Aqueous solutions that model the cytosol: Studies on polarity, chemical reactivity and enzyme kinetics

Article abstract of DOI:10.1039/b402886d

Concentrated solutions of a series of organic compounds have been prepared and the effects of these solutes on the properties of the solvent system assessed as a function of their concentration and nature. Polarity, as measured by Reichardt's E_T

Full text of DOI:10.1039/b402886d

View Article Online / Journal Homepage / Table of Contents for this issue
Aqueous solutions that model the cytosol: studies on polarity,
chemical reactivity and enzyme kinetics†
Nabil Asaad, Marie Jetta den Otter and Jan B. F. N. Engberts*
Physical Organic Chemistry Unit, Stratingh Institute, University of Groningen, Nijenborgh 4,
9747 AG Groningen, The Netherlands. E-mail: j.b.f.n.engberts@chem.rug.nl;
Fax: 31(0)50 3634296; Tel: 31(0)50 3634242
Received 25th February 2004, Accepted 22nd March 2004
First published as an Advance Article on the web 6th April 2004
Concentrated solutions of a series of organic compounds have been prepared and the effects of these solutes on the
properties of the solvent system assessed as a function of their concentration and nature. Polarity, as measured by
Reichardt’s ET(30) probe, exhibits a linear variation with both solute and water concentration for simple solutes.
Non-linear behaviour was also observed and is associated with preferential solvation or binding of the ET(30) probe
molecule by the added solute. The observed trends in polarity are mirrored in the effects of these solutes on chemical
reactivity and enzyme kinetics. Environmental effects on the kinetics of hydrolysis of 4-nitrophenyl dichloroacetate,
the hydronium-ion catalysed hydrolysis of 2-(4-nitrophenoxy)-tetrahydropyran, the acyl transfer reaction between
4-nitrophenyl acetate and TRIS, the Diels–Alder reaction between 1,4-naphthoquinone and cyclopentadiene and
the trypsin-catalysed hydrolysis of 4-nitrophenyl acetate are reported and discussed in terms of the properties of
the solutes and the mechanistic requirements of these reactions. Linear correlations were observed between the
logarithms of the rate constants for the acetal hydrolysis, acyl transfer and Diels–Alder reactions with water
concentration. Since the latter varies linearly with ET(30), this indicates a linear free energy relationship between
solution polarity and chemical reactivity.
Solutes were chosen that are freely soluble in water and
Introduction
unreactive towards the chemical probes used. The amide
The cytosol is an aqueous region, surrounded by a lipid bilayer
group is ubiquitous in biochemistry, hence a series of amides
membrane, crowded with macromolecules and other solutes
was chosen with the same basic functionality, but a progres-
totalling several hundred grams per litre.1 This environment is
sive increase in hydrophobic character: urea, formamide,
markedly different to the dilute aqueous solutions generally
N,N-dimethylacetamide (DMA), N-tert-butyl acetoacetamide
used as background comparisons for biochemical processes,2
hence is expected to have different solvent properties. In the
(NBAA) and N-cyclohexyl pyrrolidin-2-one (N–CHP). Sorbi-
tol was chosen as a mimic for sugars which does not contain
the reactive acetal functionality, and bovine serum albumin
(BSA) was used as a model protein. Solutions of poly(ethylene
cellular environment, the rate of diffusion, especially of large
molecules, is reduced3 and macromolecular crowding agents
promote protein folding and association.4
glycol) (PEG), which has been widely employed as a solubilis-
The cytosol may be modelled by aqueous solutions of polar
ing agent and stabiliser in chemo-enzymatic synthesis,7 were
organic molecules, salts, sugars and their macromolecular
also studied. Whilst, obviously less appropriate as a mimic for
analogues. This constitutes a reasonable analogue since the
biological systems, the constitution of PEG is roughly in-line
tertiary structure of proteins places hydrophilic groups outer-
most into solution, nucleic acids contain anionic phosphate
groups and are generally associated with proteins, and bilayer
with the other solutes studied and it is available in a range
of well-defined molecular weights. This potentially allows
the effects of solute size and macromolecular crowding to be
assessed.
structures similarly have polar headgroups disposed at the
aqueous interface.5 We have previously reported a study into the
effects of added solutes at high concentrations on the trypsin-
catalysed hydrolysis of 4-nitrophenyl acetate.6 This report
Results and discussion
describes a more general investigation into the chemical proper-
Since the test systems generally contain a large proportion of
water, the best measure of composition is the water concen-
tration. This prevents distortion of the results by the size of
the cosolute – for example, pure formamide has a formamide
concentration of 25.2 M, whereas the concentration of pure
N-cyclohexyl pyrrolidin-2-one is 6.0 M. Therefore a scale of
solute concentration would have a different end-point depend-
ing on the solute. Using a water concentration scale reverses the
direction of the trend relative to a solute concentration scale,
but all solutions range in concentration from zero to 55.3 M at
25 ЊC. Using molar water concentration as the default scale for
solution composition also removes the complication of how to
treat mixtures containing different quantities of more than one
solute. In practice, molar solute concentration and water con-
centration could be used interchangeably since, for the solutes
and concentration ranges studied, they have an inverse linear
relationship with each other. The solution density also increases
ties of aqueous solutions containing large concentrations of
solutes. The effects of the changing environment on polarity
and chemical and enzyme reactivity have been assessed as
a function of solute concentration. The general procedure
followed was to prepare a series of solutions of the solute of
interest, ranging from pure water to biologically significant
concentrations of several hundred grams per litre. The
change in a property such as polarity over this concentration
range was measured and the dependence of this variation on
the nature of the solute analysed. The solutes used in this study
were selected to mimic the cellular environment as closely
as possible whilst still allowing for controlled variation of
properties.
† Electronic supplementary information (ESI) available: further results
and discussion and tables of experimental data. See http://www.rsc.org/
suppdata/ob/b4/b402886d/
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 0 4 – 1 4 1 2
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4
1404

View Article Online
linearly with solute concentration indicating that there is no
preferential association of water to the charged ground state of
significant change in volume associated with mixing.
the ET(30) probe at very low water concentrations.
Non-linear variation of ET(30) with concentration was
observed for N-cyclohexyl pyrrolidin-2-one (N–CHP) (Fig. 2).
This compound is capable of acting as a hydrotrope,11 hence
solubilises hydrophobic compounds in water. The concen-
trations of water and N–CHP in binary mixtures and the
density of the solution are linearly dependent, again indicating
that there is no appreciable change in molecular size. However,
the non-linear variation in ET(30) of the mixture indicates
non-specific binding of the ET(30) probe.
Solution polarity
Polarity was measured using Reichardt’s ET(30) probe.8 The
zwitterionic form of the molecule exhibits a charge transfer
band in the visible region. The energy of this transition, the
ET(30), is sensitive to the ability of the probe’s environment to
solvate hydrophobic groups, stabilise charges and donate
hydrogen-bonds. Since the ground state of the molecule is
charged and the excited state is effectively a diradical, an
environment capable of stabilising charges will lower the
ground state relative to the excited state, increasing the transi-
tion energy and shifting the charge transfer band to shorter
wavelength.
Each aqueous test solution was made alkaline with the same
concentration of NaOH and the ET(30) value was measured as
a function of solution composition. Interestingly, simple sol-
utes exhibited a linear variation of ET(30) with water concen-
tration over a large concentration range (Fig. 1). This suggests
that the solutes mix homogenously with water.
Fig. 2 Variation of ET(30) with water concentration for solutions of
N-cyclohexyl pyrrolidin-2-one.
N–CHP may form micelle-like aggregates in aqueous
solution above a critical concentration of around 750 mM
(approximately 50 M in water). These aggregates would bind
the hydrophobic ET(30) probe preferentially, giving rise to two
regions in the ET(30)-concentration profile. At low water con-
centrations, the probe is fully bound to the N–CHP aggregates
and the small change in ET(30) on further increasing the
N–CHP concentration is presumably the effect of the increas-
ing proportion of N–CHP aggregates in the bulk solution. In
the dilute region, the observed dependence of ET(30) on con-
centration is similar to that of the other solutes. The decline in
ET(30) with water concentration in this region is much steeper
than that of the other solutes studied, which may indicate
hydrophobic association.
Fig. 1 Variation of ET(30) with water concentration for a series of
solutions of simple organic solutes [PEG-number refers to the average
molecular weight of the poly(ethylene glycol)].
It is apparent that those solutes with more hydrocarbon
character give rise to a more pronounced decrease in ET(30)
with water concentration, as would be expected from the ET(30)
values of the pure compounds. The addition of sorbitol has
a very small effect on solution polarity. A similar pattern is
observed for solutions of sugars9 and is consistent with the
“camouflage effect”.10 The highly polar solutes urea and form-
amide exert only a small effect on polarity. Taken with the still
greater effect of DMA, a picture emerges where solution polar-
ity is dependent on the proportion by volume of polar and
apolar functionality – the more alkyl group character in the
molecule, the greater the decrease in polarity per volume of
water replaced.
Actual binding of the ET(30) probe by the solute was
observed when bovine serum albumin (BSA) was used in an
attempt to mimic the presence of proteins in the cytosol (Fig. 3).
The results obtained with the series of PEG’s indicate that
the size of the solute is relatively unimportant, though it should
be noted that the smaller PEG’s have a much greater relative
proportion by volume of terminal hydroxyl groups, hence a
smaller effect on polarity.
The experimental ET(30) value of pure water (Fig. 1, closed
square) was not always equal to that obtained by extrapolation
to zero solute concentration. This is probably due to (weak)
aggregation of the poorly water-soluble probe in the absence of
organic cosolutes. Head-to-tail stacking of the probe molecule
could give rise to additional stabilisation of charges hence a
higher apparent ET(30) value. Similarly, an extrapolation to
zero water concentration overestimates the ET(30) value of the
pure solutes (not shown). This effect may be the result of
Fig. 3 ET(30) values of solutions of bovine serum albumin (BSA) at
low concentration, indicating binding of the ET(30) probe molecule to
the protein (probe concentration 28 × 10Ϫ6 M).
Initial studies on the polarity of BSA solutions showed that
all the test solutions in the region of 50–200 g LϪ1 of BSA had
essentially the same ET(30) value. It was necessary to study
solutions of BSA in the micromolar concentration range in
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 0 4 – 1 4 1 2
1405

View Article Online
order to follow the change in ET(30) value from that of pure
water to this plateau value of 55.3 kcal molϪ1
.
Several titrations of BSA into a solution of Reichardt’s dye
were carried out. A plateau on the ET(30)-concentration profile
was obtained above a BSA concentration of 2 × 10Ϫ5 M, indi-
cating complete binding of the ET(30) probe to the protein.
Given that saturation occurred at a concentration of BSA lower
than the concentration of ET(30) probe (2.8 × 10Ϫ5 M), it is
clear that one BSA molecule is capable of binding more than
one ET(30) probe molecule. This is reasonable since serum
albumins are known to contain multiple hydrophobic binding
sites.12 Attempts to ascertain the number of equivalents of
bound ET(30) probe by a Job plot analysis13 were generally
unsatisfactory, though it is clear that more than two equivalents
of ET(30) probe are bound per BSA (Fig. 4).
Scheme
(4NPD).
1
Neutral hydrolysis of 4-nitrophenyl dichloroacetate
Fig. 5 Variation in the rate constants for hydrolysis of 4-nitrophenyl
dichloroacetate with water concentration upon addition of N,N-
dimethylacetamide (DMA, ᭺) and N-cyclohexyl pyrrolidin-2-one
(N–CHP, ᭝).
that the reaction rate is dependent on the polarity of the
medium and that the relatively hydrophobic kinetic probe
reports the same environmental characteristics as the ET(30)
probe.
Fig. 4 Job-type plot for binding of the ET(30) probe to bovine serum
albumin (BSA) [second-order polynomial fit].
A free energy relationship between polarity and reactivity for
a (pseudo)first-order process would require a correlation of log
(kobs/sϪ1) with ET(30). This is indeed the case over part of the
concentration range studied (not shown). Several complications
emerge: firstly, the Gibbs energy of activation of the reaction
depends on the chemical potential of water rather than its con-
centration (the solutions studied fall well outside the concen-
tration range for which ideal behaviour can be assumed), hence
a non-linear dependence of thermodynamic water activity on
concentration is likely to contribute to the observed trend in
reactivity. Secondly, it is possible that, at lower water concen-
trations, a hydrolysis mechanism involving only one water
molecule dominates; it should be noted however that the sol-
The data does not fit two intersecting straight lines, but
rather a non-linear distribution with a minimum at a mole frac-
tion of BSA of approximately 0.46. This behaviour may be due
to the fact that the binding sites of BSA are not all equivalent
and also that a small concentration of the host was titrated into
the guest rather than vice versa.
These experiments did not give insight into the bulk polarity
of protein solutions, however, they did give the interesting
result that the ET(30) values of the BSA binding sites are in the
region of 55.3 kcal molϪ1. This is virtually the same polarity as
neat methanol and gives an indication of the likely “solvent
polarity” inside the active sites of enzymes.
vent kinetic isotope effect (kH O/kD O = 3.1)15 does not change
2
over the DMA concentration range2studied, implying that there
is no change in mechanism. It is further possible that, at low
water concentration, general base catalysis by the amide solutes
occurs. Furthermore, the rate constant for hydrolysis and the
pKa’s of the general base(s) are expected to vary with solution
composition. Binding of the probe to N–CHP aggregates must
also be considered in order to deconvolute the above hydrolysis
data.16 It was therefore decided to switch to a series of simpler
test reactions and inert solutes in order to probe the changes in
reactivity.
Chemical reactivity
Chemical processes of interest in this research fall into three
general categories: reactions that directly involve water, other
reactions that are sensitive to their environment and reactions
of biomolecules. Reactions that directly involve water would be
expected to respond to the addition of solutes that interact with
water. In conjunction with a study of the effects of solutes on
medium-sensitive reactions in aqueous solution, such processes
can be used to give insight into the nature of cellular water. The
effects of added solutes on hydrolysis, acyl transfer and Diels–
Alder reactions have been studied to this end. Furthermore, a
study of environmental effects on biochemical reactions gives
direct insight into in vivo-processes.
Initial studies on chemical reactivity focussed on the effect of
amide solutes on the hydrolysis of the activated ester, 4-nitro-
phenyl dichloroacetate (4NPD). The neutral hydrolysis reaction
in solutions of 2 < pH < 4 involves two attacking water mole-
cules in the rate-limiting step – one as a nucleophile and the
other acting as a general base (Scheme 1).14
Three further probe reactions were studied: the acyl
transfer reaction between 4-nitrophenyl acetate (4NPA) and
tris(hydroxymethyl)aminomethane (TRIS) (Scheme 2), the
hydronium-ion catalysed hydrolysis of 2-(4-nitrophenoxy)-
tetrahydropyran (Scheme 3) and the Diels–Alder reaction
between cyclopentadiene and 1,4-naphthoquinone (see Scheme
4).
Nucleophilic attack by TRIS on 4NPA17 (Scheme 2, Nu–H =
TRIS) is analogous to the hydrolysis of 4NPD (Scheme 1), with
Changes in the rate of 4NPD hydrolysis upon addition of
N,N-dimethylacetamide (DMA) or N-cyclohexyl pyrrolidin-2-
one (N–CHP) mirrored the variation in polarity with water
concentration observed for these solutes (Fig. 5), indicating
Scheme 2 Nucleophilic displacement of 4-nitrophenolate from 4-
nitrophenyl acetate.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 0 4 – 1 4 1 2
1406

View Article Online
The logarithms of the second-order rate constants for the
TRIS-catalysed reaction correlate linearly with water concen-
tration. The former is proportional to the Gibbs energy of
activation of the reaction and, since the ET(30) value of the
solution varies linearly with water concentration (Fig. 1), the
Gibbs energy of activation also correlates with solution polar-
ity. There is therefore a linear free energy relationship between
chemical reactivity and the polarity of the medium, as meas-
ured by the ET(30) probe (see Fig. 8). That the two properties –
ET(30) polarity and reaction rate constant – show the same
general trend with solution composition is not unexpected:
chemical reactivity is quantified in terms of the Gibbs energy
difference between the initial and transition states and the
ET(30) value is the energy difference between a charged ground
state and charge transfer excited state. However, the linear
correlation is noteworthy.18
The hydronium-ion catalysed hydrolysis of 2-(4-nitrophen-
oxy)-tetrahydropyran (4NPTHP) (Scheme 3) in the presence of
PEG (MW 400) was also studied. This reaction is known to be
sensitive to its medium19 and proceeds via an A1 mechanism,20
hence does not involve water until after the rate-determining
step. The first step is an equilibrium protonation, which is fol-
lowed by rate-limiting elimination of 4-nitrophenol. Water then
attacks the generated oxonium species at the 2-position forming
a hemi-acetal that may undergo further hydrolysis.
Scheme 3 Hydronium-ion catalysed hydrolysis of 2-(4-nitrophenoxy)-
tetrahydropyran (4NPTHP).
the attacking nucleophile taking the place of the two water
molecules. The initial state is charge-neutral, hence nucleophilic
attack is accompanied by separation of charge and is favoured
by a more polar environment. The addition of solutes that are
less polar than water would therefore be expected to retard the
reaction. Lewis acid catalysis by hydrogen-bond donation to
the carbonyl oxygen also stabilises the charge-separated transi-
tion state, therefore the addition of a solute that is a poorer
hydrogen-bond donor than water or competes for hydrogen-
bonds would also disfavour the reaction. Poly(ethylene glycol)
(PEG) lowers solution polarity and is a hydrogen-bond
acceptor only, hence is expected to decelerate the reaction via
both effects.
As expected, the observed pseudofirst-order rate constants
vary linearly with TRIS concentration, i.e. the reaction is first-
order in TRIS, at all PEG concentrations studied (0–341 g LϪ1
)
(Fig. 6, top). The intersection of the linear fits at a TRIS
concentration of approximately 0.8 M is not indicative of
an inversion in the trend in reactivity with respect to PEG con-
centration, but rather a reflection of the dependence of the
background (zero-buffer) specific base-catalysed reaction on
polarity. This may be related to an increase in the nucleo-
philicity of hydroxide with PEG concentration, or to a decline
in the ionic product of water.
The initial state of the reaction is charge-neutral, so although
the transition state involves delocalisation of the formal posi-
tive charge on the phenolic oxygen, it is more polar than the
initial state and hence is relatively stabilised by polar solvent
systems.
The second-order rate constants increase with water concen-
tration (Fig. 6, bottom), which is proportional to the ET(30)
polarity, i.e. the rate of acyl transfer increases with the polarity
of the environment.
The expected linear dependence of pseudofirst-order rate
constant on acid concentration is observed at all concentrations
of PEG studied (0–431 g LϪ1) (Fig. 7, top). The second-order
rate constants also follow a trend of increasing with water
Fig.
6
Variation in pseudofirst-order rate constant with TRIS
Fig.
7
Variation in pseudofirst-order rate constant with acid
concentration (top) and second-order rate constant with water
concentration (bottom) for the addition of TRIS to 4-nitrophenyl
acetate at pH 7.7 upon addition of PEG (MW 400).
concentration (top) and second-order rate constant with water
concentration (bottom) for the specific acid-catalysed hydrolysis of
2-(4-nitrophenoxy)-tetrahydropyran upon addition of PEG (MW 400).
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 0 4 – 1 4 1 2
1407

View Article Online
concentration (Fig. 7, bottom). This indicates an analogous
dependence of chemical reactivity on solution polarity to that
of the acyl transfer reaction.
The reaction is therefore expected to exhibit a similar sensi-
tivity to the addition of PEG as the activated ester hydrolysis
discussed above. The replacement of a large concentration of
water by a hydrogen-bond acceptor diminishes Lewis-acid
catalysis and the decline in polarity upon addition of PEG
would also be expected to lead to a decline in reactivity.
Addition of PEG does indeed diminish the reaction rate, but
there is a break in the correlation at a water concentration of
approximately 38 M (Fig. 9).
A good linear correlation between the logarithm of the
second-order rate constant and water concentration is also
obtained for this reaction (Fig. 8). Both trends in reactivity may
be rationalised in terms of the differential effects of the chang-
ing environment on an apolar initial state relative to a charge-
separated transition state. However, that two mechanistically
distinct reactions exhibit the same trend in reactivity with solu-
tion composition suggests that this property is general and the
observed linear correlations are striking.
Interestingly, the relative effect of PEG on the acetal
hydrolysis is greater than that on the acyl transfer reaction of
the ester (Fig. 8). It is likely that the reduction in the ability of
the medium to stabilise charges is the dominant factor. This is
more important for the relatively polar transition state in the
acetal hydrolysis than in the relatively early transition state for
attack of TRIS at the ester carbonyl.
Fig. 9 Variation in Gibbs energy of activation for the Diels–Alder
reaction between 1,4-naphthoquinone and cyclopentadiene with water
concentration upon addition of PEG (MW 400).
A break in a linear free energy relationship can be indicative
of a change in mechanism. However, it is hardly conceivable
that this Diels–Alder reaction proceeds via anything other than
a concerted electrocyclic mechanism. One possible explanation
for this behaviour is that 1,4-naphthoquinone has two carbonyl
groups conjugated to the dienophilic double bond, each of
which has two oxygen lone pairs capable of accepting a hydro-
gen-bond. It is therefore possible that the observed rate depends
on the contributions of five equilibrium hydrogen-bonded-
states, each with different rate constants for the Diels–Alder
reaction (although it is likely that only two of these states would
be important in the observed kinetics). If this was the case, then
an identical break would be expected in the rate constant-
concentration profile of the retro-Diels–Alder reaction. How-
ever, no such break is present.22 The significance of a water
concentration of ∼38 M could be that this is the critical concen-
tration above which there are microscopic regions of “bulk”
water. However, such partitioning of water would be expected
also to cause an analogous break in the ET(30)-water concen-
tration profile, which is not observed. It appears likely that this
behaviour is related to the large negative entropy of activation
of the Diels–Alder reaction in the forward direction. A previous
study into the effects of mixed aqueous-organic solvent systems
on activated acetal hydrolysis attributed the decline in rate with
decreasing water concentration to the greater entropic demands
for solvating the transition state.23 However, a satisfactory
explanation of the origin of the observed behaviour requires a
more detailed study into the kinetics of bimolecular reactions
over this concentration range.
Fig. 8 Relative Gibbs energies of activation for the nucleophilic
substitution of 4-nitrophenyl acetate (4NPA) and the specific acid-
catalysed hydrolysis of 2-(4-nitrophenoxy)-tetrahydropyran (4NPTHP)
in the presence of PEG (MW 400).
Both reactions exhibit linear free energy relationships with
polarity. In principle, the gradient of a plot of Gibbs energy of
activation or log (k) against ET(30) could be used as a measure
of the polarity dependence of reactions where a linear corre-
lation between the two is obtained. In practice, it may not be
possible to apply this approach generally: for a particular reac-
tion in a series of solvent systems, the correlation with ET(30) is
unlikely to hold across a significant change in solvent properties
such as going from a protic to an aprotic environment. How-
ever, the approach should be generally applicable to aqueous
solutions.
The Diels–Alder reaction between 1,4-naphthoquinone and
cyclopentadiene (Scheme 4) in the presence of PEG was also
studied. Strong rate enhancements have been observed for the
Diels–Alder reaction in protic solvents, and especially in water,
relative to apolar organic solvents.21 This rate acceleration has
been attributed to enforced hydrophobic interactions and to the
stabilisation of electron density at oxygen in the transition
states for carbonyl-substituted dienophiles. Furthermore, a
general trend of increasing reaction rate in more polar solvents
has been observed.
Interestingly, the above Diels–Alder reaction is slower in
neat methanol than in a PEG solution with an ET(30) value of
55.4 kcal molϪ1, which is the ET(30) value of neat methanol.
Methanol has a greater hydrogen-bond donating ability than
the PEG solution, but a lower polarisibility,24 indicating that
the Diels–Alder reaction is more sensitive to the latter.
Attempts to study the environmental and binding effects of
BSA on this Diels–Alder reaction failed since the protein reacts
with the dienophile, presumably in a Michael addition, at a rate
comparable to the desired reaction. General base catalysis of
the Kemp elimination by BSA has previously been reported and
exhibits Michaelis–Menten kinetics.25
Scheme 4 Diels–Alder reaction between 1,4-naphthoquinone and
cyclopentadiene.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 0 4 – 1 4 1 2
1408

View Article Online
Enzyme kinetics
The sensitivity of an enzyme to its environment has provoked
much interest both for its immediate relevance to biochemistry
and for its importance in the use of enzymes in chemical syn-
thesis.26 Previous studies into environmental effects on enzyme
chemistry have highlighted the importance of water concen-
tration,27 water activity and co-solute nature28 and dielectric
constant with simultaneous competitive inhibition,29 however
no clear picture has emerged.
It is conventional to compare enzymatic rate enhancements
to the uncatalysed reaction in dilute buffer solution. However, a
fairer comparison would be that between the uncatalysed and
enzyme-mediated reactions under biologically relevant condi-
tions. It would be reasonable to assume that, as long as the
active site of the enzyme is shielded from the bulk solution, the
catalytic rate constant (kcat) would be unaffected by the addition
of co-solutes or crowding agents up until they cause denatur-
ation of the enzyme or affect the diffusional rate-limit. The
substrate binding constant (KM) is expected to be sensitive to
solvent composition since it depends on the chemical potential
of the substrate free in solution. Substrate binding also requires
the dehydration of both the active site and the substrate,
hence might be dependent on water activity. However, our
study into the trypsin-catalysed hydrolysis of 4-nitrophenyl
acetate (4NPA) in the presence of PEG’s (MW 8000 and 1500)
and N-tert-butyl acetoacetamide (NBAA) showed that binding
was essentially unaffected by the addition of these solutes,
whereas kcat decreased with water concentration (Figs. 10 and
11).6 In order to confirm that our previous results were not the
result of denaturation of the enzyme, we have carried out a
further set of experiments using urea, a known denaturant, as
the solute.
Fig. 11 Michaelis–Menten parameters for the trypsin-catalysed
hydrolysis of 4-nitrophenyl acetate in the presence of different solutes.
observed kinetics must therefore be attributable to an environ-
mental effect on the enzyme, the substrate or both.
That KM is essentially unaffected by these solutes indicates
strictly that any change in the chemical potentials of the free
enzyme and substrate (relative to pure water) is equal to the
change in chemical potential of the enzyme–substrate complex.
If it is assumed as an approximation that any change in KM
results from the change in the chemical potential of the free
substrate, it is clear that the concentrated test solutions of PEG
and NBAA have a negligible effect on the neutral substrate
(4NPA). However, a substantial effect on KM could be expected
for a charged substrate. Taking the ET(30) probe in pure water
as a reference, the most concentrated PEG solution studied
destabilises the zwitterionic state by ∼16 kJ molϪ1 relative to the
neutral excited state. If this effect was mirrored on the chemical
potential of the free substrate only, then a 1000-fold decrease in
KM would result.
Fig. 10 Eadie–Hofstee plots for the trypsin-catalysed hydrolysis of
4-nitrophenyl acetate in a series of PEG (MW 400) solutions.
Trypsin catalysis was measured for a series of solutions ran-
ging from no added solute up to 395 g LϪ1 of PEG, up to 269 g
LϪ1 of NBAA and up to 522 g LϪ1 of urea. Monitoring the
concentration of the product rather than the substrate allowed
the use of four convenient initial substrate concentrations in the
region of KM, at the expense of having to calibrate against the
extinction coefficient of 4-nitrophenolate in each test solution.
Kinetic parameters kcat and KM were determined by the Eadie-
Hofstee method.30 Michaelis–Menten kinetics were observed
for all solutes at all concentrations studied.
It is apparent that urea has a significantly different effect on
the enzyme than do PEG and NBAA. The variation in Michae-
lis–Menten parameters for the latter two solutes is of the form
classically expected for noncompetitive inhibition of an
enzyme31 (Fig. 10). However, it is unlikely that the two vastly
different solutes have identical mechanisms of inhibition.
A crowding-induced change in enzyme conformation has
previously been proposed to account for lowering in enzyme
activity32 but, unlike PEG, NBAA is not a crowding agent. The
The apparently linear decline in kcat with water concentration
for these solutes mirrors the previously observed dependence of
reactivity on polarity. [Over this relatively small change in kcat
,
it is impossible to distinguish between a linear change in kcat
and in log (kcat/sϪ1) for these results]. This dependence can be
rationalised since the active site of trypsin is relatively exposed
to the solvent,33 hence a similar dependence of the enzyme-
catalysed reaction can be expected to that of the acyl transfer
reaction discussed above.34
The effect of urea on the enzyme-catalysed reaction was
significantly different to those of the other two solutes. Surpris-
ingly, urea concentrations in excess of 500 g LϪ1 do not com-
pletely inhibit the enzyme. Binding and catalysis are observed,
however there is a large decline in binding (i.e. an increase in
KM) on addition of urea (Fig. 10, bottom), but a concurrent
increase in kcat (Fig. 10, top). The change in KM can be rational-
ised in terms of partial denaturation of the enzyme. This could
potentially open up the structure of the enzyme, exposing
amine or carboxylate amino acid side chains that could catalyse
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 0 4 – 1 4 1 2
1409

View Article Online
the hydrolysis of, or be alkylated by the substrate, releasing
4-nitrophenolate and leading to an apparent increase in kcat
It should be noted that the background TRIS-catalysed
reaction is essentially independent of urea concentration (not
shown).
the effects of such changes on competing processes and the
use of additives to facilitate organic chemistry in aqueous
solutions.
The trypsin-catalysed hydrolysis of 4-nitrophenyl acetate was
remarkably insensitive to the addition of large concentrations
of solute. The substrate binding constant (KM) was essentially
unaffected by the addition of poly(ethylene glycol) and N-tert-
.
Most importantly, we have shown that so long as the struc-
ture of the enzyme is not perturbed by the altered environment,
the presence of large concentrations of inert solutes such as
PEG and NBAA has only a modest effect on enzyme activity –
less than an order of magnitude reduction in kcat and a neg-
ligible effect on KM. Apparently, the catalytic process hardly
reacts to the large reduction in the concentration of water in the
model system. This indicates that the enzyme is relatively stable
to changes in its environment, which is presumably related to its
role in vivo as a digestive enzyme. The insignificant effect on KM
indicates that these solutes have a minimal effect on the chem-
ical potential of the neutral substrate in the bulk solution and
do not compete strongly with the enzyme for solvating water. It
is noteworthy that the decline in kcat is linear with increasing
additive concentration. Comparison to solvent polarity data
and the nucleophilic reaction of the substrate with TRIS
indicates that this appears to be a medium effect on enzyme
chemistry.
butyl acetoacetamide, but the catalytic rate constant (kcat
)
decreased linearly with solute concentration, by less than one
order of magnitude up to 400 g LϪ1. This indicates a small
polarity-dependence of the reaction, presumably due to the
active site being partially exposed to the bulk solution. The
addition of urea disfavoured substrate binding and increased
the apparent catalytic rate constant, again linearly with concen-
tration, but in the case of urea, these effects may be attributable
to partial denaturation of the enzyme. It is likely that the effect
of the environment on enzyme binding will be far more pro-
nounced for a charged substrate binding to an active site that
is well shielded from the bulk solvent. Further research is
called for, ideally using an intracellular enzyme with an ionic
substrate, to assess the effects of the environment on enzyme
chemistry in vivo.
Experimental
Conclusions
Materials
For simple solutes, ET(30) polarity varied linearly with both
solute and water concentration, indicating homogeneous mix-
ing and a linear dependence of polarity on volume composition
of the test solution over the concentration range studied. As
expected, the least polar solutes gave the largest changes in
ET(30) as water concentration diminished.
The ET(30)-concentration data for N-cyclohexyl pyrrolidin-
2-one indicated aggregation of the solute and binding of the
probe molecule, hence the probe reported the polarity of a
hydrophobic domain rather than that of the solution as a
whole.
4-Nitrophenol, 4-nitrophenyl acetate, dichloroacetyl chloride,
deuterium oxide, dextrans, N,N-dimethylacetamide, N-cyclo-
hexyl pyrrolidin-2-one and N-tert-butylacetoacetamide, were
purchased from Aldrich. 1,4-Naphthoquinone, poly(ethylene
glycol)s and acetonitrile were purchased from Acros. Trypsin
(hog pancreas), bovine serum albumin, hen egg-white albumin
and TRIS buffer solution were purchased from Fluka. Urea
was purchased from J. T. Baker.
Methods
It was not possible to obtain the bulk polarity of solutions of
bovine serum albumin, since the probe molecule was preferen-
tially bound by the protein. However, the measured ET(30)
value of the BSA binding sites – in the region of 55.3 kcal molϪ1
– gives an indication of the polarity that can be expected inside
the active site of an enzyme. This region is “methanol-like” in
terms of polarity.
The acyl transfer reaction of 4-nitrophenyl acetate with
TRIS and the specific-acid catalysed hydrolysis of 2-(4-
nitrophenoxy)-tetrahydropyran exhibited linear relationships
between the logarithms of the rate constants and water concen-
tration or ET(30) i.e. linear free energy relationships between
reactivity and polarity. The effect on reaction rate on going
from pure water to biologically significant solute concen-
trations of around 400 g LϪ1 was within one order of
magnitude for both reactions.
The Diels–Alder reaction between 1,4-naphthoquinone and
cyclopentadiene also showed a linear correlation between the
logarithm of the rate constant and water concentration, but
there was a break in the linear free energy relationship at a
water concentration of ∼38 M in water. This reaction is
retarded more by methanol (relative to the reaction in pure
water) than by a poly(ethylene glycol) solution with the same
ET(30) polarity as methanol, indicating a stronger dependence
of this reaction on the polarisability of the solvent than on
hydrogen-bond donating ability. The gradient of a log (k)/
ET(30) graph for mixed aqueous solutions such as those
described here may prove a useful tool for the characterisation
of chemical reactions.
UV-visible spectra and absorbance-time measurements were
recorded using a Perkin-Elmer Lambda 12 spectrophotometer,
or a Shimadzu Multispec-1501 UV-visible spectrophotometer
fitted with diode array detector, attached to a Haake K15/DC50
thermostatted water bath. All masses were measured on a
Sartorius BP211D five decimal place balance and densities were
measured using a Mettler Toledo KEM DA-100M specific
gravity meter. Sonication was carried out using a Bandelin
Sonotex RK255 bath sonicator. Water activity was measured
using an AquaLab 3TE vapour pressure osmometer.
Graph-fitting was done using GraFit v3.00 (Erithacus
Software Ltd) or Origin v. 6.0 (Microcal Software Inc.).
Wavelengths of maximal absorption and absorbance values
were estimated using UV WinLab v. 2.80.03 (Perkin-Elmer
Corporation).
All solvents were distilled prior to use. Aqueous solutions
were prepared from de-ionised water, doubly distilled in an all-
quartz apparatus. Stock solutions of HCl, NaOH and NaCl
were prepared by appropriate dilutions of Merck Titrisol con-
centrates in volumetric glassware. Sugar, sorbitol and PEG’s
were dried over P2O5 prior to use; formamide, N,N-dimethyl-
acetamide and N-cyclohexyl pyrrolidin-2-one were distilled off
CaO; urea was recrystallised from water–ethanol and dried over
P2O5. Cyclopentadiene was prepared by thermal cracking of its
dimer and used immediately. Other compounds were used as
supplied. 4-Nitrophenyl dichloroacetate was prepared by the
method of Fife and McMahon.35 2-(4-Nitrophenoxy)-tetra-
hydropyran was prepared by a procedure based on that of
Woods and Kramer.36
Taking the trends in polarity and reactivity with solution
composition together, it is clear that varying the concentration
of an additive can have a pronounced effect on chemical
reactivity. We suggest that further research is merited into
Test solutions for ET(30), water activity and chemical reactiv-
ity measurements were prepared by mass and their density
measured. Test solutions for enzyme studies were prepared by
dilution of stock solutions to keep the concentrations of
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 0 4 – 1 4 1 2
1410

View Article Online
enzyme, buffer and salts constant across the range of added
Determination of kinetic parameters for the trypsin-catalysed
solute concentration.
hydrolysis of 4-nitrophenyl acetate
Accurate stock solutions of 4-nitrophenol and 4-nitrophenyl
acetate were prepared (ca. 3 mg and ca. 7 mg in 3.00 cm3
acetonitrile, respectively). An accurate stock solution contain-
ing trypsin (ca. 800 mg LϪ1), CaCl2.2H2O (10.0 g LϪ1) and TRIS
buffer solution (0.04 M, pH 7.7) was prepared and stored in the
freezer. All trypsin solutions were used within two days of
preparation. Test solutions were made by five-fold dilution of
this stock. Two samples of test solution (2 × 2 cm3) were pre-
heated to 25 ЊC and a baseline UV-visible scan recorded. 4-
Nitrophenol stock solution (10 µL) was injected, the solution
was briefly sonicated to ensure mixing and the UV-visible
spectrum recorded in triplicate. The wavelength of maximal
absorbance of 4-nitrophenolate (between 400 and 410 nm) and
the extinction co-efficient at that wavelength were determined.
Four further samples of test solution (4 × 2 cm3) were pre-
heated to 25 ЊC and 4-nitrophenyl acetate stock solution (20,
40, 50 and 70 µL, respectively) injected into each. The solutions
were briefly sonicated to ensure mixing and the absorbance was
recorded for a minimum of 3600 s at the wavelength of maxi-
mal absorbance of 4-nitrophenolate in that test solution. Initial
rates were determined by linear regression of the first 2000 s
of absorbance-time data, converting to concentration units
using the previously determined extinction co-efficient of the
product.
General method for the determination of ET(30)
Two samples of test solution (2 × 2 cm3) were preheated to
25 ЊC and a baseline wavelength scan was recorded.37 Solutions
of the ET(30) probe (10 µL, 7.0 mM in EtOH) and NaOH
(10 µL, 2.00 M in water) were added to each and their
UV-visible spectra recorded in triplicate at 240 nm per min with
6 nm smoothing and the ET(30) obtained using the equation:8
ET(30) = 28591/λmax(CT).
ET(30) probe-BSA binding experiment
ET(30) probe (10 µL, 7.0 mM in EtOH) and NaOH (10 µL,
1.00 M in water) were added to water (2.5 cm3) and the solution
preheated to 25 ЊC. BSA solution (105 g LϪ1 in water) was
added in aliquots from a microlitre syringe and the UV-visible
spectrum was recorded in triplicate after each addition. The
ET(30) value was calculated as described above. The combined
data from four titrations is presented in Fig. 3.
General method for chemical reactivity measurements
A stock solution of kinetic probe (ca. 4 × 10Ϫ3 M in
acetonitrile) was prepared. Samples of test solution (2 cm3)
were pre-heated to 25 ЊC and the kinetic probe solution (5 µL)
injected. An appropriate wavelength at which to follow the
kinetics was selected by recording repetitive absorbance-
wavelength scans at a number of time intervals until >95%
completion of the probe reaction in a sample of test solution. A
suitable wavelength is one at which there was a large change in
absorbance over the course of the reaction in a relatively flat
region of the absorbance-wavelength profile. (All repetitive
absorbance-wavelength scans exhibited at least one isosbestic
point). Absorbance-time data was then recorded at that
wavelength and fitted to a first-order rate law.
Acknowledgements
This research was supported financially by the NRSC-C. We
are grateful to Professor Dr Christian Reichardt for the gift of
the ET(30) probe used in this work.
References
1 (a) S. Cayley, B. A. Lewis, H. J. Guttman and M. T. Record Jr.,
J. Mol. Biol., 1991, 222, 281–300; (b) F. C. Neihardt and
H. E. Umbarger, Chemical Composition of Escherischia coli,
inEscherischia coli and Salmonella typhimurium: Cellular and
Molecular Biology, American Society for Microbiology, Washington
DC, 1987.
2 (a) A. P. Minton, J. Biol. Chem., 2001, 276, 10577–10580; (b) R. J.
Ellis, Trends Biochem. Sci., 2001, 26, 596–604.
3 A. S. Verkman, Trends Biochem. Sci., 2002, 27, 27–33.
4 (a) A. S. Morar, X. Wang and G. J. Pielak, Biochemistry, 2001, 40,
281–285; (b) G. Rivas, J. A. Fernandez and A. P. Minton,
Biochemistry, 1999, 38, 9379–9388.
Neutral hydrolysis of 4-nitrophenyl dichloroacetate
As described in the above general procedure, except 2.5 cm3 of
test solution was used and acidified with 5–30 µL of 0.1 M HCl.
The increase in absorbance at 322.2 nm was followed for a
minimum of six half-lives at a minimum of four different acid
concentrations to verify that all measurements were taken in the
region of neutral hydrolysis.
5 L. Stryer, Biochemistry, 4th edn., Freeman, New York,
1995.
Acyl-transfer reaction between 4-nitrophenyl acetate and TRIS
6 N. Asaad and J. B. F. N. Engberts, J. Am. Chem. Soc., 2003, 125,
As described in the above general procedure. Six different TRIS
(50% free base) concentrations in the range 0.04–0.60 M were
used (pH 7.7 0.1). The ionic strength was not adjusted. The
increase in absorbance at 400.0 nm was followed in triplicate for
a minimum of six half-lives.
6874–6875.
7 A. Matsushima, Y. Kodera, M. Hiroto, M. Nishimura and Y. Imada,
J. Mol. Catal. B: Enzymol., 1996, 2, 1–17.
8 C. Reichardt, Chem. Rev., 1994, 94, 2319–2358.
9 The ET(30) values of microcrystalline sugars have been estimated:
K. Fischer and S. Spange, Macromol. Chem. Phys., 2000, 201, 1922–
1929.
Hydronium-ion catalysed hydrolysis of 2-(4-nitrophenoxy)-
tetrahydropyran
10 J. B. F. N. Engberts and M. J. Blandamer, J. Phys. Org. Chem., 1998,
11, 841–846.
11 S. E. Friberg, Curr. Opin. Colloid Interface Sci., 1997, 2, 490–
494.
As described in the above general procedure. Five different acid
concentrations in the range 0.01–0.10 M were used and the
ionic strength adjusted to 0.10 M with NaCl. The increase
in absorbance at 330.0 nm was followed in triplicate for a
minimum of six half-lives.
12 A. A. Battacharya, T. Grune and S. Curry, J. Mol. Biol., 2000, 303,
721–732.
13 C. Y. Huang, R. X. Zhou, S. C. H. Yang and P. B. Chock, Biophys.
Chem., 2003, 100, 143–149.
14 W. Karzijn and J. B. F. N. Engberts, Tetrahedron Lett., 1978, 1787–
1790.
15 J. F. J. Engbersen and J. B. F. N. Engberts, J. Am. Chem. Soc., 1975,
97, 1563–1568.
Diels–Alder reaction between 1,4-naphthoquinone and
cyclopentadiene
16 (a) A previous study into the neutral hydrolysis of 1-benzoyl-3-
phenyl-1,2,4-triazole showed that dilute solutions of hydrotropes
strongly retard the reaction: N. J. Buurma, M. J. Blandamer and
J. B. F. N. Engberts, Adv. Synth. Catal., 2002, 344, 413–420;
(b) E. Iglesias, J. Phys. Chem. B., 2001, 105, 10287–10294.
As described in the above general procedure. Test solutions
contained 430 µM in cyclopentadiene. The decline in absorb-
ance at 340.0 nm was followed in triplicate for a minimum of six
half-lives.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 0 4 – 1 4 1 2
1411

View Article Online
17 A. C. Hengge and R. A. Hess, J. Am. Chem. Soc., 1994, 116, 11256–
11263.
18 For a review of attempts to correlate chemical reactivity with solvent
properties see: C. Reichardt, Solvents and Solvent Effects in Organic
Chemistry, VCH, Weinheim, 1988, ch. 5.
19 T. H. Fife and L. K. Jao, J. Am. Chem. Soc., 1968, 90, 4081–4085.
20 P. A. Sykes, Guidebook to Mechanism in Organic Chemistry,
6th ed., Longman, Harlow, 1986.
1997, 15, 97–101; (c) A. J. Mesiano, E. J. Beckman and A. J. Russel,
Chem. Rev., 1999, 99, 623–633.
27 D. G. Dzingeleski and R. Wolfenden, Biochemistry, 1993, 32, 9143–
9147.
28 G. Bell, A. E. M. Janssen and P. J. Halling, Enzyme Microb. Technol.,
1997, 20, 471–477.
29 M. Mares-Guia and A. F. S. Figueirdo, Biochemistry, 1972, 11,
2091–2098.
21 J. B. F. N. Engberts, Pure Appl. Chem., 1995, 67, 823–828.
22 T. Rispens, Ph.D. Thesis, University of Groningen, in preparation.
23 J. R. Haak and J. B. F. N. Engberts, J. Org. Chem., 1984, 49, 2387–
2391.
24 I. W. Kim, M. D. Jang, Y. K. Ryu, E. H. Cho, Y. K. Lee and J. H.
Park, Anal. Sci., 2002, 18, 1357–1360.
25 (a) K. Kikuchi, S. N. Thorn and D. Hilvert, J. Am. Chem. Soc., 1996,
118, 8184–8185; (b) F. Hollfelder, A. J. Kirby and D. S. Tawfik,
Nature, 1996, 383, 60–63; (c) F. Hollfelder, A. J. Kirby, D. S. Tawfik,
K. Kikuchi and D. Hilvert, J. Am. Chem. Soc., 2000, 122, 1022–
1029.
26 (a) A. Fersht, Enzyme Structure and Mechanism, 2nd edn.,
Freeman, New York, 1985; (b) A. M. Klibanov, Trends Biotechnol.,
30 B. H. J. Hofstee, Nature, 1959, 184, 1296–1298.
31 A. Fersht, Structure and Mechanism in Protein Science, Freeman,
New York, 1998.
32 T. G. A. Lonhienne and S. Winzor, Biochemistry, 2001, 40, 9616–
9622.
33 M. Nakasako, J. Mol. Biol., 1999, 289, 547–564.
34 R. A. Hess, A. C. Hengge and W. W. Cleland, J. Am. Chem. Soc.,
1998, 120, 2703–2709.
35 T. H. Fife and D. M. McMahon, J. Am. Chem. Soc., 1969, 91, 7481–
7485.
36 G. F. Woods and D. N. Kramer, J. Am. Chem. Soc., 1947, 69, 2246.
37 BSA solutions are yellow and some of the other solutes also absorb
light in the region of λmax(CT).
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 0 4 – 1 4 1 2
1412