Protein engineering of pyruvate carboxylase
Investigation on the function of acetyl-CoA and the quaternary structure
Shinji Sueda, Md. Nurul Islam and Hiroki Kondo
Department of Biochemical Engineering and Science, Kyushu Institute of Technology, Japan
Pyruvate carboxylase (PC) from Bacillus thermodenitrificans
was engineered in such a way that the polypeptide chain was
divided into two, between the biotin carboxylase (BC) and
carboxyl transferase (CT) domains. The two proteins thus
formed, PC-(BC) and PC-(CT+BCCP), retained their
catalytic activity as assayed by biotin-dependent ATPase
and oxamate-dependent oxalacetate decarboxylation, for
the former and the latter, respectively. Neither activity was
dependent on acetyl-CoA, in sharp contrast to the complete
reaction of intact PC. When assessed by gel filtration
chromatography, PC-(BC) was found to exist either in
dimers or monomers, depending on the protein concentra-
tion, while PC-(CT + BCCP) occurred in dimers for the
most part. The two proteins do not associate spontaneously
or in the presence of acetyl-CoA. Based on these observa-
tions, this paper discusses how the tetrameric structure of
PC is built up and how acetyl-CoA modulates the protein
structure.
Keywords: acetyl-CoA; biotin; biotin-dependent carboxy-
lase; protein engineering; pyruvate carboxylase.
Pyruvate carboxylase (PC) is a biotin-dependent enzyme
and is involved in gluconeogenesis by converting pyruvate
to oxalacetate [1–3]. There are two forms of PC, single
polypeptide chain type and subunit type, but a large
majority belongs to the former class [1,4–7]. This form of
PC is made of about 1200 amino acids and is distributed
widely in both eukaryotes and some prokaryotes. The

Because of the lack of a three-dimensional structure, the
detailed mechanism of carboxylation and regulation of
PC remains obscure. Obviously, elucidation of the three-
dimensional structure of PC will unveil much of this
uncertainty and in fact such an undertaking is under way
in this laboratory. Additionally, a protein engineering
approach would be useful to examine the two partial
reactions individually. In this study, PC from Bacillus
thermodenitrificans (previously Bacillus stearothermophilus)
was engineered in such a way as to divide the protein
into two at the boundary of the BC and CT domains
(Fig. 1). The properties of the resulting two proteins,
PC-(BC) and PC-(CT + BCCP), were examined and
compared with those of the intact PC in order to gain
insight into the domain organization, the function of
acetyl-CoA and the reaction mechanism of PC.
Experimental procedures
Materials
Inorganic salts and common organic chemicals were
obtained from commercial sources. Acetyl-coenzyme A
was from Wako Pure Chemical (Osaka, Japan) and avidin
was from ProZyme (San Leandro, CA, USA). Reagents
for genetic engineering, such as restriction enzymes, were
purchased from Takara (Kyoto, Japan). Oligonucleotides
were custom synthesized by Hokkaido Science (Sapporo,
Japan). The TOPO TA cloning kit was the product of
Invitrogen.
Correspondence to S. Sueda, Department of Biochemical Engineering
and Science, Kyushu Institute of Technology, Kawazu 680-4, Iizuka
820-8502, Japan. Fax: + 81 948 29 7801, Tel.: + 81 948 29 7834,

gene manipulation. A 19 bp segment of the PC gene (shown
above in italics) containing an EcoRI site was incorporated
into the reverse primer, Trc 2. The PCR reaction was
conducted under the following conditions: The reaction
mixture contained 5 units of Ex Taq
TM
(Takara), 1· Ex Taq
buffer, 200 l
M
each of the four dNTPs, 1 l
M
each of the
primers and 10 ng of pTrc99A in a final volume of 100 lL.
After denaturation at 94 °C for 5 min, the samples were
subjected to 30 cycles of denaturation (94 °C, 1 min),
annealing (58 °C, 1 min) and extension (72 °C, 1 min),
and subsequently subjected to additional extension (72 °C,
10 min). The PCR products were TA cloned and sequenced.
The plasmid thus prepared was digested with ApaIand
EcoRI, and the resulting fragment was ligated into the ApaI/
EcoRI sites of pPC. The second amino acid of native PC
is converted to glutamic acid from lysine because of the
introduction of the NcoI site into the start codon region of
this recombinant. This plasmid allowed E. coli to express
PC at a much higher level and the enzyme produced was
as active as native PC. Hence, this PC is called intact PC
despite the mutation of the second amino acid residue.
Construction of over-expression plasmids for PC-(BC)
and PC-(CT + BCCP)
The boundary of the BC and CT domains of B. thermo-

CCATGGCACGCCGGAAAGACGGAACGA
AAATG-3¢ and CT2, 5¢-CCGATCCCAC
GGATCCTCT
TTTAAAAAGCG-3¢ (restriction enzyme sites are under-
lined). The forward primer, CT1, harbored an NcoIsite
introduced for placing the start codon and cloning, and the
reverse primer, CT2, harbored a BamHI site present on the
PC gene. As a result of the engineering, the second amino
acid residue is converted from proline to alanine. PCR
conditions were the same as those described above, and
the PCR product was TA cloned and sequenced. Likewise,
a fragment representing the downstream region from the
BamHI site to the end of the open reading frame was
prepared (S. Sueda, unpublished observation). These two
fragments were cloned into pTrc99A through multiple steps
to yield a recombinant plasmid, pPC-(CT + BCCP), which
allowed E. coli to express the desired CT plus BCCP
domain of PC to a high level.
Purification of proteins
E. coli JM109 transformed with either one of the over-
expression plasmids prepared above was grown in Luria-
Bertani medium containing 50 lgÆmL
)1
ampicillin and
1 lgÆmL
)1
D
-biotin, where a biotin-binding domain was
present. Cells were harvested by centrifugation, suspended
in 0.12

a salt gradient from buffer A to buffer B (buffer A + 0.5
M
NaCl). The desired fractions, inspected by SDS/PAGE,
were collected and dialyzed against buffer A. The samples
were applied to gel filtration chromatography on Super-
Fig. 1. Schematic representation of the domain structures of intact PC
and engineered proteins, PC-(BC) and PC-(CT + BCCP).
1392 S. Sueda et al.(Eur. J. Biochem. 271) Ó FEBS 2004
dex
TM
200 (Amersham), eluted with 50 m
M
KP
i
buffer,
pH 7.0, containing 0.1
M
NaCl, 0.1 m
M
EDTA and 0.1 m
M
DTT, and the desired fractions collected. Intact PC and PC-
(CT+BCCP) were further purified by monomeric avidin-
Sepharose affinity chromatography [12–14] as reported
previously [15]. The samples were applied at a flow rate of
0.5 mLÆmin
)1
onto the monomeric avidin column equili-
brated with running buffer (50 m
M

TM
HR 5/5 (Amersham).
Protein was eluted by a salt gradient from buffer C (20 m
M
Tris/HCl, pH 7.5) to buffer D (buffer C+0.35
M
NaCl).
The desired fractions were collected and dialyzed against
5m
M
KP
i
buffer, pH 7.0, with 0.1 m
M
EDTA and 0.1 m
M
DTT, and stored at 4 °C. The specific activity of PC,
determined below, was 9.5 UÆmg
)1
, where 1 U is defined as
the amount of enzyme to produce 1 lmol of oxalacetate per
min, and the protein concentration was determined from the
amino acid composition.
Pyruvate carboxylase assays
Pyruvate carboxylase activity was measured by monitoring
the oxalacetate formation using the coupled reaction with
malate dehydrogenase according to the methods described
previously [16–18]. Oxidation of NADH in the malate
dehydrogenase reaction was followed spectrophotometri-
cally at 340 nm. All assays were carried out at 30 °C, and

M
at
fixed concentrations of bicarbonate (100 m
M
) and pyruvate
(5 m
M
), where the enzyme was 77% and 92% saturated with
them, respectively. In addition, free Mg
2+
concentration
was kept constant, with MgCl
2
,at3m
M
in excess of ATP;
free Mg
2+
concentration was approximated to be its ana-
lytical concentration minus that of ATP, as the true concen-
tration of free Mg
2+
calculated based on the dissociation
constant for MgATP of 0.0143 m
M
[19] was only 1%
different from the approximate value. At high ATP concen-
trations, substrate inhibition was evident, and thus two kinds
of analysis were applied for the data on ATP. First, the
simple Michaelis–Menten equation was fitted to the kinetic

The K
m
value for bicarbonate was determined by varying
its concentration from 0.5–100 m
M
at fixed concentrations
of ATP (2 m
M
) and pyruvate (5 m
M
). As the endogenous
level of bicarbonate is known to be 0.5 m
M
at pH 8.0
[20], the concentration of bicarbonate was corrected for
this value. Likewise, the K
m
value for pyruvate was
determined by varying its concentration from 0–5 m
M
at
fixed concentrations of ATP (2 m
M
) and bicarbonate
(100 m
M
). The simple Michaelis–Menten equation was
used for the analysis of the data for bicarbonate and
pyruvate.
ATP cleavage assays

NADH, 5 units of lactate
dehydrogenase and 5 units of pyruvate kinase. In the case of
the PC-(BC) assay, 50 m
M
free
D
-biotin was added to the
above reaction mixture.
The kinetic parameters, K
m
and V
max
,fortheATPase
reaction of PC-(BC) were determined as follows: the
kinetic parameters for ATP were obtained by varying its
concentration from 0–5 m
M
at fixed concentrations of
biotin (100 m
M
) and bicarbonate (100 m
M
), where the
enzyme is 68% and 62% saturated with them, respect-
ively. Obviously, this situation is not ideal for the accurate
estimation of kinetic parameters, but concentrations
higher than this will deviate too much from those of
physiological conditions. Accordingly, experiments were
carried out under these subsaturating conditions with
respect to biotin and bicarbonate. In the kinetics for ATP,

0
):
Ó FEBS 2004 Protein engineering of pyruvate carboxylase (Eur. J. Biochem. 271) 1393
v ¼ V
max
Â
½S
K
m
þ½Sþv
0

Eqn ð2Þ
Oxalacetate decarboxylase assays
Oxalacetate decarboxylase activity of intact PC and PC-
(CT + BCCP) was measured with oxamate as the stimu-
lant, according to the procedures previously reported [22].
The reactions were monitored by measuring the formation
of pyruvate which was then reduced to lactate by lactate
dehydrogenase, and the concomitant oxidation of NADH
was monitored at 340 nm. All assays were performed at
30 °C, and the reaction mixture contained the following
components, unless otherwise stated: 100 m
M
Tris/HCl
(pH 8.0), 5 m
M
MgCl
2
,100m

P
i
-Tween, and then immersed in NaCl/Tris buffer contain-
ing 0.4 UÆmL
)1
alkaline phosphatase-conjugated streptavi-
din (Boehringer Mannheim) for 20 min. The membrane
was then washed with NaCl/Tris-Tween three times, before
being developed by 0.78 m
M
4-nitroblue tetrazolium chlor-
ide and 0.40 m
M
5-bromo-4-chloro-3-indolylphosphate in
20 m
M
Tris/HCl, pH 9.5, containing 100 m
M
NaCl and
50 m
M
MgCl
2
.
For determining the amino-terminal sequence of the
proteins, electrophoresed samples were electroblotted onto
a poly(vinylidene difluoride) membrane (Atto, Tokyo,
Japan) according to the conventional procedure. Pieces
of the membrane containing the desired bands, as visualized
by ponceau S, were used for sequencing by Edman

M
,and20 lLeachwas
applied to the column.
Molecular mass determination by mass spectrometry
The molecular masses of PC-(BC) and PC-(CT + BCCP)
were determined by MALDI TOF mass spectrometry with
a Voyager DE-STR mass spectrometer (PerSeptive Bio-
systems, Framingham, MA, USA)
7
. 4-Hydroxyazobenzene-
2¢-carboxylic acid (10 mgÆmL
)1
in 0.1% (v/v) trifluoroacetic
acid in 70 : 30 water/acetonitrile) was used as the MALDI
matrix. Samples were prepared by mixing the protein
solution with the matrix solution. One microliter of this
mixture was deposited on the sample plate, dried at ambient
temperature and analyzed.
Results
Construction and purification of the engineered
proteins of PC
The boundary of the BC and CT domains of PC was
estimated as follows: the BC subunit of E. coli acetyl-CoA
carboxylase is catalytically active and the three-dimensional
structure is known [24,25]. Its C-terminus appeared to
correspond to residue 460 of B. thermodenitrificans PC by
sequence alignment [26]. Likewise, the amino acid sequences
of PCs from various sources, including those of subunit-
type PCs, were aligned to reveal that the N-terminus of CT
seemed to reside at residue 470 of B. thermodenitrificans PC.

the original lysine, was confirmed. Likewise, the amino
acid sequence of PC-(CT + BCCP) was determined to be
ARRKDRGTKM, and this sequence was consistent with
that deduced from its DNA sequence except for the absence
of the first amino acid methionine. It was also confirmed that
the second residue was properly converted to alanine from
original proline, because of the design of the expression
plasmid.
In SDS/PAGE, the bands of intact PC and PC-(BC) were
observed at the positions corresponding to the molecular
masses deduced from their sequences, 128.5 and 51.4 kDa,
respectively, while that of PC-(CT + BCCP) was observed
at a position (65 kDa) considerably smaller than that
expected (77.1 kDa). To confirm the integrity of PC-
(CT + BCCP), this protein was analyzed by MALDI
TOF mass spectrometry together with PC-(BC). The mass
(m/z value) obtained was 77 047 ± 76 for PC-(CT +
BCCP) and 51 428 ± 53 for PC-(BC) (mean ± SD from
three determinations). These values are identical, within
experimental error, to the molecular masses deduced from
their sequences, 77 082 and 51 438 Da, respectively, prov-
ing that the two engineered proteins have the correct
structure.
Molecular properties of the engineered proteins of PC
Association states of the proteins were investigated by high
performance gel filtration chromatography. Apparent
molecular masses of the samples were estimated on the
basis of the calibration curve obtained by using a set of
standard proteins (Fig. 3). Typical elution profiles of intact
PC, PC-(BC) and PC-(CT + BCCP) are shown in Fig. 4.

a dimer. On the other hand, at a low concentration (5 l
M
), a
major peak was observed at 19.41 min (Fig. 4D) and the
molecular mass estimated from it was 66.2 ± 1.1 kDa,
which appeared to represent the monomer.
Moreover, the mixtures of PC-(BC) and PC-(CT +
BCCP) at various ratios were analyzed to study their
interaction, but no new peak was observed other than those
derived from the constituent proteins, suggesting that
Fig. 3. Estimation of the molecular masses of intact and engineered PC
by gel filtration chromatography on the TSK G3000SWXL column. The
molecular masses of the proteins used for construction of the calib-
ration curve (s) were ribonuclease A (13.7 kDa), chymotrypsino-
gen A (23 kDa), ovalbumin (43 kDa), albumin (67 kDa), aldolase
(158 kDa) and thyroglobulin (669 kDa). Elution times of intact PC
and engineered proteins are represented by d, as labelled.
Fig. 2. SDS/PAGE (A) and avidin-blot analysis (B) of purified intact
PC and engineered proteins. (A) SDS/PAGE was run with 12.5%
polyacrylamide and 0.5 lg of proteins: M, marker; lane 1, intact PC;
lane 2, PC-(BC); lane 3, PC-(CT + BCCP). (B) SDS/PAGE was run
with 0.1 lg of proteins and electroblotted onto the membrane. The
proteins carrying biotin were detected by the reaction with alkaline
phosphatase-conjugated avidin: lane 1, intact PC; lane 2, PC-
(CT + BCCP).
Ó FEBS 2004 Protein engineering of pyruvate carboxylase (Eur. J. Biochem. 271) 1395
PC-(BC) and PC-(CT + BCCP) do not interact signifi-
cantly under the experimental conditions employed. Acetyl-
CoA had almost no effect on the association of the two
proteins, as the elution profile was hardly affected by

insignificant was 0.46 ± 0.06 m
M
, while the value obtained
from the whole data based on Eqn (1) that takes into
account substrate inhibition was 0.87 ± 0.11 m
M
.TheK
m
for ATP of PC from the same source, obtained by simple
Michaelis–Menten analysis on the data where substrate
inhibition is not evident, was reported as 0.38 m
M
[28], close
to the corresponding value of the present work. Also, the
effect of acetyl-CoA on the pyruvate carboxylase reaction
was nearly the same among the present work and literature;
the activity was greatly increased upon addition of acetyl-
CoA and the activity in the absence of acetyl-CoA was
approximately 0.3% of the maximum (Table 1).
PC is known to catalyze the cleavage of ATP in the
absence of pyruvate [21]. It is hence possible to study the
reaction of BC (Scheme 1) independently of the CT reaction
(Scheme 2) by measuring this activity. This activity of PC
was determined under essentially the same conditions as the
complete reaction except for the omission of pyruvate
(Table 1). The rate of the ATP cleavage reaction is about
0.2% of that of the complete reaction of PC, which almost
coincides with that reported for chicken liver enzyme [21].
Fig. 4. Typical elution profiles for intact PC (A), PC-(CT + BCCP)
(B) and PC-(BC) (C and D) on the TSK G3000SWXL gel filtration

-biotin as intact PC, suggesting that its three-dimensional
structure remains intact even in the absence of other
domains. The enzymic activity of PC-(BC) in the presence of
various concentrations of biotin, bicarbonate and ATP are
depicted in Fig. 5. As expected from the reaction mechan-
ism proposed for BC [2,29,30], this enzymic reaction was
completely dependent on three substrates; it is worthy of
noting that biotin is necessary for this reaction to proceed
(the activity in the absence of biotin is about 2% of
maximum in its presence). This subject is discussed in more
detail below. Kinetic parameters for the three substrates
were determined from the data shown in Fig. 5. The K
m
for
bicarbonate was 62.2 ± 5.3 m
M
, which was comparable
to that for the complete reaction of intact PC
(29.9±1.4m
M
). In the kinetics for ATP, substrate inhi-
bition was observed just like in intact PC, and the K
m
values
determined based on the simple Michaelis–Menton equa-
tion and Eqn (1), were 0.54 ± 0.04 m
M
and
1.03 ± 0.15 m
M

tion, as expected (Fig. 6A). The activity in the presence of a
saturating concentration of oxamate was about 40 times
higher than that in its absence (Table 2). The effect of
oxalacetate concentration on the decarboxylation reaction
at a fixed concentration of oxamate is shown in Fig. 6B. In
this case, substrate inhibition at high concentration of
oxalacetate is evident from the profile. Such a phenomenon
Fig. 5. Kinetic analysis for the ATP cleavage activity of PC-(BC). Activity of PC-(BC) (0.22 mg in 1 mL) was assayed with free biotin as the
substrate in 100 m
M
Tris/HCl (pH 8.0) containing 5 m
M
MgCl
2
,100m
M
KCl, 0.1 m
M
acetyl-CoA, 0.5 m
M
phosphoenol pyruvate, 0.15 m
M
NADH, 5 units of lactate dehydrogenase, 5 units of pyruvate kinase and variable concentrations of ATP, bicarbonate and biotin at 30 °C. (A)
Biotin was the variable substrate with 2 m
M
ATP and 100 m
M
bicarbonate; the K
m
for biotin was 50.9 ± 5.4 m

and V
max
2.39 ± 0.10 UÆlmol
)1
, while those determined with the data from 0–5.0 m
M
on the basis of Eqn (1) were as follows: K
m
1.03 ± 0.15 m
M
and V
max
3.91 ± 0.40 UÆlmol
)1
. The theoretical curve shown in this figure was drawn on the basis of Eqn (1). In each case, the standard errors in V
max
and
K
m
were determined from the nonlinear regression analysis.
Table 1. Effect of 0.1 m
M
acetyl-CoA on the pyruvate carboxylation of
intact PC and on the ATP cleavage reactions of intact PC and PC-(BC).
One unit of enzyme activity was defined as the amount of enzyme
required to catalyze the formation of 1 lmol of each product per min.
Values are the means ± SD from three separate experiments.
Protein Activity
Enzymic activity (UÆlmol
)1

from the elution profiles of gel filtration HPLC, PC-
(CT + BCCP) was found to exist mainly as a dimer, while
PC-(BC) was found to exist as a monomer or a dimer
depending on its concentration. In other words, both
engineered proteins associate with themselves to form
homodimers, and the association of PC-(CT + BCCP)
seems to be stronger than that of PC-(BC). In addition, the
association between PC-(BC) and PC-(CT + BCCP) was
not observed under the experimental conditions examined,
demonstrating that they do not possess strong affinity for
each other. Given that the same applies to intact PC as well,
it is deduced that the tetrameric form of PC is built up in the
following way: first, a dimer of PC is formed through the
association of each (CT + BCCP) domain of two proto-
mers of PC, and subsequently, individual BC domains of
the resulting two dimers associate to form a tetramer. In
other words, the tetrameric structure of PC appears to be
constructed through the interaction of the same domains,
namely BC with BC and (CT + BCCP) with (CT +
BCCP). This hypothesis awaits verification by X-ray
crystallographic analysis, which is under way in this
laboratory.
As for the reaction of BC, the following mechanism
involving the formation of an enzyme–carboxylphosphate
complex seems to be the most plausible one [2,29,30]:
ATP þ HCO
À
3
þenz-biotin Ðð
À2

(Scheme 3) of PC-(BC) was investigated with free biotin
Fig. 6. Oxalacetate decarboxylation reaction of PC-(CT + BCCP).
Activity of PC-(CT + BCCP) (0.39 mg in 1 mL) was assayed in
100 m
M
Tris/HCl (pH 8.0) containing 5 m
M
MgCl
2
,100m
M
KCl,
0.1 m
M
acetyl-CoA, 0.15 m
M
NADH, 5 units of lactate dehydro-
genase and variable concentrations of oxalacetate and oxamate at
30 °C. (A) Oxamate was the varied substrate with 0.1 m
M
oxalacetate.
(B) Oxalacetate was the varied substrate with 1.0 m
M
oxamate. Error
bars represent the standard deviations from the mean of three deter-
minations.
Table 2. Effect of 0.1 m
M
acetyl-CoA on the oxalacetate decarboxy-
lase activity (UÆlmol

Based on these arguments, it is tempting to propose the
following hypothesis: in the absence of acetyl-CoA, the
active site of the BC domain cannot interact with biotin of
the BCCP domain due to the spatial separation between the
active site and biotin; however, upon binding of acetyl-CoA,
a conformational change is induced, so that biotin can reach
the active site to carry out the catalytic reaction. In the
reaction of PC-(BC) with free biotin, such a steric constraint
is absent; as a result, its acetyl-CoA dependence may be lost.
Conformation changes of PC induced by acetyl-CoA have
been observed by various means such as electron micros-
copy [35,36], ultracentrifugation [37] and others [38]. In
order to verify the above hypothesis, further investigation is
needed and studies using other engineered proteins as well
as X-ray crystallographic analysis are under way in this
laboratory.
Acknowledgements
The authors are grateful to Ms Tomoko Ishiguro and Ms Masayo
Nonaka for their assistance with construction of the recombinant
plasmids.
References
1. Jitrapakdee, S. & Wallace, J.C. (1999) Structure, function and
regulation of pyruvate carboxylase. Biochem. J. 340, 1–16.
2. Attwood, P.V. (1995) The structure and the mechanism of action
of pyruvate carboxylase. Int. J. Biochem. Cell. Biol. 27, 231–249.
3. Wallace, J.C., Jitrapakdee, S. & Chapman-Smith, S. (1998)
Pyruvate carboxylase. Int. J. Biochem. Cell. Biol. 30, 1–5.
4. Barden, R.E., Taylor, B.L., Isohashi, F., Frey, W.H., Zander, G.,
Lee, J.C. & Utter, M.F. (1975) Structural properties of pyruvate
carboxylases from chicken liver and other sources. Proc. Natl

13. Kohanski, R.A. & Lane, M.D. (1990) Monovalent avidin affinity
column. Methods Enzymol. 184, 194–200.
14. Buckley, J.J., Libor, S. & Sundaram, T.K. (1979) Biotin subunits
of acetyl CoA carboxylase and pyruvate carboxylase from a
thermophilic Bacillus. Arch. Biochem. Biophys. 192, 396–404.
15. Jitrapakdee, S., Walker, M.E. & Wallace, J.C. (1999) Functional
expression, purification, and characterization of recombinant
human pyruvate carboxylase. Biochem. Biophys. Res. Commun.
266, 512–517.
16. Modak, H.V. & Kelly, D.J. (1995) Acetyl-CoA-dependent pyru-
vate carboxylase from the photosynthetic bacterium Rhodobacter
capsulatus: rapid and efficient purification using dye-ligand affinity
chromatography. Microbiol. 141, 2619–2628.
17. Mukhopadhyay, B., Stoddard, S.F. & Wolfe, R.S. (1998)
Purification, regulation, and molecular and biochemical charac-
terization of pyruvate carboxylase from Methanobacterium ther-
moautotrophicum strain DH. J. Biol. Chem. 273, 5155–5166.
18. Mukhopadhyay, B. & Purwantini, E. (2000) Pyruvate carboxylase
from Mycobacterium smegmatis: stabilization, rapid purification,
molecular and biochemical characterization and regulation of the
cellular level. Biochim. Biophys. Acta 1475, 191–206.
19. Morrison, J.F. (1979) Approaches to kinetic studies on metal-
activated enzymes. Methods Enzymol. 63, 257–294.
20. Levert, K.L., Lloyd, R.B. & Waldrop, G.L. (2000) Do cystein 230
and lysine 238 of biotin carboxylase play a role in the activation of
biotin? Biochemistry 39, 4122–4128.
21. Attwood, P.V. & Graneri, B.D.L.A. (1992) Bicarbonate-
dependent ATP cleavage catalysed by pyruvate carboxylase in the
absence of pyruvate. Biochem. J. 287, 1011–1017.
22. Attwood, P.V. & Cleland, W.W. (1986) Decarboxylation of

481–486.
31. Polakis,S.E.,Guchhait,R.B.,Zwergel,E.E.&Lane,M.D.(1974)
Acetyl coenzyme A carboxylase system of Escherichia coli. J. Biol.
Chem. 249, 6657–6667.
32. Attwood, P.V. (1993) Locus of action of acetyl CoA in the biotin-
carboxylation reaction of pyruvate carboxylase. Biochemistry 32,
12736–12742.
33. Phillips, N.F.B., Sonswell, M.A., Chapman-Smith, A., Keech,
D.B. & Wallace, J.C. (1992) Isolation of a carboxylphosphate
intermediate and the locus of acetyl-CoA action in the pyruvate
carboxylase reaction. Biochemistry 31, 9445–9450.
34. Attwood, P.V. & Wallace, J.C. (1986) The carboxybiotin complex
of chicken liver pyruvate carboxylase. A kinetic analysis of
the effects of acetyl-CoA, Mg
2+
ions and temperature on its
stability and on its reaction with 2-oxobutyrate. Biochem. J. 235,
359–364.
35. Rohde, M., Lim, F. & Wallace, J.C. (1986) Pyruvate carboxylase
from Saccharomyces cerevisiae. Quaternary structure, effects of
allosteric ligands and binding of avidin. Eur. J. Biochem. 156,
15–22.
36. Attwood, P.V., Mayer, F. & Wallace, J.C. (1986) Avidin as a
probe of the conformational changes induced in pyruvate car-
boxylase by acetyl-CoA and pyruvate. FEBS Lett. 203, 191–196.
37. Taylor, B.L., Frey, W.H., Barden, R.E., Scrutton, M.C. & Utter,
M.F. (1978) The use of the ultracentrifuge to determine the cata-
lytically competent forms of enzymes with more than one oligo-
meric structure. J. Biol. Chem. 253, 3062–3069.
38. Frey, W.H. & Utter, M.F. (1977) Binding of acetyl-CoA to