Kinetics of C(2α)-proton abstraction from 2-benzylthiazolium salts leading to enamines relevant to catalysis by thiamin-dependent enzymes

Article abstract of DOI:10.1021/ja9633528

The kinetics of proton transfer from the C(2α) position of 2-(1-methoxybenzyl)thiazolium salts was studied for the p-H and p-N(CH₃)₃⁺ derivatives as models for the protonation of the enamine/C(2α)-carbanion, a key intermed

Full text of DOI:10.1021/ja9633528

2356
J. Am. Chem. Soc. 1997, 119, 2356-2362
Kinetics of C(2R)-Proton Abstraction from 2-Benzylthiazolium
Salts Leading to Enamines Relevant to Catalysis by
Thiamin-Dependent Enzymes
Gabriel L. Barletta,† Yu Zou, W. Phillip Huskey,* and Frank Jordan*
Contribution from the Department of Chemistry, Rutgers, The State UniVersity of New Jersey,
Newark, New Jersey 07102
ReceiVed September 24, 1996X
Abstract: The kinetics of proton transfer from the C(2R) position of 2-(1-methoxybenzyl)thiazolium salts was studied
+
for the p-H and p-N(CH3)3 derivatives as models for the protonation of the enamine/C(2R)-carbanion, a key
intermediate in many thiamin diphosphate-dependent enzymatic reactions. The reactions were studied by rapid mixing
of the salts with sodium hydroxide in a stopped-flow instrument while monitoring the progress of enamine formation
and decomposition in the visible region of the spectrum. It was demonstrated that under the conditions of the
1
experiments, the thiazolium ring opens to give a product with a characteristic H-NMR spectrum consistent with
both cis and trans stereochemistry about the amide bond generated in the ring opening. Analysis of the progress
curves for the p-trimethylammonium-substituted compound (1) gave the following rate constants (25 °C): 21 M-1
s-1 for deprotonation at C(2R), 300-540 s-1 for reprotonation, and 3.4 M-1 s-1 for thiazolium ring opening. For
the unsubstituted compound (2), the respective rate constants were 0.019 M-1 s-1, 1 s-1, and 0.5 M-1 s-1. Through
comparisons with results for the C(2R)-deuterated form of 1, the primary hydrogen isotope effect on deprotonation
of 1 was found to be 4-6. The C(2R)-H pKa values were estimated to be 15.0-15.5 for 1 and 15.7 for 2. When
compared with the turnover number for benzoylformate decarboxylase catalysis, the rate constant for reprotonation
of the enamine derived from 2 leads to an estimate of 4500 M as a minimum effective molarity for the enzyme.
The bioorganic chemistry of thiamin diphosphate (ThDP, the
Scheme 1. The Counterion in Each Case Was BF4-
vitamin B1 coenzyme)1 is distinguished by the reactivity of two
carbanions: the C(2) carbanion which is the conjugate base of
the thiazolium ring and the C(2R) carbanion/enamine which is
the conjugate base of 2-R-hydroxyethyl(or aryl)thiamin diphos-
phate. Because these carbanions are widely postulated to be
intermediates in reactions catalyzed by enzymes that utilize
ThDP, their thermodynamic, kinetic, and structural properties
are particularly important. To date, the C(2) carbanion has not
been observed directly, although there is evidence from kinetics
that the pKa of the conjugate acid is in the 17-19 range.2-4
C(2R) carbanions/enamines have been directly observed on
pyruvate decarboxylase using conjugated substrate analogs
(XC6H4CHdCHCOCOOH)5-15 and in the absence of enzyme
† Present address: Department of Chemistry, University of Puerto Rico,
Humacao University College, C.H.U. Station, Puerto Rico, 00791.
X Abstract published in AdVance ACS Abstracts, February 15, 1997.
(1) Kluger, R. Chem. ReV. 1987, 87, 863-876.
(2) Washabaugh, M. W.; Jencks, W. P. Biochemistry 1988, 27, 5044-
5053.
(3) Washabaugh, M. W.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111,
674-683.
(4) Washabaugh, M. W.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111,
683-692.
(5) Kuo, D. J.; Jordan, F. Biochemistry 1983, 22, 3735-3740.
(6) Kuo, D. J.; Jordan, F. J. Biol. Chem. 1983, 258, 13415-13417.
(7) Kuo, D. J.; Dikdan, G.; Jordan, F. J. Biol. Chem. 1986, 261, 3316-
3319.
(8) Jordan, F.; Kudzin, Z. H.; Kuo, D. J. Ann. N. Y. Acad. Sci. 1986,
471, 308-309.
(9) Jordan, F.; Adams, J.; Farazami, B.; Kudzin, Z. H. J. Enz. Inh. 1986,
1, 139-149.
(10) Jordan, F.; Akinoyosoye, O.; Dikdan, G.; Kudzin, Z. H.; Kuo, D.
J. In Thiamine Pyrophosphate Biochemistry; Schellenberger, A., Schowen,
R. L., Eds.; CRC Press: Boca Raton, FL, 1988; Vol. 1, pp 79-92.
(11) Annan, N.; Paris, R.; Jordan, F. J. Am. Chem. Soc. 1989, 111, 8895-
8901.
(12) Annan, N.; Jordan, F. J. Am. Chem. Soc. 1990, 112, 3222-3223.
(13) Zeng, X.; Chung, A.; Haran, M.; Jordan, F. J. Am. Chem. Soc. 1991,
113, 5842-5849.
in nonaqueous16 and aqueous solutions.17 The enamines derived
from C(2)-alkylthiazolium compounds have characteristic UV
absorbance maxima near 300 nm, while the C(2)-benzyl
compounds give enamines with λmax in the visible spectrum
above 380 nm.
We now report rate constants obtained by monitoring the
enamine absorbance generated in aqueous hydroxide using the
thiazolium salts 1 and 2 shown below in Scheme 1. Rate
constants for 1 were communicated by us a few years ago.17
(14) Menon-Rudolph, S.; Nishikawa, S.; Zeng, X.; Jordan, F. J. Am.
Chem. Soc. 1992, 114, 10110-10112.
(15) Zeng, X.; Farrenkopf, B.; Hohmann, S.; Jordan, F. Biochemistry
1993, 32, 2704-2709.
(16) Jordan, F.; Kudzin, Z. H.; Rios, C. B. J. Am. Chem. Soc. 1987,
109, 4415-4416.
(17) Barletta, B. G.; Huskey, W. P.; Jordan, F. J. Am. Chem. Soc. 1992,
112, 7607-7608.
S0002-7863(96)03352-5 CCC: $14.00 © 1997 American Chemical Society

Kinetics of C(2R)-Proton Abstraction
J. Am. Chem. Soc., Vol. 119, No. 10, 1997 2357
Scheme 2. Mechanism for the Reaction of 1 (k4 ) 0) and 2
(k4 * 0) with Hydroxide Ion
nitrogen is lost. The 4-methyl peak in panel B is at a very
small chemical shift (0.91 ppm) for allylic hydrogens, perhaps
because the trans (about the amide bond) form of the ring-
opened product can adopt a conformation in which the 4-methyl
group is situated under the substituted phenyl group where it
can experience ring-current shielding.23 Minor peaks in the
panel B spectrum are consistent with the cis (amide) isomer of
the ring-opened product.22
Figure 1. 1H-NMR (200 MHz, D2O) spectrum of 3 (structure shown
in panel A). Panel B shows the spectrum a few minutes after adding
enough NaOD/D2O to make the solution 120 mM in NaOD. Panel C
shows the spectrum64 after the a few drops of DCl was added to the
sample used to obtain the spectrum in panel B.
Additional support for our explanation of the spectra can be
derived from the results shown in Table 1. The spectra for
compounds 1 and 4 are qualitatively similar to the spectra for
3, suggesting that they all undergo a similar thiazolium ring
opening. The spectra for 5 are slightly different, because this
compound has a methyl group in the 5 position of the thiazolium
ring. After adding base, the signal (1.62 ppm) for this 5-methyl
group does not show the large shielding experienced by the
4-methyl group (to 0.84 ppm), indicating that only the 4-methyl
position can experience the ring-current effect discussed above.
For all of the substituted benzyl compounds (1-5), the intensity
of the signal for the C(2R) proton is greatly reduced, or
eliminated, upon addition of the base. This finding is in
agreement with the mechanism shown in Scheme 2 in which
C(2R) proton removal (and exchange for D2O in the NMR
experiments) competes with thiazolium ring opening. The
compound with no benzyl group (6) shows little or no exchange
over the time of the experiment (ca. 30 min). Coupling between
the C(2R) proton and the C(2R)-methyl protons is apparent both
in neat D2O and in 100 mM NaOD, providing additional support
for our assignments for these signals and for the lack of C(2R)
proton exchange.
Kinetics. Reactions were studied by monitoring the absor-
bance corresponding to the enamine intermediate in a stopped-
flow apparatus. Assignment of the absorbance to the enamine
is based on the reaction of the thiazolium salt precursor with a
non-nucleophilic base in aprotic solvents (DMSO or DMF).16
For 1h, it was reported previously that λmax was 410 nm in both
DMSO and water.17 The appearance and loss of the enamine
absorbance is presented in Figures 2 and 3 for compounds 1h
and 1d in the presence of various concentrations of hydroxide
ion. Similar results for 2 are shown in Figure 5. These results
can be explained using the simple mechanism shown in Scheme
2. For 1h, the initial rise in the absorbance reflects the increase
in enamine concentration as hydroxide ion removes the C(2R)-
proton in a step with a pseudo-first-order rate constant k2H. The
slow loss of enamine absorbance can be accounted for by a
competing decomposition (k3H) of 1h in the presence of
hydroxide ion.
We present a complete study, including kinetic isotope effects
1
for the reaction of 1, along with H NMR evidence for the
structure of the product of a reaction, thiazolium ring opening,
which competes with C(2R)-proton removal. The results,
especially the rate constants for reprotonation of the enamine,
have implications for mechanistic studies of ThDP-dependent
enzymes such as pyruvate decarboxylase (PDC) and benzoyl-
formate decarboxylase (BFD). A high-resolution X-ray crystal-
lographic structure is available for PDC,18,19 and for BFD, a
high-resolution structure is emerging.20
Results
1H-NMR Spectra. The influence of base on the benzylthi-
azolium salts listed in Scheme 1 was first examined by analyzing
1
the H-NMR spectra of the compounds in the presence and
absence of base (NaOD). Figure 1 shows spectra for one of
the compounds, the p-Cl derivative (3). The labels in Figure 1
and in Scheme 1 identify several of the peaks in the spectra.
Panel A of Figure 1 shows the spectrum of the compound in
the absence of base (neat D2O), panel B shows the spectrum
after base was added to give a solution that was 120 mM NaOD,
and panel C shows the spectrum of the same solution after it
was neutralized by the addition of DCl. The panel B spectrum
is consistent with the ring-opened structure shown on the figure
and in Scheme 2. Ring-opened products have been reported
previously for the reaction of hydroxide ion with other thiazo-
lium-containing compounds.21,22 On addition of base, the
change (panel A to B) in the 5-H chemical shift is in the
direction expected if aromaticity in the thiazolium ring is lost.
The 3-methyl peak moves as expected if the charge on the
(18) Dyda, F.; Furey, W.; Swaminathan, S.; Sax, M.; Farrenkopf, B.;
Jordan, F. Biochemistry 1993, 32, 6165-6170.
(19) Arjunan, P.; Umland, T.; Dyda, F.; Swaminathan, S.; Furey, W.;
Sax, M.; Farrenkopf, B.; Gao, Y.; Zhang, D.; Jordan, F. J. Mol. Biol. 1996,
256, 590-600.
(20) Hasson, M. S.; Muscate, A.; Henehan, G. T. M.; Guidinger, P. F.;
Petsko, G. A.; Ringe, D.; Kenyon, G. L. Prot. Sci. 1995, 4, 955-959.
(21) Hopmann, R.; Brugnoni, G. P. J. Am. Chem. Soc. 1982, 104, 1341-
1344.
The C(2R)-deuterated substrate 1d shows a much slower
increase in enamine absorbance (Figure 3). The difference in
(22) Chriss, D.; Miller, R. H.; Echols, R. E.; Vessel, E. Org. Mag. Res.
1984, 22, 75-79.
(23) Becker, E. D. High Resolution NMR; Academic Press: New York,
1980.

2358 J. Am. Chem. Soc., Vol. 119, No. 10, 1997
Barletta et al.
Table 1. 1H-NMR Spectra (200 MHz) of Compounds Listed in Scheme 1 in D2O and in NaOD/D2O
chemical shift,a ppm
compdb
4-Me (s, 3 H)c
OMe (s, 1 H)
3-Me (s, 3 H)
C2RH (s, 1 H)
5H (s, 1 H)
2RCH3
4
1
3
5
6
D2O
95 mM NaOD
D2O
390 mM NaOD
D2O
120 mM NaOD
D2O
2.35
3.36
3.62
5.32
5.15
6.00
not seen
5.92
5.22
5.85
not seen
5.06
q, J ) 6.0
4.18, 4.31
6.00
7.63
0.89, 1.88d
2.37
3.29, 3.18
3.40
2.76, 2.88
3.67
6.17e
7.64
0.84e
2.38
0.91, 1.90
2.33
0.84, 1.69
2.41
3.26e
2.70e
6.17e
3.41
3 .31, 3.22
3.36
3.20, 3.18
3.40
3.67
2.76, 2.79
3.61
2.73, 2.85
3.86
7.68
6.20, 6.35
2.23f
95 mM NaOD
D2O
1.62, 1.97
7.62
1.53
d, J ) 6.0
1.31, 1.17
100 mM NaOD
D2O
1.82, 1.78
2.37
3.17, 3.29
3.52
2.93, 2.87
3.75
6.28
10
2.48f
100 mM NaOD
0.94, 2.02
3.42, 3.32
2.86, 2.98
not seen
1.95, 2.07
a Chemical shifts are relative to TMS as measured by fixing the water chemical shift at 4.67 ppm. b See Scheme 1 for structures and keys for
identifying chemical shifts. c Integration refers to the major isomer (see note d). d The second chemical shift of any pair of values listed in the table
refers to the minor isomer of the product. The chemical shift of the major isomer is listed first. e In these cases, assignments for the cis isomer
could not be made. The peaks were either small or obscured by other signals. f These entries for compounds 5 and 10 refer to the signal for the
5-methyl group (see Scheme 1).
Figure 3. Progress curves observed for the reaction of 1d (0.00156
Figure 2. Progress curves observed for the reaction of 1h (0.00115
M) with aqueous sodium hydroxide (concentrations on the figure) at
M) with aqueous sodium hydroxide (concentrations on the figure) at
25 °C. Concentrations refer to values after mixing in the stopped-flow
25 °C. Concentrations refer to values after mixing in the stopped-flow
instrument. The absorbance was monitored at 400 nm. The curves drawn
instrument. The absorbance was monitored at 400 nm. The curves drawn
through the data correspond to the fits to eqs 1-5 (with k4 ) 0) giving
through the data correspond to the fits to eqs 1-5 giving the parameters
the parameters shown in Table S2 (Supporting Information) and Figure
shown in Table S1 (Supporting Information). The results of the fits
4.
are shown in Figure 4. Only half of the runs used to construct Table
S1 and Figure 4 are shown in Figure 2.
with numerical integration.25 For fits26 of progress curves for
1h and 1d, the extinction coefficient for the enamine (ꢀE) was
the results for 1h and 1d can be readily explained if there is a
fixed at 17 000 M-1 cm-1 (determined in DMSO; see the next
significant isotope effect on the rate constant for proton
section), and the extinction coefficient for the product (ꢀP) was
abstraction by hydroxide ion, k2. The inset on Figure 3 shows
fixed at 14 M-1 cm-1 (determined by comparing the baseline
that there is also a small, fast component to the absorbance
increase, which arises from the reaction of a small amount (0.1-
0.2%, based on fitting results) of contaminating 1h in the sample
and final absorbances of long runs). The initial concentration
of enamine (and the initial concentration of 1d for the 1d fits)
was included as a parameter. Initial concentration of 1h was
of 1d. No return from the enamine to 1d is shown in Scheme
computed from mass conservation assuming that no product had
2. If the HOD produced in the deuteron abstraction from 1d
accumulated at the zero time of the instrument. All other initial
by hydroxide ion is rapidly lost to the solvent HOH before the
concentrations were set to 0. These initial concentrations refer
enamine can return to 1d, this is a valid approximation. As is
the case with 1h, the gradual loss of enamine is accounted for
in Scheme 2 through the steps identified with the rate constants
to the zero time of the instrument and account for enamine
produced during the mixing time. The initial absorbance (zero
time of the instrument) was computed in the analysis as ꢀE ×
k1H and k3H
.
[Enamine]0. For fits to the 1h data, k′2D and k′3D were fixed at
Analysis of Progress Curves. The progress curves shown
in Figures 2, 3, and 5 were fit to the absorbance, A, obtained
from integration of the rate equations corresponding to Scheme
2 (eqs 1-5). A program was written to replace the function
evaluations needed in a standard nonlinear least-squares method24
(25) Hindmarsh, A. C. In Scientific Computing: Applications of Math-
ematics and Computing to the Physical Sciences; IMACS Transactions on
Scientific Computation Vol. 1; Stepleman, R. S., Ed.; North-Holland
Publishing Co.: Amsterdam, 1983; pp 55-64; Subroutine lsode (and
supporting routines) available at http://www.netlib.org.
(26) Selected examples of this approach to fitting can be found in (a)
Zimmerle, C. T.; Frieden, C. Biochem. J. 1989, 258, 381-387 and in (b)
Adam, L. C.; Fabian, I.; Susuki, K.; Gordon, G. Inorg. Chem. 1992, 31,
3534-3541.
(24) Garbow, B. S.; Hillstrom, K. E.; More, J. J. Minpack Project;
Technical Report, Argonne National Laboratory. Subroutine lmfdif1 (and
supporting routines) available at http://www.netlib.org.

Kinetics of C(2R)-Proton Abstraction
J. Am. Chem. Soc., Vol. 119, No. 10, 1997 2359
second pathway for decomposition of the enamine (k′4). With
k′4 fixed at zero, as was done for the fits of 1h and 1d, k′1H for
compound 2 showed a very strong linear dependence on
[NaOH]. According to our model in Scheme 2, there should
be no such dependence on hydroxide ion. A second decom-
position pathway (k′4) was therefore included in the model to
allow for a hydroxide-dependent loss of enamine. Fitting of
the data for 2 to the model with six parameters allowed to float
(k′1H, k′2H, k′3H, k′4, ꢀP, and initial enamine concentration) gave
acceptable results but with very large error estimates. In order
to demonstrate the acceptability of our model, the k′1H was fixed
at 1.0 M-1 s-1 for all concentrations of hydroxide. This value
is the intercept of a plot against [NaOH] of k′1H obtained from
fits in which k′4 was fixed at zero. The results of the fit are the
curves drawn through the data in Figure 5.
Influence of the Enamine Extinction Coefficient. The
extinction coefficients fixed for the enamine (ꢀE) were values
determined for the enamine in DMSO. The short lifetime of
the enamine made it impossible for us to determine ꢀE in water,
the solvent used for the rate experiments. In order to get a
measure of the sensitivity of ꢀE to solvent, a value (14 000 M-1
cm-1, 410 nm) was determined in 90% (volume) dioxane/
DMSO. Fortunately, the rate constants are not very sensitive
to the choice of extinction coefficient. In a series of fits to the
1h data, when ꢀE was varied over the range of 13 000-19 000
M-1 cm-1 only k2 showed a significant, but small, dependence
on ꢀE ranging from 27-18 M-1 s-1. Instead, the initial enamine
concentration is the parameter that is most sensitive to ꢀE.
Dependence of k′ on [NaOH]. Figure 4 shows plots of rate
constants corresponding to Scheme 2 against [NaOH] for 1h
and 1d. The slopes and intercepts of the least-squares lines
through these data are shown at the bottom of Tables S1 and
S2 (Supporting Information). The resulting second-order rate
Figure 4. Dependence of first-order rate constants on [NaOH] for 1h
and 1d. Slopes and intercepts of the least-squares lines drawn through
the data are shown at the bottom of Tables S1 and S2. Note that k′1H
was divided by 100 for plotting.
constants are, for 1h, k2H ) 21.1 ( 0.7 M-1 s-1 and k3H
)
3.41 ( 0.09 M-1 s-1, and, for 1d, k2D ) 3.46 ( 0.42 M-1 s-1
and k3D ) 2.84 ( 0.17 M-1 s-1. Attempts to control ionic
strength were not successful because the large salt concentrations
that were needed changed the viscosity of the solutions to such
an extent that the stopped-flow experiments could not catch the
initial rise in the absorbance. The values reported here for 1h
are not significantly different from our previous report in which
a slightly different fitting procedure was used.27 The slight
dependence of k′1H on [NaOH] could be a consequence of fitting
to a model that does not incorporate a hydroxide-dependent
decomposition of the enamine (k4). Figure 6 shows the
dependence of fitted rate constants vs. [NaOH] for 2. The slopes
of these lines give second-order rate constants for 2: k2 )
0.0186 ( 0.0006 M-1 s-1 and k4 ) 7.11 ( 0.15 M-1 s-1. Fitted
values of k′3 for 2 had large estimated errors (Table S3) for the
lower hydroxide runs, so the second-order rate constant k3 is
not well determined. The line shown on Figure 6 is a least
squares fit forced through a zero intercept (slope ) 0.50 ( 0.06).
An unconstrained linear fit gave a slope of 1.39 ( 0.10 M-1
s-1 with a large negative intercept (-0.41 s-1).
Figure 5. Progress curves for the reaction of 2 (0.00340 M) with
aqueous sodium hydroxide at 25 °C. Absorbance was monitored at 380
nm. The curves through the data correspond to the fits to eqs 1-5
with k′1H fixed at 0.1 s-1 (see text). Parameters obtained from each fit
are shown in Table S3 and Figure 6. Only half of the runs used for
Table S3 and Figure 6 are shown above in Figure 5.
0. For 1d k′1H, k′2H, and k3H were fixed at values obtained from
fits for 1h. Tables S1 and S2 show the results of these least-
squares fits. k′4 was fixed at zero for fits to 1h and 1d data.
This extra enamine-decomposition pathway was needed only
to fit the data for 2 as is described below.
d[E]/dt ) -k1H[E] + k′2H[SH] + k′2D[SD] - k′4[E] (1)
d[SD]/dt ) -(k′2D + k′3D)[SD]
d[SH]/dt ) -(k′2H + k′3H)[SH] + k1H[E]
d[P]/dt ) k′3H[SH] + k′3D[SD] + k′4[E]
dA/dt ) ꢀE d[E]/dt + ꢀP d[P]/dt
(2)
(3)
(4)
(5)
Discussion
The progress curves for reactions of 1h, 1d, and 2 with
hydroxide are described well by the mechanism of Scheme 2
in which the k4 pathway for enamine decomposition was needed
only to explain the hydroxide dependence of rate constants
Fits to the progress curves for 2 in Figure 5 required two
modifications to the procedures described for 1. First, it was
necessary to allow the extinction coefficient for the product to
be a parameter in the fits. The results (Table S3) showed that
ꢀP varied from about 20 M-1 cm-1 at 0.30 M NaOH to nearly
28 M-1 cm-1 (ꢀE was fixed at 14000 M-1 cm-1 for all fits of
compound 2 runs). Second, it was necessary to allow for a
(27) In our previous report,17 rate constants were obtained by fitting to
the closed-form integrated expression of the rate equations with a time-
shift parameter included to account for the fact that some reaction had
occurred before the stopped-flow instrument recorded the first time. For
all concentrations of hydroxide, least-squares fits gave a time shift of about
0.003 s.

2360 J. Am. Chem. Soc., Vol. 119, No. 10, 1997
Barletta et al.
a methyl group instead of the 5-(4′-amino-2′-methylpyrimidyl)-
methyl group, and in place of a C(2R)-hydroxyl group, our
compounds have a methoxy group. Addition of NaOD gave a
clean thiazolium ring-opened product containing a mixture of
cis and trans (about the amide bond) isomers as identified by
1H-NMR. Furthermore, the ring-opening appears to be revers-
ible as was demonstrated by the addition of acid after ring
opening in the presence of base (spectrum C of Figure 1).
Our rate constants for the ring-opening reaction are in the
range of values reported for many related thiazolium-ring
hydrolyses. Washabaugh et al.29,34 and Knoche et al.35 list rate
constants from much of the older literature on thiazolium ring-
opening reactions. Their lists, along with rate constants from
recent work,36,37 show that rate constants for the hydroxide-
dependent reaction are 3-117 M-1 s-1 for a broad range of
substrates and experimental conditions. For 1h, we find k3 )
3.41 ( 09 M-1 s-1 from the slope of the line through k′3 in
Figure 4 (Table S1). The corresponding rate constant for 2 is
not as well determined, because the fits required more param-
eters (see the Results section).
Figure 6. Dependence of first-order rate constants on [NaOH] for 2.
Slopes and intercepts of the least-squares lines are shown at the bottom
of Table S3. The line through the k′3H data is a fit constrained to have
a zero intercept. Note that the k′2H data was multiplied by 100 for
plotting.
obtained for 2. Simulations of the Scheme 2 mechanism using
rate constants obtained from least-squares fits to progress curves
predict that the unknown product of the k4 pathway should be
a very minor product. For compound 2 in 0.3 M NaOH,
simulations predict 7.4% of the product should be produced from
the k4 pathway, while only 2.2% of the product should be the
unknown material in 0.6 M NaOH. We have been unable to
characterize what is predicted to be a minor product in the
scheme for 2.28 The additional decomposition pathway was not
needed to explain the results obtained with 1, which could be
studied at lower concentrations of hydroxide ion.
Hydroxide-Dependent Thiazolium Ring Opening. While
there have been many studies of the hydrolysis of thiamin-related
compounds, only a modest effort has been made to identify the
products of hydrolysis. The subject is often complicated by
the presence of the 4′-aminopyrimidine ring which can lead to
the formation of tricyclic thiamin (bond formation between the
nucleophilic amino group and the C(2) carbon),29,30 and in some
cases, the 4′-aminopyrimidine ring can be eliminated from
thiamin derivatives.31-33 Additional complications for product
studies are introduced in models with a C(2R)-hydroxyl group.
Elimination of the thiazolium ring to produce aldehydes, or as
was mentioned above, elimination of the 4′-aminopyrimidine
ring are important reactions under specific hydrolytic conditions
for compounds containing a C(2R)-hydroxyl group.
The mechanism of thiazolium ring opening is widely thought
to involve the intermediacy of the pseudobase form of thiazo-
lium ions (the product of hydroxide addition at C2), but there
are disparate conclusions about the rate-limiting step for the
reaction.33 Washabaugh et al.29,34,36 argue that their observations
are best explained using a mechanism with rate-limiting
decomposition of the pseudobase, while many earlier
researchers35,37-41 favored mechanisms involving rate-limiting
formation of the pseudobase. Our observations do not shed
significant new light on the controversy; they are consistent with
either rate-limiting pseudobase formation or rate-limiting pseudo-
base decomposition for the ring-opening step.
Proton Transfer from C(2r). Our results are consistent with
the classification of the 2-benzylthiazolium salts as typical
carbon acids characterized by rate-limiting proton transfer in
their acid/base behavior. It is clear from the vast qualitative
differences in the progress curves for 1h (Figure 2) and 1d
(Figure 3) that there must be significant isotope effects on one
or more of the rate constants of Scheme 2. Using the least-
squares slopes of k′ vs. [NaOH] isotope effects (H/D) of 6.1 (
0.8 and 1.2 ( 0.1 can be estimated for k2 and k3, respectively.
Given our uncertainty about ionic strength effects on the rate
constants, the linear fits through the data shown in Figure 4
may not be entirely appropriate for evaluating the second-order
rate constants k2 and k3. Ratios of rate constants as a function
of [NaOH] offer another measure of the isotope effects on our
system. From this analysis, we find that k′2H/k′2D is about 4
while k′3H/k′3D is about 0.7. We conclude that the isotope effect
on the second-order rate constant k2 is 4-6, while the isotope
effect on k3 is near unity. A kinetic isotope effect of 4-6 is
typical of isotope effects for proton transfers from carbon acids
to hydroxide ion.42
The compounds used in our studies do not have these
complications: the thiazolium-ring nitrogen is quaternized with
(28) Compound 10 (Scheme 1 and Table 1) was prepared for use in
attempts to identify unknown products generated in the presence of sodium
hydroxide. The only difference between 10 and 2, the compound for which
the k4 pathway was needed, is the 5-methyl group on the thiazolium ion
ring. The spectrum of 10 after adding enough NaOH to give a 0.1 M solution
was consistent with the ring-open product (last entry in Table 1). When
acid was added the spectrum was again consistent with 10. Based on
inspection of peak heights, we estimate that ca. 5% of the products were
unknowns. Some unknown side products formed when base was added,
and some remained or resulted from the addition of acid. We were unable
to make structural assignments for these minor side products. We observed
that in generating the enamine with base, cleaner spectra were obtained
when the parent compound has an electron withdrawing group on the phenyl
ring. This finding lends qualitative support to the conclusion that a
hydroxide-dependent pathway (k4) for enamine decomposition is needed
to explain the kinetics for 2 but not for 1.
(34) Washabaugh, M. W.; Yang, C. C.; Stivers, J. T.; Lee, K.-S. Bioorg.
Chem. 1992, 20, 296-312.
(35) Heiber-Langer, I.; Winter, I.; Knoche, W. J. Chem. Soc., Perkin
Trans. 2 1992, 1551-1557.
(36) Washabaugh, M. W.; Gold, M. A.; Yang, C. C. J. Am. Chem. Soc.
1995, 117, 7657-7664.
(37) Barrabass, S.; Heiber-Langer, I.; Knoche, W. J. Chem. Soc., Perkin
Trans. 2 1994, 131-134.
(29) Washabaugh, M. W.; Yang, C. C.; Hollenbach, A. D.; Chen, P.
Bioorg. Chem. 1993, 21, 170-191.
(38) Haake, P.; Duclos, J. M. Tetrahedron Lett. 1970, 6, 461-464.
(39) Duclos, J. M.; Haake, P. Biochemistry 1974, 13, 5358-5362.
(40) El Hage Chahine, J.-M.; Dubois, J.-E. J. Am. Chem. Soc. 1983,
1983, 2335-2340.
(30) Hopmann, R. F. W. Ann. N. Y. Acad. Sci. 1982, 378, 32-50.
(31) Kluger, R.; Lam, J. F.; Kim, C.-S. Bioorg. Chem. 1993, 21, 1-28.
(32) Kluger, R.; Lam, J. F.; Pezacki, J. P.; Yang, C.-M. J. Am. Chem.
Soc. 1995, 117, 11383-11389.
(41) Zoltewicz, J. A.; Uray, G. J. Org. Chem. 1980, 45, 2101-2108.
(42) Casamassina, T. E.; Huskey, W. P. J. Am. Chem. Soc. 1993, 115,
14-20.
(33) Zoltewicz, J. A.; Uray, G. Bioorg. Chem. 1994, 22, 1-28.

Kinetics of C(2R)-Proton Abstraction
J. Am. Chem. Soc., Vol. 119, No. 10, 1997 2361
Our results compare favorably with results obtained by
Washabaugh, Stivers, and Hickey43 using iodination initial rate
assays and a substrate similar in structure to ours. These authors
found k ) 2.9 M-1 s-1 for C(2R) proton removal from 2-(1-
hydroxy-1-phenylmethyl)oxythiamin by hydroxde ion; our cor-
responding rate constants are k2 ) 21 M-1 s-1 for 1h and 0.019
M-1 s-1 for 2. Washabaugh, Stivers, and Hickey concluded
that their reactions were rate-limited by proton transfer based
largely on their observation of kH/kD ) 1.5 and kH/kT ) 1.8 for
removal of the C(2R) by cacodylate ion (conjugate acid pKa )
6.2744). Our isotope effect is larger (kH/kD ) 4-6) suggesting
a more symmetrical transition-state structure for the proton
transfer to hydroxide ion (conjugate acid pKa ) 15.744). It is
interesting to note that, according to Washabaugh and Stiv-
ers,45,46 the rate-limiting step for C(2R) proton removal from
2-(1-hydroxyethyl)thiazolium ions by hydroxide ion or buffer
bases (isotope effects are near unity) is diffusional separation
of product rather than proton transfer. In the yeast pyruvate
decarboxylase catalyzed formation of acetoin, however, proton
transfer from 2-(1-hydroxyethyl)thiamin is rate limiting with
kH/kD ≈ 447,48 and coupled to carbonyl-addition (there is
evidence suggesting a large 14C isotope effect) according to the
mechanism of Stivers and Washabaugh.48
Additional insights into the nature of the proton transfer
reaction can be obtained though a comparison of compounds 1
and 2. The benzyl compound (2) reacted much more slowly
than did the p-trimethylammoniumbenzyl compound (1), but
our analysis shows that the apparent pKa’s of the two compounds
are very similar. Allowing for uncertainty in ionic strength
effects and ꢀE we used k′1H ) 300-540 s-1 (the intercept and
the highest hydroxide point in Figure 4) and k2H ) 18-28 M-1
s-1 (the range of k2H observed when ꢀE was varied (see the
Results section)) to estimate a pKa for 1h of 15.0-15.5 (using
Kw ) 10-14 M2).49 A similar analysis of the data in Figure 6
gives an estimate of 15.7 for the pKa of 2, in agreement with
an estimate of 15 ( 1 reported for 2-(1-hydroxybenzyl)-
oxythiamin.43 Thus the equilibrium constants for proton
removal to generate the enamines are very similar (they may
differ by about one pK unitsone order of magnitude in K). The
rate constants k2 differ by three orders of magnitude. The slope
of k′2H vs. [NaOH] in Figure 4 for compound 1h is 21 M-1
s-1, while the same slope for 2 is only 0.019 M-1 s-1. The
difference in sensitivity of the equilibrium and rate constants
to substitution in the benzyl ring may reflect another feature of
carbon acids with resonance-stabilized conjugate basesstransition-
state “imbalance” in accord with Bernasconi’s principle of
imperfect synchronization.50-53 The rate-limiting transition-state
may experience very little resonance stabilization from the
thiazolium nitrogen and thereby be subject to substantial
substituent effects. The relative insensitivity of the equilibrium
constants to substitution may reflect a very enamine-like (as
opposed to carbanion-like) conjugate base.
Protonation of the Enamine. Among the rate constants
determined by us, k1 (Scheme 2), for the protonation of the
enamine by water is the rate constant most relevant to catalysis
by benzoylformate decarboxylase. Protonation of the enamine
is very likely an essential step in the enzymatic reaction
following decarboxylation of the intermediate 2-(1-carboxy-1-
hydroxy-1-phenylmethyl)thiamin. We found k1 to be 1 s-1 for
the enamine of 2-(1-hydroxy-1-phenylmethyl)-3,4-dimethylthi-
azolium ion (2). Dividing by 55.5 M, to represent the
concentration of the proton donor water gives an estimated
second-order rate constant of 0.018 M-1 s-1. The enzyme must
protonate the enamine with a rate constant greater than kcat (81
s-1).54 The ratio of these two numbers gives a minimum
effective molarity55 of 4500 M for the enzyme in this proton-
transfer process. Effective molarities have been estimated to
be greater than 60 000 M in carefully-designed systems involv-
ing intramolecular proton transfers to carbon,56 but values of
1-10 M are typical of less sophisticated nonenzymic intramo-
lecular systems involving general acid-base catalysis.57 Our
estimate of a minimum value of 4500 M for the proton transfer
in benzoylformate decarboxylase catalysis is sufficiently large
to suggest that the enzyme does not allow the enamine to be
protonated by bulk solvent. Instead, an acidic residue must be
poised to effect catalysis of this proton-transfer step.
Experimental Section
Most chemicals were purchased from Aldrich Chemicals, Inc. and
were used without further purification, unless specified otherwise below.
4-Methylthiazole was purchased from Pyrazine Specialties. Syntheses
of compounds similar to 1-6 have been previously reported.16,58,59 In
brief, the compounds used in the present report were prepared by
addition of a thiazole to an aldehyde, followed by methylation using
of the thiazole nitrogen and the C(2R) hydroxyl group to yield
fluoroborate salts of substituted thiazolium ions. The syntheses and
analyses of 2h, 4, and 6 used in this study have been published.60
2-[1-Hydroxy-1-(p-dimethylaminophenyl)methyl]-4-methyl-
thiazole (7). Compound 7 was prepared using p-dimethylaminoben-
zaldehyde in place of benzaldehyde according to the procedure
published for 2-(1-hydroxy-1-phenylmethyl)-4-methylthiazole.60 1H
NMR (200 MHz, DMSO-d6-DMSO): d 7.17 (d, 2H, J ) 8.8 Hz,
Ar), 6.65 (d, 2H, J ) 8.8 Hz, Ar), 7.07 (s, 1H, C5-H), 6.39(s, 1H,
C2a-H), 5.72 (br, 1H, OH), 2.83 (s, 6H, Ar-N-Me2), 2.25 (s, 3H, C4-
CH3).
2-[1-Methoxy-1-(p-dimethylaminophenyl)methyl]-4-methyl-
thiazole (8). To a stirred solution of 7 (10.9 mmol) in 35 mL of dry
THF at 0 °C under Ar was added NaH (0.43 g, 14.2 mmol) in one
portion. The mixture was stirred for 30 min, and then a solution of
CH3I (6.2 g, 43.7 mmol) in THF was added in two portions (half was
added 1 h after the first portion). After 2 h, the reaction was quenched
with 20 mL of water:ethanol (1:1) and concentrated using a rotary
evaporator to afford an oil which was extracted with diethyl ether and
water. The organic layers were dried (Na2SO4) and concentrated to
yield. The crude product was purified using a column of silica gel
eluted with ether-hexanes (2:1). The product oil (1.0 g) was obtained
in 64% yield. 1H NMR (200 MHz, CDCl3-CHCl3): δ 7.29 (d, 2H, J
(43) Washabaugh, M. W.; Stivers, J. T.; Hickey, K. A. J. Am. Chem.
Soc. 1994, 116, 7094-7097.
(44) Jencks, W. P.; Regenstein, J. In Handbook of Biochemistry and
Molecular Biology (Physical and Chemical Data), 3rd ed.; Fasman, G. D.,
Ed.; CRC Press: Cleveland, 1976; pp 305-351.
(45) Stivers, J. T.; Washabaugh, M. W. Bioorg. Chem. 1991, 19, 369-
383.
(46) Stivers, J. T.; Washabaugh, M. W. Bioorg. Chem. 1992, 20, 155-
(54) Reynolds, L. J.; Garcia, G. A.; Kozarich, J. W.; Kenyon, G. L.
Biochemistry 1988, 27, 5530-5538.
172.
(47) Chen, G. C.; Jordan, F. Biochemistry 1984, 23, 3576-3582.
(48) Stivers, J. T.; Washabaugh, M. W. Biochemistry 1993, 32, 13472-
13482.
(55) Kirby, A. J. AdV. Phys. Org. Chem. 1980, 17, 183-278.
(56) Kirby, A. J.; O’Carroll, F. J. Chem. Soc., Perkin Trans. 2 1994,
649-654.
(49) We estimated this pKa to be 15.4 in our previous report17 in which
a slightly different fitting procedure27 was used, and we used the average
value of k′1H over the range of values obtained at various concentrations of
hydroxide.
(57) Kirby, A. J.; Williams, N. H. J. Chem. Soc., Perkin Trans. 2 1994,
643-648.
(58) Rastetter, W. H.; Adams, J.; Frost, J. W.; Nummney, L. W.;
Frommer, J. E.; Roberts, K. B. J. Am. Chem. Soc. 1979, 101, 2752-2753.
(59) Bordwell, F. G.; Satish, A. V.; Jordan, F.; Rios, C. B.; Chung, A.
C. J. Am. Chem. Soc. 1990, 112, 792-797.
(50) Bernasconi, C. F. Tetrahedron 1985, 41, 3219-3234.
(51) Bernasconi, C. F. Acc. Chem. Res. 1987, 20, 301-308.
(52) Bernasconi, C. F. AdV. Phys. Org. Chem. 1992, 27, 119-238.
(53) Bernasconi, C. F. Acc. Chem. Res. 1992, 25, 9-16.
(60) Barletta, G.; Chung, A. C.; Rios, C. B.; Jordan, F.; Schlegel, J. M.
J. Am. Chem. Soc. 1990, 112, 8144-8149.

2362 J. Am. Chem. Soc., Vol. 119, No. 10, 1997
Barletta et al.
) 8.9 Hz, Ar), 6.69 (d, 2H, J ) 8.9 Hz, Ar), 6.78 (s, 1H, C5-H, 5.42
(s, 1H, C2R-H), 3.39 (s, 3H, OCH3), 2.91 (s, 6H, Ar-N-Me2), 2.38 (s,
3H, C4-CH3).
2-[1-Methoxy-1-(p-chlorophenyl)methyl-3,4,5-trimethylthiazoli-
um Fluoroborate (5). Following the same procedures described for
the preparation of 7, 8, and 2h, 1.14 g (80%) of 5 was prepared from
p-chlorobenzaldehyde and 4,5-dimethylthiazole. 1H NMR (200 MHz,
D2O-H2O): δ 7.40 (d, 2H, J ) 8.6 Hz, Ar), 7.30 (d, 2H, J ) 8.6 Hz,
Ar), 5.86 (s, 1H, C(2RH)), 3.62 (s, 3H, N-CH3), 3.37 (s, 3H, OCH3),
2.33 (s, 3H, C4-CH3), 2.23 (s, 3H, C5-CH3).
2-[1-Methoxy-1-(p-trimethylammoniumphenyl)methyl-3,4-di-
methylthiazolium Fluoroborate (1h). Using the procedure described
for the synthesis of 2h,60 compound 8 was converted into a fluoroborate
salt using 2 equiv of (CH3)3OBF4. The salt was obtained as a white
powder (0.95 g, 60%) which was purified on a cellulose plate eluted
with CH3CN. 1H NMR (200 MHz, D2O-H2O): δ 7.80 (d, 2H, J )
8.0 Hz, Ar), 7.62 (s, 1H, C5-CH), 7.57 (d, 2H, J ) 8.0 Hz, Ar), 5.94
(s, 1H, C2R-H), 3.62 (s, 3H, N-CH3), 3.48 (s, 9H, ArN+-Me3), 3.35 (s,
3H, OCH3), 2.31 (s, 3H, C4-CH3). Anal. Calcd for
C16H24N2OS‚[BF4]2‚0.75H2O: C, 40.07; H, 5.30; N, 5.85; S, 6.68.
Found: C, 39.99; H, 5.27; N, 5.87; S, 6.96.
2-(1-Hydroxy-1-phenylmethyl)-4,5-dimethylthiazole (11). To a
stirred solution of 4,5-dimethylthiazole (3.5 g, 13.2 mmol) in 50 mL
of dry THF at -78 °C n-BuLi (16.5 mmol) was added dropwise under
N2. The mixture was stirred for 1 h, and then a solution of
benzaldehyde (40 mmol) in THF was added. After stirring the mixture
at -78 °C for 30 min, the reaction temperature was allowed to rise to
room temperature, and the reaction proceeded for another 45 min. The
reaction was then quenched with water (40 mL) and extracted with
ethyl ether (3 × 20 mL). The organic layer was dried (MgSO4) and
concentrated. The crude product was purified using a silica gel column
eluted with ethyl acetate-petroleum ether (1:3). After crystallization,
the product was obtained as a pale yellow solid in 41% yield.
2-(1-Methoxy-1-phenylmethyl)-4,5-dimethylthiazole (12). To a
stirred solution of 11 (1.5 g, 6.8 mmol) in 30 mL of dry THF was
added NaH (0.43 g, 18 mmol) at 0 °C under N2. The mixture was
stirred for 30 min at 0 °C, and then CH3I (2.0 g, 13.9 mmol) was added
dropwise at room temperature. After 3-4 h, the reaction was quenched
with 20 mL water and extracted with ethyl ether (3 × 15 mL). The
organic layer was dried (MgSO4) and concentrated. The crude product
was purified on a silica gel column eluted with ethyl acetate-petroleum
ether (3:7). A crystalline product was obtained in 83% yield. 1H NMR
(500 MHz, CDCl3/TMS): δ 7.45 (d, 2H, Ar), 7.35 (t, 2H, Ar), 7.30
(m, 1H, Ar), 5.5 (s, 1H, C(2RH)), 3.42 (s, 3 H, O-CH3), 2.30 (s, 3H,
2.29 C4-CH3 or C5-CH3) 2.29 (s, 3H, C4-CH3 or C5-CH3).
2-(1-Methoxy-1-phenylmethyl)-3,4,5-trimethylthiazolium Tri-
fluoromethanesulfonate (10). To a stirred solution of 12 (0.22 g, 0.94
mmol) in 10 mL of dry CH2Cl2 was added methyl trifluoromethane-
sulfonate (0.15 g, 0.94 mmol) at 0 °C under N2. The mixture was
stirred for 0 °C for 30 min, then the temperature was allowed to rise,
and the reaction was stirred for another 3 h at room temperature. The
reaction mixture was concentrated and then purified on a silica gel
column with methanol-methylene chloride (1:9). The product was
obtained as an oil (0.32 g) in 86% yield. 1H NMR (500 MHz, D2O):
δ 7.55 (m, 3H, Ar), 7.45 (m, 2H, Ar), 5.98 (s, 1H, C(2RH)), 3.72 (s,
3H, N-CH3), 3.48 (s, 3H, O-CH3), 2.46 (s, 3H, C5-CH3), 2.31, 3H, C4-
CH3).
2-(1-Deutero-1-methoxy-1-phenylmethyl)-3,4-dimethylthiazo-
lium Fluoroborate (2d). The same series of reactions used to prepare
2h60 was repeated using [C7-2H]benzaldehyde in place of benzaldehyde.
In 38% overall yield, 1.14 g of 2d was obtained as a yellow oil. 1H
NMR (200 MHz, D2O-H2O): δ 7.62 (s, 1H, C5-H), 7.41-7.35 (m,
5H, Ar), 3.61 (s, 3H, N-CH3), 3.37 (s, 3H, OCH3), 2.33 (s, 3H, C4-
CH3). Anal. Calcd for C13H15NOS‚BF4D‚0.5H2O: C, 47.15; H, 4.87;
N, 4.23; S, 9.68. Found: C, 47.32; H, 5.00; N, 4.15; S, 9.89.
p-Dimethylamino-[C7-2H]benzaldehyde (9). Compound 9 was
prepared from the reduction of an acid chloride based on related
literature methods.61,62 To a 0.023 mol (3.8 g) solution of p-
dimethylaminobenzoic acid in 50 mL of anhydrous CH2Cl2, ClCOCOCl
(5.84 g, 0.046 mol) was added dropwise under argon gas over a 30
min period. After the solution was allowed to reflux for 2 h, the solvent
was removed leaving a green residue containing the acid chloride
product. The acid chloride was reduced using LiAl(t-BuO)3D prepared,
according to standard methods,63 from LiAlD4 and t-BuOH. Over a 1
h period, a 100 mL diglyme solution containing 0.023 mol of LiAl(t-
BuO)3 was added to a diglyme solution of the acid chloride (4.2 g,
0.023 mol) at -78 °C. 1H NMR (200 MHz, CDCl3-CHCl3): δ 7.72
(d, 2H, J ) 8.9 Hz, Ar), 6.68 (d, 2H, J ) 8.8 Hz, Ar), 3.07 (s, 6H,
N(CH3)2.
2-[1-Deutero-1-methoxy-1-(p-trimethylammoniumphenyl)methyl]-
3,4-dimethylthiazolium fluoroborate (1d). Using p-dimethylamino-
[C7-2H]-benzaldehyde (9), compound 1d was prepared using the
procedures described above for 1h. The product was obtained as an
oil which was purified by thin-layer chromatography using a cellulose
plate eluted with EtAc-CH3CN (9:1). The product was washed out
of the cellulose with CH3CN. After removing the solvent using rotary
evaporation, 0.9 g of 1d was obtained in 51% overall yield from 11.
1H NMR (200 MHz, D2O-H2O): δ 7.82 (d, 2H, J ) 8.9 Hz, Ar), 7.64
(s, 1H, C5-H), 7.60 (d, 2H, J ) 8.8 Hz, Ar), 3.65 (s, 3H, N-CH3), 3.51
(s, 9H, Ar-N(CH3)2), 3.37 (s, 3H, OCH3), 2.34 (s, 3H, C4-CH3). Anal.
Calcd for C16H23N2OSB2F8D‚2H2O: C, 38.20; H, 5.41; N, 5.57; S, 6.37.
Found: C, 38.31; H, 5.42; N, 5.76; S, 6.65.
Kinetics. A Hi-Tech model stopped-flow instrument was used to
monitor the enamine absorbance produced in the reaction of 1 (410
nm) or 2 (380 nm) with hydroxide ion at 25 °C. The stopping syringe
was adjusted to collect 250 mL (125 mL each of substrate and NaOH
solutions). A 1 cm path-length configuration was used. Distilled,
deionized, and degassed water was used to prepare reaction solutions.
Concentrations of NaOH solutions were determined by titration to
phenolphthalein end points using potassium hydrogen phthalate.
Syringe concentrations of the substrate solutions were 2.3 mM (1h),
3.1 mM (1d), 6.8 mM (2h), and 2.2 mM (2d). Progress curves were
analyzed as described in the Results section.
2-[1-Methoxy-1-(p-chlorophenyl)methyl-3,4-dimethylthiazoli-
um Fluoroborate (3). Following the same procedures described for
the preparation of 7, 8, and 2h, 0.80 g (70%) of 3 was prepared from
p-chlorobenzaldehyde. 1H NMR (200 MHz, D2O-H2O): δ 7.55 (s,
1H, C5-H), 7.35 (d, 2H, J ) 8 Hz, Ar), 7.26 (d, 2H, J ) 8 Hz, Ar),
5.82 (s, 1H, C(2RH)), 3.58 (s, 3 H, N-CH3), 3.32 (s, 3H, OCH3), 2.29
(s, 3 H, C4-CH3).
Acknowledgment. This work was supported by Grants NSF-
CHE 86-17087 (F.J.), NIH-GM50380 (F.J.), NIH-MBRS at
Rutgers-Newark (Barry Komisaruk, P.I.), the Donors of the
Petrolum Research Fund, administered by the American Chemi-
cal Society (W.P.H.), and the Rutgers University Busch
Biomedical Fund. Dedicated to Professor Hugh W. Thompson,
a valued colleague, on the occasion of his 60th birthday.
(61) Brown, H. C.; Subba Rao, B. C. J. Am. Chem. Soc. 1958, 80, 5377-
5380.
Supporting Information Available: Tables S1, S2, and S3
containing parameters obtained from the least-squares fits of
each of the progress curves obtained at various [NaOH] for
compounds 1h, 1d, and 2. Also included are Figures S1a, S1b,
(62) Schnettler, R. A.; Dage, R. C.; Paiopoli, F. P. J. Med. Chem. 1986,
29, 860-862.
(63) Brown, H. C.; McFarlin, R. F. J. Am. Chem. Soc. 1958, 80, 5372-
5376.
(64) We cannot identify the peak at 2.4 ppm; it may arise from an
impurity in the DCl solution used to acidify the sample, or it may be due
to an unknown product formed once the acid is added. The latter possibility
is considered unlikely since only a single unknown peak was observed.
The remainder of the high-intensity peaks match the peaks seen in panel
A.
1
S1c, and S2 containing H NMR spectra of compound 10 in
D2O, NaOD, and DCl (6 pages). See any current masthead
page for ordering and Internet access instructions.
JA9633528