Comparisons of phosphorothioate with phosphate transfer reactions for a monoester, diester, and triester: Isotope effect studies

Article abstract of DOI:10.1021/ja0340026

Phosphorothioate esters are sometimes used as surrogates for phosphate ester substrates in studies of enzymatic phosphoryl transfer reactions. To gain better understanding of the comparative inherent chemistry of the two types of esters, we have measured equilibrium and kinetic isotope effects for several phosphorothioate esters of p-nitrophenol (pNPPT) and compared the results with data from phosphate esters. The primary ¹⁸O isotope effect at the phenolic group (¹⁸k_bridge), the secondary nitrogen-15 isotope effect (¹⁵k) in the nitro group, and (for the monoester and diester) the secondary oxygen-18 isotope effect (¹⁸k_nonbridge) in the phosphoryl oxygens were measured. The equilibrium isotope effect (EIE) ¹⁸k_nonbridge for the deprotonation of the monoanion of pNPPT is 1.015 ± 0.002, very similar to values previously reported for phosphate monoesters. The EIEs for complexation of Zn²⁺ and Cd²⁺ with the dianion pNPPT^2- were both unity. The mechanism of the aqueous hydrolysis of the monoanion and dianion of pNPPT, the diester ethyl pNPPT, and the triester dimethyl pNPPT was probed using heavy atom kinetic isotope effects. The results were compared with the data reported for analogous phosphate monoester, diester, and triester reactions. The results suggest that leaving group bond fission in the transition state of reactions of the monoester pNPPT is more advanced than for its phosphate counterpart pNPP, while alkaline hydrolysis of the phosphorothioate diester and triester exhibits somewhat less advanced bond fission than that of their phosphate ester counterparts.

Full text of DOI:10.1021/ja0340026

Published on Web 05/30/2003
Comparisons of Phosphorothioate with Phosphate Transfer
Reactions for a Monoester, Diester, and Triester: Isotope
Effect Studies
Irina E. Catrina† and Alvan C. Hengge*
Contribution from the Department of Chemistry and Biochemistry, Utah State UniVersity,
Logan, Utah 84322-0300
Received January 1, 2003; E-mail: hengge@cc.usu.edu
Abstract: Phosphorothioate esters are sometimes used as surrogates for phosphate ester substrates in
studies of enzymatic phosphoryl transfer reactions. To gain better understanding of the comparative inherent
chemistry of the two types of esters, we have measured equilibrium and kinetic isotope effects for several
phosphorothioate esters of p-nitrophenol (pNPPT) and compared the results with data from phosphate
esters. The primary 18O isotope effect at the phenolic group (18kbridge), the secondary nitrogen-15 isotope
effect (15k) in the nitro group, and (for the monoester and diester) the secondary oxygen-18 isotope effect
18
(18knonbridge) in the phosphoryl oxygens were measured. The equilibrium isotope effect (EIE)
k
for
nonbridge
the deprotonation of the monoanion of pNPPT is 1.015 ( 0.002, very similar to values previously reported
for phosphate monoesters. The EIEs for complexation of Zn2+ and Cd2+ with the dianion pNPPT2- were
both unity. The mechanism of the aqueous hydrolysis of the monoanion and dianion of pNPPT, the diester
ethyl pNPPT, and the triester dimethyl pNPPT was probed using heavy atom kinetic isotope effects. The
results were compared with the data reported for analogous phosphate monoester, diester, and triester
reactions. The results suggest that leaving group bond fission in the transition state of reactions of the
monoester pNPPT is more advanced than for its phosphate counterpart pNPP, while alkaline hydrolysis of
the phosphorothioate diester and triester exhibits somewhat less advanced bond fission than that of their
phosphate ester counterparts.
Introduction
Phosphorothioate esters are counterparts of phosphate esters
in which an oxygen atom in the phosphoryl group has been
replaced by sulfur. Such compounds have been used as stereo-
chemical probes and in mechanistic studies of enzymatic phos-
phoryl transfer. The uncatalyzed reactions of phosphorothioate
esters have been less extensively studied than phosphates. We
previously compared the kinetic and thermodynamic parameters,
and solvent effects, for hydrolysis reactions of phosphate and
phosphorothioate monoesters.1 The present study extends our
investigation of the chemistry and the hydrolysis mechanisms
of phosphorothioate esters. We report the equilibrium 18O
isotope effects (EIEs) for deprotonation and for zinc and
cadmium ion complexation with the monoester p-nitrophenyl
phosphorothioate (pNPPT) and the kinetic isotope effects (KIEs)
for the hydrolysis of the dianion and monoanion of pNPPT,
the diester ethyl p-nitrophenyl phosphorothioate (EtOpNPPT),
and the triester dimethyl p-nitrophenyl phosphorothioate
((MeO)2pNPPT). Figure 1 shows the positions at which isotope
effects in each compound were measured and the notation used.
Comparisons of these data with previously determined results
for kinetic isotope effects and equilibrium isotope effects with
phosphate esters can shed light on the consequences of sulfur
Figure 1. Phosphorothioate esters for which isotope effect measurements
are reported: red, nonbridge phosphoryl oxygen atoms (18knonbridge); blue,
bridge oxygen atom, site of bond fission (18kbridge); green, nitrogen atom of
the leaving group (15k).
substitution on the comparative mechanistic chemistry and the
transition states of these compounds and supplement information
from linear free energy relationships (LFER).
Figure 2 shows the mechanistic extremes for phosphoryl and
thiophosphoryl transfer reactions. In a fully dissociative mech-
anism (A) a (thio)metaphosphate intermediate is formed in the
rate-determining step. In the opposite extreme of a fully asso-
ciative mechanism (C) a pentacoordinate phosphorane inter-
mediate is formed. Mechanism B is concerted; in such a mech-
anism the transition state could in principle be loose (resembling
that of mechanism A) or tight (resembling mechanism C).
Considerable experimental evidence that has been reviewed2
suggests that phosphate (X ) O) monoesters undergo reactions
by a concerted mechanism having a very loose transition state,
while phosphorothioate (X ) S) monoesters are believed to form
† Present address: Department of Chemistry, University of Rochester,
Rochester, NY 14627-0216.
(1) Catrina, I. E.; Hengge, A. C. J. Am. Chem. Soc. 1999, 121, 2156-2163.
9
7546
J. AM. CHEM. SOC. 2003, 125, 7546-7552
10.1021/ja0340026 CCC: $25.00 © 2003 American Chemical Society

Chemistry and Hydrolysis of Phosphorothioates
A R T I C L E S
Figure 2. Mechanistic possibilities for phosphoryl (X ) O) or thiophos-
phoryl (X ) S) transfer: the fully dissociative mechanism (A) and the fully
associative (C) and intermediate concerted (B) mechanisms. The P-O bond
order to the nonbridge oxygen atoms increases in the transition state in
mechanism A and decreases in C. By contrast the bending modes will loosen
in the transition state of A but tighten in C.
Figure 3. Mixtures of isotopic isomers used for measurements of the 18
O
a thiometaphosphate intermediate. Both phosphoryl and thio-
phosphoryl transfer reactions are thought to become more
associative in the progression from monoester to diester to
triester. Linear free energy relationships (LFER) that examine
the interrelationship between ânuc and âleaving group 3 indicate that
with good (aryloxy) leaving groups even reactions of phospho-
triesters remain concerted, although the transition state is tight
and phosphorane-like.
18
isotope effects. A and C are used for measurement of
k
in the
nonbridge
monoester and diester, respectively. Mixtures represented by B, D, and E
18
are used in measurement of
k
in the monoester, diester, and triester.
bridge
Synthesis. [14N]-p-nitrophenol, [15N]-p-nitrophenol, and [15N,18O]-
p-nitrophenol were synthesized as previously described.5
(A) Isotopic Isomers of the Monoester p-Nitrophenyl Phospho-
rothioate. Natural abundance p-NPPT was synthesized from p-ni-
trophenol and PSCl3 with hydrolysis of the p-nitrophenyl thiophos-
phorodichloridate intermediate as previously reported.1
In the same manner, [15N,nonbridge-18O2]-p-nitrophenyl phospho-
rothioate was synthesized, but with the use of [15N]-p-nitrophenol and
then H218O to hydrolyze the intermediate. The most economical use of
labeled water was obtained when the p-nitrophenyl thiophosphoro-
dichloridate intermediate was isolated6 followed by hydrolysis in the
presence of pyridine using a 3:1 ratio of 18O-labeled water:intermediate.
31P NMR analysis showed the product to be 87 ( 1% 18O2, with the
remainder 18O16O.
The isotope effects depicted in Figure 1 give information
regarding the transition sates for the hydrolysis reactions of
the phosphorothioate esters shown. The secondary 15k and the
18
primary
k
isotope effects are sensitive to the charge on
bridge
the leaving group and the degree of P-O bond fission, re-
18
spectively. The secondary
k
both gives informa-
nonbridge
tion about hybridization changes occurring at phosphorus
during the reaction, and is sensitive to the protonation state of
the reactant.
[14N]-p-Nitrophenyl phosphorothioate was made in the same way,
using [14N]-p-nitrophenol and natural abundance water.
Experimental Section
To closely reconstitute the natural abundance of 15N (as confirmed
Materials. Reagents and solvents were commercial products and
were used as received unless otherwise noted. Pyridine was distilled
from calcium hydride. Tetrahyrofuran (THF) was distilled from sodium
and benzophenone.
by isotope ratio mass spectrometry), [15N,nonbridge-18O2]-p-nitrophenyl
phosphorothioate and [14N]-p-nitrophenyl phosphorothioate were mixed,
18
and this mixture (A in Figure 3) was used to measure
k
.
nonbridge
18
To prepare the mixture needed for measurement of
k
(B in
bridge
Isotopically Labeled Compounds Needed for the Measurement
of Isotope Effects. The 18O kinetic isotope effects were measured by
the remote label method, using the nitrogen atom in the nitro group as
a reporter for isotopic fractionation in the labeled oxygen positions.4
Figure 3 shows the mixtures of isotopically labeled compounds needed
for these experiments. For the KIE in the nonbridging oxygen atoms,
the compound synthesized with 18O in these positions and 15N in the
reporter position is mixed with the isotopic isomer bearing the natural
Figure 3) [14N]-p-nitrophenol and [15N,18O]-p-nitrophenol were mixed
to closely reconstitute the natural abundance of 15N, and this mixture
was converted to the phosphorothioate ester as described above.
(B) Isotopic Isomers of the Diester Ethyl-p-nitrophenyl Phos-
phorothioate. Ethyl [14N]-p-nitrophenyl phosphorothioate (EtOpNPPT)
was synthesized by adding a solution of 5.5 mmol of anhydrous ethanol
in 3 mL of dry THF dropwise to a solution of 5.5 mmol of [14N]-p-
nitrophenyl thiophosphorodichloridate and 5.5 mmol of dry pyridine
in 11.4 mL of dry THF. After 10-15 min of stirring at room
temperature under nitrogen, a solution of 8.4 mL of water and 5.7 mmol
of dry pyridine was added, and this mixture was stirred for an additional
10 min. THF was then removed by rotary evaporation, and the residue
was diluted with water. The mixture was filtered, concentrated in vacuo,
and then purified by reverse-phase column chromatography,7 eluting
with a solution of 1:1 acetonitrile:TEAB (triethylammonium bicarbon-
ate), 0.1 M buffer, pH 7.5. EtOpNPPT elutes first, and after solvent
abundance of oxygen and 14N in the reporter position (A for the
18
monoester and C for the diester). For measurement of
k
, the
bridge
mixtures of isotopic isomers represented in B, D, and E are needed.
Natural abundance compounds are used for measurements of 15k. All
of the synthesized phosphorothioate esters had 1H and 31P NMR spectra
consistent with their assigned structures.
(2) Hengge, A. C. In ComprehensiVe Biological Catalysis: A Mechanistic
Reference; Sinnott, M., Ed.; Academic Press: San Diego, CA, 1998; Vol.
1, pp 517-542. Thatcher, G. R. J.; Kluger, R. AdV. Phys. Org. Chem. 1989,
25, 99-265.
(3) Ba-Saif, S. A.; Waring, M. A.; Williams, A. J. Am. Chem. Soc. 1990,
112, 8115-8120. Ba-Saif, S. A.; Waring, M. A.; Williams, A. J. Chem.
Soc., Perkin Trans. 2 1991, 1653-1659.
(5) Hengge, A. C.; Edens, W. A.; Elsing, H. J. Am. Chem. Soc. 1994, 116,
5045-5049.
(6) Tolkmith, H. J. Org. Chem. 1958, 23, 1685-1690.
(7) Kuhler, T. C.; Lindsten, G. R. J. Org. Chem. 1983, 48, 3589-3591.
(4) Hengge, A. C. Acc. Chem. Res. 2002, 35, 105-112.
9
J. AM. CHEM. SOC. VOL. 125, NO. 25, 2003 7547

A R T I C L E S
Catrina and Hengge
evaporation the TEAB is removed by repeated addition/evapora-
tions of methanol. Stock solutions of the diester in water were stored
at -20 °C.
In a similar way, [15N,18O-nonbridge]-ethyl p-nitrophenyl phospho-
rothioate was synthesized. A solution of 0.83 mmol ethanol in 0.45
mL of dry THF was added dropwise to a solution of 0.83 mmol of
[15N]-p-nitrophenyl thiophosphorodichloridate and 0.83 mmol of dry
pyridine in 1.9 mL of dry THF. This final solution was stirred for 10
min under nitrogen. Subsequently a mixture of 600 µL of H218O with
70 µL of dry pyridine was added and the reaction stirred for another
10 min. The purification procedure is the same as that described above.
This compound and [14N]-EtOpNPPT were mixed to reconstitute the
15N natural abundance (C in Figure 3) and were the reactants for
18
determining
k
.
nonbridge
18
The mixture used to measure
k
with ethyl p-nitrophenyl
bridge
phosphorothioate (D in Figure 3) was prepared in the same manner as
14N-EtOpNPPT but with use of the mixture of [14N]-p-nitrophenol and
[15N,18O]-p-nitrophenol (at 15N natural abundance).
To measure the 15N isotope effect, the natural abundance diester
was used, which was synthesized from natural abundance p-nitrophenol
as described for [14N]-EtOpNPPT.
(C) Isotopic Isomers of the Triester Dimethyl-p-nitrophenyl
Phosphorothioate. Natural abundance dimethyl-p-nitrophenyl phos-
phorothioate was prepared by adding a mixture of 8.2 mmol of
p-nitrophenol, 0.8 mmol of dimethylamino pyridine, and 8.2 mmol of
triethylamine in 5 mL of THF to a solution of commercial dimethyl-
chlorothiophosphate (8.2 mmol) in 10 mL of THF dropwise over 5
min under nitrogen. After 4 h the reaction mixture was added to 15
mL of water and extracted with ether (3 × 25 mL). The ether layers
were dried over MgSO4 and concentrated by rotary evaporation. The
product was isolated as a white solid after purification by flash
Figure 4. (A) Plot of the difference in chemical shifts between nonbridge
18O2-labeled and unlabeled pNPPT as a function of pH, used for calculation
of the equilibrium isotope effect for the deprotonation of pNPPT-. (B)
Chemical shift of unlabeled pNPPT as a function of pH, used for the pKa
determination.
details of the data analysis and the experimental procedures used for
the hydrolysis reactions of the pNPPT monoester monoanion and
dianion, the ethyl p-nitrophenyl phosphorothioate diester, and the
dimethyl p-nitrophenyl phosphorothioate triester are given in the
Supporting Information.
chromatography eluting with hexane/ethyl acetate.
18
The mixture used to measure
k
with dimethyl-p-nitrophenyl
bridge
phosphorothioate (E in Figure 3) was prepared in the same manner,
but with use of the mixture of [14N]-p-nitrophenol and [15N,18O]-p-
nitrophenol.
Determinations of Equilibrium Isotope Effects for Deprotonation
and Metal Ion Complexation of pNPPT. The changes in chemical
shifts with pH for unlabeled pNPPT were used to calculate the second
pKa of the pNPPT monoanion (see Supporting Information). No attempt
was made to measure the first pKa, which is <1.8
Results
pKa of pNPPT and the Equilibrium Isotope Effect for
Deprotonation. The 18O isotope-induced differences in the 31
P
chemical shift of unlabeled and nonbridge-18O2-labeled pNPPT
as a function of pH are shown in Figure 4. If one or both
oxygens in a phosphate or phosphorothioate ester are substituted
with 18O, the 31P NMR chemical shift is shifted upfield by about
0.02 ppm/atom.11 Protonation of the phosphorothioate mono-
ester causes the 31P chemical shift to move downfield. Since
the 18O-labeled species is preferentially protonated, the isotope
effect causes the 31P signals of the 16O- and 18O-labeled pNPPT
to move together as the thiophosphoryl group is titrated. The
separation reaches a minimum at the pKa and then increases as
titration becomes complete.
The differences in 31P chemical shift between 16O and nonbridge-
18O pNPPT were used to calculate the isotope effect for deprotonation
and for complexation of pNPPT with zinc and cadmium. This technique
was introduced by Ellison and Robinson to determine the EIE for
deprotonation of formic acid9 and has also been used to determine the
EIE for deprotonation of phosphate esters.10 See Supporting Information
for full details.
Kinetic Isotope Effect Determinations. The experimental proce-
dures used to measure these isotope effects were similar to those used
to measure KIEs in phosphoryl transfer reactions in which the leaving
group is p-nitrophenol.4 In summary, the reaction is initiated and
allowed to proceed to from 30 to 70% of completion. The reaction is
then stopped or slowed dramatically by adjusting the pH and/or
temperature and assayed to determine the precise fraction of reaction.
The p-nitrophenol produced in this first portion of the reaction is
separated from the unreacted ester. The unreacted ester is then
completely hydrolyzed, and the portion of p-nitrophenol released in
the second portion of the reaction is also isolated. Both samples of
p-nitrophenol are purified and subjected to isotope ratio analysis. The
The rapid hydrolysis of pNPPT at low pH 1 prevented
collection of data at pH < 2. A fit of the data as described in
the Supporting Information gives a value for the EIE for
deprotonation of 1.015 ( 0.002. A fit of the variation in 31P
chemical shift with pH gave a value for the first pKa of pNPPT
of 3.70 ( 0.03, close to a previously reported value of 3.6 8
and within experimental uncertainty of the kinetically deter-
mined value of 3.68 ( 0.08.1
(8) Domanico, P.; Mizrahi, V.; Benkovic, S. J. In Mechanisms of Enzymatic
Reactions: Stereochemistry; Frey, P. A., Ed.; Elsevier: New York, 1986;
pp 127-137.
(11) Frey, P. A.; Sammons, D. Science 1985, 228, 541-545.
(12) Czyryca, P. G.; Hengge, A. C. Biochim. Biophys. Acta 2001, 1547, 245-
253.
(13) Hengge, A. C.; Cleland, W. W. J. Am. Chem. Soc. 1990, 112, 7421-7422.
(14) Caldwell, S. R.; Raushel, F. M.; Weiss, P. M.; Cleland, W. W. Biochemistry
1991, 30, 7444-7450.
(9) Ellison, S. L. R.; Robinson, M. J. T. J. Chem. Soc., Chem. Commun. 1983,
745-746.
(10) Knight, W. B.; Weiss, P. M.; Cleland, W. W. J. Am. Chem. Soc. 1986,
108, 2759-2761.
9
7548 J. AM. CHEM. SOC. VOL. 125, NO. 25, 2003

Chemistry and Hydrolysis of Phosphorothioates
A R T I C L E S
Table 1. 31P Chemical Shifts (ppm) of 18O-Labeled and Unlabeled
dianions has not been reported, precluding a direct comparison
of the fraction of bond fission in the two reactions, though the
large magnitudes of both âlg values suggest a high degree of
bond fission in both reactions. Reported values of âlg are also
similar between phosphorothioate and phosphate diesters. The
alkaline hydrolysis of uridine 3′-aryl phosphorothioates gives
a âlg value of -0.55 ( 0.05,19 within experimental error of the
âlg ) -0.54 for the reaction of the corresponding 3′-aryl
phosphates.20 A value of âlg ) -0.35 is observed for the alkaline
hydrolysis of diethyl aryl phosphorothioate triesters, while the
value for phosphate analogues is -0.43.21
As for monoesters, âeq values have not been reported for
phosphorothioate diesters or triesters to our knowledge. The
similarity of the data is suggestive of at least generally similar
transition states. However, there is reason to suspect a different
âeq value may hold for the sulfur analogues. Phosphorothioate
monoesters have 31P chemical shifts that are about 40 ppm
downfield of their phosphate counterparts, indicating a difference
in the electron density around phosphorus in the two types of
compounds. In addition, as noted and discussed elsewhere in
this paper, protonation of a phosphorothioate monoester results
in a downfield change in the 31P chemical shift, in contrast to
a movement upfield in phosphate monoesters. Last, it is known
that the charge distribution differs in phosphorothioate dianions
in that a full or nearly full negative charge resides on the sulfur
atom while the remaining charge is shared among the two
oxygen atoms.11,22 Changes that affect the electron distribution
in phosphoryl groups can result in changes in âeq; for example,
the effective charge on the aryl group in aryl diethyl phosphate
triesters (+0.87) is significantly different than that in aryl
diphenyl analogues (+0.33).15 Thus the KIE measurements
reported here add to existing knowledge from LFER data. The
EIE data also adds to our understanding of the differences and
similarities between phosphates and phosphorothioates, as
discussed below.
3 mM pNPPT as a Function Zn2+ Concentrationa
Zn2+ (M)
16Oδ
18Oδ
16Oδ − 18Oδ
0
0.5
1.5
41.235
32.612
32.416
41.181
32.551
32.354
0.054
0.061
0.062
a The last column is the difference between the two chemical shifts.
Table 2. 31P Chemical Shifts (ppm) of 18O-Labeled and Unlabeled
3 mM pNPPT as a Function Cd2+ Concentrationa
Cd2+ (mM)
16Oδ
18Oδ
16Oδ − 18Oδ
0
5
15
45
60
100
130
150
41.184
34.971
32.915
32.786
33.290
33.908
34.479
34.268
41.116
34.900
32.844
32.713
33.218
33.834
34.408
34.197
0.068
0.071
0.071
0.073
0.072
0.074
0.071
0.071
a The last column is the difference between the two chemical shifts. From
the known stability constant for the Cd2+:pNPPT complex,1 full complex-
ation occurs.
The 18O-isotope-induced differences in 31P chemical shift data
of unlabeled and nonbridge-18O2-labeled pNPPT as a function
of Zn2+ concentration are given in Table 1 and data for Cd2+
complexation in Table 2. In both experiments, the negligible
changes observed in 16Oδ - 18Oδ compared to the deprotonation
experiment (compare with Figure 4A) indicate the absence of
a significant EIE for complexation in aqueous solution. Analysis
of the Cd2+ data as described in the Supporting Information
yields a calculated EIE of unity within experimental error (1.000
( 0.001). The ratios of Cd2+:pNPPT used are more than
sufficient to achieve full complexation, based on the previously
determined stability constant for complexation.1
Kinetic Isotope Effects. The KIE results for phosphorothioate
esters are presented in italics in Table 3. For comparison,
previously reported KIE results for the hydrolysis of the
phosphate monoester pNPP, the diester EtOpNPP and the range
of KIE data for other phosphodiesters with the same leaving
group, and the triester diethyl-pNPP are also shown.
Expected Ranges of the Isotope Effects. The small absolute
magnitudes of heavy-atom isotope effects, which are never more
than a few percent, necessitate the use of an isotopic counting
method, such as isotope ratio mass spectrometry, for their accu-
rate measurement. Such isotope effects have proven very useful
as mechanistic tools in the study of many reactions, including
those of phosphate esters.4 This prior work gives benchmarks
for evaluations of the isotope effects shown in Figure 1.
Discussion
What Is Known from Linear Free Energy Relationships.
The Brønsted correlation of rate constants with leaving group
pKa (âlg) is one measure of the degree of leaving group bond
fission in the rate-limiting transition state. To normalize the
âlg value for a specific reaction, the Brønsted correlation of
the equilibrium constant as a function of leaving group (âeq)
for the reaction of interest is needed. As has been reviewed,15
a generally used approach is to take the ratio of these values
(âlg/âeq) as an indication of the fraction of bond fission to the
leaving group in the transition state.16
The âlg for the uncatalyzed hydrolysis of aryl phosphorothio-
ate monoester dianions is -1.1,17 similar to the value of -1.2
for aryl phosphates.18 The âeq value for aryl phosphate dianions
is +1.36, reflecting an “effective charge” of +0.36 felt by the
aryl group in such compounds.15 The âeq for phosphorothioate
With the p-nitrophenyl leaving group, the secondary 15k
isotope effect is a measure of the negative charge borne in the
rate-determining transition state.4 Thus, it gives information as
to whether the leaving group departs as the anion or if
protonation of the leaving group has neutralized all or part of
the negative charge resulting from P-O bond fission. In a
transition state with extensive P-O bond fission and no charge
neutralization 15k reaches a maximum value of about 1.003 at
room temperature.4
18
The primary isotope effect
k
gives a measure of the
bridge
degree of fission of the P-O bond in the transition state. If
P-O bond fission is extensive,18kbridge is around 1.03 at room
(19) Almer, H.; Stromberg, R. J. Am. Chem. Soc. 1996, 118, 7921-7928.
(20) Davis, A. M.; Hall, A. D.; Williams, A. J. Am. Chem. Soc. 1988, 110,
5105-5108.
(21) Hong, S. B.; Raushel, F. M. Biochemistry 1996, 35, 10904-10912.
(22) Iyengar, R.; Eckstein, F.; Frey, P. A. J. Am. Chem. Soc. 1984, 106, 8309-
8310.
(15) Williams, A. AdV. Phys. Org. Chem. 1992, 27, 1-55.
(16) Leffler, J. E.; Grunwald, E. Rates and Equilibria of Organic Reactions;
Wiley: New York, 1963.
(17) Hollfelder, F.; Herschlag, D. Biochemistry 1995, 34, 12255-12264.
(18) Kirby, A. J.; Varvoglis, A. G. J. Am. Chem. Soc. 1967, 89, 415-423.
9
J. AM. CHEM. SOC. VOL. 125, NO. 25, 2003 7549

A R T I C L E S
Catrina and Hengge
Table 3. Kinetic Isotope Effects for the Hydrolysis of Phosphate and Phosphorothioate Esters (Phosphorothioate Data in Italics)a
18
18
reactant
15k
k
k
nonbridge
bridge
pNPP- (95 °C)
1.0004 ( 0.0002 5
1.0005 ( 0.0001
1.0028 ( 0.0002 5
1.0005 ( 0.0001
1.0027 ( 0.0001
1.0010 ( 0.0001
1.0007 - 1.0016
1.0010 ( 0.0002
1.0007 ( 0.0001 13
1.0004 ( 0.0002
1.0087 ( 0.0003 5
not determined
1.0189 ( 0.0005 5
1.0091 ( 0.0007
1.0237 ( 0.0007
1.0042 ( 0.0009
1.0042 - 1.0063
1.0020 ( 0.0006
1.0060 14
1.0184 ( 0.0005 5
1.0199 ( 0.0003
0.9994 ( 0.0005 5
1.0221 ( 0.0004
1.0135 ( 0.0013
not determined
1.0028 - 1.0056
1.0019 ( 0.0004
1.0063 ( 0.0002 14
pNPP- (30 °C)12
pNPP2- (95 °C)
pNPPT- (30 °C)b
pNPPT2- (50 °C)b
EtOpNPP- (95 °C)
range of phosphodiester data4
EtOpNPPT- (95 °C)
(EtO)2pNPP (25 °C)
(MeO)2pNPPT (30 °C)
1.0045 ( 0.0006
a Referenced data are from previously reported work. The KIEs for the monoester monoanion reactions were measured at the pH optima for these reactions
(pH 2.0 for pNPPT, pH 3.5 for pNPP). The monoester dianion KIEs were measured within the respective pH-independent ranges (at pH 10.0 and 11.0 for
pNPPT and pH 11 for pNPP). KIEs for the diester and triester reactions were measured in the region where the hydrolyses are first order in [-OH]: 1 N
[-OH] for EtOpNPP and EtOpNPPT, 0.2 N [-OH] for (MeO)2pNPPT, and 0.1 N [-OH] for (EtO)2pNPP. b The large difference in reactivity between the
monoanion and dianion of pNPPT prevented measurement of the KIEs for these two reactions at the same temperature.
temperature; protonation of the leaving group in the transition
state reduces this value.
18
Secondary
k
isotope effects have proven useful in
nonbridge
distinguishing the loose transition states such as those in
mechanism A of Figure 2, which are typical of uncatalyzed
monoester reactions, from associative transition states typical
of triesters.4 This is an R-secondary KIE adjacent to an atom
undergoing a hybridization change. In general, such isotope
effects are determined by alterations in bond order as well as
in bending and torsional vibrational modes; the latter effects
can be the dominant contributors to R-secondary KIEs.23 For
example, R-secondary deuterium isotope effects are normal for
carbon hybridization changes of the type sp3 to sp2 or sp2 to sp
due to loosening of bending modes.23 The P-O bond order
Figure 5. Contributing resonance structures in a phosphorothioate mo-
noester as the dianion (top) and monoanion (bottom). The resonance form
with a P-S double bond is a negligible contributor in the dianion but is
more important (though still the lesser contributor) in the monoanion
species.26 This would result in a reduction in bond order to the isotopically
labeled oxygen atoms (marked with asterisks) in the monoanion compared
to the dianion and a smaller isotope shift in the 31P NMR signal.
chemical shift, analogous to that seen in phosphate esters.
However the chemical shifts respond differently to protonation
in the two esters. Protonation of a phosphate monoester, or of
inorganic phosphate, causes the 31P chemical shift to move
upfield.10 In contrast, protonation of pNPPT causes the 31P
chemical shift to move downfield (Figure 4B). It was also noted
that the 18Oδ - 16Oδ for pNPPT at low pH, where it exists as
the monoanion, does not return to its value in the dianion (Figure
4A). In similar experiments with phosphate monoesters, this
isotopic separation returns to its original value when titration
is complete.10 Both differences in behavior likely originate from
dissimilarities in charge localization and bond order in phos-
phorothioates compared to phosphates. Experimental and com-
putational evidence indicate that in dianions of phosphorothioate
esters close to a full negative charge resides on the sulfur atom,
while the remaining charge is shared between the two nonbridge
oxygens11,26 (Figure 5). In a computational study of inorganic
thiophosphate26 it was found that essentially no P-S double
bond character is present in the trianion, consistent with
experimental data.27 The computational results predict gradually
more π bonding to sulfur as the protonation state increases, with
the expected P-S double bond manifested in the neutral
species.26 Extending these conclusions to pNPPT, somewhat
greater P-S bond order is expected in the monoanion than in
the dianion, resulting in a decrease in P-O bond order. Because
the magnitude of the 18Oδ - 16Oδ isotope shift is correlated
with P-O bond order, this should result in a smaller isotope
shift in the monoanion. The fit of the data in Figure 4 gives a
changes of the nonbridge oxygen atoms in the mechanisms of
18
Figure 2 would lead one to expect an inverse
k
for
nonbridge
dissociative transition states and a normal value for associative
ones. The contributions to the KIE from the bending modes
should be in the opposite direction. From the trends in reported
18
k
data for phosphate mono-, di-, and triesters, bond order
nonbridge
changes seem to be the dominant contributors. Values for
18
k
are slightly inverse (near unity) for reactions of
nonbridge
monoesters, which have dissociative transition states, while
18
k
becomes normal for diester (with a single exception,
nonbridge
which may be anomalous) and larger normal for triester
reactions, respectively.5,14,24
Equilibrium Isotope Effect for Protonation of pNPPT. The
equilibrium isotope effect of 1.015 for the deprotonation of
pNPPT is the same, within experimental uncertainty, of EIEs
reported for phosphate monoesters.10 The data confirm that
protonation occurs at oxygen rather than at sulfur, as expected
from relative basicity considerations. The similarity of the EIEs
indicates that sulfur substitution does not affect this isotope
effect. This result confirms the notion that the EIE for
deprotonation is dominated by differences in 16O-H and 18O-H
stretching modes25 with minimal influence from bending modes
that might be influenced by the larger size of sulfur.
In this work, as previous researchers have observed, 18O
substitution resulted in a small upfield change in the 31P
(23) Melander, L.; Saunders: W. H. Reaction Rates of Isotopic Molecules;
Wiley: New York, 1980; pp 172-174.
(24) Sowa, G. A.; Hengge, A. C.; Cleland, W. W. J. Am. Chem. Soc. 1997,
119, 2319-2320. Hengge, A. C.; Tobin, A. E.; Cleland, W. W. J. Am.
Chem. Soc. 1995, 117, 5919-5926. Hengge, A. C.; Cleland, W. W. J.
Am. Chem. Soc. 1991, 113, 5835-5841. Anderson, M. A.; Shim, H.;
Raushel, F. M.; Cleland, W. W. J. Am. Chem. Soc. 2001, 123, 9246-
9253.
(25) Deng, H.; Wang, J.; Callender, R.; Ray, W. J. J. Phys. Chem. B 1998, 102,
3617-3623.
(26) Liang, C.; Allen, L. C. J. Am. Chem. Soc. 1987, 109, 6449-6453.
(27) Arnold, J. R. P.; Lowe, G. J. Chem. Soc., Chem. Commun. 1986, 1487-
1486.
9
7550 J. AM. CHEM. SOC. VOL. 125, NO. 25, 2003

Chemistry and Hydrolysis of Phosphorothioates
A R T I C L E S
reduction in 18Oδ - 16Oδ from 0.066 ppm in the dianion to
0.054 ppm in the monoanion.
It is uncertain why protonation of pNPPT causes the 31P
chemical shift to move downfield, in contrast to the movement
upfield observed with phosphates. Calculations have been re-
ported of the comparative Mulliken atomic charges on phosphate
and thiophosphate in all four possible protonation states (from
the trianion to the neutral species).26 In these calculations the
charge on phosphorus is more altered in phosphate than in
thiophosphate by protonation; in phosphate, from the trianion
PO43- to the neutral species H3PO4 the Mulliken atomic charge
on phosphorus changed in the sequence 1.23, 1.30, 1.38, to 1.56.
The calculated changes in thiophosphate were in the same
direction, but smaller: 1.28, 1.29, 1.31, to 1.44.26 Another study
using different computational methods found that the conjugate
bases of phosphoric acid and thiophosphoric acid respond
Figure 6. Schematic diagram of the transition-state structure for the
hydrolysis of pNPPT dianion. P-O bond fission is extensive, and nearly a
full negative charge resides on the leaving group.
Inconsistencies in the synchronization of development of
charge and its delocalization by resonance have been well-
documented.32 Nitro compounds are particularly susceptible to
this phenomenon, and such transition-state imbalances in the
deprotonation of nitroalkanes (termed the “nitroalkane anomaly”)
have been central to the development of the principle of
nonperfect synchronization.32 One example is the deprotonation
of arylnitroalkanes (ArCH2NO2) by hydroxide ion in water. The
Hammett F value for the deprotonation rate is 1.28, while F for
the equilibrium is 0.83.33 This is unusual in that the rate of
deprotonation is more sensitive to substituent effects than the
equilibrium, implying that in the transition state more charge
is felt by the aryl group than in the fully deprotonated ion. The
most widely accepted explanation is that the transition state is
imbalanced in that delocalization of the negative charge into
the nitro group lags considerably behind deprotonation and
charge development.
Why there should be a difference in the synchronization of
charge development and delocalization between the reaction of
the phosphorothioate compared with the phosphate ester is
uncertain. This may arise from a difference in the electrostatic
description of the phosphoryl group compared with thiophos-
phoryl. In both transition states the leaving group resembles
p-nitrophenolate ion and is in proximity to the negatively
charged (thio)phosphoryl group. The extent to which charge is
delocalized, and presumably the synchronization with P-O bond
fission, may be influenced by electrostatic interactions between
these moieties.
-
oppositely to protonation.28 Protonation of H2PO4 caused the
computed gross atomic charge on phosphorus to change from
1.42 to 1.48, while protonation of H2PO3S- caused a change
from 1.01 to 0.85.28 Perhaps further computational study will
reveal which of these tendencies is correct and shed more light
on the reason for our NMR results.
Equilibrium Isotope Effect for Complexation of pNPPT.
In complexes with phosphorothioates, cadmium coordinates
virtually completely with sulfur. In zinc complexes with
phosphorothioates, about three-fourths of the species in solution
are sulfur coordinated, with the remainder coordinated to
oxygen.29 The lack of a significant change in 18Oδ - 16Oδ even
at the highest Zn2+ concentration indicates there is no significant
isotope effect. For Cd2+ the results were similar, indicating the
EIE is negligible (unity within experimental error). For com-
parison, the reported EIE for Ca2+ complexation with pNPP is
also near unity, 18K ) 0.997.30 The results indicate that
complexation in aqueous solution with either zinc or cadmium
does not result in significant changes in bonding within the
phosphoryl group. Cobalt(III) complexes with phosphate esters
have been found to result in significant EIEs,31 not unexpected
in light of the stronger coordinating properties of this metal ion.
Kinetic Isotope Effects. (A) Hydrolysis of the pNPPT
Dianion. The kinetic isotope effects in the leaving group in the
reaction of the dianion of pNPPT are indicative of significant
P-O bond fission and nearly a full negative charge on the
leaving group in the transition state. These KIEs were measured
at 50 °C in contrast to 95 °C for the pNPP data. Higher
Nonetheless, the18kbridge and 15k data together suggest that a
similar and substantial negative charge is delocalized in the
leaving groups of both reactions, even though P-O bond fission
is slightly greater in the reaction of pNPPT. Stereochemical
results and other data indicate that reactions of phosphorothioate
dianions such as pNPPT are stepwise DN + AN mechanisms in
which a thiometaphosphate intermediate forms,34 while phos-
phate esters react by concerted mechanisms.35 The loose
transition state implied by the isotope effects is shown schemati-
cally in Figure 6.
temperature reduces the magnitudes of KIEs, and the difference
18
in the magnitudes of
k
in the reactions of pNPP and
bridge
pNPPT is partly due to this effect, but this temperature difference
should only account for a small change (compare the 18knonbridge
The 18knonbridge data reported here for phosphorothioate esters
differ from previous data for phosphate esters in two respects.
values for the pNPP monoanion at 95 and at 30 °C in Table 3).
18
The value for
k
is normal, not inverse, for the loose
nonbridge
18
The larger value for
k
in the pNPPT dianion reaction
bridge
transition state of the pNPPT reaction, and it changes toward
the inverse direction as the transition state becomes tighter in
the diester reaction. This trend is logical if bending modes are
compared to the pNPP reaction implies the former reaction has
a larger degree of P-O bond fission in the transition state.
Evidently this occurs without a significant difference in the
amount of charge delocalization into the nitro group in the
transition states of the two reactions, as the values of 15k are
the same.
(31) Rawlings, J.; Hengge, A. C.; Cleland, W. W. J. Am. Chem. Soc. 1997,
119, 542-549.
(32) Bernasconi, C. F. AdV. Phys. Org. Chem. 1992, 27, 119-238.
(33) Bordwell, F. G.; W. J. Boyle, J. J. Am. Chem. Soc. 1972, 94, 3907-3911.
(34) Cullis, P. M.; Misra, R.; Wilkins, D. J. J. Chem. Soc., Chem. Commun.
1987, 1594-1596. Cullis, P. M.; Iagrossi, A. J. Am. Chem. Soc. 1986,
108, 7870-7871. Burgess, J.; Blundell, N.; Cullis, P. M.; Hubbard, C. D.;
Misra, R. J. Am. Chem. Soc. 1988, 110, 7900-7901.
(28) Basch, H.; Krauss, M.; Stevens, W. J. J. Mol. Struct. (THEOCHEM) 1991,
235, 277-291.
(29) Sigel, R. K. O.; Song, B.; Sigel, H. J. Am. Chem. Soc. 1997, 119, 744-
755.
(35) Herschlag, D.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111, 7579-7586.
Buchwald, S. L.; Friedman, J. M.; Knowles, J. R. J. Am. Chem. Soc. 1984,
106, 4911-4916.
(30) Hoff, R. H.; Mertz, P.; Rusnak, F.; Hengge, A. C. J. Am. Chem. Soc. 1999,
121, 6382-6390.
9
J. AM. CHEM. SOC. VOL. 125, NO. 25, 2003 7551

A R T I C L E S
Catrina and Hengge
18
the dominant contributor to
k
for phosphorothioate
advanced for the phosphorothioate triester, though the difference
is not large.
nonbridge
esters, in contrast to the situation with phosphates. A likely
reason for the greater importance of bending modes in phos-
phorothioates is the larger van der Waals radius of sulfur.
(B) Monoanion Hydrolysis. In reactions of the monoanions
of phosphate and phosphorothioate monoesters a proton is
transferred from the phosphoryl group to the leaving group in
the transition state. Protonation of p-nitrophenolate results in
an inverse 18O isotope effect of 0.985.36 A transition state in
which P-O bond cleavage and proton transfer are both far
advanced would be expected to show an observed 18kbridge that
Conclusions
Kinetic isotope effects in the nonbridging oxygen atoms in
phosphorothioate esters compared to phosphate esters respond
oppositely to hybridization changes at phosphorus. In phospho-
rothioates the trend is similar to that seen in R-secondary
hydrogen isotope effects on atoms undergoing hybridization
changes, in which bending modes are the predominant deter-
minants of the isotope effect: normal for changes of the type
sp3 to sp2 and inverse in the direction sp2 to sp3. Changes in π
bonding occur in reactions involving the phosphoryl group that
are absent in reactions involving hydrogen. Thus, changes in
bond order can have a greater impact in the former case, and
this is the most likely explanation for the trends observed in
R-secondary nonbridge 18O KIEs for phosphate ester reactions.
The fact that the opposite trend is seen in the phosphorothioate
reactions suggests that the larger size of sulfur causes bending
mode changes to predominate.
The EIE for deprotonation of a phosphorothioate monoester
is within experimental error of those for phosphate monoesters,
indicating that this primary EIE is determined by O-H
stretching considerations and is unaffected by bending modes.
31P NMR data reveal differences in the response of the pNPPT
to protonation compared to the response of phosphate mo-
noesters, suggestive of differences in the effect of protonation
on the atomic charges on phosphorus in the two types of
compounds. Protonation of the dianion of pNPPT is ac-
companied by an increase in π bonding between phosphorus
and sulfur, resulting in lessened P-O bond order involving the
nonbridge oxygen atoms in the monoanion compared to the
dianion, consistent with computational predictions.
No significant EIEs for metal ion complexation are observed,
indicating that in aqueous solution these interactions are weak
and do not affect bonding within the thiophosphoryl group.
The kinetic isotope effects in the pNPPT dianion reaction
suggest slightly greater P-O bond fission than in the reaction
of the corresponding phosphate ester. A difference in the
synchronization of charge development on the leaving group
arising from P-O bond fission and its delocalization was noted
in the reactions of the phosphate and phosphorothioate. The
pNPPT monoanion reaction proceeds with a late transition state
in which the proton is largely but not completely transferred
from the thiophosphoryl group to the leaving group. The KIEs
measured for a phosphorothioate diester and triester suggest that
leaving group bond fission is somewhat less advanced than in
reactions of their phosphate counterparts.
is approximately the product of 0.985 and the 1.0237 value for
18
k
observed for the dianion hydrolysis, which would be
bridge
1.008. The observed result of 1.0091 is close to this value. The
magnitude of 15k is also significantly reduced relative to the
dianion reaction, consistent with nearly full charge neutralization
of the leaving group by protonation.
The normal value of 18knonbridge primarily reflects the normal
isotope effect for deprotonation. If the thiophosphoryl group
resembles thiometaphosphate in the monoanion reaction as it
does in the dianion one, then the observed 18knonbridge should be
close to the product of the equilibrium isotope effect for
deprotonation (1.015) and 18knonbridge for the hydrolysis of pNPPT
dianion (1.0135), or 1.0287. The observed
is relatively close but smaller. The lower value may be due to
incomplete proton transfer from the thiophosphoryl group in
the transition state. The small but measurable 15k in this reaction
also indicates that proton transfer to the leaving group, while
far advanced, is not complete.
(C) Alkaline Hydrolysis of the EtOpNPPT Diester and the
(MeO)2pNPPT Triester. LFER data suggest that phospho-
rothioate diesters react by a mechanism with a tighter transition
state than monoesters. The leaving group isotope effects 18kbridge
and 15k in the diester and triester reactions are smaller than in
the hydrolysis of pNPPT2-, confirming the notion that leaving
group bond fission is less advanced. A comparison with previous
18
k
of 1.0221
nonbridge
data for phosphate diesters with the same leaving group (Table
3) reveals that the EtOpNPPT diester
range. This suggests that leaving group bond fission is less
advanced in the phosphorothioate diester reaction.
18
k
lies below this
bridge
18
The
k
obtained for the alkaline hydrolysis reaction
nonbridge
of EtOpNPPT is smaller compared with that for the monoester
pNPPT. This supports the hypothesis that bending modes are
the dominant contributor to this isotope effect in phosphorothio-
ate esters, in contrast to the situation with phosphate esters. In
a more associative transition state the bending modes will
contribute to a more inverse isotope effect, analogous to
R-deuterium KIEs in situations where carbon hybridization
changes from sp2 to sp3. In the transition state of the present
reaction, the geometry of the phosphorus atom changes from
tetrahedral to trigonal-bipyramidal. The small KIEs in the
leaving group, in conjunction with the smaller âlg data, indicate
less bond fission to the leaving group, confirming the notion
that the transition state becomes more crowded in a manner
that will compress the bending modes of the nonbridging sulfur
and oxygen atoms.
Acknowledgment. The authors thank the NIH for financial
support of this research (Grant GM47297).
Supporting Information Available: Experimental details and
the equations used for calculation of the pKa and EIEs for
pNPPT, details of procedures used in KIE experiments, and
equations used for calculating corrected KIEs (PDF). This ma-
terial is available free of charge via the Internet at http://
pubs.acs.org.
The alkaline hydrolysis of (MeO)2pNPPT triester gives values
18
for
k
and 15k that are each smaller than their phosphate
bridge
JA0340026
triester counterparts. As with the diester results, this suggests
that leaving group bond fission in the transition state is less
(36) Hengge, A. C.; Hess, R. A. J. Am. Chem. Soc. 1994, 116, 11256-11263.
9
7552 J. AM. CHEM. SOC. VOL. 125, NO. 25, 2003