Crystal structures of Acetobacter aceti succinyl-coenzyme A (CoA):Acetate CoA-transferase reveal specificity determinants and illustrate the mechanism used by class i CoA-transferases

Article abstract of DOI:10.1021/bi300957f

Coenzyme A (CoA)-transferases catalyze transthioesterification reactions involving acyl-CoA substrates, using an active-site carboxylate to form covalent acyl anhydride and CoA thioester adducts. Mechanistic studies of class I CoA-transferases suggested that acyl-CoA binding energy is used to accelerate rate-limiting acyl transfers by compressing the substrate thioester tightly against the catalytic glutamate [White, H., and Jencks, W. P. (1976) J. Biol. Chem. 251, 1688-1699]. The class I CoA-transferase succinyl-CoA:acetate CoA-transferase is an acetic acid resistance factor (AarC) with a role in a variant citric acid cycle in Acetobacter aceti. In an effort to identify residues involved in substrate recognition, X-ray crystal structures of a C-terminally His₆-tagged form (AarCH6) were determined for several wild-type and mutant complexes, including freeze-trapped acetylglutamyl anhydride and glutamyl-CoA thioester adducts. The latter shows the acetate product bound to an auxiliary site that is required for efficient carboxylate substrate recognition. A mutant in which the catalytic glutamate was changed to an alanine crystallized in a closed complex containing dethiaacetyl-CoA, which adopts an unusual curled conformation. A model of the acetyl-CoA Michaelis complex demonstrates the compression anticipated four decades ago by Jencks and reveals that the nucleophilic glutamate is held at a near-ideal angle for attack as the thioester oxygen is forced into an oxyanion hole composed of Gly388 NH and CoA N2″. CoA is nearly immobile along its entire length during all stages of the enzyme reaction. Spatial and sequence conservation of key residues indicates that this mechanism is general among class I CoA-transferases.

Full text of DOI:10.1021/bi300957f

Article
pubs.acs.org/biochemistry
Crystal Structures of Acetobacter aceti Succinyl-Coenzyme A
(CoA):Acetate CoA-Transferase Reveal Specificity Determinants and
Illustrate the Mechanism Used by Class I CoA-Transferases
Elwood A. Mullins†,‡ and T. Joseph Kappock*,†
†Department of Biochemistry, Purdue University, West Lafayette, Indiana 47907-2063, United States
‡Department of Chemistry, Washington University, St. Louis, Missouri 63130-4899, United States
S
* Supporting Information
ABSTRACT: Coenzyme A (CoA)-transferases catalyze trans-
thioesterification reactions involving acyl-CoA substrates, using
an active-site carboxylate to form covalent acyl anhydride and
CoA thioester adducts. Mechanistic studies of class I CoA-
transferases suggested that acyl-CoA binding energy is used to
accelerate rate-limiting acyl transfers by compressing the
substrate thioester tightly against the catalytic glutamate
[White, H., and Jencks, W. P. (1976) J. Biol. Chem. 251,
1688−1699]. The class I CoA-transferase succinyl-CoA:acetate
CoA-transferase is an acetic acid resistance factor (AarC) with
a role in a variant citric acid cycle in Acetobacter aceti. In an effort to identify residues involved in substrate recognition, X-ray
crystal structures of a C-terminally His6-tagged form (AarCH6) were determined for several wild-type and mutant complexes,
including freeze-trapped acetylglutamyl anhydride and glutamyl-CoA thioester adducts. The latter shows the acetate product
bound to an auxiliary site that is required for efficient carboxylate substrate recognition. A mutant in which the catalytic glutamate
was changed to an alanine crystallized in a closed complex containing dethiaacetyl-CoA, which adopts an unusual curled
conformation. A model of the acetyl-CoA Michaelis complex demonstrates the compression anticipated four decades ago by
Jencks and reveals that the nucleophilic glutamate is held at a near-ideal angle for attack as the thioester oxygen is forced into an
oxyanion hole composed of Gly388 NH and CoA N2″. CoA is nearly immobile along its entire length during all stages of the
enzyme reaction. Spatial and sequence conservation of key residues indicates that this mechanism is general among class I CoA-
transferases.
binary enzyme·acyl-CoA complex is converted into a CoA
oenzyme A (CoA)-transferases (EC 2.8.3) perform acyl
Ctransfer reactions involved in the metabolic activation of thiolate complex that also contains an acylglutamyl anhydride
carboxylate substrates.1 Class I CoA-transferases, typified by
adduct. CoA thiolate then attacks the internal carbonyl to form
a complex containing the carboxylate product and a glutamyl-
CoA thioester. After carboxylate exchange, the second half-
reaction proceeds by the opposite reaction sequence to form
the acyl-CoA product complex. One or more of the binary
complex interconversions illustrated in Figure 1 is rate-limiting
in SCOT catalysis.2 The acyl-CoA substrate binding energy is
harnessed to accelerate catalysis in a process that immobilizes
the acyl moiety7,8 and requires an intact covalent link to
unreacting parts of the acyl-CoA substrate.9−11 Crystal
structures show constriction of the enzyme around CoA,12,13
but these interactions are not sufficient to account for rate
accelerations achieved by the enzyme·acyl-CoA substrate
complex.1 While glutamyl-CoA thioester adduct structures are
available (PDB entries 2ahw, 2oas, and 3oxo),14,15 structures of
a class I CoA-transferase bound to acyl-CoAs, carboxylates, and
mitochondrial succinyl-CoA:3-oxoacid CoA-transferase
(SCOT), form multiple covalent adducts involving an essential
glutamate residue (Figure 1).2−6 In the first half-reaction, the
Figure 1. General half-reaction for class I CoA-transferases. Two
tetrahedral oxyanion intermediates differ by whether CoA becomes
attached to the external carbonyl, provided by the acyl-CoA/
carboxylate substrate, or the internal carbonyl, provided by the
essential active-site glutamate. Following exchange of the carboxylate
product (dashed box), the second half-reaction proceeds in the reverse
order of the first half-reaction.
Received: July 17, 2012
Revised: September 27, 2012
Published: October 2, 2012
© 2012 American Chemical Society
8422
dx.doi.org/10.1021/bi300957f | Biochemistry 2012, 51, 8422−8434

Biochemistry
Article
anhydride adducts are still needed to explain how these
enzymes transmute acyl-CoA binding energy into rate
accelerations.16
site-directed mutagenesis kit (Stratagene), plasmid template
pJK385, and appropriate ODN pairs were used to create
plasmids for the expression of AarCH6 mutants S71A
(pJK564), R228E (pJK512), E294A (pJK513), N347A
(pJK524), E435A (pJK514), E435D (pJK517), and E435Q
(pJK516). Plasmid DNA sequenced in both directions by the
staff of the Purdue University Genomics Core Facility
contained the expected DNA sequences. Experimental details
for AarC and AarCH6 production, isolation, and character-
ization by circular dichroism (CD) are given in the Supporting
Information.
Steady-State Kinetic Analysis. Citrate synthase-depend-
ent VisF/VisR assays and high-performance liquid chromatog-
raphy (HPLC)-coupled LCF/LCR assays were performed as
described previously,28 except that the concentrations of
enzymes and substrates and the durations of the reactions
were adjusted as required for AarCH6 mutants. A unit is
defined as the amount of enzyme that forms 1 μmol of product
per minute.
Alternate carboxylate substrates (CoA acceptors) were
identified using modified LCR assays containing 50 mM
potassium phosphate (pH 8.0), 100 mM potassium chloride, 20
mM carboxylate (pH 8.0), 1 mM acetyl-CoA, and 5 μg of
AarCH6. Aliquots (0.1 mL) were withdrawn prior to the
addition of enzyme and 5 and 30 min after the addition of
enzyme, quenched with 6.25% trichloroacetic acid (0.4 mL),
vortexed briefly, and centrifuged at 16100g for 3 min. The
soluble portion was transferred to an autosampler vial and
analyzed by reverse-phase ion-pair HPLC. Compounds
identified as alternate substrates were examined in saturation
experiments employing modified LCR assays in which succinate
was replaced by the candidate CoA acceptor.
CoA and CoA thioesters were resolved on a Breeze HPLC
system (Waters) equipped with a Symmetry C18 column [4.6
mm × 75 mm, 3.5 μm (Waters)]. Acid-quenched reaction
mixtures (25−100 μL) were applied using a 717plus
autosampler (Waters). The column was developed at 30 °C
and 0.8 mL min−1 with detection at 260 nm. A mixture of 200
mM sodium phosphate and 150 mM sodium acetate (pH 4.6)
was used in a water/methanol solvent system. The gradient
program was as follows: 4.5% methanol (isocratic) from 0 to 5
min, 4.5 → 15% methanol (linear gradient) from 5 to 15 min,
15% methanol (isocratic) from 15 to 17 min, 15 → 4.5%
methanol (linear gradient) from 17 to 22 min, and 4.5%
methanol (isocratic) from 22 to 27 min. The initial and final
isocratic steps were performed in buffers without methanol for
the analysis of reaction mixtures containing formate, glycolate,
glyoxylate, oxalate, or tartrate, while the intermediate isocratic
step was lengthened by 5 min for the analysis of reaction
mixtures containing butyrate. Acyl-CoA peak areas were
compared to an acetyl-CoA standard as described previously.28
Kinetic constants were obtained by nonlinear least-squares
fitting to the Michaelis−Menten model, a substrate inhibition
model (eq 1), or a competitive inhibition model (eq 2) using
gnuplot 4.4. In these equations, Ki is an inhibition constant and
C is the inhibitor concentration.
Class I CoA-transferases have similar sequences, structures,
and ping-pong kinetic mechanisms.2 Class III CoA-transferases
differ in all of these but use a chemically analogous reaction
sequence involving anhydride and thioester adducts of an
essential aspartate residue.17 Crystal structures of adducts and
complexes of Oxalobacter formigenes formyl-CoA transferase
(FRC) have allowed the assembly of a structural model for the
entire class III CoA-transferase reaction.18,19 This reaction,
however, relies upon substantial movement of the CoA
pantetheine moiety. Conversely, evidence suggests that the
reactive portion of the acyl-CoA substrate is tightly held in a
single conformation throughout the class I reaction.7,16 This
difference may be indicative of two unique and class-specific
catalytic mechanisms.
The class I CoA-transferase succinyl-CoA:acetate CoA-
transferase (SCACT, AarC) is an acetic acid resistance factor
that is required for acetate resistance20,21 by vinegar factory
strain Acetobacter aceti 1023.22,23 Acetic acid bacteria (AAB) are
acidophilic acetic acid producers that use diverse methods to
cope with the constant influx of acetic acid from the culture
medium.24 Several AAB25−27 use AarC in a variant citric acid
cycle (CAC) that “overoxidizes” acetic acid to CO2, which then
diffuses into the acidic culture medium.28 A privileged role in
the A. aceti central metabolism and the link between CAC flux
and the pivotal acetic acid resistance phenotype24 would have
imposed strong evolutionary selection for efficient catalysis by
AarC.
Here we report the kinetic and structural characterization of
AarC. These data were used to assemble a structural model for
the AarC mechanism that is congruent with kinetic findings and
can be applied to all class I CoA-transferases. Protein
conformational changes assemble a chemically competent
active site around a bound acyl-CoA substrate and regulate
access to the catalytic pocket. In the closed conformation, the
acyl-CoA thioester is pressed tightly against the active-site
glutamate. Subsequent conversion to the glutamyl-CoA
thioester and carboxylate product occurs with little motion of
the enzyme or substrate. We propose that an auxiliary
carboxylate binding site, located just outside the AarC catalytic
pocket, contributes to the efficient recognition and conversion
of the physiological carboxylate substrates.
EXPERIMENTAL PROCEDURES
■
Materials and Methods. Oligodeoxynucleotides [ODNs
(Table S1 of the Supporting Information)] were obtained from
IDT and used without further purification. DNA manipulations
were performed using enzymes from New England Biolabs.
DNA purification kits were from Qiagen. Plasmid pREP4-
groESL was a gift from P. Cole.29 Other reagents were from
Sigma-Aldrich or Fisher Scientific. Dethiaacetyl-CoA was
synthesized as described previously.30 A. aceti citrate synthase
with a C-terminal hexahistidine affinity tag was isolated as
described previously.30 Protein concentrations were measured
using a Bio-Rad Bradford assay with crystalline bovine serum
albumin as the standard.31 Absorbance measurements were
recorded on a model 8453 UV−visible spectrophotometer
(Agilent Technologies).
VS
Protein Production and Characterization. AarC ex-
pression plasmid pJK357 and AarCH6 expression plasmid
pJK385 have been described previously.28 The QuikChange II
v =
S
Ki
Km + S 1 +
(
)
(1)
8423
dx.doi.org/10.1021/bi300957f | Biochemistry 2012, 51, 8422−8434

Biochemistry
Article
VS
ature factor refinement, and automated water picking. Later
rounds of refinement consisted of coordinate optimization,
isotropic temperature factor refinement, and automated water
picking. Noncrystallographic symmetry (NCS) restraints were
applied for automated building of the AarCH6·CoA·citrate
(PDB entry 4eu7) model and for all stages of the refinement of
the AarCH6·CoA (PDB entry 4eu4), AarCH6-R228E·CoA
(PDB entry 4eu9), and AarCH6-E294A·dethiaacetyl-CoA
(PDB entry 4euc) models. Manual building and water addition
were conducted using Coot37 after each round of automated
refinement. Ligand coordinates and restraints were obtained
from HIC-Up38 and modified using PHENIX.
v =
C
Ki
S + Km 1 +
(
)
(2)
The ability of AarCH6 to process multiple substrates was
quantitated by the substrate promiscuity index I, the relative
probability of substrate conversion in an equimolar mixture of a
set of alternative substrates, which ranges from 0 (perfect
specificity, conversion of only one substrate) to 1 (perfect
promiscuity, equal conversion of all substrates), and the
weighted substrate promiscuity index J.32 Weighting was
based on the chemical diversity of the substrate set, enumerated
using a list of 92 chemical descriptors that has been described
previously.32
All other AarC(H6) structures were determined by
molecular replacement using the refined coordinates of the
AarCH6·CoA·citrate complex (PDB entry 4eu7, chain A,
residues 2−514) as the search model.
Crystallization and Data Collection. AarC and AarCH6
crystals were grown at room temperature (∼22 °C) by the
hanging drop vapor diffusion method. Reservoirs contained 0.5
mL of screening solution (Wizard I and II kits, Emerald
BioSystems) supplemented with 25 mM 2-mercaptoethanol.
Drops consisted of 2 μL of protein solution [5.6 mg mL−1 AarC
in 45 mM potassium phosphate (pH 8.0), 90 mM potassium
chloride, and 2 mM CoA or 6.0 mg mL−1 AarCH6 in 45 mM
Tris-HCl (pH 8.0), 90 mM potassium chloride, and 2 mM
CoA] and 2 μL of reservoir solution.
Orthorhombic crystals appeared in 0.8−1.0 M sodium
citrate, 0.1 M imidazole (pH 8.2), and 25 mM 2-
mercaptoethanol. Crystals were soaked in 15% (w/v) sorbitol,
1.1 M sodium citrate, 0.1 M imidazole (pH 8.2), 25 mM 2-
mercaptoethanol, and 2 mM CoA for 24 h at room temperature
prior to being flash-cooled by rapid immersion in liquid N2.33
Water, acetyl-CoA (2 mM), succinyl-CoA (2 mM), or
dethiaacetyl-CoA (10 mM) was substituted for CoA in drops
and cryoprotectants in an effort to obtain unliganded or
alternate liganded structures. Crystals used for the trapping of
covalent adducts were soaked in cryoprotectant solutions
lacking ligand for 22 h, transferred to a cryoprotectant solution
containing acetyl-CoA (2 mM) for 2 h, and flash-cooled.
Hexagonal crystals appeared in 1.7−2.0 M ammonium
sulfate, 0.2 M sodium chloride, 0.1 M sodium cacodylate (pH
6.5), and 25 mM 2-mercaptoethanol. These crystals were
processed as described above, except that the cryoprotectant
consisted of 15% (w/v) sorbitol, 2.2 M ammonium sulfate, 0.25
M sodium chloride, 0.1 M sodium cacodylate (pH 6.5), 25 mM
2-mercaptoethanol, and 2 mM CoA.
All structural images were generated with PyMOL.39
Electrostatic potential maps were computed using the
PDB2PQR40 web server (pH 7 and 298 K) and the APBS41
plugin for PyMOL.
Ligand Conformation Calculations. Michaelis complexes
and the glutamylsuccinyl anhydride adduct were manually
constructed in active sites containing the crystallographically
observed acetylglutamyl anhydride or glutamyl-CoA thioester
intermediates [both of which are present in the AarCH6·-
CoA·acetate complex (PDB entry 4eu6)]. A single water
(HOH517) and all residues (Phe35, Thr36, Ala70, Ser71,
Phe92, Leu114, Gln267, Gly269, Val270, Glu294, Val295,
Asn361, Gly387, Gly388, and Ser389) that directly contacted
modeled atoms were included during building and subsequent
energy minimization. Hydrogen atoms were added in
appropriate riding positions before equilibrium conformer
calculations incorporating a Monte Carlo search method
(5000 K) and MMFF94 force field were performed. Only the
positions of hydrogen atoms and modeled non-hydrogen atoms
were refined during energy minimization. All non-hydrogen
protein atoms and selected ligand atoms were “frozen”.
Corrections for implicit solvent were not applied at any point
during refinement. All molecular mechanics calculations were
performed with Spartan '02 (Wavefunction).
RESULTS
■
Protein Production, Isolation, and Biophysical Char-
acterization. AarC was purified to near-homogeneity using a
procedure involving ammonium sulfate fractionation, gel-
filtration chromatography, ion-exchange chromatography, and
dye-affinity chromatography (Table S2 of the Supporting
Information). Kinetic analysis of AarC (Figure S1 of the
Supporting Information) yielded steady-state kinetic constants
for acetate saturation [kcat = 290 30 s−1, Km = 110 20 mM,
kcat/Km = (2.6 0.5) × 103 M−1 s−1, and Ki = 1600 500 mM]
that are similar to those reported for AarCH6,28 indicating that
the C-terminal fusion in AarCH6 does not interfere with
catalysis. (As previously discussed,28 high acetate Km values are
physiologically reasonable in A. aceti.)
X-ray diffraction data were collected at the Advanced Photon
Source (Argonne National Laboratory, Argonne, IL). Dif-
fraction data were indexed, integrated, and scaled in
HKL2000.34
Determination and Refinement of Structure. The
structure of the AarCH6·CoA·citrate complex (PDB entry
4eu7) was determined by molecular replacement using a
threaded AarC model corresponding to a single subunit
(residues 3−505 in GenBank entry ACD85596). The model
was computed using SWISS-MODEL35 based on coordinates
from a biochemically uncharacterized CoA-transferase (PDB
entry 2g39) from Pseudomonas aeruginosa that is 48% identical
to AarC. All refinement steps were performed using
PHENIX.36 Automated building (PHENIX AutoBuild) was
used to improve the initial model for AarCH6·CoA·citrate
(PDB entry 4eu7) and AarCH6-S71A·CoA (PDB entry 4eu8)
complexes. The subsequent early refinement procedure
consisted of successive rounds of rigid-body refinement,
simulated annealing, coordinate optimization, isotropic temper-
Heterologously overproduced AarC(H6) was ∼90% in-
soluble,28 even when the C41(DE3) expression host42 or
lower expression temperatures43 were employed. A ∼10-fold
increase in the yield of soluble AarCH6 was observed when
AarCH6 was co-overproduced with Escherichia coli GroES and
GroEL, which are presumed to function as folding aides.29
Wild-type (wt) and mutant AarCH6 forms were routinely co-
overproduced with GroESL and purified to near-homogeneity
8424
dx.doi.org/10.1021/bi300957f | Biochemistry 2012, 51, 8422−8434

Biochemistry
Article
using the same procedure. GroESL co-overproduction failed to
ameliorate the complete insolubility of AarCH6-E435A and
AarCH6-E435Q.
Crystal Structures of AarC and AarCH6. Two AarCH6
crystal forms were identified by sparse-matrix screening:
bipyramidal hexagonal crystals that appeared in solutions
containing ammonium sulfate and clusters of thin orthorhom-
bic crystals that appeared in solutions containing sodium citrate
(Figure S6 of the Supporting Information). Faster crystal-
lization of both forms was noted for AarCH6 co-overproduced
with GroESL and subsequently gel-filtered. Orthorhombic
crystals gave X-ray diffraction patterns extending beyond 2.0 Å.
In contrast, the larger hexagonal crystals failed to yield X-ray
diffraction data extending beyond 2.8 Å. Hexagonal crystals
soaked in a cryoprotectant solution for <24 h produced
diffraction patterns with dark ice rings.
ESI-TOF-MS analysis was consistent with 504-residue
(AarC) or 513-residue (wt and mutant AarCH6) proteins
lacking Met1 (Table S3 of the Supporting Information).
The diminution of CD signal intensity was used to monitor
the thermal unfolding of AarC and AarCH6 as a function of pH
(Figure S4 of the Supporting Information). The stability of
each protein decreased sharply as the pH approached 3.8. AarC
remained folded under physiological conditions encountered in
the acidic cytoplasm (pH 3.8−6.8)44 of the relatively
thermophilic (≤35 °C) strain A. aceti 102322,23 but seems to
be less stable at pH ≤4 than other characterized A. aceti
proteins.25,45,46 AarCH6 was somewhat less stable than AarC,
but only under low-pH conditions that would likely result in
the protonation of surface-exposed histidine residues in the
affinity tag.
Substrate Spectrum. Interception of the glutamyl-CoA
thioester adduct by noncognate carboxylates accounts for the
substrate promiscuity attributed to many CoA-transferases. All
of the carboxylates that supported detectable AarCH6 activity
were ≥10-fold poorer substrates than the physiological
substrates acetate and succinate (Table 1). Lower kcat values
A structure of AarCH6 bound to CoA and citrate (PDB
entry 4eu7) was obtained from an orthorhombic crystal by
molecular replacement and refined to a model with good
statistics (Table 2). The sole Ramachandran outlier in most
a
Table 2. Summary of Crystallographic Results
PDB
entry
resolution
(Å)
b
c
protein
ligand(s)
Rcryst
Rfree
d
4eu3
4eu4
4eu5
4eu6
4eu7
4eu8
4eu9
4eua
4eub
4euc
4eud
wt
none
1.58
2.80
1.74
1.99
1.70
1.81
1.48
2.40
1.97
2.64
1.95
0.156
0.178
0.167
0.164
0.156
0.180
0.157
0.152
0.154
0.168
0.159
0.189
0.223
0.209
0.209
0.194
0.226
0.179
0.209
0.209
0.225
0.206
wt
CoA
wt
CoA
e
wt
CoA, acetate
CoA, citrate
CoA
a
Table 1. Alternate Substrates for AarCH6
wt
S71A
R228E
R228E
E294A
E294A
no tag
carboxylate
succinate
relative kcat/Km (%)
CoA
100
5.1
b
CoA
acetate
b
CoA
acetoacetate
propionate
D-malate
0.35
0.34
0.28
0.20
dethiaacetyl-CoA
CoA, citrate
a
All proteins crystallized in space group P212121, except for PDB entry
fumarate
c
4eu4, which crystallized in space group P6222. Tables S8−S11 of the
L-malate
0.10
Supporting Information contain a complete set of X-ray data collection
formate
0.033
0.011
0.010
0.0012
b
and refinement values. Rcryst = ∑|Fo − ⟨Fc⟩|/∑Fo, where the
oxaloacetate
c
summation is over the data used for refinement. Rfree is defined in the
DL-methylsuccinate
glutarate
same way as Rcryst but was calculated using the 5% of the data that were
d
excluded from refinement. The active site of subunit B contains
a
e
Full details are given in Table S4 of the Supporting Information. No
citrate. Acetylglutamyl anhydride (subunit A) and glutamyl-CoA
activity (≤0.01 unit/mg) was detected with glycolate, glyoxylate,
oxalate, trifluoroacetate, DL-lactate, L-lactate, malonate, pyruvate,
maleate, butyrate, D-tartrate, L-tartrate, α-ketoglutarate, citrate, or DL-
isocitrate. Acetyl-CoA was the CoA donor except where noted.
Succinyl-CoA was the CoA donor. Combined activity for two
products that are presumed to be regioisomers.
thioester (subunit B) adducts are also present.
chains, Pro91, was modeled into excellent electron density. The
asymmetric unit contained the expected biological dimer of
subunits, each with a topology (Figure S7 of the Supporting
Information) similar to those of other class I CoA-trans-
ferases.51
b
c
The AarC(H6) monomer is composed of distinct N- and C-
terminal domains, each displaying an open α/β-fold (Figure 2).
The domains are joined by two rigid loops (residues 171−178
and 182−215) that depart from the N-terminal domain to add
strand β8 to the core C-terminal β-sheet and a flexible domain
linker (residues 224−235) that spans a deep, funnel-shaped
cleft. The active-site glutamate (Glu294) is located at the
bottom of the cleft. The positively charged active-site region is
encircled by a negatively charged zone. Other regions on the
AarC surface appear to be largely neutral (Figure S9A−C of the
Supporting Information).
The extensive, highly hydrated monomer−monomer inter-
face buries 3500 Å2 of surface area per subunit (17% of the total
surface area). Four chloride ions are present at the interface,
associated with buried and otherwise uncompensated arginine
residues from each monomer (Arg120 and Arg354). When
for nonphysiological substrates accounted for most of the
difference (Table S4 of the Supporting Information), while
many Km values were comparable to that determined for
acetate.
A pattern consistent with weak competitive inhibition against
succinate (Ki = 150 mM) was observed in LCR assays modified
by the addition of citrate (Figure S5 of the Supporting
Information). Citrate is not an AarCH6 substrate, nor is it
capable of inducing acetyl-CoA hydrolysis.
Qualitatively, AarCH6 appears to have greater fidelity and
higher catalytic activity than other characterized class I CoA-
transferases.7,14,47−50 Substrate promiscuity indices, I = 0.12
and J = 0.32 for AarCH6 with the substrate set examined,
indicate a moderate level of promiscuity (Tables S5−S7 of the
Supporting Information).
8425
dx.doi.org/10.1021/bi300957f | Biochemistry 2012, 51, 8422−8434

Biochemistry
Article
Figure 2. Crystal structure of the AarCH6·CoA·citrate complex (PDB
entry 4eu7). AarCH6 is a dimer with four chloride ions (green
spheres) at the monomer−monomer interface. CoA and citrate are
bound in a deep, funnel-shaped active-site cleft between the N- and C-
terminal domains of each subunit. The molecular surface is depicted in
silhouette with the pseudo-2-fold axis lying vertical in the plane of the
page.
these ions were modeled as water molecules, the resulting B
factors were substantially lower than those of surrounding
protein atoms and residual density remained in the σA-
weighted mFo − DFc difference map. Weak anomalous density
centered around each chloride ion (PDB entry 4eua) provided
direct evidence of a third-row element at these positions
(Figure S8 of the Supporting Information). Given the buried
locations of Arg120 and Arg354 and the lack of nearby
negatively charged residues, the four chloride ions appear to be
important for AarC(H6) dimerization.
Hexagonal AarCH6 crystals also contained an asymmetric
unit corresponding to the biological dimer (PDB entry 4eu4).
These crystals had higher solvent content and diffracted to
lower resolution (Table S8 of the Supporting Information) but
had reasonable packing interactions among asymmetric units.
The orthogonal and hexagonal structures had no significant
differences (Cα rmsd = 0.54 Å for residues 2−505), which
suggests that crystal packing interactions do not significantly
perturb the structure of AarCH6.
AarCH6 that crystallized in the orthorhombic lattice
regularly possessed a complete and ordered hexahistidine tag
(residues 506−514) in subunit A, forming a 310-helix that is
involved in crystal packing interactions with an adjacent
asymmetric unit. Ordered hexahistidine tags are rare and are
often attributed to stabilizing packing contacts,52 which are
lacking for subunit B in the orthorhombic lattice and both
subunits in the hexagonal lattice, neither of which showed
density corresponding to the tag. Orthorhombic AarC·CoA·ci-
trate crystals (PDB entry 4eud) displayed no significant
structural differences relative to the same complex formed
with AarCH6 (PDB entry 4eu7; Cα rmsd = 0.17 Å for residues
2−505), which suggests that the addition of a hexahistidine tag
does not significantly perturb the structure of AarC. All other
wt and mutant AarCH6 structures were determined from
orthorhombic crystals (Table S8 of the Supporting Informa-
tion).
Figure 3. Electron density corresponding to bound ligands. Each panel
contains a σA-weighted 2mFo − DFc map, calculated with Glu294 or
Ala294 and ligands omitted, with a 2 Å carve radius. Adduct
attachment points are indicated with arrows. (A) AarCH6·CoA·citrate
complex (PDB entry 4eu7, subunit B) contoured at 1.0σ. The thiolate
moiety of CoA is modeled in two conformations (conf A and conf B).
(B) AarCH6-E294A·dethiaacetyl-CoA complex (PDB entry 4euc)
contoured at 1.0σ. (C) Acetylglutamyl anhydride adduct of the
AarCH6·CoA·acetate complex (PDB entry 4eu6, subunit A)
contoured at 0.9σ. An alternate view of the electron density in this
active site is provided in Figure S11 of the Supporting Information.
(D) Glutamyl-CoA thioester adduct of the AarCH6·CoA·acetate
complex (PDB entry 4eu6, subunit B) contoured at 0.6σ.
hydrophobic pocket near the entrance to the cleft, while the
pantetheine arm extends deeply into the active site, terminating
near Glu294. Citrate contacts only residues in the N-terminal
domain and the domain linker, forming extensive polar contacts
with the side chains of Ser71, Thr94, and Arg228. While citrate
is not an AarCH6 substrate (Table S4 of the Supporting
Information), its resemblance to succinate and acetate suggests
that this region might bind one or both physiological
carboxylate substrates. However, the large gap between citrate
and Glu294 (>7 Å) appears to eliminate the possibility that the
citrate binding site is the site of acetate or succinate binding in
either Michaelis complex.
CoA and Citrate Binding. Electron density consistent with
CoA and citrate was observed in the funnel-shaped cleft
between the N- and C-terminal domains of AarCH6 [PDB
entry 4eu7 (Figure 3A)]. With the exception of a single
hydrophobic interaction, CoA contacts only residues in the C-
terminal domain. The adenine moiety is located in a
Protein Conformational Dynamics. AarCH6 was ob-
served in several distinct conformations ranging from fully open
to fully closed (Figure 4). The active site in the fully open
conformation is readily accessible to acyl-CoA substrates.
Interactions with the CoA moiety trigger a progressive
constriction that can be envisioned to occur in three stages.
8426
dx.doi.org/10.1021/bi300957f | Biochemistry 2012, 51, 8422−8434

Biochemistry
Article
and Val295, that isolates the catalytic glutamate and CoA
thiol(ate) from bulk solvent. An accompanying rotation of the
domain linker places the side chain of Phe232 between α9 and
the mobile loop connecting β10 and α8. In the final stage,
continued rotation of the domain linker toward the N-terminal
domain pushes the side chain of Asn229 into the active site
while pulling the side chain of Arg228 onto the protein surface,
effectively removing the citrate binding site. Mobile regions
near the active site have relatively high B factors, consistent
with multiple conformations in the crystal.
CoA adopts a very similar conformation in most structures,
except for the thiolate moiety, which rotates inward as the
enzyme closes (Figure 4). The CoA C7″−N2″−C8″−C9″
dihedral angle is −106° (major conformer; −115° for the
minor conformer) in the mostly open conformation (PDB
entry 4eu7), −170° in the mostly closed conformation (PDB
entry 4eu5), and 177° in the fully closed conformation (PDB
entry 4eub). This progression suggests that enzyme closure
constrains the acyl-CoA thioester in a near-attack configuration.
Complexes Formed from Acyl-CoA Substrates or
Dethiaacetyl-CoA. Attempts to cocrystallize AarCH6 with
acetyl-CoA (or succinyl-CoA) failed to yield covalent anhydride
and CoA thioester intermediates that are characteristic of the
class I CoA-transferase reaction (Figure 1), probably because of
the short half-lives of the thioester substrates at pH 8 [∼10 h
(E. A. Mullins, unpublished observations)]. However, covalent
Figure 4. Stepwise constriction of AarCH6. The funnel-shaped
AarCH6 active-site cleft narrows upon binding CoA. The cleft opening
is in the foreground, and the cleft base, defined by Glu294, is in the
background. Residues in the fully open conformation are colored
green (PDB entry 4eu3), residues in the mostly open conformation
are colored cyan (PDB entry 4eu7), residues in the mostly closed
conformation are colored purple (PDB entry 4eu5), and residues in
the fully closed conformation are colored yellow (AarCH6-E294A,
PDB entry 4eub). CoA and citrate from all models are depicted in ball-
and-stick representation. All are from subunit B of the indicated
structures.
a
adducts with Glu294 were detected by freeze-trapping an
AarCH6 crystal soaked with acetyl-CoA (PDB entry 4eu6).
Subunit A contains the acetylglutamyl anhydride adduct and
CoA thiolate (Figure 3C and Figure S11 of the Supporting
Information). Subunit B contains acetate, bound to the side
chains of Ser71, Thr94, and Arg228, and the glutamyl-CoA
thioester adduct (Figure 3D). Acetate occupies the same site as
the pro-R acetyl moiety of citrate (PDB entry 4eu7), which is
too remote from Glu294 to represent a Michaelis complex
binding site. The only known source of acetate is acetyl-CoA
hydrolysis, possibly in the adjacent catalytic pocket. The CoA
conformation is the same in both freeze-trapped adducts,
except for an outward rotation of the thiolate moiety in the
anhydride form (C7″−N2″−C8″−C9″ dihedral angle of
−161°, compared to 167° in the glutamyl-CoA thioester). In
this conformation, the CoA thiolate is 3.4 and 3.2 Å from the
internal and external carbonyl carbons of the acetylglutamyl
anhydride, respectively, and stabilized by hydrogen bonding
interactions with Val270 NH and HOH517.
In the first stage, the active site remains largely open, but small,
local structural shifts improve hydrogen bonding, electrostatic,
and hydrophobic interactions with the adenine and ribose rings,
the 3′-phosphate, and the β-alanyl oxygen [O7″, acyl-CoA
numbering53 given in Figure 5 (Figure S10A of the Supporting
AarCH6-E294A was cocrystallized with dethiaacetyl-CoA,54
a
competitive inhibitor of AarCH6 that closely resembles acetyl-
CoA and binds with a Kd of 0.8 μM.28 Each active site in the
resulting structure (PDB entry 4euc) contains dethiaacetyl-CoA
in a nearly identical conformation (Figure 3B). The
dethiaacetyl-CoA “acetyl” group curves back toward the
pantoate moiety so that the ketone O1 atom forms hydrogen
bonds to dethiaacetyl-CoA N2″ and Gly388 NH (2.55 Å
average O1−N distance). The dethiaacetyl-CoA C7″−N2″−
C8″−C9″ dihedral angle (171°) is close to that in the glutamyl-
CoA thioester (PDB entry 4eu6, subunit B). When superposed
on a wt structure, dethiaacetyl-CoA C1 would be ∼2 Å from
the nearest Glu294 carboxylate oxygen, with an O−C1−O1
angle of 113°. This arrangement should allow acetyl-CoA to
adopt a near-optimal configuration for nucleophilic attack by
Glu29455 and to form two hydrogen bonds that stabilize the
oxyanion of the resulting tetrahedral intermediate.
Figure 5. Acetyl-CoA and the nonhydrolyzable analogue dethiaacetyl-
CoA are nearly isosteric except for altered geometries in the vicinity of
the atomic substitution. Indicated interatomic distances (in italics,
angstroms) and angles are based on equilibrium geometries computed
using the Hartree−Fock model and a 6-31G* basis set for ethyl
thioacetate (top) and 2-pentanone (bottom). Atom numbering follows
a previously described scheme for acyl-CoAs.53
Information)]. In the second stage, the active site shrinks
substantially. Helices α8−α11 and the loop connecting β10 and
α8 translate toward CoA to maximize hydrogen bonding and
electrostatic interactions with the diphosphate moiety and the
adjacent pantoate C3″ hydroxyl group. The concomitant
relocation of Val270 creates a catalytic pocket, defined by
hydrophobic residues Phe35, Thr36, Ala70, Phe92, Leu114,
8427
dx.doi.org/10.1021/bi300957f | Biochemistry 2012, 51, 8422−8434

Biochemistry
Article
Figure 6. Structural depiction of the SCACT reaction, illustrated in the reverse biosynthetic direction (acetyl-CoA + succinate → acetate + succinyl-
CoA). Crystallographically observed intermediates (ball-and-stick representation) and modeled intermediates (stick representation) were combined
to produce a complete structural description of the SCACT mechanism. (A) Unliganded. (B) Acetyl-CoA complex. The acetyl-CoA analogue
dethiaacetyl-CoA is colored black. (C) Acetyl anhydride adduct and CoA thiolate. (D) CoA thioester adduct and acetate. Acetate is depicted in two
positions. Position 2 outside the catalytic pocket was crystallographically observed (Figure 3C), and position 1 inside the catalytic pocket was
modeled. (E) CoA thioester adduct and succinate. (F) Succinyl anhydride adduct and CoA thiolate. (G) Succinyl-CoA complex. HOH518,
HOH519, and HOH520 (panel A) are displaced by the acyl/carboxylate moieties (except in panel D, model 2). Modeled intermediates were
manually positioned within the active site, and equilibrium conformations were determined using molecular mechanics calculations. Atoms shown in
ball-and-stick representation were either frozen or omitted during optimization. Yellow and red dotted lines indicate hydrogen bonds and possible
paths of nucleophilic attack, respectively.
Model Building. Models for missing reaction intermediates
were built using closely related crystallographic observations as
starting points. All were manually built prior to geometrical
optimization of a small number of atoms. In all cases, the polar
and hydrophobic contacts present in manually built models
were retained in the geometrically optimized final models
(Figure 6).
A model for the glutamylsuccinyl anhydride adduct [model F
(Table S15 of the Supporting Information)] was based on the
acetylglutamyl anhydride adduct (PDB entry 4eu6, subunit A).
A carboxymethyl group was appended to the acetyl anhydride
to create a fully extended succinyl anhydride. This con-
formation allowed the distal carboxylate to form hydrogen
bonding interactions with Thr36 NH and Ser71 Oγ and NH.
The optimized model made hydrophobic contacts with the side
chains of Phe35, Thr36, Ala70, Phe92, Leu114, Val270, and
Val295 and retained a nearly trans geometry.
A model for the acetyl-CoA Michaelis complex [model B
(Table S12 of the Supporting Information)] was based on the
dethiaacetyl-CoA complex (PDB entry 4euc). Hydrogen
bonding interactions with Gly388 NH and CoA N2″ were
retained. The only significant alterations in the optimized
model were associated with the replacement of methylene
C10″ with sulfur (Figure 5). (Acetyl-CoA C1 is ∼2 Å from the
nearest Glu294 carboxylate oxygen, with an O−C1−O1 angle
of ∼110°.) A model for the succinyl-CoA Michaelis complex
[model G (Table S16 of the Supporting Information)] was
based on model B and altered to contain the succinyl moiety
from model F in the place of the acetyl moiety.
Models for the Michaelis complexes containing acetate
[model D1 (Table S13 of the Supporting Information)] or
succinate [model E (Table S14 of the Supporting Informa-
tion)] were built adjacent to the observed glutamyl-CoA
thioester complex (PDB entry 4eu6, subunit B). Atoms from
the acyl moieties in the corresponding anhydride adducts (PDB
entry 4eu6, subunit A and model F, respectively) were used to
manually construct the models. The optimized models retained
the hydrogen bonding interactions and hydrophobic contacts
made by the anhydride adducts.
Structural Comparisons and Phylogenetic Analyses.
Paralogous groups within the class I CoA-transferase super-
family have diverse subunit and domain arrangements
consistent with low levels of sequence identity between groups.
Each group contains one or more enzyme families that
recognize different pairs of acyl-CoA and carboxylate substrates.
However, the superfamily is mechanistically homogeneous
(Figure 1), and its members have related backbone topologies
that conserve the three-dimensional arrangement of the four
key catalytic residues. Structural alignments (Figure 7A and
Table S17 of the Supporting Information) show that Glu294,
b
Asn347, and Gly388 homologues maintain a fixed relative
orientation in diverse class I CoA-transferases. Val270
homologues were crystallized in different locations along the
path this residue traverses as the enzyme closes.
8428
dx.doi.org/10.1021/bi300957f | Biochemistry 2012, 51, 8422−8434

Biochemistry
Article
proposed to form a hydrogen bond to the substrate γ-hydroxyl
moiety.59 The corresponding AarC residue is Gly34. AarC
residues that interact with citrate are not conserved in the
4HBCT group.
Kinetic and Structural Characterization of AarCH6
Mutants. Several soluble AarCH6 mutants were kinetically
(Table 3 and Table S19 of the Supporting Information) or
structurally (Table 2) characterized.
a
Table 3. Kinetic Data for AarCH6 Mutants
kcat/Km (mM−1 s−1
)
kcat/Km (μM−1 s−1
)
b
AarCH6 form
acetate
succinate
acetyl-CoA
succinyl-CoA
c
wt
4
79
0.030
0.3
3.4
0.7
9
S71A
0.029
0.6
0.074
d
e
e
R228E
2.3
0.13
f
E294A
∼0
0.16
∼0
2.7
∼0
0.7
5.2
∼0
0.39
9
N347A
E435D
3
73
a
Table S19 of the Supporting Information contains a full set of kinetic
parameters for each enzyme form with detectable catalytic activity.
Figure 7. Comparisons of class I CoA-transferase sequence and
structure. (A) Constellation of conserved active-site residues in 10
phylogenetically diverse class I CoA-transferases. AarC side chains and
ligands are depicted in ball-and-stick representation. The isolated red
sphere indicates a solvent atom located at a conserved location in
structures determined at sufficiently high resolution [HOH517 in
AarC(H6) structures]. Carbon atoms and ribbons are colored by PDB
entry: 4eu7 (AarCH6), black; 2nvv, teal; 2g39, green; 2oas, white;
3gk7,59 salmon; 3oxo,15 slate; 3qli,74 orange; 2ahv,14 yellow; 1k6d,82
gray; 1poi,51 magenta. Residue numbers, sequence accession codes,
and protein functional data are given in Tables S17 and S18 of the
Supporting Information. (B) Residues located near ligands in the
AarCH6·CoA·citrate complex (PDB entry 4eu7). A partial structure-
based sequence alignment highlights residues involved in enzyme
chemistry (red triangles), including the covalent attachment point
Glu294 (filled triangle), or that are part of the binding site for citrate
b
Activity measurements could not be performed for the insoluble
c
AarCH6-E435A and AarCH6-E435Q mutants. From ref 28.
d
AarCH6-R228E could not be fully saturated with either carboxylate
substrate. Values were estimated from the initial slopes of the substrate
e
saturation plots (Figure S18 of the Supporting Information). Acyl-
CoA kinetic parameters were determined at subsaturating, fixed
concentrations of succinate (550 mM) or acetate (750 mM).
f
AarCH6-E294A did not support activity above the estimated
∼0.009 s−1 detection limit.
Glu294 is the highly conserved site of covalent catalysis in
class I CoA-transferases. Mutations at this position have either
obliterated catalytic activity or caused aberrant thioesterase
activity.59,60 Introduction of a Glu294 → Ala mutation into
AarCH6 reduced the specific activity by >10000-fold. A crystal
structure of AarCH6-E294A shows no structural alterations
apart from a slight inward rotation of Thr36 (PDB entry 4eub).
Glu435 is conserved throughout class I CoA-transferases. Its
location outside the active-site cleft, in a conserved β-hairpin
motif, is inconsistent with an early proposal that it plays a role
in covalent catalysis in AarC.60 The catalytic properties of
AarCH6-E435D are nearly equivalent to those of wt AarCH6,
while E435A and E435Q mutants are completely insoluble.
Glu435 appears to have an important but purely structural role.
Asn347 is conserved in many class I CoA-transferases and
has been proposed to orient the key glutamate side chain for
nucleophilic attack14 or stabilize oxyanion intermediates formed
on the internal carbonyl.51 AarCH6-N347A had impaired
catalytic activity, but the apparent affinities for all four
substrates were largely unaffected. These findings are consistent
with a defect in catalysis that could arise from disruption of
either or both of its proposed roles.
Ser71 is highly conserved within the SCACT group but not
in other class I CoA-transferases (Figure S13 of the Supporting
Information). In AarCH6 structural models, Ser71 makes polar
side chain contacts to the distal carboxylate of succinate (Figure
6E−G). Ser71 also forms part of the citrate binding site,
interacting with the same pro-R carboxylate oxygen that
contacts Arg228 NH2.
AarCH6-S71A had impaired catalytic activity associated with
lower kcat values. A crystal structure of the AarCH6-S71A·CoA
complex (PDB entry 4eu8) revealed that disruption of the three
○
(c), succinate or the acyl group of succinyl-CoA (s), or CoA ( ).
Absolutely conserved residues are shown with black backgrounds.
Differentially conserved residues are shown with colored letters. The
full alignment is given in Figure S13 of the Supporting Information.
Phylogenetically diverse bacteria contain AarC paralogues
[SCACT group (Figure 7B and Figure S12 of the Supporting
Information)]. Biochemically characterized members of this
group recognize succinate/succinyl-CoA and are able to
perform the SCACT reaction.28,48,56,57 AarC residues with
important roles in catalysis, structure, CoA binding, or
interactions with the distal carboxylate in succinate/succinyl-
CoA models are highly conserved in the SCACT group. AarC
residues that interact with citrate are not widely conserved.
Structure-based sequence alignments, performed as described
in the Supporting Information, were used to identify proteins
that closely resemble AarC but have different substrate
preferences. The SCACT group is highly similar in structure
and sequence to the 4-hydroxybutyrate CoA-transferase
(4HBCT) group (Figure 7B,C). However, succinate is not a
substrate for Clostridium aminobutyricum 4HBCT,58 which
suggests that residues involved in recognizing distal portions of
the C4 substrate are different in the SCACT and 4HBCT
groups. AarC residues with important roles in catalysis,
structure, or CoA binding are highly conserved in the
4HBCT group, but those that interact with the distal
carboxylate in succinate/succinyl-CoA models are not. In C.
aminobutyricum 4HBCT, the side chain of His31 has been
8429
dx.doi.org/10.1021/bi300957f | Biochemistry 2012, 51, 8422−8434

Biochemistry
Article
hydrogen bonds involving Ser71 induced a partial collapse of
the catalytic pocket, in a region that would appear to be
necessary to accommodate succinate and possibly acetate
(Figure S16 of the Supporting Information). In one subunit, a
shift also occurred in a surface-exposed unit of the N-terminal
domain (Figure S17 of the Supporting Information).
Arg228 is found in only AarC and several closely allied
SCACT group sequences (red taxa in Figure S12 of the
Supporting Information). As described above, Ser71 and
Arg228 interact with each other and with acetate or citrate in
the auxiliary binding site outside the catalytic pocket. Arg228 is
initially displaced from the active-site cleft as the enzyme
constricts around acyl-CoA (PDB entry 4eub) but returns to
allow partial reopening and carboxylate exchange from the
glutamyl-CoA thioester adduct.
Figure 8. Stabilization of tetrahedral oxyanion intermediates within
the AarCH6 active site. (A) Intermediate formed by the attack of
Glu294 on the acyl-CoA carbonyl (Figure 6, B → C) and by the attack
of CoA thiolate on the external carbonyl of the mixed anhydride
adduct (F → G). (B) Intermediate formed by the attack of CoA
thiolate on the internal carbonyl of the mixed anhydride adduct (C →
D) and by the attack of the carboxylate substrate on the glutamyl
thioester carbonyl (D → E).
AarCH6-R228E was not saturated at a concentration of 0.5
M of either physiological carboxylate substrate (Figure S18 of
the Supporting Information). An acetyl-CoA saturation curve
with a normal shape, obtained at the highest achievable
succinate concentration, furnished a specificity constant
comparable to that of wt AarCH6, indicating AarCH6-R228E
had a specific defect in its ability to bind both carboxylate
substrates. (A modest defect in succinyl-CoA binding cannot be
ruled out.)
Crystallographic analysis of AarCH6-R228E (PDB entry
4eu9) showed that the side chain of Glu228 is well-ordered and
is displaced from the citrate binding region (Figure S14 of the
Supporting Information). Nonetheless, the charge-reversed
mutation diminishes the positive electrostatic potential in the
active-site cleft (Figure S9 of the Supporting Information). A
lower affinity for carboxylate substrates could be the result of
diminished electrostatic attraction to the active-site cleft,
disruption of a carboxylate binding function within the auxiliary
binding site, perturbation of the conformational equilibrium, or
a combination of multiple effects. Several carboxylates with
charge-reversed distal moieties (e.g., 4-aminobutyrate) were not
accepted as substrates by AarCH6-R228E (data not shown).
hole. This restrictive conformation places the reacting atoms in
a Burgi−Dunitz configuration55 suitable for attack by the syn
̈
orbital of the carboxylate63 and exerts strain that would be
relieved in the tetrahedral oxyanion intermediate and transition
state.7 In the modeled acetyl-CoA Michaelis complex, the
Glu294 carboxylate oxygen and thioester carbonyl carbon
atoms are in van der Waals contact. The same CoA
conformation is observed in complexes containing dethiaace-
tyl-CoA, the acetylglutamyl anhydride adduct, and the
glutamyl-CoA thioester adduct, which means minimal motion
is required for the conversion of an acyl-CoA Michaelis
complex to a carboxylate Michaelis complex (Figure 6, panels B
→ D and G → E). Previous studies of on-enzyme equilibria
indicate that the acylglutamyl anhydride adduct may be less
stable than the glutamyl-CoA thioester adduct.6,7,9 Trapping of
the acetylglutamyl anhydride adduct (Figure 3C and Figure S11
of the Supporting Information) suggests that anhydride
cleavage is suppressed within the crystalline environment.
Acylglutamyl anhydride adduct thiolysis occurs at two
carbonyls separated by only 2.5 Å, performed by two regions
of the catalytic pocket that (arguably) represent separate active
sites. These reactions are facilitated by discrete external and
internal oxyanion holes that work together (Figure 8). The
leaving group potential of the carboxylate product, formed by
thiolysis of the internal anhydride carbonyl, is increased by
interactions with CoA N2″ and Gly388 NH [the external
oxyanion hole, serving a different function (Figure 6D, position
1)]. The leaving group potential of Glu294, formed by thiolysis
of the external anhydride carbonyl, is increased by interactions
with Asn347 Nδ [the internal oxyanion hole, serving a different
function (Figure 6G)]. The leaving group potential of the CoA
thiolate expelled by thioester cleavage is increased by
interactions with Val270 NH and HOH517 (Figure 6C,F).
Protein conformational changes triggered by acyl-CoA binding
preorganize Val270 NH and CoA N2″ appropriately. With
these hydrogen bond donors in their proper places, the enzyme
can perform both reactions without much substrate motion,
using the adjacent oxyanion holes as a kind of two-stroke
engine. This chemical strategy manifests an unprecedented
catalytic role for CoA N2″, substrate destabilization,16,64
enzyme selection for near-attack configurations,65 and the
electrostatic preorganization of active sites,66 each requiring
precise protein conformational control. Transferase enzymes in
general undergo larger conformational changes than hydro-
lases,67 even when enzymes from the same superfamily are
compared,68 a finding that underscores the widely accepted role
DISCUSSION
■
The kinetic mechanism and covalent intermediates of the
archetypal class I CoA-transferase SCOT were established more
than 35 years ago.2,3,5−8 Subsequent studies predicted that
protein conformational changes driven by substrate binding7,8
allow enzyme−substrate interactions remote from the site of
chemistry9−11 to contribute to catalytic accelerations.16 Crystal
structures of AarC supplemented with conservative models
illustrate how class I CoA-transferases use binding energy to
induce a conformational change that accelerates elegant tandem
acyl transfer reactions and to select for appropriate acyl-CoA
and carboxylate substrates (Figure 6).
Mechanistic Strategies Used by AarC and Other Class
I CoA-Transferases. The structural data presented here are
consistent with the Jencks model for catalysis by class I CoA-
transferases.16 Acyl-CoA binding leads to the assembly of an
oxyanion hole for the first tetrahedral intermediate (Figure 8A),
a case of the “gear-shift” type61 of induced fit between the
enzyme and substrate.62 As the active site constricts, the acyl-
CoA (represented by dethiaacetyl-CoA) adopts a curled
conformation, observed here for the first time, that is clamped
against the nucleophilic glutamate carboxylate in an orientation
suitable for nucleophilic attack. During this process, Glu294
shifts slightly and a rotation around the acyl-CoA C8″−C9″
bond places the thioester oxygen into a preorganized oxyanion
8430
dx.doi.org/10.1021/bi300957f | Biochemistry 2012, 51, 8422−8434

Biochemistry
Article
of protein motions in regulating access to labile intermediates.
The dynamic assembly of a competent tandem catalytic site
may provide yet another tool for suppressing unproductive side
reactions by class I CoA-transferases.
Furthermore, the conserved protein fold and constellation of
active-site residues strongly imply a common mechanism for all
class I CoA-transferases.
Comparison to Class III CoA-Transferases. The kinetic
mechanisms for class I and class III CoA-transferases
differ,18,69,70 with the latter forming a pivotal aspartyl-CoA
thioester complex that contains two carboxylates.19 The results
presented here show that there are differences all along the
reaction coordinate. Structures of Oxalobacter formigenes FRC
complexes containing CoA, either alone (PDB entry 1p5r) or
with acylaspartyl anhydride adducts (PDB entries 1t4c and
2vjm), bind CoA in the same location, but the CoA moiety
shifts substantially in the aspartyl-CoA thioester adduct (PDB
entries 2vjk and 2vjm).18,19,71 The motion propagates along the
entire length of CoA, with some atoms moving 9 Å. In each
anhydride adduct, CoA N2″ is at least 5 Å from either
anhydride carbonyl oxygen and moves 5 Å in the conversion
between anhydride and CoA adducts. These observations are
incompatible with any catalytic role for CoA N2″ in class III
CoA-transferase chemistry and are very different from those for
the fixed CoA conformation observed here in comparable AarC
adducts.
Relative to AarC, FRC appears to move very little during the
reaction sequence, outside a mobile, glycine-rich loop that is
proposed to protect labile enzyme intermediates71 and binds
oxalate at a substantial remove from the site of catalysis.19,72
This region may perform a role similar to that of the AarC
auxiliary site, Val270 gate, or both (see below). Even though
they use the same sequence of reactions, class I and class III
CoA-transferases employ different kinetic mechanisms. More-
over, these enzyme superfamilies appear to use fundamentally
different chemical strategies that are not the result of
convergent evolution.
Substrate Selection. Class I CoA-transferase enzyme
chemistry is closely coordinated with a protein conformational
cycle that has been likened to breathing.73 Enzyme closure
induced by acyl-CoA binding (Figure 4) creates an environ-
ment suitable for acyl transfer chemistry and protects labile
intermediates (Figure 9B,C). Within the catalytic pocket, class I
CoA-transferases have been proposed to recognize distal
functional groups in preferred carboxylate or acyl-CoA
substrates using chemically diverse side chains (e.g., Ser71)
identified by model building.12,13,74 Each set of residues is
conserved only within a subset of functionally related class I
CoA-transferases, consistent with roles in substrate selection.
Good substrates allow complete enzyme closure (Figure 9) and
the resulting immobilization of reactive groups.7 Poor
substrates that are too large [e.g., methylsuccinate (Table 1)]
may interfere with the critical clamping, desolvation, and
hydrogen bonding functions of Val270.
Figure 9. Regulation of carboxylate access to the active-site glutamate.
Val270 forms part of a dynamic aperture separating the auxiliary
carboxylate binding site from the catalytic pocket. The active-site cleft
is shown in cross section, narrowing from the top left to the catalytic
pocket adjacent to Glu294. The clipping plane for all panels is placed
at the narrowest point (∼3.5 Å) of the aperture. Ser71 is located at the
bottom of each panel between the viewer and the clipping plane.
Ligand atoms and backbone atoms are rendered in ball-and-stick and
ribbon representations, respectively. Val270 Cβ and Cγ are colored
purple. The interior of the protein is colored black. Structures in
panels B−D were aligned on the structure in panel A. Citrate from the
complex in panel A is shown as thin green lines in other panels; in all
panels, the central carboxylate is hidden by the clipping plane. (A) The
AarCH6·CoA·citrate complex (PDB entry 4eu7, subunit B) adopts the
open conformation. (B) The AarCH6-E294A·dethiaacetyl-CoA
complex (PDB entry 4euc, subunit B) adopts the closed conformation
with Arg228 extruded from the active site. The clipping plane was
moved toward the viewer by 0.8 Å for the acetyl atoms only.
Additional structures possessing Arg228 or Glu228 in the outward
orientation are illustrated in Figure S14 of the Supporting Information.
(C) The AarCH6·CoA·acetate complex (acetylglutamyl anhydride
adduct, PDB entry 4eu6, subunit A) adopts the closed conformation.
(D) The AarCH6·CoA·acetate complex (glutamyl-CoA thioester
adduct, PDB entry 4eu6, subunit B) adopts the closed conformation.
To the right of the aperture, the catalytic pocket is empty. High
relative B factors for the loop containing Val270 suggest conforma-
tional motion that could have allowed acetate to leave the catalytic
pocket.
Carboxylate product formation (Figure 1) accompanies
partial enzyme reopening, to allow carboxylate exchange
(Figure 9D). Rapid carboxylate exchange may suppress
unproductive hydrolysis of the glutamyl-CoA thioester adduct.
Val270 defines one side of the aperture, which is wide enough
to admit acetate and succinate to the catalytic pocket but not
large carboxylates like citrate. Like Val227 in Yersinia pestis
RipA,74 Val270 appears to be a dynamic “gating” residue.75−77
Therefore, Val270 has a dual influence on carboxylate substrate
selectivity, as a gate and as a clamp.
8431
dx.doi.org/10.1021/bi300957f | Biochemistry 2012, 51, 8422−8434

Biochemistry
Article
Class I CoA-transferases appear to use positive potential in
the active-site cleft to recruit carboxylate substrates,51,74 like
several enzymes that catalyze diffusion-limited conversions of
anionic substrates.78−81 AarC has presumably been subject to
strong selection for the efficient conversion of acetate to acetyl-
CoA, consistent with a high kcat value (Table S4 of the
Supporting Information).28 One such adaptation may be an
auxiliary carboxylate site located outside the catalytic pocket
near gating residue Val270, which contains the dynamic, AarC-
specific residue Arg228 (Figure 9A). Studies of AarCH6-R228E
indicate that Arg228 has an important kinetic role in
carboxylate substrate binding. The auxiliary site nonselectively
binds carboxylates at the threshold of the catalytic pocket, while
selectivity is enforced by the conserved gating residue Val270
and the interior of the catalytic pocket. Binding of ≤2 mM
acetate produced by acetyl-CoA cleavage (Figure 3D) suggests
that the auxiliary site is traversed by carboxylates entering or
exiting the catalytic pocket. Conformational changes that expel
Arg228 from the vicinity of the active site may further increase
the rate of carboxylate product dissociation.
This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: kappock@purdue.edu. Phone: (765) 494-8383. Fax:
(765) 494-1897.
Funding
This work was supported by the Herman Frasch Foundation
for Research in Agricultural Chemistry (531-HF02), the
National Science Foundation (MCB 0936108 and MCB
0347250), and Purdue University Agricultural Research
Programs. Use of the Advanced Photon Source was supported
by the U.S. Department of Energy, Office of Science, Office of
Basic Energy Sciences, under Contract DE-AC02-06CH11357.
Use of LS-CAT Sector 21 was supported by the Michigan
Economic Development Corp. and the Michigan Technology
Tri-Corridor for the support of this research program (Grant
085P1000817).
Notes
SUMMARY
The authors declare no competing financial interest.
■
AarC evolved to play a primary metabolic role as SCACT under
the selective pressure induced by high concentrations of
membrane-permeant acetic acid, producing a CoA-transferase
with relatively high substrate specificity and high turnover
numbers. A novel oxyanion hole configuration involving CoA
stabilizes one of two tetrahedral intermediates formed less than
3 Å apart. The proposal by Jencks that CoA binding energy is
diverted into protein conformational changes that accelerate
catalysis is fully substantiated by crystallographic depictions of
what seems to be a common mechanism for class I CoA-
transferases. An auxiliary binding site just outside the catalytic
pocket appears to preconcentrate carboxylate substrates and
might explain the relatively efficient conversion of the
physiological substrate acetate.
ACKNOWLEDGMENTS
■
We thank Hong Jiang and Aaron Ransome for preparing
dethiaacetyl-CoA, Kayleigh Nyffeler for help with mutant
protein isolations, Kelly Sullivan for gel-filtration analysis,
Spencer Anderson for collecting anomalous X-ray data, and
Barb Golden for advice on refinement methods.
ABBREVIATIONS
■
4HBCT, 4-hydroxybutyrate CoA-transferase; AAB, acetic acid
bacteria; aar, acetic acid resistance; AarC, SCACT from A. aceti;
AarCH6, AarC with a C-terminal hexahistidine tag; CAC, citric
acid cycle; CD, circular dichroism; CoA, coenzyme A; ESI-
TOF-MS, electrospray ionization time-of-flight mass spectrom-
etry; FRC, formyl-CoA transferase; NCS, noncrystallographic
symmetry; ODN, oligodeoxynucleotide; PCT, propionate
CoA-transferase; PDB, Protein Data Bank; rmsd, root-mean-
square deviation; SCACT, succinyl-CoA:acetate CoA-trans-
ferase; SCOT, succinyl-CoA:3-oxoacid CoA-transferase; wt,
wild-type.
ASSOCIATED CONTENT
■
S
* Supporting Information
Methods for protein purification, CD analysis, ESI-TOF-MS
analysis, and alignment of structures and sequences, oligodeox-
ynucleotide sequences (Table S1), isolation of AarC (Table
S2), ESI-TOF-MS data (Table S3), kinetic data for alternate
substrates (Table S4), substrate promiscuity data (Tables S5−
S7), X-ray crystallographic data collection and refinement
statistics (Tables S8−S11), atomic coordinates for modeled
atoms (Tables S12−S16), conserved catalytic residues in
diverse class I CoA-transferases (Table S17), class I CoA-
transferase structural and functional data (Table S18), steady-
state kinetic data for AarCH6 mutants (Table S19), acetate
saturation curves (Figure S1), CD analysis (Figures S2−S4),
citrate inhibition plot (Figure S5), crystal images (Figure S6),
AarC topology diagram (Figure S7), chloride-site electron
density maps (Figure S8), electrostatic surface renderings
(Figure S9), CoA/citrate binding modes (Figure S10),
acetylglutamyl anhydride adduct electron density maps (Figure
S11), SCACT/4HBCT phylograms (Figure S12), structure-
based sequence alignment (Figure S13), cutaway views of
AarCH6 mutant active-site clefts (Figure S14), concentration-
dependent AarCH6-S71A activity plot (Figure S15), conforma-
tional motions of AarCH6-S71A (Figures S16 and S17), and
carboxylate saturation curves for AarCH6-R228E (Figure S18).
ADDITIONAL NOTES
■
a
The mismatched active sites may be the result of tight crystal
packing in the vicinity of subunit A (primarily acetylglutamyl
anhydride adduct), in which the CoA 3′-phosphate interacts
with Nε of Lys240B in an adjacent asymmetric unit. The active
site in subunit B (primarily glutamyl-CoA thioester adduct) is
more accessible. A similar phenomenon may have allowed the
detection of an aspartyloxalyl anhydride adduct in one subunit
of the class III CoA-transferase FRC.18
In some distantly related class I CoA-transferases, Asn347
appears to be functionally replaced by Gln99 (SCOT
numbering).
b
REFERENCES
■
(1) Jencks, W. P. (1973) In The Enzymes (Boyer, P. D., Ed.) Vol. 9,
Chapter 11, pp 483−496, Academic Press, San Diego.
(2) Hersh, L. B., and Jencks, W. P. (1967) Coenzyme A transferase.
Kinetics and exchange reactions. J. Biol. Chem. 242, 3468−3480.
8432
dx.doi.org/10.1021/bi300957f | Biochemistry 2012, 51, 8422−8434

Biochemistry
Article
(3) Hersh, L. B., and Jencks, W. P. (1967) Coenzyme A transferase.
Isolation and properties of an enzyme-coenzyme A intermediate. J.
Biol. Chem. 242, 3481−3486.
(4) Hersh, L. B., and Jencks, W. P. (1967) Isolation of an enzyme-
coenzyme A intermediate from succinyl coenzyme A-acetoacetate
coenzyme A transferase. J. Biol. Chem. 242, 339−340.
(24) Nakano, S., and Fukaya, M. (2008) Analysis of proteins
responsive to acetic acid in Acetobacter: Molecular mechanisms
conferring acetic acid resistance in acetic acid bacteria. Int. J. Food
Microbiol. 125, 54−59.
(25) Francois, J. A., and Kappock, T. J. (2007) Alanine racemase
from the acidophile Acetobacter aceti. Protein Expression Purif. 51, 39−
48.
(5) Solomon, F., and Jencks, W. P. (1969) Identification of an
enzyme-γ-glutamyl coenzyme A intermediate from coenzyme A
transferase. J. Biol. Chem. 244, 1079−1081.
(26) Azuma, Y., Hosoyama, A., Matsutani, M., Furuya, N., Horikawa,
H., Harada, T., Hirakawa, H., Kuhara, S., Matsushita, K., Fujita, N., and
Shirai, M. (2009) Whole-genome analyses reveal genetic instability of
Acetobacter pasteurianus. Nucleic Acids Res. 37, 5768−5783.
(27) Sakurai, K., Arai, H., Ishii, M., and Igarashi, Y. (2011)
Transcriptome response to different carbon sources in Acetobacter
aceti. Microbiology 157, 899−910.
(28) Mullins, E. A., Francois, J. A., and Kappock, T. J. (2008) A
specialized citric acid cycle requiring succinyl-coenzyme A (CoA):a-
cetate CoA-transferase (AarC) confers acetic acid resistance on the
acidophile Acetobacter aceti. J. Bacteriol. 190, 4933−4940.
(29) Cole, P. A. (1996) Chaperone-assisted protein expression.
Structure 4, 239−242.
(30) Francois, J. A., Starks, C. M., Sivanuntakorn, S., Jiang, H.,
Ransome, A. E., Nam, J.-W., Constantine, C. Z., and Kappock, T. J.
(2006) Structure of a NADH-insensitive hexameric citrate synthase
that resists acid inactivation. Biochemistry 45, 13487−13499.
(31) Bradford, M. M. (1976) A rapid and sensitive method for the
quantitation of microgram quantities of protein utilizing the principle
of protein-dye binding. Anal. Biochem. 72, 248−254.
(32) Nath, A., and Atkins, W. M. (2008) A quantitative index of
substrate promiscuity. Biochemistry 47, 157−166.
(6) Pickart, C. M., and Jencks, W. P. (1979) Formation of stable
anhydrides from CoA transferase and hydroxamic acids. J. Biol. Chem.
254, 9120−9129.
(7) White, H., and Jencks, W. P. (1976) Mechanism and specificity of
succinyl-CoA:3-ketoacid coenzyme A transferase. J. Biol. Chem. 251,
1688−1699.
(8) White, H., Solomon, F., and Jencks, W. P. (1976) Utilization of
the inactivation rate of coenzyme A transferase by thiol reagents to
determine properties of the enzyme-CoA intermediate. J. Biol. Chem.
251, 1700−1707.
(9) Moore, S. A., and Jencks, W. P. (1982) Formation of active site
thiol esters of CoA transferase and the dependence of catalysis on
specific binding interactions. J. Biol. Chem. 257, 10893−10907.
(10) Fierke, C. A., and Jencks, W. P. (1986) Two functional domains
of coenzyme A activate catalysis by coenzyme A transferase.
Pantetheine and adenosine 3′-phosphate 5′-diphosphate. J. Biol.
Chem. 261, 7603−7606.
(11) Whitty, A., Fierke, C. A., and Jencks, W. P. (1995) Role of
binding energy with coenzyme A in catalysis by 3-oxoacid coenzyme A
transferase. Biochemistry 34, 11678−11689.
(33) Teng, T.-Y. (1990) Mounting of crystals for macromolecular
crystallography in a free-standing thin film. J. Appl. Crystallogr. 23,
387−391.
(12) Bateman, K. S., Brownie, E. R., Wolodko, W. T., and Fraser, M.
E. (2002) Structure of the mammalian CoA transferase from pig heart.
Biochemistry 41, 14455−14462.
(34) Otwinowski, Z., and Minor, W. (1997) Processing of X-ray
diffraction data collected in oscillation mode. Methods Enzymol. 276,
307−326.
(13) Tammam, S. D., Rochet, J.-C., and Fraser, M. E. (2007)
Identification of the cysteine residue exposed by the conformational
change in pig heart succinyl-CoA:3-ketoacid coenzyme A transferase
on binding coenzyme A. Biochemistry 46, 10852−10863.
(14) Rangarajan, E. S., Li, Y., Ajamian, E., Iannuzzi, P., Kernaghan, S.
D., Fraser, M. E., Cygler, M., and Matte, A. (2005) Crystallographic
trapping of the glutamyl-CoA thioester intermediate of family I CoA
transferases. J. Biol. Chem. 280, 42919−42928.
(35) Guex, N., and Peitsch, M. C. (1997) SWISS-MODEL and the
Swiss-PdbViewer: An environment for comparative protein modeling.
Electrophoresis 18, 2714−2723.
(36) Adams, P. D., et al. (2010) PHENIX: A comprehensive Python-
based system for macromolecular structure solution. Acta Crystallogr.
D66, 213−221.
(37) Emsley, P., Lohkamp, B., Scott, W. G., and Cowtan, K. (2010)
Features and development of Coot. Acta Crystallogr. D66, 486−501.
(38) Kleywegt, G. J. (2007) Crystallographic refinement of ligand
complexes. Acta Crystallogr. D63, 94−100.
(39) DeLano, W. L. (2002) The PyMOL molecular graphics system,
DeLano Scientific, Palo Alto, CA.
(40) Dolinsky, T. J., Nielsen, J. E., McCammon, J. A., and Baker, N.
A. (2004) PDB2PQR: An automated pipeline for the setup of Poisson-
Boltzmann electrostatics calculations. Nucleic Acids Res. 32, W665−
W667.
(41) Baker, N. A., Sept, D., Joseph, S., Holst, M. J., and McCammon,
J. A. (2001) Electrostatics of nanosystems: Application to micro-
tubules and the ribosome. Proc. Natl. Acad. Sci. U.S.A. 98, 10037−
10041.
(42) Miroux, B., and Walker, J. E. (1996) Over-production of
proteins in Escherichia coli: Mutant hosts that allow synthesis of some
membrane proteins and globular proteins at high levels. J. Mol. Biol.
260, 289−298.
(43) Francis, D. M., and Page, R. (2001) In Current Protocols in
Protein Science (Coligan, J. E., Dunn, B. M., Ploegh, H. L., Speicher, D.
W., and Wingfield, P. T., Eds.) Chapter 61, pp 5.24.1−5.24.29, Wiley,
New York.
(44) Menzel, U., and Gottschalk, G. (1985) The internal pH of
Acetobacterium wieringae and Acetobacter aceti during growth and
production of acetic acid. Arch. Microbiol. 143, 47−51.
(45) Constantine, C. Z., Starks, C. M., Mill, C. P., Ransome, A. E.,
Karpowicz, S. J., Francois, J. A., Goodman, R. A., and Kappock, T. J.
(2006) Biochemical and structural studies of N5-carboxyaminoimida-
(15) Fraser, M. E., Hayakawa, K., and Brown, W. D. (2010) Catalytic
role of the conformational change in succinyl-CoA:3-oxoacid CoA
transferase on binding CoA. Biochemistry 49, 10319−10328.
(16) Jencks, W. P. (1987) Economics of enzyme catalysis. Cold
Spring Harbor Symp. Quant. Biol. 52, 65−73.
(17) Heider, J. (2001) A new family of CoA-transferases. FEBS Lett.
509, 345−349.
(18) Jonsson, S., Ricagno, S., Lindqvist, Y., and Richards, N. G. J.
(2004) Kinetic and mechanistic characterization of the formyl-CoA
transferase from Oxalobacter formigenes. J. Biol. Chem. 279, 36003−
36012.
(19) Berthold, C. L., Toyota, C. G., Richards, N. G. J., and Lindqvist,
Y. (2008) Reinvestigation of the catalytic mechanism of formyl-CoA
transferase, a class III CoA-transferase. J. Biol. Chem. 283, 6519−6529.
(20) Fukaya, M., Takemura, H., Okumura, H., Kawamura, Y.,
Horinouchi, S., and Beppu, T. (1990) Cloning of genes responsible for
acetic acid resistance in Acetobacter aceti. J. Bacteriol. 172, 2096−2104.
(21) Fukaya, M., Takemura, H., Tayama, K., Okumura, H.,
Kawamura, Y., Horinouchi, S., and Beppu, T. (1993) The aarC gene
responsible for acetic acid assimilation confers acetic acid resistance on
Acetobacter aceti. J. Ferment. Bioeng. 76, 270−276.
(22) Ohmori, S., Masai, H., Arima, K., and Beppu, T. (1980)
Isolation and identification of acetic acid bacteria for submerged acetic
acid fermentation at high temperature. Agric. Biol. Chem. 44, 2901−
2906.
(23) Ohmori, S., Uozumi, T., and Beppu, T. (1982) Loss of acetic
acid resistance and ethanol oxidizing ability in an Acetobacter strain.
Agric. Biol. Chem. 46, 381−389.
8433
dx.doi.org/10.1021/bi300957f | Biochemistry 2012, 51, 8422−8434

Biochemistry
Article
zole ribonucleotide mutase (PurE) from the acidophilic bacterium
Acetobacter aceti. Biochemistry 45, 8193−8208.
(46) Mullins, E. A., Starks, C. M., Francois, J. A., Sael, L., Kihara, D.,
and Kappock, T. J. (2012) Formyl-coenzyme A (CoA):oxalate CoA-
transferase from the acidophile Acetobacter aceti has a distinctive
electrostatic surface and inherent acid stability. Protein Sci. 21, 686−
696.
(66) Kamerlin, S. C. L., Chu, Z. T., and Warshel, A. (2010) On
catalytic preorganization in oxyanion holes: Highlighting the problems
with the gas-phase modeling of oxyanion holes and illustrating the
need for complete enzyme models. J. Org. Chem. 75, 6391−6401.
(67) Koike, R., Amemiya, T., Ota, M., and Kidera, A. (2008) Protein
structural change upon ligand binding correlates with enzymatic
reaction mechanism. J. Mol. Biol. 379, 397−401.
(68) Koike, R., Kidera, A., and Ota, M. (2009) Alteration of
oligomeric state and domain architecture is essential for functional
transformation between transferase and hydrolase with the same
scaffold. Protein Sci. 18, 2060−2066.
(69) Dickert, S., Pierik, A. J., Linder, D., and Buckel, W. (2000) The
involvement of coenzyme A esters in the dehydration of (R)-
phenyllactate to (E)-cinnamate by Clostridium sporogenes. Eur. J.
Biochem. 267, 3874−3884.
(70) Leutwein, C., and Heider, J. (2001) Succinyl-CoA:(R)-
benzylsuccinate CoA-transferase: An enzyme of the anaerobic toluene
catabolic pathway in denitrifying bacteria. J. Bacteriol. 183, 4288−4295.
(71) Ricagno, S., Jonsson, S., Richards, N., and Lindqvist, Y. (2003)
Formyl-CoA transferase encloses the CoA binding site at the interface
of an interlocked dimer. EMBO J. 22, 3210−3219.
(72) Gruez, A., Roig-Zamboni, V., Valencia, C., Campanacci, V., and
Cambillau, C. (2003) The crystal structure of the Escherichia coli YfdW
gene product reveals a new fold of two interlaced rings identifying a
wide family of CoA transferases. J. Biol. Chem. 278, 34582−34586.
(73) Macieira, S., Zhang, J., Buckel, W., and Messerschmidt, A.
(2012) Crystal structure of the complex between 4-hydroxybutyrate
CoA-transferase from Clostridium aminobutyricum and CoA. Arch.
Microbiol. 194, 157−166.
(74) Torres, R., Swift, R. V., Chim, N., Wheatley, N., Lan, B.,
Atwood, B. R., Pujol, C., Sankaran, B., Bliska, J. B., Amaro, R. E., and
Goulding, C. W. (2011) Biochemical, structural and molecular
dynamics analyses of the potential virulence factor RipA from Yersinia
pestis. PLoS One 6, e25084.
(75) Zhou, H.-X., Wlodek, S. T., and McCammon, J. A. (1998)
Conformation gating as a mechanism for enzyme specificity. Proc. Natl.
Acad. Sci. U.S.A. 95, 9280−9283.
(76) Zhou, H.-X., and McCammon, J. A. (2010) The gates of ion
channels and enzymes. Trends Biochem. Sci. 35, 179−185.
(77) McCammon, J. A. (2011) Gated diffusion-controlled reactions.
BMC Biophys. 4, 4.
(78) Sheridan, R. P., and Allen, L. C. (1981) The active site
electrostatic potential of human carbonic anhydrase. J. Am. Chem. Soc.
103, 1544−1550.
(79) Getzoff, E. D., Tainer, J. A., Weiner, P. K., Kollman, P. A.,
Richardson, J. S., and Richardson, D. C. (1983) Electrostatic
recognition between superoxide and copper, zinc superoxide
dismutase. Nature 306, 287−290.
(80) Ripoll, D. R., Faerman, C. H., Axelsen, P. H., Silman, I., and
Sussman, J. L. (1993) An electrostatic mechanism for substrate
guidance down the aromatic gorge of acetylcholinesterase. Proc. Natl.
Acad. Sci. U.S.A. 90, 5128−5132.
(81) Tan, R. C., Truong, T. N., McCammon, J. A., and Sussman, J. L.
(1993) Acetylcholinesterase: Electrostatic steering increases the rate of
ligand binding. Biochemistry 32, 401−403.
(82) Korolev, S., Koroleva, O., Petterson, K., Gu, M., Collart, F.,
Dementieva, I., and Joachimiak, A. (2002) Autotracing of Escherichia
coli acetate CoA-transferase α-subunit structure using 3.4 Å MAD and
1.9 Å native data. Acta Crystallogr. D58, 2116−2121.
(47) Buckel, W., Dorn, U., and Semmler, R. (1981) Glutaconate
CoA-transferase from Acidaminococcus fermentans. Eur. J. Biochem. 118,
315−321.
́
(48) Haller, T., Buckel, T., Retey, J., and Gerlt, J. A. (2000)
Discovering new enzymes and metabolic pathways: Conversion of
succinate to propionate by Escherichia coli. Biochemistry 39, 4622−
4629.
(49) Selmer, T., Willanzheimer, A., and Hetzel, M. (2002)
Propionate CoA-transferase from Clostridium propionicum. Cloning
of gene and identification of glutamate 324 at the active site. Eur. J.
Biochem. 269, 372−380.
(50) Fleck, C. B., and Brock, M. (2008) Characterization of an acyl-
CoA:carboxylate CoA-transferase from Aspergillus nidulans involved in
propionyl-CoA detoxification. Mol. Microbiol. 68, 642−656.
(51) Jacob, U., Mack, M., Clausen, T., Huber, R., Buckel, W., and
Messerschmidt, A. (1997) Glutaconate CoA-transferase from Acid-
aminococcus fermentans: The crystal structure reveals homology with
other CoA-transferases. Structure 5, 415−426.
(52) Carson, M., Johnson, D. H., McDonald, H., Brouillette, C., and
Delucas, L. J. (2007) His-tag impact on structure. Acta Crystallogr.
D63, 295−301.
(53) Wu, W.-J., Anderson, V. E., Raleigh, D. P., and Tonge, P. J.
(1997) Structure of hexadienoyl-CoA bound to enoyl-CoA hydratase
determined by transferred nuclear Overhauser effect measurements:
Mechanistic predictions based on the X-ray structure of 4-
(chlorobenzoyl)-CoA dehalogenase. Biochemistry 36, 2211−2220.
(54) Martin, D. P., Bibart, R. T., and Drueckhammer, D. G. (1994)
Synthesis of novel analogs of acetyl coenzyme A: Mimics of enzyme
reaction intermediates. J. Am. Chem. Soc. 116, 4660−4668.
(55) Burgi, H. B., Dunitz, J. D., and Shefter, E. (1973) Geometrical
̈
reaction coordinates. II. Nucleophilic addition to a carbonyl group. J.
Am. Chem. Soc. 95, 5065−5067.
(56) Allen, S. H. G., Kellermeyer, R. W., Stjernholm, R. L., and
Wood, H. G. (1964) Purification and properties of enzymes involved
in the propionic acid fermentation. J. Bacteriol. 87, 171−187.
(57) Sohling, B., and Gottschalk, G. (1996) Molecular analysis of the
̈
anaerobic succinate degradation pathway in Clostridium kluyveri. J.
Bacteriol. 178, 871−880.
(58) Scherf, U., and Buckel, W. (1991) Purification and properties of
4-hydroxybutyrate coenzyme A transferase from Clostridium amino-
butyricum. Appl. Environ. Microbiol. 57, 2699−2702.
(59) Macieira, S., Zhang, J., Velarde, M., Buckel, W., and
Messerschmidt, A. (2009) Crystal structure of 4-hydroxybutyrate
CoA-transferase from Clostridium aminobutyricum. Biol. Chem. 390,
1251−1263.
(60) Mack, M., and Buckel, W. (1997) Conversion of glutaconate
CoA-transferase from Acidaminococcus fermentans into an acyl-CoA
hydrolase by site-directed mutagenesis. FEBS Lett. 405, 209−212.
(61) Hammes, G. G. (2002) Multiple conformational changes in
enzyme catalysis. Biochemistry 41, 8221−8228.
(62) Koshland, D. E., Jr. (1960) The active site and enzyme action.
Adv. Enzymol. Relat. Subj. Biochem. 22, 45−97.
(63) Gandour, R. D. (1981) On the importance of orientation in
general base catalysis by carboxylate. Bioorg. Chem. 10, 169−176.
(64) Chan, K. K., Wood, B. M., Fedorov, A. A., Fedorov, E. V., Imker,
H. J., Amyes, T. L., Richard, J. P., Almo, S. C., and Gerlt, J. A. (2009)
Mechanism of the orotidine 5′-monophosphate decarboxylase-
catalyzed reaction: Evidence for substrate destabilization. Biochemistry
48, 5518−5531.
(65) Bruice, T. C. (2002) A view at the millennium: The efficiency of
enzymatic catalysis. Acc. Chem. Res. 35, 139−148.
8434
dx.doi.org/10.1021/bi300957f | Biochemistry 2012, 51, 8422−8434