Destruction of vitamin B1 by benzaldehyde. Reactivity of intermediates in the fragmentation of N1'-benzyl-2-(1- hydroxybenzyl)thiamin

Article abstract of DOI:10.1021/ja000194i

Thiamin (vitamin B 1) combines with benzaldehyde in alkaline solutions to form 2-(1-hydroxybenzyl)thiamin (HBzT), a reactive intermediate in the thiamin-catalyzed benzoin condensation. In neutral solutions, HBzT fragments into pyrimidine and thiazole constituents by cleavage of the bridging methylene-thiazole bond. The fragmentation is promoted by protonation of the pyrimidine moiety of HBzT. The N1'-benzyl derivative of HBzT (4 in Scheme 4, BHT) also undergoes fragmentation in neutral and alkaline solutions, consistent with fragmentation being driven by positive charge on the pyrimidine derived from thiamin. Anionic Bronsted bases catalyze the reaction (β = 0.5, for a series of substituted acetates). The dependence of the observed first-order rate coefficient for fragmentation of BHT on buffer concentration is nonlinear, becoming buffer-independent at concentrations above 0.05 M. This is consistent with a change in rate-determining step with buffer concentration from proton removal to subsequent fragmentation of the conjugate base of BHT. The solvent isotope effect is inverse, also consistent with reversible formation of the conjugate base. Analysis of the kinetic data reveals that the fragmentation step is very fast (k(f) = 1.2 x 10⁵ s^-1 at 40 °C). Such a low barrier is consistent with electron-shift mechanisms for the fragmentation step.

Full text of DOI:10.1021/ja000194i

J. Am. Chem. Soc. 2000, 122, 6145-6150
6145
Destruction of Vitamin B1 by Benzaldehyde. Reactivity of
Intermediates in the Fragmentation of
N1′-Benzyl-2-(1-hydroxybenzyl)thiamin
Ronald Kluger* and Ian F. Moore
Contribution from the Lash Miller Laboratories, Department of Chemistry, UniVersity of Toronto,
Toronto, Canada M5S 3H6
ReceiVed January 18, 2000
Abstract: Thiamin (vitamin B1) combines with benzaldehyde in alkaline solutions to form 2-(1-hydroxybenzyl)-
thiamin (HBzT), a reactive intermediate in the thiamin-catalyzed benzoin condensation. In neutral solutions,
HBzT fragments into pyrimidine and thiazole constituents by cleavage of the bridging methylene-thiazole
bond. The fragmentation is promoted by protonation of the pyrimidine moiety of HBzT. The N1′-benzyl
derivative of HBzT (4 in Scheme 4, BHT) also undergoes fragmentation in neutral and alkaline solutions,
consistent with fragmentation being driven by positive charge on the pyrimidine derived from thiamin. Anionic
Brønsted bases catalyze the reaction (â ) 0.5, for a series of substituted acetates). The dependence of the
observed first-order rate coefficient for fragmentation of BHT on buffer concentration is nonlinear, becoming
buffer-independent at concentrations above 0.05 M. This is consistent with a change in rate-determining step
with buffer concentration from proton removal to subsequent fragmentation of the conjugate base of BHT.
The solvent isotope effect is inverse, also consistent with reversible formation of the conjugate base. Analysis
5
-1
of the kinetic data reveals that the fragmentation step is very fast (kf ) 1.2 × 10 s at 40 °C). Such a low
barrier is consistent with electron-shift mechanisms for the fragmentation step.
Thiamin (vitamin B1) can be used as an alternative to cyanide
mate decarboxylase from the conjugate of benzoylformate and
thiamin diphosphate by loss of carbon dioxide (Scheme 2).
1
-4
8,9
to catalyze the formation of benzoin from benzaldehyde.
5
Breslow’s mechanistic analysis of thiamin reactions revealed
that catalysis involves base-catalyzed addition of the C-2
conjugate base of thiamin (an ylide) to benzaldehyde, producing
the intermediate 2-(1-hydroxybenzyl)thiamin, HBzT (1 in
Scheme 1).
Jordan has been able to generate and observe enamines derived
1
0-14
from thiamin adducts.
Recent work has shown that there are complications in the
reaction possibilities of HBzT. In neutral and acidic solutions,
HBzT does not revert to thiamin and benzaldehyde but
fragments irreversibly into thiazole and pyrimidine derivatives
Loss of the C2R proton (derived from benzaldehyde) from
HBzT produces an enamine, which is functionally equivalent
to a benzoyl carbanion. This adds to the carbonyl group of a
second molecule of benzaldehyde, leading to formation of
benzoin by elimination of thiamin.
1
5,16
(Scheme 3).
We have been studying the mechanism of the fragmentation
reaction and the factors that divert the reaction from the benzoin
1
7
condensation pathway. A key observation is that the rate of
fragmentation of HBzT is proportional to the amount of material
that is present as the conjugate acid (protonated at N1′ of the
pyrimidine of HBzT). The effect of the positive charge on the
pyrimidine ring is consistent with development of carbanionic
character at the cleavage site. The observation that the N1′-
methyl(pyrimidinium) derivative of HBzT undergoes base-
catalyzed fragmentation and no elimination demonstrates the
Mieyal and Sable established the existence of the enamine
derived from HBzT by observing that the C2R proton of HBzT
6,7
is exchanged for a deuteron in basic deuterium oxide. Similar
intermediates also occur in enzymic reactions that involve
thiamin diphosphate as a cofactor. For example, the C2R
enamine of the diphosphate of HBzT is formed by benzoylfor-
(
1) Gould, E. S. Structure and Mechanism in Organic Chemistry; Holt:
New York, 1959; pp 394-397.
2) Ugai, T.; Tanaka, S.; Dokawa, S. J. Pharm. Soc. Jpn. 1943, 63, 269-
72.
3) Mohrig, J. R.; Hammond, C. N.; Morrill, T. C.; Neckers, D. C.
Experimental Organic Chemistry; Freeman: New York, 1998; pp 419-
27.
4) Mohrig, J. R.; Neckers, D. C. Laboratory Experiments in Organic
Chemistry, 2nd ed.; Van Nostrand: New York, 1973; p 868.
1
6
profound effect of the localized charge.
(
Kinetic patterns of the fragmentation reaction implicate initial
removal of the proton at C2R to form the enamine intermediate,
as in the benzoin condensation. Exchange of the C2R proton in
2
(
4
(10) Kuo, D. J.; Jordan, F. J. Biol. Chem. 1983, 258, 13415-13417.
(11) Jordan, F.; Kudzin, Z. H.; Rios, C. B. J. Am. Chem. Soc. 1987,
109, 4415-4416.
(
(
5) Breslow, R. J. Am. Chem. Soc. 1958, 80, 3719-3726.
(12) Barletta, G.; Huskey, W. P.; Jordan, F. J. Am. Chem. Soc. 1992,
114, 7607-7608.
(6) Mieyal, J. J.; Bantle, G.; Votaw, R. G.; Rosner, I. A.; Sable, H. Z. J.
Biol. Chem. 1971, 246, 5213.
7) Mieyal, J. J.; Votaw, R. W.; Krampitz, L. O.; Sable, H. Z. Biochim.
Biophys. Acta 1967, 141, 205-208.
8) Dirmaier, L. J.; Garcia, G. A.; Kozarich, J. W.; Kenyon, G. L. J.
Am. Chem. Soc. 1986, 108, 3149.
9) Hasson, M. S.; Muscate, A.; McLeish, M. J.; Polovnikova, L. S.;
Gerlt, J. A.; Kenyon, G. L.; Petsko, G. A.; Ringe, D. Biochemistry 1998,
7, 9918-9930.
(13) Jordan, F.; Li, H.; Brown, A. Biochemistry 1999, 38, 6369-6373.
(14) Jordan, F.; Adams, J.; Farzami, B.; Kudzin, Z. H. J Enzym. Inhib.
1986, 1, 139-149.
(
(
(15) Kluger, R.; Lam, J. F.; Kim, C.-S. Bioorg. Chem. 1993, 21, 275-
283.
(
(16) Kluger, R.; Lam, J. F.; Pezacki, J. P.; Yang, C.-M. J. Am. Chem.
Soc. 1995, 117, 11383-11389.
(17) Kluger, R. Pure Appl. Chem. 1997, 69, 1957-1967.
3
1
0.1021/ja000194i CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/13/2000

6146 J. Am. Chem. Soc., Vol. 122, No. 26, 2000
Kluger and Moore
Scheme 1
Scheme 2
to 0.15 M). The pD readings were obtained by adding 0.4 to the
observed value on the meter.19 The values of acidity constants for
buffers were obtained from published tables.20
The rate of exchange of the proton for a deuteron at C2R of BHT
1
was measured by the decrease of the integrated singlet in the H NMR
1
at δ 6.6. H NMR was also used to follow the fragmentation of BHT
as the decrease in the integrated peak at δ 5.9 (due to C6′ H of the
pyrimidinium moiety, which is shifted in 6). The chemical shifts of
the C2R and C6′ protons are solvent dependent and assignments were
confirmed by HSQC NMR. Integrated areas for the signals of the C2R
and C6′ protons were measured in triplicate and compared to the area
of the same peaks in the initial spectrum. Data were collected over at
least 5 half-lives. First-order rate constants were determined from the
slopes of linear regression fits of the data to the integrated rate law.
The temperature of the instrument’s probe was set at 40.0 °C with
calibration using ethylene glycol prior to each run.
deuterium oxide is faster than the fragmentation process, yet
1
6
anionic Brønsted bases catalyze the fragmentation. This is
kinetically equivalent to fragmentation from the enamine
intermediate being catalyzed by Brønsted acids. However, the
role for an acid catalyst is not apparent. Thus, we have
undertaken further investigations into the reactivity of interme-
diates in the fragmentation and the roles of catalysts. For these
Synthesis of N1′-Benzyl-2-(1-hydroxybenzyl)thiamin (4). 2-(1-
Hydroxybenzyl)thiamin chloride hydrochloride (1) was prepared by
condensation of benzaldehyde and thiamin.7
,21-23
The product was
suspended in ethanol and 1 equiv of sodium hydroxide was added (0.50
M). The solvent was removed under vacuum and the resulting white
solid was lyophilized. The dry solid (2.6 g) was dissolved in 6 mL of
anhydrous DMSO. The reaction flask was purged with dry nitrogen
and benzyl bromide (0.90 mL, 1.2 equiv) was added. The reaction was
kept in the dark at room temperature. After 4 days, the reaction mixture
was washed with ether and the DMSO layer was separated. The product
was precipitated from DMSO by the addition of acetone (15 mL)
followed by ether (40 mL). Excess solvent was decanted and residual
solvent was removed under high vacuum. The crude product was
recrystallized twice from distilled water and converted to the bis-
perchlorate salt by the addition of aqueous sodium perchlorate (0.75
1
8
studies we synthesized the N1′-benzyl derivative of HBzT,
BHT (4, Scheme 4). We have examined the kinetics and reaction
patterns of the C2R conjugate base (enamine) of BHT (5) over
a range of conditions. The results show that the identity of the
rate-determining step changes with buffer concentration and that
there is a very low barrier to fragmentation from the enamine
of BHT.
g/mL). BHT (0.96 g, 23%) was recovered as a pale yellow solid with
Experimental Section
1
mp 172-176 °C dec. H NMR (500 MHz, DMSO-d
6
): δ 9.5 (s, 1H),
Methods. (a) Kinetics of Reactions in Water. The rate of
fragmentation of BHT to 2-benzoyl-4-methyl-5-(2-hydroxyethyl)-1,3-
thiazole (2 in Scheme 4) and 4-amino-1-benzyl-2,5-dimethyl-1,3-
pyrimidine (6 in Scheme 4) was monitored with use of buffer solutions
maintained at 40.0 °C in a jacketed beaker. The appearance of 2 was
followed at 328 nm (ꢀ ) 10 000). The pH electrode was standardized
against reference buffers at 40 °C. The ionic strength of all reaction
solutions was maintained at 0.10 or 1.0 by the addition of sodium
chloride or potassium chloride. Data were collected with an interfaced
computer. First-order rate constants were calculated from nonlinear
regression fitting of the data to the integrated first-order rate expression.
Once the kinetic order was determined for a set of conditions, the data
for additional sets of reactions were acquired over a relatively short
interval and analyzed by the method of initial rates.
8
5
.7 (s, 1H), 7.7 (s, 1H), 7.4-7.1 (m, 10H), 6.9 (s, 1H), 6.3 (s, 1H),
.37 (dd, 2H, J ) 18 Hz), 5.22 (dd, 2H, J ) 16 Hz), 3.7 (m, 2H), 3.0
13
6
(m, 2H), 2.4 (s, 3H), 2.3 (s, 3H). C NMR (125 MHz, DMSO-d ): δ
1
1
1
1
77.7, 161.4, 160.0, 143.1, 141.7, 137.9, 134.4, 133.6, 129.3, 129.2,
29.0, 128.7, 127.8, 127.3, 109.1, 70.0, 59.5, 56.5, 46.5, 29.6, 21.4,
13
1.3. C DEPT NMR (125 MHz, DMSO-d ): δ 141.7 (+), 129.3 (+),
6
29.2 (+), 129.0 (+), 128.7 (+), 127.8 (+), 127.3 (+), 70.0 (+), 59.5
(-), 56.5 (-), 46.5 (-), 29.6 (-), 21.4 (+), 11.3 (+). FABMS (high
2+
30 4 2
resolution) [C26H N O S] : calcd 462.2092; found 462.2057
Preparative Scale Fragmentation of BHT. A larger scale reaction
of BHT was conducted under conditions similar to those of the kinetic
(
(
19) Glasoe, P. K.; Long, F. A. J. Phys. Chem. 1960, 64, 188-190.
20) Sober, H. A. CRC Handbook of biochemistry; selected data for
molecular biology; CRC Inc.: Cleveland, 1968.
21) Crane, E. J. I.; Washabaugh, M. W. Bioorg. Chem. 1991, 19, 351-
68.
(22) Oka, Y.; Kishimoto, S.; Hirano, H. Chem. Pharm. Bull. 1970, 18,
a
(b) Kinetics in Deuterium Oxide. The pK ′ of potassium dideute-
(
rium orthophosphate in deuterium oxide (40 °C, µ ) 1.0) was obtained
from potentiometric measurements with a glass electrode by using a
pH meter, at several buffer ratios (total buffer concentration from 0.10
3
527-533.
(23) Oka, Y.; Kishimoto, S.; Hirano, H. Chem. Pharm. Bull. 1970, 18,
(18) Karimian, K.; Mohanazadeh, F. Synthesis 1986, 1065-1067.
534-541.

Destruction of Vitamin B1 by Benzaldehyde
J. Am. Chem. Soc., Vol. 122, No. 26, 2000 6147
Scheme 3
Scheme 4
Table 1. Observed First-Order Rate Constants for Fragmentation
of BHT in Dilute Buffer
studies to isolate and identify the products. The bis-perchlorate of BHT
(0.095 g) was dissolved in 7 mL of water, final pH 8.0. The solution
k0 (s-1)
kB- (M-1 s-1
)
acidity was maintained by addition of 0.5 M potassium hydroxide from
an automated buret. After 1 h the reaction was stopped by addition of
hydrochloric acid (1 M) to give an acid concentration in the sample of
buffer
pH [buffer], M
-6
-5
acetate
5.00 0.01-0.1
(3.0 ( 0.2) × 10
(2.3 ( 0.3) × 10
succinate 5.90 0.004-0.04 (2.22 ( 0.04) × 10-5 (2.7 ( 0.1) × 10-4
-
-
-
4
4
4
-3
-3
-3
0
.01 M. The product was extracted with dichloromethane and dried
phosphate 6.40 0.008-0.08 (1.2 ( 0.1) × 10
6.90 0.005-0.04 (2.4 ( 0.1) × 10
(1.1 ( 0.2) × 10
(3.3 ( 0.2) × 10
(5.8 ( 0.8) × 10
3
over magnesium sulfate. After the solvent was removed, the yellow
oil was purified by flash chromatography on silica gel eluting with a
7
.40 0.005-0.04 (8.4 ( 0.4) × 10
-
HEPES
POPSO
bicine
7.60 0.01-0.1
8.10 0.004-0.04 (3.2 ( 0.3) × 10
(1.09 ( 0.07) × 10
1
:1 combination of hexanes and ethyl acetate (R
f
) 0.28) and
-
-
3
2
concentrated (0.03 g, yellow oil). This was identified as the ketone (2)
formed in the fragmentation of HBzT.
CDCl
6
8.60 0.01-0.1
(1.1 ( 0.1) × 10
1
5,16,23 1
H NMR (300 MHz,
): δ 8.41-8.37 (m, 2H), 7.62-7.45 (m, 3H), 3.87 (t, 2H, J )
3
1
3
.4 Hz), 3.06 (t, 2H, J ) 6.4 Hz), 2.47 (s, 3H). C NMR (50 MHz,
positive charge on the pyrimidine in promoting the fragmenta-
tion process. Analysis of the dependence of rate on pH shows
that the reaction is first order in hydroxide ion concentration
CDCl ): δ 184.0, 163.7, 152.2, 137.2, 135.5, 133.3, 131.0, 128.3, 62.7,
3
3
0.3, 15.3.
The aqueous layer was freeze-dried to yield 0.065 g of a white
powder, which was identified as a mixture of N1′-benzyl-2,5-dimethyl-
(eight points between pH 5 and 8.5, r ) 0.99, Table 1). The
1
5,16,22
1
fragmentation reaction requires the loss of the C2R proton
(derived from the aldehyde group on benzaldehyde). As with
N1′-methyl HBzT, at the relatively high buffer concentrations
necessary for NMR kinetic analysis in deuterium oxide, we
observe that the C2R proton exchanges from BHT faster than
the overall fragmentation occurs. For reactions followed by UV
spectroscopy (with more dilute buffer solutions in water) the
basic components of anionic buffers accelerate the fragmentation
process (Table 1). Analysis of the dependence of the observed
rate on the buffer ratio shows that only the basic component is
catalytically active. A Brønsted plot of the kinetic data for dilute
formate and substituted acetate buffers (Figure 1) gives a good
fit with a slope of Brønsted â ) 0.5.
4
-aminopyrimidine
(6) and potassium chloride. H NMR (400
MHz, DMSO-d
6
): δ 9.14 (s, 1H), 8.52 (s, 1H), 8.35 (s, 1H), 7.45-
1
3
7
.27 (m, 5H), 5.39 (s, 2H), 2.51 (s, 3H), 2.08 (s, 3H). C NMR (100
MHz, DMSO-d ): δ 163.3, 160.5, 145.3, 134.4, 129.1, 128.4, 127.1,
6
1
13.5, 56.5, 21.5, 13.4.
Results
The preparation of BHT was accomplished by reaction of
HBzT with benzyl bromide in DMSO, following the general
18
procedure of Karimian. Consistent with the reaction patterns
seen for N1-methyl HBzT,16 BHT undergoes fragmentation in
alkaline solutions (HBzT reverts to thiamin and benzaldehyde
under these conditions),1
5,21
confirming the importance of the

6148 J. Am. Chem. Soc., Vol. 122, No. 26, 2000
Kluger and Moore
Table 2. Derived Rate Constants and Ratios for Fragmentation of
BHT in Water and Deuterium Oxidea
(M- s-1
1
)
k
BH/k
(M-1
56
6.9
)
solvent
k
B
f
w f
k /k
-
-
3
H
D
2
O
O
7.4 × 10
0.5
0.1
3
2
3.1 × 10
a
Conditions described in the caption to Figure 2
Table 3. Second-Order Rate Constants for General Base Catalysis
B- (M- s-1
1
buffer
pK
a
k
)
-
-
-
-
-
-
6
6
6
5
5
5
chloroacetate
methoxyacetate
formate
glycolate
acetate
2.90
3.61
3.77
3.83
4.77
4.88
3.11 × 10
5.75 × 10
6.07 × 10
1.07 × 10
2.30 × 10
3.60 × 10
propionate
Discussion
Figure 1. Brønsted plot for fragmentation of BHT catalyzed by
substituted acetates and formate at 40 °C and I ) 0.1.
The plot in Figure 2 provides a unifying basis for understand-
ing catalytic steps in the fragmentation of BHT and this extends
to HBzT. The UV kinetic data for the rates of fragmentation at
the high buffer concentrations are clearly invariant while at low
concentrations there is a linear dependence. Earlier NMR studies
at high buffer concentrations revealed that the rate of fragmenta-
tion does not change with changes in buffer concentration.16
We can see from the present results that the invariant rate was
observed in the earlier studies because the high concentration
caused the proton removal step to be faster than the subsequent
uncatalyzed fragmentation step.
The change in the slope of the plot of observed rate coefficient
versus buffer concentration for fragmentation of BHT (Figure
2
) is consistent with a single mechanism having a change in
rate-determining step with increasing buffer concentration. At
low concentrations, the reaction is subject to Brønsted base
catalysis but at high buffer concentrations this becomes limited
by a subsequent process that does not involve a rate-determining
proton transfer. Thus, the rapid proton exchange of BHT
(compared to fragmentation) observed by NMR occurs in
Figure 2. Dependence of the observed first-order rate coefficient for
fragmentation of BHT on the concentration of hydrogen phosphate ion
in water (O) and deuterium phosphate ion in deuterium oxide (b) at
solutions where the observed rate coefficient for fragmentation
is not increased by additional buffer concentrations. (Since the
1
5
fragmentation process in HBzT is much slower than proton
exchange at C2R, it is reasonable that the early discovery of
the exchange reaction at C2R was not complicated by frag-
-
5
4
s
0 °C, I ) 1.0. The solid lines are plots of eq 1 (k
∞
) 6.6 × 10
-
1
-6 -1
-4 -1
and k
o
) 2.3 × 10
s
in water; k
∞
) 2.2 × 10
s
and k )
o
-
5
-1
1
.6 × 10
s
in deuterium oxide).
7
mentation. ) Thus, for BHT, removal of the C2R proton is the
step that is susceptible to general base catalysis in the
fragmentation process, consistent with the slope of the Brønsted
plot of 0.5. The curvature of the buffer plot occurs at the buffer
base concentration where the rate of the reverse reaction,
protonation of the enamine by the acid form of the buffer, is
equal to the rate of fragmentation of the intermediate that is
unlikely to be catalyzed. (Since the enamine is formed as a
reactive steady-state intermediate we do not expect to be able
Since UV kinetic data in previous studies were acquired only
for dilute buffers and NMR kinetic data only for more
concentrated buffers,15,16 we extended the range of buffer
concentrations used for the UV studies to encompass the
concentrations used for NMR kinetics. Hydrogen phosphate
buffer was used for these studies. The fragmentation is followed
in the UV by changes in absorbance at 328 nm due to the
formation of the ketone product (2). The data obtained from
these studies are shown in Figure 2. The observed first-order
rate coefficients for fragmentation show a saturating dependence
on hydrogen phosphate buffer concentration. The observed first-
order rate coefficient in deuterium oxide at the same buffer ratio
as in water also shows a linear increase and saturation (Figure
to observe it as has been done for nonaqueous conditions where
the opportunity for proton transfer is suppressed.11,12,24)
The Brønsted plot (Figure 1) for general base catalysis of
fragmentation of BHT by the buffers listed in Table 2 has a
slope of 0.5, consistent with a rate-determining step in which a
proton is abstracted from a carbon acid having a delocalized
2
). The maximum rate in deuterium oxide is significantly larger
25
conjugate base. The rate constant for proton removal by
than that in water, giving an inverse solvent isotope effect (kDOD/
kHOH ) 3.0). The solvent isotope effect in the region where the
rate increases linearly with base concentration is also inverse
3
-1 -1
hydroxide from BHT is 1 × 10 M s . The data point for
catalysis by hydroxide ion is well above the extrapolated
Brønsted plot for carboxylate buffers. This is consistent with a
rate-determining proton transfer from carbon as the solvent lyate
(
kDOD/kHOH ) 1.3 measured at the lowest buffer concentration).
(
24) Barletta, G. L.; Zou, Y.; Huskey, W. P.; Jordan, F. J. Am. Chem.
(25) Bell, R. P. The Proton in Chemistry, 2nd ed.; Cornell University
Press: Ithaca, New York, 1973; pp 194-225.
Soc. 1997, 119, 2356-2362.

Destruction of Vitamin B1 by Benzaldehyde
J. Am. Chem. Soc., Vol. 122, No. 26, 2000 6149
Scheme 5
Scheme 6
ion can react through the solvation shell while the buffer
components must react directly.25,26
(5) associated with kf is very fast, competing with the kinetically
significant rate for protonation of the conjugate base. The inverse
solvent isotope effects on kB provide further insights. At low
buffer concentrations, proton removal is rate determining. This
is consistent with a normal isotope effect on dissociation of the
conjugate acid of the catalytic base. The high inverse solvent
isotope effect at saturation is consistent with deuteration of the
conjugate base of BHT being expectedly slower than is
protonation of the same species in Scheme 7. (The reverse
reaction thus displays a normal solvent isotope effect, similar
These results can be accommodated by a mechanism in which
removal of the C2R proton from BHT is rate-determining for
fragmentation in dilute buffers but a subsequent step is rate-
determining at higher buffer concentrations. The dependence
of the observed first-order rate constant on buffer components
and solution acidity according to the mechanism in Scheme 3
under steady-state conditions is given by eq 1 (terms shown in
Scheme 5).
to what was reported by Keefe and Jencks in their study of
-
27,29
kobs ) (k [B] + k [OH ]) k /(k [BH] + k + k ) (1)
E1CB reactions.
)
B
OH
f
BH
w
f
The low barrier to the unimolecular fragmentation of the
ionized intermediate suggests that such a process would be a
problem in benzoylformate decarboxylase where the conjugate
base of the pyrophosphate of HBzT is formed by the loss of
carbon dioxide from the precursor derivative of thiamin diphos-
phate (Scheme 2). A low net dielectric of the medium at the
active site would accelerate decarboxylation and suppress
fragmentation because protonation of the pyrimidine would be
suppressed (the conjugate acid is a dication).
Following the method outlined by Keefe and Jencks,27
nonlinear regression of the data in Figure 2 with eq 1 was used
to find the rate constant at zero buffer and at saturation. We
-
3
-1 -1
obtain kB ) 7.4 × 10
M
s
for removal of the C2R proton
from BHT by the basic component of the buffer as well as the
rate constant ratios, kBH/kf ) 56 M and kw/kf ) 0.5. With use
of a specific value for the concentration of water of 55.5 M
-
1
-
3
-1
and kw ) kHOH[HOH], then kHOH/kf ) 9 × 10
M
and kBH/
3
kHOH ) 6 × 10 . Table 2 contains a summary of the derived
kinetic parameters for reactions in water and deuterium oxide.
From the values of the rate constants in the ratio expression
in Table 1 and the reported pKa of 15.4 for C2R of the related
methoxybenzyl methylthiazolium salts28 (which we assume will
The mechanism of the fragmentation process of the conjugate
base of BHT remains an interesting question. We now know
that it is not subject to Brønsted acid-base catalysis, so it is
unlikely to involve kinetically significant proton transfers. Since
the conversion to the fragmentation products is very fast, the
formation of free forms of highly energetic intermediates is
6
-1 -1
be similar for BHT), we obtain kBH ) 6.7 × 10 M
s
and
5
-1
27
kf ) 1.2 × 10 s by substitution into eq 1. The ratio establishes
unlikely and may involve assisted proton transfers. The
that the fragmentation process from the enamine intermediate
electron-shift arrows in Scheme 7 summarize the steps in a six-
electron electrocyclic process (related to a [1,5]-sigmatropic
rearrangement), which could be consistent with the very low
activation barrier for the reaction. There are intriguing alternative
(
26) Kluger, R.; Wong, M.; Dodds, A. K. J. Am. Chem. Soc. 1984, 106,
113-1117.
27) Keeffe, J. R.; Jencks, W. P. J. Am. Chem. Soc. 1983, 105, 265-
79.
28) Bordwell, F. G.; Satish, A. V.; Jordan, F.; Rios, C. B.; Chung, A.
C. J. Am. Chem. Soc. 1990, 112, 792-797.
1
2
(
(
(29) Keeffe, J. R.; Jencks, W. P. J. Am. Chem. Soc. 1981, 103, 2457-
2459.

6150 J. Am. Chem. Soc., Vol. 122, No. 26, 2000
Kluger and Moore
Scheme 7
processes that can accomplish the transformation, including a
suggestion by a referee that involves addition of the carbanion
at C2R to the adjacent pyrimidinium ion, followed by cleavage
in analogy to the mechanism by which sulfite cleaves thiamin
tation on buffer concentration permits an analysis that shows
that the steps leading from the conjugate base to the fragmented
products are very fast, suggesting the possibility of an electro-
cyclic mechanism that splits the thiazole and pyrimidine
components. Finally, it is clear that enzymes that involve
reactions via similar intermediates need to provide a means to
avoid the fragmentation reaction. Structural information suggests
that the active site of benzoylformate decarboxylase is relatively
30
through initial addition to the pyrimidine.
Conclusions
BHT, like the N1′-methyl derivative of HBzT,16 does not
release thiamin in neutral and alkaline solutions but fragments
into thiazole and pyrimidine components. Our kinetic study of
fragmentation of BHT over a wide range of concentrations
establishes that the observed buffer catalysis arises from the
first step, transferring the C2R proton to a base. Catalysis of a
similar step will explain the general base catalysis observed with
9
nonpolar. This environment will lower the apparent basicity
of the pyrimidine derived from hydroxybenzylthiamin diphos-
phate. Since the fragmentation is critically dependent on the
16
pyrimidine being positively charged, the normal catalytic
process may avoid competition from fragmentation.
1
5
fragmentation of hydroxybenzylthiamin and its N1′-methyl
1
6
Acknowledgment. We thank the Natural Sciences and
Engineering Research Council of Canada for support through
an operating grant and Dr. John Pezacki for helpful discussions.
Ian Moore is the recipient of an Ontario Graduate Scholarship
in Science and Technology.
analogue. Therefore, the fragmentation process from the
conjugate base probably occurs without catalysis. The inverse
solvent isotope effect on the rate at saturating buffer concentra-
tion confirms the existence of a carbanion equivalent as an
intermediate. The nonlinear dependence of the rate of fragmen-
(30) Zoltewicz, J. A.; Uray, G. Bioorg. Chem. 1994, 22, 1-28.
JA000194I