Hydrolysis of phosphotriesters: Determination of transition states in parallel reactions by heavy-atom isotope effects

Article abstract of DOI:10.1021/ja011025g

The remote label method was used to measure primary and secondary ¹⁸O isotope effects in the alkaline hydrolysis of O,O-diethylphosphorylcholine iodide (DEPC) and the primary ¹⁸O effect in the alkaline hydrolysis of O,O-diethyl-m-nitrobenzyl phosphate (DEmNBP). Both the leaving group of interest (choline or m-nitrobenzyl alcohol) and ethanol can be ejected during hydrolysis due to the similarity of their pK values. The heavy-atom isotope effects were measured by isotope ratio mass spectrometry. Parallel reaction and incomplete labeling corrections were made for both systems. DEPC has a primary ¹⁸O isotope effect of 1.041 ± 0.003 and a secondary ¹⁸O isotope effect of 1.033 ± 0.002. The primary ¹⁸O isotope effect for DEmNBP was 1.052 ± 0.003. These large effects suggest a highly associative transition state in which the nucleophile approaches very close to the phosphorus atom to eject the leaving group. The large values are also indicative of a large compression, or general movement, on the reaction coordinate.

Full text of DOI:10.1021/ja011025g

9246
J. Am. Chem. Soc. 2001, 123, 9246-9253
Hydrolysis of Phosphotriesters: Determination of Transition States in
Parallel Reactions by Heavy-Atom Isotope Effects
Mark A. Anderson,† Hyunbo Shim,‡ Frank M. Raushel,‡ and W. W. Cleland*,†
Contribution from the Institute for Enzyme Research and the Department of Biochemistry,
UniVersity of WisconsinsMadison, Madison, Wisconsin 53705, and Department of Chemistry,
Texas A & M UniVersity, College Station, Texas 77843
ReceiVed April 23, 2001
Abstract: The remote label method was used to measure primary and secondary 18O isotope effects in the
alkaline hydrolysis of O,O-diethylphosphorylcholine iodide (DEPC) and the primary 18O effect in the alkaline
hydrolysis of O,O-diethyl-m-nitrobenzyl phosphate (DEmNBP). Both the leaving group of interest (choline or
m-nitrobenzyl alcohol) and ethanol can be ejected during hydrolysis due to the similarity of their pK values.
The heavy-atom isotope effects were measured by isotope ratio mass spectrometry. Parallel reaction and
incomplete labeling corrections were made for both systems. DEPC has a primary 18O isotope effect of 1.041
( 0.003 and a secondary 18O isotope effect of 1.033 ( 0.002. The primary 18O isotope effect for DEmNBP
was 1.052 ( 0.003. These large effects suggest a highly associative transition state in which the nucleophile
approaches very close to the phosphorus atom to eject the leaving group. The large values are also indicative
of a large compression, or general movement, on the reaction coordinate.
Introduction
compounds go through SN2 type mechanisms with inversion of
configuration with respect to the phosphorus.12 Indeed most, if
not all, of the studies done on monoesters have been used to
determine the reaction mechanism of phosphodiester hydrolysis
also.13 Both esters have been studied enzymatically and have
added greatly to the understanding of the chemistry that takes
place in the active site of the protein.14
The chemical hydrolysis of triesters has not been studied as
extensively as that of the less ligated phosphoesters, and the
principal reason is that there are no naturally occurring phos-
photriesters in nature. This is rather strange when one considers
that phosphotriesterase activity was discovered in the soil
microbe Pseudomonas diminuta in 1974. Where it came from
and how it came to be are still unknown, but at the time of its
discovery activity on the pesticide parathion (O,O-diethyl-O-
p-nitrophenyl phosphorothioate) was observed.15 Since then,
phosphotriesterase and the enzymatic hydrolysis of phospho-
Phosphoryl transfer is enormously important in biological
systems. With the exception of water, adenosine triphosphate
(ATP) is easily the most important chemical in the world for
animal life. Indeed, ATP supplies the energy for muscular
movement, neural activity, biosynthesis, and active transport.
As such, the chemistry of phosphomonoesters and phosphodi-
esters has been studied intensely for many years. In 1955,
Westheimer1 and Bunton2 independently proposed a metaphos-
phate intermediate in the hydrolysis of phosphomonoesters.
Several years later, Herschlag3 showed that, while there is
considerable metaphosphate-like character in the transition state,
the evidence is against formation of a true metaphosphate
intermediate in aqueous solution. For phosphomonoesters the
kinetics,1,2,4,5 pH-rate profiles,6 effect of metal ions,7 ab initio
calculations,8 stereochemistry,9 and isotope effects,10 and reac-
tivity11 have all been determined at one time or another.
Phosphodiesters, the most stable of the phosphate esters, have
also been studied extensively. It is widely accepted that these
(10) Hengge, A. C.; Edens, W. A.; Elsing, H. J. Am. Chem. Soc. 1994,
116, 5045. Knight, W. B.; Weiss, P. M.; Cleland, W. W. J. Am. Chem.
Soc. 1986, 108, 2759. Weiss, P. M.; Knight, W. B.; Cleland, W. W. J. Am.
Chem. Soc. 1986, 108, 2761. Weiss, P. M.; Cleland, W. W. J. Am. Chem.
Soc. 1989, 111, 1928.
† University of Wisconsin.
‡ Texas A & M University.
(1) Butcher, W. W.; Westheimer, F. H. J. Am. Chem. Soc. 1955, 77,
2420.
(11) Bromilow, R. H.; Kirby, A. J. J. Chem. Soc., Perkin Trans. 2 1972,
2, 149. Kirby, A. J.; Varvoglis, A. G. J. Am. Chem. Soc. 1966, 88, 1823.
Kirby, A. J.; Jencks, W. P. J. Am. Chem. Soc. 1965, 87, 3209.
(12) Hengge, A. C.; Cleland, W. W. J. Am. Chem. Soc. 1991, 113, 5835.
(13) Hengge, A. C.; Tobin, A. E.; Cleland, W. W. J. Am. Chem. Soc.
1995, 117, 5919. Hengge, A. C.; Cleland, W. W. J. Am. Chem. Soc. 1990,
112, 7421. Jones, D. R.; Lindoy, L. F.; Sargeson, A. M. J. Am. Chem. Soc.
1983, 105, 7327. Ja¨rvinen, P.; Oivanen, M.; Lo¨nnberg, H. J. Org. Chem.
1991, 30, 7444. Breslow, R.; Xu, R. J. Am. Chem. Soc. 1993, 115, 10705.
Lim, C.; Karplus, M. J. Am. Chem. Soc. 1990, 112, 5872. Dejaegere, A.;
Lim, C.; Karplus, M. J. Am. Chem. Soc. 1991, 113, 4353. Ba-Saif, S. A.;
Waring, M. A.; Williams, A. J. Chem. Soc., Perkin Trans. 2 1991, 11, 1653.
Davis, A. M.; Hall, A. D.; Williams, A. J. Am. Chem. Soc. 1988, 110,
5105. Loran, J. S.; Naylor, R. A.; Williams, A. J. Chem. Soc., Perkin Trans.
2 1977, 4, 418.
(2) Barnard, P. W. C.; Bunton, C. A.; Liewellyn, D. R.; Oldham, K. G.;
Silver, B. L.; Vernon, C. A. Chem. Ind. (London). 1955, 760.
(3) Herschlag, D.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111, 7587.
(4) Kirby, A. J.; Jencks, W. P. J. Am. Chem. Soc. 1965, 87, 3217.
(5) Westheimer, F. H. Chem. ReV. 1981, 81, 313. Todd, A. R. Proc.
Natl. Acad. Sci. U.S.A. 1959, 45, 1389. Bruice, T. C.; Benkovic, S. In
Bioorganic Mechanisms; Benjamin: New York, 1966; Vol. II, Chapter 5.
Jencks, W. P. Catalysis in Chemistry and Enzymology; McGraw-Hill: New
York, 1969; p 169. Kirby, A. J.; Warren, S. G. The Organic Chemistry of
Phosphorus; Elsevier: London, 1967; p 284.
(6) Kirby, A. J.; Varvoglis, A. G. J. Am. Chem. Soc. 1967, 89, 415.
(7) Herschlag, D.; Jencks, W. P. J. Am. Chem. Soc. 1987, 109, 4665.
(8) Rajca, A.; Rice, J. E.; Streitweiser, A., Jr.; Schaefer, H. F., III. J.
Am. Chem. Soc. 1987, 109, 4189.
(14) Hengge, A. C.; Sowa, G. A.; Wu, L.; Zhang, Z,-Y. Biochemistry
1995, 34, 13982. Jones, J. P.; Weiss, P. M.; Cleland, W. W. Biochemistry
1991, 30, 3634. Hengge, A. C.; Cleland, W. W. J. Org. Chem. 1991, 56,
1972. Culp, J. S.; Butler, L. G. Arch. Biochem. Biophys. 1986, 246, 245.
(9) Buchwald, S. L.; Friedman, J. M.; Knowles, J. R. J. Am. Chem. Soc.
1984, 106, 4911. Friedman, J. M.; Freeman, S.; Knowles, J. R. J. Am. Chem.
Soc. 1988, 110, 1268.
10.1021/ja011025g CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/31/2001

Hydrolysis of Phosphotriesters
J. Am. Chem. Soc., Vol. 123, No. 38, 2001 9247
Scheme 1a
a Only the relevant (N-containing) products are shown. Diethyl phosphate and ethanol are also products of both reactions.
triesters have been studied at length.16 Use of man-made
pesticides such as parathion and malathion (S-1,2-bis(ethoxy-
carbonyl)ethyl O,O-dimethylphosphorodithioate) is one of the
primary reasons why phosphotriesters are coming under closer
scrutiny.16 The mechanism of their chemical breakdown,17 and
the products derived from it, have become quite important from
an ecological as well as a business standpoint.
In previous studies of triester hydrolysis, we used the remote
label method to determine the isotope effects for diethyl
phosphotriesters with reasonably good leaving groups (p-
nitrophenol, pK 7.0, and p-carbamoylphenol, pK 8.6). These
compounds underwent hydrolysis at basic pH with expulsion
of the phenol group only. The transition-state structures for the
reactions were probed by measuring the 18O isotope effects for
the breaking of the P-Ophenol bond (the primary isotope effect),
which depends on the extent of bond cleavage in the transition
state, and the secondary 18O isotope effect, which is determined
by the change in bond order between the phosphorus and the
phosphoryl oxygen.18
In the present work, we have synthesized diethyl phospho-
triesters with leaving group pK values that are much larger
(choline, pK 13.9,19 and m-nitrobenzyl alcohol, pK 14.920) in
order to compare the transition state of the hydrolysis of a
substrate with a good leaving group versus a poor one. As a
consequence of the pK values being relatively close to that of
ethanol (pK 16.019), the hydrolysis of diethylphosporylcholine
iodide (DEPC) and diethyl-m-nitrobenzyl phosphate (DEmNBP)
results in the ejection of some ethanol as well as choline from
DEPC, and m-nitrobenzyl alcohol (m-NBA) from DEmNBP.
These are thus parallel reactions (Scheme 1).
Data Analysis for a Simple System Using Nitrogen
Isotope Effects
For a reaction that involves a one-step process with no side
or competing reactions, determining isotope effects in systems
containing a single N atom is straightforward by the internal
competition method.
k1
A
98 P
(1)
We will assume that a C-N bond is broken in the rate-limiting
step of the reaction and that after bond scission the portion with
the N atom will be the isolated product. No labeled syntheses
are necessary, and changes in the natural abundance ratio of
15N/14N in the compound will be sufficient to determine the
isotope effect of the reaction. The reaction is run from 10 to
70% completion, and the amount of initial substrate that reacts,
the fraction of reaction f, is determined accurately (NMR, UV-
vis, etc.). The product and residual substrate are separated,
purified, and converted to N2 gas (combustion, Kjeldahl acid
digestion/hypobromite oxidation, etc.). Each gas sample is
analyzed individually by IRMS to determine its isotopic ratio
compared to a known standard (we are using N as an example
here, but the equation is valid for any element):
δ ) 1000[(15Nsample/14Nsample)/(15Nstandard/14Nstandard) - 1]
(15) Caldwell, S. R.; Newcomb, J. R.; Schlecht, K. A.; Raushel, F. M.
Biochemistry 1991, 30, 7438. Munnecke, D. M.; Hsieh, D. P. H. Appl.
Microbiol. 1974, 28, 212.
(2)
(16) Raushel, F. M.; Holden, H. M. AdV. Enzymol. 2000, 74, 51 and
A change of 1δ is equal to a change of 0.1% of the 15N/14N
ratio. To determine the isotope effect, the samples are converted
to an R value defined as
references therein.
(17) Bromilow, R. H.; Khan, S. A.; Kirby, A. J. J. Chem. Soc., Perkin
Trans. 2 1972, 7, 911. Kirby, A. J.; Bromilow, R. H.; Khan, S. A. J. Chem.
Soc. B 1971, 6, 1091.
(18) Caldwell, S. R.; Raushel, F. M.; Weiss P. M.; Cleland, W. W.
Biochemistry 1991, 30, 7444.
R ) (δsample/1000) + 1
(3)
(19) Jencks, W. P.; Regenstein, J. Ionization Constants of Acids and
Bases. Handbook of Biochemistry and Molecular Biology; CRC Press:
Cleveland, OH, 1976; p 315.
(20) Sowa, G. A.; Hengge, A. C.; Cleland, W. W. J. Am. Chem. Soc.
1997, 119, 2319.
R values for the reaction product, Rp, and residual substrate,
Rs, are used in the following equations to arrive at the isotope
effect (IE).

9248 J. Am. Chem. Soc., Vol. 123, No. 38, 2001
IEprod ) ln(1 - f)/ln(1 - fRp/Ro)
Anderson et al.
y is the the fraction of heavy discriminating label in doubly
labeled material, and Q is the degree to which light material in
the remote labeled mixture is depleted below natural abundance,
defined as
(4)
(5)
IEsub ) ln(1 - f)/ln[(1 - f)(Rs/Ro)]
In these equations, Ro is the isotopic ratio of the starting material
(found by direct combustion or complete reaction of the starting
material). The isotope effect can also be found by using a
comparison of the product and residual substrate R values by
eq 6.
Q ) (1 - b)z/bx
(9)
where b is the the fraction of doubly labeled material in the
mixture (determined by IRMS), z is the the fraction of heavy
label present in the remote label position of light material, and
x is the the fraction of heavy label in the remote label position
of the doubly labeled material.
IEsub/prod ) ln(1 - f)/ln{(1 - f)/[1 - f + (fRp/Rs)]} (6)
In a simple system of this type, the apparent isotope effects
for Rs and Rp are determined by using IRMS and then corrected
for incomplete isotopic substitution by using eqs 7-9. Hydroly-
sis reactions of labeled material leading to the extrusion of only
the remote labeled group are good examples of this process.
Reactions such as these are the hydrolysis of phosphomonoesters
(monoanionic and dianionic), where the system has only one
leaving group, and phosphotriesters, where the leaving group
of interest has a pK value much lower than those of the other
two ester substituents.
If the starting material is pure and the separation process is good,
the isotope effect should be easy to determine and should be
quite reproducible when varying the fraction of reaction.
Remote Label Method
The remote label method is used to measure primary and
secondary isotope effects in systems where the atom that is
actually responsible for the effect is extremely difficult to isolate
and study. Oxygen is one such atom because in CO2 the O of
interest exchanges with the O of water. Even a slight water
contamination can lead to 18O washout and useless isotopic
results. Since the approach to studying hydrolysis reactions of
phosphotriesters is to measure primary 18O isotope effects in
the leaving groups and secondary isotope effects in the
nonbridging phosphoryl oxygen, the easiest way to follow these
reactions is with a remote label.21 Remote labeling allows the
determination of isotope effects by using a substrate labeled in
two positions, the site of chemical interest and another position
which can be isolated and measured isotopically. In these
Parallel Reactions
When the rates of hydrolysis for two different groups of a
phosphotriester are similar, the system is known as a parallel
reaction. The schemes for the hydrolysis of the diethylphos-
phorylcholine (DEPC) and diethyl m-nitrobenzyl phosphate
(DEmNBP) point out the dual nature of these reactions. In both
cases the labeled substituent is not the only possible leaving
group. Here, the pK’s of the possible leaving group substituents
are relatively close to one another (ethanol 16.0,19 choline 13.9,19
m-nitrobenzyl alcohol 14.920), and ethanol can be extruded as
well as choline in the case of DEPC and m-nitrobenzyl alcohol
in the case of DEmNBP. Isotope effects for such a cleavage
may be as isotopically sensitive as for the ligand of interest
(choline and m-NBA).23 Here, the kinetic scheme becomes more
complex, as shown below using the hydrolysis of DEPC as an
example.
studies, the leaving group of the triester is synthesized with 18
O
at the point of bond breaking and 15N at the remote site on the
molecule. A larger amount of the compound is synthesized with
16O at the point of bond breaking22 and 14N (i.e., nitrogen from
which the 0.37% natural abundance of 15N has been removed)
at the remote site. The 18O,15N molecule is mixed with the 14
N
one in the natural abundance ratio. The experimental isotope
effect with this material is the product of the 18O primary isotope
effect and any effect from 15N in the remote position. The 15
N
k1
98 P
effect alone is determined by using natural abundance substrate,
and the ratio of the two isotope effects gives the 18O one.
Determination of the secondary 18O isotope effect is done by
A
P is ethyl choline phosphodiester after the release of ethanol
upon hydrolysis of DEPC.
the same method, using doubly labeled substrate which has 18
O
at the phosphoryl O and an 15N label in the remote position.
If the syntheses yielded isotopically pure compounds, the
determination of the isotope effects would occur as described
above. However, syntheses almost never result in completely
labeled compounds. Therefore, the observed isotope effects must
be corrected for the amount of label in each position by using
the following equations:
k2
98 Q
A
Q is choline released upon hydrolysis of DEPC.
k1/xk1
Ax
8 Px
IE per atom of label ) {(T1/i - 1)/[1 - T1/i(1 - y)/i]} + 1
The rate constant for the heavy labeled compound for the release
of ethanol upon hydrolysis of DEPC is reduced by k1.
x
(7)
k2/xk2
where T is defined as
Ax
8 Qx
T ) (P/R)/[1 - Q(P/R - 1)]
(8)
The rate constant for the heavy labeled compound for the release
x
of choline upon hydrolysis of DEPC is reduced by k2 (x )
and P is the the isotope effect with remote labeled substrate, R
is the isotope effect in the remote label position with natural
abundance material, i is the the number of discriminating atoms,
double 18O,15N or single 15N label). The definitions of the R
values for the hydrolysis of DEPC are more complex than for
a simple system.
(21) O’Leary, M. H.; Marlier, J. F. J. Am. Chem. Soc. 1979, 101, 3300.
(22) Oxygen containing the natural abundance of 17O and 18O. The small
amounts of the heavier atoms have no effect on the measured isotope effects.
(23) Rawlings, J.; Hengge, A. C.; Cleland, W. W. J. Am. Chem. Soc.
1997, 119, 542.

Hydrolysis of Phosphotriesters
J. Am. Chem. Soc., Vol. 123, No. 38, 2001 9249
kapp
18,15k2(k1 + k2)/(k2 + k118,15k2)
Rp ) Qx/Q
(the 15N/14N ratio in the extruded choline at a known
)
and
fraction of reaction, f)
18,15k2 ) kappk2/(k1 + k2 - k1kapp
)
(12)
Rs ) (Qx∞ - Qx)/(Q∞ - Q)
Results
(the 15N/14N ratio in unreacted DEPC at a known
Synthesis of Labeled and Depleted Phosphotriesters. The
synthetic routes for the 18O,15N-labeled and the 14N phospho-
triesters are shown in Schemes 2-4, and the details of the
fraction of reaction, f)
R∞ ) (Qx∞/Q∞)
1
syntheses are given in Supporting Information. The H, 13C,
(the 15N/14N ratio in choline after complete hydrolysis
and 31P NMR spectra of the three separate labeled DEPC
compounds and the two labeled DEmNBP compounds matched
those of natural abundance counterparts.
of the triester)
Hydrolysis of Phosphotriesters in 18O/16O H2O. Although
most evidence indicates that the nucleophilic attack of the
hydroxyl takes place at the P center of a phosphotriester,18,24,25
it is not impossible that some could be taking place at the
methylenic carbons attached to the bridging oxygens of the
system.
Rp∞ ) Px∞/P∞
(the 15N/14N ratio in the phosphodiester after complete
hydrolysis of the triester)
The integrated equations describing the concentrations of the
various molecules at a known fraction of reaction, or at infinite
time are
Since isotope effect studies have not been done on the
hydrolysis of compounds with high pK values, it is important
to verify that the mechanism of attack is the same as for
compounds with good leaving groups. The hydrolysis of both
triesters was followed by 13C NMR in an alkaline solution
consisting of H218O/H216O (30/70). Attack at the -CH2- group
by 18OH- would result in a ∼0.04 ppm upfield shift for that
carbon in the 13C NMR spectrum of a product.
A ) Ao e-(k +k )t
Ax ) Axo e-k′t
1
2
where
k ′ ) k1/xk1 + k2/xk2
When DEPC and DEmNBP were subjected to these condi-
tions and the 13C spectrum of the H218O/H216O reaction mixture
was compared to that in 100% H216O, no difference was seen
in the chemical shifts of interest in either the released ethanol,
choline, or m-nitrobenzyl alcohol (DEPC 66.29, ethyl CH2 of
ethanol; 66.86, cholyl CH2 of the choline; DEmNBP 65.79, ethyl
CH2 of ethanol; 68.88, CH2 of the m-nitrobenzyl alcohol).
31P NMR was then used to determine the whereabouts of the
18O atom. High-resolution NMR of the hydrolyzed samples
clearly showed the upfield shift of the P-18O in the diester
products (DECP P-16O diethyl phosphate, 0.62 ppm; P-18O
diethyl phosphate, 0.59 ppm; P-16O ethyl choline phosphate,
-0.27 ppm; P-18O ethyl choline phosphate, -0.30 ppm;
DEmNBP P-16O diethyl phosphate, 0.48 ppm; P-18O diethyl
phosphate, 0.45 ppm; P-16O ethyl m-nitrobenzyl phosphate,
0.22 ppm; P-18O ethyl m-nitrobenzyl phosphate, 0.19 ppm).
When the overlapping peaks were deconvoluted and the areas
were calculated, the ratio of the P-16O peaks to the P-18O
peaks was 69%/31% ((3%). The lack of change in any of the
13C NMR shifts coupled with the 0.03 ppm upfield shifts of
the P-18O diesters and the correct 30/70 ratio of the P-18O/
P-16O areas in the 31P NMR shows conclusively that the
hydrolysis of the two systems involves attack of the hydroxyl
group at the phosphorus center.
Isotope Effects. The primary and secondary effects for the
alkaline hydrolysis of DEPC and the primary effect for the
alkaline hydrolysis of DEmNBP were determined by using the
remote label method.21 The observed isotope effects were
determined from comparative isotopic analysis of the residual
substrate versus the products of the hydrolysis (choline and ethyl
choline phosphodiester, or m-nitrobenzyl alcohol and ethyl-m-
nitrobenzyl phosphodiester), which were collected and com-
busted as one product. The hydrolyses were run to 30-70%
completion (f values of 0.3-0.7). The observed 18O isotope
Q ) k2Ao(1 - e-(k +k )t)/(k1 + k2)
Qx ) k2Axo(1 - e-k′t)/xk2k ′
1
2
P ) k1Ao(1 - e-(k +k )t)/(k1 + k2)
Px ) k1Axo(1 - e-k′t)/xk1k′
1
2
Using the experimentally determined isotope ratios, an apparent
isotope effect can be calculated starting with the following
equations:
xkapp ) ln(1 - f)/ln(1 - fRp/R∞)
) ln(1 - f)/ln[(1 - f)(Rs/R∞)]
(10)
(11)
xkapp ) xk2(1 + k1/k2)/(1 + k1 k2/(k2 k1))
x
x
When the ratio of k1/k2 is known, the above isotope effect can
be used along with the ratio of xk2/xk1 to determine xk1 and xk2.
x
The ratio of the isotope effects k2/xk1 is equal to Rp∞/R∞.
This methodology was used by Rawlings to determine heavy-
atom isotope effects on reaction of Co(III)-bound p-nitrophenyl
phosphate.23 However, for the DEPC and DEmNBP systems,
the apparent R values for the ethyl choline phosphodiester and
ethyl m-nitrobenzyl phosphodiesters could not be accurately
determined since the samples were too small. Instead, an
assumption was made that the value of the isotope effect for
the loss of ethanol, 18,15k1, was equal to unity. This assumption
is probably valid, since no appreciable bonding changes occur
in the P-Ocholine bond when ethanol is the leaving group. The
apparent isotope effect (from the IRMS) is equal to
kapp
and if it is assumed that 18,15k1 ) 1.000, then
)
18,15k2(k1 + k2)/[k2 + k1(18,15k2/18,15k1)]
(24) Caldwell, S. R.; Raushel, F. M. J. Am. Chem. Soc. 1991, 113, 730.
(25) Cleland, W. W.; Hengge, A. C. FASEB J. 1995, 9, 1585.

9250 J. Am. Chem. Soc., Vol. 123, No. 38, 2001
Anderson et al.
Scheme 2
Using the same assumptions for the diethyl-p-carbamoyl-
phenyl phosphate (secondary isotope effect 2.5%), the bond
order of PdO is ∼1.375, and the sum of the axial bond orders
becomes ∼1.625. The sum of the axial bond orders has increased
by 40% when the leaving group pK value is increased by 1.6
units.
When the estimates are done for DEPC and DEmNBP
(assuming the same secondary isotope effect for both com-
pounds), the PdO bond order is 1.175 and the sum of the axial
bond orders is 1.825, an increase of 12% from the diethyl-p-
carbamoylphenyl phosphate. This does not seem like a large
increase in the axial bond order for a difference in pK value of
5.6 units. However, it may be that the axial bond order is as
large as possible in these systems. It may only be possible to
see a phosphoryl bond order change from 2 to 1 in a cyclic
phosphotriester when a true phosphorane intermediate is formed
prior to pseudorotation.25,28 Even though phosphorane interme-
diates have been hypothesized in acyclic phosphotriester hy-
drolysis, the evidence has shown that, for acyclic phosphotri-
esters at least, these mechanisms are in-line displacements
without phosphorane intermediates.25
With high pK leaving groups, the transition state presumably
is approximately symmetrical, and the attacking nucleophile
must get in much closer to the phosphorus center to eject the
leaving group. The Hammond postulate supports this conclusion,
since the incoming nucleophile and the leaving group have
roughly the same pK values (14.9-16.0), so that the transition
state will be symmetrical, or at least much more symmetrical
than in the diethyl-p-nitrophenyl phosphate or the diethyl-p-
carbamoylphenyl phosphate systems.29 The secondary isotope
effect values show a noticeable change in the bond order of the
phosphoryl group which can only be attributed to the hydroxyl
group moving in very close to the phosphorus center to dislodge
the leaving group. The fact that the leaving groups are unable
to stabilize themselves as ions (lack of resonance forms) or to
be protonated in the transition state suggests that the nucleophile
would need to move in much closer than van der Waals contact
distance in order to displace the leaving group. A comparison
of hypothesized energy diagrams for diethyl phosphotriesters
with either p-nitrophenol or m-nitrobenzyl alcohol as leaving
groups is shown in Figure 1. In the hydrolysis of diethyl-p-
nitrophenyl phosphate, the ability of the leaving group to
stabilize itself allows the phenol to move away from the
phosphorus early in the reaction due to a slight association with
the nucleophile. Conversely, in the DEmNBP reaction, the
nucleophile and leaving group are both closer to the phosphorus
center, and the transition state is nearly symmetrical and centered
on the reaction coordinate.
effects were calculated by using eqs 4-6 and were corrected
for incomplete label incorporation by using eqs 7 and 8. The
correction for parallel reactions was calculated assuming 18,15k1
) 1 and using eq 12, where k1/k2 was the ratio of diester to
choline or m-nitrobenzyl alcohol.
The corrected primary isotope effects for DEPC and DEm-
NBP were 1.041 ( 0.003 and 1.052 ( 0.002, respectively. The
secondary isotope effect for DEPC was 1.033 ( 0.002. The
secondary isotope effect for DEmNBP was not determined. The
isotope effects for these compounds are compared to those for
phosphotriesters with good leaving groups in Table 1.
Discussion
It is known that phosphotriesters undergo hydrolysis with
associative transition states26 and considerable shortening of the
reaction coordinate.25 This may seem intuitive until it is realized
that, in some reactions with dissociative transition states, the
reaction coordinates are also compressed. An example is hy-
drolysis of phosphomonoesters, where the reactants start at the
van der Waals contact distance and the entering and leaving
groups do not move during the reaction. If the reaction coor-
dinate is slightly shortened and the phosphoryl oxygen remains
motionless (as possibly occurs in enzyme mechanisms), the axial
bond order will be ∼0.15, illustrating reaction coordinate
compression in a dissociative mechanism.25
In the case of a triester with a low pK value leaving group,
the axial bond order is much higher. In diethyl-p-nitrophenyl
phosphate, the pK’s of the possible leaving group groups are
7.0 (p-nitrophenol) and 16.0 (ethanol). Here, a parallel reaction
is not a possibility, and the hydrolysis will lead to the ejection
of only the p-nitrophenol due to the wide discrepancy in the
pK values. By assuming that the total bond order of the P
remains at 5 and that the bond order for the EtO- groups
remains at 1, the secondary isotope effect can be used to estimate
the sum of the axial bond orders. The calculated isotope effect
for reducing the phosphoryl bond order from 2 to 1 is 4%.27
The secondary isotope effect of 0.63% for the nitrophenolic
triester gives a bond order of 1.85 for the PdO component.
The total axial bond order sum can then be estimated to be
∼1.15 (5 - 1.85 - 2). While much larger than for phospho-
monoesters, this still translates to only a slightly associative
transition state.25
The size of the primary isotope effects in these systems also
points to much more movement on the reaction coordinate than
in the phenolic systems. For phosphomonoester hydrolysis with
dissociative transition states, the major primary isotope effect
comes from the zero point energy difference between ground
and transition states. With the low (8-15%) bond order along
the reaction path between phosphorus and entering and leaving
groups, this accounts for up to 2% of the observed isotope effect.
By contrast, the imaginary frequency factor is small because
the coupling of the two stretches in off-diagonal position of
the force field matrix is small because of the low bond orders.
(26) Benkovic, S. J.; Schray, K. J. The Mechanism of Phosphoryl
Transfer. In Transition States of Biochemical Processes; Candour, R. D.,
Schowen, R. L., Eds.; Plenum Press: New York, 1978; p 493.
(27) Unpublished calculations by W. W. Cleland, using the BEBOVIB
program and force constants that correctly reproduce the Raman frequencies
for PO4 (Weiss, P. M.; Knight, W. B.; Cleland, W. W. J. Am. Chem.
Soc. 1986, 108, 2761). This value may not be the same in the trigonal
bipyramidal transition-state structure of a phosphotriester hydrolysis as in
the simple tetrahedral phosphate ion.
(28) Mislow, K. Acc. Chem. Res. 1970, 3, 321. Musher, J. I. J. Am. Chem.
Soc. 1972, 94, 5662. Westheimer, F. H. Acc. Chem. Res. 1968, 1, 70. Ugi,
I.; Marquarding, D.; Klusacek, H.; Gillespie, P.; Ramirez, F. Acc. Chem.
Res. 1971, 4, 288.
(29) Hammond, G. S. J. Am. Chem. Soc. 1955, 77, 334. Farcasiu, J.
Chem. Educ. 1975, 52, 76.
3-

Hydrolysis of Phosphotriesters
J. Am. Chem. Soc., Vol. 123, No. 38, 2001 9251
Scheme 3a
a The synthesis of 6, 15N,(phosphoryl)18O diethyl choline phosphotriester iodide, was done in the same manner except 2 was hydrolyzed by
excess water with natural oxygen isotope abundance. Also diethylchlorophosphine was reacted with H218O to produce the labeled diethylchloro-
phosphate used in the reaction with N,N-dimethylaminoethanol.31
Scheme 4a
a The syntheses of 14N m-nitrobenzaldehyde (7) and 14N m-nitrobenzyl alcohol (8) were done with 14NH414NO3 using the same methodology.
Table 1. Kinetic Isotope Effects on the Alkaline Hydrolysis of Diethyl Phosphotriesters
leaving group
pK of leaving group
primary 18O IE %
secondary 18O IE %
transition state
ref
p-nitrophenola
7.0
8.6
13.9
14.9
0.6
0.63 ( 0.01
2.5 ( 0.2
slightly associative
associative
highly associative
highly associative
18
18
this work
this work
p-carbamoylphenola
choline iodideb
2.7 ( 0.2
4.1 ( 0.3c
5.2 ( 0.3c
3.3 ( 0.2c
m-nitrobenzyl alcoholb
a The phenol is the only leaving group. b The leaving groups are accompanied by some ethanol in ratios of 3/1 for choline and 6.6/1 for m-nitrobenzyl
alcohol. c Corrected for incomplete isotopic labeling in parallel reactions; values are the average of six determinations with f values between 0.3
and 0.7.
With phosphotriester hydrolysis with an associative transition
state, there is less difference in zero point energies of ground
and transition states because of the higher bond orders along
the reaction path. Thus, this part of the isotope effect is smaller
than with monoesters. The imaginary frequency factor, however,
is much larger because of the tight off-diagonal coupling of
the stretches resulting from the higher bond order.
We have data for primary isotope effects with p-nitrophenol
(pK 7),18 p-carbamoylphenol (pK 8.6),18 and for choline (pK
13.9)30 and m-nitrobenzyl alcohol (pK 14.9).30 The imaginary
frequency factor increases with the associative character and
(30) This work.
(31) Reynolds, M. A.; Gerlt, J. A.; Demou, P. C.; Oppenheimer, N. J.;
Kenyon, G. L. J. Am. Chem. Soc. 1983, 105, 6475.

9252 J. Am. Chem. Soc., Vol. 123, No. 38, 2001
Anderson et al.
0.12). Deconvolution of the peaks and area integration were done with
Bruker software.
Procedure for the Hydrolysis, Isolation of Products, and Deter-
mination of 15N/14N by Isotope Ratio Mass Spectrometry of
Diethylphosphorylcholine Iodide (DEPC). 15N,18O DEPC and 14N-
labeled DEPC were mixed to natural abundance (0.37% by IRMS),
and this solution was taken to dryness by rotary evaporation. The dry
powder was suspended in ∼50 mL of anhydrous methyl acetate. To
this was added freshly distilled CH3CN until the solid just dissolved.
The solution was refrigerated at -10 °C for 24 h and then the solvent
was removed via cannula and N2 pressure. Cold anhydrous methyl
acetate (25 mL) was added to the flask via cannula to wash the
crystalline product. This wash solution was removed, fresh anhydrous
methyl acetate and CH3CN were added, and the recrystallization process
was repeated. The crystals were washed with three 50-mL aliquots of
cold methyl acetate and were then dried under vacuum for several hours.
NMR studies showed no discernible impurities.
The DEPC was dissolved in 10 mL of water. Aliquots of 1 mL (∼65
mg of triester) were placed in separate 3-mL vials. To the 1-mL
solutions of aqueous DEPC was added a 70% molar equivalent of KOH.
The solution was stirred for 24 h, and the reaction was quenched to
pH 7 using 0.1 N HCl. Percent completion (f value) and product
structure were determined by 31P and 1H NMR. The sample was taken
to dryness by rotary evaporation, and ∼5 mL of freshly distilled
CH3CN was added. The mixture was allowed to stir for 1 h and then
filtered through a 2-mL medium glass frit to remove the KCl, and the
solvent was removed by rotary evaporation. The sample was dis-
solved in ∼2 mL of the eluent used in the C18 chromatography (45
mM KxHxPO4 {39/1, KH2PO4/K2HPO4}, 27.5% MeOH, 1.5% THF,
pH 6.0). The sample solution (∼250 µL, 12-15 mg hydrolysis
products) was eluted at 9.9 mL min-1 from a Microsorb preparative
C18 column (2.2 cm i.d. × 25 cm). The individual product peaks (ethyl
choline phosphate and choline) were collected together between 6 and
8 min, and the residual substrate was collected between 14 and 20 min.
The individual samples were reduced to dryness and then dissolved in
a minimum of H2O (8-10 mL). A large excess of freshly distilled
acetone (100 mL) was added to the filtrate slowly with stirring. A white
precipitate (KPi buffer salt) was formed immediately and was removed
by suction filtration through a medium glass frit. The solution was rotary
evaporated to dryness, leaving a small amount of precipitate, and the
solid was stirred with 25 mL of freshly distilled MeOH. The MeOH
solution was suction-filtered through a medium glass frit, and the filtrate
was taken to dryness by rotary evaporation. The products were dissolved
in freshly distilled methanol and transferred to quartz tubes. Aspirator
vacuum was used initially to remove the majority of the methanol,
and then the samples were evacuated at ∼5 × 10-3 Torr to remove
any trace of volatile compounds. When the samples were dry, CuO
(3-4 g), Cu (∼500 mg), ditomaceous earth (∼500 mg), and silver
(∼200 mg) were added, and the tubes were placed under high vacuum
(<1 × 10-3 Torr) and flame-sealed. The samples were combusted at
775 °C to convert all N products to N2 gas. The samples were distilled
on a high-vacuum line, and the N2 was trapped on molecular sieves at
-196 °C. The isotope mass ratio was determined on a Finnegan isotope
ratio mass spectrometer.
Procedure for the Hydrolysis, Isolation of Products, and Deter-
mination of 15N/14N by Isotope Ratio Mass Spectrometry of Diethyl
m-Nitrobenzyl Phosphate (DEmNBP). 15N,18O DEmNBP and 14N
DEmNBP were mixed to natural abundance (0.37%). The triester was
cleaned on Florosil (4:1 ethyl acetate:petroleum ether, 4 cm × 35 cm).
DEmNBP was dissolved in aqueous methanol (90)/10). Three 1-mL
aliquots (∼65 mg/mL) were placed in vials with 1 mL of 1.005 M
KOH (10% MeOH/90% H2O), and the solutions were stirred from 8
to 48 h. The reactions were quenched to pH 7 using 1 N HCl. Percent
completion (f value) and product structure were determined from 31P
and 1H NMR. The sample was taken to dryness by rotary evaporation,
and ∼5 mL of freshly distilled CH3CN was added. The mixture was
allowed to stir for 1 h and then filtered through a 2-mL medium glass
frit to remove the KCl, and the solvent was removed by rotary
evaporation. The sample was dissolved in ∼2 mL of the eluent used
in the C18 chromatography (25% i-PrOH, 1.5% THF, 73.5% H2O) and
then filtered through a 0.22 µm syringe filter. The sample solution was
Figure 1. Reaction coordinates and transition-state structures for
alkaline hydrolysis of phosphotriesters with p-nitrophenol or m-nitro-
benzyl alcohol as leaving groups.
reaches its highest value with the high pK leaving groups. The
zero point energy difference goes the other way, being maximum
for lower pK’s. p-Nitrophenol is an exception, since electron
delocalization into the nitro group in the transition state makes
the phenolic oxygen-ring bond have some double bond
character.
Thus, the 4% primary O-18 isotope effects we see come from
the high imaginary frequency factor, rather than from a zero
point energy difference.
Conclusions
The hydrolysis of phosphotriesters with high pK value leaving
groups involves parallel reactions. The large primary and sec-
ondary 18O isotope effects suggest a very associative transition
state, and the Hammond postulate suggests that it is fairly
symmetrical as well. As the reaction progresses, the nucleophile
moves in very close to the P center, and the leaving group is
ejected with some difficulty. The large size of the primary
isotope effects indicates large compression, or general move-
ment, along the reaction coordinate.
Experimental Section
Materials. Potassium phthalimide (15N, 98%+) was obtained from
Cambridge Isotope Laboratory. H218O was obtained from Aldrich (95
atom % 18O) and Isotec (97 atom % 18O). 15N-depleted ammonium
nitrate was purchased from Monsanto. 15N-enriched ammonium nitrate
was purchased from Isotec (99.7 atom % 15N). All of the other chemicals
and solvents were commercially available and used without further
purification, unless otherwise stated. The syntheses of labeled com-
pounds are shown in Schemes 2-4, and the details are in Supporting
Information.
18O Incorporation in the Hydrolysis of Phosphotriesters. 13C
NMR and 31P NMR Experiments. The NMR samples were prepared
in the same manner and at the same time. In a small vial, 100 mg of
DEPC or 100 mg of DEmNBP was dissolved in 500 µL of CD3OD.
To this were added 100 µL of 10 M KOH and 50 µL of either H216O
or H218O, making the samples 1.5 M in KOH. Four samples were made
with each compound, having one sample of 100% H216O, and the other
70% H216O and 30% H218O. The reaction was allowed to proceed for
1 week, and then the samples were studied by using 13C NMR on a
Bruker 250 MHz instrument.
The high-resolution 31P experiments (162 MHz) were run on a Bruker
400 MHz NMR at ambient temperatures. The spectral width was 1615
Hz, acquisition time 2.53 s, 90° pulse; 16K data points; resolution 0.197
1
Hz/point; H decoupled; Gaussian multiplication (LB, 0.08 Hz; GB,

Hydrolysis of Phosphotriesters
J. Am. Chem. Soc., Vol. 123, No. 38, 2001 9253
eluted at 9.9 mL min-1 from a Microsorb preparative C18 column (2.2
cm i.d. × 25 cm). The m-NBA and ethyl m-nitrobenzyl phosphate peaks
were collected together at 11-14 and 19-22 min, respectively, and
residual DEmNBP at 26-33 min. The individual samples were reduced
to dryness and then dissolved in 12 mL of freshly distilled MeOH and
transferred to 25-mL conical flasks. The volume was reduced to ∼2
mL by rotary evaporation, and the samples were transferred to quartz
tubes. The solvent was removed by aspirator vacuum, and the samples
were then evacuated at ∼5 × 10-3 Torr for several hours to remove
the rest of the MeOH. When the samples were dry, CuO (3-4 g), Cu
(∼500 mg), and diatomaceous earth (∼500 mg) were added, and the
tubes were placed under high vacuum (∼1 × 10-3 Torr) and flame-
sealed. The samples were combusted at 775 °C to convert all N products
to N2 gas. The samples were distilled on a high-vacuum line and trapped
on molecular sieves at -196 °C. The isotope mass ratio was determined
on a Finnegan isotope ratio mass spectrometer.
study made use of the National Magnetic Research Facility at
Madison, which is supported by NIH Grant RR02301 from the
Biomedical Research Technology Program, National Center for
Research Resources. Equipment in the facility was purchased
with funds from the University of Wisconsin, the NSF Biologi-
cal Instrumentation Program (Grant DMB-8415048), NIH
Biological Instrumentation Program (Grant RR02301), NIH
Shared Instrumentation Program (Grant RR02781), and the U.S.
Department of Agriculture. The authors express their apprecia-
tion to Professor Alvin C. Hengge of Utah State University for
his help in the synthesis of the labeled m-nitrobenzyl alcohols
and his helpful discussions on parallel reactions.
Supporting Information Available: Syntheses and purifica-
tion of labeled compounds (PDF). This information is available
free of charge via the Internet at http://pubs.acs.org.
Acknowledgment. This work was supported by NIH Grants
GM18938 to W.W.C. and GM33894 to F.M.R. and by Grant
A-840 from the Robert A. Welch Foundation to F.M.R. This
JA011025G