The linkage of catalysis and regulation in enzyme action: Oxidative diversion in the hysteretically regulated yeast pyruvate decarboxylase

Article abstract of DOI:10.1016/S0968-0896(98)00269-7

The reaction catalyzed by the thiamin-diphosphate-dependent yeast pyruvate decarboxylase, which is hysteretically regulated by pyruvate, undergoes paracatalytic oxidative diversion by 2,6-dichlorophenolindophenol, which traps a carbanionic intermediate an

Full text of DOI:10.1016/S0968-0896(98)00269-7

Bioorganic & Medicinal Chemistry 7 (1999) 887±894
The Linkage of Catalysis and Regulation in Enzyme Action:
Oxidative Diversion in the Hysteretically Regulated Yeast
Pyruvate Decarboxylase
Gholamhossein Hajipour, K. Barbara Schowen and Richard L. Schowen*
Department of Chemistry and Higuchi Biosciences Center, University of Kansas, Lawrence, KS 66047, USA
Received 24 September 1998; accepted 17 November 1998
AbstractÐThe reaction catalyzed by the thiamin-diphosphate-dependent yeast pyruvate decarboxylase, which is hysteretically
regulated by pyruvate, undergoes paracatalytic oxidative diversion by 2,6-dichlorophenolindophenol, which traps a carbanionic
intermediate and diverts the product from acetaldehyde to acetate (Christen, P. Meth. Enzymol. 1977, 46, 48). This reaction is now
shown to exhibit an oxidant on-rate constant somewhat faster than that for pyruvate in the normal catalytic cycle and a product
o€-rate constant about 60-fold smaller than that for acetaldehyde. Both on-rates and o€-rates exhibit an inverse solvent isotope
e€ect of 1.5-2, observed in normal catalysis as a signal of sulfhydryl addition to the keto group of pyruvate at the allosteric reg-
ulatory site. The ®ndings are consistent with a model for regulation in which the sulfhydryl-addition process mediates access to a
fully catalytically competent active site, the oxidative-diversion reaction being forced to make use of the normal entry±exit
machinery. # 1999 Elsevier Science Ltd. All rights reserved.
Introduction
acid cycle and the electron-transport chain, a metabolic
pathway that produces a high yield of energy stored in
adenosine triphosphate. The organism is also capable in
the absence of oxygen of fermentative metabolism to
produce a lower energy yield. SCPDC occupies the
branch point between these metabolic routes and, per-
haps as a result, is regulated by its own substrate in an
unusual hysteretic, or time-dependent, manner. The
regulatory action may protect the organism from losses
of energy that would occur if pyruvate from glycolysis
were partially diverted into the fermentative pathway
when oxygen is available for the more productive elec-
tron-transport pathway.
The interface of chemistry and biology is a region of
which Professor Sir Derek H. R. Barton was a pioneer
explorer and a region of which he was a comfortable
and in¯uential inhabitant. His major instrument of
exploration was the explanation of biological phenom-
enology in the rigorous language of chemistry, com-
monly at the level of molecular structure and often at
the submolecular level. In this paper, which we dedicate
respectfully to his memory, we attempt on a scale very
modest by Barton standards a chemical examination of
features linking catalysis and regulation in enzyme
action.
The hysteretic regulatory feature of SCPDC renders the
enzyme completely inactive upon initial contact with the
substrate pyruvate (i.e., the asymptotic rate of product
formation is zero at the time of exposure to pyruvate).
In a period of typically 10±30 seconds, the rate of pro-
duct formation relaxes to the steady-state value. The
rate of relaxation is saturable in pyruvate. The steady-
state velocity is second order in pyruvate at low con-
centrations, re¯ecting the requirement for the enzyme
®rst to combine with a `regulatory pyruvate' and then
with a `catalytic pyruvate'. At higher concentrations as
the regulatory interaction becomes saturated, the
steady-state velocity becomes ®rst order in pyruvate
and, ®nally, when the catalytic reaction becomes satu-
rated, the rate becomes zero order in pyruvate. Surro-
gate regulators that possess a carbonyl group but do not
Yeast pyruvate decarboxylase1
The pyruvate decarboxylase (EC 4.1.1.1) of the yeast
Saccharomyces cerevisiae (SCPDC) is a thiamin-dipho-
sphate (ThDP)-dependent enzyme that catalyzes the
conversion of pyruvic acid to carbon dioxide and acet-
aldehyde (Fig. 1). Its thoroughly investigated catalytic
mechanism and its striking regulatory properties make
it an e€ective system in which to explore how catalysis
and regulation are linked in enzyme action. The yeast is
capable in the presence of oxygen of the oxidative
metabolism of carbohydrates through the tricarboxylic-
*
5
Corresponding author. Tel.: +1-785-864-4080; fax: +1-785-864-
349; e-mail: rschowen@ukans.edu
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(98)00269-7

8
88
G. Hajipour et al. / Bioorg. Med. Chem. 7 (1999) 887±894
Figure 1. The normal catalytic cycle for SCPDC. Note the structure of thiamin diphosphate (ThDP) with a diphosphate arm and a pyrimidine arm.
Ionization of the acidic C2±H bond leads to the `®rst carbanion'. Addition of this species to the substrate pyruvate produces decarboxylation and
formation of the `second carbanion'. Protonation of the latter and elimination to form acetaldehyde completes the catalytic cycle. The `second car-
banion' is denoted SE*M in the system20 used herein for mechanistic descriptions (see Figs 2 and 5).
undergo decarboxylation (pyruvamide and ketomalo-
nate) can replace pyruvate in the regulatory interaction.
The steady-state velocity is then hyperbolically depen-
dent on pyruvate.
oxylase of the bacterium Zymomonas mobilis,9±12 in
which a cysteine residue is not present at the position
corresponding to Cys-221 in SCPDC. Earlier work had
1
3
already implicated sulfhydryl groups in regulation.
Surrogate regulators of SCPDC can replace the `reg-
ulatory pyruvate', as already mentioned. All such spe-
cies possess a carbonyl group and this fact, together
with the strong evidence for the role of Cys-221 have led
to the concept that at some point in the regulatory pro-
cess, the sulfhydryl group of Cys-221 adds to the car-
bonyl group of the regulatory molecule.
Catalysis by SCPDC1±3
Catalysis by yeast pyruvate decarboxylase is typical of
ThDP-dependent decarboxylases (Fig. 1). The enzyme
possesses three domains.4
domain and a pyrimidine-binding domain serve to
A
diphosphate-binding
anchor the cofactor with its thiazolium C ±H bond
2
directed toward the presumed pyruvate binding site.
The third domain is known as the regulatory domain;
we will return to it below. Ionization of the C ±H bond
Models for PDC regulation: active-site reorganization
2
leads to addition across the pyruvate carbonyl group. In
the adduct, ®ssion of the C±C bond in decarboxylation
is promoted by electron delocalization into the thiazo-
lium nucleus. The `second carbanion' formed as the ®rst
product of decarboxylation is then protonated. An
elimination reaction produces the product acetaldehyde
and regenerates the cofactor.
Jordan, Furey, and their co-workers have made use of
mutagenesis experiments14 and of the crystal structure
to propose a model for the regulatory activation of
SCPDC. They employ the consensus hypothesis that
addition to the keto group of pyruvate by Cys-221-SH
is important, and they further suppose that this is the
event that triggers activation of the enzyme. They have
shown that residues that form a network of interactions
reaching from Cys-221 to the active site are important
for activation and they therefore propose that the net-
work transfers the information to the active site that the
activation event has occurred, the resulting active-site
reorganization then giving rise to full catalytic activity.
4
Regulation of SCPDC5±7
Hysteretic regulation of pyruvate is critically depen-
8
dent on the residue Cys-221, which is located in the
regulatory domain about 2 nm from the thiazolium
nucleus. When this residue is mutated to serine, about
1
5% of the enzyme activity remains but the kinetics
Models for PDC regulation: active-site access
become simply hyperbolic and the initial relaxation
period disappears: the enzyme has become unregulated.
Some natural pyruvate decarboxylases are also unregu-
lated, displaying hyperbolic kinetics and no initial
relaxation period. One of these is the pyruvate decarb-
We have favored a di€erent model,15,16 illustrated in
Figure 2. According to this model, the time-dependent
activation represents a `seating' of the regulatory pyr-
uvate, perhaps by coordination of its carboxylate group

G. Hajipour et al. / Bioorg. Med. Chem. 7 (1999) 887±894
889
Figure 2. The active-site access model for the hysteretic, allosteric regulation of SCPDC. The slow, hysteretic activation is assumed to involve pla-
cement of the `regulatory pyruvate' at the allosteric site of the free enzyme E, in which active-site access is denied. Binding at the allosteric site is
indicated by writing S to the left of E, and completion of the slow activation process by the asterisk in SE*. During the catalytic cycle, addition of
the sulfhydryl group of Cys-221 to the keto group of the regulator opens the active site and elimination recloses it. The substrate enters the active site
of SE* during an open period to produce SE*S. The active site is presumed closed during decarboxylation, which yields SE*M. Protonation and
elimination then gives the product complex and product departs during an open period.
to the active-site magnesium ion (which is also coordi-
nated to the diphosphate unit of ThDP). Thereafter
sulfhydryl addition to the keto group moves a loop that
denies access to the active site, thus opening it for sub-
strate entry or product departure. The main evidence
for this view is that an inverse solvent isotope e€ect
around 2 (i.e., the rate is twofold faster in deuterium
oxide than protium oxide) is observed both in substrate
binding to the activated enzyme (in every catalytic cycle)
and in product release to regenerate the activated
enzyme (also in every catalytic cycle). An inverse iso-
alytic machinery of SCPDC is fully functional in the
inactive enzyme but the enzyme cannot bind substrate
or release product.
Paracatalytic intervention in SCPDC action by oxidizing
agents
A useful approach to contributing to a discrimination
between the two regulatory models is o€ered by an
1
9
observation of Christen and his co-workers. They
noted that oxidizing agents such as dichloro-
phenolindophenol (D; Fig. 3) can reduce the rate of
acetaldehyde formation during the catalytic cycle of
tope e€ect of around twofold is expected for sulfhydryl
addition reactions.1
7,18
According to this view, the cat-

8
90
G. Hajipour et al. / Bioorg. Med. Chem. 7 (1999) 887±894
Figure 3. The concept of paracatalytic oxidative diversion during SCPDC action.19 The oxidant dichlorophenolindophenol (D) traps the product of
decarboxylation, the `second carbanion' SE*M. Oxidation leads to acetyl-ThDP which then hydrolyzes to release acetate and complete the catalytic
cycle.
SCPDC although the agents have no e€ect on SCPDC
Figure 4 shows speci®c velocities as a function of the
concentration of D in both protium and deuterium oxi-
des. The data in each case were ®tted to eq (1). The
value of the inhibition constant Kinh is poorly deter-
mined in both cases (339‹130 mM in protium oxide,
307‹109 mM in deuterium oxide), as is also the case for
the similar constant of similar magnitude that is
observed in the normal kinetics of SCPDC action. As
19
in the absence of pyruvate. Christen termed such
reactions paracatalytic reactions; it seems very likely
that an intermediate in the normal enzymic reaction is
being oxidatively trapped and diverted to the formation
of alternative products. In the case of D, the product
was shown to be acetic acid. Thus the likely route of
diversion is trapping of the `second carbanion' that
results from decarboxylation, as shown in Figure 3. This
converts the intermediate to acetyl-ThDP and alters the
mechanism of the product release event. Solvent isotope
e€ects on the unnatural product release event could
therefore be informative about the regulation of pro-
duct release.
2
0
was done in the treatment of the normal kinetics,
a
constant value equal to the mean (323 mM) of the ®tted
values for the two isotopic solvents was therefore used
for both data sets. The values of the parameters of eq
(1) and the isotope e€ects resulting from them are
shown in Table 1.
We report here the kinetic and isotope-e€ect character-
ization of the oxidative diversion reaction of D in the
action of SCPDC.
Results
The oxidative diversion of the SCPDC-catalyzed dec-
arboxylation of pyruvate by dichlorophenolindophenol
(D), followed by the spectral change at 520 nm atten-
dant upon reduction of D, exhibits roughly hyperbolic
saturation (Fig. 4). The concentration of pyruvate in
these experiments was 30 mM so that the normal cata-
lytic kinetics were fully saturated. Some inhibition by D
appears to set in at high concentrations. The velocity n
can be described by eq (1), where e is the enzyme con-
centration (molarity of tetrameric SCPDC) and the
other constants are de®ned below.
o
Figure 4. Speci®c velocity of the oxidative diversion reactions versus
oxidant concentration at a saturating concentration of pyruvate. Fil-
led circles refer to reaction in deuterium oxide, ®lled squares to reac-
tion in protium oxide. The solid lines are plots of eq (1) with the
parameters shown in Table 1.
È
�
È
ÉꢀÉ
n=eo ˆ kox‰DŠ= Kox ‡ ‰DŠ 1 ‡ Kinh=‰DŠ
ꢀ1†

G. Hajipour et al. / Bioorg. Med. Chem. 7 (1999) 887±894
891
Table 1. Kinetic constants for the SCPDC-catalyzed oxidative
diversion of pyruvate by dichlorophenolindophenol in protium and
deuterium oxides at 30 C and pH 6.2 and equivalent
The expression for k /K can also be considered rea-
ox ox
sonable. The ®rst factor in brackets gives the steady-
state second-order rate constant for conversion of
SE*S+D to DH . The second factor in brackets allows
ꢀ
Kinetic constanta
H
2
O
D
2
O
Isotope e€ects
O/D
2
H
2
2
O
for the possible accumulation of SE*M, which would be
inhibitory for the oxidative-diversion reaction on this
model. We know from studies of the normal catalytic
reaction that this factor is about 1/2 in protium oxide
�
1
k
ox, s
10.7‹0.3 14.6‹0.3 1/(1.37‹0.04)
26.4‹2.7 21.6‹1.7 1.22‹0.16
6.8‹0.6 1/(1.68‹0.22)
(323)b
6
1
1
1
0 Kox, M
�
5
� 1 � 1
ox/Kox, M
0
k
s
4.0‹0.4
(323)b
6
20
0 Kinh, M
±
and about 1/3 in deuterium oxide. The third factor is
then the fraction of oxidative diversion of the `second
carbanion' SE*MD.
a
De®ned by eq (1) of the text.
Fixed at the mean of values for H O and D O.
2 2
b
The simplest variations on this model, also consistent
with the data of Figure 4, involve the addition of D to
the enzyme at other points in the catalytic cycle. The
mechanistic data, now to be presented, appear to us to
favor the version of Figure 5 over these alternatives.
Discussion
Kinetic model for the oxidative-diversion reaction
The results in Figure 4 are consistent with several mod-
els of the oxidative-diversion reaction, but we would
like to give special consideration to the model shown in
Figure 5. This model omits the weak inhibition seen at
D 0.1 mM, which we do not regard as mechanistically
signi®cant. This allows the kinetic law of eq (1) to be
simpli®ed to that of eq (2), with the parameters de®ned
as shown in eqs (3) and (4) in terms of the microscopic
rate constants of Figure 3.
Values of the observed rate constants
Table 2 compares the numerical values of the kinetic
parameters for oxidative diversion (the regulatory acti-
vation of the enzyme having been e€ected by pyruvate)
with related parameters for the normal catalytic reac-
2
0
tions of pyruvate (regulatory activation by pyruvate)
and of ¯uoropyruvate (regulatory activation by ¯uor-
opyruvate).21 The ¯uoropyruvate reaction presents a
useful comparison because in the action of SCPDC on
È
É
n=e ˆ k ‰DŠ= K ‡ ‰DŠ
ꢀ2†
o
ox
ox
kox ˆ jko3kor=ꢀjko3 ‡ kor†
ꢀ3a†
¯uoropyruvate a ¯uoride ion is eliminated at the `second-
carbanion' stage to generate acetyl-ThDP in the active
site.2
2,23
Acetyl-ThDP is also generated in the oxidative
where:
diversion reaction. Thus in both cases, product release
of acetate ion requires the hydrolysis of acetyl-ThDP.
kor ˆ k k =ꢀk ‡ k †
ꢀ3b†
ꢀ3c†
o5 o7
o5
o7
j ˆ ko5=ꢀkCD ‡ k
The most striking result may well be the constant k /
ox
o5†
5
� 1 � 1
s
, the value of k/B for the
È
ÉÈ
É
K
ox, with a value of 4Â10 M
. This value is slightly
kox=Kox ˆ ko1ko3=ꢀko2 ‡ ko3† kc5=ꢀkc3 ‡ kc5† j ꢀ4†
5
� 1 � 1
larger than 2Â10 M
s
20
normal catalytic reaction of pyruvate. As eq (5)
shows, the rate constant k /K contains the actual rate
ox ox
These rate-constant expressions appear complex but are
rational upon examination. The de®nition of kox in eq
(3a) describes the steady-state ®rst-order rate constant
for a sequential process in which decarboxylation in the
constant for binding of D to SE*S, {ko1 k /(k +k )},
o3
o2
o3
multiplied by two fractions that are less than or equal to
unity. Thus the second order rate constant for binding
5
� 1 � 1
s . Pre-
of the oxidant is still larger than 4Â10 M
presence of oxidant D (ko3 step) is followed by oxida-
tion and product release (rate constant k , de®ned in
vious work20 shows that the factor {k /(k +k )} is
about 1/2 so that the binding rate constant for the oxi-
c5 c3 c5
or
5
� 1 � 1
terms of the two rate constants ko5 and ko7 in eq (3b)).
The contribution of the rate constant ko3 is modi®ed by
the fraction j, which measures the relative rate of oxi-
dative diversion of SE*MD (see the caption to Fig. 4 for
an explanation of the symbols for enzyme forms)
against reaction of SE*MD with a proton and elimina-
tion to generate acetaldehyde.
dant is no smaller than 8Â10 M
s . Furthermore,
the rate constant for conversion in the normal catalytic
cycle20 for pyruvate of SE*+S to the transition state for
5
� 1 � 1
decarboxylation is of the order of 10Â10 M
s ,
nearly equal to the rate constant for reaction of SE*S
with D. Two reasonable conclusions then suggest them-
selves. They ¯ow from the fact that the rate constant
Table 2. Comparison of kinetic parameters for normal catalysis and for oxidative diversion by dichlorophenolindophenol in the action of SCPDC
ꢀ
at 30 C and pH 6.2 and equivalent
Kinetic constant
Pyruvate20
Fluoropyruvate11
Dichlorophenol±indophenol
�
5
1
k, s
�
320
2.1
1510
1.6
1/(1.8)
44
1.1
417
1.9
1/(1.3)
11
4.0
26
1/(1.4)
1/(1.7)
� 1 � 1
k/K, M s
1
0
6
1
DOD
0 K, M
k
DOD(k/K)

8
92
G. Hajipour et al. / Bioorg. Med. Chem. 7 (1999) 887±894
Figure 5. Kinetic model for the oxidative-diversion reaction under conditions of saturating pyruvate. Subscripts c refer to the normal catalytic cycle,
and subscripts o to the oxidative diversion. Note that rate constants are in some cases aggregate rate constants for several serial events. The normal
catalytic cycle connecting SE*S and SE*M is shown at the bottom (free or unactivated enzyme are not present because of the saturating con-
centration of S). The oxidative-diversion branch extends upward from SE*S, beginning with entry of D into the active site and continuing with
decarboxylation in the presence of D to generate SE*MD. The mechanism again branches, with protonation of the `second carbanion' and elim-
2 2
ination yielding acetaldehyde while oxidative diversion competitively generates a complex SE*ADH . This complex contains in the active site DH
2
and acetyl-ThDP; the latter hydrolyzes and acetate and DH are released.
kox/Kox cannot be greater than the rate constant for
conversion of SE*S+D to the transition state for dec-
arboxylation, the ®rst irreversible step in the reaction
sequence. First, unless decarboxylation is somehow
accelerated by the presence of the oxidant D in the
active site, the oxidant must be entering the active site
Isotope e€ects
The solvent isotope e€ects are given in Table 2 for the
reactions of pyruvate (normal catalytic reaction with
allosteric regulation by pyruvate),20 ¯uoropyruvate
(diversion to acetyl-ThDP and ¯uoride with allosteric
regulation by ¯uoropyruvate),21 and the paracatalytic
oxidative-diversion reaction (allosteric regulation by
pyruvate). The isotope e€ects on the k/K terms are all
inverse (faster in deuterium oxide) by factors of 1.3 to
1.8, while those for the k terms are normal (faster in
protium oxide) by factors of 1.6 and 1.9 for pyruvate
and ¯uoropyruvate, respectively, but inverse by a factor
of 1.4 for the oxidative-diversion reaction.
rapidly and reversibly and the rate constant k /K
ox ox
must be that for conversion of SE*S+D to the transi-
tion state for decarboxylation. Second, the value of j in
eq (4) must be very close to unity. If it were sensibly
smaller than unity, the value of {k k /(k +k )}
o1 o3
o2
o3
(
describing conversion of SE*S+D to the transition
state for decarboxylation) would become larger than the
rate constant for conversion of SE*+S to the transition
state for decarboxylation in the normal catalytic cycle.
This seems unreasonable, so we conclude that (1) kox/
Kox describes formation of the decarboxylation transi-
tion state, and (2) that the product of decarboxylation,
SE*M, is eciently trapped by the oxidant D (j=ca.1).
The isotope e€ects for the normal catalytic reaction of
2
0
pyruvate have been most thoroughly analyzed. The
current model we are employing for these isotope e€ects
is this:
The picture for kox is illuminated by these conclusions,
which suggest that decarboxylation is not retarded by
the presence of D, and that trapping by the oxidant is
ecient. Thus in eq (3a), we may take j=1 and
. Inverse e€ect of about 2 on k/B: This e€ect arises
from a rapid, reversible addition of the sulfhydryl
group of Cys-221 to pyruvate bound at the allos-
teric site (Fig. 2). This addition±elimination equi-
librium is coupled to a rapid opening±closing of
access to the active site. The rate-limiting process
is approximately the entry of pyruvate into the
active site during an open-access period.
. Normal e€ect of about 1.5 on k: Proton-inventory
analysis suggests that this e€ect has a complex
origin. The value of 1.5 is an average of 1 (no iso-
tope e€ect) on the decarboxylation reaction and 2
for the product-release reaction, which are equally
�
1
ko3=640 s , the value of the corresponding rate con-
stant in the normal catalytic cycle. Since the overall rate
�
1
� 1
constant kox is only 11 s , k ꢁ 640 s and k =k
or
ox
or
and the product-release manifold thus governs the value
of the observed rate constant. The observed rate constant
must then describe some combination of oxidation, acetyl-
ThDP hydrolysis, acetate release, and DH2 release,
following rapid decarboxylation. The isotope e€ects
suggest a somewhat higher de®nition for this model.

G. Hajipour et al. / Bioorg. Med. Chem. 7 (1999) 887±894
893
rate-limiting. The value of 2 for product release is
the product of an inverse isotope e€ect of 2 and a
normal isotope e€ect of 4. The inverse isotope
e€ect of 2 is exactly similar to that for substrate
entry into the active site: we take it to arise from a
sulfhydryl addition±elimination equilibrium at the
allosteric site coupled to active-site opening and
closing, with rate-limiting departure of acetalde-
hyde from the active site during an open period.
Concurrent with the export of acetaldehyde is the
import of a proton required in the stoichiometry,
and this process generates a normal isotope e€ect
of 4.
j is about unity. The inverse isotope e€ect on {ko1 ko3/
(k +k )} which should approximate that for oxidant
entry into the active site is therefore 1.7Â1.5=2.6.
o2
o3
Conclusion
The oxidative-diversion reaction of the SCPDC-pyr-
uvate system by the oxidant dichlorophenolindophenol
appears to occur by entry of the oxidant into the active
site with a second-order rate constant k /K and sol-
ox ox
vent isotope e€ect very similar to that for the binding of
pyruvate to the active site. The solvent isotope e€ect
arises from sulfhydryl addition at the allosteric site,
coupled to opening of access to the active site. This
process is required for admission of the oxidant to the
active site, as well as the normal substrate pyruvate and
the abnormal substrate ¯uoropyruvate.
A proton-inventory study has also been carried out for
the solvent isotope e€ect of 1.9 (Table 2) seen with
1
¯
uoropyruvate.2 In this case, there is no contribution of
an inverse isotope e€ect and if such an e€ect is required
for product release in the ¯uoropyruvate reaction, then
product release must be a rapid, non-rate-limiting pro-
cess. The proton inventory is consistent with two steps
that are about equally rate-limiting, one with no solvent
isotope e€ect and the other with a normal solvent iso-
tope e€ect of about 3.4. The former may be ¯uoride
elimination, the latter either ketonization of the elim-
ination product to generate acetyl-ThDP or hydrolysis
of acetyl-ThDP, each of the two contributing processes
The ®rst-order rate constant rate kox re¯ects a quite
slow release of the hydrophobic product of oxidative
diversion DH , also with an inverse solvent isotope
2
e€ect arising sulfhydryl addition at the allosteric site,
coupled to opening of the active site to permit product
departure.
The inverse isotope e€ects for both oxidant entry into
the active site and reduced-product exit from the active
site are therefore readily reconciled with the model of
allosteric regulation according to which sulfhydryl
addition by Cys-221 to the keto group of a pyruvate or
other ketone at the allosteric site drives a rapid open-
ing±closing of access to the active site. It is less easy to
see how the observations can be reconciled with a model
for regulation according to which events at the allosteric
site e€ect reorganization of the catalytic machinery at
the active site.
�
1
occurring with a rate constant of about 88 s . If actual
release of acetate occurs with a rate constant similar to
that for acetaldehyde release in the normal catalytic
�
1
cycle (640 s ), then it is about sevenfold faster than the
rate-limiting steps.
In the oxidative-diversion reaction, the inverse isotope
e€ect of around 2 is most easily explained if it is
assumed that product release, coupled to sulfhydryl
addition at the allosteric site, is now rate-limiting.
However, unless some unusual e€ect is at work, the rate
of acetate release cannot be rate-limiting here because
its rate constant in the ¯uoropyruvate reaction is much
Experimental
�
1
larger than the observed rate constant of 44 s , whereas
the observed rate constant in the oxidative diversion
Materials. Thiamin diphosphate hydrochloride (`cocar-
boxylase'), bu€ers, salts, and NADH were purchased
from Sigma, standard solutions, anhydrous citric acid
from Fisher, sodium citrate dihydrate and magnesium
sulfate (anhydrous) from J. T. Baker. Deuterium oxide
was obtained from Aldrich (99.9% D). Sodium pyr-
uvate was obtained from Sigma in 99% pure crystalline
form.
�
1
reaction is only 11 s . This leaves only the release of
DH . It does seem likely that this hydrophobic dye
2
molecule may have such a large anity for the active
site that it departs less frequently by about 60-fold than
the polar acetate molecule. We therefore suggest that
the rate-limiting event for kox is release of DH , coupled
2
to sulfhydryl addition at the allosteric site, which pro-
duces the inverse isotope e€ect.
Enzyme. Pyruvate decarboxylase (EC 4.1.1.1) from
Saccharomyces cerevisiae (speci®c activity 12±14 units/
mg) was purchased from Sigma, suspended in a solution
of 5% glycerol, 3.2 M ammonium sulfate, 5 mM potas-
sium phosphate, 1 mM magnesium acetate, 0.5 mM
EDTA, and 25 FM ThDP, pH 6.5.
The isotope e€ects on k/K are all inverse, in agreement
with the view that this constant is determined by events
during, or subsequent to, the entry of substrate or oxi-
dizing agent into the active site during an open period in
the rapid open±closed cycling driven by reversible addi-
tion of Cys-221 to pyruvate at the allosteric site. It
should be noted that the true inverse isotope e€ect for
oxidant entry in oxidative diversion may be larger than
2,6-Dichlorophenolindophenol (D). Obtained from
Sigma, dissolved in 1 M HCl, and extracted into ether.
The sodium salt was re-extracted into 2% sodium
bicarbonate solution and precipitated with sodium
chloride. The precipitate was washed with sodium chlo-
the measured value of 1.7 for k /K . The rate constant
ox ox
k /K is given by {ko1 k /(k +k )}{k /(k +k )}j
ox ox
o3
o2
o3
c5
c3
c5
where, as has been noted, the second factor is about 1/2
in protium oxide and about 1/3 in deuterium oxide and
ꢀ
ride solution, dried at 100 C, and stored in a dessicator.

8
94
G. Hajipour et al. / Bioorg. Med. Chem. 7 (1999) 887±894
ꢀ
Solutions in deionized water were stable at 4 C for two
weeks. Concentrations of D were determined by spec-
trophotometric titration at 520 nm (lmax 600 nm) with
standard ascorbic acid (prepared by iodate titration).
The extinction coecient of D in citrate bu€er at pH 6.2
2. Kluger, R. In The Enzymes, 3rd ed.; Sigman, D. S. Ed.;
Academic: New York, 1992; 20, 271.
3. Kluger, R. Chem. Rev. 1987, 87, 863.
4. Arjunan, P.; Umland, T.; Dyda, F.; Swaminathan, S.;
Furey, W.; Sax, M.; Farrenkopf, B.; Gao, Y.; Zhang, D.; Jor-
dan, F. J. Molec. Biol. 1996, 256, 590.
3
� 1
� 1
was measured as (8.28‹0.02)Â10 M cm . DH has
2
5
6
1
7
. Boiteux, A.; Hess, B. FEBS Lett. 1970, 9, 293.
no measurable absorbance at this wavelength.
. Hubner, G.; Fischer, A.; Schellenberger, A. Z. Chem. 1970,
È
1, 436.
. Ullrich, J.; Bonner, I. Hoppe-Seyler's Z. Physiol. Chem.
Kinetics. The absorbance at 520 nm was followed after
injection into a solution containing 30 mM pyruvate,
.6 U/mL PDC, 0.2 mM ThDP, 10 mM magnesium ion
1970, 351, 1026±1029.
8. Baburina, I.; Gao, Y.; Jordan, F.; Hohmann, S.; Furey, W.
Biochemistry 1994, 33, 5630.
0
and 0.1 M citrate bu€er of an aliquot of the stock solu-
tion of D such that its concentration was between 10
and 170 mM. Absorbance-time curves were initially lin-
ear and the slope was taken to obtain the rate. Rates
obtained in this manner were strictly proportional to the
enzyme concentration. Speci®c rates versus [D] were
then ®tted to eq 1 by non-linear least-squares proce-
dures. Reactions in deuterium oxide were carried out at
the same bu€er ratio as in protium oxide, assuring the
corresponding location in the two solvents on the
9
. Neale, A. D.; Scopes, R. K.; Wettenhall, R. E. H.; Hoo-
genraad, N. J. J. Bacteriol. 1987, 169, 1024.
0. Hoppner, T. C.; Doelle, H. W. Eur. J. Appl. Microbiol.
Biotechnol. 1983, 17, 152.
1. Sun, S.; Duggleby, R. G.; Schowen, R. L. J. Am. Chem.
Soc. 1995, 117, 7317.
2. Dobritzsch, D.; Ko
Biol. 1998, 273, 20196.
13. Hubner, G.; Konig, S.; Schellenberger, A. Biomed. Bio-
1
1
1
È
nig, S.; Schneider, G.; Lu, G. J. Mol.
È
È
chim. Acta 1988, 47, 9.
14. Baburina, I.; Li, H.; Furey, W.; Bennion, B.; Jordan, F.
Biochemistry 1998, 37, 1235.
1
7,18
pH(D)-rate pro®le (`corresponding pL').
1
5. Hong, J.; Sun, S.; Stewart, B. J.; Schowen, K. B.; Scho-
Background reactions. Each measured rate was cor-
rected for a slow background reaction of D with ThDP.
No change in absorbance could be detected when D was
incubated with enzyme, pyruvate, or bu€er.
wen, R. L. In The Biochemistry and Physiology of Thiamin
Diphosphate Enzymes; Bisswanger, H.; Schellenberger, A.,
Eds.; Intemann: Prien, Germany, 1996; p 70.
1
Schowen, R. L. Biochim. Biophys. Acta 1998, 1385, 187.
6. Hong, J.; Sun, S.; Derrick, T.; Larive, C.; Schowen, K. B.;
1
8
7. Schowen, K. B.; Schowen, R. L. Methods Enzymol. 1982,
7c, 551±606.
Acknowledgements
We are grateful to Professor Chi-Huey Wong for the
opportunity to contribute to this memorial issue. This
work was supported by a grant from the National
Institute of General Medical Sciences. RLS thanks the
Alexander von Humboldt Foundation.
18. Quinn, D. M.; Sutton, L. D. In Enzyme Mechanism from
Isotope E€ects; Cook, P. F., Ed.; CRC Press: Boca Raton, FL,
1
1
2
991.
9. Christen, P. Meth. Enzymol. 1977, 46, 48.
0. Alvarez, F. J.; Ermer, J.; Hubner, G.; Schellenberger, A.;
È
Schowen, R. L. J. Am. Chem. Soc. 1991, 113, 8402.
1. Sun, S.; Smith, G. S.; O'Leary, M. H.; Schowen, R. L. J.
Am. Chem. Soc. 1997, 119, 1507.
2. Leung, L. S.; Frey, P. A. Biochem. Biophys. Res. Commun.
2
References
2
1
. Schowen, R. L. In Comprehensive Biological Catalysis;
Sinnott, M. L., Ed.; Academic: London, 1998; Vol. III, p
17.
1978, 81, 274.
23. Gruys, K. J.; Halkides, C. J.; Frey, P. A. Biochemistry
1987, 26, 7575.
2