Báo cáo khoa học: Reaction mechanisms of thiamin diphosphate enzymes: defining states of ionization and tautomerization of the cofactor at individual steps docx

MINIREVIEW
Reaction mechanisms of thiamin diphosphate enzymes:
defining states of ionization and tautomerization of the
cofactor at individual steps
Natalia S. Nemeria, Sumit Chakraborty, Anand Balakrishnan and Frank Jordan
Department of Chemistry, Rutgers, The State University of New Jersey, Newark, NJ, USA
Introduction
Mindful of the fact that there are several reviews on
the enzymology of thiamin diphosphate (ThDP, the
vitamin B1 coenzyme; for structures of small molecules
mentioned in the present review, see Fig. 1) available
in the literature [1–15], the present review aims to con-
centrate on the tautomeric and ionization states of
ThDP on enzymes, which is a fascinating and, in some
respects, perhaps unique aspect of thiamin enzymology.
Keywords
1¢,4¢-iminopyrimidine tautomeric form of
thiamin; benzaldehyde lyase;
benzoylformate decarboxylase; CD; enamine
intermediate; pyruvate decarboxylase;
pyruvate dehydrogenase; thiamin
diphosphate
Correspondence
N. S. Nemeria, 73 Warren Street, Newark,
NJ 07102, USA
Fax: +1 973 353 1264
Tel: +1 973 353 5727
E-mail:
F. Jordan, 73 Warren Street, Newark,
NJ 07102, USA
Fax: +1 973 353 1264

Tel: +1 973 353 5470
E-mail:
(Received 23 October 2008, revised 4
February 2009, accepted 9 February 2009)
doi:10.1111/j.1742-4658.2009.06964.x
We summarize the currently available information regarding the state of
ionization and tautomerization of the 4¢-aminopyrimidine ring of the thia-
mine diphosphate on enzymes requiring this coenzyme. This coenzyme
forms a series of covalent intermediates with its substrates as an electro-
philic catalyst, and the coenzyme itself also carries out intramolecular pro-
ton transfers, which is virtually unprecedented in coenzyme chemistry.
An understanding of the state of ionization and tautomerization of the
4¢-aminopyrimidine ring in each of these intermediates provides important
details about proton movements during catalysis. CD spectroscopy, both
steady-state and time-resolved, has proved crucial for obtaining this infor-
mation because no other experimental method has provided such atomic
detail so far.
Abbreviations
3-PKB, (E)-4-(pyridine-3-yl)-2-oxo-3-butenoic acid; AcP
)
, acetylphosphinate; AP, the canonical 4¢-aminopyrimidine tautomer of ThDP or its
C2-substituted derivatives; APH
+
, the N1-protonated 4-aminopyrimidinium form of ThDP or its C2-substituted derivatives; BAL, benzaldehyde
lyase; BFDC, benzoylformate decarboxylase; E1ec, the first component of the Escherichia coli pyruvate dehydrogenase complex; E1h, the
first component of the human pyruvate dehydrogenase complex; GCL, glyoxylate carboligase; HBThDP, C2a-hydroxybenzylThDP, the adduct
of benzaldehyde and ThDP; HEThDP, C2a-hydroxyethylThDP, the adduct of acetaldehyde and ThDP; IP, 1¢,4¢-iminopyrimidine tautomer of
ThDP or its C2-substituted derivatives; LThDP, C2a-lactylThDP, the adduct of pyruvic acid and ThDP; MAP, methyl acetylphosphonate; MBP,
methyl benzoylphosphonate; PAA, (E)-3-(pyridine-3-yl) acrylaldehyde; PLThDP, C2a-phosphonolactylThDP, the adduct of MAP and ThDP;
POX, pyruvate oxidase from Lactobacillus plantarum; ThDP, thiamin diphosphate; TK, transketolase; Yl, C2-carbanion ⁄ ylide ⁄ carbene form

conjugate base of ThDP; YPDC, yeast pyruvate decarboxylase from Saccharomyces cerevisiae.
2432 FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS
This issue has come to the fore relatively recently, but
its understanding is made more urgent and more sig-
nificant by some recent X-ray crystal structure determi-
nations of ThDP enzymes. Briefly, the question is
related to the conundrum that any plausible mecha-
nism suggested for ThDP-dependent enzymes, be they
2-oxoacid decarboxylases or carboligases [examples of
a non-oxidative decarboxylase yeast pyruvate decar-
boxylase (YPDC; EC 4.1.1.1), an oxidative decarboxyl-
ase, the pyruvate dehydrogenase complex (EC 1.2.4.1),
and a carboligase benzaldehyde lyase (BAL;
EC 4.1.2.38) are given in Schemes 1–3], requires some
proton transfer steps. On the basis of the accumulated
understanding of enzyme mechanisms, such proton
transfers are likely to be mediated by general acid ⁄ base
catalysts, such as His, Asp and Glu, and perhaps Cys,
Lys and Tyr, with the understanding that the enzyme
active center could modulate the aqueous pK
a
of these
side chains, as needed.
Several groups, including our own [16], have spent
considerable time trying to assign acid ⁄ base functions
to such residues on ThDP enzymes, with limited suc-
cess. Very recently, Yep et al. [17] carried out satura-
tion mutagenesis experiments probing the function of
two active center histidine residues (His70 and
His281) on benzoylformate decarboxylase (BFDC;

EC 4.1.1.7), long believed to participate in acid ⁄ base
reactions [18]. Surprisingly, their results indicated that
hydrophobic residues could replace the His281 with little
penalty, and the His70Thr or His70Leu substitutions
Scheme 1. Mechanism of yeast pyruvate decarboxylase YPDC.
N
S
Me
R2
Me
HO
CO
2
–
N
S
Me
R2
N
S
Me
Me
HO
Me
N
S
Me
R2
N
S

Me
R2
+
yli de, Yl
LThDP, IP
Me
O
–
enamine/ C2α-carbanion, AP(or APH
+
)
+
+
R1
R1
+
–
k
2
k
3
k
5
R1 = 4'-amino-2-methyl-5-pyrimidyl
R2 = β-hydroxyethyldiphosphate
OH
S8-acetyldihydrolipoyl-E2
R2
C
H

3
C
O
C
O
2
-
CO
2
k
–MM
R1
R1
R1
k
4
lipoyl-E2
2-AcThDP, AP (or APH
+
)
S S
E2
SH S
E2
CoASH
CH
3
COSCoA
dihydrolipoyl-E2
SHHS

E2
E3 +FAD+NAD
+
N
S
Me
R2
+
R1
–
k
M
M
pyruvate.
Michaelis complex
k
–2
HN
N
N
S
NH
Me
Me
R2
H
N
N
N
S

NH
2
Me
Me
H
+
4'-aminopyrimidinium, APH
+
+
1',4'-iminopyrimidine, IP
R2
N
N
N
S
NH
2
Me
Me
H
4'-aminopyrimidine, AP
+
R2
thiazolium
-H1', pK
1'
–H4'
1'
4'
2

3'
+
H
–H4',
pK
4'
–H2, pK
2
Ke q
K
tautomerization
H
3
COC
MM, AP
N
S
Me
R2
Me
HO
H
+
R1
HEThDP, IP
k
6
k
–6
Scheme 2. Mechanism of E coli and human pyruvate dehydrogenase complex with role of ThDP.

N. S. Nemeria et al. Enzyme-bound imino tautomer of thiamin diphosphate
FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS 2433
only led to a 30-fold penalty on k
cat
⁄ K
m
. A reason-
able question in the interpretation of such findings is
what is the appropriate contribution from His, Asp
or Glu to reflect general acid ⁄ base reactivity on the
enzyme? There appear to be two well-explored exam-
ples that could provide benchmark values, although
the precise interpretation of these numbers is not only
risky, but also depends on the particular substitution
used to arrive at them [19]: (a) serine proteases,
where substitution of either His (a presumed general
acid ⁄ base catalyst) or Ser (a nucleophilic catalyst) by
Ala in the well-characterized Asp-His-Ser catalytic
triad of subtilisin leads to an approximate 2 · 10
6
reduction in k
cat
, with little impact on k
cat
⁄ K
m
[20]
and (b) ketosteroid isomerase (EC 5.3.3.1), where sub-
stitution of the catalytic Asp38 by Asn leads to a
10

5.6
decrease in k
cat
[21], whereas substitution of the
same residue by Ala only reduced the k
cat
by 140
[22].
Complicating this issue on ThDP enzymes is that
the pH dependence of the steady-state kinetic parame-
ters does not provide clear evidence for the participa-
tion of such residues in the rate-limiting step(s). For
example, all potential active center acid ⁄ base residues
were substituted on YPDC [16], with little perturbation
of the pH dependence of such plots, perhaps with the
exception of the substitution at the conserved gluta-
mate. Therefore, the 100- to 500-fold reduction in
steady-state kinetic constants could not be unequivo-
cally attributed to acid ⁄ base function, whereas such
numbers are certainly consistent with hydrogen-bond-
ing interactions.
Relevant to the issue of acid ⁄ base catalysis, the
structure of two interesting ThDP-dependent lyases
was solved with unusual characteristics. The enzyme
BAL carries out reversible decomposition of (R)-ben-
zoin to two molecules of benzaldehyde according to
the mechanism given in Scheme 3; in the reverse direc-
tion, the enzyme is a carboligase. The BAL structure
reported contained only two acid ⁄ base residues sur-
rounding the ThDP at the active center [23–25]: a

highly conserved Glu50 within hydrogen-bonding dis-
tance of the N1¢ atom of the 4¢-aminopyrimidine (AP)
ring and a His29 residue. The residue His29 is too far
from the thiazolium C2 atom to be of value in the first
steps of the reaction and was suggested to have a
function in removing the b-hydroxyl proton of the
ThDP-bound benzoin to assist in releasing the first
benzaldehyde molecule. In the authors’ view, this
enzyme provides the clearest interpretation of the pH
dependence of the steady-state kinetic parameters of
any ThDP enzymes to date. There is a pK
a
= 5.3 at
the acidic side of either the k
cat
-pH or k
cat
⁄ K
m
-pH pro-
file, almost certainly corresponding to the highly con-
served glutamate residue [26]. With this information in
hand, the pH dependence of kinetic parameters on
YPDC could be re-examined, suggesting that the
conserved glutamate affected the behavior similarly.
The second case reported even greater surprises: the
enzyme glyoxylate carboligase (GCL; EC 4.1.1.47)
carries out a carboligation reaction after decarboxyl-
ation of the first molecule of glyoxal to the enamine
intermediate. This enzyme is not only devoid of acid ⁄

base groups at its active center within hydrogen-bond-
ing distance of ThDP, but it is also lacking the highly
conserved Glu and, in its place, there is a hydrophobic
valine residue [27].
These two case studies suggest that our understand-
ing of ThDP enzymes is not nearly as complete as was
previously assumed, and certainly suggest that the
N
+
S
R
2
HO
Ph
N
S
R
2
N
S
R
2
Ph
HO
Ph
HO
N
S
R
2

ylide
Mechanism of benzaldehyde lyase
–
C2α-carbanion/enamiine
+
R
1
R
1
+
–
k
2
HN
N
N
S
NH
R
2
H
N
N
N
S
NH
2
H
+
4'-aminopyrimidinium

+
1',4'-iminopyrimidine
R
2
N
N
N
S
NH
2
4'-aminopyrimidine
+
R
2
thiazolium
–H1'
–H4'
1'
4'
2
3'
k
–2
R
1
R
1
N
+
S

R
2
Ph
OH
R
1
+
H
k
1
/k
–1
PhCHO
HBThDP
k
–4
k
4
k
–5
k
5
PhCHO
AP
APH
+
IP
λ
max
380 nm

Ph
Ph
O
OH
Ph
OH
DDEThDP
PhCHO
PhCHO
k
3
k
–3
Ph = C
6
H
5
Scheme 3. Mechanism of benzaldehyde lyase.
Enzyme-bound imino tautomer of thiamin diphosphate N. S. Nemeria et al.
2434 FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS
ThDP cofactor has a much more dramatic impact on
the reaction pathway than hitherto accepted. With
results such as those described above, the coenzyme
and its chemical reactivity need to be scrutinized from
a newer vantage point.
Early evidence indicating a catalytic
function for the AP ring
The chemistry and enzymology of ThDP is intimately
dependent on three chemical moieties comprising the
coenzyme: a thiazolium ring, a 4-aminopyrimidine

ring and the diphosphate side chain (Fig. 1). From
the large number of high-resolution X-ray structures
available over the past 16 years, starting with the
structures of transketolase [28] (TK; EC 2.2.1.1), pyru-
vate oxidase [29] (POX; EC 1.2.3.3) from Lactobacil-
lus plantarum and YPDC [30,31], it has become clear
that the diphosphate serves to bind the cofactor to
the protein. This is achieved via electrostatic bonds
of the a and b phosphoryl group negative charges
with the required Mg
2+
or Ca
2+
, the divalent metal
serving as an anchor in a highly tailored environment
with a universally conserved GDG recognition site
and the diphosphate-Mg
2+
binding motif consisting
of a GDG-X
26
-NN sequence of amino acids, as sug-
gested by the Hawkins et al. [32]. As shown in a series
of seminal studies by Breslow, the thiazolium ring is
central to catalysis [33], as a result of its ability to
form a key nucleophilic center at the C2 atom, the
C2-carbanion ⁄ ylide or carbene, depending on one’s
viewpoint with respect to the relative importance of
the resonance contributions. The demonstration that
the thiazolium C2H can undergo exchange with D

2
O,
and that thiazolium salts per se, even in the absence
of the AP ring, can induce benzoin condensations in a
manner analogous to the cyanide ion catalyzed ben-
zoin condensation, led to the proposal of the pathway
involving thiazolium-bound covalent intermediates, as
also shown in Schemes 1–3. Thus, is there anything
else to thiamin catalysis? It was reported that the pro-
tein environment of YPDC provides a catalytic rate
acceleration of 10
12
–10
13
[34]. Is this simply a result
of juxtaposition of amino acid side chains to provide
the general acid ⁄ base catalysis, or an enzymatic sol-
vent effect [10,14] and does it include a contribution
Fig. 1. Compounds under discussion.
N. S. Nemeria et al. Enzyme-bound imino tautomer of thiamin diphosphate
FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS 2435
from the special properties of ThDP when enzyme
bound?
Starting in the 1960s, Schellenberger and his princi-
pal associate Hu
¨
bner, and their colleagues in Halle,
examined the role of the AP ring [8]. Most notably,
they undertook de novo synthesis of thiamin diphos-
phate analogs replacing each of the three nitrogen

atoms of the AP ring in turn. They then tested each of
these deaza analogs for coenzyme activity on a number
of enzymes. The results clearly indicated that the N1¢
atom and the N4 ¢ -amino group are absolutely
required, with the N3¢ atom to a lesser extent. On the
basis of application of this powerful probe to a num-
ber of ThDP enzymes, the group from Halle made the
totally reasonable suggestion that the AP ring has cat-
alytic role, and does not serve simply as an anchor to
hold the coenzyme in place. The idea was further elab-
orated at Rutgers with a synthetic model in which the
mobile proton at the N1 ¢ position (the principal site of
first protonation of the AP) was replaced by a methyl
group, creating N1¢-methylthiaminium and N1¢-meth-
ylpyrimidinium salts, consequently demonstrating that
the positive charge installed at the N1¢ position con-
verted the amino group to a weak acid with a pK
a
of
almost 12–12.5 in aqueous solution [35]. This raised
the possibility of the existence of the 1¢,4¢-iminopyrimi-
dine (IP) tautomer for the first time. This was impor-
tant because the earlier model for AP reactivity
typically assumed that the amino group, as a base,
would accept a proton. As more information became
available about protonation sites in aminopyridines
and aminopyrimidines, such as the nucleic bases, it
became clear that ring nitrogen protonation is pre-
ferred over protonation of the exocyclic amino group.
The hypothesis suggesting the AP moiety as an impor-

tant contributor to catalysis and the possibility for its
participation in acid ⁄ base catalysis [35] has gained
wider acceptance subsequent to the appearance of the
X-ray structures of ThDP enzymes. The following gen-
eralizations could be made on the basis of structural
observations that hold in virtually all of the ThDP
enzyme structures: (a) strong hydrogen bonds from the
protein to both the N1¢ atom (via a conserved Glu
with the exception of the enzyme GCL so far) and to
the N4¢-amino nitrogen atom on the side of the N3¢
atom of the ring; (b) an unusual V conformation
(describing the disposition of the AP and thiazolium
rings with respect to the bridging methylene group)
[36] rarely observed in model ThDP structures [37],
and predicted to be in a high energy region in van der
Waals conformational maps [38]; and (c) a surprisingly
short < 3.5 A
˚
distance between the AP amino nitro-
gen atom and the thiazolium C2 atom.
Detection of intermediates on ThDP
enzymes in solution
A number of methods now exists to monitor the
kinetic fate of each covalent ThDP-substrate interme-
diate along the catalytic cycle of various ThDP
enzymes represented by examples in Schemes 1–3
[10,14,15,39]. The three ThDP-bound intermediates in
Scheme 1 could be classified as: a pre-decarboxylation
intermediate C2a-lactylThDP (LThDP) or its analogs,
the first post-decarboxylation intermediate (the enam-

ine), and the second post-decarboxylation intermediate
C2a-hydroxyethylThDP (HEThDP) or its analogs. The
last one could also be construed as a product-ThDP
adduct for decarboxylases. A distinguishing feature of
these three intermediates is that the first (LThDP) and
third (HEThDP) have tetrahedral substitution at the
C2a atom, whereas the enamine being conjugated
should be trigonal planar at this position. Below, a
brief summary is given of the presence of various
ThDP intermediates on the enzymes, and the informa-
tion that has emerged regarding the state of ionization
and tautomerization of the AP ring on these intermedi-
ates. Understanding these issues is important with
respect to monitoring proton movements during
catalysis.
A convenient way to view ThDP-related and ThDP-
bound intermediates is to classify them as pre-, or
post-substrate (or substrate analog) binding.
ThDP-related intermediates prior to substrate
addition
For reasons mentioned earlier, during the recent past,
a need arose for the direct detection of various inter-
mediates shown in Schemes 1–3. Although the NMR
method developed by Tittmann and Hu
¨
bner [39] could
identify most of the covalent ThDP-bound substrates
and products on the pathway, the tautomeric forms
and ionization states of the 4¢-aminopyrimidine ring
along the reaction pathway and under the reaction

conditions remained to be elucidated.
The AP form of ThDP
The signature for this species is a negative CD band
centered near 320–330 nm and is well illustrated by the
enzyme BAL (Fig. 2). Although this CD band has long
been observed on the enzyme TK [40], it had been
suggested to be the result of a charge transfer transi-
tion between ThDP and an amino acid side chain on
TK, although early reports attributed it to the ThDP
itself. A number of studies at Rutgers on YPDC and
Enzyme-bound imino tautomer of thiamin diphosphate N. S. Nemeria et al.
2436 FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS
the first component of the Escherichia coli pyruvate
dehydrogenase complex (E1ec; EC 1.2.4.1.) and their
variants, as well as chemical model studies, strongly
suggest that this CD band is due to a charge transfer
transition between the AP ring as donor and the thia-
zolium ring as acceptor [41,42]. This CD band has
now been observed on a number of ThDP enzymes
(Table 1) and its detection strongly depends on pH
and, to a significant extent, on the enzyme environ-
ment.
The IP form of ThDP [41–45]
The notion that the AP could exist in the IP tauto-
meric form was suggested earlier by models attempting
to mimic the reactivity of such a tautomer. In the N1¢-
methylpyrimidinium, the pK
a
of the exocyclic amine is
reduced to approximately 12–12.5 [35,45], offering

rationalization for the presence of conserved glutamate
as a catalyst for the amino–imino tautomerization.
The positive charge on the 4¢-aminopyrimidinium ring
also induced differential exchange rates for the two
amino protons and the exchange was found to be buf-
fer catalyzed [46]. The first evidence for the possibility
that the IP tautomer may have a spectroscopic signa-
ture was found on the slow E477Q variant of YPDC
[43]. Inspired by these results, the old models were
dusted off and, in a series of chemical model studies,
Jordan et al. [43] and later Baykal et al. [45] showed
that an appropriate chemical model for the IP will give
rise to a UV absorption in the 300–310 nm range. Ser-
endipitously, the
15
N chemical shifts of the three
species on the left hand side of Schemes 1 and 2, the
two neutral and one positively charged forms of the
Fig. 2. CD detection of the AP form of ThDP on BAL. Inset: pH
dependence of the amplitude of the band for the AP form of ThDP.
Determination of pK
a
for the ([AP]+[IP]) ⁄ [APH
+
] equilibrium on BAL
[45].
Table 1. Assignment of the state of tautomerization of ThDP during the reaction pathway. ND, not detected.
ThDP intermediates IP positive CD, 300–314 nm AP negative CD, 320–330 nm References
ThDP E1h E1h [26,27,44]
POX POX

V51D GCL BAL (pH > 6.0)
BFDC (pH > 7.0)
Michaelis–Menten complex ND E91D YPDC-MAP [42,44,52]
E51D YPDC-MAP
YPDC-AcP
-
E571A E1ec-Py
E401K E1ec-Py
POX-AcP
)
Pre-decarboxylation reaction intermediate analog E91D YPDC-MAP [14,25,42,44,58]
E51D YPDC-MAP
YPDC-AcP
)
E1ec-MAP
E1ec-AcP
)
E1h-AcP
)
POX-AcP
)
BAL-BF
BAL-PPy
BFDC-MBP
BAL-MBP
Enamine (stopped-flow photodiode array) ND ND
Post-decarboxylation BAL-PAA ND [61]
YPDC+acetaldehyde
N. S. Nemeria et al. Enzyme-bound imino tautomer of thiamin diphosphate
FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS 2437

AP, are quite distinct [45]; early
15
N NMR experi-
ments on this issue were conducted by Cain et al. [47].
The recognition that the CD bands corresponding to
the AP and IP forms have different phases enables the
simultaneous observation of the two tautomeric forms,
notwithstanding the proximity of the bands to each
other, and also make CD the method of choice for
such studies. The signature for this IP species is a posi-
tive CD band centered near 300–314 nm (Table 1) and
is well illustrated on the first component of the human
pyruvate dehydrogenase complex (E1h), where both
the IP and AP tautomeric forms can be observed
simultaneously (Fig. 3).
To the authors’ knowledge, no electronic absorption
characteristic of the N1-protonated 4-aminopyrimidinium
form of ThDP or its C2-substituted derivatives
(APH
+
) or the ylide (Yl) has yet been proposed.
Determination of pK
a
for the enzyme-bound APH
+
form [26]
As the pH is lowered, the amplitude of the band for
the AP form diminishes and titrates with an apparent
pK
a

= 7.42 for the ([AP]+[IP]) ⁄ [APH
+
] equilibrium
on BAL (Fig. 2). This pK
a
in water for ThDP is 4.85
[47], whereas, on the enzymes, it is in the range 5.6–7.5
(Table 2) [26]. From the data provided in Table 2, it
was concluded that the pK
a
for the APH
+
coincides
with the pH of optimum activity for each enzyme,
indicating that all three forms (IP, AP and APH
+
)
must be readily accessible during the catalytic cycle.
The pK
a
elevation on the enzymes could be rational-
ized by the presence of the highly conserved glutamate
near the N1¢ position of ThDP, which would tend to
make the AP ring more basic. The tautomeric equilib-
rium constant K
tautomer
, in conjunction with the pK
a
led to a novel insight regarding ThDP catalysis, best
viewed by the thermodynamic box for enzymes that

are not substrate activated (Scheme 2, left hand side),
such as E1h and POX from L. plantarum. For these
enzymes, both the IP and AP forms could be moni-
tored over a wide pH range, providing both pK
a
and
K
tautomer
within reasonable error limits. The equilibria
shown in Schemes 1 and 2 are valid prior to addition
of substrate and lead to the tantalizing conclusions:
(a) on POX and E1h, pK
1¢
and pK
4¢
have similar mag-
nitudes; the enzymes shifted the pK
4
¢ from 12 in water
[35] to 5.6 and 7.0, respectively (see left triangle in
Schemes 1 and 2), and (b) with a known forward rate
constant from APH
+
to the Yl of approximately
50 s
)1
determined for E1h [48], and assuming a diffu-
sion-controlled reverse protonation rate constant of
10
10

s
)1
Æm
)1
(giving a pK
2
of 8.3 on E1h compared to
an estimate in water of 17–19) [49], it is possible to
speculate about the right triangle in Schemes 1 and 2.
The most important conclusion is that the proton-
transfer equilibrium constant for [IP] ⁄ [Yl] is 10
1
–10
2
on E1h. These thermodynamic parameters are the first
estimates on any ThDP enzyme and should be gener-
ally applicable to ThDP enzymes. The results also
suggest conditions under which a significant fraction
of the thiazolium ring may be in the conjugate base
ylide form.
The results provided in Table 2 also indicate that,
when the AP form is observable, below the pK
a
, the
APH
+
form likely exists, which comprises a form with
no known spectroscopic signature as far as we aware.
The C2-carbanion ⁄ ylide ⁄ carbene
According to the findings of Breslow, proton loss at

the thiazolium C2 position is required to initiate the
catalytic cycle. In 1997, there were two studies
reporting significant implications regarding this issue:
(a) Arduengo et al. [50] showed that the conjugate
Fig. 3. CD spectra of E1h titrated with ThDP. The spectra revealed
the presence of both the IP (at 305 nm) and AP (at 330 nm) tauto-
meric forms of ThDP [44].
Table 2. Correlation of pKa of enzyme bound APH+ and pH opti-
mum of enzyme activity.
Enzyme
pH optimal
activity
pKa for
([AP] + [IP]) ⁄ (APH
+
)
BAL 6.5–7.5 7.42 ± 0.02
BFDC 6.0–8.5 7.54 ± 0.11
POX 5.6–6.2 5.56 ± 0.03
E1h 7.0–7.5 7.07 ± 0.07
Enzyme-bound imino tautomer of thiamin diphosphate N. S. Nemeria et al.
2438 FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS
bases of imidazolium and indeed of thiazolium salts
could be generated and it was possible to study their
structure by NMR methods. In the intervening years,
some of these carbenes have been used in organometal-
lic reactions, including olefin metathesis. Arduengo
et al. [50] showed that the
13
C chemical shift of the C2

resonance shifted from 157 to 253 p.p.m. on conver-
sion of their model thiazolium compound to its conju-
gate base, thereby providing the all important guide
for future attempts to observe the ylide. (b) At the
same time, the group at Halle reported
13
C measure-
ments with specifically-labeled ThDP, according to
which, on the YPDC, the thiazolium ring C2H of
bound ThDP is in its undissociated state, both in the
absence and presence of the substrate activator surro-
gate pyruvamide (this enzyme has long been known to
be substrate activated); in other words, no evidence
was found for the presence of the conjugate base in
the activated or unactivated forms of YPDC [51].
It is important to emphasize that determination of
the state of ionization and tautomerization of enzyme-
bound ThDP by solution NMR methods poses several
challenges, both in the absence and presence of substit-
uents at the C2 atom: (a) the size of ThDP enzymes
(> 120 kDa) leads to broadened lines; (b) for many
ThDP enzymes, it is difficult to reversibly remove
ThDP and replace it with labeled coenzyme; and
(c) de novo synthesis required for specific labeling of
ThDP is time consuming and expensive.
Thiamin-bound intermediates with substrate or
substrate analog present
The Michaelis–Menten complex
Our earliest detection of an Michaelis–Menten com-
plex was on addition of a substrate analog methyl

acetylphosphonate (MAP) and acetylphosphinate
(AcP
)
) to several ThDP enzymes (Table 1). An exam-
ple is shown with AcP
)
added to YPDC (Fig. 4)
leading to a negative CD band at approximately 325–
335 nm, which is very reminiscent of the band
observed for the AP form [44].
Similar results were also seen when low concentra-
tions of pyruvate were added to E1ec [42]. Clear evi-
dence for the formation of the Michaelis–Menten
complex with a negative CD band near 320 nm was
also provided when adding pyruvate to the ‘inner loop’
E1ec variants [52]. Especially valuable support for the
claim that the Michaelis–Menten complex was indeed
being detected is provided by kinetic measurements:
stopped-flow photodiode array spectra in the absorp-
tion mode, as well as stopped-flow CD spectra at the
appropriate wavelength, showed formation of the
absorbance ⁄ CD band attributed to Michaelis–Menten
complex formation, within the dead-time of the
stopped-flow instruments (< 1 ms), as expected of a
noncovalent Michaelis–Menten complex [52].
From these results, we conclude that the Michaelis–
Menten complex is in the AP form.
The covalent substrate-ThDP pre-decarboxylation
complex (LThDP and analogs)
Observation of pre-decarboxylation intermediate derived

from aromatic substrates
In some favorable cases, such as with BAL, the posi-
tive CD band at 300–314 nm (Table 1) for the pre-
decarboxylation intermediate (via the IP form) could
be observed from the slow substrates benzoylformate
or phenylpyruvic acid [53]. This is plausible because
BAL, although a carboligase ⁄ lyase enzyme, also cata-
lyzes the decarboxylation of aromatic 2-oxoacids,
albeit very slowly.
Observation of stable pre-decarboxylation intermediates
derived from substrate analog phosphonates and
phosphinates
The initial identification of the IP form (positive CD
band, 300–314 nm) resulted from formation of a stable
pre-decarboxylation adduct of ThDP with: (a) MAP
[41,42] or AcP
)
[44], with pyruvate-specific enzymes
and (b) the aromatic 2-oxo acid analog methyl ben-
zoylphosphonate (MBP) with BFDC and BAL [25,53],
Fig. 4. CD spectra of YPDC in the presence of AcP
)
. The spectra
revealed the presence of the Michaelis–Menten complex in the AP
form (325–335 nm) and of the 1¢,4¢-iminophosphinolactyl-ThDP
covalent pre-decarboxylation intermediate in IP form (302 nm).
Inset: dependence of 1¢,4¢-iminophosphinolactyl-ThDP formation at
302 nm on [AcP
)
] [44].

N. S. Nemeria et al. Enzyme-bound imino tautomer of thiamin diphosphate
FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS 2439
according to Scheme 4. With six ThDP enzymes tested
so far (Table 1), the IP form appeared on the stopped-
flow time scale (either absorption or CD mode; for the
E1ec reaction with AcP
)
; see Fig. 5, top). The reaction
is efficiently catalyzed by all of the enzymes tested (for
E1h with AcP
)
, see Fig. 5, bottom; Scheme 4). An
important additional finding is shown in Fig. 4, result-
ing from mixing YPDC and AcP
)
[44]: because we are
observing evidence for the coexistence of the Michael-
is–Menten complex and the covalent pre-decarboxyl-
ation intermediate, the results are consistent with
‘alternating active site reactivity’ in a functional dimer,
as suggested for YPDC and BFDC [54–56]. We sug-
gested that, although one active center catalyzes the
pre-decarboxylation step, the other catalyzes the post-
decarboxylation events [55,56,57].
Formation of C2a-phosphonomandelylThDP on
BFDC from MBP and ThDP was also confirmed in
solution (FT-MS) [58], and that of C2a-phospho-
nolactylThDP (from MAP.ThDP) by X-ray methods
on E1ec [59] and POX [60].
Observation of pre-decarboxylation adducts of ThDP

with chromophoric substrate analogs
Recently, in three enzymes, YPDC, BFDC [58,61] and
BAL [53], the formation of the pre-decarboxylation
adduct formed with ThDP from a chromophoric sub-
strate analog (E)-2-oxo-4(pyridine-3-yl)-3-butenoic acid
(3-PKB) (as well as its ortho- and para isomers) was
also observed. In a series of studies on BAL [53],
BFDC [61] and YPDC (Fig. 6), the compound 3-PKB
provided outstanding information about the rates of
formation of two important intermediates, the pre-
decarboxylation LThDP analog and the enamine,
which were not readily available from other
experiments. At the same time, using (E)-3(pyridine-3-
yl)-2-propenal (PAA, the product of decarboxylation
of 3-PKB), provided not only information about the
second post-decarboxylation intermediate, but also
enabled us to assign the IP tautomeric form to both
tetrahedral, LThDP and HEThDP analogs (see
below).
The first post-decarboxylation intermediate: the
enamine ⁄ C2a-carbanion
According to Schemes 1 and 2, the enamine is the
only covalent thiamin-bound intermediate capable of
being conjugated. Electronic spectral observation of
the enzyme-bound enamine derived from aliphatic
substrates is difficult due to the expected k
max
near
290–295 nm, according to thiazolium-based models
[10,14].

With YPDC, BFDC and BAL, the enamine could
be observed directly near 430 nm with 3-PKB as alter-
nate substrate, as shown in Fig. 6 for YPDC.
The enamine intermediate derived from benzoylfor-
mate has been observed directly on the enzyme
BFDC at 390 nm [61]. We had modeled this enamine
with a k
max
of 380 nm) [10,14]. When BFDC was
reacted with the benzaldehyde product, there was
absorbance (and a CD band) at 390 nm, as predicted
by the chemical models, but no CD band was evident
in the 300–310 nm region, suggesting that the enam-
ine is not in the IP form [61]. Also, when (R)-benzoin
was added to BAL, the same CD band was formed
at 390 nm, indicating the slow release of the first
benzaldehyde, and the stability of the enamine in the
forward direction [53]. These experiments provided
fundamental information: (a) the ‘real’ enamine could
Scheme 4. Mechanism of formation of LThDP and analogue adducts.
Enzyme-bound imino tautomer of thiamin diphosphate N. S. Nemeria et al.
2440 FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS
be observed (due to its long k
max
at 390 nm) for the
first time derived from benzoin or benzaldehyde;
(b) the enamine may be in its APH
+
form, but not
in its IP form; and (c) because it gives rise to a CD

signal, the enamine is chiral on the enzyme by virtue
of the chirality induced by the enzyme, even though
it is planar and conjugated.
The enamine has also been detected indirectly using
the Tittmann and Hu
¨
bner method [39]. The method is
demonstrated with the E401K active center inner loop
variant of the E1ec (Fig. 7), where we used the synthetic
[C2,C6¢-
13
C
2
]ThDP enabling measurement of the rate of
enamine formation via HEThDP (unpublished results).
The labeled ThDP allowed observation of only those
protons directly attached to
13
C nuclei, simplifying
analysis in this otherwise busy aromatic region, espe-
cially for the pyruvate dehydrogenase complex, in which
there are three additional aromatic moieties (FAD,
NADH, CoA). Accumulation of the enamine ⁄ HEThDP,
but not of LThDP, suggests that decarboxylation is
faster than LThDP formation. Furthermore, for the
E401K E1ec variant, assembly to the complex appears to
accelerate the rate by a modest factor (Fig. 7).
The second post-decarboxylation intermediate, the
product-ThDP complex (HEThDP, C2a-hydroxy-
benzylThDP)

Clear evidence was obtained for HEThDP analog for-
mation from reacting PAA (i.e. the product of decar-
boxylation of 3-PKB) with BAL or BFDC [61]. The
structure of BFDC with both PAA and 3-PKB was
solved to high resolution [61]. The structure with PAA
clearly indicated: (a) covalent binding to ThDP as the
C2a-hydroxymethyl derivative with the vinylpyridyl
substituent attached to the C2a atom, (b) a tetrahedral
rather than trigonal environment at that atom because
10
4
3
2
Ellipticity (mdeg)
1
0
0123
Time (s)
45
8
6
4
2
Ellipticity (mdeg)
Ellipticity (mdeg)
0
–2
–4
–6
300 320 340

330 nm
305 nm
1–150 µ
M AcP
[AcP]/[E1 h] (µ
M/µM)
8
6
4
2
0
012345678910
360
Wavelen
g
th (nm)
380 400 420 440
Fig. 5. Formation of the pre-decarboxylation intermediate on the
PDHc-E1 component from AcP
)
. CD detection of the covalent
1¢,4¢-iminophosphinolactyl-ThDP intermediate on E1h from acetyl-
phosphinate (bottom) and the rate of 1¢,4¢- iminophosphinolactyl-
ThDP formation on E1ec by stopped-flow CD (top). Rate constants
of k
1
= 4.44 ± 0.34 s
)1
and k
2

= 0.593 ± 0.064 s
)1
were calculated
[44].
Wavelen
g
th (nm)
400 450 500 550 600
Relative absorbance
0.00
0.05
0.10
0.15
0.20
0.25
LThDP analogue (
ma x
473 nm )
Enamine
(
ma x
435 nm
)
Time (s)
0 10 20 30 40 50 60 7
0
Concentration (
M
)
0

2
4
6
8
10
12
14
16
18
[ES]
Enamine
LThDP analogue
k
2
= 0.507 + 0.002 s
–1

k
3
= 0.118 + 0.013 s
–1

Fig. 6. Reaction of YPDC with 3-PKB. Left: direct observation of the enamine at 435 nm on YPDC derived from 3-PKB by stopped-flow pho-
todiode array spectroscopy. Right: time course of intermediate formation after deconvolution of the spectrum. (S. Chakraborty, unpublished
data).
N. S. Nemeria et al. Enzyme-bound imino tautomer of thiamin diphosphate
FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS 2441
the planar thiazolium and vinylpyridine rings were not
coplanar with each other, as would be expected if the
electron density corresponded to the enamine (Fig. 8)

[61].
A striking confirmation of HEThDP formation
resulted from mixing acetaldehyde, the product of
pyruvate decarboxylation, with YPDC (Fig. 9), giving
the characteristic absorption for the IP form,
k
max
= 310 nm. Here, there is simply no alternative
assignment than to the IP form of HEThDP
(Scheme 2).
2-Acetylthiamin diphosphate and the
C2a-hydroxyethylidene cation radical
There is no evidence yet regarding their state of tauto-
merization ⁄ protonation.
Assignment of the state of ionization
and tautomerization to each
intermediate on the pathway
Perhaps our most significant observation is that, even
in the absence of pyruvate, both the AP and IP forms
are present on E1h and POX, but not on YPDC or
E1ec. The results on POX and E1h provide strong
direct support for interaction of ThDPs at the active
centers. The evidence reveals the presence of the IP
tautomer ready to initiate the intramolecular proton
transfer to produce the ylide prior to arrival of
substrate.
On four enzymes, where the AP form was detected,
it could be shown that the pK
a
of the 4¢-aminopyrimid-

inium on the enzyme coincides with the pH optimum
of the enzyme activity.
Time (s)
0 100 200 300
[HEThDP]/[active sites]
0.0
0.2
0.4
0.6
0.8
1.0
1.2
PDHc-ec k = 0.734 s
–1
E401K E1ec k = 0.0812
13
C
13
C
Fig. 7.
1
H NMR Distribution of ThDP intermediates on the E401K E1ec variant. Top: E401K inner loop E1ec variant reconstituted with E2ec
and E3ec and pyruvate analyzed by 1D HSQC NMR spectroscopy using the characteristic C6¢-H chemical shifts. Control, before adding pyru-
vate; 30 s after addition of pyruvate at 25 °C. Bottom: time course of HEThDP formation by E401K E1ec, and by the same variant reconsti-
tuted with the E2ec and E3ec components. To either E401K E1ec or PDHc reconstituted with this variant [C2,C6-
13
C
2
]ThDP and pyruvate
were added and, at the indicated times, the reaction was quenched into 12.5% TCA in 1

M DCl. (A. Balakrishnan, unpublished data).
Enzyme-bound imino tautomer of thiamin diphosphate N. S. Nemeria et al.
2442 FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS
Addition of phosphinate or phosphonate analogs of
pyruvate or benzoylformate to the seven enzymes
tested so far leads to highly efficient synthesis of stable
1¢,4¢-iminopyrimidinyl tautomeric analogs of LThDP.
These and other data, along with high resolution
X-ray structures, led us to conclude that ThDP-bound
intermediates with tetrahedrally substituted C2a atoms
exist in their IP or 1¢4¢-iminopyrimidinyl tautomeric
forms.
Of the other intermediates, we suggest that the Micha-
elis–Menten complex is in its AP form, whereas the
enamine is not in its IP form. We further suggest that
the 2-acylThDP and the C2a-hydroxyethylideneThDP
radical also are not in the IP form on account of the
hybridization at C2a in these intermediates.
An asymmetry of active centers is revealed by sev-
eral findings: (a) on POX and E1h (Fig. 3), the IP and
AP forms of ThDP coexist even in the absence of sub-
strates; (b) on POX and YPDC (Fig. 4) in the presence
AcP
)
, one active center is filled with IP (1¢,4¢-imi-
nophosphinolactyl-ThDP), the other with the M ichaelis–
Menten complex in the AP form.
With added AcP
)
, the four enzymes comprise two

distinct groups: E1ec and E1h with two active centers
revealed the presence of the 1¢,4¢-iminophosphinolac-
tyl-ThDP, whereas YPDC and POX with four active
centers revealed the additional presence of a Michaelis–
Menten complex. This could reflect the alternating
active center mechanism in homotetramers, already
identified on YPDC and BFDC (also a homotetrameric
ThDP enzyme).
We note that there are several recent examples in
the literature, both from our own work and that of
others, suggesting that, with tetrahedral substitution at
the C2a atom, the C2-C2a bond may be out of the
plane of the thiazolium ring [25,53,58,59]. This cer-
tainly suggests, but does not prove, that there is van
der Waals repulsion between the C2a substituent and
the 4¢-imino nitrogen under these conditions.
Table 1 summarizes the nature of the tautomeric
form suggested for all of the intermediates on the
ThDP pathways and provides the supporting evidence
for them. Curiously, the mechanisms could be
expounded invoking solely the IP and APH
+
forms,
whereas, to date, the canonical AP form has only been
identified in the Michaelis–Menten complex. The
results suggest that, for several steps, there are proton
transfers in the reaction pathway that are required to
ensure the presence of the appropriate tautomeric form
for the intermediate. This is evident from Scheme 2,
where the tautomeric ⁄ ionization state of the AP is

assigned to each intermediate.
Prospects
Further examples, and alternative methods to confirm
the electronic spectroscopic assignments would be
Fig. 9. Stopped-flow photodiode array spectra of YPDC with product
of pyruvate decarboxylation, acetaldehyde. Spectra show formation
of the IP form of HEThDP (maximum at 310 nm) (S. Chakraborty,
unpublished data).
AB
Fig. 8. Crystal structure of BFDC with PAA
added. (a) Evidence for covalent PAA-ThDP
adduct formation. (b) Tetrahedral environ-
ment around C2a supported by the finding
that the pyridyl and thiazolium rings are not
coplanar with each other, thereby ruling out
the enamine [61].
N. S. Nemeria et al. Enzyme-bound imino tautomer of thiamin diphosphate
FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS 2443
desirable, although we believe that the spectral assign-
ments of the IP and AP forms are on a firm footing in
view of the similar observations made so far for eight
enzymes with several substrates and substrate analogs.
Of course, it would also be desirable to be able to
detect the APH
+
and the ylide forms.
As with any novel finding, the observation of the
1¢,4¢-imino tautomeric form on the addition of ThDP
itself to the enzymes, even in the absence of substrate
or substrate analog, also raises many interesting ques-

tions, some of which are summarized below.
First, why does the negative CD band at 320–
330 nm correspond to the AP form observed in some
and but not in other enzymes? A plausible explanation
is that, in the absence of an Asp or Glu negative
charge proximal to the thiazolium ring, the charge
transfer interaction (AP donor and thiazolium accep-
tor) is stronger; hence, the band is visible. This is cer-
tainly the case with BAL, BFDC, E1h and POX.
Although, in TK, substitution of the negatively-
charged side chain closest to the thiazolium ring
reduced the amplitude to 20% of its original ampli-
tude, the band was not eliminated [62,63]. By contrast,
on both E1ec and YPDC, there are several potentially
negatively-charged residues proximal to the thiazolium
ring and, indeed, we do not see the band correspond-
ing to the AP form.
Second, why is the IP form that is observed in some
enzymes usually accompanied by the AP form?
Third, what is the tautomeric form of ThDP on
GLC, as there is no conserved glutamate present?
Fourth, what is the energetic cost to the enzyme of
stabilizing the IP form at the active center and from
where is that energy obtained?
Fifth, does the simultaneous presence of the IP and
AP forms on POX and E1h imply half-of-the-sites
reactivity?
Finally, where both the IP and AP forms are present
simultaneously, is it the reflection of the so-called ‘pro-
ton wire’ mechanism [64] via acidic residues connecting

the ThDPs in adjacent active centers? According to
this hypothesis, two active centers on ThDP enzymes
are synchronized by reversible proton transfer.
Some of these questions could be addressed by
experiments, whereas others will need to be addressed
by computational methods.
Acknowledgements
The authors are grateful to the National Institutes of
Health for grant GM050380 for financial support.
They also are grateful to all the collaborators whose
names appear on the publications.
References
1 Krampitz LO (1969) Catalytic functions of thiamin
diphosphate. Annu Rev Biochem 38, 213–240.
2 Krampitz LO (1970) Thiamin Diphosphate and its Cata-
lytic Functions. Marcel Dekker, New York.
3 Sable HZ & Gubler CJ (1982) Thiamin, 20 years of
progress. Ann N Y Acad Sci 378, 7–12.
4 Kluger R (1987) Thiamin diphosphate: a mechanistic
update on enzymic and nonenzymic catalysis of decar-
boxylation. Chem Rev 87, 863–876.
5 Schellenberger A & Schowen RL (1988) Thiamin Pyro-
phosphate Biochemistry, Vol. 1. CRC Press, Boca
Raton, FL.
6 Bisswanger H & Ullrich J (1991) Biochemistry and Phys-
iology of Thiamin Diphosphate Enzymes. VCH Publish-
ers, Weinheim.
7 Bisswanger H & Schellenberger A (1996) Biochemistry
and Physiology of Thiamin Diphosphate Enzymes. A.u.C.
Intemann, Wissenschaftlicher Verlag, Prien.

8 Schellenberger A (1998) Sixty years of thiamin diphos-
phate biochemistry. Biochim Biophys Acta 1385, 177–186.
9 Jordan F (1999) Interplay of organic and biological
chemistry in understanding coenzyme mechanisms:
example of thiamin diphosphate-dependent decarboxy-
lations of 2-oxo acids. FEBS Lett 457, 298–301.
10 Jordan F (2003) Current mechanistic understanding of
thiamin diphosphate-dependent enzymatic reactions.
Nat Prod Rep 20, 184–201.
11 Jordan F & Patel MS (2004) Thiamine. Catalytic Mech-
anisms in Normal and Disease States. Marcel Dekker.
Inc., New York.
12 Pohl M, Sprenger GA & Mu
¨
ller M (2004) A new per-
spective on thiamine catalysis. Curr Opin Biotechnol 15,
335–342.
13 Frank RA, Leeper FJ & Luisi BF (2007) Structure,
mechanism and catalytic duality of thiamine-dependent
enzymes. Cell Mol Life Sci 64, 892–905.
14 Jordan F & Nemeria N (2005) Experimental observa-
tion of thiamin diphosphate-bound intermediates on
enzymes and mechanistic information derived from
these observations. Bioorg Chem 33, 190–215.
15 Kluger R & Tittmann K (2008) Thiamin diphosphate
catalysis: enzymic and nonenzymic covalent intermedi-
ates. Chem Rev 108, 1797–1833.
16 Liu M, Sergienko EA, Guo F, Wang J, Tittmann K,
Hu
¨

bner G, Furey W & Jordan F (2001) Catalytic acid-
base groups in yeast pyruvate decarboxylase I. Site
directed mutagenesis and steady-state kinetic studies on
the enzyme with the D28A, H114F, H115F and E477Q
substitutions. Biochemistry 40, 7355–7368.
17 Yep A, Kenyon GL & McLeish MJ (2008) Saturation
mutagenesis of putative catalytic residues of benzoylfor-
mate decarboxylase provides a challenge to the accepted
mechanism. Proc Natl Acad Sci USA 105, 5733–5738.
Enzyme-bound imino tautomer of thiamin diphosphate N. S. Nemeria et al.
2444 FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS
18 Polovnikova ES, McLeish MJ, Sergienko EA, Burgner
JT, Anderson NL, Bera AK, Jordan F, Kenyon GL &
Hasson MS (2003) Structural and kinetic analysis of
catalysis by a thiamin diphosphate-dependent enzyme,
benzoylformate decarboxylase. Biochemistry 39, 13862–
13869.
19 Fersht A. (1998) Structure and Mechanism in Protein Sci-
ence, A Giude to Enzyme Catalysis and Protein Folding.
W.H. Freeman and Co, New York, NY, pp. 474–475.
20 Carter P & Wells JA (1988) Dissecting the catalytic
triad of a serine protease. Nature 332, 564–568.
21 Kuliopulos A, Talalay P & Mildvan AS (1990) Com-
bined effects of two mutations of catalytic residues on
the ketosteroid isomerase reaction. Biochemistry 29,
10271–10280.
22 Henot F & Pollack RM (2000) Catalytic activity of the
D38A mutant of 3-oxo-D-5-steroid isomerase: recruit-
ment of aspartate-99 as the base. Biochemistry 39,
3351–3359.

23 Mosbacher TG, Mu
¨
ller M & Schulz GE (2005) Struc-
ture and mechanism of the ThDP-dependent benzalde-
hyde lyase from Pseudomonas fluorescencs. FEBS J 272,
6067–6076.
24 Kneen MM, Pogozheva ID, Kenyon GL & McLeish
MJ (2005) Exploring the active site of benzaldehyde
lyase by modeling and mutagenesis. Biochim Biophys
Acta Proteins Proteom 1753, 263–271.
25 Brandt GS, Nemeria N, Chakraborty S, McLeish MJ,
Yep A, Kenyon GL, Petsko GA, Jordan F & Ringe D
(2008) Probing the active center of benzaldehyde lyase
with substitutions and the pseudosubstrate analogue
benzoylphosphonic acid methyl ester. Biochemistry 47,
7734–7743.
26 Nemeria N, Korotchkina L, McLeish MJ, Kenyon GL,
Patel MS & Jordan F (2007) Elucidation of the chemis-
try of enzyme-bound thiamin diphosphate prior to sub-
strate binding: defining internal equilibria among
tautomeric and ionization states. Biochemistry 46,
10739–1044.
27 Kaplun A, Binshtein E, Vyazmensky M, Steinmetz A,
Barak Z, Chipman DM, Tittmann K & Shaanan B
(2008) Glyoxylate carboligase lacks the canonical active
site glutamate of thiamine-dependent enzymes. Nat
Chem Biol 4, 113–118.
28 Lindqvist Y, Schneider G, Ermler U & Sundstro
¨
mM

(1992) Three-dimensional structure of transketolase,
a thiamine diphosphate dependent enzyme, at 2.5 A
˚
res-
olution. EMBO J 11, 2373–2379.
29 Muller Y & Schulz G (1993) Structure of the thiamine-
and flavin-dependent enzyme pyruvate oxidase. Science
259, 965–967.
30 Dyda F, Furey W, Swaminathan S, Sax M, Farrenkopf
B & Jordan F (1993) Catalytic centers in the thiamin
diphosphate dependent enzyme pyruvate decarboxylase
at 2.4-A
˚
resolution. Biochemistry 32, 6165–6170.
31 Arjunan P, Umland T, Dyda F, Swaminathan S, Furey
W, Sax M, Farrenkopf B, Gao Y, Zhang D & Jordan
F (1996) Crystal structure of the thiamin diphosphate-
dependent enzyme pyruvate decarboxylase from the
yeast Saccharomyces cerevisiae at 2.3 A
˚
resolution.
J Mol Biol 256, 590–600.
32 Hawkins CF, Borges A & Perham RN (1989) A com-
mon structural motif in thiamin pyrophosphate-binding
enzymes. FEBS Lett 255, 77–82.
33 Breslow R (1957) Rapid deuterium exchange in thiazoli-
um salts. J Am Chem Soc 79, 1762–1763.
34 Alvarez FJ, Ermer J, Hu
¨
bner G, Schellenberger A &

Schowen RL (1995) The linkage of catalysis and regula-
tion in enzyme action. Solvent isotope effects as probes
of protonic sites in the yeast pyruvate decarboxylase
mechanism. J Am Chem Soc 117, 1678–1683.
35 Jordan F & Mariam YH (1978) N1¢-Methylthiaminium
diiodide. Model study on the effect of a coenzyme
bound positive charge on reaction mechanisms requiring
thiamin pyrophosphate. J Am Chem Soc 100, 2534–
2541.
36 Muller YA, Lindqvist Y, Furey W, Schulz GE, Jordan
F & Schneider GA (1993) Thiamin diphosphate binding
fold revealed by comparison of the crystal structures of
transketolase, pyruvate oxidase and pyruvate decarbox-
ylase. Structure 1, 95–103.
37 Sax M, Pulsinelli P & Pletcher J (1974) Stereochemistry
of intermediates in thiamine catalysis. I. Crystal struc-
tures of 2-(a-hydroxyethyl)-3,4-dimethylthiazolium bro-
mide and dl-2-(a-hydroxyethyl)thiamine chloride
hydrochloride. J Am Chem Soc 96, 155–165.
38 Jordan F (1976) Theoretical calculations on thiamine
and related compounds. II. Conformational analysis
and electronic properties of 2-(a-hydroxyethyl)thiamine.
J Am Chem Soc 98, 808–813.
39 Tittmann K, Golbik R, Uhlemann K, Khailova L,
Schneider G, Patel M, Jordan F, Chipman DM, Dugg-
leby RG & Hu
¨
bner G (2003) NMR analysis of covalent
intermediates in thiamin diphosphate enzymes. Biochem-
istry 42, 7885–7891.

40 Kochetov GA & Usmanov RA (1970) Charge transfer
interactions in transketolase-thiamine pyrophosphate
complex. Biochem Biophys Res Commun 41, 1134–
1140.
41 Jordan F, Nemeria NS, Zhang S, Yan Y, Arjunan P &
Furey W (2003) Dual catalytic apparatus of the thiamin
diphosphate coenzyme: acid-base via the 1¢,4¢-iminopyr-
imidine tautomer along with its electrophilic role. JAm
Chem Soc 125, 12732–12738.
42 Nemeria N, Baykal A, Joseph E, Zhang S, Yan Y,
Furey W & Jordan F (2004) Tetrahedral intermediates
in thiamin diphosphate-dependent decarboxylations
exist as a 1¢,4¢-imino tautomeric form of the coenzyme,
unlike the Michaelis complex or the free coenzyme. Bio-
chemistry 43, 6565–6575.
N. S. Nemeria et al. Enzyme-bound imino tautomer of thiamin diphosphate
FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS 2445
43 Jordan F, Zhang Z & Sergienko E (2002) Spectrosco
pic evidence for participation of the 1¢,4¢- imino tauto-
mer of thiamin diphosphate in catalysis by yeast pyru-
vate decarboxylase. Bioorg Chem 30, 188–98.
44 Nemeria N, Chakraborty S, Baykal A, Korotchkina
LG, Patel MS & Jordan F (2007) The 1¢,4¢-iminopyrim-
idine tautomer of thiamin diphosphate is poised for
catalysis in asymmetric active centers on enzymes. Proc
Natl Acad Sci USA 104, 78–82.
45 Baykal AT, Kakalis L & Jordan F (2006) Electronic
and nuclear magnetic resonance spectroscopic features
of the 1¢,4¢-iminopyrimidine tautomeric form of thiamin
diphosphate, a novel intermediate on enzymes requiring

this coenzyme. Biochemistry 45, 7522–7528.
46 Jordan F (1982) Role of the aminopyridine ring in thia-
min-catalyzed reactions. II. Proton NMR evidence for
high barriers to amino group rotation in 4-aminopyrim-
idines, including thiamin, at low pH in water. J Org
Chem 47, 2748–2753.
47 Cain AH, Sullivan GR & Roberts JD (1977) The pro-
tonation site of vitamin B1 as determined from natural-
abundance nitrogen-15 nuclear magnetic resonance
spectra. J Am Chem Soc 99, 6423–6425.
48 Seifert F, Golbik R, Brauer J, Lilie H, Schro
¨
der-
Tittmann K, Hinze E, Korotchkina LG, Patel MS &
Tittmann K (2006) Direct kinetic evidence for half-of-
the-sites reactivity in the E1 component of the human
pyruvate dehydrogenase multienzyme complex through
alternating sites cofactor activation. Biochemistry 45,
12775–12785.
49 Washabaugh MW & Jencks WP (1988) Thiazolium
C(2)-proton exchange: structure-reactivity correlations
and the pK
a
of thiamin C(2)-H revisited. Biochemistry
27, 5044–5053.
50 Arduengo AJ, Goerlich JR & Marshall WJ (1997) A
stable thiazol-2-ylidene and its dimer. J Liebigs Ann
Chem 2, 365–374.
51 Kern D, Kern G, Neef H, Tittmann K, Killenberg-Jabs
M, Wikner C, Schneider G & Hu

¨
bner G (1997) How
thiamine diphosphate is activated in enzymes. Science
275, 67–70.
52 Kale S, Ulas G, Song J, Brudvig GW, Furey W & Jor-
dan F (2008) Efficient coupling of catalysis and dynam-
ics in the E1 component of Escherichia coli pyruvate
dehydrogenase multienzyme complex. Proc Natl Acad
Sci USA 105, 1158–63.
53 Chakraborty S, Nemeria N, Yep A, McLeish MJ, Ken-
yon GL & Jordan F (2008) Mechanism of benzaldehyde
lyase studied via thiamin diphosphate-bound intermedi-
ates and kinetic isotope effects. Biochemistry 47, 3800–
3809.
54 Sergienko EA & Jordan F (2001) Catalytic acid-base
groups in yeast pyruvate decarboxylase. 3. A steady-
state kinetic model consistent with the behavior of both
wild-type and variant enzymes at all relevant pH values.
Biochemistry 40, 7382–7403.
55 Sergienko EA & Jordan F (2002) New model for activa-
tion of yeast pyruvate decarboxylase by substrate con-
sistent with the alternating sites mechanism:
demonstration of the existence of two active forms of
the enzyme. Biochemistry 41, 3952–3967.
56 Sergienko EA, Wang J, Polovnikova L, Hasson MS,
McLeish MJ, Kenyon GL & Jordan F (2000) Spectro-
scopic detection of transient thiamin diphosphate-bound
intermediates on benzoylformate decarboxylase. Bio-
chemistry 39, 13862–13869.
57 Jordan F, Nemeria NS & Sergienko EA (2005) Multiple

modes of active center communication in thiamin
diphosphate-dependent enzymes. Acc Chem Res 38,
755–763.
58 Brandt GS, Kneen MM, Chakraborty S, Baykal A,
Nemeria N, Yep A, Ruby DI, Petsko GA, Kenyon GL,
McLeish MJ et al. (2009) Snapshot of a reaction inter-
mediate: analysis of benzoylformate decarboxylase in
complex with a benzoylphosphonate inhibitor. Biochem-
istry (in press).
59 Arjunan P, Sax M, Brunskill A, Chandrasekhar K,
Nemeria N, Zhang S, Jordan F & Furey W (2006) A
thiamin-bound, pre-decarboxylation reaction intermedi-
ate analogue in the pyruvate dehydrogenase E1 subunit
induces large scale disorder-to-order transformations in
the enzyme and reveals novel structural features in the
covalently bound adduct. J Biol Chem 281, 15296–
15303.
60 Wille G, Meyer D, Steinmetz A, Hinze E, Golbik R &
Tittmann K (2006) The catalytic cycle of a thiamin
diphosphate enzyme examined by cryocrystallography.
Nat Chem Biol 2, 324–328.
61 Chakraborty S, Nemeria NS, Balakrishnan A, Brandt
GS, Kneen MM, Yep A, McLeish MJ, Kenyon GL,
Petsko GA, Ringe D et al. (2009) Detection and time-
course of formation of all thiamin diphosphate-bound
covalent intermediates derived from a chromophoric
substrate analogue on benzoylformate decarboxylase.
Biochemistry 48, 981–994.
62 Wikner C, Meshalkina L, Nilson U, Nikkola M, Lindq-
vist Y, Sundstro

¨
m M & Schneider G (1994) Analysis of
an invariant cofactor-protein interaction in thiamin
diphosphate-dependent enzymes by site-directed muta-
genesis. J Biol Chem 269, 32144–32150.
63 Meshalkina L, Nilsson U, Wikner C, Kostikova T &
Schneider G (1997) Examination of the thiamin diphos-
phate binding site in yeast transketolase by site-directed
mutagenesis. Eur J Biochem 244, 646–652.
64 Frank RA, Titman CM, Pratap JV, Luisi BF & Perham
RN (2004) A molecular switch and proton wire syn-
chronize the active sites in thiamine enzymes. Science
306, 818–820.
Enzyme-bound imino tautomer of thiamin diphosphate N. S. Nemeria et al.
2446 FEBS Journal 276 (2009) 2432–2446 ª 2009 The Authors Journal compilation ª 2009 FEBS