Theoretical Biology and Medical

Modelling BioMed Central

Research Open Access

Dynamic simulation of red blood cell metabolism and its application
to the analysis of a pathological condition
Yoichi Nakayama, Ayako Kinoshita and Masaru Tomita*

Address: Institute for Advanced Biosciences, Keio University, Tsuruoka, 997-0017, Japan
Email: Yoichi Nakayama - ynakayam@[Link]; Ayako Kinoshita - ayakosan@[Link]; Masaru Tomita* - mt@[Link]
* Corresponding author

Published: 09 May 2005 Received: 19 November 2004

Accepted: 09 May 2005
Theoretical Biology and Medical Modelling 2005, 2:18 doi:10.1186/1742-4682-2-18
This article is available from: [Link]
© 2005 Nakayama et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License ([Link]
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

kineticsmetabolism

Abstract
Background: Cell simulation, which aims to predict the complex and dynamic behavior of living
cells, is becoming a valuable tool. In silico models of human red blood cell (RBC) metabolism have
been developed by several laboratories. An RBC model using the E-Cell simulation system has been
developed. This prototype model consists of three major metabolic pathways, namely, the
glycolytic pathway, the pentose phosphate pathway and the nucleotide metabolic pathway. Like the
previous model by Joshi and Palsson, it also models physical effects such as osmotic balance. This
model was used here to reconstruct the pathology arising from hereditary glucose-6-phosphate
dehydrogenase (G6PD) deficiency, which is the most common deficiency in human RBC.
Results: Since the prototype model could not reproduce the state of G6PD deficiency, the model
was modified to include a pathway for de novo glutathione synthesis and a glutathione disulfide
(GSSG) export system. The de novo glutathione (GSH) synthesis pathway was found to compensate
partially for the lowered GSH concentrations resulting from G6PD deficiency, with the result that
GSSG could be maintained at a very low concentration due to the active export system.
Conclusion: The results of the simulation were consistent with the estimated situation of real
G6PD-deficient cells. These results suggest that the de novo glutathione synthesis pathway and the
GSSG export system play an important role in alleviating the consequences of G6PD deficiency.

Introduction ear PDE/ODE/Algebraic systems that can represent the

Many attempts have been made to simulate molecular cellular geometry; and DBsolve [3], which combines con-
processes in cellular systems. Perhaps the most active area tinuation and bifurcation analysis.
of cellular simulation is the kinetics of metabolic path-
ways. Various software packages that quantitatively simu- The E-Cell project [4,5], which aims to model and simu-
late cellular processes and are based on numerical late various cellular systems, was launched in 1996 at Keio
integration of rate equations have been developed. These University. The first version of the E-Cell simulation sys-
include GEPASI [1], which calculates steady states as well tem, a generic software package for cell modeling, was
as reaction time behavior; V-Cell [2], a solver of non-lin- completed in 2001. E-Cell version2, which is a Windows

Page 1 of 11
(page number not for citation purposes)
Theoretical Biology and Medical Modelling 2005, 2:18 [Link]

version of the first E-Cell system, is now also available [6]. these simulation models consisted only of glycolysis and
E-Cell version 3, which enables multi-algorithm simula- the pentose phosphate pathway. We found that including
tion, is the latest version [7]. The E-Cell system allows the the glutathione (GSH) biosynthesis pathway and the glu-
user to define spatially discrete compartments such as tathione disulfide (GSSG) export system, which are
membranes, chromosomes and the cytoplasm. The collec- involved in suppressing oxidative stress, improved the
tions of molecules in all cellular compartments are repre- ability of the model to reflect the real diseased RBC. This
sented as numbers of molecules, which can be converted suggests that these pathways may compensate for the con-
to concentrations, and these can be monitored and/or sequences of G6PD deficiency in human RBCs.
manipulated by employing the various graphical user
interfaces. In addition, the E-Cell system enables the user Methods
to model not only deterministic metabolic pathways but Development of the prototype model and simulation
also other higher-order cellular processes, including sto- experiments
chastic processes such as gene expression, within the same The E-Cell system version 1.1 was used as the simulation
framework. By using the E-Cell system, a virtual cell with platform in this work. The software can be downloaded
127 genes that are sufficient for "self-support" [4] was from [Link] Our prototype model of the
developed. This gene set was selected from information RBC was developed on the basis of the whole-cell model
about Mycoplasma genitalium genomic sequences and of Joshi and Palsson [8-11] with slight modifications (Fig-
includes genes for transcription, translation, the glycolysis ure 1). We modified the model to represent the oxidant-
pathway for energy production, membrane transport, and induced decrease of hexokinase and pyruvate kinase, and
the phospholipid biosynthesis pathway for membrane the maximum activity of these enzymes was allowed to
production. change according to the ratio of GSH and GSSG. The
equations and parameters used are derived from the liter-
On the basis of existing models of single pathways and ature [17]. The parameters and kinetic equations in the
enzymes, various in silico models of human red blood cell original model of Joshi and Palsson were replaced with
(RBC) metabolism were first developed by Joshi and Pals- those obtained from the literature [17,20,21] (Table 1) in
son [8-11]. Subsequently, other groups developed RBC order to fit the model to the measured concentrations dur-
models [12-15]. The RBC is thought to be a good target for ing the calculation of the steady state. The steady state
biosimulation because extensive studies over the last three obtained had concentrations of many metabolites that
decades have generated extensive biochemical data on its were very close to those in real RBCs (Table 2). However,
enzymes and metabolites. Moreover, the RBCs of many the concentrations of several metabolites, namely adeno-
species, including humans, do not contain a nucleus or sine, hypoxanthine, inosine, 5-phosphoribosyl 1-phos-
carry genes. This means that gene expression can be phate and ribose 1-phosphate, differed from the
excluded from the model, which greatly simplifies the experimental values. These differences were due to the
biosimulation. RBCs take up glucose from the plasma and kinetic parameters and equations used, and because the
process it by glycolysis, which generates the ATP mole- nucleotide metabolism in the original model was repre-
cules that are used in other cellular metabolic processes. sented as simple first-order kinetics or equilibrium.
The ATP molecules are mostly consumed by the ion trans-
port systems that maintain the osmotic balance of the cell. The parameters from the work of Jacobasch et al. [30] were
used in the experiments simulating G6PD deficiency
Here we describe our computer model of the human RBC, (Table 3). Since the rate equation of G6PD deficiency is
which we developed on the basis of previous models [8- the same as that in the normal cell, the parameters were
13]. Our prototype model of the human RBC consisted simply replaced in the deficiency experiment. We adopted
only of glycolysis, the pentose phosphate pathway, nucle- the We.G variant of G6PD deficiency because its parame-
otide metabolism and simple membrane transport sys- ters are well described in the literature and its phenotype
tems such as the Na+/K+ antiport channel. Here, we have is rather severe. As with the original model, the oxidative
employed this prototype model to reproduce the patho- load is represented as the conversion of GSH to GSSG, and
logical condition of glucose-6-phosphate dehydrogenase the equation is expressed as a simple first-order kinetics.
(G6PD) deficiency. This is the most common hereditary
enzyme deficiency in RBCs; it causes anemia, and more Expansion of the prototype model and simulation
than 400 varieties of G6PD deficiency have been identi- experiments
fied [16]. The deficiency is known to exert only mild The de novo GSH synthesis and GSSG export pathways
effects as it does not cause clinically significant problems (Figure 3) were added to the prototype model. The kinetic
in most cases, except upon exposure to medications and equations and parameters of these pathways were
foods that cause hemolysis. Computer simulations for obtained from the literature [31-33] (Table 4). Since these
analyzing this deficiency have been reported [17-19], but pathways have very low activity in normal cells, the

Page 2 of 11
(page number not for citation purposes)
Theoretical Biology and Medical Modelling 2005, 2:18 [Link]

Figure 1 map of the prototype RBC model

Metabolic
Metabolic map of the prototype RBC model. The circles are metabolic intermediates and ions. These molecular species
are defined as "Substance" in the E-Cell system. The boxes are enzymes and reaction processes. Their rate expressions are
defined as "Reactor" whereas the enzyme molecules are defined as "Substance".

concentrations of metabolites at the steady state were Results and Discussion

almost unchanged in the expanded model. The concentra- Simulation of G6PD deficiency using the prototype model
tions listed in Table 2 were used as the steady state concen- The prototype model was used to simulate the effects of
trations. The conditions employed to simulate G6PD G6PD deficiency. G6PD is a key enzyme in the pentose
deficiency using this expanded model were the same as phosphate pathway that converts glucose 6-phosphate
those of the prototype model. It is known that multidrug into gluconolactone 6-phosphate (GL6P); this simultane-
resistance-associated proteins (MRP1) and the cystic ously generates NADPH. The metabolic intermediate
fibrosis transmembrane conductance regulator (CFTR) GL6P is then metabolized into ribulose 5-phosphate
are expressed in human RBC and involved in GSH and/or (Ru5P) acid via gluconate 6-phosphate (GO6P). This
GSH conjugates transport [35]. However, their rate equa- process also generates NADPH. This reduction power is
tions and parameters are unavailable, so these proteins employed by various other intracellular processes, in par-
were not included in this model. ticular the reduction of GSSG. A major function of GSH in

Page 3 of 11
(page number not for citation purposes)
Theoretical Biology and Medical Modelling 2005, 2:18 [Link]

Table 1: Enzymes and rate equations of the prototype model

Enzymes Abbreviation Group Reaction mechanism Reference

Glutathione turnover GSHox PPP Chemical reaction 24

Glutathione reductase (NADPH) GSSGR PPP Ordered Bi Ter mechanism 24
Glutathione reductase (NADH) GSSGR2 PPP Michaelis Menten mechanism 24
Glucose 6-phosphate dehydrogenase G6PD PPP Ordered Bi Bi mechanism 17
6-Phosphogluconolactonase 6PGLase PPP Michaelis Menten mechanism 17
6-Phosphogluconate dehydrogenase 6PGLDH PPP Ordered Bi Ter mechanism 24
Ribose 5-phosphate isomerase R5PI PPP Uni Uni mechanism 25
Xylulose 5-phosphate isomerase X5PI PPP Uni Uni mechanism 25
Transketolase I TK1 PPP Ping-Pong Bi Bi mechanism 25
Transketolase II TK2 PPP Ping-Pong Bi Bi mechanism 25
Transaldolase TA PPP Ping-Pong Bi Bi mechanism 25
Hexokinase HK Glycolysis 26
Phosphoglucoisomerase PGI Glycolysis Uni Uni mechanism 25
Phosphofructokinase PFK Glycolysis 27
Aldolase ALD Glycolysis Ordered Uni Bi mechanism 25
Triose phosphate isomerase TPI Glycolysis Uni Uni mechanism 25
Glyceraldehyde phosphate dehydrogenase GAPDH Glycolysis Chemical reaction 20
Phosphoglycerate kinase PGK Glycolysis Chemical reaction 20
Diphosphoglycerate mutase DPGM Glycolysis Michaelis Menten mechanism 20
Diphosphoglycerate phosphatase DPGase Glycolysis Michaelis Menten mechanism 20
Phosphoglyceromutase PGM Glycolysis Chemical reaction 20
Enolase EN Glycolysis Chemical reaction 20
Pyruvate kinase PK Glycolysis 28
Pyruvate transport process PYRtr Transport Michaelis Menten mechanism 22
Lactate dehydrogenase LDH Glycolysis Chemical reaction 20
Lactate transport process LACtr Transport Michaelis Menten mechanism 22
Leak of Potassium K_Leak Transport 9
Leak of Sodium Na_Leak Transport 9
Sodium/potassium pump Pump Transport 9
Adenosine transport process ADEtr Transport Chemical reaction 13
AMP phosphohydrolase AMPase NM Chemical reaction 20
Adenosine deaminase ADA NM Michaelis Menten mechanism 20
Adenosine kinase AK NM Michaelis Menten mechanism 20
Adenylate kinase APK NM Chemical reaction 20
Adenosine triphosphate phosphohydrolase ATPase NM Chemical reaction 8
Adenosine monophosphate deaminase AMPDA NM Michaelis Menten mechanism 20
Inosine monophosphatase IMPase NM Michaelis Menten mechanism 8
Purine nucleotide phosphorylase PNPase NM Chemical reaction 23
Phosphoribosyl pyrophosphate synthetase PRPPsyn NM 8
Adenine phosphoribosyl transferase ADPRT NM Michaelis Menten mechanism 8
Hypoxanthine-guanine phosphoryl transferase HGPRT NM Michaelis Menten mechanism 8
Hypoxanthine transport process HXtr NM 29

PPP, Pentose phosphate pathway; NM, Nucleotide metabolism.

the RBC is to eliminate superoxide anions and organic Ru5P produced was mainly converted to fructose 6-phos-
hydroperoxides. Peroxides are eliminated through the phate (F6P), and this metabolite consumed ATP to make
action of glutathione peroxidase, which yields GSSG. fructose 1,6-diphosphate (FDP). The FDP production led
to an accumulation of dihydroxy acetone phosphate
The simulation experiments were carried out with steady (DHAP), and the metabolite was not used to provide ATP.
state concentrations corresponding to those in the normal The high GO6P concentration could sustain normal levels
RBC. Sequential changes in the quantities of NADPH, of GSH concentration at the first stage of the simulation,
GSH and ATP were observed (Figure 2). There is a negative but after the depletion of GO6P the rate of Ru5P produc-
peak in ATP concentration before 10 h. This was due to tion was drastically reduced. This decrease in Ru5P con-
the shutting down of the pentose phosphate pathway. The centration led to decreased F6P concentrations.

Page 4 of 11
(page number not for citation purposes)
Theoretical Biology and Medical Modelling 2005, 2:18 [Link]

Table 2: Steady state of the RBC model.

Concentration (mM)

Metabolic intermediate Abbreviation Steady stateb Literaturec

1,3-Diphosphoglycerate 13DPG 1.83E-04 4.00E-04

2-Phosphoglycerate 2PG 4.16E-03 1.40E-02 ± 5.00E-03
3-Phosphoglycerate 3PG 4.62E-02 4.50E-02
Adenosine ADO 8.93E-06 1.20E-03 ± 3.00E-04
Dihydroxy acetone phosphate DHAP 1.35E-01 1.40E-01 ± 8.00E-02
Erythrose 4-phosphate E4P 1.17E+00 -
Fructose 6-phosphate F6P 6.39E-02 1.60E-02 ± 3.00E-03
Fructose 1,6-diphosphate FDP 1.14E-02 7.60E-03 ± 4.00E-03
Glucose 6-phosphate G6P 1.96E-01 3.80E-02 ± 1.20E-02
Glyceraldehyde 3-phosphate GA3P 6.24E-03 6.70E-03 ± 1.00E-03
Gluconolactone 6-phosphate GL6P 7.62E-06 -
Gluconate 6-phosphate GO6P 2.72E+00 -
Glutathione GSH 3.21E+00 3.21E+00 ± 1.50E+00
Glutathione GSSG 1.03E-04 -
Hypoxanthine HXi 9.32E-06 2.00E-03
Inosine monophosphate IMP 5.03E-03 1.00E-02
Inosine INO 3.32E-08 1.00E-03
Potassium Ki 1.26E+02 1.35E+02 ± 1.00E+01
Lactate LACi 1.20E+00 1.10E+00 ± 5.00E-01
Nicotinamide adenine dinucleotide NAD 8.87E-02d -
Nicotinamide adenine dinucleotide NADH 3.13E-04d -
Nicotinamide adenine phosphate NADP 8.06E-05d -
Nicotinamide adenine phosphate NADPH 6.58E-02d 6.58E-02
Sodium Nai 2.27E+01 1.00E+01 ± 6.00E+00
Phosphoenolpyruvate PEP 1.89E-02 1.70E-02 ± 2.00E-03
5-Phosphoribosyl 1-phosphate PRPP 6.91E-05 5.00E-03 ± 1.00E-03
Pyruvate PYRi 6.00E-02 7.70E-02 ± 5.00E-02
Inorganic phosphate Pi 1.30E-01 1.00E+00
Ribose 1-phosphate R1P 2.12E-05 6.00E-02
Ribose 5-phosphate R5P 2.81E-04 -
Ribulose 5-phosphate RU5P 1.48E-04 -
Sedoheptulose 7-phosphate S7P 7.49E-02 -
Xylulose 5-phosphate X5P 4.30E-04 -
2,3-Diphosphoglycerate 2,3-DPG 4.21E+00 4.50E+00 ± 5.00E-01
Adenosine diphosphate ADP 2.20E-01 2.70E-01 ± 1.20E-01
Adenosine monophosphate AMP 2.42E-02 8.00E-02 ± 9.00E-03
Adenosine triphosphate ATP 1.57E+00 1.54E-00 ± 2.50E-01

The values are given in scientific notation; E-01 denotes multiplication by 10-1.
aThe initial data set was from experimental data in the literature and from predictions of previous simulation models [12].
bThe simulation was run for more than 1,000,000 seconds in simulation time until the model reached steady state.
c Biochemical experimental data taken from the literature and reported in Joshi and Palsson [11].
d NAD(H) and NADP(H) pools are kept constant.

Table 3: Parameters for normal and deficient enzymes

t/2 (day) Vmax (mkat/l cells) KmG6P KmNADP (mM) KiNADPH KiATP Ki2,3DPG

Normal 27 575 67 3.7 3.1 749 2289

We.G. 2.5 10 152 3.8 0.62 180 520

These values are based on experimental data taken from the literature [10]

Page 5 of 11
(page number not for citation purposes)
Theoretical Biology and Medical Modelling 2005, 2:18 [Link]

A B C

D E F

Figure 2for the de novo of GSH and the GSSG export system
Pathway
Pathway for the de novo of GSH and the GSSG export system. γ-GCS, γ-glutamyl cysteine synthetase; γ-CS, γ-glutamyl
cysteine.

At around 20 h, ATP was rapidly consumed and depleted. After the expansion of the prototype model to include de
Since ATP concentrations less than half the normal con- novo GSH synthesis and GSSG export, the ATP levels were
centration have never been observed in enzyme deficien- maintained at 80% of normal and the cell was longer
cies [36], cells in this condition will probably be lived (Figure 4). In addition, the GSH/GSSG ratio was
destroyed. Although the half-life of the real G6PD-defi- higher (Figure 5). This indicates that the de novo GSH syn-
cient We.G type RBC is known to be 2.5 days [30], the lon- thesis pathway can partially compensate for the lowered
gevity of our computer model turned out to be much GSH concentrations resulting from G6PD deficiency, and
shorter (Table 3). Since data on the concentration of that the concentration of GSSG can be kept at a very low
metabolites in RBCs with G6PD deficiency are not availa- level due to the active export system. These observations
ble, it was not possible to determine whether the metabo- suggest that these reactions could alleviate the anemia
lite concentrations arising in our simulation experiments resulting from G6PD deficiency. It is known that people
reflected those observed in real cells. with this deficiency are not normally anemic and display
no evidence of the disease until the RBCs are exposed to
Simulation of G6PD deficiency using the expanded model oxidant stress. The compensatory effect of the de novo GSH
It is obvious that decreased pentose phosphate pathway synthesis and GSSG export pathways may thus help to
activity leads to faster cell death, and that the difference explain why many varieties of G6PD deficiency have no
between the simulated cell and the real cell regarding the evident phenotype. Moreover, it has been proposed that
timing of cell death could be caused by the lack of a path- the high frequency of G6PD deficiency may be due to its
way producing GSH. This pathway may compensate for ability to protect against malaria. Our observations sug-
the decrease in GSH. A mature RBC normally contains 2 gest that the compensatory mechanism we have eluci-
mM GSH but contains only several µM GSSG. Although dated may have aided this spread of G6PD deficiency, as
GSSG reductase plays a prominent role in maintaining a it counterbalances the worst effects of the deficiency, thus
stable GSH/GSSG ratio, other processes, including de novo decreasing its severity and promoting the propagation of
GSH synthesis and GSSG export pathways, may generate the disease during evolution.
GSH in the G6PD-deficient cell.

Page 6 of 11
(page number not for citation purposes)
Theoretical Biology and Medical Modelling 2005, 2:18 [Link]

Table 4: Rate equations and parameters of GSH synthesis and GSSG export that were used in the expanded model.

Rate equation for γ-glutamyl cysteine synthetase

Vmax[ATP][Glu][Cys]
α KmATP Km ’Glu KmCys
v=
[Glu] [Glu][Cys] [Glu][ATP] [Glu][Cys][ATP]
1+ + + +
Km ’Glu Km ’Glu KmCys Km ’Glu KmATP α Km ’Glu KmCys KmATP
( Ordered Ter Mechanism )
Parameters for γ-glutamyl cysteine synthetase
Parameter Value Reference
Vmax 141.57 mM/h 31, 32
α 0.2 31
Kmglu 1.8 mM 31
Kmcys 0.1 mM 31
KiGSH 3.4 mM 31
KmATP 0.4 mM 31

Rate equation for glutathione synthetase

Vmax[γ _ GC][Gly][ATP]
α Kmγ _ GC KmGly KmATP
v=
[γ _ GC] [γ _ GC][Gly] [γ _ GC][ATP] [γ _ GC][Gly][ATP]
1+ + + +
Kmγ _ GC Kmγ _ GC KmGly Kmγ _ GC KmATP α Kmγ −GC KmGly KmATP
( Ordered Ter Mechanism )
Parameters for glutathione synthetase

Parameter Value Reference

Kmγ_GC 0.99 mM 33
KmGly 1,37 mM 33
KmATP 0,23 mM 33
α 2.6 33
Vmax 88.4 mM/h 33

Rate equation for GSSG export

[GSSG] [MgATP]
v = Vmax1( )( )
GSSG + KmGSSG1 MgATP + KmATP
Parameters for GSSG export

Parameter Value Reference

KmGSSG1 0.1 mM 34
KmATP 0.63 mM 34
Vm1 20 µM/h 34

Determination of a range of metabolic pathways for by others contain only three metabolic pathways, namely,
modeling the glycolysis pathway, the pentose phosphate pathway
These results showed that the de novo GSH synthesis path- and the nucleotide metabolic pathway. Although these
way and the GSSG export system are essential for accurate models are sufficient for representing the normal state of
simulation of G6PD deficiency in human RBCs. Previous the human RBC, they are not adequate for simulating
simulations of this deficiency have not included these irregular conditions such as deficiencies, because they lack
pathways [17] and the results they generated were similar alternative pathways that may normally not be particu-
to those obtained using our prototype model (Figure 2). larly active but can compensate for the deficiency to some
Our prototype model and the previous models developed extent. Indeed, our results indicate that all the metabolic

Page 7 of 11
(page number not for citation purposes)
Theoretical Biology and Medical Modelling 2005, 2:18 [Link]

Figure 3 simulation time-course of metabolic intermediates

Computer
Computer simulation time-course of metabolic intermediates. Changes in the concentrations of ATP (A), GO6P (B),
GSH (C), GSSG (D), NADP (E) and NADPH (F) during the RBC simulation. The simulation was run for 200,000 seconds
(Approx. 55 h) in simulation time. Concentrations change when G6PD kinetic parameters are shifted from the normal to path-
ological values (Table 3). ATP became depleted at around 20 h.

pathways in the cell will be needed to develop a general The mathematical steady state may not be the normal
purpose model that can be used to simulate any condi- state of real cells
tion. However, dynamic simulation based on kinetic During this simulation analysis, we realized that the lon-
equations requires a large variety of rate equations and gevity of enzymes should be considered in long-term sim-
kinetic parameters, and unfortunately, such data are rarely ulation experiments. While in our model the activities of
available as a complete set. Recently, our laboratory enzymes are decreased by oxidants, enzymes also gener-
proposed a novel simulation method that reduces the ally become degraded over time. This natural decrease is
need for this kind of information [37]. This hybrid not included in our model. As shown in this work, the
dynamic/static simulation method combines dynamic prototype model was able to achieve a steady state. How-
rate equations with a flux-based approach and as a result ever, this mathematical steady state, which is when the
reduces the numbers of rate equations and parameters rates of the production and consumption of all metabolic
that are needed by up to 70–80%. It may solve the intermediates become equal, may not exactly represent
problems associated with developing a model that simu- the condition of the RBCs in the human body. Such a
lates all the cellular metabolic pathways. "mathematical steady state" never occurs in living organ-
isms, especially in higher multicellular organisms. Rather,

Page 8 of 11
(page number not for citation purposes)
Theoretical Biology and Medical Modelling 2005, 2:18 [Link]

A B C

E F
D

Figure 4 of G6PD deficiency using the expanded model

Simulation
Simulation of G6PD deficiency using the expanded model. Changes in the concentrations of ATP (A), GO6P (B), GSH
(C), GSSG (D), NADP (E) and NADPH (F) during RBC simulation. Broken lines are the results of the prototype model, while
solid lines are the results of the expanded model during the same parameter shift as described in Figure 2. The simulation was
run for 200,000 seconds (Approx. 55 h) in simulation time.

A B

Figure
The GSH/GSSG
5 ratio of the prototype and expanded models
The GSH/GSSG ratio of the prototype and expanded models. The prototype model (A) and the expanded model (B).

Page 9 of 11
(page number not for citation purposes)
Theoretical Biology and Medical Modelling 2005, 2:18 [Link]

homeostasis in multicellular organisms is maintained by 3. Goryanin I, Hodgman TC, Selkov E: Mathematical simulation and
analysis of cellular metabolism and regulation. Bioinformatics
replacing the loss of disposable cells with additional cells. 1999, 15:749-758.
It is possible that these disposable cells never reach a 4. Tomita M, Hashimoto K, Takahashi K, Shimizu TS, Matsuzaki Y, Miy-
mathematical steady state. Thus, a model that can tolerate oshi F, Saito K, Tanida S, Yugi K, Venter JC, Hutchison CA 3rd: E-
CELL: software environment for whole-cell simulation. Bioin-
long-term simulation for analyzing the pathology of formatics 1999, 15:72-84.
human diseases should not approximate the 5. Tomita M: Whole-cell simulation: a grand challenge of the
"mathematical steady state". Moreover, in the case where 21st century. Trends Biotechnol 2001, 19:205-210.
6. Takahashi K, Ishikawa N, Sadamoto Y, Sasamoto H, Ohta S, Shiozawa
the system reaches a steady state with a certain oscillation, A, Miyoshi F, Naito Y, Nakayama Y, Tomita M: E-Cell 2: Multi-plat-
it is impossible to obtain a mathematical steady state form E-Cell simulation system. Bioinformatics 2003,
19:1727-1729.
using an accurate model. It is known, for example, that 7. Takahashi K, Kaizu K, Hu B, Tomita M: A multi-algorithm, multi-
some key enzymes in glycolysis bind to the Band III pro- timescale method for cell simulation. Bioinformatics 2004,
tein, an abundant membrane protein in the human RBC 20:538-546.
8. Joshi A, Palsson BØ: Metabolic dynamics in the human red cell.
[38-40]. The interaction between glycolytic enzymes and Part I – A comprehensive kinetic model. J Theor Biol 1989,
Band III varies depending on the ratio of oxyhemoglobin 141:515-528.
9. Joshi A, Palsson BØ: Metabolic dynamics in the human red cell.
to deoxyhemoglobin, and it is believed that this interac- Part II – Interactions with the environment. J Theor Biol 1989,
tion is responsible for some oscillations in metabolic 141:529-545.
pathways in the human RBC. 10. Joshi A, Palsson BØ: Metabolic dynamics in the human red cell.
Part III – Metabolic reaction rates. J Theor Biol 1990, 142:41-68.
11. Joshi A, Palsson BØ: Metabolic dynamics in the human red cell.
Conclusion Part IV – Data prediction and some model computations. J
We developed a computer model of the human RBC that Theor Biol 1990, 142:69-85.
12. Ni TC, Savageau MA: Model assessment and refinement using
is based on a previous model but was expanded by intro- strategies from biochemical systems theory: application to
ducing a GSH synthesis pathway and a GSSG export sys- metabolism in human red blood cells. J Theor Biol 1996,
179:329-368.
tem. With this expansion, the model maintained high ATP 13. Ni TC, Savageau MA: Application of biochemical systems the-
concentrations in G6PD deficiency. This suggests that ory to metabolism in human red blood cells. Signal propaga-
these pathways may play an important role in alleviating tion and accuracy of representation. J Biol Chem 1996,
271:7927-7941.
the consequences of G6PD deficiency. It also indicates 14. Jamshidi N, Edwards JS, Fahland T, Church GM, Palsson BØ:
that sub-pathways that are normally not particularly Dynamic simulation of the human red blood cell metabolic
highly activated may play important roles in abnormal network. Bioinformatics 2001, 17:286-7.
15. Mulquiney PJ, Kuchel PW: Model of 2,3-bisphosphoglycerate
conditions such as deficiencies. metabolism in the human erythrocyte based on detailed
enzyme kinetic equations: computer simulation and meta-
bolic control analysis. Biochem J 1999, 15:597-604.
Authors' contributions 16. Fiorelli G, Martinez di Montemuros F, Cappellini MD: Chronic non-
Nakayama contributed mostly to the model develop- spherocytic haemolytic disorders associated with glucose-6-
ment, Kinoshita contributed to the analysis, and Tomita phosphate dehydrogenase variants. Baillieres Best Pract Res Clin
Haematol 2000, 13:39-55.
developed the basic ideas and directed the project. 17. Schuster R, Jacobasch G, Holzhütter HG: Mathematical modelling
of metabolic pathways affected by an enzyme deficiency.
Energy and redox metabolism of glucose-6-phosphate-dehy-
Competing interests drogenase-deficient erythrocytes. Eur J Biochem 1989,
The author(s) declare that they have no competing 182:605-612.
interests. 18. Schuster R, Jacobasch G, Holzhütter H: Mathematical modelling
of energy and redox metabolism of G6PD-deficient
erythrocytes. Biomed Biochim Acta 1990, 49:S160-5.
Acknowledgements 19. Holzhütter HG, Schuster R, Buckwitz D, Jacobasch G: Mathemati-
We thank Ryo Matsushima and Kazunari Kaizu for providing technical cal modelling of metabolic pathways affected by an enzyme
deficiency. Biomed Biochim Acta 1990, 49:791-800.
advice. This work was supported in part by a grant-in-aid from the Ministry 20. Schauer M, Heinrich R, Rapoport SM: Mathematical modelling of
of Education, Culture, Sports, Science and Technology (the leading project glycolysis and adenine nucleotide metabolism of human
for biosimulation and the 21st Century Center of Excellence (COE) Pro- erythrocytes. I. Reaction-kinetic statements, analysis of in
gram: Understanding and Control of Life's Function via Systems Biology), vivo state and determination of starting conditions for in
vitro experiments. Acta Biol Med Ger 1981, 40:1659-1682.
and in part by a grant from New Energy and Industrial Technology Devel-
21. Mulquiney PJ, Kuchel PW: Model of the pH-dependence of the
opment and Organization (NEDO) of the Ministry of Economy, Trade and concentrations of complexes involving metabolites, haemo-
Industry of Japan (Development of a Technological Infrastructure for Indus- globin and magnesium ions in the human erythrocyte. Eur J
trial Bioprocess Project). Biochem 1997, 245:71-83.
22. Halestrap AP: Transport of pyruvate and lactate into human
erythrocytes. Evidence for the involvement of the chloride
References carrier and a chloride-independent carrier. Biochem J 1976,
1. Mendes P: GEPASI: a software package for modelling the 156:193-207.
dynamics, steady states and control of biochemical and 23. Henderson JF, Patterson ARP: Nucleotide Metabolism: An Introduction
other systems. Comput Appl Biosci 1993, 9:563-571. Academic Press; 1973.
2. Schaff J, Fink CC, Slepchenko B, Carson JH, Loew LM: A general 24. Thorburn DR, Kuchel PW: Regulation of the human-erythro-
computational framework for modeling cellular structure cyte hexose-monophosphate shunt under conditions of oxi-
and function. Biophys J 1997, 73:1135-1146. dative stress. A study using NMR spectroscopy, a kinetic

Page 10 of 11
(page number not for citation purposes)
Theoretical Biology and Medical Modelling 2005, 2:18 [Link]

isotope effect, a reconstituted system and computer

simulation. Eur J Biochem 1985, 50:371-86.
25. McIntyre LM, Thorburn DR, Bubb WA, Kuchel PW: Comparison of
computer simulations of the F-type and L-type non-oxida-
tive hexose monophosphate shunts with 31P-NMR experi-
mental data from human erythrocytes. Eur J Biochem 1989,
180:399-420.
26. Gerber G, Preissler H, Heinrich R, Rapoport SM: Hexokinase of
human erythrocytes. Purification, kinetic model and its
application to the conditions in the cell. Eur J Biochem 1974,
45:39-52.
27. Otto M, Heinrich R, Kuhn B, Jacobasch G: A mathematical model
for the influence of fructose 6-phosphate, ATP, potassium,
ammonium and magnesium on the phosphofructokinase
from rat erythrocytes. Eur J Biochem 1974, 49:169-78.
28. Holzhütter HG, Jacobasch G, Bisdorff A: Mathematical modelling
of metabolic pathways affected by an enzyme deficiency. A
mathematical model of glycolysis in normal and pyruvate-
kinase-deficient red blood cells. Eur J Biochem 1985, 149:101-11.
29. Lassen UV: Hypoxanthine transport in human erythrocytes.
Biochim Biophys Acta 1967, 135:146-54.
30. Jacobasch G, Buckwitz D, Jurowski R, Gerth C, Plonka A, Kuckelkorn
U: Heterogeneity of glucose-6-phosphate dehydrogenase
enzymopathies in the GDR. Biomed Biochim Acta 1987,
46:S177-181.
31. Misra I, Griffith OW: Expression and purification of human
gamma-glutamylcysteine synthetase. Protein Expr Purif 1998,
13:268-276.
32. Ristoff E, Augustson C, Geissler J, de Rijk T, Carlsson K, Luo JL,
Andersson K, Weening RS, van Zwieten R, Larsson A, Roos D: A
missense mutation in the heavy subunit of gamma-glutamyl-
cysteine synthetase gene causes hemolytic anemia. Blood
2000, 95:2193-6.
33. Njalsson R, Carlsson K, Olin B, Carlsson B, Whitbread L, Polekhina
G, Parker MW, Norgren S, Mannervik B, Board PG, Larsson A:
Kinetic properties of missense mutations in patients with
glutathione synthetase deficiency. Biochem J 2000, 349:275-279.
34. Kondo T, Dale GL, Beutler E: Glutathione transport by inside-
out vesicles from human erythrocytes. Proc Natl Acad Sci U S A
1980, 77:6359-6362.
35. Homolya L, Varadi A, Sarkadi B: Multidrug resistance-associated
proteins: Export pumps for conjugates with glutathione, glu-
curonate or sulfate. Biofactors 2003, 17:103-14.
36. Schuster R, Holzhütter HG: Use of mathematical models for
predicting the metabolic effect of large-scale enzyme activ-
ity alterations. Application to enzyme deficiencies of red
blood cells. Eur J Biochem 1995, 229:403-18.
37. Yugi K, Nakayama Y, Tomita M: A hybrid static/dynamic simula-
tion algorithm: Towards large-scale pathway simulation
[abstract]. Proceedings of the Third International Conference on Systems
Biology: 13-15 December 2002, Stockholm:235 .
38. Jenkins JD, Madden DP, Steck TL: Association of phosphofruc-
tokinase and aldolase with the membrane of the intact
erythrocyte. J Biol Chem 1984, 259:9374-8.
39. Jenkins JD, Kezdy FJ, Steck TL: Mode of interaction of phosphof-
ructokinase with the erythrocyte membrane. J Biol Chem 1985,
260:10426-10433.
40. von Ruckmann B, Schubert D: The complex of band 3 protein of
the human erythrocyte membrane and glyceraldehyde-3-
phosphate dehydrogenase: stoichiometry and competition
by aldolase. Biochim Biophys Acta 2002, 1559:43-55. Publish with Bio Med Central and every
scientist can read your work free of charge
"BioMed Central will be the most significant development for
disseminating the results of biomedical researc h in our lifetime."
Sir Paul Nurse, Cancer Research UK

Your research papers will be:

available free of charge to the entire biomedical community
peer reviewed and published immediately upon acceptance
cited in PubMed and archived on PubMed Central
yours — you keep the copyright

Submit your manuscript here: BioMedcentral

[Link]

Page 11 of 11
(page number not for citation purposes)