Deuterium labeling as a test of intramolecular hydride mechanisms in the fragmentation of 2-(1-hydroxybenzyl)-N1′-methylthiamin

Article abstract of DOI:10.1139/v05-146

2-(1-Hydroxybenzyl)-N1′-methylthiamin (1b) is a model for the addition intermediate in the thiamin catalyzed benzoin condensation. However, N-alkylation alters the reactivity of the compound: instead of undergoing base-catalyzed formation of benzaldehyde and N1′-methylthiamin, it rapidly forms trimethyl amino pyrimidine (2b) and phenylthiazole ketone (3). The base-catalyzed fragmentation process is faster than the analogous enzymic reaction (in benzoylformate decarboxylase) under the same conditions. One possible mechanism for the rapid fragmentation is an internal hydride transfer from α-C2 to the methylene bridge between the heterocycles. To test the hydride mechanism we prepared α-C2-deuterated 1b and conducted the fragmentation reaction in normal water. Spectroscopic analysis revealed that the trimethyl aminopyrimidine product does not contain any deuterium, ruling out a hydride transfer mechanism. This supports a mechanism for fragmentation that proceeds instead via a proton transfer from α-C2. Since protonation (and hence, deprotonation) of that site is part of the normal catalytic cycle of benzoylformate decarboxylase, the enzyme must divert the reaction from the lowest energy pathway since it would share a common intermediate with the fragmentation process.

Full text of DOI:10.1139/v05-146

1277
Deuterium labeling as a test of intramolecular
hydride mechanisms in the fragmentation of 2-(1-
hydroxybenzyl)-N1′-methylthiamin1
Glenn Ikeda and Ronald Kluger
Abstract: 2-(1-Hydroxybenzyl)-N1′-methylthiamin (1b) is a model for the addition intermediate in the thiamin cata-
lyzed benzoin condensation. However, N-alkylation alters the reactivity of the compound: instead of undergoing base-
catalyzed formation of benzaldehyde and N1′-methylthiamin, it rapidly forms trimethyl amino pyrimidine (2b) and
phenylthiazole ketone (3). The base-catalyzed fragmentation process is faster than the analogous enzymic reaction (in
benzoylformate decarboxylase) under the same conditions. One possible mechanism for the rapid fragmentation is an
internal hydride transfer from α-C2 to the methylene bridge between the heterocycles. To test the hydride mechanism
we prepared α-C2-deuterated 1b and conducted the fragmentation reaction in normal water. Spectroscopic analysis re-
vealed that the trimethyl aminopyrimidine product does not contain any deuterium, ruling out a hydride transfer mecha-
nism. This supports a mechanism for fragmentation that proceeds instead via a proton transfer from α-C2. Since
protonation (and hence, deprotonation) of that site is part of the normal catalytic cycle of benzoylformate decarboxyl-
ase, the enzyme must divert the reaction from the lowest energy pathway since it would share a common intermediate
with the fragmentation process.
Key words: thiamin, fragmentation, benzoylformate decarboxylase, proton transfer, hydride shift.
Résumé : La 2-(1-hydroxybenzyl)-N1′-méthylthiamine (1b) est un modèle pour l’intermédiaire d’addition dans la
condensation de la benzoïne catalysée par la thiamine. Toutefois, la N-alkylation altère la réactivité du composé qui, au
lieu de subir une réaction catalysée par les bases conduisant à la formation de benzaldéhyde et de N1′-méthylthiamine
conduit à la formation rapide de triméthylaminopyrimidine (2b) et phénylthiazole cétone (3). Le processus de fragmen-
tation catalysé par les bases est plus rapide que la réaction enzymatique analogue (en présence de décarboxylase de
benzoylformate) dans les mêmes conditions. Un mécanisme possible pour la fragmentation rapide implique un transfert
interne d’hydrure de la position α-C2 vers le pont méthylène entre les hétérocycles. Afin de tenter de confirmer ce mé-
canisme impliquant un hydrure, le composé 1b deutéré en α-C2 a été préparé et soumis à une fragmentation dans de
l’eau normale. Une analyse spectroscopique a révélé que le produit triméthyl aminopyrimidine ne contient pas de deu-
térium, ce qui élimine le mécanisme par transfert d’hydrure, mais qui supporte un mécanisme de fragmentation qui se
produirait plutôt par le biais d’un transfert de proton à partir de α-C2. Puisque la protonation (et, par voie de consé-
quence, la déprotonation) de ce site fait partie du cycle catalytique normal de la décarboxylase du benzoylformate,
l’enzyme doit faire dévier la réaction de la voie impliquant l’énergie minimale puisqu’il doit partager un intermédiaire
commun avec le processus de fragmentation.
Mots clés : thiamine, fragmentation, décarboxylase de benzoylformate, transfert de proton, déplacement d’hydrure.
[Traduit par la Rédaction] Ikeda and Kluger 1280
Introduction
(3–5). In the enzymic system, ionization of the hydroxyl
group facilitates the elimination of benzaldehyde. However,
the major products from decomposition of HBnT in solution
are an aminopyrimidine (AP, 2a) and a phenylthiazole
ketone (PTK, 3) (6). These products were originally reported
by Oka et al. for different conditions (7). Base-catalyzed
fragmentation overwhelms elimination, being over 103 times
faster at pH 7 (8). To observe fragmentation without compe-
tition from elimination, the N1′-methylated derivative of
HBnT (MHBT, 1b) was prepared, an analogue of the BFD
intermediate with N1′ protonated on the enzyme (5).
The fragmentation pathway of HBnT is consistent with
the first step being the loss of a proton from α-C2 to gener-
ate the delocalized conjugate base, which undergoes the
fragmentation reaction as shown (Fig. 1, mechanism 1). The
carbanion can also be accessed by decarboxylation (9, 10).
The base-catalyzed addition of thiamin to benzaldehyde
produces 2-(1-hydroxybenzyl)thiamin (HBnT, 1a), following
the mechanism deduced by Breslow (1, 2).
HBnT also resembles HBnT diphosphate, an intermediate
on the pathway of benzoylformate decarboxylase (BFD)
Received 30 March 2005. Published on the NRC Research
Press Web site at http://canjchem.nrc.ca on 21 October 2005.
G. Ikeda and R. Kluger.2 Davenport Chemical Laboratories,
Department of Chemistry, University of Toronto, Toronto,
ON M5S 3H6, Canada.
1This article is part of a Special Issue dedicated to organic
reaction mechanisms.
2Corresponding author (e-mail: rkluger@chem.utoronto.ca).
Can. J. Chem. 83: 1277–1280 (2005)
doi: 10.1139/V05-146
© 2005 NRC Canada

1278
Can. J. Chem. Vol. 83, 2005
Fig. 1. Mechanisms and deuterium tracing in the fragmentation of HBnT.
N
R2
N
H2N
N
-O2C
HO
R1
S
-CO2
mechanism 1
-H+
N
R2
N
R2
N
N
R2
N
N
H2N
HO
H2N
HO
N
H2N
R2
AP
N
H2N
HO
2
N
N
+
R1
R1
C5'
N
N
S
S
R1
O
H(D)
R1
S
S
PTK
C2α
-H+
1
3
mechanism 2
N
N
R2
N
R2
H2N
HBnT: R1 = -CH2CH2OH
BFD: R1 = -CH2CH2OPO3
N
H2N
2-
H
C
H
N
H(D)
4
D
R1
O
a, R2 = -H
N
R2
S
b, R2 = -CH3
N
+
H2N
3
+
N
R1
mechanism 3
O
H(D)
S
NH2
N
NH2
NH2
N
N
N
R2
N
N
N
D
O
R2
R2
OH
D
+
D
S
5
3
An alternative pathway involves the kinetically equivalent
oxyanion, obtained from rapid proton transfer from oxygen
to carbon that may derive additional stabilization from the
positively charged pyrimidine. Fragmentation then would re-
sult from an internal Cannizzaro-type reaction (11) involv-
ing a hydride shift from α-C2 (Fig. 1, mechanisms 2 and 3).
If the hydride mechanism was responsible for the fragmenta-
tion, it might involve a manifold of intermediates that are
clearly distinct from those formed in the enzymic process. In
these mechanisms, the conjugate base of MHBT would react
in analogy to the conjugate base of the hydrate of benzalde-
hyde (Fig. 2). The distinction of hydride and proton transfer
mechanisms is reminiscent of the alternatives that were con-
sidered as possibilities in the normal catalytic process of
glyoxylase I (12–14), which were resolved by the work of
Jordan, Kozarich, and their co-workers (13, 14).
Fig. 2. The hydroxide induced Cannizzaro reaction as a model
for hydride transfer in MHBT.
O
OH
H*
H
O
O
O
H*
OH
-OH
Ph
H
O
H*
+
conducted the reaction in normal water. The α-C2 carbon
acid exchanges very slowly under the conditions of our stud-
ies of the fragmentation reaction, so loss of the deuteron will
not occur prior to the fragmentation process. The hydride
mechanisms require incorporation of the deuterium into the
methyl or aryl groups of the pyrimidine product, and the rel-
ative rates assure us that this would be detected. Further-
more, proton–deuterium exchange does not occur into the
products, so that any loss of deuterium comes from the ini-
tial reaction.
The proposed mechanism could potentially give C5′-
deuterated AP from the fragmentation of MHBT-d in water
if the aromatization step occurred with internal deuterium
transfer. However, tautomerization would be subject to an
The hydride transfer mechanism must give a direct trans-
fer to the acceptor site while a proton transfer mechanism
leads to equilibration with the solvent. Thus, we produced
the α-C2 deuterated analogue of MHBT (MHBT-d) by addi-
tion of thiamin addition to deuterobenzaldehyde, and then
© 2005 NRC Canada

Ikeda and Kluger
1279
Fig. 3. Summary of the results from the deuterium labeling ex-
isotope effect and could be solvent-mediated. Therefore,
only partial incorporation of deuterium into the C5′ hydro-
gen is possible in this mechanism.
periments in water and deuterium oxide.
CH3
N
CH3
N
In the case of BFD, a carbanion must be generated from
loss of carbon dioxide from the intermediate generated by
addition of thiamin diphosphate to benzoylformate. The
same carbanion would be generated from the loss of a pro-
ton from the conjugate of thiamin diphosphate and benzalde-
hyde (HBnT diphosphate). If the fragmentation reaction
occurs by a hydride shift mechanism, then this would com-
pete with elimination of the TDP ylide in the enzyme reac-
tion. The enzyme could promote formation of benzaldehyde
by controlling the direction of elimination in formation of
the carbonyl group, forming the aldehyde rather than the
ketone.
CH3
N
N
CH3
H2N
H2N
pH 8.0
CH3
CH2CH2OH
N
H3C
D
HO
S
+
3
CH3
N
CH3
N
N
CH3
N
CH3
H2N
HO
D2N
DH2C
pD 8.0
CH3
CH2CH2OH
N
H
Experimental section
S
+
The syntheses of MHBT and MHBT-d have been previ-
ously reported (8, 15). The bisperchlorate of MHBT-d
(0.3 mmol) was dissolved in 10 mL of water at 25 °C. The
solution acidity was maintained at pH 8.0 by addition of
0.1 mol/L potassium hydroxide by an automated burette. Af-
ter 2 h, the reaction was stopped by addition of hydrochloric
acid (1 mol/L) to give an acid concentration in the sample of
0.01 mol/L. The product was purified by extraction with two
portions of dichloromethane. Lyophilization of the aqueous
layer yielded a mixture of AP and potassium chloride (6–8).
A similar procedure was used for the reaction of MHBT
in deuterium oxide. We converted pH meter readings to
those for pD by adding 0.4 to the displayed results (16).
Fully deuterated acid–base reagents were used. This pro-
duced the C5′-monodeuterated analog of AP. 1H NMR
(500 MHz, CD3OD) δ: 8.0 (s, 1H), 3.8 (s, 3H), 2.6 (s, 3H),
2.1 (dt, 3H). EI-MS (high resolution) calcd. for
[N7H10DN3]+: 138.1016; found: 138.1018.
3
mechanism 1. Monodeuteration of AP from the fragmenta-
tion in deuterium oxide is also consistent with mechanism 1
in which C—N bond cleavage occurs with the ionization of
C5′ followed by solvent-mediated protonation. Regiospecific
deuterium incorporation demonstrates that the isotope comes
from solvent and is not transferred internally. The fact that
proton–deuterium exchange is not observed in AP indicates
that the C5′ methyl group is only weakly acidic, which is
consistent with the fragmentation being highly exothermic.
The evidence against an internal hydride transfer supports
a mechanism involving the conjugate base at carbon (α-C2).
If fragmentation involves the delocalized conjugate base of
MHBT, C—N bond cleavage, oxidation of the hydroxyl, and
protonation of the C5′ carbon must occur, in steps or in con-
cert. The surprisingly low barrier associated with fragmenta-
tion (104 s–1) (15) suggests that the transition state of the
C—N cleavage process has special stabilization. We note
that BFD avoids fragmentation of a similar carbanionic
intermediate that is generated by decarboxylation of the
addition intermediate from benzoylformate and thiamin
diphosphate. The enzyme’s normal catalytic route has a
higher barrier (kcat = 102 s–1) than the fragmentation reaction
(17). Since the enzyme-catalyzed reaction necessarily in-
volves a carbanionic structure upon loss of carbon dioxide,
the mechanism of fragmentation avoidance is clearly a result
of the effects of the enzyme environment (10). The protein
manifold of the active site must increase the barrier to frag-
mentation while lowering that for protonation since the anal-
ogous substrates readily undergo fragmentation in solution.
Results and discussion
No deuterium incorporation occurs into AP from the frag-
mentation of MHBT-d in water (1H NMR and EI-MS (high
resolution) calcd. for [N7H11N3]+: 137.0953; found:
137.0953). As a control, AP was characterized from the
fragmentation of MHBT in deuterium oxide. EI-MS indi-
cates the incorporation of a single deuterium. The location
1
of the deuteron was determined by H NMR to be at the C5′
position, which is derived from the methylene bridge be-
tween the two heterocycles. The signal for the C5′-methyl
group initially consisted of a doublet of triplets (2JH-D
=
4
2.1 Hz, JH-H = 0.9 Hz, CD3OD) from coupling to the
geminal deuteron and neighbouring C6′ proton. After con-
1
firming the latter assignment by H–1H COSY, signal de-
Acknowledgements
coupling produced the characteristic 1:1:1 triplet of the
monodeuterated methyl group. As the protons of AP do not
exchange in deuterium oxide, deuterium incorporation in
water would have been detectable under these conditions.
Owing to the absence of deuterated AP products 4 and 5
from the fragmentation of MHBT-d in water, we can rule out
hydride transfer mechanisms (Fig. 3). Deuterium would be
scrambled into solvent from deprotonation of the weak car-
bon acid to generate the carbanion that is consistent with
We thank the Natural Sciences and Engineering Research
Council of Canada (NSERC) for support through an operat-
ing grant and for a postgraduate scholarship to GI.
References
1. R. Breslow. Chem. Ind. (London), 893 (1957).
2. R. Breslow. J. Am. Chem. Soc. 79, 1762 (1957).
© 2005 NRC Canada

1280
Can. J. Chem. Vol. 83, 2005
3. P.M. Weiss, G.A. Garcia, G.L. Kenyon, W.W. Cleland, and P.F.
Cook. Biochemistry, 27, 2197 (1988).
4. L.J. Dirmaier, G.A. Garcia, J.W. Kozarich, and G.L. Kenyon.
10. Q. Hu and R. Kluger. J. Am. Chem. Soc. 126, 68 (2004).
11. C.G. Swain, A.L. Powell, W.A. Sheppard, and C.R. Morgan. J.
Am. Chem. Soc. 101, 3576 (1979).
J. Am. Chem. Soc. 108, 3149 (1986).
12. I.A. Rose. Biochim. Biophys. Acta, 25, 214 (1957).
13. S.S. Hall, A.M. Doweyko, and F. Jordan. J. Am. Chem. Soc.
100, 5934 (1978).
14. R.V.J. Chari and J.W. Kozarich. J. Biol. Chem. 256, 9785
(1981).
15. G. Ikeda and R. Kluger. J. Phys. Org. Chem. 17, 507 (2004).
16. P.K. Glasoe and F.A. Long. J. Phys. Chem. 64, 188 (1960).
17. E.S. Polovnikova, M.J. McLeish, E.A. Sergienko, J.T. Burgner,
N.L. Anderson, A.K. Bera, F. Jordan, G.L. Kenyon, and M.S.
Hasson. Biochemistry, 42, 1820 (2003).
5. M.S. Hasson, A. Muscate, M.J. McLeish, L.S. Polovnikova,
J.A. Gerlt, G.L. Kenyon, G.A. Petsko, and D. Ringe. Biochem-
istry, 37, 9918 (1998).
6. R. Kluger, J.F. Lam, and C.-S. Kim. Bioorg. Chem. 21, 275
(1993).
7. Y. Oka, S. Kishimoto, and H. Hirano. Chem. Pharm. Bull. 18,
527 (1970).
8. R. Kluger, J.F. Lam, J.P. Pezacki, and C.-M. Yang. J. Am.
Chem. Soc. 117, 11383 (1995).
9. Q. Hu and R. Kluger. J. Am. Chem. Soc. 124, 14858 (2002).
© 2005 NRC Canada