Biochemical and Biophysical Research Communications 405 (2011) 388–392

Contents lists available at ScienceDirect

Biochemical and Biophysical Research Communications

journal homepage: www.elsevier.com/locate/ybbrc

What is the true enzyme kinetics in the biological system? An investigation

of macromolecular crowding effect upon enzyme kinetics of
glucose-6-phosphate dehydrogenase
Matthew G.S. Norris, Naglis Malys ⇑
Manchester Centre for Integrative Systems Biology, Faculty of Life Sciences, Manchester Interdisciplinary Biocentre, The University of Manchester, 131 Princess
Street, Manchester M1 7DN, UK

a r t i c l e i n f o a b s t r a c t

Article history: Enzyme kinetic parameters for rate equations are vital in metabolic network simulation, a major part of
Received 24 December 2010 systems biology research efforts. Measurements of Michaelis–Menten kinetic parameters Km and Kcat
Available online 13 January 2011 have been performed for enzymes glucose-6-phosphate dehydrogenase (G6P DH) under crowded condi-
tions using molecular crowding agents bovine serum albumin (BSA) and polyethylene glycol (PEG) of
Keywords: 8000 Da molecular weight. An increase in Kcat was observed at very low concentrations of crowding
Enzyme kinetics agent, and also at high crowder concentrations when the experiment was performed at 45 °C with
Macromolecular crowding
PEG. The observed pattern in Kcat for G6P DH at high crowder concentrations has been explained via mod-
Modelling
Systems biology
elling using excluded volume theory. An increase in rate was observed at 45 °C for G6P DH versus 30 °C;
Metabolic network this has been modelled via the Arrhenius equation.
Glucose-6-phosphate dehydrogenase Ó 2011 Elsevier Inc. All rights reserved.

1. Introduction Ficolls and dextrans of different sizes to study the kinetics of alka-
line phosphatise catalysed hydrolysis of p-nitrophenyl phosphate.
When performing metabolic network simulation, an important A reduction rate was observed in all cases where crowding agent
part of systems biology research, enzyme rate equations are used was added. Interestingly the rate decrease was higher for crowding
to calculate flux through different points in a metabolic network. agents with greater molecular weight. Table 1 provides a summary
For enzymes that obey Michaelis–Menten kinetics, rate equation of existing efforts to characterise enzymes under crowded
parameters Km and Vmax must be obtained. Typical enzyme kinetics conditions.
experiments measure such parameters under non-crowded condi- Molecular crowding is known to help thermally stabilise pro-
tions that are considerably different to those seen in vivo. This teins. Steadman et al. [11] used differential scanning calorimetry
work aims to assay molecular crowding effects upon the kinetic to assess the thermal stability of bovine lens crystallins. Increased
parameters of G6P DH using crowding agents BSA and PEG. These thermal stability was noted at high protein concentrations, this
compounds have seen use in several similar experiments previ- was attributed to the excluded volume effect of molecular crowd-
ously [1–7]. ing, described below. Stagg et al. [12] found that molecular crowd-
The effect of molecular crowding upon enzyme kinetics has ing via Ficoll 70 at 25% volume occupancy could thermally stabilise
been investigated in many earlier works. Jiang and Guo [5] inves- flavodextrin. Zhua et al. [13] applied dextran as a crowding agent
tigated the effects of crowding upon the activity of enterobactin- to study the thermal stability of rabbit muscle creatine kinase,
specific isochorismate synthase. They observed an increase in cat- using far-ultraviolet circular dichroism to detect secondary struc-
alytic efficiency (Kcat/Km) with increased PEG, dextran and Ficoll ture. They found that increased crowder concentrations helped sta-
concentrations. Minton and Wilf [8] studied the effects of crowding bilise the enzyme against denaturation.
via agents ribonuclease A, b-lactoglobulin, BSA and PEG upon the A major player in molecular crowding is the effect of excluded
kinetics and oligomer states of glyceraldehyde-3-phosphate dehy- volume, which can cause more compact structures to be favoured
drogenase. They found that observed change in rate could be ac- under crowded conditions [14,15]. Previous work [15] has estab-
counted for by a simple model of excluded volume which took lished a theoretical framework for studying this effect, which has
into account enzyme oligomer states. Homchaudhuri et al. [9] used been utilised for this work. If a molecule, such as an enzyme is
placed in a solution containing a crowding agent, a free energy
⇑ Corresponding author. Fax: +44 161 306 4556. change occurs, equivalent to forming a cavity in which no crowd-
E-mail address: n.malys@manchester.ac.uk (N. Malys). ing agent is present. When an enzyme has monomeric and dimeric

0006-291X/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.bbrc.2011.01.037
 M.G.S. Norris, N. Malys / Biochemical and Biophysical Research Communications 405 (2011) 388–392 389

Table 1
Summary of crowding agent effects upon kinetic parameters of different enzymes.

Enzyme Measurement techniques Crowding agent Kcat Km Kcat /Km Model to explain crowding effect Reference
change change change
Hexokinase Isothermal titration BSA ;30% ;30% None Conformational change during [4]
calorimetry (ITC) catalysis and/or diffusion of product
from enzyme-substrate complex.
Isochorismate synthase Absorbance Ficoll ;20% ;55% "75% Change in ratio of activity coefficients [5]
Dextran PEG 6000 ;20% ;60% "100% between native enzyme and
"30% ;70% "200% enzyme-substrate complex.
ADP-sugar pyrophosphatase Absorbance PEG 6000 "500% ;70% "2000% Crowding effect suggested to be a [6]
regulatory mechanism
Glucosidase II Absorbance BSA "80% ;35% "420% Conformational change resulting in [7]
cleavage reaction enhancement
Multi-copper oxidase Absorbance Dextran 20 "150% "2700% ;90% Change in active site conformation [10]
Ficoll 70 "350% "1200% ;70% via restriction of internal dynamics

Increases and decreases shown are relative to values without crowding agent. ‘‘"’’ denotes increased value, ‘‘;’’ decreased value, and ‘‘None’’ no significant change. Percentages
are the greatest changes found and are approximate.

forms, and the dimer is more compact than individual monomers, measurements using different crowding agent concentrations, as
the total size of the required cavity will be smaller for the dimeric described previously at 30 °C. Another experiment entailed 1 h of
form, and the dimer may be favoured. Similarly if a monomeric en- enzyme incubation at 45 °C with 0%, 0.00000975%, 0.000195%,
zyme has both a compact and extended conformation, the equilib- 0.0781%, 5%, 10%, 15%, 20% w/w PEG, 50 mM Tris–HCl pH 8.0,
rium can be pushed towards the compact conformation by 5 mM NADP+ and 0.4 mM MgCl2. Measurements were performed
excluded volume effects. The enzyme studied below has interest- at 45 °C for 80 min. This experiment was repeated twice to obtain
ing properties which may be perturbed by molecular crowding. two kinetic measurements for each crowding agent concentration.
Glucose-6-phosphate dehydrogenase (G6P DH) catalyses the A third experiment repeated the aforementioned experiment twice
oxidation of glucose-6-phosphate (G6P) to gluconolactone-6-phos- using higher PEG concentrations of 0%, 10%, 20%, 25% and 32% w/w.
phate, with NADP+ simultaneously accepting hydrogen to form Another experiment simultaneously tested stabilisation by NADP+
NADPH. It is part of the pentose phosphate pathway and important and crowding at high temperatures. Enzyme was incubated at
for maintaining NADPH levels in cells. The enzyme’s active form is 45 °C with 50 mM Tris–HCl and 0.4 mM MgCl2 for 1 h, with or
dimeric, with monomeric and tetrameric forms inactive. Under without NADP+ and with or without 0.781% w/w PEG, leading to
normal conditions and at low NADP+ and NADPH levels, the en- a total of 4 combinations of conditions. After incubation, measure-
zyme exists as an equilibrium of dimers and tetramers, with high ments were made at 45 °C for 80 min.
NADP+ and NADPH shifting the equilibrium to inactive tetramers
[16]. Irreversible inactivation follows heating the enzyme to 2.2. Analysis
46 °C, which breaks dimers into monomers. The dimer is stabilised
by two inter subunit contacts. Breakage of one of the contacts dur- Each rate was calculated by fitting a gradient to the linear part of
ing heating results in a less active labile form, with destruction of the absorbance plot from the plate reader software. Substrate con-
the remaining contact inducing monomer formation [17]. centrations and rates were fitted to the Michaelis–Menten equation
using GraphPad Prism, to establish Km and Vmax values. Vmax values
2. Methods and materials were converted to product concentrations by mapping from OD s1
to mM s1 using calibration curves for NADPH (Sigma).
2.1. Glucose-6-phosphate dehydrogenase
2.3. Modelling – excluded volume
For G6P DH experiments, reaction volumes of 100 uL were used
in a 96 well plate format. Each volume contained 50 mM Tris–HCl A model for excluded volume effect developed by Zhou et al.
at pH 8.0, 5 mM NADP+ (from Roche) and 0.4 mM MgCl2. Glucose- [15] was used to calculate Keq values for the monomer/dimer equi-
6-phosphate (G6P) substrate (from Sigma) concentrations included librium of G6P DH under different crowding agent concentrations.
25, 5, 1, 0.4, 0.2, 0.1, 0.05 and 0.025 mM. 0.0025 U of G6P DH The model uses scaled particle theory of hard particle fluids from
(lyophilised powder from Sigma) was used per reaction volume. Reiss et al. [18]. Crowding agent is represented as a hard sphere
NADPH absorbance at 340 nM was measured using a BMG POLAR- with radius rc. The free energy of transferring a solute, modelled
Star Omega plate reader for 80 min. Exploratory experiments as a spherocylinder, into crowded medium is then calculated.
determined a range of BSA and PEG (both obtained from Sigma) The spherocylinder is a cylinder capped at both ends by a hemi-
concentrations; 0%, 0.00000975%, 0.000195%, 0.0781%, 5%, 10%, sphere with a radius of rsc = Rscrc, and a cylindrical length of 2Lscrsc.
15%, 20% w/w. Each substrate concentration was a row on the plate Rsc is used in Eq. (1) below as the ratio rsc/rc, i.e. the ratio between
and each crowding agent concentration a column. the solute and crowding agent radii. u is the proportion of volume
An initial experiment entailed incubation of the plate for 5 min occupied by crowding agent:
at 30 °C prior to starting reaction by adding substrate to wells, fol-
lowed by measurements at 30 °C. This assayed the effects of both DF crowd =RT ¼ lnð1  uÞ þ A1 Q þ A2 Q 2 þ A3 Q 3
SC
BSA and PEG crowding agents. Measurements of the reaction rates
Q ¼ u=ð1  uÞ
associated with each crowding agent concentration were repeated
three times for each BSA concentration and twice for each PEG A1 ¼ R3SC þ 3R2SC þ 3RSC þ 1:5LSC ðR2SC þ 2RSC þ 1Þ ð1Þ
concentration. A2 ¼ 1:5ð2R3SC þ 3R2SC Þ þ 4:5LSC ðR2SC þ RSC Þ
An initial experiment involved incubation of enzyme in 50 mM
A2 ¼ 3R3SC þ 4:5LSC ðR2SC Þ
Tris–HCl buffer and 0.4 mM MgCl2 for 1 h at 45 °C, followed by
390 M.G.S. Norris, N. Malys / Biochemical and Biophysical Research Communications 405 (2011) 388–392

Eq. (1) can be combined with Eqs. (2) and (3) below to calculate Results for the experiment performed on G6P DH at 30 °C are
the change in Keq for a monomer dimer equilibrium in the presence shown in Fig. 1. It can be seen that PEG and BSA have similar effects
of different volumes of crowding agent, u. A and B represent indi- upon Kcat, in that both mediate a significant increase in catalytic
vidual spherical monomers and AB represents the dimer form, activity which begins at a crowding agent concentration of approx-
modelled as a spherocylinder. imately 0.000195% w/w. At this concentration there are approxi-
mately 277 molecules of crowding agent per molecule of
DDF ¼ DF crowd
AB  DF crowd
A  DF crowd
B ð2Þ enzyme. The increase in activity may be caused by the active dimer
form being preferred over the inactive tetramer [17] when crowd-
K eq ¼ K 0eq eðDDF AB =RTÞ ð3Þ ing agent is present, due to the tetramer occupying a larger effec-
tive volume than two individual dimers. Although it appears Km
A modified Arrhenius equation (Eq. (4)) has been used to model the
increases with higher crowding agent concentrations, this increase
dependence of rate constants upon temperature:
    is not significant.
K2 Ea 1 1 Active dimeric G6P DH is known to dissociate into two inactive
Ln ¼  ð4Þ
K1 R T2 T1 monomers at higher temperatures [17]:
Where T1 and T2 are temperatures, k1 and k2 are rate constants for E2 ! 2E1 ð5Þ
respective temperatures, Ea is reaction activation energy and R the
To measure the effect of crowding upon dissociation, experi-
universal gas constant.
ments with incubation of enzyme at elevated temperatures were
performed. An initial thermal deactivation experiment, as men-
3. Results and discussion tioned under Section 2 above, involved incubating G6P dehydroge-
nase in buffer and MgCl2 for 1 h at 45 °C, followed by measurement
In order to evaluate many different conditions simultaneously of activity for 80 min at 30 °C with different concentrations of PEG.
in a high throughput manner, a microplate reader was used with No activity was detected in this experiment, indicating complete
96 well plates. An advantage of the plate reader is that it allows thermal deactivation when the enzyme is heated without NADP+
absorbance to be measured at regular intervals from multiple and crowding agent.
wells, with measurements typically around 1 min apart for each A second thermal deactivation experiment entailed incubation
well, with problems arising due to unreliability of the plate reader of G6P DH at 45 °C for 1 h with MgCl2, buffer, NADP+ and different
machine. When calculating rates, a gradient is fitted to the absor- PEG concentrations, followed by measurement of activity for
bance plot produced by BMG POLARStar Omega plate reader 80 min at 45 °C. It was found that incubation with NADP+ and
software. crowding agent instead of without dramatically stabilises the en-

Fig. 1. Km (A) and Kcat values (B) for G6P DH at different crowding agent concentrations. Kinetic parameters for BSA are shown in blue and those for PEG in red. A log scale is
used on the right side for clarity. Error bars represent one standard deviation. (For interpretation of the references to colour in this figure legend, the reader is referred to the
web version of this article).
 M.G.S. Norris, N. Malys / Biochemical and Biophysical Research Communications 405 (2011) 388–392 391

Fig. 2. Km (A) and Kcat values (B) for G6P DH at different PEG concentrations. Kinetic parameters for a 30 °C experiment are shown in red. Results for an experiment performed
at 45 °C at lower PEG concentrations are shown in green and those for a separate identical experiment with higher PEG concentrations are in blue. Error bars represent one
standard deviation. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

zyme, allowing it to remain active. Crowding agent concentrations agent as a sphere with a radius 0.563 times smaller than the radius
between 0% and 20% w/w appear to linearly increase Kcat values of spherical monomers. This dimension was calculated according
(Fig. 2). to the molecular weights of the crowding agent and monomer. Di-
The third thermal deactivation experiment was performed to mer was represented as a spherocylinder with a volume twice that
determine individual stabilising effects of NADP+ and crowding of monomers, and with Lsc = 0.928. After experimentation it was
agent. Results from this experiment (Table 2) interestingly show found that representing the dimer with a slightly larger volume
that NADP+ and PEG have independent stabilising effects at high than two individual monomers produced an approximately linear
temperature. When there is no NADP+ or PEG, the enzyme is com- dependence of Keq change upon crowding agent volume u between
pletely deactivated by incubation at 45 °C, with stabilising effects 0 and 0.2, as inferred from the low PEG concentration 45 °C exper-
being strong with PEG and weaker with NADP+. iment results shown in Fig. 2. Fig. 3 shows a range of dimer sphe-
Taken together, the thermal inactivation results infer the di- rocylinder radii rsc and their effect upon Keq with different values of
meric active form of the enzyme can be deactivated by heat to pro- u. rsc of 1.11 produces Keq values which match activity seen in the
duce the monomeric form, and that this process is irreversible, as experiment best. u is assumed to be directly proportional to the w/
indicated by the lack of activity in the first experiment. This agrees
with previous research [17]. Results also show that crowding agent
dramatically stabilises the enzyme at high temperature, but that
crowding agent is unable to restore activity once irreversible deac-
tivation has taken place.
The dependence of enzyme activity upon crowding agent con-
centration seen with incubation at 45 °C can be explained by the
scaled particle theory developed by Minton, described under Sec-
tion 2 above. Modelling was performed by representing crowding

Table 2
Kcat and Km values for G6P DH that had been incubated for 1 h at 45 °C with or
without NADP+ and with or without 0.0781% w/w PEG. ‘‘–‘‘ denotes no activity.

Incubation with NADP+ Incubation with PEG Km (mM) Kcat (s1)

Yes Yes 1.37 289.40
Yes No 0.178 36.67
No Yes 1.57 356.93 Fig. 3. Effect of crowding agent volume (u) and dimer rsc (different lines) upon
No No – – modelled Keq of G6P DH dimer formation from monomers. Calculated as described
in the text.
392 M.G.S. Norris, N. Malys / Biochemical and Biophysical Research Communications 405 (2011) 388–392

that can be considered crowding agent. Future studies could use

heterogeneous crowder media to mimic the cellular environment
more closely. Models which take mixed crowded solutions into ac-
count should also be developed.
Crowding agent Monomer, enzyme Dimer with small Dimer with larger
effective volume effective volume
Acknowledgments
Fig. 4. Graphical explanation showing how dimers can occupy a smaller or larger
effective volume than individual monomers. The effective volume is represented by We acknowledge the support of the BBSRC/EPSRC Grant BB/
a dashed line, and represents the volume excluded to crowding agent, as opposed to C008219/1 ‘‘The Manchester Centre for Integrative Systems Biol-
the real volume of the macromolecule, represented by the shaded area. ogy (MCISB)’’ and of the Doctoral Training Center ISBML by BBSRC,
EPSRC and The University of Manchester.
w crowding agent concentration. It was noted that at volume occu-
pancies above 0.2, a peak and dip in dimer Keq is predicted by the References
model. A second thermal deactivation experiment was therefore
performed using higher PEG concentrations to probe this region. [1] J.M. Rohwer, P.W. Postma, B.N. Kholodenko, H.V. Westerhoff, Implications of
macromolecular crowding for signal transduction and metabolite channelling,
Results interestingly infer a plateau in the expected area (Fig. 2). Proc. Natl. Acad. Sci. USA 95 (1998) 10547–10552.
Representing the dimer with a slightly larger volume than two [2] L.A. Munishkina, E.M. Cooper, V.N. Uversky, A.L. Fink, The effect of
individual monomers can be justified by considering that the effec- macromolecular crowding on protein aggregation and amyloid fibril
formation, J. Mol. Recognit. 17 (2004) 456–464.
tive volume excluded to crowding agent by dimers can be either [3] D.S. Banks, C. Fradin, Anomalous diffusion of proteins due to molecular
larger or smaller than that of individual monomers (Fig. 4), since crowding, Biophys. J. 89 (2005) 2960–2971.
cavities in the enzyme structure can be inaccessible to crowding [4] S.N. Olsen, Applications of isothermal titration calorimetry to measureenzyme
kinetics and activity in complex solutions, Thermochim. Acta 448 (2006) 12–
agent. 18.
Eq. (4) has been used to calculate the level of rate constant in- [5] M. Jiang, Z. Guo, Effects of macromolecular crowding on the intrinsic catalytic
crease, k2/k1, for G6P dehydrogenase when temperature is in- efficiency and structure of enterobactin-specific isochorismate synthase, J. Am.
Chem. Soc. 129 (2007) 730–731.
creased from 30 to 45 °C. According to Glaser and Brown [19],
[6] M.T. Morán-Zorzano, A.M. Viale, F.J. Muñoz, N. Alonso-Casajús, G.G. Eydallín, B.
the reaction activation energy for this enzyme is 7.1 kcal/mol or Zugasti, E. Baroja-Fernández, J. Pozueta-Romero, Escherichia coli AspP activity
29,719.18 J/mol. The value of k2/k1 under these conditions is is enhanced by macromolecular crowding and by both glucose-1,6-
1.74. This increase in the rate constant appears to approximately bisphosphate and nucleotide-sugars, FEBS Lett. 581 (2007) 1035–1040.
[7] K. Totani, Y. Ihara, I. Matsuo, Y. Ito, Effects of macromolecular crowding on
match the increase in Kcat observed with the 45 °C experiments. glycoprotein processing enzymes, J. Am. Chem. Soc. 130 (2008) 2101–2107.
To conclude, we have tested the effects of molecular crowding [8] A.P. Minton, J. Wilf, Effect of macromolecular crowding upon the structure and
upon glucose-6-phosphate dehydrogenase. The effects of crowding function of an enzyme: glyceraldehyde-3-phosphate dehydrogenase,
Biochemistry 20 (1981) 4821–4826.
upon G6P DH have been modelled in detail using an excluded vol- [9] L. Homchaudhuri, N. Sarma, R. Swaminathan, Effect of crowding by dextrans
ume framework by Zhou et al. [15]. Modelling has also been per- and Ficolls on the rate of alkaline phosphatase-catalyzed hydrolysis: a size-
formed using the Arrhenius equation for temperature dependent investigation, Biopolymers 83 (2006) 477–486.
[10] I. Pozdnyakova, P. Wittung-Stafshede, Non-linear effects of macromolecular
dependence. Interestingly with the excluded volume model, at u crowding on enzymatic activity of multi-copper oxidase, Biochim. Biophys.
values greater than 0.25 and corresponding crowding agent con- Acta 1804 (2010) 740–744.
centrations, a drop in Keq is predicted by the model. Experimental [11] B.L. Steadman, P.A. Trautman, E.Q. Lawson, M.J. Raymond, D.A. Mood, J.A.
Thomson, C.R. Middaugh, A differential scanning calorimetric study of the
results show a plateau in this region which approximately agrees bovine lens crystallins, Biochemistry 28 (1989) 9653–9658.
with the model. Further work could test the effects of PEG concen- [12] L. Stagg, S.Q. Zhang, M.S. Cheung, P. Wittung-Stafshede, Molecular crowding
trations higher than 32% to verify the drop in Keq predicted by the enhances native structure and stability of alpha/beta protein flavodoxin, Proc.
Natl. Acad. Sci. USA 104 (2007) 18976–18981.
model. However, no significant change in catalytic efficiency, Kcat/
[13] J. Zhua, H. Hea, S. Lib, Macromolecular crowding enhances thermal stability of
Km, was observed. rabbit muscle creatine kinase, Tsingh. Sci & Tech. 13 (2008) 454–459.
The mechanism behind the increase in Kcat for G6P DH at very [14] A.P. Minton, How can biochemical reactions within cells differ from those in
low crowding agent concentrations is unclear but may be caused test tubes?, J Cell. Sci. 119 (2006) 2863–2869.
[15] H.X. Zhou, G. Rivas, A.P. Minton, Macromolecular crowding and confinement:
by active dimers being preferred over inactive tetramers [17]. Mea- biochemical, biophysical, and potential physiological consequences, Annu.
surement of activity using a technique which does not require cou- Rev. Biophys. 37 (2008) 375–397.
pling enzyme, such as isothermal titration calorimetry, may allow [16] F.A. Hasmann, D.B. Gurpilhares, I.C. Roberto, A. Converti, A. Pessoa, New
combined kinetic and thermodynamic approach to model glucose-6-
rate dependence at very low concentrations of crowding agent to phosphate dehydrogenase activity and stability, Enzymol. Microbiol. Tech.
measured, as was the case for G6P DH, which did not require cou- 40 (2007) 849–858.
pling enzyme. The excluded volume model employed here allows [17] E.A. Zaitseva, E.S. Chukhrai, O.M. Poltorak, Stabilisation mechanism of glucose-
6-phosphate dehydrogenase, Khimiya 41 (2000) 127–129.
prediction of the crowding effect that was seen experimentally. A [18] H. Reiss, H.L. Frisch, J.L. Lebowitz, Statistical mechanics of rigid spheres, J.
more sophisticated method could involve attempting to model Chem. Phys. 31 (1959) 369–380.
crowding effects using more precise 3D shapes for enzyme and [19] L. Glaser, D.H. Brown, Purification and properties of D-glucose-6-phosphate
dehydrogenase, J. Biol. Chem. 216 (1955) 67–79.
crowding agent, rather than the simple spheres and spherocylin-
ders used in this work. Real cells have diverse mixture of solutes