A concerted mechanism for the transfer of the thiophosphinoyl group from aryl dimethylphosphinothioate esters to oxyanionic nucleophiles in aqueous solution

Article abstract of DOI:10.1021/ja0501565

Earlier work on the hydrolysis of aryl phosphinothioate esters has led to contradictory mechanistic conclusions. To resolve this mechanistic ambiguity, we have measured linear free energy relationships (β_nuc and β_Ig) and kinetic isot

Full text of DOI:10.1021/ja0501565

Published on Web 05/06/2005
A Concerted Mechanism for the Transfer of the
Thiophosphinoyl Group from Aryl Dimethylphosphinothioate
Esters to Oxyanionic Nucleophiles in Aqueous Solution
Ikenna Onyido,*,† Krzysztof Swierczek,‡,§ Jamie Purcell,‡ and Alvan C. Hengge*,‡
Contribution from the Department of Chemistry and Center for Agrochemical Technology,
UniVersity of Agriculture, Makurdi, Nigeria, and Department of Chemistry and Biochemistry,
Utah State UniVersity, Logan, Utah 84322-0300
Received January 10, 2005; E-mail: hengge@cc.usu.edu; ikennaonyido@yahoo.com
Abstract: Earlier work on the hydrolysis of aryl phosphinothioate esters has led to contradictory mechanistic
conclusions. To resolve this mechanistic ambiguity, we have measured linear free energy relationships
(ânuc and âlg) and kinetic isotope effects for the reactions of oxyanions with aryl dimethylphosphinothioates.
For the attack of nucleophiles on 4-nitrophenyl dimethylphosphinothioate, ânuc ) 0.47 ( 0.05 for phenoxide
nucleophiles (pKa < 11) and ânuc ) 0.08 ( 0.01 for hydroxide and alkoxide nucleophiles (pKa g 11). Linearity
of the plot in the range that straddles the pKa of the leaving group (4-nitrophenoxide, pKa 7.14) is indicative
of a concerted mechanism. The much lower value of ânuc for the more basic nucleophiles reveals the
importance of a desolvation step prior to rate-limiting nucleophilic attack. The reactions of a series of
substituted aryl dimethylphosphinothioate esters give the same value of âlg with the nucleophiles HO- (â
) -0.54 ( 0.03) and PhO- (â ) -0.52 ( 0.09). A significantly better Hammett correlation is obtained with
σ
- than with σ or σ°, as expected for a transition state involving rate-limiting cleavage of the P-OAr bond.
The 18O KIE at the position of bond fission (18k ) 1.0124 ( 0.0008) indicates the P-O bond is ∼40%
broken, and the 15N KIE in the leaving group (15k ) 1.0009 ( 0.0003) reveals the nucleofuge carries about
a third of a negative charge in the transition state. Thus, both the LFER and KIE data are consistent with
a concerted reaction and disfavor a stepwise mechanism.
Introduction
σ° substituent constant is based on phenyl acetic acid and its
substituted derivatives, originally devised by Taft, to correlate
Phosphoryl transfer reactions are widespread in biological
systems where they are involved in energy transport, transmis-
sion of genetic information, and a host of other essential
regulatory processes at the cellular level. As a consequence of
their ubiquity, an understanding of the mechanism of biological
phosphoryl transfer has been an important preoccupation in
chemical biology. Although phosphinate esters do not occur
naturally, they are structurally related to phosphate and phos-
phonate esters that function as pesticides, neurotoxins, and other
biologically active substances.1 A number of appropriately
substituted phosphinic acids have found use as metalloprotease
inhibitors.2 Investigations of the reactions of phosphinates with
a variety of nucleophiles have been undertaken to characterize
the influence of structure on the reactivity of organophosphorus
esters, as part of the effort to understand the detailed mechanism
of uncatalyzed phosphoryl transfer.3
purely inductive effects. The σ and σ- substituent constants are
the more familiar Hammett constants based on the ionization
of benzoic acids.) This result is consistent with an associative,
addition-elimination (AN + DN) mechanism in which the
formation of a pentacoordinate (phosphorane) intermediate6 is
rate determining (see Pathway A, Scheme 1). On the other hand,
data from these studies4,5 also provide Brønsted-type correlations
with âlg values in the range of -0.30 to -0.50 in water and
50% water-50% ethanol solvent. The magnitude of the âlg
values could be interpreted to indicate little-to-moderate separa-
tion of the leaving group in the TS of a concerted (ANDN)
reaction7 (Pathway B, Scheme 1). Thus, a mechanistic ambiguity
exists for these reactions. The discrepancy noted above may be
due in part to differences in the mixed solvent systems used in
some of these studies and, in other cases, to limited numbers
of substituents used in the construction of the LFER.
Earlier reports4,5 concluded that the rates of hydrolysis of aryl
dimethylphosphinothioate esters and related substrates correlate
well with σ and σ° substituent constants rather than σ-. (The
(4) Eliseeva, G. D.; Istomin, B. I.; Kalabina, A. V. J. Gen. Chem., USSR 1979,
1684-1685; Istomin, B. I.; Eliseeva, G. D.; Kalabina, A. V. J. Gen. Chem.,
USSR 1979, 49, 966-970.
(5) Istomin, B. I.; Eliseeva, G. D. J. Gen. Chem., USSR 1981, 51, 2063-
2075.
† University of Agriculture.
‡ Utah State Univeristy.
(6) Hengge, A. C. Transfer of the PO32- Group. 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.
§ Current address: NPS Pharmaceuticals, Salt Lake City, UT.
(1) Toy, A. D. F. Phosphorus Chemistry in EVeryday LiVing; American
Chemical Society: Washington, DC, 1987.
(2) Caldwell, C. G. et al. Bioorg. Med. Chem. Lett. 1996, 6, 323-328.
(3) Hengge, A. C.; Onyido, I. Curr. Org. Chem. 2005, 9, 61-74.
(7) Bourne, N.; Chrystiuk, E.; Davis, A. M.; Williams, A. J. Am. Chem. Soc.
1988, 110, 1890-1895.
9
10.1021/ja0501565 CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 7703-7711
7703

A R T I C L E S
Onyido et al.
Scheme 1. Alternative Mechanisms, an Addition-Elimination (AN + DN) Pathway with the Formation of a Pentacoordinate (Phosphorane)
Intermediate (Pathway A) and a Concerted (ANDN) Reaction (Pathway B)
Literature data show that the rates of hydrolysis of the oxygen
analogues of the title compounds, the aryl dimethylphosphinate
esters 1, correlate satisfactorily with σ- substituent constants.8
This result is consistent with either a concerted mechanism or
an addition-elimination mechanism in which the departure of
the leaving group is rate limiting. A kinetic imperative for the
latter mechanism requires that the condition k-1/k2 > 1 holds
in Pathway A, Scheme 1. Since hydroxide ion is a poorer leaving
group than aryloxide ions, a concerted mechanism was advo-
cated for this system in preference to rate-limiting separation
of the leaving group in an addition-elimination mechanism.8
Furthermore, the transition states for the hydrolysis of alkyl9
and aryl10 phosphate and phosphorothioate triesters are very
similar,11 even though the thio triester hydrolyzes at a slower
rate than its oxygen counterpart (the so-called “thio effect”).
By analogy, it seemed unlikely to us that the aryl dimethylphos-
phinothioate esters of the type depicted in 2 would react by a
stepwise, associative mechanism when a concerted mechanism
has been demonstrated for 1.
of charge will be located on the O atom of the nucleophile in
the transition states for bond formation (k1) and bond cleavage
(k2).7 We have also measured the dependence of rate on leaving
group pKa for reactions of 2 with HO- and PhO-, to enable a
determination of the extent of cleavage of the P-OAr bond in
the TS of the reaction.
LFER experiments have been supplemented with heavy-atom
(18O and 15N) kinetic isotope effect (KIE) measurements as
probes of TS structure. The primary 18O KIE at the position of
18
bond fission,
k
, gives a measure of the extent of P-O
bridge
bond cleavage in the TS. The 15N KIE in the leaving group
NO2, 15k, is sensitive to the amount of charge delocalized into
the aromatic ring of the nucleofuge in the TS. The origin of 15
k
lies in contributions from a quinonoid resonance form in the
nitrophenolate anion, and differences in stiffness of N-O bonds
and N-C bonds. Because N-O bonds are stiffer, the nitrogen
atom is more tightly bonded in neutral 4-nitrophenol than in
the phenolate anion. The 15K equilibrium isotope effect for
deprotonation of 4-nitrophenol is thus normal, 1.0023 (
0.0002.13
To obtain definitive information regarding the mechanism
and structure of the TS for these reactions, and to resolve the
discrepancy noted above, we have undertaken a study of the
dependence of rate on nucleophile pKa for the reaction of
4-nitrophenyl dimethylphosphinothioate (2a) with a series of
oxyanionic nucleophiles, which include phenoxides, alkoxides,
and hydroxide ion. The pKa values of these nucleophiles straddle
that of the leaving group, 4-nitrophenoxide (pKa ) 7.14). Such
LFER experiments have been found to have diagnostic utility
in determining the relative timing of bond formation and bond
The corresponding isotope effects have been measured for
other group transfer reactions with the 4-nitrophenyl leaving
group. These include acyl transfer reactions of 4-nitrophenyl
acetate;14 the hydrolysis of 4-nitrophenyl phosphate monoester,
and diesters and triesters with the same nucleofuge;15 and the
hydrolysis of 4-nitrophenyl sulfate monoester and diesters.16 The
maximum value of 18kbridge, with the 4-nitrophenyl leaving group,
is in the range of 1.02-1.03 in a late transition state with
significant bond fission, such as the hydrolysis of 4-nitrophenyl
sulfate and 4-nitrophenyl phosphate.15 In the late TS of the
hydrolysis of 4-nitrophenyl phosphate dianion, where close
proximity to the anionic phosphoryl group in the TS enhances
delocalization of charge, 15k ) 1.0028 ( 0.0002.17 In hydrolysis
reactions of phosphodiesters and triesters, which are concerted
with tighter transition states and less bond fission to the
fission in a number of nucleophilic substitution reactions.12
change in the rate-limiting step will be expected to occur in the
mechanism of Pathway A, Scheme 1 when ∆pKa ) pKnuc
A
-
pKlg ) 0, especially when the nucleophiles involved are
structurally related. Such a change will manifest a nonlinear
Brønsted plot, with a break at ∆pKa ) 0, since different amounts
(8) Douglas, K. T.; Williams, A. J. Chem. Soc., Perkin Trans. 2 1976, 515-
521.
(9) Fanni, T.; Taira, K.; Gorenstein, D. G.; Vaidyanathaswamy, R.; Verkade,
J. G. J. Am. Chem. Soc. 1986, 108, 6311-6314.
(10) Hong, S. B.; Raushel, F. M. Biochemistry 1996, 35, 10904-10912.
(11) Catrina, I. E.; Hengge, A. C. J. Am. Chem. Soc. 2003, 125, 7546-7552.
(12) Williams, A. Free Energy Relationships; Royal Society of Chemistry:
Cambridge, UK, 2003; p 297.
(13) Hengge, A. C.; Cleland, W. W. J. Am. Chem. Soc. 1990, 112, 7421-7422.
(14) Hengge, A. C.; Hess, R. A. J. Am. Chem. Soc. 1994, 116, 11256-11263.
(15) Hengge, A. C. Acc. Chem. Res. 2002, 35, 105-112.
(16) Younker, J. M.; Hengge, A. C. J. Org. Chem. 2004, 69, 9043-9048.
(17) Hengge, A. C.; Edens, W. A.; Elsing, H. J. Am. Chem. Soc. 1994, 116,
5045-5049.
9
7704 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Hydrolysis of Aryl Dimethylphosphinothioates
A R T I C L E S
nucleofuge, 18kbridge ranges from 1.004 to 1.006, and 15k ranges
from 1.0007 to 1.0016.15
substituted phenolates as products, initial repetitive scans of the reaction
mixture (substrate + nucleophile) were carried out prior to the kinetic
determination to ascertain the appropriate wavelength for monitoring
the appearance of the product. All reactions involving the alcoholates
and hydroxide ion as nucleophiles were monitored for 10 half-lives.
For these reactions, there was good agreement between the experimental
and theoretical absorbances of the product at infinite reaction time.
Pseudo-first-order rate constants, kobs, were obtained by linear regression
analysis, as slopes of plots of ln(A∞ - At) vs time. Second-order rate
constants, knuc, were reckoned from plots of kobs vs [Nuc-], which gave
very good linearity in all cases. The rate constants of the reactions
with the phenolate nucleophiles were determined by the initial rate
method. The theoretical absorbance of the product at infinity was used
to calculate the pseudo-first-order rate constant in such cases. Division
of the initial rate of absorbance change by the theoretical absorbance
at infinity gave the kobs. In a few cases, e.g. reactions involving
phenoxide and 4-methylphenoxide ions as nucleophiles, an internal
check on the technique was made by allowing the reactions to go to
completion. The rate constants obtained by both methods were in very
good agreement.
If the reaction is concerted, large KIE values would be
expected at both positions. If, on the other hand, the reaction is
stepwise, then formation of the intermediate is expected to be
rate limiting in the region where nucleophile pKa . leaving
group pKa (7.14), and the KIEs will then reflect formation of
the intermediate. In this case, no 15N KIE should be observed,
and the 18O KIE would likely be inverse, reflecting compression
of bending modes.
LFER and KIE methods are both powerful tools for the
diagnosis of reaction mechanism, and relatively few reactions
have been studied combining both methodologies. The results
obtained in the LFER and KIE studies, herein reported and
discussed below, provide good evidence for a concerted
mechanism in the reaction of these phosphinothioate esters and
enable a description of the TS structure.
Experimental Section
Kinetic Isotope Effect Measurements. The solvent isotope effects
on the reaction of 4-nitrophenyl dimethylphosphinothioate with one
phenol and one alcohol were obtained. For the reaction with phenol,
ester concentration was 3.4 × 10-5 M, and phenol concentration was
0.004 M. For the nucleophile 4,4,4,3,3,2,2-heptafluorobutanol, con-
centrations of 0.004 and 0.009 M were used, with an ester concentration
of 1.8 × 10-5 M. The isotope effects were obtained from separate
kinetic runs monitored at 400 nm in H2O and D2O, at 25 °C.
Materials. Distilled, deionized water was degassed under vacuum
just prior to use. 1,4-Dioxane was passed through an alumina column
to remove peroxides. Phenols and alcohols were commercial products
and were purified by either recrystallization from a suitable solvent or
by distillation. Analytical grade NaOH was standardized, using phe-
nolphthalein as indicator. Buffer materials used were of analytical
reagent grade. The syntheses of the compounds 2a-g, including the
isotopic isomers used for kinetic isotope effects experiments, are
reported in Supporting Information.
The experimental procedures employed for measurement of 15k and
18
k
were very similar to previously reported methods used to
bridge
Kinetic Measurements. Kinetic measurements were made on a Cary
50 Bio spectrophotometer equipped with a Lauda E100 thermostat.
Reactions in which HO- was the nucleophile were carried out in CAPS
buffer solutions in the pH range of 10.57-11.01. For alcoholates and
phenolates, except 2,4,5-Cl3PhO- and 2,3,5,6-F4PhO-, self-buffered
solutions of the nucleophiles (Nuc-) were obtained by partially
neutralizing the conjugate acid of the relevant nucleophile (NucH) with
NaOH, such that a NucH:Nuc- ratio of 2:1 resulted. In the case of
2,4,5-Cl3PhO- and 2,3,5,6-F4PhO-, the phenolate nucleophiles were
obtained by dissolving the parent phenols in MOPS and Bis-Tris buffer
solutions, respectively. The nucleophile solution in each case was
prepared from two stock solutions, one containing the phenol and the
buffer component and the other being an identical solution of the buffer
adjusted to the same pH as the former but without the phenol. These
two stock solutions were mixed in the appropriate proportions to obtain
different concentrations of the phenolates, while the parameters of buffer
concentration, ionic strength, and pH remained constant. All nucleophile
solutions for kinetic runs were maintained at ionic strength, I ) 1 M
(KCl).
Reaction was initiated in each case by injecting 25-50 µL of a stock
solution of the substrate in 1,4-dioxane into a 1 cm cuvette containing
3 mL of the desired run concentration of the nucleophile. The run
solution of the nucleophile was equilibrated thermally at 25.0 ( 0.1
°C in the cuvette holder of the spectrophotometer for 30 min before
the kinetic run was initiated. It was ensured that the substrate
concentration in the kinetic runs was kept in the region of 3.0-4.0 ×
10-5 M and that the nucleophile concentration was in sufficient excess
of the substrate to maintain pseudo-first-order conditions. Nucleophile
concentrations were kept in the range of 0.001-0.005 M for the
alcoholates and 0.005-0.025 M for the phenolates.
measure KIEs in similar systems where 4-nitrophenol is the leaving
group.15 The product isolation procedures used below have been used
previously, and control experiments have shown they do not result in
isotopic fractionation. Reactions to measure 15k were carried out using
approximately 100 µmol (25 mg) of natural abundance 2a in 100 mL
of a 1:1 dioxane to buffer solution in capped beakers. (The cosolvent
was required to solubilize the necessary quantity of the reactant.) The
buffer solution used was 0.1 M CHES at pH 9.6. This pH was chosen
to yield a convenient rate of reaction at ambient temperature.
The hydrolysis reactions were allowed to react at 25 °C for
approximately one half-life (2 h). The reactions were then stopped by
extraction three times with methylene chloride, which resulted in a two-
phase mixture with the unconsumed reactant in the organic layer. The
aqueous layer, containing the liberated 4-nitrophenolate, was titrated
to pH 4 using 1 N HCl and extracted with diethyl ether to isolate the
4-nitrophenol product. The unreacted ester, which had been extracted
into the organic layer, was completely hydrolyzed by addition of excess
base (20 mL of 0.1 N NaOH), slow evaporation of the methylene
chloride, and allowing the reaction to proceed at ambient temperature
for >10 additional half-lives. The 4-nitrophenol liberated in the
complete hydrolysis reaction mixture was also isolated via ether
extractions, as previously described. Upon termination of the reaction,
absorbance values were obtained at 410 nm. These values, in combina-
tion with the absorbance values obtained from assays of the 4-nitro-
phenol liberated at partial hydrolysis, were used to determine the fraction
of reaction. All reactions fell within 46-48% of completion. (The rate
in the 1:1 buffer:dioxane system is about 2.5-fold slower than that
expected in aqueous buffer at the same pH as the buffer component of
the mixed solvent. This most likely arises from pKa shifts resulting
from addition of the aprotic cosolvent, which would be expected to
reduce the concentration of hydroxide ion.)
The reactions in which 4-nitrophenyl dimethylphosphinothioate was
the substrate were monitored at 400 nm, the wavelength of maximum
absorption of 4-nitrophenolate. With weakly basic phenolates as
nucleophiles, the pH of the reaction medium was such that 4-nitrophenol
was the product; under this condition, the reaction was monitored at
350 nm. In the reactions of the other substrates yielding the variously
All reactions were performed in triplicate, and the 4-nitrophenol
samples were purified by sublimation and analyzed by isotope ratio
mass spectrometry using an ANCA-NT combustion system in tandem
with a Europa 20-20 isotope ratio mass spectrometer.
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7705

A R T I C L E S
Onyido et al.
Figure 1. Plot of log knuc vs pKa (nucleophile) for the reaction of 2a with oxygen nucleophiles in water at 25 °C. The line for nucleophiles with pKa < 11
is defined by eq 1, while the line for more strongly basic nucleophiles (pKa g 11) is defined by eq 2.
18
was measured similarly, using the mixture of [14N]
Table 1. Second-Order Rate Constants (kNuc
)
for the Reaction of
a
The
k
bridge
4-nitrophenyl and [15N]-4-nitrophen[18O]yl compounds. This method
utilizes the remote label method,18 in which the isotope ratio of the
nitrogen atom serves as a reporter for oxygen isotope ratios. The
observed isotope effect was corrected for 15k and for incomplete isotopic
incorporation (see Supporting Information).
Several Oxygen Nucleophiles with 4-Nitrophenyl
Dimethylphosphinothioate in Water at 25 °C
b
1
entry
nucleophile
pKa
k
nuc/M-1s-
1
HO-
15.74
12.89
12.43
11.4
6.63 ( 0.23
2
CHCl2CH2O-
CF3CH2O-
4.28 ( 0.25
3
3.81 ( 0.12
4
CF3CF2CF2CH2O-
4-MeOPhO-
PhO-
3.01 ( 0.03
Results
5
10.20
9.95
2.23 ( 0.03 × 10-1
5.40 ( 0.17 × 10-2
2.93 ( 0.03 × 10-2
2.04 ( 0.29 × 10-2
1.25 ( 0.24 × 10-2
1.08 ( 0.01 × 10-2
1.73 ( 0.04 × 10-3
8.20 ( 0.03 × 10-4
The solvent isotope effect for the reaction of 4-nitrophenyl
dimethylphosphinothioate with phenoxide is 1.00 ( 0.02. The
solvent isotope effect for the reaction of 4-nitrophenyl dimethyl-
phosphinothioate with 4,4,4,3,3,2,2-heptafluorobutanol is 1.01
( 0.05.
6
7
4-ClPhO-
9.38
8
3-CNPhO-
8.61
9
4-CNPhO-
7.95
10
11
12
2,5-Cl2PhO-
2,4,5-Cl3PhO-
2,3,5,6-F4PhO-
7.51c
6.72c
5.53c
Second-order rate constants for the reaction of 2a with a
variety of oxyanions at 25 °C, obtained as described in the
Experimental Section, are given in Table 1, together with the
pKa values of the conjugate acids of the nucleophiles in water.
A value of k-OH ) 6.63 M-1 s-1 was obtained in this study for
the hydrolysis of 2a in aqueous solution at 25 °C. Istomin and
Eliseeva reported a value of 5.12 M-1 s-1 under the same
conditions.5 A Brønsted-type plot of log knuc vs pKa of the
conjugate acid of the nucleophile is displayed in Figure 1. The
plot is linear for the series of structurally related phenoxides in
the pKa range 5.5-10.2, which straddles the pKa of the
4-nitrophenol leaving group (pKa ) 7.14). A ânuc value of 0.47
( 0.05 (R ) 0.968) is obtained in this region. A change in
slope is observed in the more basic region (pKa g 11) where
the nucleophiles are hydroxide and alkoxide ions; in this region,
the much smaller value of ânuc ) 0.08 ( 0.01 (R ) 0.996) is
calculated.
a Measured at ionic strength, I ) 1.0 M (KCl). b Values of pKa at 25
°C, taken from Jencks and Regenstein.19 c Values of pKa at 25 °C, taken
from Bourne et al.7
Polar substituent effects in the nucleophilic reactions of
aryldimethylphosphinothioate esters 2a-g for the attacking
nucleophiles PhO- and HO- were evaluated at 25 °C. The
relevant second-order rate constants are assembled in Table 2.
Construction of a Brønsted-type plot of log knuc vs pKa of the
leaving group results in the linear relationships shown in Figure
2. Values of âlg ) -0.54 ( 0.03 (R ) 0.991) and -0.52 (
0.09 (R ) 0.974) were obtained for PhO- and HO- nucleo-
philes, respectively.
Hammett plots were also constructed using the data in Table
2, in conjunction with literature values of σ substituent
constants.20,21 The plots obtained using σ- constants give a
significantly better correlation in each case than those involving
σ and σ° substituent constants (see Table 3). The plots using
(18) O’Leary, M. H.; Marlier, J. F. J. Am. Chem. Soc. 1979, 101, 3300-3306.
9
7706 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Hydrolysis of Aryl Dimethylphosphinothioates
A R T I C L E S
Figure 2. Plot of log knuc vs pKa (leaving group) for the reactions of HO- and PhO- with aryl dimethylphosphinothioate esters 2a-g in water at 25 °C. The
lines for HO- and PhO- are defined by eqs 4 and 5, respectively.
a
Table 2. Second-Order Rate Constants (knuc
)
for the Reactions
Table 3. Hammett F Values with Correlation Coefficients (R) for
Leaving Group Variation in the Reactions of HO- and PhO- with
Substituted Aryl Dimethylphosphinothioates in Water at 25 °C (see
text)
of HO- and PhO- with a Series of Substituted
Dimethylphosphinothioates in Water at 25 °C
b
1
entry
leaving group
pKa
k
nuc/M-1 s-
F
value (R)
A. Nucleophile ) HO-
substituent
constant20, 22
HO-
PhO-
1
2
3
4
5
6
7
4-NO2PhO-
7.14
6.63 ( 0.23
2.94 ( 0.24
1.66 ( 0.07
1.25 ( 0.03
0.48 ( 0.04
0.44 ( 0.01
0.23 ( 0.01
4-Cl,3-NO2PhO-
3-NO2PhO-
3-CNPhO-
3-ClPhO-
4-ClPhO-
PhO-
7.78
8.35
8.61
9.01
9.30
9.95
σ
1.43 ( 0.26 (0.926)
1.62 ( 1.02 (0.745)
1.45 ( 0.10 (0.996)a
1.41 ( 0.24 (0.933)
1.37 ( 0.11 (0.993)a
1.17 ( 0.07 (0.991)
0.93 ( 0.02 (0.9997)a
σ°
σ-
1.80 ( 0.95 (0.802)
1.05 ( 0.27 (0.940)
B. Nucleophile ) PhO-
1
2
3
4
4-NO2PhO-
7.14
7.78
8.35
8.61
5.40 ( 0.17 × 10-2
3.65 ( 0.0.18 × 10-2
1.22 ( 0.02 × 10-2
1.10 ( 0.02 × 10-2
a Calculated from the data of Istomin and Eliseeva.5
4-Cl,3-NO2PhO-
3-NO2PhO-
3-CNPhO-
The absence of a significant solvent isotope effect for the
reaction of 4-nitrophenyl dimethylphosphinothioate with either
phenoxide or with 4,4,4,3,3,2,2-heptafluorobutanol rules out the
possibility that these nucleophiles act as general bases, since
such a mechanism should result in a normal solvent isotope
effect. Since the other aryl oxide and alcoholate nucleophiles
follow the same Brønsted correlations as these two respective
nucleophiles, it is concluded that the other reactions also proceed
by nucleophilic attack at phosphorus.
LFER Correlations. (a) Brønsted-Type Correlations. Brøn-
sted-type correlations have proved an invaluable mechanistic
tool for probing the properties of the TS of numerous reaction
types.12,22-25 Values of ânuc and âlg obtained from such plots
have been used to estimate TS structures and to evaluate the
a Measured at ionic strength, I ) 1.0 M (KCl). b Values of pKa at 25
°C, taken from Jencks & Regenstein.19
σ- for the reactions of the nucleophiles PhO- and HO- with
aryl dimethylphosphinothioates are shown in Figure 3.
The kinetic isotope effects were measured for the hydroxide
attack on compound 2a, at the oxygen atom of the scissile P-O
bond and at the nitrogen atom in the nitro group (Figure 4).
These KIEs were found to be: 15k ) 1.0009 ( 0.0003 and
18
k
) 1.0124 ( 0.0008.
bridge
Discussion
The kinetic behavior observed in these reactions is consistent
either with nucleophilic attack or with general base catalysis.
(22) Jencks, W. P. Chem. ReV. 1985, 85, 511-527.
(23) Grunwald, E. J. Am. Chem. Soc. 1985, 107, 125-133; Grunwald, E. J.
Am. Chem. Soc. 1985, 107, 4710-4715.
(19) Jencks, W. P.; Regenstein, J. In Handbook of Biochemistry and Selected
Data for Molecular Biology, 2nd ed.; CRC: Cleveland, OH, 1970; Section
J-187.
(20) Hansch, C.; Leo, A.; Taft, R. W. Chem. ReV. 1991, 91, 165-195.
(21) Taft, R. W. J. Phys. Chem. 1960, 64, 1805-1815.
(24) Jencks, D. A.; Jencks, W. P. J. Am. Chem. Soc. 1977, 99, 7948-7960.
(25) Leffler, J. E.; Grunwald, E. Extrathermodynamic free energy relationships.
In Rates and Equilibria of Organic Reactions; Wiley: New York, 1963;
Chapter 7.
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7707

A R T I C L E S
Onyido et al.
Figure 3. Hammett plots for the reactions of 2a-g with HO- and PhO- in water at 25°C, utilizing σ- substituent constants (see text).
Scheme 2. Compounds 1(a-j) Were the Subject of a Previous
Study8 Concluding that Nucleophilic Attack Occurs by a Concerted
Mechanisma
Figure 4. Compound 2a, with the positions and magnitudes of KIE
measurements indicated, and the notation used.
response of TS structures to changing substitution patterns and
reaction conditions, especially when the nucleophiles are
structurally related and possess the same atom at the nucleophilic
site.23,25-28
The Brønsted correlation in Figure 1 is linear for the
phenoxide nucleophiles; the more basic alkoxide and hydroxide
nucleophiles follow a different linear correlation, with a much
smaller slope. The solid line defined by phenoxide nucleophiles
in Figure 1 is given by eq 1, while eq 2 defines the solid line
obtained with hydroxide and alkoxide nucleophiles, giving ânuc
) 0.47 ( 0.05 (R ) 0.968) and 0.08 ( 0.01 (R ) 0.994),
respectively, in the two regions.
a Reactions of compounds 2(a-g) have also been studied previously,
leading to contradictory mechanistic conclusions, and are the subject of
this study.
to ânuc ) 0.41 measured in 90% water-10% dioxane for the
hydrolysis of the oxygen analogue, 1a, which also reacts via a
concerted mechanism.8
log knuc ) (0.47 ( 0.05)pKa - (5.70 ( 0.41)
log knuc ) (0.08 ( 0.01)pKa - (0.37 ( 0.08)
(1)
(2)
The break at high nucleophile basicity is too sharp to be
ascribed to a gradual change of selectivity with increasing
nucleophilic reactivity, as would be expected from a Hammond
effect.29 As is evident from Figure 1, the break does not occur
where ∆pKa ) pKnuc - pKlg ) 0, which might imply a change
in rate-limiting step in a stepwise mechanism. These data are
consistent with the well-documented effect of solvation on
nucleophilic reactivity in hydroxylic solvents.24,30,31 Brønsted
The linear plot for phenoxide nucleophiles in the pKa range
5.53-10.2, which straddles the basicity of the leaving group
(pKa of 4-nitrophenol ) 7.14), is good evidence that the reaction
of these nucleophiles with 2a occurs by a concerted mechanism.
The value of ânuc ) 0.47 obtained in this study may be compared
(26) Williams, A. Acc. Chem. Res. 1984, 17, 425-430.
(27) Williams, A. AdV. Phys. Org. Chem. 1992, 27, 1-55.
(28) Williams, A. Chem. Soc. ReV. 1994, 23, 93-100.
(29) Hupe, D. J.; Jencks, W. P. J. Am. Chem. Soc. 1977, 99, 451-464.
9
7708 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Hydrolysis of Aryl Dimethylphosphinothioates
A R T I C L E S
Scheme 3. More Effective Solvation of the Strongly Basic
Nucleophiles Necessitates a Kinetically Significant Desolvation
Step before Nucleophilic Attack on the Substrate Occurs
group in the TS, irrespective of the apparent difference in ânuc
for phenoxides and alkoxides/HO-. This is entirely consistent
with the conclusion reached above, namely that the amount of
bond formation in the TS is the same for the two groups of
nucleophiles, phenoxides and alkoxides/HO-, and that the
different values of ânuc mirror the difference in the ground state
of the two nucleophile types due to strong solvation of the
strongly basic nucleophiles.
plots for reactions of phenoxide and alkoxide nucleophiles are
frequently discontinuous, with a significantly smaller slope for
the more basic alkoxides. The diminished slope is explained
by more effective solvation of HO- and strongly basic alkoxides,
in comparison to the more weakly basic phenolate anions.
Desolvation of the strongly basic nucleophiles, according to
Scheme 3, is a requirement before nucleophilic attack on the
substrate occurs in the concerted reaction;31 Kd in Scheme 3 is
the equilibrium constant for the desolvation step.
The foregoing considerations suggest that HO- and the
alkoxide ions react by the same mechanism as the phenoxides.
According to this model, the decreased Brønsted-type slope for
the reaction of the strongly basic nucleophiles is a manifestation
of negative deviations of log knuc at high basicity according to
eq 3.31,32
To summarize, the Brønsted-type LFER data obtained in this
study have yielded values of ânuc ) 0.47 and âlg ) -0.53. These
values indicate that there is moderate bond formation and bond
cleavage in the TS of a concerted mechanism for the reaction
of aryl dimethylphosphinothioates with oxyanion nucleophiles.
As will also be shown in the sections that follow using other
mechanistic criteria, the data are not consistent with an
associative, stepwise mechanism.
(b) Hammett Correlations. Hammett correlations have
featured prominently in the discussion of the mechanisms of
the reactions of aryl phosphinates.3,8,37,38 Reactions in which
rates are satisfactorily correlated with σ or σ° substituent
constants are deemed to react by an associative mechanism
involving rate-limiting formation of the phosphorane intermedi-
ate. On the other hand, correlation of rates with σ- constants
point to either a concerted mechanism or an associative
mechanism in which the decomposition of the phosphorane
intermediate is rate-limiting. In the latter case, additional criteria
are needed to decide for either mechanism.
Hammett plots have been constructed for the reactions of
HO- and PhO- with differently substituted aryl dimethylphos-
phinothioates in water, using the data in Table 2 in conjunction
with values of σ, σ°, and σ- obtained from the literature.20,21
The results of these plots are summarized in Table 3, while
Figure 3 shows the plots obtained for both nucleophiles using
σ- substituent constants. We have also included in Table 3 the
result of an analysis of the Hammett plots obtained with the
different substituent constants from the limited data of Istomin
and Eliseeva for this system in pure water.5 It is evident from
Table 3 that significantly better correlations are obtained in our
work with σ- constants than with σ or σ° constants, as would
be expected when cleavage of the P-OAr bond is involved in
the rate-limiting step. Thus, the results of the Hammett plots
reinforce the conclusions drawn above on the basis of Brønsted
LFER correlations.
Istomin and Eliseeva4 concluded that the correlation of rates
with σ° indicated a stepwise AN + DN mechanism. Hammett
plots using the rate constant data of Istomin and Eliseeva4 with
σ, σ°, and σ- substituent constants are all linear (see Table 3),
the difference in correlation coefficients (R) of the different plots
being within the range of the experimental error in the data.
The fit is good in all cases, but within the narrow range of
variation of R, the correlation with σ- actually gives the best
fit. Under these circumstances, we do not think that the scope
of the original data of these authors warrants their definitive
conclusion that the mechanism of reaction is AN + DN, since
they relied on Hammett plots as the sole criterion of mechanism.
(c) Kinetic Isotope Effects. The kinetic isotope effects are
those for hydroxide attack on 2a. These data also indicate partial
fission of the P-O bond in the transition state of the rate-limiting
∆log knuc ) (1 - ânuc) log Kd
(3)
Such solvational imbalance30,31 has been observed in phosphoryl
transfer reactions31,33 and other group transfers,29,34 as well as
in the solvolysis of 1-phenylethyl carbocations.35 Bernasconi36
has proposed a model in which both partial desolvation of the
nucleophile and bond formation occur in the same step, to
account for decreased reactivity of strong bases in these
processes. In this instance, the imbalance results from the
necessity for desolvation of the nucleophile to progress ahead
of nucleophilic attack in the TS.
Data for the dependence of rate of the reactions of the
nucleophiles HO- and PhO- on leaving group pKa are given in
Table 2 and are displayed graphically in Figure 2. Equations 4
and 5, respectively, define the straight lines obtained in Figure
2 for HO- and PhO- as nucleophiles, giving âlg ) -0.54 (
0.03 (R ) 0.991) for HO- and -0.52 ( 0.09 (R ) 0.974) for
PhO-.
log knuc ) (-0.54 ( 0.03)pKa + (4.67 ( 0.28)
log knuc ) (-0.52 ( 0.09)pKa + (2.48 ( 0.68)
(4)
(5)
It is noted that âlg is similar for both nucleophiles, hence, for
purposes of subsequent discussion, the average value of âlg
)
-0.53 is adopted. The similar sensitivity of the reaction to
substituents in the leaving group for attack by HO- and PhO-
indicates that nucleophilic attack on phosphorus causes the same
magnitude of decrease in the effective charge on the leaving
(30) Jencks, W. P. Effects of solvation on nucleophilic reactivity in hydroxylic
solvents. In Nucleophilicity, Harris, J. M., McManus, S. P., Eds.; American
Chemical Society: Washington, DC, 1987; pp 155-167.
(31) Jencks, W. P.; Brant, S. R.; Gandler, J. R.; Fendrich, G.; Nakamura, C. J.
Am. Chem. Soc. 1982, 104, 7045-7051.
(32) Buncel, E.; Cannes, C.; Chatrousse, A.-P.; Terrier, F. J. Am. Chem. Soc.
2002, 124, 8766-8767.
(33) Jencks, W. P.; Haber, M. T.; Herschlag, D.; Nazaretian, K. L. J. Am. Chem.
Soc. 1986, 108, 479-483; Omakor, J. E.; Onyido, I.; vanLoon, G. W.;
Buncel, E. J. Chem. Soc., Perkin Trans. 2 2001, 324-330.
(34) Jencks, W. P.; Gilchrist, M. J. Am. Chem. Soc. 1962, 84, 2910-2913.
(35) Richard, J. P.; Jencks, W. P. J. Am. Chem. Soc. 1984, 106, 1373-1383.
(36) Bernasconi, C. F. AdV. Phys. Org. Chem. 1992, 27, 119-238.
(37) Buncel, E.; Albright, K. G.; Onyido, I. Org. Biomol. Chem. 2004, 601-
610.
(38) Cook, R. D.; Rahhal-Arabi, L. Tetrahedron Lett. 1985, 26, 3147-3150.
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7709

A R T I C L E S
Onyido et al.
Figure 5. Effective charges in the transition state of the identity reaction in which the nucleophile and leaving group are 4-nitrophenol.
18
step. The magnitude of
k
, 1.0124, is ∼40% of the
The quantity ꢀTS is the effective charge on the nucleophile
in the TS, and is related to ânuc by eq 7, where ꢀR is the effective
bridge
maximum value of this isotope effect, and 15k is consistent with
about a third of a negative charge on the nucleofuge. These
data are entirely consistent with the LFER results. These data
are not consistent with an addition-elimination mechanism, in
which the first step should be rate-limiting, as a result of the
much higher pKa of hydroxide compared to 4-nitrophenol. In
the rate-limiting TS of such a mechanism no negative charge
should arise on the nucleofuge, and the P-O bond is little
affected. Such a mechanism would exhibit KIEs for the rate-
ꢀTS ) ânuc + ꢀR
(7)
charge on the nucleophile in the ground state, i.e. -1. For the
identity reaction, when nucleophile ) leaving group ) 4-nitro-
phenoxide, the charge on the nucleophile and the leaving group
will be the same; hence, the effective charge on both equals
-0.53 (Figure 5).
The total charge of nucleophile plus Me2P(S)-OPhX in the
reactant state is -1.0. In the TS, the charges on entering and
leaving groups sum to -1.06. This leaves a charge imbalance
of -0.06, which is balanced by +0.06 charge unit on the
thiophosphinoyl group to maintain overall charge balance. The
effective charges are displayed in the TS structure shown in
Figure 5.
limiting formation of the intermediate. This could reasonably
be expected to give rise to a small inverse value for
18
k
,
bridge
due to the compression of bending modes in the pentacoordinate
intermediate. A precedent for such an effect has been observed
in the amide-15N isotope effect for the alkaline hydrolysis of
4-nitroanilide.39 Under conditions where formation of a tetra-
hedral intermediate is rate-limiting, 15k ) 0.995 in contrast to
15k ) 1.035 when breakdown and accompanying C-N bond
fission is rate-limiting.39 The inverse KIE for intermediate
formation in the amide reaction indicates that the isotope effect
from compression of bending modes outweighs any reduction
in C-N bond order from loss of amide resonance, which would
give a normal isotope effect. In the present case, if a phospho-
rane intermediate forms, while some weakening of the P-O
bond is expected on the basis of the observation of slightly
longer P-O bonds in crystal structures of phosphoranes
ânuc and âlg values are normalized to obtain the Leffler
indices25-27,42
R
formation ) ânuc /âeq and Rfission ) âlg/âeq, which
measure the extent of bond formation and bond fission,
respectively, in the TS. Since âeq ) 1.00, Rformation ) ânuc
)
0.47 and Rfission ) âlg ) 0.53. These values locate the position
of the TS as Tq along the tightness diagonal23,25,43 on the More
O’Ferrall-Jencks diagram (Figure 6). The position of Tq is only
slightly displaced from the intersection of the synchronous route
and the tightness diagonal, showing that the reaction in question
is a concerted one in which bond cleavage is only very
marginally advanced over bond formation in the TS. The present
results are therefore clearly inconsistent with the formation of
a phosphorane intermediate, as has been advocated for this
reaction.5 No intermediate, derived from an associative or a
dissociative process, will be stable enough to have any real
existence at this point of the potential energy surface.
Comparison of the Behavior of the Me2P(O)-OPhX and
Me2P(S)-OPhX Systems. It is instructive to compare the
behavior of the present system [Me2P(S)-OPhX] with that of
its oxygen analogue [Me2P(O)-OPhX]8 to assess the influence
of the thio effect on the mechanism and TS of the reaction of
the title compounds. The rate constant ratio kp)o/kP)S is found
to be generally small, calculated as 2.4, 4.1, 5.0, 5.2, and 3.5
with 4-NO2, 3-NO2, 4-Cl, 3-Cl, and H, respectively, as the
substituents in the phenoxide leaving group. These values for
the thio effect may be compared with values of 4-11 and 12.4
obtained for the reaction of the methyl 2,4-dinitrophenyl
compared to analogous bonds in phosphate esters, this would
not be enough to account for such a large normal
especially when compression of bending modes would give an
inverse contribution to the observed KIE.
18
k
,
bridge
Effective Charge Distribution and Transition-State Struc-
ture. Having established a concerted mechanism for the reaction
under discussion, we now proceed to determine the effective
charge in the TS (ꢀTS) on the entering nucleophile and leaving
group to enable a discussion of the TS structure, using the
methodology developed by Williams12,27,28,40 and Jencks.22,24,41
This procedure requires knowledge of the value of âeq, according
to eq 6, provided that the reactions for which ânuc and âlg have
been measured are the microscopic reverse of each other, and
that both parameters relate to the same rate-determining step.
The reactions studied here utilize aryl oxides as the entering
and leaving groups. The parameter âeq is the overall charge
change on the nucleofuge (or nucleophile) between the reactant
and product states for the overall reaction, and is calculated
according to eq 6 to be +1.00. Since in the product the
nucleofuge has a defined charge of -1.0, this means that the
aryl group has virtually no effective charge in the reactant.
(39) Hengge, A. C.; Stein, R. L. Biochemistry 2004, 43, 742-747.
(40) Bourne, N.; Williams, A. J. Am. Chem. Soc. 1983, 105, (10), 3357-3358;
Williams, A. J. Am. Chem. Soc. 1985, 107, 6335-6339.
(41) Herschlag, D.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111, 7587-7596;
Herschlag, D.; Jencks, W. P. J. Am. Chem. Soc. 1990, 112, 1951-1956.
(42) Leffler, J. E. Science 1953, 117, 340-341.
âeq ) ânuc - âlg
(6)
(43) Albery, W. J.; Kreevoy, M. M. AdV. Phys. Org. Chem. 1978, 16, 87-157.
9
7710 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Hydrolysis of Aryl Dimethylphosphinothioates
A R T I C L E S
) O or S), Me2P(Y)-OPhX (Y ) O or S), Ph(Me)P(S)-SPhX,
and Ph(MeSO2CH2)P(O)-OPhX), have appeared in the litera-
ture.4,5,7,8,37,38,50,51 The features of these reactions have been
summarized.3 The stepwise mechanism involving rate-limiting
formation of the pentacoordinate intermediate has been advo-
cated for the transfer of the (thio)phosphinoyl group in most of
these reactions, based mainly on the results of Hammett
correlations. The notable exceptions to this general trend are
the alkaline hydrolysis of aryl dimethylphosphinates51 and the
reaction of aryl diphenylphosphinates with phenoxides in water,7
which were shown to react by a concerted mechanism.
The present study, which has relied on Brønsted LFER,
Hammett correlations, and heavy atom kinetic isotope effects
as criteria for mechanism, has provided unambiguous evidence
for a concerted mechanism for a reaction that was earlier
proposed5 to occur by an associative, stepwise mechanism. The
question that needs be asked is: how general is the stepwise
mechanism in the reactions of phosphinates? The corollary to
this question is: how uncommon is the concerted mechanism?
There are also subsidiary issues that need to be addressed. What
factors promote one mechanism over another and at what point
can a mechanistic changeover be anticipated? Clearly, more
work is required to answer these questions and elucidate the
factors that select for particular mechanisms.
Figure 6. Location of the TS on a More O’Ferrall-Jencks diagram of the
reaction of Me2P(S)-OAr with nucleophiles.
phosphate/phosphorothioate system with a variety of nucleo-
philes44 and the alkaline hydrolysis of the triethyl phosphate/
phosphorothioate system,9 respectively. The present system is
apparently less sensitive to S-substitution (P ) S versus P )
O) than phosphates.3 It is noted that Cook and co-workers45
have reported a small rate-retarding effect, less than a factor of
10, for the replacement of O by S in a series of alkyl
phosphinates.
The Leffler indices Rformation ) 0.47 and Rfission ) 0.53 have
been calculated37 from the data of Douglas and Williams8 for
the hydrolysis of Me2P(O)-OPhX. The same values of these
parameters have been obtained in the present study for the
transfer of the thiophosphinoyl group from Me2P(S)-OPhX to
oxyanions. The implication of this finding is that both substrate
series demonstrate the same amount of bond making and bond
fission in the TS of their reactions. This leads to the conclusion
that S-substitution in the P ) O bond of Me2P(O)-OPhX, while
exerting a marginal rate-retarding effect, does not alter the TS
structure in the reaction with oxygen nucleophiles. In this regard,
the behavior of phosphinate/phosphinothioate system is similar
to the phosphate/phosphorothioate system. In general, the TS
for the hydrolysis of phosphates and phosphorothioates are
similar,3,11,46,47 the exception being the reaction of monoesters,
where nucleophilic participation, that is absent in the DN + AN
reactions of phosphorothioates,48 is a feature of the ANDN
reactions of phosphate monoesters.49
Conclusions
The Brønsted ânuc and âlg parameters, measured for the
reaction of aryl dimethylphosphinothioate esters with oxyanion
nucleophiles in aqueous solution, and the Leffler indices derived
from them describe a one-step process in which the bond to
the incoming nucleophile is 47% formed and the bond to the
nucleofuge is 53% cleaved in the TS. Correlation of the second-
order rate constants for the reactions of HO- and PhO- with
σ- substituent constant supports a mechanism in which the bond
to the nucleofuge is broken in the TS. This conclusion is strongly
reinforced by the primary 18O and secondary 15N kinetic isotope
effects measured in the hydrolysis reaction, thus removing the
ambiguity in the literature concerning the mechanism of
hydrolysis of aryl dimethylphosphinothioates.
Acknowledgment. We thank the National Science Foundation
for support of this research through the U.S.-Nigeria Cooperative
Research Program (Grant INT-0217688) and the Human
Frontier Science Program for a Short-Term Fellowship (I.O.)
during which part of this work was accomplished.
Supporting Information Available: Synthetic procedures and
spectral data for compounds 2a-g, preparation of isotopically
labeled compounds, the equations used for calculation of the
kinetic isotope effects, and complete citation to ref 2. This
material is available free of charge via the Internet at
http://pubs.acs.org.
The Associative Mechanism and Reactions of Phosphi-
nates with Aryloxy Leaving Groups. Prior to this work, a
number of reports on the reactions of oxygen nucleophiles with
a range of phosphinate and phosphinothioate esters bearing
aryloxy or arylthiolate leaving groups (e.g., Ph2P(Y)-OPhX (Y
JA0501565
(44) Herschlag, D.; Piccirilli, J. A.; Cech, T. R. Biochemistry 1991, 30, 4844-
4854.
(45) Cook, R. D.; Farah, S.; Ghawi, L.; Itani, A.; Rahil, J. Can. J. Chem. 1986,
64, (8), 1630-1637.
(50) Cevasco, G.; Thea, S. J. Chem. Soc., Perkin Trans. 2 1993, 1103-1106;
Dunn, E. J.; Buncel, E. Can. J. Chem. 1989, 67, (9), 1440-1448; Haake,
P.; Diebert, C. E.; Marmor, R. S. Tetrahedron Lett. 1968, (50), 5247-
5250; Istomin, B. I.; Sakhdrukova, N. A.; Kalabina, A. V.; Sukhorkov, I.
J. Gen. Chem., USSR 1982, 52, 1787-1795; Pregel, M. J.; Dunn, E. J.;
Nagelkerke, R.; Thatcher, G. R. J.; Buncel, E. Chem. Soc. ReV. 1995, 24,
449-455; Zhdankovich, E. L.; Istomin, B. I.; Voronkov, M. G. J. Gen.
Chem., USSR 1981, 998-1001; Zhdankovich, E. L.; Istomin, B. I.;
Voronkov, M. G. J. Gen. Chem., USSR 1982, 764-767.
(46) Kirby, A. J.; Varvoglis, A. G. J. Am. Chem. Soc. 1967, 89, 415-423.
(47) Hollfelder, F.; Herschlag, D. Biochemistry 1995, 34, 12255-12264.
(48) Burgess, J.; Blundell, N.; Cullis, P. M.; Hubbard, C. D.; Misra, R. J. Am.
Chem. Soc. 1988, 110, 7900-7901; Cullis, P. M.; Iagrossi, A. J. Am. Chem.
Soc. 1986, 108, 7870-7871; Cullis, P. M.; Misra, R.; Wilkins, D. J. J.
Chem. Soc., Chem. Commun. 1987, 1594-1596.
(49) Herschlag, D.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111, 7579-7586.
(51) Williams, A.; Naylor, R. A. J. Chem. Soc. B 1971, 1967-1972.
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7711