Biochem. J. (2013) 454, 387–399 (Printed in Great Britain)
387
doi:10.1042/BJ20130389
Effects of viscosity and osmotic stress on the reaction of human
butyrylcholinesterase with cresyl saligenin phosphate, a toxicant related
to aerotoxic syndrome: kinetic and molecular dynamics studies
Patrick MASSON*†‡1 , Sofya LUSHCHEKINA§, Lawrence M. SCHOPFER* and Oksana LOCKRIDGE*
*Eppley Institute, University of Nebraska Medical Center, Omaha, NE 68198-5950, U.S.A., †Département de Toxicologie, Institut de Recherche Biomédicale des Armées (IRBA)-Centre
de Recherches du Service de Santé des Armées (CRSSA), 24 av des Maquis du Grésivaudan, 38702 La Tronche, France, ‡Laboratoire de Biophysique Moléculaire, Institut de Biologie
Structurale, 41 rue Jules Horowitz, 38027 Grenoble, France, and §Modeling of Biomolecules Laboratory, N.M. Emanuel Institute of Biochemical Physics, Russian Academy of Sciences,
4 Kosygina St., 119334 Moscow, Russia
CSP (cresyl saligenin phosphate) is an irreversible inhibitor
of human BChE (butyrylcholinesterase) that has been involved
in the aerotoxic syndrome. Inhibition under pseudo-first-order
conditions is biphasic, reflecting a slow equilibrium between
two enzyme states E and E . The elementary constants for
CSP inhibition of wild-type BChE and D70G mutant were
determined by studying the dependence of inhibition kinetics
on viscosity and osmotic pressure. Glycerol and sucrose were
used as viscosogens. Phosphorylation by CSP is sensitive to
viscosity and is thus strongly diffusion-controlled (kon ≈108
M − 1 · min − 1 ). Bimolecular rate constants (ki ) are about equal to
kon values, making CSP one of the fastest inhibitors of BChE.
Sucrose caused osmotic stress because it is excluded from the
active-site gorge. This depleted the active-site gorge of water.
Osmotic activation volumes, determined from the dependence
of ki on osmotic pressure, showed that water in the gorge of
the D70G mutant is more easily depleted than that in wild-type
BChE. This demonstrates the importance of the peripheral site
residue Asp70 in controlling the active-site gorge hydration. MD
simulations provided new evidence for differences in the motion
of water within the gorge of wild-type and D70G enzymes.
The effect of viscosogens/osmolytes provided information on the
slow equilibrium E˙E , indicating that alteration in hydration
of a key catalytic residue shifts the equilibrium towards E .
MD simulations showed that glycerol molecules that substitute
for water molecules in the enzyme active-site gorge induce a
conformational change in the catalytic triad residue His438 , leading
to the less reactive form E .
INTRODUCTION
[8]. TOCP has been linked to a delayed neuropathy caused by
consumption of adulterated alcohols during Prohibition [9], and
later to OP-induced delayed neuropathy [10]. Fumes escaping into
the bleed air of aircraft cabins through leaky oil seals in jet engines
contain low levels of TOCP [11]. These fumes are hypothesized to
be the cause of aerotoxic syndrome, an ensemble of neurological
symptoms associated with air travel [12]. It has been proposed that
TOCP is potentially one of the toxicants that causes aerotoxic
syndrome. Sublethal, acute, chronic and long-term exposure to
TOCP could result in phosphorylation of a variety of enzymes
by CSP [13]. However, there is no physiological evidence that
AChE is inhibited in aerotoxic syndrome. On the other hand,
plasma BChE is phosphorylated by CSP in people exposed
to low levels of TOCP. Phosphorylated BChE is a biomarker
for TOCP exposure [14]. BChE may act as an endogenous
stoichiometric bioscavenger [1,6] to protect humans against low
doses of CSP following exposure to TOCP during air travel.
Because conversion of TOCP into CSP depends on cytochrome
P450, genetic polymorphism of cytochrome P450 may contribute
to the level of CSP in the bloodstream. The CSP level should
also depend on endogenous catalytic OP-detoxifying enzymes.
However, two of them, paraoxonase-1 and prolidase, do not
hydrolyse CSP [6].
Previous results have demonstrated that human plasma
BChE reacts biphasically with CSP under pseudo-first-order
BChE (butyrylcholinesterase) (EC 3.1.1.8) is present in most
organs and tissues in humans. Unlike AChE (acetylcholinesterase)
(EC 3.1.1.7) that plays a key role in the cholinergic system
by terminating the action of acetylcholine, BChE has no clear
physiological function. Yet BChE is toxicologically and pharmacologically important: it hydrolyses numerous ester-containing
drugs, and scavenges poisonous OPs (organophosphates),
including nerve agents [1]. OPs act as irreversible inhibitors
of ChEs (cholinesterases) by phosphylating their catalytic serine
residues [2]. In particular, it was found that ChEs are irreversibly
phosphylated by CSP (cresyl saligenin phosphate), also known
as CBDP [2-(o-cresyl)-4H-1,2,3-benzodioxaphosphoran-2-one]
[3–5] (Figure 1A). CSP is an irreversible inhibitor of serine
hydrolases such as carboxylesterases, neuropathy target esterase,
fatty acid ethyl ester synthase/esterase, and chymotrypsin, trypsin
and other serine proteases are also inhibited (see [6] and references
therein).
CSP is the toxic metabolite of TOCP (tri-o-cresyl phosphate)
(Figure 1B). CSP results from in vivo cyclization of TOCP
mediated by liver cytochrome P450 and plasma albumin [7].
TOCP is a component of TCP (tri-cresyl phosphate), which is
a mixture of isomers that is used as an anti-wear/extreme pressure
agent and flame retardant in hydraulic fluids and jet engine oils
Key words: aerotoxic syndrome, butyrylcholinesterase (BChE),
cresyl saligenin phosphate (CSP), diffusion control, organophosphate, osmotic stress, viscosity.
Abbreviations used: AChE, acetylcholinesterase; BChE, butyrylcholinesterase; BTC, butyrylthiocholine iodide; CBDP, 2-(o -cresyl)-4H -1,2,3benzodioxaphosphoran-2-one; ChE, cholinesterase; CSP, cresyl saligenin phosphate; DFP, di-isopropylfluorophosphate; OP, organophosphate; PAS,
peripheral anionic site; TOCP, tri-o -cresyl phosphate; wt, wild-type.
1
To whom correspondence should be addressed (email [email protected]).
c The Authors Journal compilation c 2013 Biochemical Society
388
P. Masson and others
Figure 1
Chemical structures of OPs
(A) CSP (also known as CBDP). (B) TOCP.
Scheme 1
Mechanistic model for inhibition of BChE by CSP
8
conditions with bimolecular rate constants of (1.5 +
− 0.2) × 10
−1
−1
8
−1
−1
+
M · min and (0.25 − 0.07) × 10 M · min [4]. Biphasic
phosphorylation of BChE by CSP was interpreted in terms
of hysteretic behaviour of the enzyme [1,15] (Scheme 1). In
Scheme 1, the enzyme is shown as two interconvertible forms
E and E in slow equilibrium. k0 and k − 0 are the first-order rate
constants for the reversible transition between E and E . In this
minimum reaction scheme, CSP, noted as I, binds and reacts
with both E and E , but with different bimolecular rate constants
(ki > ki ).
The dissociation constants for both enzyme–inhibitor
complexes (E.I and E .I), are respectively:
KI =
koff
kon
and
KI =
koff
kon
(1)
kp and k p are the rate constants of phosphorylation. Since
[Etot ]0 = [E]0 + [E ]0 , the enzyme inactivation rate is:
vin = d[Etot ]0 /dt = kobs,app [Etot ]0 = d([E]0+ [E ]0 )/dt
(2)
Then, the apparent overall inactivation rate constant is:
kobs,app = kobs
1
k0 /k−0
= kobs
1 + k0 /k−0
1 + k0 /k−0
(3)
in which both rate constants for inactivation of E and E are:
kobs =
kp [I]
K I + [I]
=
and kobs
kp [I]
K I + [I]
(4)
If there is a true equilibrium (koff kp ), then the bimolecular rate
constants ki and k i are kp /K I and k p /K I respectively. On the other
hand, if koff kp and k off k p , then ki and k i are approximately
equal to kon and k on respectively. Therefore, starting with a total
activity [Etot ]0 at t0 , the remaining activity at time t, [E]t , can be
described by the sum of two pseudo-first-order processes (eqn 5):
[E]t = [E]0 e
−kobs t
+ [E ]0 e
t
−kobs
(5)
Such biphasic pseudo-first-order inhibition was observed in
some cases for carbamylation and phosphorylation of ChEs. It
c The Authors Journal compilation c 2013 Biochemical Society
has been explained by reaction of two enzyme forms, E and E ,
differing in reactivity (ki >k i ) (see [15] and references therein).
To determine the individual constants in Scheme 1, kon , k on , the
partition coefficients koff /kp , k off /k p , and the characteristics of
the slow equilibrium between E and E , we studied the dependence
of the phosphorylation rates of wt (wild-type) and D70G mutant
BChE on both viscosity and osmotic pressure. Results showed
that inhibition is strongly diffusion-controlled, and that CSP is
one of the fastest phosphorylating agents known for BChE.
The catalytic active centre of BChE is located at the bottom of
a deep gorge [16]. Asp70 is located at the rim of the gorge and is
involved in the PAS (peripheral anionic site) (Figure 2) which
is the initial transient binding site for substrates and other ligands
binding to BChE [17,18]. Asp70 has also been shown to play a
role in the control of hydration of the active-site gorge [19]. For
these reasons, we compared the effect of the viscosogens glycerol
and sucrose on phosphorylation by CSP of the D70G mutant and
wt BChE. Characteristics for inhibition of D70G by CSP were
similar to those for wt BChE. However, results showed that the
water network in the gorge of D70G is weaker than the water
network in the gorge of wt BChE. This finding confirmed the
importance of Asp70 in controlling the hydration of the active-site
gorge.
MD simulations were used to clarify the effect of viscosogens
on the kinetic observations. First, it was shown that changes in
the number of water molecules present in the gorge of the wt
enzyme are rather concerted, whereas in the D70G mutant, they
are chaotic. Secondly, the simulations indicated that a change in
the hydration of the catalytic His438 induced by sucrose-driven
osmotic pressure effects is responsible for the shift of E˙E
towards the less reactive form E . Thirdly, MD simulations showed
that the presence of glycerol in the gorge induces a change in the
conformation of His438 and that it disrupts the water network in
the gorge. These changes hinder inhibitor/substrate traffic to the
active site and hamper formation of Michaelian complexes E.I
and E .I. In this regard, glycerol acts as a competitive inhibitor.
Glycerol was also found to occasionally form hydrogen bonds
with the catalytic triad side chains, thereby acting as a noncompetitive inhibitor.
MATERIALS AND METHODS
Chemicals
CSP (CAS number 1222-87-3) was synthesized by the Defence
Research and Development Center Suffield (Medicine Hat,
Alberta, Canada). Analysis by NMR and MS indicated that CSP
was 100 % pure with a molecular mass of 276.2 Da. A stock
solution of CSP (0.1 M) was made in acetonitrile and stored at
− 70 ◦ C. This stock solution, when kept at − 70 ◦ C, was stable for
at least 2 years. Working solutions of CSP (1 and 0.1 μM) were
made in anhydrous methanol and stored at − 20 ◦ C. Under these
conditions, solutions of CSP were stable for months. All other
chemicals were of biochemical grade.
CSP is a highly toxic OP. Handling of this chemical
requires suitable personal protection, training and facilities. These
requirements are the same as those for other poisonous OPs.
Human plasma BChE
Wt BChE (UniProt accession number P06276) was highly purified
from human plasma of blood donors obtained from a blood bank
as described in [20]. Details of the analysis of the preparation can
be found in [1,20]. BChE had an activity of 41 units/ml, using
Viscosity and osmotic stress effects on phosphorylation of butyrylcholinesterase
Figure 2
389
View of the active-site gorge of human BChE [16] showing its volume and the hydrogen-bond network of water molecules
(A) The full gorge from the enzyme mouth including the PAS; the rim amino acids are shown in grey. Tyr332 and Asp70 form the PAS. Water molecules located at the entrance to the gorge are
hydrogen-bonded to Asp70 . The surface in light blue is the molecular surface that determines the total volume of the full gorge including the PAS (1500 Å3 ). The catalytic centre includes the
π-cation-binding site, Trp82 , and the catalytic nucleophile, Ser198 . Ser198 belongs to a catalytic triad Ser198 . . . His438 . . . Glu325 . Glu197 is involved in stabilization of transition states. The PAS and
the catalytic centre are linked through an -loop (purple). Glu325 and Tyr332 are connected through the E-helix (grey). (B) The catalytic gorge located below Asp70 . Detailed explanations about the
volume of the gorge are in the Supplementary Online Data at http://www.biochemj.org/bj/454/bj4540387add.htm.
1 mM BTC (butyrylthiocholine iodide) as the substrate at pH 8.0
and 25 ◦ C (1 unit = 1 μmol of substrate hydrolysed per min) [21].
The enzyme was stored for several years at + 4 ◦ C in 30 mM
Tris/HCl (pH 7.5) containing 0.1 M NaCl and 0.02 % sodium
azide without significant loss of activity.
of the inhibition mixture. The dilution together with the presence
of substrate stopped the progressive action of the inhibitor. The
dilution also reversed the effects of the viscosogen on BChE
hydrolysis of BTC.
D70G mutant of human BChE
Kinetic analysis of inhibition
The D70G mutant of human BChE was made by site-directed
mutagenesis, expressed in HEK (human embryonic kidney)-293
cells and highly purified as described in [17]. The activity of the
preparation was 3.8 units/ml with 1 mM BTC at pH 8.0 and 25 ◦ C.
Stored for years at + 4 ◦ C, the preparation did not lose activity.
Inhibition of BChE by CSP does not follow simple first-order
inhibition kinetics [4]. The inhibition time course can be described
by eqn (5). The observed rates for inhibition of the two enzyme
forms are designated kobs and k obs respectively, where kobs >k obs .
According to mechanistic models for enzymes that display a
hysteretic behaviour [15], an additional slow equilibrium between
E.I and E .I cannot be ruled out (see Scheme 1), but the rate
constants for that step would be much lower than the inhibition
rates. Thus a slow equilibrium between the two enzyme–inhibitor
complexes would not affect kobs and k obs , or the enzyme activity
remaining during inhibition.
Analysis of progressive inhibition was performed by plotting
log residual activity against the time for incubation with CSP.
The biphasic inhibition time course was deconvoluted graphically
(Figure 3). The slow phase was first-order and was defined
by the linear points late in the time course. The slow phase
was extrapolated to the y-axis origin and the extrapolation was
subtracted from the observed time course to yield the fast-phase
time course which was linear and therefore also first-order. These
observations are consistent with eqn (5), which predicts two
first-order lines (ln[E]t /[E]0 and ln[E ]t /[E ]0 against time), one
for each phase. These lines extrapolate to lnE0 and lnE 0 at
t = 0. Since [Etot ]0 = [E]0 + [E ]0 , these ordinate values give the
fractional abundance of each enzyme form [Etot ]0 and [E tot ]0 , in
the preparation at t = 0 (Figure 3). The slopes of the firstorder lines give kobs and k obs for each inhibitor concentration.
The bimolecular rate constants of inhibition, ki and k i , were
Enzyme inhibition by CSP in buffers of increasing viscosity
Inhibition of wt BChE ([E]0 = 1.0 × 10 − 11 M) was carried out
under pseudo-first-order conditions with different concentrations
of CSP ranging from 0.18 × 10 − 8 M to 0.77 × 10 − 8 M in 100 mM
phosphate buffer (pH 8.0) at 25 ◦ C as described previously [4].
The methanol concentration in each assay was kept constant at
9 %. Glycerol or sucrose were used to increase the viscosity
(η) of the reaction buffer. Viscosogen concentration ranged
from 3.125 % (w/w) to 50 % (w/w). Most experiments were
performed in the absence or presence of 12.5, 25, 37.5 and 50 %
viscosogen. Determination of dynamic and relative (ηrel = η/η0 )
viscosities of buffers is in the Supplementary Online Data at
http://www.biochemj.org/bj/454/bj4540387add.htm.
The time-dependent inhibition of BChE was carried out for
up to 20 min. BChE was incubated with CSP. At different times,
25 μl of incubation medium were removed, and the activity of
BChE was immediately assayed using the method of Ellman et
al. [21] (at 412 nm, 25 ◦ C in 50 mM phosphate buffer, pH 8.0,
using 1 mM BTC as the substrate). The volume of the Ellman
assay mixture was 1540 μl which resulted in a 63-fold dilution
c The Authors Journal compilation c 2013 Biochemical Society
390
P. Masson and others
Figure 3 Semi-log plot showing the time-dependence for the inhibition
of wt BChE by CSP (0.0077 μM) at different concentrations of glycerol (%,
w/w): 0 (䊉), 12.5 (䊏), 25 (䉱), 37.5 (䉫) and 50 (䉬)
The data are from single experiments. For clarity, only the biphasic curve for inhibition kinetics in
the absence of glycerol (0 %) is represented. The continuous lines through the points represent
slow phases, and the continuous and broken lines with no experimental points represent the
deconvoluted fast phases (from 0 % to 50 % glycerol). The slopes of the lines yield − k obs and
− k obs respectively. The ordinate intercepts are ln[E]0 for the fast phase and ln[E 0 ] for the slow
phase, where [E]0 + [E ]0 = [Etot ]0 . Inset: plot of the observed rate constants (k obs and k obs ) at
0 % (continuous lines) and 12.5 % (broken lines) glycerol as a function of CSP concentration
for determination of k i and k i .
determined by plotting kobs and k obs against the CSP concentration
(eqn 4).
Analysis of inhibition kinetics in media of increasing viscosity and
osmotic pressure
The dependence of elementary kinetic constants on solvent
viscosity provides information on the extent to which a reaction
is diffusion-controlled [22–24]. This strategy was used in studies
on AChE-catalysed substrate hydrolysis [25–27]. However, this
approach was never implemented for the investigation of OP
inhibition of ChEs.
The observed bimolecular rate constants (ki and k i ) for
phosphorylation of BChE forms E and E by CSP are a
combination of rate constants for two steps: the diffusion of CSP
to the binding site to form the productive reversible complex E.I
(or E .I), and the chemical step, leading to phosphorylation of the
active-site serine residue. Assuming steady-state formation of the
reversible complex E.I, then [28]
ki =
kon kp
kon
=
koff + kp
1 + kkoffp
(6)
ki depends on the partition coefficient P = koff /kp . If koff kp , there
is a true equilibrium, the reaction is not limited by diffusion and
ki =
kon kp
kp
=
koff
KI
(7)
If koff kp , P = 0 and the reaction is diffusion-limited. In that
case, the phosphorylation would occur with a rate greater than
c The Authors Journal compilation c 2013 Biochemical Society
Figure 4
gorge
Osmotic effect of sucrose on water molecules in the active-site
Sucrose molecules located at the entrance to the gorge act as a semi-permeable membrane.
Water molecules are represented as red balls. Arrows show the direction of water molecule
transfer from the gorge to the bulk solution. The blue surface is the smoothed solvent-accessible
surface of the sucrose molecules. The grey mesh is the protein surface.
the rate of formation of complex E.I, and there would be direct
phosphorylation of the enzyme, i.e. ki ≈ kon . If kp ≈ koff , the
reaction is partially diffusion-limited.
Determination of the dependence of ki on the relative viscosity
(ηrel = η/η0 ) theoretically leads to k0 on and the partition ratio
k0 off /kp . These constants can be determined by plotting the
reciprocal of ki against the relative viscosity [22] (eqn 8). For
theoretical background, see the Supplementary Online Data.
1
k0
1
= 0 ηrel + 0off
ki
kon
kon kp
(8)
However, this analysis can be complicated if the viscosogen
is excluded from the active-site gorge, as is sucrose. In that
case, the compound acts not only as a viscosogen, but also as
an osmolyte. For BChE, this concept manifests itself as a semipermeable membrane across the mouth of the active-site gorge,
inducing an osmotic stress that sucks water molecules out of the
gorge [19,29,30] (Figure 4). The effect of osmotic pressure, π, on
kinetic constants, k, can be described by eqn (9) [19,31]:
‡
Vosm
∂ ln k
=−
∂π
RT
(9)
In this equation, π = [osmolyte] × RT, V ‡ osm is the osmotic
volume of activation (ml/mol), R is the gas constant
(82.1 ml · atm · K − 1 · mol − 1 where 1 atm = 1.013 bar = 0.1 MPa),
and T is the absolute temperature (298 K). Since osmolytes cause
the removal of water molecules from the active-site gorge, V ‡ osm
can be used to probe hydration changes in the gorge and enzyme
surface accompanying the reaction. Estimation of the maximum
Viscosity and osmotic stress effects on phosphorylation of butyrylcholinesterase
391
number of water molecules (nw ) stripped off the hydration layer
and out of the active-site gorge, as a function of π is possible:
nw =
V‡osm
vw
(10)
where vw is the average molar volume of water in the confined
environment of the active-site gorge which is equal to the average
molecular volume of water multiplied by Avogadro’s number.
MD simulations
For molecular modelling of wt BChE, co-ordinates of the protein
and water molecules from the reported PDB crystallographic
structure 1P0I were used [16]. For molecular modelling of
the D70G mutant, the 1P0I structure was manually converted
into D70G by removing the Asp70 side chain. Six systems were
prepared: wt BChE and D70G in water solution without cosolvent, wt BChE and D70G in the presence of 30 % (w/w)
glycerol, and both enzymes in the presence of 30 % (w/w) sucrose.
MD simulations were performed with the NAMD program using
the Lomonosov Moscow State University supercomputer [32].
More details of the systems preparation and computational and
analysis protocol are provided in the Supplementary Online Data.
RESULTS AND DISCUSSION
Inhibition of wt BChE and D70G by CSP
Inhibition of both wt BChE and its D70G mutant showed
similar biphasic behaviour with fast (kobs ) and slow (k obs ) phases
corresponding to the inhibition of two enzyme forms, E and E ,
in slow equilibrium [4,15] (Figure 3). Plots of kobs or k obs against
CSP concentration for wt BChE were linear in the concentration
range used (0.0018–0.0077 μM; Figure 3, inset), indicating that
the highest CSP concentration was far below K I . Thus eqn (4)
reduces to: kobs = kp [CSP]/K I , where kp /K I = ki because koff kp
refer to eqn (7). Thus:
= ki [CSP]
kobs = ki [CSP] or kobs
(11)
The bimolecular rate constants ki and k i were determined by
linear regression analysis of kobs (k obs ) against CSP concentration
plots. Experimental values for wt BChE and D70G mutant are
given in Supplementary Tables S1 and S2 (http://www.biochemj.
org/bj/454/bj4540387add.htm). For wt BChE at zero viscosogen
k0 i = 3.50 × 108 M − 1 · min − 1 and k0 i = 0.75 × 108 M − 1 · min − 1 .
Inhibition of D70G was measured at a single CSP
concentration (6 nM). Estimations of ki and k i were made
using eqn (11): ki,est = kobs /[CSP]. Estimated values for D70G
at zero viscosogen were: k0 i = 3.20 × 108 M − 1 · min − 1 and
k0 i = 0.25 × 108 M − 1 · min − 1 , which are very close to the
corresponding values for the wt enzyme. This confirms previous
results that Asp70 is not important for binding neutral OPs [33].
ki for inhibition of human BChE by potent OP nerve agents
are fast; for instance, 1.1 × 107 M − 1 · min − 1 for VXS ( − ) [34]
and 3.8 × 108 M − 1 · min − 1 for cyclohexylsarin [35]. Values of
ki and k i for inhibition of BChE by CSP are comparable and
are thus among the highest values for reaction of OP agents
with ChEs, making CSP one of the most potent irreversible
inhibitors of human BChE that is known. In addition, owing to
an ultrafast aging reaction (dealkylation of the OP adduct) [3–5],
CSP-inhibited BChE is non-reactivatable by nucleophilic agents
(oximes) currently used in treatment of OP poisoning. Before
the present study, no data were available on the elementary rate
Figure 5 Dependence of 1/k i (the fast-phase rate constant) on relative
viscosity, η/η0
The viscosogens were glycerol (open circles) and sucrose (closed circles). The enzymes were
wt BChE (black) and the D70G mutant (grey). The apparent k i values for D70G were estimated
from k obs /[CSP] for a CSP concentration of 6 nM.
constants and partition coefficients for the reaction of BChE with
CSP. There are no data on the elementary rate constants for the
reaction of any ChE with any of the OP agents. This situation has
begun to be corrected with the current studies using the effect of
viscosity on the kinetics for inhibition of BChE by CSP.
Effect of viscosity on the inhibition kinetics of wt and D70G mutant
BChE
Experiments performed in buffers of increasing viscosity showed
a decrease in rate constants for inhibition as the viscosity was
increased (Supplementary Tables S1 and S2). In theory, when the
reciprocals of the bimolecular inhibition rate constants (1/ki ) are
plotted as a function of ηrel (cf. eqn 8), the plots should be linear.
In the case of wt and D70G BChE, the viscosity dependencies of
ki and k i were not linear, and the degree of non-linearity was very
different depending on whether glycerol or sucrose was used as
the viscosogen (Figures 5 and 6).
Consistent with theoretical expectations, experiments in
glycerol gave linear plots up to 37.5 % glycerol for D70G BChE
and up to 50 % glycerol for wt BChE (Figures 5 and 6). However,
above 37.5 % glycerol, the D70G plots became non-linear. On
the other hand, plots for experiments performed in sucrose were
distinctly non-linear for both enzymes. They were hyperbolic for
the wt enzyme, and biphasic for D70G. The change in slope occurs
at approximately 12.5 % sucrose, ηrel = 1.4, for both enzymes.
Non-linear behaviour suggests that the mechanism by which
the viscosogen affects ki undergoes a change as the viscosogen
concentration is increased. This effect is particularly pronounced
for sucrose where the change in mechanism leads to a decrease
in sensitivity to the sucrose concentration. According to eqn (8),
insensitivity to viscosity may occur if koff kp , because the ki
becomes constant: ki = kon kp /koff = kp /K I . In that case, the reaction
is not diffusion limited. Applying that interpretation to the sucrose
experiments implies that at ηrel ≈1.4, there is a change in the
relative rates of koff and kp such that koff kp . However, the
kinetic behaviour of both wt and D70G BChE in the presence of
glycerol shows that ki remains sensitive to viscosity out to ηrel = 3,
therefore viscosity-induced changes in koff and/or kp at ηrel = 1.4
c The Authors Journal compilation c 2013 Biochemical Society
392
P. Masson and others
High bimolecular rate constant for reaction of human BChE by CSP
Figure 6 Dependence of 1/k i (the slow-phase rate constant) on relative
viscosity, η/η0
The viscosogens were glycerol (open circles) and sucrose (closed circles). The enzymes were
wt BChE (black) and the D70G mutant (grey). The apparent k i values for D70G were estimated
from k obs /[CSP] for a CSP concentration of 6 nM.
are unlikely. An alternative interpretation can be offered, on the
basis of the fact that sucrose is large enough that it is excluded
from the active-site gorge (Figure 4). Consequently, experiments
in the presence of sucrose turn into osmotic stress experiments.
For analysis of sucrose data in terms of osmotic stress, see below.
At high concentrations of glycerol, plots of 1/ki against ηrel also
deviate from linearity, suggesting a change in mechanism. This
behaviour is discussed below.
Influence of glycerol on diffusion-limited components of the
bimolecular inhibition rate constants
Plots of ki − 1 and k i − 1 against ηrel are largely linear for both wt and
D70G BChE when the changes in viscosity are due to glycerol
below ηrel = 3 (<37.5 % glycerol) (Figures 5 and 6, open circles).
However, at high concentration, preferential interactions between
glycerol and catalytic and non-catalytic residues cause deviations
from linearity. Therefore eqn (8) could be employed to determine
kon , k on , and the partition coefficients P = koff /kp and P = k off /k p
for the E and E forms of both enzymes for concentrations of
glycerol below 37.5 % (Table 1). Results show that ki ≈ kon for
both forms of both enzymes. Because the reaction of BChE with
CSP is linearly dependent on viscosity, it must be at least partially
diffusion-controlled, but ki (k i ) values are some three orders of
magnitude lower than the diffusion-controlled limit which is in
the range 6 × 1010 –6 × 1012 M − 1 · min − 1 for aqueous media [28].
Thus inhibition is partially limited by diffusion of the inhibitor
to the active site. This is in agreement with results obtained for
reaction of AChE with good substrates [25–27,36]. In particular,
the viscosity-dependence of kcat /K m for human AChE-catalysed
hydrolysis of acetylthiocholine in sodium phosphate buffer (ionic
strength 0.17) containing glycerol showed that kcat /K m (1.86 × 109
M − 1 · min − 1 ) is solely limited by kon = 3.42 × 109 M − 1 · min − 1
(these values were obtained by recalculation of data from Figure 2
and eqn 5 of [27]). No data are available concerning diffusion
control of BChE-catalysed hydrolysis of substrates.
c The Authors Journal compilation c 2013 Biochemical Society
At 3.50 × 108 M − 1 · min − 1 and 0.75 × 108 M − 1 · min − 1 , k0 i and
k0 i for the reaction of human BChE with CSP are certainly
among the highest known for the reaction of BChE with OPs.
However, these values are three orders of magnitude lower
than the highest bimolecular reaction rate constant (kcat /K m )
ever determined for a ChE reaction, i.e. for the ionic strengthsensitive hydrolysis of the cationic substrate acetylthiocholine
by electric eel AChE (2.5 × 1011 M − 1 · min − 1 ) at zero ionic
strength [26,37]. They are only one order of magnitude less
than kcat /K m for human AChE reacting with acetylthiocholine in
50 mM sodium phosphate buffer (1.86 × 109 M − 1 · min − 1 ) [27]
and are similar to the kcat /K m of this enzyme for hydrolysis of
the neutral substrate phenylacetate (4.85 × 108 M − 1 · min − 1 ) [37].
Also, kon = 3.42 × 109 M − 1 · min − 1 for acetylthiocholine reacting
with human AChE in 50 mM phosphate buffer [27] is only one
order of magnitude higher than our k0 on values for reaction of
human BChE with CSP.
kobs for inhibition of BChE by CSP was linearly dependent
on CSP concentration, therefore the CSP experiments were
performed at a CSP concentration far lower than K I . The
highest CSP concentration was 0.77 × 10 − 8 M. It follows
that K I 0.77 × 10 − 8 M. In addition, P = 0, i.e. koff kp .
For these two reasons, experimental determination of kp was
not possible. However, a minimum value can be tentatively
estimated by rearranging eqn (7) to k0 p = k0 i K I and assuming that
K I 0.77 × 10 − 8 M, then yields a minimum value for k0 p 2.7
min − 1 . This value places the rate for phosphorylation, kp , of
human BChE by CSP within an order of magnitude of kp of
this enzyme by most OPs under the same conditions. Typical
values for kp at 25 ◦ C pH 7.0–8.0 are: ∼ 30 min − 1 for DFP (diisopropylfluorophosphate) [38], >6 min − 1 for sarin [39], >6
min − 1 for VX [39] and 0.5 min − 1 for paraoxon [33]. This tentative
minimum estimation for kp confirms the conclusion based on kon
that CSP is one of the most reactive OPs for human BChE. Taking
kon = 4.76 × 108 M − 1 · min − 1 for the E form of wt BChE, with
koff 2.7 min − 1 , it follows that K I 5.6 × 10 − 9 M. K I values for
most OP inhibitors of BChE fall into a range between 10 − 7 and
10 − 4 M at 25 ◦ C [2,33,39].
Anomalies in the effect of viscosogens on the BChE reaction with
CSP
Plots of k0 i /ki against ηrel and k0 i /k i against ηrel describing the
effect of increasing concentrations of glycerol or sucrose on
the rate for inhibition of wt BChE or D70G by CSP are given
in Figure 7. Under all conditions, increasing the viscosogen
concentration results in a decrease in the rates for inhibition,
thereby increasing the k0 i /ki and k0 i /k i ratios (cf. eqn S3 in
the Supplementary Online Data). However, visual inspection of
these plots reveals that only the effect of glycerol on the fast
phase (Figure 7, closed grey circles) can be rationalized as being
due simply to the effects of viscosity on the bulk medium, i.e.
the effect is linear and below the maximum limit. For wt BChE
(Figure 7A), the effect of glycerol on the slow phase (closed black
circles) is always greater than predicted for diffusion-controlled
limit (broken line), and the effect of sucrose on both phases is nonlinear from the very beginning (grey and black open squares). For
D70G (Figure 7B), the data are similar to those for wt except
that the non-linearity in the effect of sucrose is more pronounced
and the effect of glycerol on the fast phase (grey filled circles)
verges on the diffusion-controlled limit (it is also non-linear at
high concentration). These qualitative observations have led us to
explore mechanisms by which sucrose and glycerol could inhibit
Viscosity and osmotic stress effects on phosphorylation of butyrylcholinesterase
393
Table 1 Differences in catalytic behaviour between E and E forms of BChE (wt and D70G mutant) from viscocity experiments in glycerol-containing buffers
and MD simulations
wt BChE
E
E
k i = 3.50 × 108 M − 1 · min − 1
k on = 4.76 × 108 M − 1 · min − 1
k on ≈ k i
k i = 0.75 × 108 M − 1 · min − 1
k on = 2.46 × 108 M − 1 · min − 1
k on >k i
k on ≈ k on
P= 0
k p k off
k p <k p
(k 0 i /k i,app )/ηrel = 3* up to ηrel = 3, then non-linearity
Reaction strongly diffusion-controlled
k i,app = 0.25 × 108 M − 1 · min − 1
k on = 0.19 × 108 M − 1 · min − 1
k i ≈ k on
P= 0
k p k off
k p <k p
(k 0 i /k i,app )/ηrel = 3.7* up to ηrel = 1.4, then non-linearity
Reaction strongly diffusion-controlled
Mutation D70G has no effect on k i and k i
Weak water network in the active-site gorge
Desolvation induces E to E shift; E : low reactivity enzyme form
P = k off /k p = 0.4
k p >k off
Viscosity control
D70G mutant
Viscosity control
Role of Asp70 in reaction of BChE with CSP
Role of Asp70
Importance of water
Catalytic machinery from MD simulations
(orientation of catalytic His438 )
Role of viscosogen from MD simulations
(k 0 i /k i,app )/ηrel = 0.75
Reaction strongly diffusion-controlled
k i,app = 3.25 × 108 M − 1 · min − 1
k on = 3.17 × 108 M − 1 · min − 1
k i ≈ k on
P= 0
k p k off
(k 0 i /k i,app )/ηrel = 1* up to ηrel = 3, then non-linearity
Reaction strongly diffusion-controlled
Asp70 is not important
Control of water network in the active site gorge
Solvation of key catalytic residue(s) yields E: optimum enzyme
form
◦
◦+ ◦
His438 χ 1 = 180◦ +
− 10 , χ 2 = 85 − 15
In glycerol solutions: observed only at the beginning of the
trajectory
◦
◦+ ◦
His438 χ 1 = 70◦ +
− 13 , χ 2 = 270 − 22
Not observed in sucrose solutions and water solution in the absence
of co-solvent
*Ratio >1 because of direct inhibition by glycerol at low concentration; non-linearity because glycerol and sucrose induced conformational changes at high concentration.
Figure 7 Plots of bimolecular rate constant ratios k 0 i /k i (grey) and k 0 i /k i (black) against relative viscosity, η/η0 , caused by glycerol (closed circles) or
sucrose (open squares) for the reaction of CSP with wt BChE (A) or the D70G mutant (B)
Apparent values of k i and k i for D70G were estimated from k obs /[CSP] and k obs /[CSP] for a CSP concentration of 6 nM. The broken lines starting from 1 correspond to the diffusion-controlled limit
(slope = 1).
the reaction of CSP with BChE other than by increasing the
viscosity of the bulk medium.
Effects of glycerol on the inhibition rate constants
Of the four plots for glycerol-containing samples (Figure 7, closed
circles), the plots for the E forms (fast phase) of both wt and D70G
BChE (grey closed circles) show slopes <1 (slope = 0.75 for wt;
slope = 1.1 for D70G which is equal to 1 within experimental
error). A slope of 0.75 (Figure 7A) indicates that reaction of wt
form E with CSP is 75 % diffusion-controlled, whereas a slope of
1 indicates that the reaction of D70G form E with CSP is 100 %
diffusion-controlled.
For the E forms (slow phase) of wt BChE and D70G (black
closed circles in Figure 7), the slope values are 3.2 and 3.5
respectively. Slopes >1 violate the physics of eqn 8 (and eqns S2
and S3 in the Supplementary Online Data) by requiring negative P
values. Therefore increased values of slopes indicate that glycerol
decreases the k i for form E of wt and D70G BChE by means
other than increasing the viscosity of the bulk medium. The
non-linear portions of the traces also suggest that there is a nonviscosity-based effect of glycerol on the inhibition rates affecting
c The Authors Journal compilation c 2013 Biochemical Society
394
P. Masson and others
the mechanism of the phosphorylation reaction. This non-linear
effect is significant beyond ηrel = 3 for wt and beyond ηrel = 1.4
for D70G. Evidence developed in the Supplementary Online Data
suggests that the non-viscosity-based effects of glycerol may
be due to classical inhibition. However, from the kinetic data
alone, we could not decide whether inhibition by glycerol was
competitive, non-competitive or of mixed type. This dilemma
was resolved by MD. MD simulations showed that glycerol
acts as a mixed-type inhibitor, hindering formation of enzyme–
inhibitor complexes and interacting with Ser198 and His438 (see the
Supplementary Online Data for details).
The non-linear portions of the traces in Figure 7 correspond to
the range of glycerol concentrations where the MD simulations
show glycerol to be acting as a mixed-type inhibitor. Changes
in slope suggest progressive changes in the mechanism and
structure of BChE as the glycerol concentration is increased. This
phenomenon is minor for wt BChE, form E (black closed circles
in Figure 7A), whereas the effect is pronounced for both E and
E forms of the D70G mutant (grey and black closed circles in
Figure 7B). For D70G, from 37.5 % glycerol and beyond the
slopes of the plots in Figure 7(B) are almost 0 for both E and
E . Such flat slopes are consistent with the partition coefficients
having become very high, i.e. koff kp . This implies that reversible
complexes E.I and E .I in Scheme 1 are in true equilibria
under these conditions. This is a major mechanistic change that
supersedes the direct competition effects. This change in kinetics
cannot be interpreted in terms of diffusional dependence of rate
constants. A highly appealing option for this additional inhibitory
effect of glycerol is interaction of glycerol with the surface of
BChE. Glycerol is known to interact with the enzyme surface,
affecting the conformational stability and the MD of the enzyme
[40].
Effect of sucrose on the bimolecular rate constants
As seen for wt BChE (grey and black open squares in Figure 7A),
increasing the sucrose concentration has a weak non-linear effect
on plots of k0 i /ki and k0 i /k i against ηrel . The non-linearity in the
plot indicates that the effect is not due to increasing the viscosity
of the bulk medium. The weakness of the effect indicates that wt
BChE is resistant to the influences of sucrose. Similarly, we found
that kinetic parameters of wt BChE-catalysed hydrolysis of BTC
were insensitive to sucrose up to 40 % [41]. MD simulations of
wt in 30 % sucrose solution revealed that sucrose molecules do
not enter the gorge, rather they interact tightly with the enzyme
surface clustering at the gorge entrance where they act as a
semi-permeable membrane (Figure 4). This suggests that sucrose
operates as an osmolyte and that osmolyte-induced dehydration
of the active-site gorge may cause the observed decrease in
the inhibition rate constant. Increasing sucrose concentration
caused a continuous small decrease in ki and k i for wt BChE
(Supplementary Table S1).
For the D70G mutant (Figure 7B) (grey and black open
squares), sucrose exerts a strong inhibitory effect at low
concentration (up to ηrel = 1.18, i.e. 6.25 % sucrose) with a slope
higher than the diffusion-controlled limit of one. As for wt BChE,
this effect can be explained in terms of osmotic stress. Below
6.5 % sucrose, this effect appears much more pronounced for
D70G than for wt. Above 6.25 %, sucrose has no additional effect
on the inhibition rate constant. This specific effect on D70G is
puzzling, and is discussed in the next section.
Assuming that the osmotic pressure (π) generated by sucrose
is the operative factor in causing the decrease in the inhibitory
rate, then the continuous small decrease in ki and k i for wt BChE
c The Authors Journal compilation c 2013 Biochemical Society
(Supplementary Table S1) suggests a moderate depletion in water
from the gorge. On the other hand, the decrease in kobs and k obs for
D70G was pronounced at the lowest concentrations of sucrose, but
stopped beyond π = 4.4–6 bar (6.25–8.33 % sucrose), suggesting
a more extensive release of water from the gorge. The difference in
behaviour suggests that for D70G, water molecules are removed
from the gorge at low π, whereas water molecules are displaced
from wt BChE more gradually, requiring greater π. This in turn
suggests that Asp70 functions as a gate to restrict efflux of the
water molecules. When the Asp70 valve is in place, efflux of water
is difficult, but when it is missing, efflux of water is easier.
Water in the active-site gorge of BChE: number, organization and
release upon reaction with CSP
π-Induced effect on inhibition provides a means for measuring
nw displaced from BChE in the course of the CSP reaction. CSP
reactivity and π are related through the osmotic activation volume
ln k
‡
= −RT ∂∂π
), i.e. the volume of water displaced when
(Vosm
CSP binds to and reacts with BChE, in the absence of sucrose,
i.e. at atmospheric pressure. V ‡ osm was determined from eqn (9)
(Figure 8), and nw was involved in the reaction from eqn (10).
Non-linearity or breaks in plots are indicative of osmolyteinduced changes in the protein structure/dynamics. Applying this
analysis to wt BChE, the negative slope gave a positive V ‡ osm
of 700 +
− 65 ml/mol, for both the E and E forms. The decrease
in ln(rate constant) with π was linear, indicating that there is
no change in mechanism or enzyme conformation as water is
removed from the active-site gorge by sucrose. Applying this
analysis to D70G gave a positive V ‡ osm of 3300 +
− 200 ml/mol,
for both the E and E forms. There was a sharp break in the
plot at π = 6 bar. Beyond this osmotic pressure, ki and k i
become independent of π. Two non-exclusive explanations can
be proposed for this sharp break: (i) all water molecules involved
in binding and reaction of CSP were displaced by π = 6 bar;
(ii) sucrose binds to D70G causing a negative osmotic volume
contribution of the same magnitude as V ‡ osm for the reaction
with CSP, making a resulting net apparent V ‡ osm = 0. This
second explanation is consistent with studies on the D70Gcatalysed hydrolysis of BTC [41]. The sucrose-induced changes
were found not to be related to either η or π. However, sucrose can
compete with water and preferentially interact with the enzyme
surface through multiple hydrogen bonds. In this way, it may
affect the conformation and/or the dynamics of the enzyme. In
particular, as hypothesized previously [41], sucrose may interact
with mobile residues located at the rim of the active-site gorge
that are connected to the active site via the -loop (Figure 2A).
In wt BChE, -loop conformational flexibility is controlled by an
interaction between PAS residues Asp70 and Tyr332 . In D70G,
such interaction is missing. Therefore sucrose interactions at
the rim of the gorge with -loop residues may allosterically
affect the activity by altering the conformation and the plasticity
of the enzyme, and by impeding the movement of BTC (or
CSP) into the active site. Therefore, for D70G, the apparent
V ‡ osm can be described by the algebraic sum of the reaction
activation volume plus volume changes resulting from removal
and/or reorganization of water on solvent-exposed residues due
to conformational changes and protein dynamics (eqn 12):
‡
‡
= Vosm,react
+
Vosm
‡
Vosm,conf/dyn
(12)
The fact that V ‡ osm was some five times higher for D70G than
for wt indicates that water is more readily released from the activesite gorge of D70G than from the gorge of the wt BChE. This is in
Viscosity and osmotic stress effects on phosphorylation of butyrylcholinesterase
395
Figure 8 Dependence of lnk i (black circles and lines) and lnk i (grey circles and lines) on sucrose-generated osmotic pressure for wt BChE (A) and
dependence of lnk obs and lnk obs for D70G at a CSP concentration of 6 nM (B)
agreement with the conclusion reached above, based on the effect
of π on the inhibition rates. It is also in agreement with previous
results obtained for the rate of aging for DFP-phosphorylated
BChE (dealkylation of one isopropyl chain) where a ratio of 5
was observed between V ‡ osm for D70G and wt enzyme [19].
These results support the statement that Asp70 at the entrance of
the gorge, hydrogen-bonded to water molecules located at the
top of the catalytic gorge (Figure 2), plays an important role in
organizing the water network, and in controlling the dynamics of
water molecules in the gorge.
A comparison of the movement of water molecules in the
catalytic gorge (i.e. below Asp70 ) for wt (Figure 9A) and D70G
(Figure 9B) BChE clearly demonstrates the organizing role of
Asp70 on water molecule motion. Figure 9 was constructed
from snapshots of the MD trajectory recorded every 10 ps. At
each time point, we calculated the nw inside the catalytic gorge
below Asp70 . For the wt enzyme (Figure 9A), an oscillating
change in the nw within the catalytic gorge can be seen. These
periodic phases, during the 50 ns trajectory, suggest that there
are alternating periods during which water is sucked down into
the gorge and then squeezed back out. These phases occur
with a period of approximately 25 ns. Approximately five water
molecules are exchanged in each phase of the cycle. If sufficient
computational resources become available, we will perform
150 ns MD simulation to validate this suggestion. We attribute
these changes to regular ‘breathing’ motions of the enzyme. For
D70G, i.e. in the absence of Asp70 , no regular change in the nw
can be observed (Figure 9B). Rather, there is a random scatter
in the calculated nw ≈10. This is probably due to thermal motion.
However, longer simulation is also needed for D70G to search for
large-scale periodic changes.
Cluster organization of the water network in the AChE gorge
has been well described [42]. As stated previously, Asp70 acts
as a one-way check valve [19]. Thus the weaker structure of
the water network in D70G compared with that of wt enzyme
facilitates stripping more water molecules out of the gorge by
osmotic pressure.
An MD study at 298 K and 1 atm showed that the maximum
nw that could be packed into the gorge for wt BChE is 67 +
− 1.5
(see the Supplementary Online Data). The crystal structure shows
29 water molecules in the gorge [16]. A difference between the
theoretical nw and the number seen in the crystal structure is
reasonable because disordered water molecules are not seen in
the crystal structure. In particular, water molecules present at the
rim of the gorge are highly mobile. In addition, the crystal was
grown in 2.1 M ammonium sulfate, thus a lack of water molecules
in the gorge is possibly a consequence of osmotic stress.
The total volume of the full gorge for wt BChE, including
the PAS, was calculated to be 1500 Å3 (1 Å = 0.1 nm) (see the
Supplementary Online Data, and Supplementary Figures S2–
S5 and Supplementary Table S3, at http://www.biochemj.org/
bj/454/bj4540387add.htm). Thus the average molecular volume
of water in the gorge is 1500/67 = 22.4 Å3 . The density
of water in confined environments is different from that
of bulk water, because the partial molar volume of water
depends on intermolecular hydrogen-bonding, and filling of void
spaces in the total volume under consideration (http://www.
lsbu.ac.uk/water/sitemap.html). Our average value (22.4 Å3 ) is
intermediate between the van der Waals volume of non-hydrogenbonded water (14.6 Å3 ) [44] and the volume of fully hydrogenbonded bulk water (30 Å3 ) [45], being close to the average
molecular volume of water confined in cavities (18 Å3 ) [46].
An estimation of nw released from wt BChE upon reaction with
CSP can be calculated from V ‡ osm = 700 +
− 65 ml/mol and the
calculated average molecular volume of a water molecule
in the gorge (22.4 Å3 ). A total of 52 +
− 5 molecules were released
from wt BChE. This number may correspond to the nw involved in
binding of CSP and phosphorylation of the enzyme. Because the
van der Waals volume of CSP is 160 Å3 , with a solvent-excluded
volume of 175.5 Å3 , it can be estimated that five to eleven water
molecules [44] are displaced when CSP enters the gorge. Then a
maximum of 41–47 water molecules could be directly involved
in the reaction of CSP with the enzyme. Only 52 of the 67 water
molecules originally in the gorge were displaced by binding and
reaction of CSP, therefore a number of water molecules (∼ 15)
originally present in the gorge are not displaced. These water
molecules are likely to be retained by strong hydrogen bonds with
amino acids lining the gorge, making them essential structural
molecules
For D70G, nw released upon reaction with CSP was calculated
to be 245 +
− 15 by assuming that the volume of the gorge for
D70G is the same as that for wt BChE. This number is far greater
than the total nw in the gorge of the enzyme, suggesting that
V ‡ osm does not solely reflect hydration changes in the activesite gorge. A large contribution to the apparent V ‡ osm thus
results from conformation/dynamics changes accompanying the
c The Authors Journal compilation c 2013 Biochemical Society
396
P. Masson and others
Figure 9 A scatter plot showing the number of the water molecules in the catalytic gorge (below Asp70 ) for wt BChE (A) and D70G mutant (B) along the MD
trajectories of these enzymes in water in the absence of sucrose
Water molecules partially lying outside the gorge boundaries were taken as a fraction of a water molecule, i.e. if one atom of a water molecule lies outside the cavity, this water molecule is considered
to be two-thirds (0.66) of a water, if only one atom is within the boundary, this water molecule is considered to be one-third (0.33) of a water.
Figure 10
The fraction of BChE present in the E form as a function of viscosogen concentration (%, w/w; open circles, glycerol; closed circles, sucrose)
(A) Wt BChE. (B) D70G mutant. Changes in the fraction of E reflect changes in the EE equilibrium.
reaction as described in eqn (12). A similar observation was made
during studies on the sucrose-induced osmotic stress of the aging
process for DFP-phosphorylated BChE [19]. In that case, taking
vw w = 22.4 Å3 , nw released from wt and D70G were calculated as
53 and 272 respectively.
A thorough analysis of osmotic stress in terms of preferential
interactions shows that V ‡ osm has to be regarded as a
phenomenological parameter [29–31]. Thus, as described in
eqn (12), V ‡ osm has to be regarded as a composite volume
change that reflects not only the nw involved in binding and
stabilization of the reaction transition state, but also the removal
of water molecules from all solute-inaccessible areas. These
areas include the active-site gorge, water-accessible cavities and
the enzyme surface. Increasing π decreases the water activity
outside the enzyme hydration layer because of protein–cosolvent preferential interactions, which in turn decreases the
overall hydration of the protein molecule, and affects both its
conformation and MD.
Thus the main conclusion to be taken from our osmotic stress
studies is that there are five times more water molecules displaced
from D70G than from wt BChE upon reaction with CSP. This
is consistent with weaker interactions of water involved in the
water network in the active-site gorge of D70G and at other sites
throughout the enzyme. Such a large difference in water released
c The Authors Journal compilation c 2013 Biochemical Society
must cause greater alterations in the dynamics and conformation
of the mutant enzyme.
MD simulation of wt and D70G BChE in sucrose solution
revealed that changes in the nw in the gorge (Figure 4) interfere
with hydration changes due to the enzyme breathing (Figure 9).
Work is in progress to separate and address both effects. There is
evidence of periodic influx and efflux of water in the presence of
sucrose, in addition to osmotic-stress-induced removal of water;
these two effects interfere. It is necessary to discuss these changes
extensively and accurately, thus it will not be part of the present
paper.
Effects of glycerol and sucrose on the equilibrium E˙E
Extrapolation of slow and fast phases of inhibition to t = 0
(Figure 3) give the initial fractions of enzyme in the E
(slow) and E (fast) forms for each concentration of viscosogen
(Figure 10). Relative amounts of E and E forms for both wt
and D70G BChE were found to change as glycerol or sucrose
concentrations increased (Supplementary Tables S1 and S2). To a
first approximation, the E˙E equilibrium is shifted towards E as
the viscosogen concentration is increased (Figure 10). However,
evolution of E is not linear and shows differences between
glycerol and sucrose. Experiments performed at different CSP
Viscosity and osmotic stress effects on phosphorylation of butyrylcholinesterase
Figure 11
397
Conformational change of His438 in MD simulation of wt BChE in 30 % glycerol solution
Torsion angles for the His438 side chain along the trajectory: χ 1 in black, χ 2 in grey (A), conformation of the catalytic triad before the flip (B), and the conformation of the catalytic triad after the flip,
taken from the snapshot with the shortest Ser198 Oγ –His438 Nε distance (C).
concentrations showed that CSP does not affect the equilibrium.
This confirms the model in Scheme 1.
In general, the effects of osmolytes on proteins result from
preferential interactions and exclusion of water molecules from
protein surfaces. Such interactions can shift protein equilibria
towards less hydrated and more compact conformations, reduce
fluctuations and increase thermodynamic conformational stability
[30,47,48]. These tendencies are reflected in the effect of sucrose
and glycerol on the E˙E . The effects of glycerol are more
pronounced than the effects of sucrose. This reflects more preferential interactions of E with glycerol than with sucrose. In
particular, sucrose is excluded from the active-site gorge whereas
glycerol is not. There is a slight shift in the equilibria towards E at
low sucrose concentrations that was unexpected, but it is likely to
be related to an osmotic effect on mobile water molecules located
in the active-site gorge and other cavities.
The effects of osmolytes are directly correlated with the effect
of the chaotropic salt LiSCN on the hysteresis of BChE (wt and
D70G) with N-methylindoxyl acetate as the substrate. Increasing
the LiSCN concentration increases the lag time needed to shift the
enzyme equilibrium toward the enzymatically active form E [49],
i.e. it stabilizes the E form. Conversely, the effects of osmolytes
are inversely correlated with the effects of hydrostatic pressure and
kosmotropic salts (ammonium sulfate and tetramethylammonium
fluoride) on the hysteretic behaviour of the enzyme. Increasing
hydrostatic pressure or kosmotropic salt concentration decreases
lag time needed to shift the enzyme equilibrium toward the
enzymatically active form [49], i.e. they stabilize the E form.
At this point, it can be said that osmotic pressure/chaotropic
salts and hydrostatic pressure/kosmotropic salts have opposite
effects. Hydrostatic pressure and kosmotropic salts increase water
structure, increasing water hydrogen-bonding at enzyme/solvent
interfaces. Osmotic pressure and chaotropic salts decrease water
structure, and reduce water hydrogen-bonding at enzyme/solvent
interfaces [50]. It follows that the E form of BChE is less hydrated
than the E form.
As it turns out, the critical difference between E and E is
very subtle and results from a change in hydration of the activesite residue His438 . Several lines of evidence suggest that the
differences in catalytic and inhibitory properties between E and E
result from changes in the conformation of His438 in the catalytic
triad [15,51]. In particular, we propose that a change in hydrogenbonding between a water molecule and His438 causes the imidazole
ring to flip which in turn disrupts the function of the catalytic
triad. Changes in the orientation of the catalytic histidine residue
have been demonstrated by X-ray crystallography for AChE and
serine proteases, and it has been suggested that the position of
His438 is an important factor in reactions with bulky substrates,
carbamylesters and OPs [52].
Support for this interpretation comes from our MD simulations
of wt and D70G BChE in 30 % glycerol solution that revealed
rotation of the His438 side chain. Details of the changes are
shown for wt BChE in Figure 11. Figure 11(A) shows that,
5 ns into the simulation, the torsion angle for the His438
side chain flips. This change is presumably induced by the
presence of glycerol. Figure 11(B) shows the conformation
before the flip (high-activity form), whereas Figure 11(C) shows
the conformation after it (low-activity form). We suggest that
this movement of His438 is responsible for the change in
activity between E and E forms. In water solution without
co-solvent, the His438 conformation maintains its initial Eform position during the whole trajectory. Structural data from
MD simulations and QM/MM (quantum mechanics/molecular
mechanics) simulations on wt and mutant BChE [51] support
c The Authors Journal compilation c 2013 Biochemical Society
398
P. Masson and others
this statement. In the absence of co-solvent (i.e. normal
hydration), this conformational change does not occur, or
occurs at far lower probability, when hydration is changed
by glycerol, the His438 conformation changes. This hydration
change is shown in Supplementary Figures S1(B) and S1(C)
at http://www.biochemj.org/bj/454/bj4540387add.htm. Normal
hydration (all water molecules surrounding His438 ) is not shown
in Figures 11(B) and 11(C) as it would obscure the His438
conformational change, and would not add any significant
information.
In addition to the specific effect of osmolytes on the hydration
and operational conformation of the catalytic triad, Kramer’s
theory for the effects of viscosity on the MD of proteins [40,53]
must be considered. According to Kramer’s theory, reaction
rates depend on viscosity through a dynamic coefficient κ.
This coefficient is introduced as a corrective pre-factor (κ<1)
in the classical transition state rate equation. κ corresponds to
the fraction of reactive trajectories that successfully cross the
activation free energy barrier (G‡ ). Thus productive trajectories
reflect the effect of viscosity on the dynamics of the enzyme
for jumping over the transition state. Results of the kinetic
experiments in the presence of osmolytes are consistent with
Kramer’s theory and suggest that an overall change in the enzyme
MD cannot be ruled out.
Conclusion
The importance of BChE as a sensitive biomarker of exposure to
low doses of TOCP has been recognized in aerotoxic syndrome.
BChE scavenges CSP, the highly toxic metabolite of TOCP.
Because the mechanism of BChE irreversible inhibition by CSP
is more complex than for most OPs, it was of great interest to
dissect the different steps of the enzyme phosphorylation by
this bulky cyclic OP. For this purpose, a methodology based
on steady-state kinetics and MD approaches was implemented
for the first time. Inhibition of BChE (wt and D70G mutant) by
CSP in buffers of increasing viscosity and/or osmotic pressure
provided new information on the diffusion control of BChE
phosphorylation, on the mechanism of phosphorylation, on the
importance of the water molecule network in the active-site gorge
on phosphorylation, on the role of Asp70 in the PAS in control
of water motion, and on hydration changes underlying the slow
equilibrium between enzyme forms E and E . The conclusions
are supported by MD simulations, which provide insight into
the E˙E interconversion mechanism, highlighting a flip in the
position of His438 . The main conclusions of the present study are
summarized in Table 1. The importance of the last conclusion goes
beyond the scope of the present study, and provides a convincing
explanation for the hysteretic behaviour of ChEs with certain
substrates and inhibitors.
AUTHOR CONTRIBUTION
Patrick Masson performed kinetic experiments and wrote the paper. Sofya Lushchekina
performed MD simulations and molecular modelling analysis. Lawrence Schopfer
participated in analysis of data and in the editing of the paper before submission. Oksana
Lockridge purified human plasma BChE, expressed the D70G mutant and participated in
the editing of the paper before submission.
ACKNOWLEDGEMENTS
We thank Dr J. Mikler (Defence Research and Development Center Suffield, Medicine Hat,
Alberta, Canada) for the gift of CSP and Dr Vladimir Mironov (Moscow State University,
Moscow, Russia) for helpful advice. We thank the Supercomputing Center of Lomonosov
c The Authors Journal compilation c 2013 Biochemical Society
Moscow State University for supercomputing time. Figures 2, 4, 11(B) and 11(C), and
Supplementary Figures S1(B) and S1(C) were prepared by means of Discovery Studio
Visualizer 3.5, freely distributed by Accelrys Software.
FUNDING
This work was partly supported by the Russian Foundation for Basic Research [project
number 12-03-31039-mol_a (to S.L.)].
REFERENCES
1 Masson, P. and Lockridge, O. (2010) Butyrylcholinesterase for protection from
organophosphorus poisons: catalytic complexities and hysteretic behavior. Arch.
Biochem. Biophys. 494, 107–120
2 Main, A. R. (1979) The mode of action of anticholinesterases. Pharmacol. Ther. 6,
579–628
3 Schopfer, L. M., Furlong, C. E. and Lockridge, O. (2010) Development of diagnostics in
the search for an explanation of aerotoxic syndrome. Anal. Biochem. 404, 64–74
4 Carletti, E., Schopfer, L. M., Colletier, J. P., Froment, M. T., Nachon, F., Weik, M.,
Lockridge, O. and Masson, P. (2011) Reaction of cresyl saligenin phosphate, the
organophosphorus agent implicated in aerotoxic syndrome, with human cholinesterases:
mechanistic studies employing kinetics, mass spectrometry, and X-ray structure analysis.
Chem. Res. Toxicol. 24, 797–808
5 Carletti, E., Colletier, J. P., Schopfer, L. M., Santoni, G., Masson, P., Lockridge, O.,
Nachon, F. and Weik, M. (2013) Inhibition pathways of the potent organophosphate CBDP
with cholinesterases revealed by X-ray crystallographic snapshots and mass
spectrometry. Chem. Res. Toxicol. 26, 280–289
6 Masson, P., Schopfer, L. M., Saxena, A., Mikler, J. and Lockridge, O. (2012) Endogenous
human plasma catalytic bioscavengers for organophosphorus compounds do not protect
against the toxicity of chemicals implicated in aerotoxic syndrome: an in vitro study.
J. Biol. Phys. Chem. 12, 89–97
7 Eto, M., Casida, J. E. and Eto, T. (1962) Hydroxylation and cyclization reactions involved
in the metabolism of tri-o -cresyl phosphate. Biochem. Pharmacol. 11, 337–352
8 Winder, C. and Balouet, J. C. (2002) The toxicity of commercial jet oils. Environ. Res. 89,
146–164
9 Kidd, J. G. and Langworthy, O. R. (1933) Paralysis following the ingestion of Jamaica
ginger extract adulterated with tri-ortho-cresyl phosphate. Bull. Johns Hopkins Hosp. 52,
39–60
10 Lotti, M. and Moretto, A. (1999) Promotion of organophosphate induced delayed
polyneuropathy by certain esterase inhibitors. Chem.–Biol. Interact. 119–120, 519–524
11 Winder, C. (2006) Hazardous chemicals on jet aircraft. Case study: jet engine oils and
aerotoxic syndrome. Curr. Top. Toxicol. 3, 65–88
12 Van Netten, C. and Leung, V. (2001) Hydraulic fluids and jet engine oil: pyrolysis and
aircraft air quality. Arch. Environ. Health 56, 181–186
13 Gupta, R. C. (ed.) (2009) Handbook of Toxicology of Chemical Warfare Agents, Academic
Press/Elsevier, Amsterdam
14 Liyasova, M., Li, B., Schopfer, L. M., Nachon, F., Masson, P., Furlong, C. and Lockridge,
O. (2011) Exposure to tri-o -cresyl phosphate in jet airplane passengers. Toxicol. Appl.
Pharmacol. 256, 337–347
15 Masson, P. (2012) Time-dependent kinetic complexities in cholinesterase-catalyzed
reactions. Biochemistry (Moscow) 77, 1147–1161
16 Nicolet, Y., Lockridge, O., Masson, P., Fontecilla-Camps, J. C. and Nachon, F. (2003)
Crystal structure of human butyrylcholinesterase and of its complexes with substrate and
products. J. Biol. Chem. 278, 41141–41147
17 Masson, P., Legrand, P., Bartels, C. F., Froment, M. T., Schopfer, L. M. and Lockridge, O.
(1997) Role of aspartate 70 and tryptophan 82 in binding of succinyldithiocholine to
human butyrylcholinesterase. Biochemistry 36, 2266–2277
18 Macdonald, I. R., Martin, E., Rosenberry, T. L. and Darvesh, S. (2012) Probing the
peripheral site of human butyrylcholinesterase. Biochemistry 51, 7046–7053
19 Masson, P., Cléry, C., Guerra, P., Redslob, A., Albaret, C. and Fortier, P. L. (1999)
Hydration change during the aging of phosphorylated human butyrylcholinesterase:
importance of residues aspartate-70 and glutamate-197 in the water network as probed by
hydroastatic and osmotic pressure. Biochem. J. 343, 361–369
20 Lockridge, O., Schopfer, L. M., Winger, G. and Woods, G. H. (2005) Large scale
purification of butyrylcholinesterase from human plasma suitable for injection into
monkeys; a potential new therapeutic for protection against cocaine and nerve agent
toxicity. J. Med. Chem. Biol. Radiol. Def. 3, nihms5095
21 Ellman, G. L., Courtney, K. D., Andres, V. and Featherstone, R. M. (1961) A new and rapid
colorimetric determination of acetylcholinesterase activity. Biochem. Pharmacol. 7, 88–95
22 Nakatani, H. and Dunford, H. B. (1979) Meaning of diffusion-controlled association rate
constants in enzymology. J. Phys. Chem. 83, 2662–2665
Viscosity and osmotic stress effects on phosphorylation of butyrylcholinesterase
23 Brouwer, A. C. and Kirsch, J. F. (1982) Investigation of diffusion-limited rates of
chymotrypsin reactions by viscosity variation. Biochemistry 21, 1302–1307
24 Hasinoff, B. B., Dreher, R. and Davey, J. P. (1987) The association reaction of yeast
alcohol dehydrogenase with coenzyme is partly diffusion-controlled in solvents of
increased viscosity. Biochim. Biophys. Acta 911, 53–58
25 Hasinoff, B. B. (1982) Kinetics of acetylcholine binding to electric eel acetylcholinesterase
in glycerol/water solvents of increased viscosity. Biochim. Biophys. Acta 704,
52–58
26 Bazelyansky, M., Robey, E. and Kirsch, J. F. (1986) Fractional diffusion-limited
component of reactions catalyzed by acetylcholinesterase. Biochemistry 25, 125–130
27 Malany, S., Sawai, M., Sikorski, R. S., Seravalli, J., Quinn, D. M., Radic, Z., Taylor, P.,
Kronman, C., Velan, B. and Shafferman, A. (2000) Transition state structure and rate
determination for the acylation stage of acetylcholinesterase catalyzed hydrolysis of
(acetylthio)choline. J. Am. Chem. Soc. 122, 2981–2987
28 Eigen, M. and Hames, G. G. (1963) Elementary steps in enzyme reactions. Adv. Enzymol.
Relat. Areas Mol. Biol. 25, 1–28
29 Parsegian, V. A., Rand, R. P. and Rau, D. C. (2000) Osmotic stress, crowding, preferential
hydration, and binding: a comparison of perspectives. Proc. Natl. Acad. Sci. U.S.A. 97,
3987–3992
30 Timasheff, S. N. (2002) Protein-solvent preferential interactions, protein hydration, and
the modulation of biochemical reactions by solvent components. Proc. Natl. Acad. Sci.
U.S.A. 99, 9721–9726
31 LiCata, V. J. and Allewell, N. M. (1998) Measuring hydration changes of proteins in
solution: applications of osmotic stress and structure-based calculations. Methods
Enzymol. 295, 42–61
32 Sadovnichy, V., Tikhonravov, A., Voevodin, V. I. and Opanasenko, V. (2013) “Lomonosov”:
supercomputing at Moscow State University. Contemporary High Performance
Computing: from Petascale toward Exascale (Chapman & Hall/CRC Computational
Science), CRC Press, Boca Raton, pp. 283–307
33 Masson, P., Froment, M.-T., Bartels, C. F. and Lockridge, O. (1997) Importance of
aspartate-70 in organophosphate inhibition, oxime re-activation and aging of human
butyrylcholinesterase. Biochem. J. 325, 53–61
34 Wandhammer, M., Carletti, E., van der Schans, M., Gillon, E., Nicolet, Y., Masson, P.,
Goeldner, M., Noort, D. and Nachon, F. (2011) Structural study of the complex
stereoselectivity of human butyrylcholinesterase for the neurotoxic V-agents. J. Biol.
Chem. 286, 16783–16789
35 Worek, F., Eyer, P. and Szinicz, L. (1998) Inhibition, reactivation and aging kinetics of
cyclohexylmethylfluorophosphate-inhibited human cholinesterases. Arch. Toxicol. 72,
580–587
36 Nolte, H., Rosenberry, T. L. and Neumann, E. (1980) Effective charge on
acetylcholinesterase active sites determined from the ionic strength dependence of
association rate constants with cationic ligands. Biochemistry 19, 3705–3711
399
37 Quinn, D. (1987) Acetylcholinesterase: enzyme structure, reaction dynamics, and virtual
transition states. Chem. Rev. 87, 955–979
38 Main, A. D. (1964) Affinity and phosphorylation constants for the inhibition of esterases
by organophosphates. Science 144, 992–993
39 Millard, C. B., Lockridge, O. and Broomfield, C. A. Design and expression of
organophosphorus acid anhydride hydrolase activity in human butyrylcholinesterase.
Biochemistry 34, 15925–15993
40 Finkelstein, I. J., Massari, A. M. and Fayer, M. D. (2007) Viscosity-dependent protein
dynamics. Biophys. J. 92, 3652–3662
41 Levitsky, V., Xie, W., Froment, M.-T., Lockridge, O. and Masson, P. (1999) Polyol-induced
activation by excess substrate of the D70G butyrylcholinesterase mutant. Biochim.
Biophys. Acta 1429, 422–430
42 Koellner, G., Kryger, G., Millard, C. B., Silman, I., Sussman, J. L. and Steiner, T. (2000)
Active-site gorge and buried water molecules in crystal structures of acetylcholinesterase
from Torpedo californica . J. Mol. Biol. 296, 713–35
43 Reference deleted
44 Sonavane, S. and Chakrabarti, P. (2008) Cavities and atomic packing in protein structures
and interfaces. PLoS Comput. Biol. 4, e1000188
45 Scherer, J. R. (1987) The partial molar volume of water in biological membranes. Proc.
Natl. Acad. Sci. U.S.A. 84, 7938–7942
46 Till, M. S. and Ullmann, G. M. (2010) McVol: a program for calculating protein volumes
and identifying cavities by a Monte Carlo algorithm. J. Mol. Model. 16, 419–429
47 Kim, Y. S., Jones, L. S., Dong, A., Kendrick, B. S., Chang, B. S., Manning, M. C.,
Randolph, T. W. and Carpenter, J. F. (2003) Effects of sucrose on conformational equilibria
and fluctuations within the native-state ensemble of proteins. Protein Sci. 12, 1252–1261
48 Vagenende, V., Yap, M. G. S. and Trout, B. L. (2009) Mechanisms of protein stabilization
and prevention of protein aggregation by glycerol. Biochemistry 48, 11084–11096
49 Masson, P., Froment, M. T., Fort, S., Ribes, F., Bec, N., Balny, C. and Schopfer, L. M.
(2002) Butyrylcholinesterase-catalyzed hydrolysis of N -methylindoxyl acetate: analysis of
volume changes upon reaction and hysteretic behavior. Biochim. Biophys. Acta 1597,
229–243
50 Cacace, M. G., Landau, E. M. and Ramsden, J. J. (1997) The Hofmeister series: salt and
solvent effects on interfacial phenomena. Q. Rev. Biophys. 30, 241–277
51 Lushchekina, S., Masson, P., Nemukhin, A. V. and Varfolomeev, S. D. (2012)
Computational modeling of hysteresis in Ala328Cys mutant of human
butyrylcholinesterase. 11th International Meeting on Cholinesterases, Kazan, Russia,
4–9 June 2012, Abstract L1-9
52 Millard, C. B., Koellner, G., Ordentlich, A., Shafferman, A., Silman, I. and Sussman, J. L.
(1999) Reaction products of acetylcholinesterase and VX reveal a mobile histidine in the
catalytic triad. J. Am. Chem. Soc. 121, 9883–9884
53 Hammes-Schiffer, S. (2013) Catalytic efficiency of enzymes: a theoretical analysis.
Biochemistry 52, 2012–2020
Received 15 March 2013/10 June 2013; accepted 20 June 2013
Published as BJ Immediate Publication 20 June 2013, doi:10.1042/BJ20130389
c The Authors Journal compilation c 2013 Biochemical Society
Biochem. J. (2013) 454, 387–399 (Printed in Great Britain)
doi:10.1042/BJ20130389
SUPPLEMENTARY ONLINE DATA
Effects of viscosity and osmotic stress on the reaction of human
butyrylcholinesterase with cresyl saligenin phosphate, a toxicant related
to aerotoxic syndrome: kinetic and molecular dynamics studies
Patrick MASSON*†‡1 , Sofya LUSHCHEKINA§, Lawrence M. SCHOPFER* and Oksana LOCKRIDGE*
*Eppley Institute, University of Nebraska Medical Center, Omaha, NE 68198-5950, U.S.A., †Département de Toxicologie, Institut de Recherche Biomédicale des Armées (IRBA)-Centre
de Recherches du Service de Santé des Armées (CRSSA), 24 av des Maquis du Grésivaudan, 38702 La Tronche, France, ‡Laboratoire de Biophysique Moléculaire, Institut de Biologie
Structurale, 41 rue Jules Horowitz, 38027 Grenoble, France, and §Modeling of Biomolecules Laboratory, N.M. Emanuel Institute of Biochemical Physics, Russian Academy of Sciences,
4 Kosygina St., 119334 Moscow, Russia
VISCOSITY EFFECTS ON BChE INHIBITION BY CSP: THEORETICAL
BACKGROUND, PRACTICAL ASPECTS AND COMPLICATIONS
The dynamic viscosity of viscosogen-containing buffers, η, at
25 ◦ C was obtained from tables for aqueous solutions of glycerol
and sucrose [1,2]. For analysis of kinetic data, the viscosity was
expressed as relative viscosity, ηrel = η/η0 [3], i.e. the ratio of
the dynamic viscosity of buffer containing viscosogen, η, to the
viscosity of reaction buffer containing no viscosogen, η0 . The
value of η0 for 0.1 M sodium phosphate buffer (pH 8.0) at 25 ◦ C,
was calculated as 1.0 cP (mPa·s) from interpolation of reported
viscosity values of diluted solutions of Na2 HPO4 and NaH2 PO4 at
this temperature [4], using the Henderson–Hasselbach equation
for pH 8.0.
As it turns out, for the reaction of CSP with BChE, koff kp ,
therefore kp is not diffusion-controlled. This conclusion is based
on the viscosity-dependence of ki which showed that both kon and
koff depend on the dynamic viscosity of the medium, η, (i.e. k0 on η =
kon η0 and k0 off η = koff η0 ) and are therefore dependent, at least in
part, on the diffusion of CSP and BChE through the medium. The
diffusion coefficients, D, for BChE and CSP are related to
the dynamic viscosity by the Stokes–Einstein equation [5]:
D=
kB T
6πηri
(S1)
In eqn (S1), kB is Boltzmann’s constant, T is the absolute
temperature, and ri is the molecular radius of the molecule
(regarded as a sphere). The inhibition rate constant, ki , depends
on viscosity according to eqn (S2) [3]:
0
0
η0
koff
0 η
ki = kon
1+
(S2)
η
kp η
In this equation, η0 is the viscosity of the buffer in the absence
of added viscosogen. A normalized plot for the dependence of ki
or k i on relative viscosity is described by eqn (S3). This equation
is the ratio of eqn (S2), in which ηrel = 1, i.e. ki 0 , to eqn (S2) in
which ηrel >1, i.e. ki [6]:
ki0
P
P
+
ηrel
=
ki
1+ P
1+ P
(S3)
The slope of eqn (S3) is d(k0 i /ki )/dηrel = 1/(1 + P). For
bimolecular reactions that are fully rate-limited by diffusion,
P = 0, and the slope of eqn (S3) should be 1. For reactions that are
not diffusion-controlled, P will be very large, making 1/(1 + P)
1
very small, making the slope effectively 0. The slope for reactions
that are partially diffusion-controlled should be between 0 and 1.
This theory describes the consequences on reaction kinetics that
arise from changing the viscosity of the bulk medium. Effects due
to viscosity should be the same regardless of the chemical nature
of the viscosogen. If the slopes for plots of k0 i /ki against ηrel
differ in the presence of different viscosogens, then perturbations
of ki in addition to those due to viscosity of the bulk medium
are occurring. Additional perturbations of ki are indicated if the
slope is non-linear or if the slope exceeds 1. Such additional
perturbations could include osmotic effects, direct inhibition by
the viscosogen on the reaction, or interactions of the viscosogen
with the surface of the enzyme that alter the structure and/or
dynamics of the enzyme.
Glycerol appears to affect the inhibition rate by direct
competition with the binding of CSP to BChE, i.e. acting as a
competitive inhibitor, and by perturbing the partition coefficient
(k off /k p ), i.e. acting as a non-competitive inhibitor. A similar
inhibition phenomenon was reported for turnover kinetics of
electric eel AChE with acetylthiocholine when the reaction was
performed with increasing glycerol concentrations. The authors
for that study modified eqn (S2) to accommodate the observations,
but the type of inhibition was not firmly established [6].
If glycerol acts as a competitive inhibitor, then introducing the
inhibition factor 1 + [glycerol]/K d lowers the apparent inhibition
constant by 1/(1 + [glycerol]/K d ). As used here, K d is the overall
dissociation constant for the binding of one or more glycerol
molecules into the active-site gorge. Since P = 0, and ki = kon ,
the slope of eqn (S3) would be increased by 1 + [glycerol]/K d ,
making it >1. Alternatively, if glycerol acts as a non-competitive
inhibitor, then the inhibition factor 1 + [glycerol]/K d would affect
the partition coefficient, koff /kp , by decreasing kp according
to kp /(1 + [glycerol]/K d ). In that case, increasing the glycerol
concentration would change the mechanism so that kp > koff or
kp ≈ koff , thereby making 0<P 1. With this scenario, the slope
of eqn (S3) would also be >1.
MD simulation resolved this dilemma. During the MD
trajectories for wt and D70G BChE in 30 % glycerol solution,
glycerol molecules entered the gorge shortly after the beginning
of the simulation (at 320 ps for wt and 120 ps for D70G BChE).
The number of glycerol molecules entering the gorge of wt BChE
during the MD simulation (full gorge and catalytic gorge below
Asp70 ) is shown in Figure S1(A). Glycerol molecules remained in
the gorge during the rest of the simulation. For a portion of the
time, i.e. in a fraction of snapshots, glycerol molecules were seen
to be interacting directly with the catalytic triad residues, making
To whom correspondence should be addressed (email [email protected]).
c The Authors Journal compilation c 2013 Biochemical Society
P. Masson and others
hydrogen bonds with the side chains of Ser198 and His438 (Figure
S1B). As such, they were acting as non-competitive inhibitors. In
all snapshots, glycerol molecules were found in the gorge between
the entrance and the catalytic triad (Figure S1C). In this location,
they hindered CSP traffic and formation of productive Michaelian
complexes. As such, they were acting as competitive inhibitors.
Thus glycerol appears to acting as a mixed-type inhibitor. It
disorganizes the water network by forming hydrogen bonds with
water molecules, so that important water molecules involved in the
functional activity of the catalytic triad are displaced or replaced
by glycerol (Figure S1B) and it physically obstructs passage
of substrates through the gorge. This inhibition appears to be
effective at the lowest glycerol concentration.
MOLECULAR DYNAMICS
The PDB structure of human BChE (code 1P0I) [7] was
used. The resolution was 2.00 Å, the structure lacks some amino
acids, i.e. Asp378 , Asp379 and Gln455 , and all the gaps are located on
the protein surface far from the gorge entrance. The positions of
Asp378 and Asp379 were taken from another BChE X-ray structure
(PDB code 2PM8) [8], and Gln455 was reconstructed by means
of the VMD psfgen module [9]. Water molecules seen in the
crystal structure were kept. To form the solvent box, TIP3P
(transferable intermolecular potential 3P) water molecules were
added by means of the VMD solvate module. To make the systems
electroneutral, an appropriate number of counterions was added
by means of the VMD autoionize module: three Cl − for wt BChE
systems and four Cl − for D70G BChE systems. Clusters of cosolvents (sucrose and glycerol) were added by means of VegaZZ
software [10]. The number of co-solvent and water molecules was
adjusted to make 30 % (w/w) solution. The total number of water
and co-solvent molecules added and the sizes of the resulting
systems are provided in Tables S1–S3.
MD simulations were performed with the NAMD 2.8
program [11] using the Lomonosov Moscow State University
supercomputer [12]. The recently published CHARMM36 forcefield program with revised parameters for proteins [13], sucrose
[14] and glycerol [15], plus the TIP3P water model was used.
Structures of wt and D70G BChE were initially saturated with
water molecules and energy-minimized (see the next section for
details). Co-solvent solutions were added to the water-saturated
structures. Full dynamics of each system were performed for
50 ns under periodic boundary conditions at constant temperature
of 300 K and constant pressure of 1 atm (the NPT ensemble).
The BChE gorge volume was determined by means of the
McVol program [16]. Structural analysis was performed with
VegaZZ [10] and VMD [9] software packages.
DEFINITION OF THE GORGE OF WT BChE AND NUMBER OF WATER
MOLECULES
The wt BChE catalytic gorge is currently defined as the space
between Asp70 located at the rim of the gorge and the catalytic
residues (Ser198 and His438 ) at the bottom. Residues located above
Asp70 are considered to be part of the PAS and gorge mouth
[17,18]. However, osmotic stress affects all water molecules
present in the gorge from the mouth to the bottom. For this reason,
we implemented two ways for determination of the gorge volume:
(i) the full gorge from the mouth to the bottom (Figure 2A in the
main paper), and (ii) the catalytic gorge from Asp70 to the bottom
(Figure 2B in the main paper). As reported in the main paper, the
volume of the full gorge was found to be 1500 Å3 , and the volume
c The Authors Journal compilation c 2013 Biochemical Society
of the catalytic gorge was calculated as 692 Å3 . These volumes
were calculated using McVol [16] software.
The number of water molecules inside the gorge was calculated
for the full gorge and for the catalytic gorge. For those water
molecules partly inside the gorge, partial values were used, i.e.
one-third of a molecule was used if one atom was inside the
gorge volume and two-thirds if two atoms were inside. The
total number of resolved water molecules in the crystallographic
structure of human BChE 1P0I [7] (the structure used for this
study) is 481; 29 of them are in the full gorge, and 19 are
in the catalytic gorge. It should be noted that highly mobile
water molecules are not seen in the crystal structure, therefore
more waters could be associated with the gorge and other parts
of the structure than are seen in the crystallographic data. In
addition, crystals were grown in 2.1 M (NH4 )2 SO4 [7], thus
osmotic displacement of water molecules out of the gorge is
possible. Consequently, it is necessary to computationally saturate
the gorge with water to obtain an accurate value for the number
of water molecules that could be present. For this purpose, a
1 ns MD simulation of BChE in water solution was performed
(using the NPT ensemble: 298 K, 1 atm pressure). The protein
co-ordinates were fixed for two reasons: (i) to avoid artificial
changes in the enzyme structure induced by low pressure in the
unsaturated gorge, and (ii) to avoid changing the number of water
molecules as a consequence of protein conformational changes.
We calculated the number of water molecules within the full gorge
and within the catalytic gorge using a three-step process. First, an
MD simulation was performed for 500 ps (Figure S2A), then the
structure was minimized (both protein and water molecules), and
finally a second run of 500 ps was performed (Figure S2B). Figure
S2(A) shows that, after the first 100 ps, the gorge was already
saturated. During the second 500 ps (Figure S2B), the number
of water molecules inside the gorge did not change significantly.
However, there is some dispersion due to thermal motion. The
average number of water molecules (calculated from the second
500 ps run) was found to be 67.2 +
− 1.5 inside the full gorge, and
29.6 +
− 1.2 inside the catalytic gorge.
COMPARISON OF OUR VALUE FOR THE CATALYTIC GORGE
VOLUME WITH PREVIOUSLY CALCULATED VOLUMES
In a recent study by Pezzementi et al. [19], the volume of the
catalytic gorge for human BChE, from Asp70 to the bottom, was
calculated to be 690 Å3 , using HOLLOW software. This value
is in excellent agreement with our value of 692 Å3 . In another
study by Saxena et al. [20], the volume of the human BChE
gorge was reported as 500 Å3 . This latter value was calculated
using the VOIDOO program [21] with a probe radius of 1.4 Å,
ten refinement cycles and other parameters apparently at their
default program values. The volume calculations were performed
for AChE, BChE and certain mutants. To convert open gorges
into closed spaces, planar covers consisting of 126 carbon atoms
were placed over the entrances of the gorges. The position of the
covers is reported as “around residues 71, 74, 276, 280, 285 and
335 in Torpedo AChE” [20]. This description is consistent with
the definition of the full gorge as we have been using the term.
To address the discrepancy between the full gorge volume we
calculated (1500 Å3 ) and the gorge volume value reported in [20]
(500 Å3 ), we calculated the gorge volume using the VOIDOO
program in two ways. First, we repeated the calculation according
the procedure described by Saxena et al. [20] for human BChE.
A probe radius of 1.4 Å and ten refinement cycles were used,
whereas all other parameters were set to their default values. We
placed a planar cover over the entrance of the gorge at BChE
Viscosity and osmotic stress effects on phosphorylation of butyrylcholinesterase
Figure S1
Glycerol molecules in the active-site gorge of human wt BChE from MD simulations using 30 % glycerol in water
(A) Number of glycerol molecules entering the full gorge (red trace) and the active-site gorge (blue trace) during a 50 ns simulation. (B) Glycerol molecules hydrogen-bonded to side chains of the
catalytic residues. Direct hydrogen bonds with Ser198 and His438 were found in 9.6 % of the snapshots from the MD trajectory for wt and 8 % of the snapshots for D70G. Glycerol hydroxy groups were
found within 3 Å of the catalytic side-chain residues Ser198 Oγ and His438 Nε in 36 % of the snapshots for wt BChE and 20 % of the snapshots for D70G. For these latter snapshots, the distances
were still favourable for hydrogen-bond formation but the O-H . . . O(N) angles deviated from linearity by more than 20◦ , making hydrogen-bond formation unfavourable. (C) Glycerol molecules in
the catalytic gorge between the entrance (defined by Asp70 and Tyr332 ) and catalytic site (defined by Ser198 , His438 and Trp82 ). In these positions, glycerol would interfere with the inhibitor/substrate
trafficking pathway.
Table S1
Effect of viscosity on inhibition of human wt BChE by CSP at pH 8.0 and 20 ◦ C with (a) glycerol and (b) sucrose as the viscosogen
Bimolecular rate constants k i and k i were determined by linear regression from plots of k obs or k obs against CSP concentration (eqn 11 in the main text). Experiments were performed in triplicate;
+
rate constants are expressed as average values; S.D. values for k i are +
−7 % of the average values, and for k i are −9 % of the average values. ND, not determined. Experiments at 3.125, 6.25 and
8.33 % sucrose were carried out at only one CSP concentration (6 nM).
(a)
Glycerol concentration (%, w/w)
ηrel
k i (M − 1 ·min − 1 )
k 0 i /k i
k i (M − 1 ·min − 1 )
k 0 i /k i
E
E
0
6.25
12.5
25
37.5
50
1
1.15
1.40
2.10
3.05
5.92
3.50 × 108
3.02 × 108
2.93 × 108
1.81 × 108
1.55 × 108
0.8 × 108
1
1.16
1.19
1.96
2.26
4.37
0.75 × 108
0.43 × 108
0.34 × 108
0.18 × 108
0.10 × 108
0.07 × 108
1
1.74
2.20
4.16
7.50
10.71
0.52
0.47
0.31
0.15
0.13
0.10
0.48
0.53
0.69
0.85
0.87
0.90
ηrel
π (bar)
k i (M − 1 ·min − 1 )
k 0 i /k i
k i (M − 1 ·min − 1 )
k 0 i /k i
E
E’
1
2.25
4.43
6.02
9.04
18.10
27.16
36.23
3.50 × 10
ND
ND
ND
2.60 × 108
2.44 × 108
1.88 × 108
1.30 × 108
1
0.75 × 10
ND
ND
ND
0.50 × 108
0.44 × 108
0.35 × 108
0.25 × 108
1
0.52
0.58
0.63
0.62
0.48
0.43
0.36
0.27
0.48
0.42
0.37
0.38
0.52
0.57
0.64
0.73
(b)
Sucrose concentration (%, w/w)
0
3.125
6.25
8.33
12.5
25
37.5
50
1
1.07
1.18
1.26
1.41
2.12
4.33
12.40
8
1.37
1.46
1.89
3.74
8
1.50
1.70
2.14
3.00
c The Authors Journal compilation c 2013 Biochemical Society
P. Masson and others
Table S2
Effect of viscosity on inhibition of the human BChE mutant D70G by CSP at pH 8.0, 25 ◦ C with (a) glycerol and (b) sucrose as the viscosogen
Inhibition was conducted at pH 8.0 and 25 ◦ C using only 6 nM CSP. Experiments were performed in triplicate; rate constants are expressed as average values; S.D. values for k obs are +
−20 % of
average values, and for k obs are +
−13 % of average values.
(a)
Glycerol concentration (%, w/w)
ηrel
k obs (min − 1 )
k 0 i /k i
k obs (min − 1 )
k 0 i /k i
E
E
wt k obs (min − 1 )
wt k obs (min − 1 )
0
6.25
12.5
25
37.5
50
1
1.15
1.40
2.10
3.05
5.92
1.918
1.648
1.087
1.070
0.559
0.492
1
1.16
1.76
1.79
3.43
3.89
0.150
0.087
0.056
0.049
0.038
0.040
1
1.72
2.67
3.06
3.94
3.75
0.65
0.59
0.60
0.36
0.30
0.25
0.35
0.41
0.40
0.64
0.70
0.75
2.172
1.196
1.132
1.151
0.931
0.392
0.401
0.254
0.130
0.086
0.064
0.034
Sucrose concentration (%, w/w)
ηrel
π (bar)
k obs (min − 1 )
k 0 i /k i
k obs (min − 1 )
k 0 i /k i
E
E’
wt k obs (min − 1 )
wt k obs (min − 1 )
0
3.125
6.25
8.33
12.5
25
37.5
50
1
1.07
1.18
1.26
1.41
2.12
4.33
12.40
1
2.25
4.43
6.02
9.04
18.10
27.16
36.23
1.918
2.234
1.292
1.199
1.423
1.189
0.936
1.309
1
0.85
1.48
1.59
1.35
1.61
2.05
1.46
0.150
0.158
0.096
0.075
0.090
0.090
0.091
0.096
1
0.95
1.56
2.00
1.66
1.66
1.64
1.56
0.65
0.71
0.72
0.73
0.71
0.62
0.55
0.48
0.35
0.29
0.28
0.27
0.29
0.38
0.45
0.52
2.172
1.533
1.304
1.378
1.264
1.644
1.119
0.713
0.401
0. 355
0.300
0.320
0.240
0.220
0.200
0.160
(b)
Table S3
The main parameters of the wt and D70G BChE systems used for MD simulations
System
BChE
Number of water molecules
Number of co-solvent molecules in the system
Total number of atoms in the system
Equilibrated cell size
Water solution
wt
D70G
wt
D70G
wt
D70G
16365
16364
17723
17722
11956
11883
–
–
1288
1288
275
275
57 399
57 392
79 505
79 498
56 547
56 324
78.9 Å × 78.0 Å × 90.7 Å
78.3 Å × 77.4 Å × 90.1 Å
92.2 Å × 87.7 Å × 92.3 Å
91.6 Å × 87.2 Å × 91.8 Å
75.5 Å × 81.7 Å × 85.3 Å
75.4 Å × 81.5 Å × 85.2 Å
30 % Glycerol solution
30 % Sucrose solution
residues homologous with the TcAChE (Torpedo californica
AChE) residues mentioned in [20]: Ile69 , Ser72 , Leu274 , Phe278 ,
Gly283 and Gly333 . These residues were chosen on the basis of the
structural alignment presented in [17]. The mouth of the BChE
gorge has a ‘depression’ at Ile69 [17], thus all of the amino acids
listed cannot be placed at one plane. For this reason, the plane was
placed over the last five amino acids listed above, and additional
carbon atoms were placed to close the indent between the plane
and Ile69 (Figure S3). The volume was then calculated, yielding
580 Å3 . We attribute the difference between this result and the
value in [20] to differences in the position of the covering plane.
We consider this 13 % difference to be acceptable. Our choice
of cover permits more water molecules from the BChE gorge to
be included in the calculated gorge volume, and thus corresponds
better with the aims of our study.
In our second calculation, we took advantage of another method
for volume calculation available in the VOIDOO program. In the
paper describing the VOIDOO program [21], it is explained that
the volume of cavity calculated by the program by default is
the volume encompassed with the probe-accessible surface: “The
accessible surface is that described by the centre of a probe of
an appropriate radius (typically, a water molecule is used with a
radius between 1.4 and 1.6 Å) when it is rolled over the protein’s
van der Waals surface” [21]. This corresponds to the solventaccessible surface definition developed by Richards [22], also
known as the Richards’ surface. The other possibility provided
by VOIDOO is calculation of the probe-occupied volume, i.e.
the volume accessible not only to the centre of a probe, but also
c The Authors Journal compilation c 2013 Biochemical Society
to any point of a probe rolling over the protein surface (Figure
S4). This corresponds to the method developed by Connolly
[23]. This surface is known as the solvent-excluded molecular
surface or the Connolly surface (a detailed historical overview of
different molecular surfaces by Michael L. Connolly is available at
http://www.netsci.org/Science/Compchem/feature14.html). Calculation of the full gorge volume defined as the probe-occupied
cavity provides a volume of 1581 Å which is very close to the
volume defined by the McVol software, reported in the main paper
(‘Effect of sucrose-induced osmotic stress on the bimolecular
rate constants’ section). Thus the apparent discrepancy with the
previously reported gorge volume is explained. In Figure S5, both
calculated surfaces are shown.
A probe-accessible (solvent-accessible) surface by definition
describes the volume where the centre of the oxygen atoms in
water molecules can be found. It does not consider the volume
occupied by the water molecule as a totality, and thus many
hydrogen atoms lie outside the volume. Furthermore, because
of an arbitrary choice for the probe radius [“The reason people
often use 1.4 Å as the radius for a water molecule is based on
the observation that the O–O distance in ice is ∼2.8 Å. While
this hardly justifies the use of a 1.4 Å radius for a solvent water,
the fact that many have used this value and are still using it
means that you can compare numbers slightly better than when
you are using another (arbitrary) number (e.g. 1.2 Å or 1.8 Å).”
VOIDOO Manual, http://xray.bmc.uu.se/usf/voidoo_man.html],
oxygen atom centres for some water molecules can be accidently
excluded from the probe-accessible area. A number of such
Viscosity and osmotic stress effects on phosphorylation of butyrylcholinesterase
Figure S4 Representation of the probe-accessible surface (solventaccessible surface) and the probe-occupied surface (solvent-excluded
surface)
Based on data taken from [22] and [23] respectively.
Figure S2
The number of water molecules in the gorge of BChE
Data for the full gorge are shown in blue and data for the catalytic gorge (below Asp70 ) are
shown in red. Water saturation was performed using MD simulations with protein co-ordinates
fixed. (A) The number of water molecules calculated during the first 500 ps of simulation, before
minimization. (B) The number of water molecules calculated during an additional 500 ps of
simulation, after minimization.
Figure S5 Gorge volumes calculated by the VOIDOO program [21] using
both the probe-accessible and probe-occupied options
The probe-accessible surface is purple and opaque. The probe-occupied surface is light blue and
transparent. Water molecules shown are the same as those in Figure 2(A) in the main paper. Three
catalytic gorge water molecules with oxygen atom centres lying outside the probe-accessible
volume are shown as ball and stick. All other water molecules are shown as sticks.
Figure S3 Covering the mouth of the BChE gorge for VOIDOO volume
calculations
Carbon atoms placed to cover the gorge are shown as grey balls. A portion of the calculated
gorge volume is shown in light blue.
water molecules can be seen at the top of the gorge in Figure
S5. This is not a big problem because the mouth of the gorge
is a transiently opened area, and it is difficult to assign water
molecules unambiguously as belonging to the gorge or to the
surrounding solvent. But in the catalytic area, it is unacceptable
to exclude water molecules from consideration because they are
crucial for BChE biophysical studies. In Figure S5, the centres
of the oxygen atoms for three of the water molecules present in
the catalytic gorge can be seen lying outside the probe-accessible
surface (shown as ball and stick). This demonstrates a deficiency
of the probe-accessible approach for defining the gorge volume
and the number of water molecules it contains.
Choice between probe-accessible and probe-occupied surfaces
is also discussed by Kleywegt and Jones [21]: “(a) probe-occupied
cavities can only be reliably used if the cavity is closed when
the atoms have their normal van der Waals radii; (b) accessible
cavities show the volume of space available to the centre of
the probe sphere, which is more meaningful when stick models
of ligands etc. are used; (c) probe-occupied cavities must be
calculated on fine grids in order to yield reliable results, which
increases the amount of CPU time needed for the calculations.”
In our case, we artificially converted the gorge into a closed
cavity by placing a cover over the entrance. Because we used a
ball-and-stick water molecule representation, the probe-occupied
volume is more meaningful. As computational resources have
been considerably improved since 1994, when Kleywegt and
Jones published their paper, the CPU time needed for volume
calculation is not a limitation any more, even for dense grids.
c The Authors Journal compilation c 2013 Biochemical Society
P. Masson and others
For these reasons, we conclude that probe-occupied volume
is more useful for our purposes as it defines the volume filled
by water molecules including the oxygen and hydrogen atoms
as spheres with the correct van der Waals radii. The full
gorge volume, determined independently using different software
(McVol and VOIDOO, using the probe-occupied volume option),
is close to 1500 Å3 . This volume encompasses all the water
molecules in the catalytic part of the gorge.
REFERENCES
1 Segur, J. B. and Oberstard, H. E. (1951) Viscosity of glycerol and its aqueous solutions.
Ind. Eng. Chem. 43, 2117–2120
2 Swindells, J. F., Snyder, C. F., Hardy, R. C. and Golden, P. E. (1958) Viscosity of sucrose
solutions at various temperatures: tables of recalculated values. Natl. Bur. Stand. Circ.
(U.S.) 440, 1–7
3 Nakatani, H. and Dunford, H. B. (1979) Meaning of diffusion-controlled association rate
constants in enzymology. J. Phys. Chem. 83, 2662–2665
4 Lide, D. R. (ed.) (2004–2005) Concentrative properties of aqueous solutions. In CRC
Handbook of Chemistry and Physics (85th edn), pp. 8-72–8-73, CRC Press, Boca Raton
5 Eigen, M. and Hames, G. G. (1963) Elementary steps in enzyme reactions. Adv. Enzymol.
Relat. Areas Mol. Biol. 25, 1–28
6 Brouwer, A. C. and Kirsch, J. F. (1982) Investigation of diffusion-limited rates of
chymotrypsin reactions by viscosity variation. Biochemistry 21, 1302–1307
7 Nicolet, Y., Lockridge, O., Masson, P., Fontecilla-Camps, J. C. and Nachon, F. (2003)
Crystal structure of human butyrylcholinesterase and of its complexes with substrate and
products. J. Biol. Chem. 278, 41141–41147
8 Ngamelue, M. N., Homma, K., Lockridge, O. and Asojo, O. A. (2007) Crystallization and
X-ray structure of full-length recombinant human butyrylcholinesterase. Acta Crystallogr.,
Sect. F: Struct. Biol. Crystal. Commun. 63, 723–727
9 Humphrey, W., Dalke, A. and Schulten, K. (1996) VMD: visual molecular dynamics.
J. Mol. Graphics 14, 33–38
10 Pedretti, A., Villa, L. and Vistoli, G. (2004) VEGA: an open platform to develop
chemo-bio-informatics applications, using plug-in architecture and script programming.
J. Comput.-Aided Mol. Des. 18, 167–173
Received 15 March 2013/10 June 2013; accepted 20 June 2013
Published as BJ Immediate Publication 20 June 2013, doi:10.1042/BJ20130389
c The Authors Journal compilation c 2013 Biochemical Society
11 Phillips, J. C., Braun, R., Wang, W., Gumbart, J., Tajkhorshid, E., Villa, E., Chipot, C.,
Skeel, R. D., Kalé, L. and Schulten, K. (2005) Scalable molecular dynamics with NAMD.
J. Comput. Chem. 26, 1781–1802
12 Sadovnichy, V., Tikhonravov, A., Voevodin, V. I. and Opanasenko, V. (2013) “Lomonosov”:
supercomputing at Moscow State University. Contemporary High Performance
Computing: from Petascale toward Exascale (Chapman & Hall/CRC Computational
Science), pp. 283–307, CRC Press, Boca Raton
13 Best, R. B., Zhu, X., Shim, J., Lopes, P. E. M., Mittal, J., Feig, M. and MacKerell, A. D.
(2012) Optimization of the additive CHARMM all-atom protein force field targeting
improved sampling of the backbone φ, ψ and side-chain χ1 and χ2 dihedral angles.
J. Chem. Theory Comput. 8, 3257–3273
14 Raman, E. P., Guvench, O. and MacKerell, A. D. (2010) CHARMM additive all-atom force
field for glycosidic linkages in carbohydrates involving furanoses. J. Phys. Chem. B 114,
12981–12994
15 Hatcher, E. R., Guvench, O. and MacKerell, A. D. (2009) CHARMM additive all-atom force
field for acyclic polyalcohols, acyclic carbohydrates, and inositol. J. Chem. Theory
Comput. 5, 1315–1327
16 Till, M. S. and Ullmann, G. M. (2010) McVol: a program for calculating protein volumes
and identifying cavities by a Monte Carlo algorithm. J. Mol. Model. 16, 419–429
17 Masson, P., Froment, M.-T., Bartels, C. F. and Lockridge, O. (1996) Asp70 in the
peripheral anionic site of human butyrylcholinesterase. Eur. J. Biochem. 235, 36–48
18 Masson, P., Clery, C., Guerra, P., Redslob, A., Albaret, C. and Fortier, P. L. (1999)
Hydration change during the aging of phosphorylated human butyrylcholinesterase:
importance of residues aspartate-70 and glutamate-197 in the water network as probed by
hydrostatic and osmotic pressures. Biochem. J. 343, 361–369
19 Pezzementi, L., Nachon, F. and Chatonnet, A. (2011) Evolution of acetylcholinesterase and
butyrylcholinesterase in the vertebrates: an atypical butyrylcholinesterase from the
medaka Oryzias latipes . PLoS ONE 6, e17396
20 Saxena, A., Redman, A. M. G., Jiang, X., Lockridge, O. and Doctor, B. P. (1997)
Differences in active site gorge dimensions of cholinesterases revealed by binding of
inhibitors to human butyrylcholinesterase. Biochemistry 36, 14642–14651
21 Kleywegt, G. J. and Jones, T. A. (1994) Detection, delineation, measurement and display
of cavities in macromolecular structures. Acta Crystallogr., Sect. D: Biol. Crystallogr. 50,
178–185
22 Richards, F. M. (1985) Calculation of molecular volumes and areas for structures of
known geometry. Methods Enzymol. 15, 440–464
23 Connolly, M. (1983) Analytical molecular surface calculation. J. Appl. Crystallogr. 16,
548–558