An assessment of desolvation on rates of acetyl transfer: Insights into enzyme catalysis

Article abstract of DOI:10.1021/ja993341p

Enzymes greatly enhance the rate of reactions by a variety of physical organic mechanisms. One of the more contentious of these has been desolvation. To get a quantitative assessment of this contribution, we examined acetyl transfer to oxydianions. This is a model reaction for enzymes in which a high-energy acyl phosphate is formed. The aqueous reaction of phosphate or phosphonates with p-nitrophenyl acetate (pNPA) shows a substantial negative deviation from the Bronsted correlation obtained with monoanionic nucleophiles and from the reactivity of other, larger, oxydianions (molybdate, arsenate and vanadate). This, and other data, suggests a significant contribution of desolvation to the activation energy. To further investigate this we studied the effect of various DMSO (dimethyl sulfoxide)/H₂O mixtures on the reaction of chloromethylphosphonate, and of molybdate, on the reaction with pNPA. Increasing the DMSO concentration from 1% to 90% (v/v) increases the second-order rate constant for each of these reactions by over a factor of 5000. This is over a 1000 times greater than the enhancing effect on the reaction of phenoxides and over 10⁵ times the (inhibiting) effect on the reaction with neutral nucleophiles (imidazole). Extrapolation to pure DMSO yields a rate enhancement of ?10⁵, relative to the reaction in water, for the oxydianions. This enhancement is over 3 orders of magnitude greater than that seen with monoanionic phenoxide nucleophiles. This suggests a significant role for desolvation in the reactions in the enzyme-catalyzed nucleophilic reactions of inorganic phosphate but a modest role in reactions with less highly charged nucleophiles.

Full text of DOI:10.1021/ja993341p

J. Am. Chem. Soc. 2000, 122, 8365-8369
8365
An Assessment of Desolvation on Rates of Acetyl Transfer: Insights
into Enzyme Catalysis
Victor C. Rucker† and Larry D. Byers*
Contribution from the Department of Chemistry, Tulane UniVersity, New Orleans, Louisiana 70118
ReceiVed September 14, 1999
Abstract: Enzymes greatly enhance the rate of reactions by a variety of physical organic mechanisms. One of
the more contentious of these has been desolvation. To get a quantitative assessment of this contribution, we
examined acetyl transfer to oxydianions. This is a model reaction for enzymes in which a high-energy acyl
phosphate is formed. The aqueous reaction of phosphate or phosphonates with p-nitrophenyl acetate (pNPA)
shows a substantial negative deviation from the Brønsted correlation obtained with monoanionic nucleophiles
and from the reactivity of other, larger, oxydianions (molybdate, arsenate and vanadate). This, and other data,
suggests a significant contribution of desolvation to the activation energy. To further investigate this we studied
the effect of various DMSO (dimethyl sulfoxide)/H2O mixtures on the reaction of chloromethylphosphonate,
and of molybdate, on the reaction with pNPA. Increasing the DMSO concentration from 1% to 90% (v/v)
increases the second-order rate constant for each of these reactions by over a factor of 5000. This is over a
1000 times greater than the enhancing effect on the reaction of phenoxides and over 105 times the (inhibiting)
effect on the reaction with neutral nucleophiles (imidazole). Extrapolation to pure DMSO yields a rate
enhancement of ∼105, relative to the reaction in water, for the oxydianions. This enhancement is over 3 orders
of magnitude greater than that seen with monoanionic phenoxide nucleophiles. This suggests a significant role
for desolvation in the reactions in the enzyme-catalyzed nucleophilic reactions of inorganic phosphate but a
modest role in reactions with less highly charged nucleophiles.
Enzymes are responsible for enormous rate enhancements.
Depending on how this catalytic power is calculated,1,2 rate
accelerations of 108-1015 are typical [kcat/knon values, where
knon is the rate constant for the, often unobserved, nonenzymatic
reaction in aqueous solution corresponding mechanistically to
the enzyme-catalyzed reaction; the current record for this ratio
is 1017 for orotidine 5′-phosphate decarboxylase, or a value of
1023 M-1 for (kcat/Km)/knon, the “catalytic proficiency” of the
enzyme2]. There is no question that enzymes achieve this
impressive catalytic efficiency through a variety of physical
organic mechanisms (e.g., nucleophilic and/or general acid-
base catalysis, substrate binding in proper proximity and
orientation, electrostatic catalysis, solvation, and desolvation).
However, just how much each of these contribute (and indeed,
how separate they are) has remained a contentious issue.3,4 This
study was undertaken to quantitatively evaluate the contribution
of desolvation to enzymic rate enhancements, a topic which has
recently received much renewed interest (e.g.,5,6). The model
reaction we chose was acetyl transfer from p-nitrophenol to
oxydianions. Displacement reactions involving oxydianions,
particularly phosphate, are of great practical importance in
biochemistry. There are over 130 enzymatic reactions in which
the phosphate dianion acts as a nucleophile. Particularly
important among these are acyl transfer reactions in which a
high-energy acyl phosphate intermediate is formed. It has long
been known that in aqueous solutions the phosphate dianion is
2 to 3 orders of magnitude less reactive than monoanionic or
neutral nucleophiles of comparable basicity.7 Desolvation of a
strongly hydrated species, such as the phosphate dianion, is
expected to make a significant contribution to rate constant (kcat)
for a phosphate nucleophile bound at the active site of an
enzyme. Indeed, it has been shown that even for a monoanion
such as Cl- the reactivity (in the SN2 reaction with CH3Br) is
strongly dependent on the nature of the solvent. As the solvation
of the nucleophile decreases (e.g., from a decrease in the solvent
polarity or hydrogen bonding ability) the second-order rate
constant is enhanced, relative to that in water, reaching a factor
of 1016 when the gas-phase reaction is compared to the aqueous
reaction.8 To get an assessment of how much the microsolvent
of an enzyme’s active site can contribute to the enzymatic rate
enhancement (when compared to the reaction in water) and to
compare this solvent effect on oxydianions with that on
monoanionic and neutral nucleophiles, we investigated the
reaction of some phosphate analogues with p-nitrophenyl acetate
in water and in water/DMSO mixtures. This solvent mixture
was chosen for a variety of reasons. DMSO is a nonprotic
solvent and does not effectively solvate anions. Furthermore,
the dielectric constant of the medium is reduced by only 30%
as the mole fraction of DMSO is increased from 0 to 0.7.9 This
later property means that the electrostatic effect of solvent
* Address correspondence to this author.
† Present Address: Department of Chemistry, California Institute of
Technology, Pasadena, CA 91125.
(1) Northrop, D. B. AdV. Enzym. Rel. Areas Mol. Biol. 1999, 73, 25-
55.
(2) Radzicka, A.; Wolfenden, R. Science 1995, 267, 90-93.
(3) Cannon, W. R.; Singleton, S. F.; Benkovic, S. J. Nat. Struct. Biol.
1996, 3, 821-833.
(4) Lightstone, F. C.; Bruice, T. C. J. Am. Chem. Soc. 1997, 119, 9103-
9113.
(5) Cannon, W. R.; Benkovic, S. J. J. Biol. Chem. 1998, 273, 26257-
26260.
(7) Johnson, S. L. AdV. Phys. Org. Chem. 1967, 5, 237-330.
(8) Tanaka, K.; Mackay, G. I.; Payzant, J. D.; Bohme, D. K. Can. J.
Chem. 1976, 54, 1643-1659.
(6) Dewar, M. J. S.; Storch, D. M. Proc. Natl. Acad. Sci. U.S.A. 1985,
82, 2225-2229.
10.1021/ja993341p CCC: $19.00 © 2000 American Chemical Society
Published on Web 08/18/2000

8366 J. Am. Chem. Soc., Vol. 122, No. 35, 2000
Rucker and Byers
variation will be minimized so that the main effect on the rate
constant should be that of solvent reorganization as the reaction
progresses from the ground state to the transition state.
To more thoroughly examine, and quantitate, this solvation
effect on the reactivity of phosphonates as acyl acceptors, we
investigated the kinetic effect of DMSO/H2O mixtures on the
reaction.
Unusual Characteristics of the Reaction of Phosphate and
Phosphonates with p-Nitrophenyl Acetate
Experimental Section
The low intrinsic reactivity of esters such as pNPA (p-
nitrophenyl acetate) with the phosphate dianion7 is also seen in
the reaction with dianionic phosphonates.10 This could reflect,
at least in part, the requirement for removal of a strongly held
solvent molecule from a nucleophilic site on the dianion.
Consistent with this is the observation that, in the Brønsted
relationship between the reaction of monoanionic and neutral
nucleophiles with pNPA in water, the phosphate dianion shows
a negative deviation (by a factor ∼102) from this correlation.11
When the reaction, albeit with a different set of nucleophiles,
is carried out in 28.5% ethanol/water the negative deviation of
the phosphate dianion from the correlation line is not as
significant.12 In this later study, Bruice and Lapinski found that
phenoxide reacts with pNPA about 103 times faster than does
the phosphate dianion while the data for the reactions in water
indicate that this ratio is about 104.
Materials. p-Nitrophenyl acetate was obtained from Sigma-Aldrich
Chemical Co. The buffers (HEPES (N-(2-hydroxyethyl)piperazine-N′-
2-ethanesulfonic acid),1 pH 7-8; BICINE (N,N-bis(2-hydroxyethyl)-
glycine), pH 8-9, and CAPS (cyclohexylaminopropanesulfonate), pH
9-10) were obtained from Sigma-Aldrich Chemical Co. as the free
acid. Solutions were adjusted to the desired pH with sodium hydroxide.
The phosphonates were either prepared or obtained, as described
earlier.10 Sodium molybdate (dihydrate) was obtained from Matheson,
Coleman and Bell and stock solutions (0.5 M) were prepared by
dissolving Na2MoO4 in buffer solutions (0.01 M HEPES, 0.1 M NaCl)
and adjusting the pH to 7.5 with HCl. At this pH there is negligible
polymerization of molybdate.18 Stock vanadate solutions were prepared
essentially as described by Gresser et al.19 by dissolving vanadium oxide
(Sigma-Aldrich) in buffer solutions (0.01 M CAPS, 0.1 M NaCl)
containing a 4-fold excess of NaOH over V2O5 and carefully adjusting
the pH (8.8 to 9.5) with HCl. At the vanadate concentrations used in
-
this study (<5 mM), the predominant species are H2VO4 and
2- 19
HVO4
.
The acyl transfers to phosphonates or phosphate have two
other notable features which distinguish them from reactions
involving acyl transfer to neutral or monoanionic nucleophiles:
1. Low ânuc: In 1960 Jencks and Carriuolo,11 extending
Bruice and Lapinski’s work,12 showed that the reactivity of a
wide variety of nucleophiles toward pNPA can be correlated
with their pKa values. Of these 50 nucleophiles, most were either
oxymonoanions or neutral amines and the general trend in
reactivity corresponds to a ânuc ≈ 0.8 (25 °C). Using a more
homogeneous series of nucleophiles (phenoxides) Williams13
obtained a ânuc ) 0.75((0.04) for the reaction with pNPA at
25 °C. This can be compared with a value of ânuc ) 0.3 for the
reaction of a series of phosphonates with either pNPA or its
thiol ester analogue at 37 °C.10 The low ânucvalue is also
consistent with a significant role of desolvation in the reaction.
The more basic the phosphonate the more tightly it is expected
to bind water.14 This will partially compensate for the greater
nucleophilic reactivity of the more basic phosphonates.
2. Unusually favorable entropy of activation: The reaction
of methylphosphonate with pNPA in aqueous solutions shows
a value of ∆Sq ) -10 eu.15 This is considerably less negative
than the value usually seen for bimolecular reactions in aqueous
solution. For example, in the acetyl transfer from p-nitrophenol
to acetate [∆Sq ) -29 eu16] or to imidazole [∆Sq ) -30 eu17]
the activation entropies are about 3 times more negative than
those for the reaction with the dianionic phosphonate. This is
also consistent with a significant contribution of removal of
highly ordered water molecules from the phosphonate as the
reactants progress toward the transition state.
Methods. Aqueous pH measurements were made with a Metrohm
(Brinkmann) combititrator. pH and pKa measurements in H2O/DMSO
mixtures were made by using the overlapping indicator method using
the phenols and data of Buncel et al.20 The phosphonate or molybdate
solutions in H2O/DMSO were prepared by first dissolving the disodium
salt in a buffered aqueous solution (10 mM HEPES, pH 7.5 with
molybdate and 10 mM CAPS, pH 9.5 with chloromethyl-phosphonate).
A stock solution with sodium sulfate was prepared in the buffer at the
same concentration as the phosphonate or molybdate. This was then
mixed with DMSO and used as the blank when measuring the kinetics
for the phosphonate and molybdate reactions. This blank solution was
also used to maintain a constant ionic strength when the kinetics were
followed by varying the nucleophile concentration. The reaction was
followed by measuring the absorbance (λ ) 400 nm for p-nitrophe-
noxide) on a Hewlett-Packard model 8452A diode array spectropho-
tometer. In water, the reactions were initiated by adding 10 µL of the
ester (10 mM in DMSO) to 1 mL of a thermally equilibrated (25 °C)
buffer solution (ionic strength maintained at 1.6 M with Na2SO4)
containing a large excess of the nucleophile. In general the increase in
the absorbance was typically followed for at least 4 half-lives and the
data analyzed by nonlinear regression to a first-order curve.21 The
internal standard deviations of the pseudo-first-order rate constants were
consistently less than 1%. For the reactions in H2O/DMSO mixtures
the solubility of the nucleophile decreases as the DMSO concentration
increases. At the highest DMSO concentrations the concentrations of
the nucleophile (∼2 mM), while still high enough to carry out the
reactions under pseudo-first-order conditions, was (in the case of the
phosphonate) sufficiently low so that the reactions were followed by
the initial rate method (i.e., pseudo-zero-order conditions). At DMSO
concentrations g60% v/v corrections in the second-order rate constants
were made for the fraction of the nucleophile present as the dianion.
All rate constants were determined at least in triplicate for each reaction.
(9) Covington, A. K.; Dickinson, T. Physical Chemistry of Organic
SolVent Systems; Plenum: London, 1973; p 17.
Results
(10) Shames, S. L.; Byers, L.D. J. Am. Chem. Soc. 1981, 103, 6170-
6177.
Brønsted Correlation. To compare the ânuc for the reaction
of pNPA with phosphonates with the corresponding values for
the reactions with other nucleophiles we have repeated, at 25
°C, the determination made at 37 °C10 with a series of
(11) Jencks, W. P.; Carriuolo, J. J. Am. Chem. Soc. 1960, 82, 1778-
1786.
(12) Bruice, T. C.; Lapinski, R. J. Am. Chem. Soc. 1958, 80, 2265-
2267.
(13) Ba-Saif, S.; Luthra, A. K.; Williams, A. J. Am. Chem. Soc. 1989,
111, 2647-2652.
(18) Aveston, J.; Anacker, E. W.; Johnson, J. S. Inorg. Chem. 1964, 3,
735-746.
(14) Blades, A. T.; Klassen, J. S.; Kebarle, P. J. Am. Chem. Soc. 1995,
117, 10563-10571.
(19) Gresser, M. J.; Tracey, A. S.; Parkinson, K. M. J. Am. Chem. Soc.
1986, 108, 6229-6234.
(15) Wikjord, B. R.; Byers, L. D. J. Org. Chem. 1992, 57, 6814-6817.
(16) Gaetjens, E.; Morawetz, H. J. Am. Chem. Soc. 1960, 82, 5328-
5335.
(20) Buncel, E.; Um, I. H.; Hoz, S. J. Am. Chem. Soc. 1989, 111, 971-
975.
(17) Akiyama, A.; Hara, Y.; Tanabe, M. J. Chem. Soc., Perkin Trans. 2
1978, 288-292.
(21) Leatherbarrow, R. J. 1987 Enzfitter: A Nonlinear Regression Data
Analysis Program for the IBM PC; Elsevier Science Publishers: New York.

DesolVation on Rates of Acetyl Transfer
J. Am. Chem. Soc., Vol. 122, No. 35, 2000 8367
Figure 1. Brønsted correlation of the rate constants for the reaction
of a series of phosphonates (R-PO32-) with p-nitrophenyl acetate in
water (1% v/v DMSO, µ ) 1.6 M, 25°) yielding a value of ânuc ) 0.3.
Also illustrated are the positive deviations from this correlation of some
phosphate analogues (O).
Figure 2. Dependence of the rate constants for the reaction of
ClCH2PO32- (9) and of MoO42- (b) with p-nitrophenyl acetate on the
log of the water concentration in H2O/DMSO solutions (25 °C). Also
illustrated are the data of Buncel et al.20 for the reaction of 4-cyanophe-
nolate with pNPA (O).
phosphonates (Figure 1). We find the following correlation
equation:
For chloromethylphosphonate (pKa2 in H2O ) 6.6)
log k2 (M-1 min-1) ) [0.34((0.02)]pKa2
+
log k2(M-1 min-1) ) -[3.0((0.2)]log[H2O] +
3.5((0.3) [r2 ) 0.984] (2b)
5.1((0.4) [n ) 8, r2 ) 0.988] (1)
For molybdate (pKa2 in H2O ) 3.8)
Thus, the ânuc is essentially the same at 25 and 37 °C. Also
indicated in Figure 1 are some additional phosphate analogues,
all showing positive deviations from the correlation line.
Particularly notable is molybdate which is 370 times more
reactive than a phosphonate of comparable pKa2. Molybdate has
been shown to be a nucleophilic catalyst for this reaction.22
It will be noted that the phosphate dianion shows no deviation
from the correlation line based on the reactivity of the
phosphonates in water. This is in contrast to the negative
deviation seen in the correlation based on monoanionic oxygen
nucleophiles.11
Effect of DMSO on Reactivity. The reactions of the
dianionic nucleophiles with pNPA could be followed conve-
niently in DMSO/H2O mixtures up to about 90% (v/v) (which
corresponds to 0.70 mol % DMSO or [H2O] ) 0.5.6 M). At
DMSO concentrations e60% (v/v) the reactions could be
followed over a wide enough range of nucleophile concentra-
tions to establish the kinetics to be first-order in acyl acceptor
(molybdate or chloromethylphosphonate). The dependence of
the second-order rate constant on the water concentration is
illustrated in Figure 2. At water concentrations less than 28 M
[≈50% (v/v)] the molybdate reaction is more sensitive to the
increasing DMSO concentration than is the phosphonate reaction
and both of these are more sensitive than are the reactions of
various phenoxides with pNPA.20 Furthermore, once the DMSO
concentration exceeds 50% (v/v) (log[H2O] < 1.44) there
appears to be a linear relationship between log k2 and the
decreasing log[H2O]. This dependence of k2 on the water
concentration (5.6 M e [H2O] e 27.9 M which corresponds to
a mole fraction of DMSO between 0.7 and 0.2) is given below
for each of the oxydianions. The data of Buncel et al.20 for
4-cyanophenolate are included for comparison:
log k2(M-1 min-1) ) -[4.6((0.4)]log[H2O] +
6.4((0.4) [r2 ) 0.982] (2c)
In contrast to the reaction with these anionic nucleophiles,
the reaction of imidazole with pNPA was found to be inhibited
by increasing DMSO concentrations. Thus, when the DMSO
concentration was increased from 1% to 90% (v/v) the second-
order rate constant for the reaction decreased 20-fold. This is
in contrast to the reaction of pNPA with 4-cyanophenylate,
which shows a 15-fold increase in the rate constant over this
DMSO concentration range,20 and 5200- and 5500-fold increases
2-
with ClCH2PO3 and MoO42-, respectively.
Effect of DMSO on Acidity. There is a linear increase in
the pKa2 of chloromethylphosphonate and of molybdate with
increasing mole fraction of DMSO (0e ø e0.7). As has been
observed for the series of phenols studied by Buncel et al.,20
the sensitivity of the pKa to increasing DMSO concentration is
proportional to the pKa of the acid in water. Thus, for molybdate
pK*a2 ) [3.4((0.1)]ø + 3.8((0.1) [r2 ) 0.995] (3a)
and for chloromethylphosphonate
pK*a2 ) [4.9((0.1)]ø + 6.5((0.1) [r2 ) 0.990] (3b)
Using these data (for 0.2 e ø e 0.7) and the method of
Buncel et al.,20 the following “Brønsted” correlations can be
obtained:
For molybdate
log k2(M-1 min-1) ) [1.8((0.1)]pK*a2 - 8.6((0.8)
[r2 ) 0.980] (4a)
For 4-cyanophenolate (pKa in H2O ) 7.95)
log k2(M-1 min-1) ) -[1.91((0.08)]log[H2O] +
2.85((0.09) [r2 ) 0.997] (2a)
and for chloromethylphosphonate
log k2(M-1 min-1) ) [0.83((0.06)]pK*a2 - 7.2((0.5)
[r2 ) 0.985] (4b)
(22) Wikjord, B.; Byers, L. D. J. Am. Chem. Soc. 1992, 114, 5553-
5554.

8368 J. Am. Chem. Soc., Vol. 122, No. 35, 2000
Rucker and Byers
Discussion
transfer to molybdate is accompanied by a somewhat greater
solvent reorganization than that of the phosphonate. In pure
water, the greater reactivity of molybdate than of the phospho-
nate must, therefore, reflect not so much the number of water
molecules in transition but rather the strength of the interaction
between the water molecules and the nucleophilic site from
which they must be removed. In this regard it is interesting to
note that the positive deviations of the dianions from the
Brønsted correlation of the phosphonates (Figure 1) is inversely
related to the calculated charge on the oxygen of the dianionic
species. Using Gaussian 94, Kish and Viola23 calculated the
following charge densities: MoO42-, -0.800; HVO42-, -0.885;
HAsO42-, -0.892; and HPO42-, -0.952. Thus, it is reasonable
to assume that it should be more difficult to remove a water
molecule from a nucleophilic site on phosphate than from the
corresponding site on these other oxydianions.
The aqueous reaction of the phosphonates (and presumably
the other oxydianions) with pNPA is, most-likely, a concerted
reaction. This has been shown to be the case for the reaction of
phenoxide nucleophiles with pNPA.13,23 The evidence for a
similar mechanism for acetyl transfers to phosphonates is as
follows: (1) The Brønsted plot (Figure 1) is linear for the
reaction of pNPA with a series of phosphonates, showing no
break at pKa ≈ 7 ()pKa of the p-nitrophenol leaving group).
(2) The phosphate dianion does not deviate from this correla-
tion. If a tetrahedral intermediate, Td (e.g., with a lifetime
>∼10-12 s) formed, it should tautomerize:
It has been implicitly argued that the change in reactivity of
acyl acceptors in media of various DMSO/water compositions
is due just to the effect of solvent on the pKa of the nucleophile.
This allowed Buncel et al.20 to propose a novel strategy for the
construction of Brønsted plots. Thus, for the reaction of
4-cyanophenolate with pNPA in various DMSO/H2O solvents,
they obtained an excellent correlation between log k2 and the
pKa* in that particular solvent. Indeed, the ânuc value they
obtained in this way ()0.6) was the same as that obtained with
several other phenoxide acyl acceptors, although it was lower
than the ânuc ()0.75) which they obtained in water using a series
of substituted phenoxides. In our case, the “ânuc” ()0.8) value
for the phosphonate reaction obtained in this manner (eq 4b) is
considerably larger than the value obtained in water ()0.3) with
a series of substituted phosphonates (Figure 1). Clearly, factors
other than just changing pKa values influence the effect of
solvent variation on reaction rates. In this regard, it is interesting
to note that the pKa2 value of molybdate changes by 2.4 units
as the solvent is changed from 1% to 90% v/v DMSO. Thus,
this solvent change makes the equilibrium constant for mono-
protonation of the molybdate dianion more favorable by a factor
of 260. The same solvent change, however, makes the rate
constant of acetyl transfer to the molybdate dianion more
favorable by a factor of 5500. A qualitatively similar situation
is seen with chloromethylphosphonate where the change in
solvent has a greater effect on the acylation rate constant than
on the pKa2. This is in spite of the fact that protonation reduces
the charge on the dianion by one unit while, in the transition
state for acylation, the charge, although more dispersed, is
probably not much changed from -2. Clearly the differential
solvation (hydration) between the species in the ground state
and in the transition state is one of the many factors which
influence the sensitivity of a reaction to solvent variation. In
enzymatic reactions the reactivity of oxydianions such as
phosphate will be significantly increased if the ionic species is
bound poorly solvated in a hydrophobic region (where it is
protected from protonation) in the active site. We can get an
estimate of this rate increase from an extrapolation of our data.
A convenient way of doing this is illustrated in Figure 3. This
is the same data as in Figure 2 except that the change in solvent
composition is expressed as the mole fraction of DMSO ()ø).
The linear dependence (in the region where ø g 0.2) of log k2
on the mole fraction of DMSO is given below for each of the
oxydianions.
Based on an estimate of the pKa difference, this tautomerization
equilibrium constant is expected to be ∼104 in favor of Td′, a
species that is not on the reaction path. Thus, if Td did form as
a discrete intermediate, the phosphate dianion should show a
negative deviation (of ∼4 log units) from the correlation based
on the phosphonate dianions (for which such a tautomerization
is not possible). (3) Using Guthrie’s approach,23 the energy of
Td (relative to that of pNPA + RPO3-2) can be estimated to
be ∼34 kcal/mol. This ∆G°′ value is over 10 kcal/mol greater
than the observed ∆Gq values ()23 kcal/mol for R ) ClCH2-
and 22 kcal/mol for R ) HO-) at 25 °C. These results are
consistent with a concerted reaction or, at least, one that does
not involve a tetrahedral intermediate.
Solvent Effect. The increase in the rate constant, with
increasing DMSO concentration, for the reaction of anionic
nucleophiles with pNPA is not surprising. What is unusual,
however, is the magnitude of this rate increase with the dianionic
nucleophiles when compared to that seen with monoanionic
nucleophiles. It is reasonable to expect that the more strongly
solvated in water a nucleophile is, the more destabilized, and
thus more reactive, it will become when the solvent is changed
to one like DMSO which does not effectively solvate anions.
An admittedly simplistic estimation of the extent of this
nucleophile solvation difference can be made from the linear
portions in Figure 2. This is based on the following model:
n is the solvation number in the ground state (i.e., the number
of water molecules in the primary solvation shell of the ester
plus that of the nucleophile) and n′ is the solvation number in
the transition state. The slope ()m) of a plot of log k2 vs log-
[H2O] (Figure 2) provides an estimate of the difference between
these numbers (i.e., m ) n′ - n) and corresponds to the net
number of water molecules which undergo reorganization as
the transition state is approached. Thus, the reaction of dianions
(molybdate and chloromethylphosphonate) involves significantly
more solvent reorganization than does the reaction of the
phenoxide monoanions (eqs 2a-c). Furthermore, the acetyl
For 4-cyanophenolate
log k2(M-1 min-1) ) [2.7((0.1)]ø - 0.47((0.05)
[r2 ) 0.997] (5a)
(23) Guthrie, J. P. J. Am. Chem. Soc. 1991, 113, 3941-3949.
(24) Kish, M. M.; Viola, R. E. Inorg. Chem. 1999, 38, 818-820.

DesolVation on Rates of Acetyl Transfer
J. Am. Chem. Soc., Vol. 122, No. 35, 2000 8369
Table 1. Solvent Effect on Nucleophilicity
k2, M-1 min-1
ø ) 0a ø ) 1b
nucleophile
kø)1/kø)0
4-cyanophenolatec
chloromethylphosphonate
molybdate
1.8
1.7 × 102
2.8 × 102
6.5 × 104
94
2.5 × 10-3
1 × 105
1.3 × 10-1
5 × 105
a Measured rate constants in water (mole fraction of DMSO ) 0).
b Calculated rate constants [eq 5] in DMSO. c Data from Buncel et al.20
results are summarized in Table 1. It can be seen that for the
chloromethylphosphonate reaction, complete stripping of the
dianion of its solvating water molecules results in a rate
enhancement of about 105-fold. Thus, for the enzymatic reaction
of dianionic nucleophiles such as phosphate we might expect a
significant contribution of “desolvation” to the enzymic rate
enhancement. However, for monoanionic nucleophiles, such as
4-cyanophenolate, this contribution is quite modest (less than
100-fold) and for neutral nucleophiles the substitution of an
aqueous solvent to a DMSO-like reaction environment is
inhibitory.
Figure 3. Replot of the data from Figure 2 in terms of the mole fraction
of DMSO: (9) ClCH2PO32-, (b) MoO42-, and (O) 4-cyanophenolate.
For chloromethylphosphonate
log k2(M-1 min-1) ) [4.3((0.2)]ø - 1.85((0.11)
[r2 ) 0.992] (5b)
Supporting Information Available: Table of log k2 and
pKa2* values in solutions of various DMSO/H2O compositions
(PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
For molybdate
log k2(M-1 min-1) ) [6.4((0.3)]ø - 1.59((0.15)
[r2 ) 0.993] (5c)
This allows for easy extrapolation to pure DMSO (ø ) 1). These
JA993341P