certain, that the compounds are acting in a
similar way at the receptor. If they are not parallel the interpretation is
more complex.
One way of analysing such data is by a Schild plot. This depends on the fact
that when an antagonist inhibits the response of a given tissue to an agonist
competitively, the log(dose)-response curve is shifted to the right so that one
obtains a plot similar to Fig. A. 1, with the tissue showing Lne same response
at a higher agonist concentration (LX’*) as it did previously at the original con-
centration (x’~). It has been shown (Arunlakshana and Schild, 1959) that:

where xB is the concentration of antagonist and KB its binding constant. If the

ratio between these agonist concentrations giving the same response is defmed
as R. then:
 QuantiJTcation of ligand-macromolecule binding 279

80

60-
Contraction l%J
40-

20-

o- r
10
Concentration (M)

The ethyl- and butyl-trimethyiammonium salts give parallel dose-response curves acting
as agonists. The heptvl and nonyl analogues, however, produce curves with totally
different shapes. This is interpreted as being the result of a gradual shift towards anta-
gonism. After Stephenson, 1956.

Figure A. 1 Log dose-response curve for contraction of guinea-pig ileum by alkyltrimethylammon~

ium ions.

(14)

and
log (R - 1) = log xB - log KB (15)
The Schild plot is log(R-1) against -log xB which will give a straight line
with slope close to unity and a negative intercept on the x axis equal to -log
KB (Fig. A.2).
The binding of a drug to a receptor normally depends on the dose in a linear
fashion. The magnitude of the tissue response is, however, often a non-linear
function of the drug concentration and reaches a maximum value, in fact a
maximum tissue response can occur with only a fraction of the receptors occu-

3-

2.
log (R-1)

1-l
01, , \
6 7 8 9 10
-Log xs

Antagonism by ligand B of a hypothetical agonist (A) at a specific receptor. The negative

log of the binding constant for the antagonist is given by the intercept on the x axis
(i.e. 10d9~).
Figure A.2 Schild plot.
280 Molecular mechanisms of drug action

pied. Accordingly, it is often assumed that the tissue contains a ftite number
of receptors with which the drug can interact; the tissue response then
depends on how many receptors are occupied by the drug which in turn
depends on the affinity of the receptor for the drug.
The concept of efficacy is also useful here. Ligands are ranked on a scale
from 0 to 1: an antagonist will bind to a receptor and elicit no tissue response
at all, efficacy = 0, whereas a full agonist will be ranked at 1. Partial agonists
will be ranked in between 0 and 1. These assumptions underlie much of what
is discussed in this book, and have been found to hold up reasonably well
in practice.

A.3 Ligand-protein binding

In order to complete this analysis, a third type of interaction must be con-

sidered - one where the binding of a ligand does not necessarily lead to any
further effect. The binding of drugs to serum albumin comes into this class
(Iusko and Gretch, 1976). In order to derive a mathematical formulation for
the interaction a number of criteria have to be obeyed:
1. The binding sites must not interact with each other.
2. The binding must be specific and reversible.
3. A steady state must be reached, with equilibrium between free and
bound ligand.
If we consider the simplest case of a macromolecule (M) with one binding
site for one molecule of ligand L, we have the following equilibrium:
M+L*ML (16)
The dissociation constant for this equilibrium is given by:

where square brackets denote concentrations and [L] is free ligand. If we con-
sider the number of moles of ligand bound (r) to one mole of macromolecule
we have:

whence

If we consider the more usual case of KZidentical but independent sites we

have the following formula:
 QuantzjIication of ligand-macromolecule binding 281

This can be rearranged in three ways to give derivations that should yield
straight lines when the following graphs are drawn; (a) l/r against l/[L]; (b)
r/[L] against r, and (c) [L]/r against [L]. The number of binding sites and the
binding constant can be measured from the plot.
(a) l/r = l/n + K&z [L] (21)
Cb) r/K1 = n/Kd - r/Kd cw
Cc) n Wr = [Ll + Kd (23)
Each of these three deviations has been used to study ligand binding to macro-
molecules. The second is probably the best known and is attributed to Scatch-
ard (1949). It is the best form mathematically, in that the plot does not rely
heavily on measurements taken at low ligand concentrations which are subject
to the greatest error in determination. The other two plots divide by Y which
will magnify the effect of inaccuracies. Furthermore, as Scatchard himself
noted, ‘double reciprocal plots tend to tempt straight lines where none exist’.
It is an interesting exercise to plot one set of data both by double-reciprocal
plot (equation 21) and by the Scatchard plot (equation 22) and compare the
figures for n and K obtained.
One major use of the Scatchard plot is in studying hormone binding to
receptor. The plot will also show up situations where more than one set of
binding sites exist with different binding constants. The plot then gives two
straight lines with different slopes cormected by a boundary region where one
curve blends into the other (Fig. A.3).
If we look at these equations in the light of enzyme kinetics and drug recep-
tor binding, it is clear that there are considerable formal similarities - despite
the fact that conversion to product occurs with an enzyme and pharmacologi-
cal effects result from the binding of a ligand to a receptor, whereas no further
change may result as a consequence of ligand binding to protein - for example
the binding of drugs to serum albumin (Jusko and Gretch, 1976). The key
similarity is that all of these interactions obey the Law of Mass Action.
Another method of plotting is due to Hill (1910) who derived an equation
to explain the binding of oxygen to haemoglobin. This plot is of value when
interaction between sites is suspected with the binding of one ligand to a
macromolecule either facilitating (positive cooperativity) or hindering
(negative cooperativity) the binding of a second. We take equation (18) and
transform it into the Hill equation:
r/Oz - r.1 = [Ll/Kd (26)
We can call the fraction of the binding sites occupied (Z), where Z = r-/n.
Then the left-hand side of equation (24) becomes:
z/(1 - z) = fLl/K,,
282 Molecular mechanisms of drug action

A. One set of binding sites.

r is the number of moles of ligand bound to one mole of macromolecule. [Ll is the free
ligand concentration. The intercept on the x axis is n (the number of binding sites). The
slope is -l/K+

B. Two sets of binding sites.

The intercept on the x axis is now the sum of the number of binding sites at each of the
two sets of site, n,+nz but the two slopes are not simply the negative reciprocals of the
respective binding constants. They are rather more complex (see Feldman, 1972 for a
mathematical analysis).
Figure A.3 Scatchard plot.

A plot of log {Z/(1 - Z)} against log [L] will give a straight line with a slope
of 1 if there is no cooperativity. If there is cooperativity, however, the slope
in the central portion of the curve will be greater than 1 for positive, and less
than 1 for negative, cooperativity. This is because at low concentrations of
drug the first binding site is being ffied and at high concentrations the last;
it is only in the middle that sites already ffied can influence those untIlled (Fig.
A.4). In fact, a Scatchard plot will also indicate cooperativity by being curved.
It is interesting to note the number of double-reciprocal plots used for
enzyme kinetics whereas for ligand binding the Scatchard plot is usually fav-
oured. This is not consistent but it may allow data to be made presentable.
Even the Scatchard plot may produce straight lines that could be the conse-
quence of obedience to a more complex equation as discussed by Klotz
(19831.

References
Arunlakshana, 0. and Schild, H.O., 1959, B&t. J. Pbarmacol., 14, 48-58.
Bush, K., 1986, Lkugs Exptl. Clin. Res.,
12, 565-76.
Cheng, Y.-C. and Prusoff, W.H., 1973, Biocbem. Pbarmacol., 22, 3099- 108.
Coulson, C.J. and Smith, V.J., 1979, Enzyme Microbial Tecbnol., 1, 193-6.
Hess, B., Boiteux, A. and Kruger, J., 1969, Adu. Enzyme Regul., 7, 149-67.
 QuanttjXcation of l&and-macromolecule binding 283

This is a Hill plot for haemoglobin in equilibrium with oxygen (p isthe partial pressure of
oxygen). The slope in the middle portion of the curve (the Hill coefficient) approximates to
2.8 whereas at both ends it is lower. The total number of binding sites for oxygen in
haemoglobin is 4, but this figure is not reached because to derive the Hill equation infinite
cooperativity is assumed. Since this is an ideal situation the Hill coefficient is always less
than the actual number of sites and in practice indicates a minimum number, in this case
3. After Koshland (1970).

Figure 0.4 Hill plot,

Hill, A.V., 1910,J Pbysiol., 40, iv-vii.

Jusko, W.J. and Gretch, M., 1976, Drug. Met&. Rev., 5, 43-140.
Klotz, I., 1983, Trends Pbarnaacol. Sci., 4, 253-5.
Koshland, D.E., 1970, In: 7be Enzymes, Vol. 1, 3rd Ed, Ed. P.D. Boyer (New York:
Academic Press), p. 341.
Scatchard, G.S., 1949, Ann. N.K Acad. Sci., 51, 660-72.
 Index

Abscess, brain 59 p-Adrenergic receptor blocker (/3-

Acetate 66, 96 blockers) 162, 163, 166, 222
Acetazolamide 141 o-Adrenergic receptors 161-3
Acetoacetate 262 p-Adrenergic receptors 163-6
Acetone 262 Adrenodoxin 108
Acetyl coenzyme A 66, 67, 96, 262 Adrenodoxin reductase 108
Acetyfcholine 153, 158, 167, 168, Adriamycin 34
188194, 237 After-depolarization 2 19
Acetylcholinesterase 153, 192-4 Agonist, definition of 2, 5
N-Acetylglucosamine 75 o-Afanine carboxypeptidase 78, 80
N-Acetylmuramic acid 75 Albendazole 258
Acetylphosphate 66 Albumin 6, 8, 280
Acetylsalicylic acid see Aspirin Alcohol dehydrogenase 135
Acetyltobramycin 55 Aldosterone 97, 117-9, 142, 147, 157
Acidosis 262 Alkaline phosphatase 136
Acquired immune deficiency syndrome Allopurinol 5, 14, 43-5
(AIDS) 18, 38 Allosteric effector 5
ACTH 107, 109 Alloxanthine 44
Actinomycin 34 Allylamines 103
Action potential, membrane (potential Ames test 70
difference) 215, 222 Amikacin 54
Active transport 7-8 Amiloride 223-5
Acycloguanosine 9 Amino acids, aromatic 7, 51
Acyclovir 9, 30-l 2-Amino-4-hydroxy-6-methylpteridine
Acyl CoA acyltransferase (ACAT) 97, 100 16, 168
Adenosine deaminase 5, 32, 41-3 p-Aminobenzoate 19
Adenosine diphosphate see ADP 4-(or y-) Aminobutyric acid (GABA)
Adenosine kinase 26, 41 receptors 159, 194-201
Adenosine monophosphate see AMP 7Arninocephalosporanic acid 78
Adenosine 3’,5’-monophosphate see Aminoglutethimide 109- 110
CAMP Aminoglycosides 54-7
Adenosine triphosphate see ATP 5-Aminoimidazole4-carboxamide ribotide
S-Adenosylhomocysteine 41 (AICAR) 15, 16, 17, 19
S-Adenosylhomocysteine hydrolase 4 1 &4rninopenicihanic acid 78, 82
S-Adenosylmethionine 4 1, 184 Amiodarone 222
Adenoviruses 26 Amoxiciflin 84
Adenylate cyclase 113, 155, 156, 161, mP 16, 51
163, 164, 166, 204, 208, 209, CAMP 155, 156, 206, 226, 237, 241
216, 237 dibutyryl 238
Adenylate kinase 41, 277 Amphotericin B 26, 244-7
ADP 226, 233 p-Amylase 135
Adrenaline 159 Anaemia, aplastic 2, 58
Adrenergic receptors 159-66 Anaesthetics 215, 218, 220

285
286 Index

local 200-21, 222 sodium/potassium 216, 217, 230,

Analgesics 148, 201 232-6
Androstenedione 107, 109 Atropine 189, 191
Angina pectoris 45-6, 132, 227, 229 Augmentin see Amoxicilhn
Angiotensin I 141, 142, 143 Autacoids 207
Angiotensin II 141, 142, 143, 146, 147, Avermectin 197-200
148, 157 Avermectin Bra 197, 199
Angiotensin-converting enzyme (ACE) Azidothymidine 38, 39
135 Aztreonam 78
inhibitors 136, 141-8
Angiotensinogen 142 Bacillus subtilis 78
Antagonists, definition of 2, 6 Baclofen 200- 1
Anthelminthic drugs 252, 257-9 Bacteria
Anthracychne antibiotics 34 gram-negative 6, 54
Anti-atrhythmic drugs 215, 218, 220, gram-positive 54
221, 222-3 Bactericidal action 6, 54
Anti-metabohte approach 13 Bacteriostatic action 6, 55
Antibacterial drugs 6, 19, 20 Bactet-oidesfragilis 58, 59, 61, 65, 68,
Antibiotics 69, 84
anthracycline 34 Bactoprenol 77
fluoroquinolone 37 Benzimidazole 236, 257-9
polyene 244-7 Benzodiazepines 195-7
Antidepressants, tricyclic 17 1, 180- 1 Benzomorphan 204
Antifungal drugs 6 Benzylamine 184
Antihypertensive drugs 2 15 Benzylpenicillin 78, 82
Bicuculline 195, 196, 199
Anxiolytic drugs 195
Am-A 30-2, 33 Binding constants 3, 275-82
Bis-chromone dicarboxyhc acid 240
Am-ATP 31, 33
B-Blockers see B-Adrenergic receptor
Ara-C 32, 33
blocker
Ara-CTP 32
Bradykinin 14 1
Arabinosyladenine (At-a-A) 30-2
Bromocryptine 167-9
Arabinosylcytosine (Am-C) 32, 33 Bronchoconstriction 240
Arabinosylhypoxanthine 32
Bufotenin 173
Arachidonic acid 123, 124, 125, 129, Burimamide 208, 209
130, 133, 157
Archaebacteria 65 Caffeine 209
Aromatase 109- 10 Calcium channel 218-9, 225-30
Arrhythmias 163, 218, 219, 227, 229 Calcium channel antagonists 227-30
Arthritis 11 Calcium channel blockers 218
rheumatoid 43, 130- 1 Calmodulin 126, 158, 206, 241
Ascaridia galli 257 Cancer 2, 11, 13, 30, 201, 244
Ascarik mum 199, 258 breast 272
Aspergillus terreus 10 1 of head and neck 256
Aspirin 1, 127, 128, 129, 130, 201 prostate 108, 272
Asthma 124, 126, 132, 161, 164, 239-41 see aZso Carcinoma; Lymphoma;
Atenolol 163 Tumour
Atherosclerosis 100, 266 Candida albicans 6, 103, 244, 245, 247
Athlete’s foot 257 Candidosis 26
ATP 23, 33, 34, 45, 51, 151, 155, 156, Captopril 145, 146, 147, 148
215, 223, 227, 229, 231, 233, Carbidopa 168
234, 236, 237 Carbonic anhydrase 135, 136-41
ATPases 230, 231-9 inhibitors 136, 138
potassium/hydrogen 236-9 Carbonic acid 136, 137
 Index 287

Carboxypeptidase A 143 Constipation 205

Carcinoma Contraceptive pill 99
of the breast 26, 34, 109, 114-5, 272 Cornea, keratinization of 31
of the gastrointestinal tract 26 Coronary heart disease 97, 100, 243
oatcell, of the hmg 34 Corticosteroids 126
of the ovary 34 Corticosterone 108
see &so Cancer; Lymphoma; Tumour Cortisol 97, 108
Cardiac glycosides 234-6 Coxsackie B virus 261
Cataracts 266 Cromoglycate 239-41
Catechol 0-methyltransferase (COMT) Cryptococcosis 26
153, 164, 184 Cryptorchidism 272
Catecholamines 153, 158, 159, 163, 188, Gushing’s syndrome 107
229 Cyclic adenosine 3’,5’-monophosphate
Catbarantbus 255 see CAMP
Ceftazidime 79, 80, 84 Cyclic nucleotide phosphodiesterase 159
Cefoxitin 61, 69, 79, 80, 84 Cyclooxygenase reaction 123
Cefuroxime 79, 84, 85 Cytidine deaminase 32
Cephalosporin C 78, 79 Cytidine triphosphate 18
Cephalosporins 54, 75, 77-8, 82, 85 Cytochrome P450 70, 100, 103, 104,
Cepbalosporium acremonium 78 107
Cephalothm 78 Cytochome P450 difference spectra,
Cephamycin C 78 ultraviolet 105, 107
‘Cheese effect’ 4 Cytochrome P-450 reductase 104
Chloramphenicol 2, 51, 54, 57-9, 61, 62 Cytosine deaminase 26
Chlordiazepoxide 195
Chloride channel and antagonists 222 Dapsone 20
Chloroquine 103 Deoxy-CTP 32
Chlorpheniramine 207, 208 Deoxy-TMP 16
Chlorpromazine 170 Deoxy-UMP 16, 21
Chlorpropamide 266, 267 Deoxyadenosine 41
Chlortetracycline 59 2-Deoxycoformycin 5, 14, 32, 41
Cholera toxin 206 11 PDeoxycorticosterone 108
Cholesterol 95, 96, 97, 100, 101, 215, Deoxycortisol 108
217, 245, 246, 247, 269 Deoxycytidine kinase 32
Cholesterol 20,22-lyase 109 Deoxycytidylate kinase 32
Choriocarcinoma 21 Deoxyinosine 41
Chymotrypsin 192 Deoxyribonucleic acid see DNA
Cimetidine 208, 209, 210, 237 Deoxyribose 15
Ciprofloxacm 36 L-Deoxystreptamine 54
Clauiceps purpurea 167 Deoxythymidine monophosphate 16, 21
Clavulanic acid 5, 83, 84 Deprenyl 184, 185
Clindamycin 54, 60, 61-2, 69 Depression 175, 176, 183
Clomiphene 115 Diabetes mellitus 261
Clonidine 161, 162 Diacylglycerol 157, 240
Clorgyline 184, 186 Diaminopyrimidines 2 1, 22
Clorobiocin 6 Diarrhoea 62, 205, 206
Clostridium dtfjficile 62, 65, 86 Diazepam 195, 199
C. pasteurianum 67 3H-Diazepam 196
C. perfringens 68 Dicyclohexylcarbodiimide (DCCD) 232
C. welcbii 67 Diethylcarbamazine 197
Coccidioides 247 Diethylstilboestrol 114, 272
Coformycin 41, 44 Digitalis 234
Colchicine 252-5, 256, 258, 259 Digitonin 161
Colitis, pseudomembraneous 60, 62 Digitoxin 234
288 Zndex
Digoxin 234, 235 Enkephalin 159
Dihydroorotate 17, 18, 19 Enterobacteriaceae 54
Dihydrofolate 21 Enzyme 2
Dihydrofolate reductase (DHFR) 4, 20, drugs binding to 3-5
21-3 inhibition 1
Dihydropteroate 16 irreversible 3, 4-5 mechanism-based
Dihydropteroate symhetase 19-20, 21 5
1,4Dihydropyridine drugs 227 reversible 3, 4 competitive 3
Dihydrostreptomycin 55 noncompetitive 3
3,4Dihydroxymandelaldehyde 183 tumour 11
Dihydroxyphenylacetic acids 171 uncompetitive 3
t.-Dihydroxyphenylalanine @-Dopa) 167, kinetics 3, 4
168 Enzyme-inhibitor complex 3
Diltiazem 227 Epipodophyllotoxins 34
Dipeptidylcarboxypeptidase 148 Epithelia 138, 230
Diphenylbutylpiperidines 170 leaky 223
Diphtheria 60 ‘tight’ 223
Dissociation constant 3, 7, 280 2,3-Epoxysqualene 103
Distribution coefficients 7 Ergocalciferol 97, 215
Diuretics 137, 223-5 Ergosterol 97
‘high ceiling’ 230 Erythromycin 54, 60, 62
loop 230 Escbericbia coli 55, 61
DNA 15, 26, 68 Ethacrynic acid 230, 231
biosynthesis 13, 23, 28-30, 68 Ethinyloestradiol 113
DNA gyrase 36-7, 68 Euphoria 148
DNA polymerase 16, 28, 32-3
mammalian 32-3 FAD 103
herpesvirus 30 Fatty acids 8
DNA topoisomerase 33-6 Fenoldopam 168
r.-Dopa 168 Ferredoxin (Fd) 65
Dopamine 4, 153, 167-74, 183, 186 Fevers 127
agonists 167-9 Flagyl see Metronidazole
antagonists 167-74 Flubendazole 258-9
receptors 167-74 Flucytosine 247
Dosage regime 4 Fhmitrazepam 196
Double reciprocal plot 281 5-Fluorocytosine 26
Doxorubicin (adriamycin) 34 5-Fluorodeoxyuridylate 5
Doxycycline 59 5-Fluorouracil (5~FU) 25
Drug nomenclature 8-9 5-Fluorouridine diphosphate (5-FUDP) 25
Drug/receptor binding, quantigcation of 5-Fluorouridine monophosphate (5-
278-82 FUMP) 25
Dynorphin A 203 Folic acid 16
Dysentery, amoebic 65 Follicle-stimulating hormone (FSH) 112,
Dyskinesia, tardive 172 115, 204, 268, 272
Dysphoria 205 Formyhnethionyl-tRNA 52
l(%Formyhetrahydrofolate 16
Ehrlich ascites tumour cells 28 Fungicidal drugs 6
Eicosatrienoic acids 99 Furosemide 225, 230, 231
Elastase 192
Enalapril 147 GABA 194, 195, 196, 197, 199, 200, 217
Enalaprilic acid 145, 146, 147 GARA aminotransferase 196
Encephalitis 3 1 GM& 195, 196, 199
Endopeptidase 142, 144 Galactorrhoea 173
Endorphin 159, 148 g-Galactosidase 264
 Index 289

Gastrin 237, 238 Hodgkin’s lymphomas 256

GDP 156, 166, 252 Homovanillic acid 171
Glaucoma 139, 141 Human immunodeficiency virus (HIV) 38
Glibenclamide @Iyburide) 266-7 Human T-celi lymphotropic virus III
GIipizide 266-7 (HTLV-III) 38
Glucagon 265-6 Hydrogenosome 6
Glucocorticoids 97 @Hydroxy-@-methylglutarylCoA (HMG
Gluconeogenesis 97, 262 CoA) reductase 100-2
Glutamate decarboxylase 194 Hydroperoxidase 129
Y-Glutamyl peptidase 133 5-Hydroperoxy-7, 9, I I, 14
Glutathione S-transferase 133 eicosatetraenoic acid 133
Glyburide 266 15-Hydroperoxy-9, 1 l-endoperoxide
Glycerol phosphate 54 (prostagIandin G2) 129
Glycinamide ribotide 16 g-Hydroxybutyrate 262
Glycogen 264 Hydroxyeicosatrienoic acids 123
Glycogenolysis 262, 265 l-(2’-HydroxyethyD-2-methyl-5-
Glycopeptide 54 nitroimidazole see Metronidazole
GMP 28 9-[2-Hydroxyethoxy)methylguanine see
cGMP 241 Acyclovir
GMP synthetase 28 5-Hydroxyindoleacetaldehyde 183
Gonadorelin see Gonadotrophin 5-Hydroxyindoleacetic acid 182
hormone releasing hormone 17cx-Hydroxyprogesterone 107-8
GnW 5-Hydroxytryptamine 5 HT see Serotonin
Gonadotrophin hormone releasing Hydroxyurea 24
hormone (GnRH) 261 Hyperglycaemia 262, 266
Gonadotrophins 112, 169, 204, 261, 268 Hyperpolarization 195, 199, 200, 231
Gout 43, 44, 252 Hypertension 146
Griseofulvin 252, 256-7 Hypogonadism 272
GTP 26, 28, 52, 53, 60, 155, 202, 252 Hypoxanthine 18, 43
GTPase 204, 254 Hypoxanthine
Guanosine monophosphate see GMP phosphoribosylpyrophosphate 43
Guanylate cyclase 45-6 Hypoxanthine phosphoribosyl-transferase
Gynaecomastia 107, 108, 119 (HPRT) 43-5
Hyoscine 189
Haemopbihs inji’uenzae 57, 59
HaIlucinations 7 Imidazoles 2 10
HaIoperidol 170 Imipramine 181
Hay fever 207 Immune suppressant drugs 6
Headaches 127 IMP see Inosine monophosphate
Heart beat 217-19 Indomethacin 128, 129
Heroin 168, 187 Intluenza 26
Herpes roster 30, 31 Influenza virus 28
Herpesvirus 26, 30, 31 Inosine 18, 41
Hexane 7 Inosine monophosphate (IMP) 28
HilI plot analysis 275, 281-2, 283 Inosine monophosphate dehydrogenase
Hirsutism 119 26-8
Histamine 153, 208, 237, 238, 239, 240, Inositol l-phosphate 157, 192
241 Insulin 261-6, 267
Histamine receptors 206-210 Interleukin 1, 130
II1 receptor antagonists 207-8 InterstitiaI cell stimulating hormone
Hz receptor antagonists 208-10 (ICSH) 268, 269
Histophsma capsdatum 247 Iodoacetamide 278
HMG CoA reductase 101 Ion charnel blocking 1
HMG CoA synthase 101 Isoprenaline 159, 163, 164
 290 Index

Ivermectin 197, 198 Lymphadenopathy-associated virus (LAV)

38
Kat inhibitors 5 Lymphoma
Kala-azar (leishmaniasis) 44 Hodgkin’s 256
Kaolin 205 non-Hodgkin’s 32
Kaposi’s sarcoma 38 Lysergic acid diethyIamide 167
Ketanserin 178 Lysozyme 76
Ketoconazole 104-7, 108
Kinetics, enzyme 3, 4, 275, 281 Malaria 14
Kininase 2 see Angiotensin converting Maprotiline 181
enzyme Measles 26
KlebsielZa pneumoniae 84 Mebendazole 257
Mecillinam 80
Labetalol 166 Meclofenamic acid 128-9
B-Lactamase 5, 58, 81-3 Melanoma 24
B-Lactams 54, 58, 75, 77-86 Membrane
Lanosterol 96, 104-6 fluidity 217
Lanosterol 14u-demethyIase 107 potential 2 18
Law of Mass Action 5 structure 2 15
Legionella pneunaopbila 60 transport 217, 218
Legionnaire’s disease 60 Meningitis 7, 57, 59, 247
Leprosy 20, 40 Meperidine 201, 205
Leucine aminopeptidase 135 Mepyramine 207, 208
Leukaemia 36 6Mercaptopurine 43
acute 32, 34 Mescaline 173
lymphocytic 41 Mestranol 113
lymphoblastic 41 Methadone 20 1, 204
myelogenous 32 Methazolamide 141
chronic granulocytic 24 Methicilhn 81
Leukotriene(s) 130, 135, 157 Methionine 52
biosynthesis 132 Methotrexate 2 l-2
Leukotriene Ad 133 binding 4
Leukotriene B* 130, 131, 133 3-Methoxytyramine 17 1
Leukotriene Cd 123, 133, 165 N-Methyl-N-nitro-N-nitrosoguanidine 68
Leukotriene D* 123, 133, 165 l-Methyl4phenyl-1,2,3,6tetrahydro-
Leukotriene Ed 133, 165 pyridine (MPTP) 168, 187-8
Leuprolide 271, 272 1-MethyI-4-phenyldihydropyridine
Lewis acids 136, 143 (MPDP) 188
Librium 195 Ma”‘-MethyIene-tetrahydrofoIate 25-6
Lidocaine 221, 222, 223 2-Methylhistamine 208
Ligand receptor complex 3 4Methylhistamine 208
Ligand/protein binding, quantiBcation of 5-Methyluracil 18
280-3 Metiamide 208
Lincomycin 61, 62 Metoprolol 163
Lineweaver-Burk plots 4 Metronidazole 9, 61, 65-70
Lipocortin 126 Mevalonate 96
Lipolysis 97, 262, 265 Mevinolin 10 1
Lipophilicity 8, 22 1 Mianserin 178
PLipotrophin 202 Michaehs-Menten analysis 227, 278
5-Lipoxygenase 123, 133 Micromonospora spp. 257
12Lipoxygenase 123 Microtubule-associated proteins (MAPS)
Loperamide 205-6 251-2, 256
Luteinizing hormone (LH) 112, 115, 205, Mifepristone 116
268-73 Migraine 176-8
 Index 291
Mineralocorticoids 97, 117 Oestrogen 114-5, 269
Minimum inhibitory concentration Oestrogen antagonists 114
(M.I.C.) 6 Oestrogen receptor agonists 111
Minocycline 59 Oestrone 109, 269
Monoamine oxidase (MAO) 152, 183-8 Oligomycin 232
Monoamine oxidase A 180 Omeprazole 238-9
Monoamine oxidase inhibitors 4, 5 Oncbocerca uolvulus 197
Monobactams 78 Onchocerciasis 197
Monoiodoamphenicol 57 Opiate receptors 201-4
Monosodium urate 43 Opiates 201
Morphine 201-5 Opioid peptides 148, 174
Murein 54 Opioids 201, 205
Muscarine 189 Orotate 17
Muscarinic agonists 191-2 Orotate phosphoribosyltransferase 14, 25
Muscarinic antagonists 190 Orotidine 5-phosphate 17
Muscarinic receptors 189 Osteoarthritis 13 1
Muscimol 199 Ouabain 234
Myasthenia gravis 193 Over-active bowel syndrome 205
Mycosamine 245 Oxfendazole 257
Mycoplasma artbritidis 13 1 Oxipurinol 44
M. pneumoniae 60 Oxytetracyclin 59
Myocardial infarction 132, 229
Pacemaker current 2 18
NADH 103 Paracoccidioides brasiliensis 247
NADH oxidase 68 Pargyline 183
NADPH 21, 23, 67, 70, 103, 104, 107, Parkhtsonism 167-9, 172, 187-9
108 Partition coefftcient 7-8
Naftifine 103 Passive transport 7
Nalidixic acid 36, 37 Penicillin-binding proteins (PBP) 78, 85
Naloxone 202, 204 Penicillins 1, 8, 60, 75, 78
Naproxen 128, 129 Penicillium notatum 78
Negative cooperativity 5 P. gri3eofulvium dierckx 256
Neostigmine 194 Penicilloic acids 82
Neuroleptic drugs 172-3 Pentagastrin 238
Neurospora crassa 2 17 Pentaglycine 76
Nicotine 189 Pentapeptide 76
Nicotimc receptors 189 Pentazocine 204-5
Nifedipine 227, 228, 229 Pentylenetetrazole 196
Nimodipine 228-9 Peptidoglycan 75-8
5-Nitroimidazole 68 Peptidoglycan endopeptidase 54, 78
Non-Hodgkin’s lymphoma 30, 256 j?-Phenethylamine 187
Non-steroidal anti-inflammatory drugs 6 Phentolamine 162
Noradrenaline 153, 183, 188 Phenylbutazone 128
Norethisterone 114 Phenylethylamine 184
Norgestrel 114 Phorbol diester 158
Novobiocin 37 Phosphatidylethanolamine 246
Nucleoside diphosphate kinase 32 Phosphatidylcholine 245
Nystatin 244 Phosphatidylinositol 157, 208
Phosphodiesterase 140
Woctanol 7 cyclic nucleotide 159
Oedema 131, 266 Phosphofructokinase 277
Oestradiol 97, 98, 270 Phospholipase AZ 155
17mestradiol 98, 108, 115, 269 Phospholipase C 155, 157, 190, 263
Oestriol 269 Phospholipase inhibition 155
 Index 293
Semicarbazide 195 Sulphonamides 8, 22
Serotonin 153, 180-4 Sulphonylureas 266-8
receptors 174-5 Suramin 197
Sex steroids 100
Similium 197 Tamoxifen 110
Skin ahergies 207 Teichoic acids 54
‘Slow reacting substance of anaphylaxis’ Terbintime 103
see Leukotriene Cd: Leukotriene Testosterone 88, 107-l 1, 269, 272
D4 Tetracychnes 49-50, 54, 59-60
Sodium charmel 218, 219-20 Tetrabymena pyriyormis 25 1
Sodium channel blockers 218, 220-22 Therapeutic ratio 2
Sodium ion channels 230 Thiabendazole 258
Sodium urate 43 Thiamphenicol 58
Sodium zincate 136 Thiazides 139
Sodium/chIoride ion channels, coupled Thioredoxin 23
230 Thioredoxin reductase 23
Somatostatin 265 6Thiouric acid 43
Spironolactone 119, 225 Thromboxane AZ 123, 132
3H-Spiroperidol 167 Thromboxane antagonists 132
Squalene 96 Thromboxane Bz 132
Squaiene epoxidase 95 Thromboxane synthetase 132
Squalene epoxide cyclase 99 Thromboxanes 123
StaphylococcaI species, methicihin- Thymidine 31
resistant 86 Thymidine kinase 31
Staphylococcus aureus 76, 78, 81, 84, Thymidine monophosphate 18, 21
86 Thymidylate synthetase 5, 21, 25-6, 27
Steroid 17, 20 lyase 107-8 Tiotidine 109
1 l&Steroid hydroxylase 108-9 Tobramycin 54, 55
Steroids 95-100 Tolbutamide 267
Az4-Sterol reductase 95 Tolnaftate 104
Streptidine 54 Topoisomerase II (mammalian) 34-6
Streptomyces Tranquillisers 195
S. antibioticus 41 TransgIycosylase SO
S. aureofaciens 59 Transition state inhibitors 5
S. auermitilis 197 Transpeptidase 78-80
S. clavuligerus 83 Tranylcypromine 187
S. erytbreus 60 TriacylgIycerols 97
S. lactamdurans 78 Triamterene 223, 225
S. lincolnensis 61 Tricbomonas vaginals 65, 67, 68
S. orientalis 86 Tricbopbyton 103, 257
S peucetius var. caesius 34 T. mentagropbytes 257
S. pneumoniae 60 Tricbostrongylus colubriformis 258
S. rimosus 59 Tricyclic antidepressants 180
S. venezuelae 57 TrigIyceride lipase 156
Streptomycetes 54 Trimethoprim 21, 22
Streptomycins 55, 57 Trimethylcolchicinic acid 254
‘Substoichiometric poisoning’ 255 Tripelennamine 204
Succinate 66 1,4,5-Triphosphate 157
Succinyl coenzyme A 66 Triticbomonas foetus 67
Succinylproline 145 Trypsin 192
Suicide substrates (inhibitors) 5 Tryptamine 184
Sulphadiazine 22 Tryptophan 8
Mphadoxine 22 TubercuIosis 183
Sulphamethoxazole 22 Tubocurarine 189
294 Index
TubuIin 25 l-8 Vanadate 232
Tumours 109 Vancomycin 76, 86-7
pituitary 167, 169 Vancosamme 86
of prostate gland 109, 110 Verapamil 228, 229
trophoblastic 21 Vidarabine see Ara-A
Tyramine 4 Vinblastine 254-5, 256
Vinca alkaloids 252, 255-6
UDP 76 Vinca rosea 255
UDP-N-acetylglucosamine 76 Vincristine 255-6
UMP pyrophosphorylase 26 Virazole see Ribavirin
Uracil 18
Urethritis 60 Xanthine 18
Uric acid 43 Xanthine oxidase 5, 43-5
Uridine diphosphate (UDP) 76 Xanthosine monophosphate 28
Uridine monophosphate (UMP) 17 Xenobiotics 8
Uridine 5-monophosphate 17
Zinc diamine chloride 136
Valium 195 Zinc metallopeptidase 142
van der Waals interactions Zovirax see Acyclovir