Generality of solvation effects on the hydrolysis rates of phosphate monoesters and their possible relevance to enzymatic catalysis

Article abstract of DOI:10.1021/jo016104p

Previous work by Kirby and co-workers revealed a significant acceleration of the rate of hydrolysis of p-nitrophenyl phosphate by added dipolar solvents such as DMSO. Activation parameters and kinetic isotope effects have been measured to ascertain the origin of this effect. The generality of this phenomenon was examined with a series of esters with more basic leaving groups. Computational analyses of the effects of desolvation of dianionic phosphate monoesters were carried out, and the possible effect of the transfer from water to the active site of alkaline phosphatase was modeled. The results are consistent with a desolvation-induced weakening of the P-O ester bond in the ground state. Other aryl phosphate esters show similar rate accelerations at high fractions of DMSO, but phenyl and methyl phosphates do not, and their hydrolysis reactions are actually slowed by these conditions.

Full text of DOI:10.1021/jo016104p

1214
J . Org. Chem. 2002, 67, 1214-1220
Gen er a lity of Solva tion Effects on th e Hyd r olysis Ra tes of
P h osp h a te Mon oester s a n d Th eir P ossible Releva n ce to En zym a tic
Ca ta lysis
Piotr K. Grzyska,† Przemyslaw G. Czyryca,† J ustin Golightly, Kelly Small, Paul Larsen,
Richard H. Hoff,‡ and Alvan C. Hengge*
Utah State University, Department of Chemistry and Biochemistry, Logan, Utah 84322-0300
hengge@cc.usu.edu
Received September 10, 2001
Previous work by Kirby and co-workers revealed a significant acceleration of the rate of hydrolysis
of p-nitrophenyl phosphate by added dipolar solvents such as DMSO. Activation parameters and
kinetic isotope effects have been measured to ascertain the origin of this effect. The generality of
this phenomenon was examined with a series of esters with more basic leaving groups.
Computational analyses of the effects of desolvation of dianionic phosphate monoesters were carried
out, and the possible effect of the transfer from water to the active site of alkaline phosphatase
was modeled. The results are consistent with a desolvation-induced weakening of the P-O ester
bond in the ground state. Other aryl phosphate esters show similar rate accelerations at high
fractions of DMSO, but phenyl and methyl phosphates do not, and their hydrolysis reactions are
actually slowed by these conditions.
In tr od u ction
Phosphoryl transfer reactions are ubiquitous in bio-
logical systems where the formation and hydrolysis of
phosphate monoesters comprises an essential regulatory
mechanism for a host of cellular processes. For most
phosphatases, the dianion of the phosphate monoester
is the substrate for catalysis. Most of what is known
about the chemistry of phosphate monoesters has come
from the study of aryl phosphates. The uncatalyzed
hydrolysis reactions of the dianions of aryl phosphate
monoesters proceed through a concerted mechanism with
a very loose transition state (Figure 1) in which the
phosphoryl group resembles metaphosphate ion, bond
cleavage to the leaving group is extensive, and bond
formation to the nucleophile is minimal. The extensive
evidence for this mechanism has been reviewed.1,2 While
bond formation to the nucleophile is not far advanced in
the transition state, data from linear free-energy rela-
tionships indicate that metaphosphate is not an inter-
mediate,3 and results with phosphates made chiral by
the use of oxygen isotopes show that the bond is suf-
ficiently far developed to effect inversion except when the
phosphoryl acceptor is the sterically hindered tert-butyl
alcohol.4,5
F igu r e 1. Loose transition state for hydrolysis reactions of
aryl phosphate ester dianions.
dianion resulted in up to a 106-fold rate acceleration when
the level of DMSO reached 95%.6 It was speculated that
the faster rate arose from disruption of hydrogen bonding
to the phosphoryl group, resulting in a weakening of the
P-O ester bond, and that a similar effect might arise
from binding to a phosphatase active site and contribute
to enzymatic catalysis. In the uncatalyzed reaction, the
negative charges on the phosphoryl group assist in
expulsion of the leaving group. Thus, it is a logical
hypothesis that desolvation of the phosphoryl group
would destabilize the reactant by making these charges
more available to assist in expulsion of the leaving group.
However, there are alternative explanations for the rate
acceleration imparted by DMSO. Reactions involving
nucleophilic attack are often faster in aprotic solvents
because the nucleophile is less solvated. Thus, under
these conditions the mechanism may change to one with
significant nucleophilic involvement by either water or
hydroxide.
In 1986 Kirby and Abell reported that added DMSO
on the hydrolysis of p-nitrophenyl phosphate (pNPP)
* Corresponding author. Fax: 435-797-3390.
† These authors contributed equally to this work.
‡ Present address: Associate Professor, Department of Chemistry,
U.S. Military Academy, West Point.
We have carried out experiments to better understand
the nature of the rate acceleration imparted by added
DMSO to aqueous solutions of pNPP. To ascertain
whether rate accelerations occur for phosphate mo-
noesters having less acidic leaving groups, the effect of
DMSO on the hydrolysis rates of several aryl phosphates,
phenyl phosphate, and methyl phosphate have been
examined.
(1) Hengge, A. C. In Comprehensive Biological Catalysis: A Mecha-
nistic Reference; Sinnott, M., Ed.; Academic Press: San Diego, CA,
1998; Vol. 1, pp 517-542.
(2) Thatcher, G. R. J .; Kluger, R. Adv. Phys. Org. Chem. 1989, 25,
99-265.
(3) Herschlag, D.; J encks, W. P. J . Am. Chem. Soc. 1989, 111, 7579-
7586.
(4) Friedman, J . M.; Freeman, S.; Knowles, J . R. J . Am. Chem. Soc.
1988, 110, 1268-1275.
(5) Buchwald, S. L.; Friedman, J . M.; Knowles, J . R. J . Am. Chem.
Soc. 1984, 106, 4911-4916.
(6) Abell, K. W. Y.; Kirby, A. J . Tetrahedron. Lett. 1986, 27, 1085-
1088.
10.1021/jo016104p CCC: $22.00 © 2002 American Chemical Society
Published on Web 01/17/2002

Solvation Effects on Phosphate Ester Hydrolysis
J . Org. Chem., Vol. 67, No. 4, 2002 1215
Sch em e 1. Dia gr a m Sh ow in g th e P osition a n d
Nom en cla tu r e of th e Isotop e Effects Mea su r ed ,
a n d th e Isotop e Effects for Hyd r olysis in Aqu eou s
Solu tion a n d in 90% DMSOa
F igu r e 2. Isotopic isomers used for measurement of the
bridge (A and C) and nonbridge (B and C) 18O isotope effects.
Exp er im en ta l Section
The disodium salts of p-nitrophenyl phosphate and of phenyl
phosphate were commercial products and were recrystallized
before use. The bis(cyclohexylammonium) salts of p-chlorophen-
yl phosphate, m-bromophenyl phosphate, and p-cyanophenyl
phosphate were prepared by the method of Bourne and
Williams.7 These were converted to their sodium salts by a
cation exchange column using SP-C25 resin equilibrated with
sodium acetate. The esters were pure by proton and 31P NMR
(see Supporting Information).
Methyl phosphate was synthesized by a modification of the
method described by Kirby.8 In a 50 mL round-bottom flask
with condenser placed in an ice water bath, 0.010 mol of dry
phosphorous acid was dissolved in 20 mL of methanol to which
solution 0.035 mol of triethylamine was added. Then, 0.015
mol of solid iodine was added very slowly in such a manner
that rapid evolution of heat was avoided. The cold bath was
removed, and the reaction was allowed to warm to room
temperature. After approximately 15 min, the reaction mixture
was poured into about 200 mL of acetone with 5 g of ground
NaOH as a suspension. A cloudy suspension of the disodium
salt of methyl phosphate was filtered off. The product was
recrystallized from methanol to yield methyl phosphate in 80%
yield as the disodium salt. The product was pure by 1H and
31P NMR.
The isotopic isomers of p-nitrophenyl phosphate used for
measurement of kinetic isotope effects are shown in Figure 2.
Natural abundance p-nitrophenyl phosphate, [14N]-p-nitro-
phenyl phosphate, [15N, nonbridge-18O3]-p-nitrophenyl phos-
phate, [14N]-p-nitrophenol, and [15N, 18O]-p-nitrophenol were
synthesized as described previously.9 [14N]-p-nitrophenol and
[15N, 18O]-p-nitrophenol were then mixed to reconstitute the
natural abundance of 15N, and then the mixture was phos-
phorylated to produce p-nitrophenyl phosphate using the
method cited above. This mixture of isotopic isomers (A and
C, Figure 2) was used for determination of 18kbridge. Isomers B
a
Standard errors in the last decimal place are in parentheses.
of ice-cold water. An aliquot was assayed for p-nitrophenol in
basic solution at 400 nm; a second aliquot was added to TRIS
buffer at pH 9 and treated with alkaline phosphatase to
completely hydrolyze the remaining reactant and then assayed
similarly. The fraction of reaction was determined from these
two measurements. The remainder of the reaction mixture was
titrated to pH 5 and extracted with diethyl ether three times
to extract the product p-nitrophenol. These ether layers were
dried over magnesium sulfate, and DMSO was removed by
passing the ether solution through a column of silica gel (320-
400 mesh). The resulting ether solution was evaporated to
dryness, and the p-nitrophenol was further purified by subli-
mation before isotopic analysis by isotope ratio mass spec-
trometry using an ANCA-NT combustion system in tandem
with a Europa 20-20 isotope ratio mass spectrometer.
The aqueous layer, containing the unreacted pNPP, was
made alkaline and, water was removed at 35-40 °C by rotary
evaporation. By the time most of the water had been removed
and the reactant mixture had once again achieved a high
proportion of DMSO, complete hydrolysis of the residual pNPP
was rapid. After >10 half-lives, this mixture was added to 200
mL of water, titrated to pH 5, and treated as described above
to recover p-nitrophenol.
and C were also mixed to reconstitute the natural abundance
18
of 15N and used for determination of
k
. The isotopic
nonbridge
abundance of the mixtures was determined by isotope ratio
mass spectrometry. Scheme 1 summarizes the positions in the
pNPP reactant and the nomenclature for the isotope effects
measured.
Isotope effects were calculated from the isotopic ratios at
partial reaction in the p-nitrophenol product (Rp), in the
residual substrate (Rs), and in the starting material (Ro).
Equations 1 and 2 were used to calculate the observed isotope
effect either from Rp and Ro or from Rs and Ro, respectively,
at the fraction of reaction f.11 Thus, each experiment yields
two independent determinations of the isotope effect.
Isotop e Effect Deter m in a tion s. Isotope effects were
measured by isotope ratio mass spectrometry. The N15 isotope
effects were measured using the natural abundance compound.
The O18 isotope effects were measured by the remote label
method10 using the nitrogen atom as a reporter for isotopic
fractionation at the bridge or nonbridge oxygen atoms.9
Isotope effects were measured at 85 °C in 90% DMSO as
follows. The disodium salt (30 mg) of the appropriate isotopi-
cally labeled form of the substrate was dissolved in 500 µL of
1 N NaOH at room temperature. This was added to 19.5 mL
of 92.3% DMSO/H2O at 85 °C. After partial hydrolysis, the
reactions were stopped by pouring the mixture into 200 mL
isotope effect ) log(1 - f )/log(1 - f(Rp/Ro))
isotope effect ) log(1 - f)/log[(1 - f )(Rs/Ro)]
(1)
(2)
Ro was determined by two methods: from unreacted pNPS
by isotope ratio mass spectrometry, and, as a control, from
p-nitrophenol isolated after complete hydrolysis of a sample
(7) Bourne, N.; Williams, A. J . Org. Chem. 1984, 49, 1200-1204.
(8) Kirby, A. J . Chem. Ind. 1963, 1877-1878.
(9) Hengge, A. C.; Edens, W. A.; Elsing, H. J . Am. Chem. Soc. 1994,
116, 5045-5049.
(10) O’Leary, M. H.; Marlier, J . F. J . Am. Chem. Soc. 1979, 101,
3300-3306.
(11) Bigeleisen, J .; Wolfsberg, M. Adv. Chem. Phys. 1958, 1, 15-
76.

1216 J . Org. Chem., Vol. 67, No. 4, 2002
Grzyska et al.
Ta ble 1. P h osp h or u s-Oxygen Ester Bon d Len gth s (Å)
for Meth yl, P h en yl, a n d p-Nitr op h en yl P h osp h a te
Dia n ion s fr om Ca lcu la tion s a t th e HF /6-31++G** Level,
Usin g th e P CM Meth od a s Im p lem en ted in Ga u ssia n 98.a
of pNPS using the same isolation and purification procedures
used in the isotope effect experiments. The agreement of these
two numbers demonstrated that, within the experimental
error, no isotopic fractionation occurs as a result of the
procedures used to isolate and purify the p-nitrophenol.
Kin etics of Ar yl P h osp h a tes a n d Meth yl P h osp h a te.
The hydrolysis of the dianion of p-nitrophenyl phosphate in
95% DMSO in the presence of 20 mM tetrabutylammonium
hydroxide, with ionic strength maintained using tetrabutyl-
ammonium bromide, was examined as described by Kirby et
al.6 The reaction rate was monitored to completion by following
the release of p-nitrophenolate at 400 nm at 283, 297, 303,
and 312 K.
Leaving Group/pKa
-OMe/
15.5
-OPh/
9.95
-OpNPh/
7.14
gas phase, HF/6-31++G**
1 molecule of H2O explicitly,
HF/6-31++G**
DMSO, HF/6-31++G**, PCM
water, HF/6-31++G**, PCM
ONIOM HF/6-31++G**:PM3
h igh level: methyl
1.730
1.709
1.825
1.785
1.989
1.866
1.680
1.663
1.678
1.725
1.699
1.866
1.729
For determination of the effect of the leaving group on the
rate, the hydrolysis rates of p-chlorophenyl phosphate, m-
bromophenyl phosphate, and p-cyanophenyl phosphate in 95%
DMSO were measured using the initial rate method. The
phosphate ester concentration was 2.5 mM for p-chlorophenyl
phosphate and for m-bromophenyl phosphate and 1.25 mM
for p-cyanophenyl phosphate. Reaction mixtures were 20 mM
in [-OH]. The dependence of the rates on hydroxide concentra-
tion rates were examined for [-OH] between 20 and 40 mM,
keeping ionic strength constant using tetrabutylammonium
bromide. The rates were followed by periodically adding
aliquots of the reaction mixture to 0.1 N NaOH and measuring
the absorbance of the liberated phenolate ion at 310, 295, and
295 nm for p-chlorophenyl phosphate, m-bromophenyl phos-
phate, and p-cyanophenyl phosphate, respectively. Endpoints
for these reactions were obtained by adding an aliquot of the
reaction mixture to a solution of E. coli alkaline phosphatase
in 100 mM TRIS buffer at pH 9, 1 mM ZnCl2, and 1 mM MgCl2
and assaying for liberated phenol after 24 h.
Because of the very small change in UV absorption that
accompanies hydrolysis of phenyl phosphate and methyl
phosphate, 31P NMR was used to monitor the hydrolysis rates
of these two esters. The reactions were carried out in sealed
quartz tubes as described by Wolfenden.12 Tubes containing
from 200 to 450 µL of the reaction solution were placed in a
thermostated silicone oil bath. After appropriate times, samples
were removed and cooled, and the contents were diluted with
D2O. For the aqueous hydrolysis of phenyl and methyl
phosphates, reactions were carried out at pH 11 in 0.3 M
carbonate buffer with an initial reactant concentration of 50
mM. The experiments with phenyl and methyl phosphates in
a DMSO/water mixture required a lower fraction of DMSO
(80%) in order to solubilize a sufficient concentration (5 mM)
of the reactant to allow accurate monitoring by NMR. The
concentration of tetrabutylammonium hydroxide was varied
from 30 to 60 mM, with ionic strength maintained at 75 mM
using tetrabutylammonium bromide. After appropriate times,
samples were removed and cooled, and the contents were
diluted with D2O and analyzed by NMR.
phosphate; low level: seven
solvating molecules of water
same system as above,
fully in HF/6-31++G**
ONIOM HF/6-31++G**:PM3
h igh level: methyl
1.674
1.707
phosphate and two Zn2+
ions; low level: residues of
alkaline phosphatase active
site
a
z-Matrix for each final structure is available in Supporting
Information
Additional calculations were carried out with the methyl
phosphate dianion (the other esters were excluded for practical
considerations). These calculations involved an evaluation of
the applicability of the ONIOM1 method14 to study the influ-
ence of the environment on the P-O bond length using a test
structure of methyl phosphate solvated with seven explicitly
introduced molecules of water and a calculation of the geom-
etry of the alkaline phosphatase-substrate complex. In the
former, methyl phosphate solvated by seven water molecules
was examined fully at the HF/6-31++G** level, and then for
comparison, using the ONIOM method (the ONIOM method
allows the model to be split into a number of layers of different
levels of theory and thus allows for reduction of the compu-
tational time and memory usage) with the HF/6-31++G**
layer used for the phosphate and PM3 for the water molecules.
Both approaches yielded basically the same results (Table 1),
which validated the applicability of the ONIOM method. Then,
the ONIOM method was used for optimization of methyl
phosphate bound to the active site of alkaline phosphatase. A
fragment of the ALK115 structure, corresponding to the active
site, was used as an initial structure (Figure 3). Free-peptide
linkages were terminated as amides, and the inorganic
phosphate in ALK1 was substituted with methyl phosphate
in its dianionic form prior to optimization. To preserve the
proper geometry of the active site, a number of atoms of the
structure were set as nonoptimizable.
Com p u ta tion a l An a lyses. The effect of solvation on the
scissile P-O ester bond of p-nitrophenyl, phenyl, and methyl
phosphates was examined computationally using the Gaussian
98 revision 7 package on the HF/6-31++G** level.13 The
structures of these three phosphate esters were optimized in
the gas phase, in the gas phase with one molecule of water
explicitly modeled, and using the polarizable continuum
models (PCM) for DMSO and for water.
Resu lts
To compare these results with prior data reported with
pNPP in 95% DMSO, and to maximize the expected rate
acceleration imparted by the dipolar aprotic solvent,
experiments were conducted in 95% DMSO whenever
possible. Some experiments required the use of lower
DMSO fractions as discussed above. Table 2 shows the
percentages of DMSO that were used in particular
experiments.
Ar yl P h osp h a te Ester s. The hydrolysis rate of p-
nitrophenyl phosphate dianion in 95% DMSO and the
independence of the rate from the concentration of
hydroxide agree with results previously reported by Kirby
et al.6 The rate of hydrolysis was measured over the
temperature range from 10 to 39.5 °C, and an Eyring plot
(12) Wolfenden, R.; Ridgway, C.; Young, G. J . Am. Chem. Soc. 1998,
120, 833-834.
(13) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; Montgomery, J . A.,
J r.; Stratmann, R. E.; Burant, J . C.; Dapprich, S.; Millam, J . M.;
Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J .;
Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo,
C.; Clifford, S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J . B.; Cioslowski, J .; Ortiz, J . V.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.;
Fox, D. J .; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; J ohnson, B. G.; Chen,
W.; Wong, M. W.; Andres, J . L.; Head-Gordon, M.; Replogle, E. S.;
Pople, J . A. Gaussian 98, revision 7.0; Gaussian, Inc.: Pittsburgh, PA,
1998.
(14) Dapprich, S.; Komaromi, I.; Byun, K. S.; Morokuma, K.; Frisch,
M. J . J . Mol. Struct.: THEOCHEM 1999, 1-21.
(15) Kim, E. E.; Wyckoff, H. W. J . Mol. Biol. 1991, 218, 449-464.

Solvation Effects on Phosphate Ester Hydrolysis
J . Org. Chem., Vol. 67, No. 4, 2002 1217
F igu r e 3. Model of the active site of ALK1. Optimizable region of the active site is shown in green. Atoms belonging to a high,
ab initio level for the ONIOM method are marked with stars; all other atoms were modeled at the PM3 level.
Ta ble 2. Su m m a r y of P er cen ta ges of DMSO Used in
Exp er im en ts w ith P h osp h a te Mon oester Dia n ion s
Ta ble 3. Activa tion P a r a m eter s for Hyd r olysis of th e
p NP P Dia n ion in Wa ter a n d in 95% DMSO/Wa ter a
∆H‡ (kcal/mole)
∆S‡ (eu)
Rate measurements and measurements of
Brønsted âlg with p-nitrophenyl phosphate,
p-chlorophenyl phosphate, m-bromophenyl
phosphate, and p-cyanophenyl phosphate
Rate measurements with phenyl phosphate and
methyl phosphate
95% DMSO
H2O
95:5 DMSO/H2O
30.6
20.7 ( 0.6
+ 3.5
-5.7 ( 0.6
80% DMSO
a
Aqueous values are from ref 17.
Kinetic isotope effects with p-nitrophenyl phosphate 90% DMSO
F igu r e 5. Brønsted plot showing the log of the pseudo-first-
order rate constant (in sec-1) versus leaving group pKa for the
hydrolysis of p-chlorophenyl phosphate, m-bromophenyl phos-
phate, p-nitrophenyl phosphate, and p-cyanophenyl phosphate
in 95% DMSO.
F igu r e 4. Eyring plot for the hydrolysis of pNPP dianion in
95:5 DMSO/water.
(Figure 4) was used to obtain the activation parameters
∆H‡ ) 20.7 ( 0.6 kcal/mol and ∆S‡39 ) -5.7 ( 0.6 entropy
units (eu) (Table 3).
The rate constants for the hydrolysis in 95% DMSO of
p-nitrophenyl phosphate, of p-chlorophenyl phosphate, of
m-bromophenyl phosphate, and of p-cyanophenyl phos-
phate were used to construct the Brønsted plot in Figure
5. If a line is fitted to all four points, a Brønsted slope of
-1.4 ( 0.2 is obtained, with r2 ) 0.948. The phosphate
ester having the leaving group with the highest pKa,
p-chlorophenol, undergoes reaction somewhat slower
than predicted by a line passing through the points for
the first three compounds. A fit to the first three points
alone gives a Brønsted slope of -1.15 ( 0.01, with r2 )

1218 J . Org. Chem., Vol. 67, No. 4, 2002
Grzyska et al.
tylammonium hydroxide was left out of the reaction
mixture or was present in a low concentration relative
to the reactant. When [-OH] was 50 mM or higher, no
hydrolysis was observed after time periods sufficient for
the aqueous reaction at the same temperature to proceed
to 50% or more of completion.
Kin etic Isotop e Effects. The kinetic isotope effects
measured for the hydrolysis of pNPP in 90% DMSO at
85 °C are shown in Scheme 1. For comparison, the
previously reported isotope effects for the aqueous hy-
drolysis of the dianion at 95 °C are also shown. Because
temperature affects the magnitude of isotope effects, it
is desirable to compare data taken at similar tempera-
tures. At 85 °C, the half-life of the reaction is less than
1 min; it was not feasible to conduct the isotope effect
reactions at a higher temperature or higher DMSO
fraction because of the need to stop the reaction after
partial hydrolysis. The temperature-induced deviation
expected from a 10° temperature difference is very small.
Com p u ta tion a l Stu d ies. Table 1 shows the results
of the computational studies of p-nitrophenyl phosphate,
phenyl phosphate, and methyl phosphate. The results
indicate a dependence of the scissile P-O ester bond
length on both the leaving group and the solvation of the
phosphoryl group. This bond length becomes shorter (and
thus the substrate more stable) with increasing solvation
or with increasing basicity of the leaving group. The effect
of solvation is most prominent in the case of p-nitrophen-
yl phosphate.
Calculations using the ONIOM method were performed
on a partially rigid fragment of the enzyme (Figure 3)
with methyl phosphate manually docked into the active
site in the initial state before optimization, as described
in the Experimental Section. The results show a rather
surprising, measurable destabilization of the substrate
in the alkaline phosphatase active site in comparison to
that in solution (Table 1, Figure 7), suggesting a potential
susceptibility of the reactant to undergo hydrolysis by a
dissociative mechanism.
F igu r e 6. Eyring plot for hydrolysis of the phenyl phosphate
dianion in aqueous solution, pH 11.
0.999. For comparison, the value obtained for the aqueous
hydrolysis of a series of aryl phosphate dianions is
-1.23.16
P h en yl a n d Meth yl P h osp h a te Ester s. NMR was
used to follow the rates of the reactions of phenyl and
methyl phosphate in aqueous and in 80% DMSO mix-
tures, and elevated temperatures were necessary to
observe measurable rates of hydrolysis.
Analyses of the reaction progress of the aqueous
hydrolysis of phenyl phosphate and methyl phosphate by
31P NMR showed that both reactions followed simple
first-order kinetics to completion. The measured rate for
the aqueous hydrolysis of methyl phosphate matched that
reported previously by Wolfenden et al.12 The rate of
hydrolysis of the phenyl phosphate dianion at pH 11,
which is in the pH-independent range, was measured at
temperatures of 140, 150, 160, and 170 °C. These data
were used to construct the Eyring plot shown in Figure
6, which yielded a value for the enthalpy of activation of
37 ( 1 kcal/mol. The value for the entropy of activation
at 39 °C was calculated for comparison with the reported
value of +3.5 eu for pNPP at the same temperature.17
Discu ssion
The value for ∆S‡ was obtained by using the extrapo-
Na tu r e of th e p NP P Rea ction in DMSO. An indica-
tion of the transition state structure can be obtained from
the kinetic isotope effects. A considerable body of data
39
lated rate constant at 39 °C to calculate ∆G‡ and then
subtracting ∆H‡ to obtain T∆S. This yields a value for
∆S‡ ) +7.3 ( 0.6 eu.
has been obtained for the uncatalyzed hydrolysis and the
39
18
enzymatic reactions of pNPP. The
k
isotope effect
The aqueous reactions were well behaved, but the 80%
DMSO reactions were accompanied by the gradual
formation of brown-colored material and the gradual loss
of the 31P NMR signal. The NMR signal could be partially
restored by treating the sample with EDTA; thus, it is
possible that metals were being leached out of the quartz
tubes by the solution. While neither of these reactions
in DMSO/water could be followed to sufficient turnover
for calculation of a rate constant, under some conditions
the formation of product could be detected before the
NMR signal was lost. The reactions were examined
across a range of hydroxide concentrations in order to
ensure that the reaction of the dianion was being
observed (in aqueous solution, the monoanionic species
undergoes hydrolysis much faster than the dianion). It
was observed that rapid hydrolysis occurred if tetrabu-
bridge
gives a measure of the degree of bond cleavage, and the
15k isotope effect is sensitive to charge delocalization in
18
the leaving group.9,18 The
k
isotope effect in
nonbridge
reactions of phosphate esters is very small and inverse
in reactions with a loose transition state in which the
phosphoryl group resembles metaphosphate and becomes
normal in mechanisms with more nucleophilic participa-
tion such as in the hydrolysis reactions of phosphodi-
esters and triesters.9,18,19 A comparison of the isotope
effects measured in 90% DMSO solution with those
previously measured in aqueous solution is shown in
Scheme 1.
The possibility that the DMSO induces a change to a
more nucleophilic mechanism is unlikely in view of the
18
unaltered magnitude of
k
. This conclusion is
nonbridge
(16) Kirby, A. J .; Varvoglis, A. G. J . Am. Chem. Soc. 1967, 89, 415-
(18) Hengge, A. C.; Tobin, A. E.; Cleland, W. W. J . Am. Chem. Soc.
1995, 117, 5919-5926.
(19) Hengge, A. C. In Enzymatic Mechanisms; Frey, P. A., Northrop,
D. B., Eds.; IOS Press: Amsterdam, 1999; pp 72-84.
423.
(17) Kirby, A. J .; J encks, W. P. J . Am. Chem. Soc. 1965, 87, 3209-
3216.

Solvation Effects on Phosphate Ester Hydrolysis
J . Org. Chem., Vol. 67, No. 4, 2002 1219
F igu r e 7. Optimized geometry of methyl phosphate at the alkaline phosphatase active site.
supported by the independence of the hydrolysis rate on
the hydroxide concentration. Although the entropy of
activation becomes slightly negative in the DMSO/water
reaction, the value is still much less negative than values
typical of bimolecular reactions such as alkaline ester
hydrolysis. Because of the significant effect that solvation
can have on entropies of activation, the change in this
parameter cannot be confidently ascribed to a mecha-
tion state or the ground state relative to the situation in
aqueous solution. The simplest explanation of all the data
is the one originally proposed by Kirby, a weakening of
the P-O ester bond in the ground state at high DMSO
concentrations relative to aqueous solution. This would
explain both the lowered enthalpy of activation and the
reduced magnitudes of the isotope effects in the leaving
group. Because of the logarithmic relationship between
bond length and bond order, a relatively small increase
in P-O bond length can result in significant weakening
of the scissile ester bond and reduced KIEs. With ap-
plication of Pauling’s rule,21 a desolvation-induced in-
crease in P-O bond length of 0.1 Å corresponds to a
decrease in bond order from 1.0 to 0.68.
Oth er Ar yl P h osp h a tes. The hydrolysis reactions of
p-chlorophenyl phosphate, of m-bromophenyl phosphate,
and of p-cyanophenyl phosphate are all substantially
faster in 95% DMSO relative to the aqueous reactions.
These results with the data for pNPP are shown graphi-
cally in the Brønsted plot of Figure 5. If all four points
are included in the correlation, the best linear fit has a
slope of -1.4; if only the first three points are used, the
slope is -1.1. Both values are close to the value of -1.23
reported for the aqueous hydrolysis of a series of aryl
phosphate dianions.16 The similarity in the slopes indi-
cates that the hydrolysis rates of these four aryl phos-
phates are accelerated to a similar degree.
P h en yl a n d Meth yl P h osp h a te. For reasons de-
scribed in the Experimental Section, the effect of DMSO
on the reactions of phenyl phosphate and of methyl
phosphate were examined in 80% DMSO rather than the
95% used in the reactions of the substituted aryl phos-
phates. Nonetheless, a rate acceleration of nearly 3 orders
of magnitude is expected if the effect of DMSO on these
esters is the same as reported for p-nitrophenyl phos-
phate.6 Hydrolysis occurs if hydroxide is left out, and
under such conditions the monoanion is most likely the
reactive species (the pKa of the phosphoryl group will be
elevated by the added DMSO). Unexpectedly, the hy-
drolysis rates of the dianions of phenyl phosphate and
of methyl phosphate in 80% DMSO are not accelerated
and in fact are slower than the aqueous reactions; the
nistic change and is more likely due to solvation effects.
18
Both
k
and 15k are reduced in magnitude by about
bridge
one-fourth in the reaction in 90% DMSO. Isotope effects
measure the difference between the ground state and
transition structure; thus, the reduced magnitudes of
these isotope effects are consistent with either an earlier
transition state with less P-O bond cleavage or a solvent-
induced weakening of the scissile P-O bond in the
reactant. In this reaction, the reactant and transition
state bear a charge of -2. In the late transition state of
the aqueous reaction, this charge is dispersed relatively
equally between the phosphoryl group and the leaving
group (Figure 1). The high DMSO content results in a
less protic solvent compared to water. This should favor
a transition state with as much, or more, dispersal of
negative charge relative to that of the aqueous reaction.20
Thus, a significantly earlier transition state, which would
have less charge dispersal, would not be expected.
However, the Hammond postulate predicts that a desta-
bilization of the reactant, assuming the stability of the
product is unaltered, should lead to a somewhat earlier
transition state. Thus, it is possible that both effects
contribute to the reduced magnitudes of 18kbridge and 15k.
Additional information comes from the enthalpy of
activation, which in 95% DMSO is substantially reduced
from its value in aqueous solution (Table 3). This change
is consistent with either destabilization of the reactant,
as proposed by Kirby, increased stabilization of the
transition state, or a combination of both effects. As
stated previously, it is not expected that a dianionic
transition state should be better stabilized in the less
protic 90% DMSO solvent mixture. The altered KIEs
indicate that, in addition to the lowered activation
energy, there are changes in the structure of the transi-
(20) March, J .; Smith, M. B. In March’s Advanced Organic Chem-
istry; Fifth ed.; Wiley: New York, 2001, p 450-451.
(21) Pauling, L.; The Nature of the Chemical Bond, third ed.; Cornell
University Press: Ithaca, NY, 1960; pp 255-260.

1220 J . Org. Chem., Vol. 67, No. 4, 2002
Grzyska et al.
rates were too slow to allow determination of the rate
constant. It is intriguing to note from the Brønsted plot
in Figure 5, that the point for p-chlorophenyl phosphate,
which has the most basic leaving group of the set,
deviates negatively from the linear behavior of the other
three aryl phosphates. The cumulative results demon-
strate that the accelerating effect of added DMSO holds
only for phosphates having leaving groups less basic than
phenol.
unexpectedly exhibits a lengthened P-O ester bond
(Figure 7, Table 1). Since enzymes are dynamic during
catalysis, and evidence has been reported for a confor-
mational change during catalysis,25 the results must be
interpreted with caution. However, they raise the pos-
sibility that the hydrogen bonding and dipole-dipole
interactions at the active site might be less stabilizing
for the phosphate ester dianion than the hydrogen
bonding network present in aqueous solution. Such
destabilization would favor a loose transition state like
that for hydrolysis in solution for phosphate monoesters
having good leaving groups.
While the experiments in DMSO/water mixtures sug-
gest that reactant destabilization will be effective only
for esters having good leaving groups, it is possible that
a phosphatase could take advantage of this mechanism
by reducing the pKa of the leaving group, either by
protonation or by complexation to a metal ion. The
computational results here only show that such desta-
bilization may be feasible. Further experiments are
necessary to determine whether such effects in fact occur
or whether they contribute to enzymatic catalysis of
phosphoryl transfer.
It is uncertain why the reaction rate is slower for
methyl and phenyl phosphate dianions in aqueous DMSO
than in water. The extrapolated rate constant for phenyl
phosphate hydrolysis in water at 39 °C is close to that
expected from a Brønsted plot for aryl phosphates.16 This
rate constant is also very close to that reported for methyl
phosphate, which, as previously noted,12 is orders of
magnitude faster that its predicted value. The similarity
of the values for the aqueous hydrolysis of these two
phosphate esters, despite their different leaving groups,
suggests that another mechanism may be operative for
the aqueous hydrolysis of phosphate esters having leav-
ing groups with basicity greater than that of phenol. A
mechanistic difference might also explain why the ac-
celerating effect from added DMSO disappears for esters
with more basic leaving groups. However, the similarity
of the entropies of activation for methyl (+3.7 eu),12
phenyl (+7.3, this work), and p-nitrophenyl phosphate
(+3.5 eu)17 hydrolysis are consistent with similar mech-
anisms. Further work is needed to resolve this question.
Com p u ta tion a l Resu lts. The computational exami-
nation of the scissile P-O ester bond indicates a trend
toward a longer, weaker bond as the leaving group pKa-
decreases (Table 1). The same trend is seen in X-ray
structures of these phosphate esters.22-24 (In these struc-
tures, the P-O ester bonds are from 0.05 to 0.08 Å
shorter than the calculated bond lengths in water. Part
of this is likely due to the close contacts made with the
ammonium and potassium counterions in these struc-
tures.)
The calculations also indicate a trend toward a longer,
weaker bond as solvation of the phosphoryl group by
hydrogen bonding is diminished. This trend is seen
whether one compares aqueous structures with those in
DMSO or with the gas-phase structures and is dimin-
ished as the leaving group becomes more basic. A
difference of 0.1 Å will represent a significant weakening
of this bond because of the logarithmic relationship
between bond length and bond order.21 The greatest effect
is predicted to occur with the labile substrate p-nitro-
phenyl phosphate, with markedly less of an effect occur-
ring with phenyl and methyl phosphates. This is the
same trend seen in the experimental results.
Con clu sion s
Both the experimental and computational results
indicate that the rate acceleration imparted by the
aprotic solvent is limited to phosphate monoesters having
leaving groups less basic than phenol. The most likely
origin of the rate acceleration is that originally proposed
by Kirby, namely, a destabilization of the P-O ester bond
induced by loss of solvation of the anionic phosphoryl
group. This is supported by the lowered enthalpic barrier,
the smaller isotope effects in the leaving group, and the
results of computational modeling. The kinetic isotope
effects rule out a change to a more nucleophilic mecha-
nism and indicate that the transition state remains loose.
The computational results indicate that this desolvation-
induced effect is most significant with activated esters
and is much smaller with an alkyl phosphate or an
phenyl phosphate. However, with phenyl and methyl
phosphate dianions, the acceleration from added DMSO
is not lost, but the hydrolysis rates are significantly
slower than that in aqueous solution. The reasons for this
are not clear, but it is pertinent that the majority of what
is known about phosphate ester hydrolysis has come from
studies of the more labile aryl phosphates.
Ack n ow led gm en t. This work was supported by
NIH Grant GM47297. We thank Richard Wolfenden for
advice on conducting the sealed tube kinetics experi-
ments.
Su p p or tin g In for m a tion Ava ila ble: Proton and phos-
phorus NMR spectra of the sodium salts of p-chloro, p-cyano,
m-bromo, and methyl phosphates, and the z-matrices for each
of the computational structures listed in Table 1. This material
is available free of charge via the Internet at http://pubs.acs.org.
Alk a lin e P h osp h a ta se. The optimized methyl phos-
phate molecule in the active site of alkaline phosphatase
(22) J ones, P. G.; Sheldrick, G. M.; Kirby, A. J .; Abell, K. W. Y. Acta
Crystallogr., Sect. C 1984, C40, 550-552.
(23) Garbassi, F.; Giarda, L.; Fagherazzi, G. Acta. Crystallogr., Sect.
B 1972, 28, 1665-1670.
(24) Caughlan, C. N.; Mazhar, U.-H. Inorg. Chem. 1967, 6, 1998-
2002.
J O016104P
(25) Halford, S. E.; Bennett, N. G.; Trentham, D. R.; Gutfreund, H.
Biochem. J . 1969, 114, 243-251.