Kinetics of CO insertion and acetyl group transfer steps, and a model of the acetyl-CoA synthase catalytic mechanism

Article abstract of DOI:10.1021/ja0627702

Acetyl-CoA synthase/carbon monoxide dehydrogenase is a Ni-Fe-S-containing enzyme that catalyzes the synthesis of acetyl-CoA from CO, CoA, and a methyl group. The methyl group is transferred onto the enzyme from a corrinoid-iron-sulfur protein (CoFeSP). The kinetics of two steps within the catalytic mechanism were studied using the stopped-flow method, including the insertion of CO into a putative Ni²⁺-CH₃ bond and the transfer of the resulting acetyl group to CoA. Neither step had been studied previously. Reactions were monitored indirectly, starting with the methylated intermediate form of the enzyme. Resulting traces were analyzed by constructing a simple kinetic model describing the catalytic mechanism under reducing conditions. Besides methyl group transfer, CO insertion, and acetyl group transfer, fitting to experimental traces required the inclusion of an inhibitory step in which CO reversibly bound to the form of the enzyme obtained immediately after product release. Global simulation of the reported datasets afforded a consistent set of kinetic parameters. The equilibrium constant for the overall synthesis of acetyl-CoA was estimated and compared to the product of the individual equilibrium constants. Simulations obtained with the model duplicated the essential behavior of the enzyme, in terms of the variation of activity with [CO], and the time-dependent decay of the NiFeC EPR signal upon reaction with CoFeSP. Under standard assay conditions, the model suggests that the vast majority of active enzyme molecules in a population should be in the methylated form, suggesting that the subsequent catalytic step, namely CO insertion, is rate limiting. This conclusion is further supported by a sensitivity analysis showing that the rate is most sensitively affected by a change in the rate coefficient associated with the CO insertion step.

Full text of DOI:10.1021/ja0627702

Published on Web 08/26/2006
Kinetics of CO Insertion and Acetyl Group Transfer Steps,
and a Model of the Acetyl-CoA Synthase Catalytic Mechanism
Xiangshi Tan, Ivan V. Surovtsev, and Paul A. Lindahl*
Contribution from the Departments of Chemistry and of Biochemistry and Biophysics,
Texas A&M UniVersity, College Station, Texas 77843
Received May 10, 2006; E-mail: lindahl@mail.chem.tamu.edu
Abstract: Acetyl-CoA synthase/carbon monoxide dehydrogenase is a Ni-Fe-S-containing enzyme that
catalyzes the synthesis of acetyl-CoA from CO, CoA, and a methyl group. The methyl group is transferred
onto the enzyme from a corrinoid-iron-sulfur protein (CoFeSP). The kinetics of two steps within the catalytic
mechanism were studied using the stopped-flow method, including the insertion of CO into a putative Ni2+
-
CH3 bond and the transfer of the resulting acetyl group to CoA. Neither step had been studied previously.
Reactions were monitored indirectly, starting with the methylated intermediate form of the enzyme. Resulting
traces were analyzed by constructing a simple kinetic model describing the catalytic mechanism under
reducing conditions. Besides methyl group transfer, CO insertion, and acetyl group transfer, fitting to
experimental traces required the inclusion of an inhibitory step in which CO reversibly bound to the form
of the enzyme obtained immediately after product release. Global simulation of the reported datasets afforded
a consistent set of kinetic parameters. The equilibrium constant for the overall synthesis of acetyl-CoA
was estimated and compared to the product of the individual equilibrium constants. Simulations obtained
with the model duplicated the essential behavior of the enzyme, in terms of the variation of activity with
[CO], and the time-dependent decay of the NiFeC EPR signal upon reaction with CoFeSP. Under standard
assay conditions, the model suggests that the vast majority of active enzyme molecules in a population
should be in the methylated form, suggesting that the subsequent catalytic step, namely CO insertion, is
rate limiting. This conclusion is further supported by a sensitivity analysis showing that the rate is most
sensitively affected by a change in the rate coefficient associated with the CO insertion step.
Introduction
of an [Fe4S4] cubane bridged via a cysteine residue to a Ni ion
called proximal Nip. This Ni is also bridged (via two other
Acetyl-CoA synthases/carbon monoxide dehydrogenases are
found in homoacetogenic bacteria, methanogenic archaea, and
CO-utilizing hydrogenogenic bacteria.1-4 These O2-sensitive
bifunctional enzymes allow such organisms to grow chemo-
autotrophically on simple inorganic compounds. The enzyme
from the homoacetogen Moorella thermoacetica (ACS/CODH)
has been studied most extensively.5,6 The â subunits of this 310
kDa R2â2 tetramer catalyze the reversible reduction of CO2 to
CO, while the R subunits catalyze the synthesis of acetyl-CoA
from CO, CoA, and a methyl group donated from a corrinoid-
iron-sulfur protein (CoFeSP). At a buffered pH, this is
represented by reaction 1.
cysteines) to a second Ni ion (distal Nid); thus, Nip is coordinated
to three bridging thiolates. Nid has an N2S2 square-planar
environment including coordination to two amide nitrogens
derived from the protein backbone.7-9 Evidence suggests that
CO and methyl groups bind to Nip during catalysis.10,11
Although aspects of the ACS catalytic mechanism remain
uncertain, our understanding of it is gradually improving. The
components of the A cluster in its most oxidized redox state
2+
2+
(called Aox) appear to be in the {[Fe4S4]2+ Nip Nid
}
electronic configuration.12,13 This state is inactive for both
(5) Ragsdale, S. W. Crit. ReV. Biochem. Mol. 2004, 39, 165-195.
(6) Drennan, C. L.; Doukov, T. I.; Ragsdale, S. W. J. Biol. Inorg. Chem. 2004,
9, 511-515.
(7) Doukov, T. I.; Iverson, T. M.; Seravalli, J.; Ragsdale, S. W.; Drennan, C.
L. Science 2002, 298, 567-572.
(8) Darnault, C.; Volbeda, A.; Kim, E. J.; Legrand, P.; Verne`de, X.; Lindahl,
P. A.; Fontecilla-Camps, J. C. Nat. Struct. Biol. 2003, 10, 271-279.
(9) Svetlitchnyi, V.; Dobbek, H.; Meyer-Klaucke, W.; Meins, T.; Thiele, B.;
Romer, P.; Huber, R.; Meyer, O. Proc. Natl. Acad. Sci. U.S.A. 2004, 101,
446-451.
CO + CoA + CH3-Co3+FeSP a
CH3-C(O)-CoA + Co1+FeSP KACS (1)
The active site for this reaction, called the A cluster, consists
(10) (a) Ragsdale, S. W.; Ljungdahl, L. G.; Dervartanian, D.V. Biochem. Biophys.
Res. Commun. 1982, 108, 658-663. (b) Ragsdale, S. W.; Ljungdahl, L.
G.; Dervartanian, D. V. Biochem. Biophys. Res. Commun. 1983, 115, 658-
665. (c) Ragsdale, S. W.; Wood, H. G.; Antholine, W. E. Proc. Natl. Acad.
Sci. U.S.A. 1985, 82, 6811-6814.
(1) Wood, H.G.; Ljungdahl, L.G. Variations in Autotrophic Life; Academic
Press: London, 1992; pp 201-250.
(2) Grahame, D.A. Trends Biochem. Sci. 2003, 28, 221-224.
(3) Lindahl, P. A.; Chang, B. Orig. Life EVol. Biosph. 2001, 31, 403-434.
(4) Lindahl, P. A.; Graham, D. E. In Nickel and Its Surprising Impact in
Nature”; Sigel, A., Sigel, H., Sigel, R. K. O., Eds.; Metal Ions in Life
Science, Vol. 2; John Wiley & Sons, Ltd.: Chichester, UK, 2007; Chapter
9. In press.
(11) Barondeau, D. P.; Lindahl, P. A. J. Am. Chem. Soc. 1997, 119, 3959-
3970.
(12) Xia, J. Q.; Hu, Z. G.; Popescu, C. V.; Lindahl, P. A.; Mu¨nck, E. J. Am.
Chem. Soc. 1997, 119, 8301-8312.
9
10.1021/ja0627702 CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 12331-12338
12331

A R T I C L E S
Tan et al.
catalysis and methyl group transfer, but it can be activated by
a 2-electron reduction corresponding to an apparent midpoint
potential of ca. -540 mV vs NHE at neutral pH.14 The resulting
reductively activated state apparently has the cubane and Nid
Co1+FeSP absorbs. Under these conditions, the reverse of
reaction 2 (i.e. starting from Ni2+-CH3 and Co1+) could barely
be detected, indicating that the equilibrium position lies on the
side of the products.21 In contrast, the reverse reaction proceeded
rapidly and to near completion when the Ni2+-CH3 was not
preincubated with Ti3+citrate. Under these conditions, the
equilibrium position appears to be on the side of the reactants
of reaction 2. This reductant-dependent shift in the kinetic and
thermodynamic properties of the methyl group transfer reaction
is not understood mechanistically.
Bhaskar et al. have examined the steady-state kinetics of the
exchange reaction between acetyl-CoA and dephospho-CoA as
catalyzed by the ACS/CODH homologue from Methanosarcina
barkeri.22 Their results indicate a ping-pong mechanism in
which the binding of acetyl-CoA to the enzyme is followed by
the release of CoA and formation of the acetyl intermediate.
They proposed that acetyl-CoA binds to the oxidized form of
the enzyme, followed by reduction. This was suggested because
partially reduced enzyme exhibited cooperative binding with
acetyl-CoA, whereas fully reduced enzyme showed simple
hyperbolic binding. However, the same behavior would be
observed if enzyme were first reduced and then bound with
acetyl-CoA. This latter scenario would be congruent with a
nucleophilic attack (e.g., by a Ni0 species) on the carbonyl of
acetyl-CoA, as in the reverse of reaction 4. The alternative
proposal of binding followed by reduction would seem to require
attack by Ni2+, a non-nucleophilic metal ion. Using two
methods, Bhaskar et al. measured the equilibrium constant for
the reverse of reaction 4 to be ∼0.2 (averaged value),22
suggesting KCoA ≈ 5 for a homologous ACS/CODH from a
methanogenic archaeon.
in the 2+ valence, suggesting the unprecedented {[Fe4S4]2+
-
Nip Nid2+} configuration.12-14 Although the occurrence of a
zero-valent Ni atom in the reductively activated state is not
established, we will use this nomenclature throughout in this
paper, for convenience if for no other reason. For a full
discussion of this issue, readers are referred to the literature.4,13-15
If preferred, “Ni0” can be viewed simply as an electron-counting
formalism indicating the Aox state to which 2e- have been
added. This formal view does not complicate or bias any
interpretation, analysis, or conclusion presented here.
0
Whether the methyl group or CO binds first to the enzyme
remains contentious, and reasonable arguments have been made
for both cases.5,6,11,16 The one-electron-reduced and CO-bound
1
state of the A cluster (the S ) /2 Ared-CO state) has been
proposed to be an intermediate of catalysis as well as an
inhibitory state. Recent evidence that ACS/CODH need not pass
through the Ared-CO state during catalysis, and evidence that
reductive activation requires 2 electrons14 compels us to favor
the case where the methyl group binds first. It is also known
that the reductively activated state accepts a methyl group in
the absence of CO, as shown in reaction 2.
Ni0 + CH3-CO3+FeSP a
k+met
Ni2+-CH3 + Co1+FeSP Kmet
)
(2)
k-met
The resulting methylated state is stable and has been character-
ized.11,14,17,18 When exposed to CO, e.g. during catalysis, CO
is thought to insert into the Ni-methyl bond in accordance with
reaction 3.
To date, no direct studies of the CO insertion, reaction 3,
have been reported, nor have the kinetics of the reductive
elimination of the acetyl group and CoA, reaction 4, been
reported. The problem in studying these reactions has been to
identify strategies for monitoring them. Using stopped-flow
kinetics, we report here that reactions 3 and 4 can be monitored
by starting with the methylated state of ACS/CODH. Resulting
traces were used to construct a simple kinetic model describing
the catalytic mechanism of acetyl-CoA synthase. In this paper
we report these results and describe the model.
k+ins
Ni2+-CH3 + CO a Ni2+-C(O)CH3 Kins
)
(3)
k-ins
Reaction of the acetyl intermediate with CoA affords acetyl-
CoA and regenerates the reductively activated state, reaction
4.
Ni2+-C(O)CH3 + CoA a
Experimental Procedures
Preparation of Proteins. M. thermoacetica cells were grown and
harvested as described previously.23 ACS/CODH, CoFeSP, and methyl
transferase were purified in a glovebox containing <1 ppm O2.11,24
Protein concentrations were determined as described previously.25 Each
protein was >90% pure, as quantified by imaging Coomassie blue (Bio-
Rad)-stained SDS-PAGE gels (Alpha Innotech Imager 2000). Portions
were thawed as needed, subjected to a 1 cm × 20 cm column of
Sephadex G25 equilibrated in 50 mM Tris (pH 8.0) + 1 mM
dithiothreitol (DTT), divided into aliquots, and either used immediately
k+CoA
CH3C(O)-CoA + Ni0 KCoA
)
(4)
k-CoA
Reactions 2-4 complete the catalytic cycle for the synthesis of
acetyl-CoA.
Of these steps, only methyl group transfer, reaction 2, has
been studied specifically.19-21 In a stopped-flow study, ACS/
CODH and Ti3+citrate were preincubated to generate the Ni0
state, and then reacted against CH3-Co3+FeSP (also preincu-
bated in Ti3+citrate) and monitored at 390 nm where the product
(19) Zhao, S.; Roberts, D. L.; Ragsdale, S.W. Biochemistry 1995, 34, 15075-
15083.
(20) Kumar, M.; Qiu, D.; Spiro, T. G.; Ragsdale, S. W. Science 1995, 270,
(13) Lindahl, P. A. J. Biol. Inorg. Chem. 2004, 9, 516-524.
(14) Bramlett, M. R.; Stubna, A.; Tan, X.; Surovtsev, I. V.; Mu¨nck, E.; Lindahl,
P. A. Biochemistry 2006, .
(15) Amara, P.; Volbeda, A.; Fontecilla-Camps, J. C.; Field, M. J. J. Am. Chem.
Soc. 2005, 127, 2776-2784.
(16) Seravalli, J.; Kumar, M.; Ragsdale, S. W. Biochemistry, 2002, 41, 1807-
628-630.
(21) Tan, X.; Sewell, C.; Lindahl, P. A. J. Am. Chem. Soc. 2002, 124, 6277-
6284.
(22) Bhaskar B.; DeMoll, E.; Grahame, D. A. Biochemistry 1998, 37, 14491-
14499.
(23) Lundie, L. L., Jr.; Drake, H. L. J. Bacteriol. 1984, 159, 700-703.
1819.
(24) Maynard, E. L.; Sewell, C.; Lindahl, P. A. J. Am. Chem. Soc. 2001, 123,
(17) Pezacka, E.; Wood, H. G. J. Biol. Chem. 1988, 263, 16000-16006.
(18) Lu, W.-P.; Harder, S. R.; Ragsdale, S. W. J. Biol. Chem. 1990, 265, 3124-
3133.
4697-4703.
(25) Pelley, J. W.; Garner, C. W.; Little, G. H. Anal. Biochem. 1978, 86, 341-
343.
9
12332 J. AM. CHEM. SOC. VOL. 128, NO. 37, 2006

Model of ACS Catalysis
A R T I C L E S
or frozen in liquid N2. Acetyl-CoA synthase and CO oxidation activities
were assayed as described previously.24 Samples had specific activities
of 1.2 µmol min-1 mg-1 and 400 units mg-1, respectively. EPR of
ACS/CODH in the presence of 1 atm CO exhibited the NiFeC signal
with a spin intensity corresponding to 0.2 spins/Râ, as determined
elsewhere.26 Ti3+citrate solutions were prepared from TiCl3 (Aldrich)
and standardized by titration against K3[Fe(CN)6].27 Co1+FeSP was
methylated using CH3-tetrahydrofolate (THF) (Sigma) and methyl-
transferase.11 ACS/CODH was methylated (forming CH3-ACS/CODH)
and then separated using a phenyl sepharose column.21 In a previous
equivalent experiment,21 CH3-Co3+FeSP contained 0.95 methyl groups
per protein, and ACS/CODH was methylated to the extent of 0.5 CH3/
Râ.
Figure 1. Kinetic model describing the mechanism by which ACS/CODH
CO Insertion Reaction. A stopped-flow kinetic experiment was
performed at 25 °C with an SF-61 DX2 double-mixing instrument and
a Xe lamp (Hi-Tech Limited, UK) installed in an Ar-atmosphere
glovebox (MBraun) containing <1 ppm O2. For all experiments in this
study, reactions were monitored in PM mode at 390 nm (and, in some
cases, at 450 nm). A 3-L Ar-filled flask with a 2-mL conical vial fused
at the bottom was sealed with a rubber septum that had been stored in
the glove box for >1 wk. An aliquot (1.2 mL) of a 20 µM solution of
CH3-ACS/CODH in Buffer A (50 mM Tris pH 8 + 100 µM DTT)
was transferred to a vial. Different volumes of oxygen-scrubbed
(Oxysorb cartridge, MG industries) research-grade CO (MG Industries)
were injected into the vessel by syringe, affording [CO] of 0, 0.1, 0.5,
1.0, 2.0, and 5.0 µM in each respective flask. The solution was stirred
slowly for 4 h, and then transferred to a stopped-flow syringe. The
other syringe contained 10 µM of Co1+FeSP and 200 µM of Ti3+citrate
in a solution of Buffer A. The experiments were repeated with CO (20
µM) included in the CoFeSP syringe rather than in the ACS/CODH
syringe.
Acetyl Group Transfer Reaction. CH3-ACS/CODH (20 µM in
Buffer A) was preincubated for 2 h in the 3-L Ar-filled flask, mentioned
above, containing either 20 or 40 µM CO and then was transferred to
a stopped-flow syringe. The other syringe contained 10 µM of CH3-
Co3+FeSP and coenzyme A at different concentrations (10, or 100, or
1000 µM, respectively) in a solution of Buffer A.
Acetyl-CoA Synthesis Reaction. A 380-µL solution of Buffer A
containing CoA (1.4 mM), CH3-Co3+FeSP (100 µM), Ti3+citrate (200
µM), CO (100 µM) (all final concentrations) were added to a reaction
vessel. The vessel was sealed with a rubber septum, removed from the
glovebox, flushed for 15 min with a mixture of Ar and CO (100 µM
CO), and returned to the box. To prepare these gases, Ar was passed
through an Oxysorb (MG Industries) filter, mixed with equivalently
treated CO using a calibrated flow meter (MG Industries, series 7941-
AS2 4 tube), and passed into the vessel. The resulting solution was
incubated for 15 min at 25 °C in dim light. Acetyl-CoA synthesis was
initiated by addition of 20 µL of an ACS/CODH solution (0.2 µM Râ,
final). Aliquots (50 µL) were removed by syringe at various times,
quenched, and analyzed for acetyl-CoA by reversed phase HPLC.24
Initial rates were determined by linear least-squares regression analysis
of [acetyl-CoA] vs time plots.
catalyzes the synthesis of Acetyl-CoA.
∆ꢀCO were +0.0100 and -0.0015 µM-1 cm-1, respectively, which were
adjusted slightly from experimentally reported values.21 Experimental
traces of A390 were simulated using the expression
A390(t) ) A390(0) + ∆ꢀmet‚Xmet(t) + ∆ꢀCO‚XCO(t)
where Xmet and XCO represent the extent of reactions 2 and 5 at any
time.
Reaction conditions used were common for all experiments (namely
200 µM Ti3+citrate, 50 mM buffer (Tris pH 8), 100 µM DTT occurring
at 25 °C), allowing a single consistent set of simulations to be obtained.
All rate coefficients and equilibrium constants reported are apparent
values specifically for these conditions. Concentrations used during the
fitting procedure were allowed to differ by (10% relative to experi-
mentally determined values (except for [CO] which was allowed to
vary (25%). This flexibility corresponds to the relative uncertainty
associated with the experimental determinations. ACS/CODH has been
reported to be heterogeneous, with active fractions varying between
20 to 50%, depending upon the property measured.26,28 Since 0.5 methyl
groups have been reported to be accepted per ACS/CODH,11,21 we have
assumed this value for all simulations. The [ACS/CODH] given for
the experiments described in the text refer to experimentally determined
concentrations of the Râ dimer, assuming a molecular mass of 154
kDa, whereas [ACS/CODH] concentrations assumed for simulations
were half of the experimental values.
Results
While analyzing the experiments described below, we
developed a simple but comprehensive kinetic model that
describes how ACS/CODH catalyzes the synthesis of acetyl-
CoA from CO, CoA, and CH3-Co3+FeSP. This model, il-
lustrated in Figure 1, consists of reactions 2, 3, 4 and an
inhibition step, reaction 5.
k+CO
Ni0 + CO a Ni:CO
KCO
)
(5)
k-CO
Kinetic Model, Simulations, and Fitting. For the model described
in the text, a set of ordinary differential equations (ODEs) was generated
assuming the Law of Mass Action. Once initial concentrations for all
species and rate coefficients associated with these reactions were
specified, the ODEs were numerically solved using a 4th-order Runge-
Kutta method to yield a set of component concentrations at any time.
Data were fit using Origin (version 7.5) software. Extinction coefficient
changes at 390 nm for reactions 2 and 5 (see below) were given by
∆ꢀmet ) (ꢀCo1+ + ꢀCH3-Ni - ꢀCH3-Co - ꢀNi0) and ∆ꢀCO ) (ꢀNi:CO - ꢀNi0),
respectively, while extinction coefficient changes for reactions 3 and
4 were presumed to be negligible. The values assigned to ∆ꢀmet and
The nomenclature Ni:CO leaves ambiguous the oxidation state
of the proximal Ni in the inhibited form and the issue of whether
CO is bound directly to the Ni in that form.
Methyl Group Transfer Reaction. Although the kinetics
of methyl group transfer have been evaluated previously,21 we
re-evaluated this reaction for the particular batch of enzyme
used. A solution of ACS/CODH and Ti3+citrate was placed in
one syringe and a solution of CH3-Co3+FeSP and Ti3+citrate
was place in the other. Upon mixing, changes in the absorbance
at 390 nm were monitored vs time, and an analytical expression
(26) Shin, W. S.; Anderson, M. E.; Lindahl, P. A. J. Am. Chem. Soc. 1993,
115, 5522-5526.
(27) Seefeldt, L. C.; Ensign, S. A. Anal. Biochem. 1994, 221, 379-386.
(28) Fraser, D. M.; Lindahl, P. A. Biochemistry 1999, 38, 15697-15705.
9
J. AM. CHEM. SOC. VOL. 128, NO. 37, 2006 12333

A R T I C L E S
Tan et al.
for bimolecular reaction 2 was used to fit the resulting
experimental traces (data not shown). The best-fit forward rate
coefficient (k+met )15 µM-1 s-1) was within error of that
reported previously, while the reverse rate coefficient (k-met
)
0.05 µM-1 s-1) was ∼30 times less than the previous estimate.21
The ratio of these two rate coefficients affords the equilibrium
constant Kmet ≈ 300. Similar to the previous study, the reverse
reaction proceeded rapidly and extensively as long as the Ni2+
-
CH3 state was not preincubated with Ti3+citrate. In this case,
the kinetic parameters were k+met ) 0.13 µM-1 s-1 and k-met
) 1.8 µM-1 s-1 affording Kmet ≈ 0.07.29
CO Insertion Reaction. Our next objective was to determine
the equilibrium constant for reaction 3, in which CO inserts
into CH3-ACS/CODH. A solution of CH3-ACS/CODH was
preincubated with CO in one syringe and then reacted with a
solution of Co1+FeSP while monitoring the absorbance at 390
nm. Depending on the [CO] and the magnitude of Kins, the
syringe that initially contained CO and CH3-ACS/CODH
should, by the time it is mixed with Co1+FeSP, contain
equilibrium proportions of CH3-ACS/CODH and CH3C(O)-
ACS/CODH. Since only the CH3-ACS/CODH form can donate
a methyl group to Co1+FeSP, the initial rate of reaction will
reflect the proportion of ACS/CODH in this form. In the absence
of CO, all of the enzyme will be in the CH3-ACS/CODH form,
and the rate of reverse methyl transfer (and requisite decline at
390 nm) will be maximal. At infinite [CO], all of the enzyme
will be in the CH3C(O)-ACS/CODH form, and no methyl
transfer will be observed.
Preliminary experiments revealed that even at concentrations
of CO comparable to those of ACS/CODH used, the CO
insertion reaction product was almost quantitatively formed.
Thus, the experiment was repeated in a large vessel such that
the final concentration of free CO was virtually identical to the
added concentration of CO, regardless of the amount of CO
that associated with the enzyme. The results (Figure 2A)
displayed the expected behavior. In Panel B, initial rates υ0 of
this reaction are plotted vs [CO]. These data-points were
determined from the expression υ0 ) ∆A390/(∆t‚∆ꢀmet) where
∆A390 is the initial absorbance change and ∆t is the initial time
interval. For the rates shown in Figure 2B, ∆t corresponds to
the first 0.05 s of reaction. The solid line in Figure 2B is the
best-fit of expression 6 to the data,
Figure 2. Stopped-flow reaction of CH3-ACS/CODH with CO and Ti3+
-
citrate-reduced Co1+FeSP. (A) CO was pre-equilibrated in the CH3-ACS/
CODH syringe. (O) Experimental traces of A390 vs time using indicated
[CO]. (s) Simulations using the model of Figure 1 and the kinetic
parameters given in Table 1 except k+met ) 0.13 µM-1 s-1 and k-met ) 1.8
µM-1 s-1. (B) (O) Initial rates obtained as described in the text. (s) Best-
fit of eq 6, obtained using KD(CO) ) 0.31 µM. Error bars reflect the estimated
uncertainties for each point. (C) CO initially in the Co1+FeSP syringe. (O)
Experimental traces of A390 vs time. (s) Simulations with kinetic parameters
as in A, except k+ins values of 100, 1, and 0.1 µM-1 s-1
.
pre-equilibration solution. Derivation of eq 6 does not include
reactions 4 and 5 since no CoA was present in the experiment
and the concentration of CO used was very low. It assumes
that reaction 3 had reached equilibrium prior to reaction with
Co1+FeSP.
Our next objective was to explore the kinetics of reaction 3.
To do this, CH3-ACS/CODH was included in one stopped-
flow syringe, while Co1+FeSP and CO were included in the
other. If CO inserted at an infinitely fast rate relative to that of
the reverse methyl transfer reaction (reaction 2, right to left),
the resulting acetyl form would not react with Co1+FeSP, and
no absorbance decline at 390 nm would be observed. On the
other extreme, if CO inserted at an infinitely slow rate, CH3-
ACS/CODH would react with Co1+FeSP, and A390 would
decline rapidly. As shown in Figure 2C, A390 declined at a rate
similar to that observed in the absence of CO. These traces were
simulated assuming the model of Figure 1, the set of experi-
mentally determined concentrations, and the k+met(no red), k-met(no
red) and KD(ins) values determined above, as well as the other
parameters determined below and listed in Table 1. Numerous
simulations were generated using various trial values for all other
required parameters. The resulting simulations (solid lines
nearest to the experimental traces) were obtained using the set
of kinetic parameters described below. The other solid lines in
Figure 2C were generated using the same parameters except
for the indicated values of k+ins. The lack of fidelity to the data
indicates that the rate of CO insertion is relatively slow under
the conditions of this experiment.
KD(ins)
3 0 [CO] + KD(ins)
ν0 ) k-met[CO1+FeSP]0[Ni2+-CH ]
(6)
affording the dissociation constant for the CO insertion reaction
3, KD(ins) ) 1/Kins ) 0.31 ( 0.11 µM. [Co1+FeSP]0 and [Ni2+
-
CH3]0 refer to the active concentrations of these proteins initially
present after mixing, while [CO] is concentration of CO in the
(29) We previously estimated a value of 2.3 for the equilibrium constant
associated with methyl group transfer.21 Although we mentioned in that
report that forward and reverse rate constants depended on the presence/
absence of reductant, we calculated the equilibrium constant using the
forward rate constant as determined in the presence of reductant and the
reVerse rate constant as determined in the absence of the reductant. In
hindsight, a more reasonable approach, which we take here, would have
been to report two Kmet values, one relevant for reactions performed in the
presence of a reductant, one for reactions performed in the absence of a
reductant. Our calculated respective values, Kmet ≈ 300 and Kmet(no
≈
red)
0.07, indicate that, in the presence of reductant, the methyl group transfer
reaction at equilibrium lies far to the side of the products, whereas in the
absence of a reductant it lies far to the side of the reactants. The kinetic
model developed here reflects the situation in the presence of reductant, as
is typical of ACS/CODH catalytic reaction conditions.
9
12334 J. AM. CHEM. SOC. VOL. 128, NO. 37, 2006

Model of ACS Catalysis
A R T I C L E S
Table 1. Kinetic Parameters used in Fitting
k+met ) 15.0 ( 1.8 µM-1 s-1
k-met e 0.05 µM-1 s-1
K+met g 300
k+ins ) 0.11 ( 0.08 µM-1 s-1
k-ins ) 0.033 s-1
K-ins ) 3.3 ( 1.1 µM-1
KCoA ) 0.66
k+CoA ) 3.8 ( 1.0 µM-1 s-1
k-CoA ) 5.8 ( 2.7 µM-1 s-1
k+CO ) 12.5 ( 7.8 µM-1 s-1
k-CO ) 64.8 ( 9.4 s-1
KCO ) 0.19 µM-1
Acetyl Group Transfer Reaction. To monitor reaction 4, a
solution containing Ni2+-CH3 and CO was placed in one
stopped-flow syringe, while a solution of CH3-Co3+FeSP and
CoA was placed in the other. The concentration of CO in the
syringe was in all cases sufficient for ∼99% of ACS/CODH to
be acetylated given KD(ins) ) 0.31 µM. The concentration of
CoA was varied from 10 to 1000 µM. Reactions were monitored
at 390 nm (Figure 3) and at 450 nm (data not shown); equivalent
results were obtained. At low [CoA] (Figure 3C), the reaction
of CoA with the acetyl intermediate was relatively slow such
that a small amount of Ni0 was generated soon after mixing.
Consequently, the rate by which Ni0 methylated (as monitored
by the increase at A390) was slow. At the highest [CoA] used,
the reaction of CoA with the acetyl intermediate was sufficiently
fast such that a larger proportion of the enzyme was converted
into the Ni0 state, and this large proportion could be methylated,
resulting in a more rapid rise of A390 (Figure 3A). The rate at
which A390 increased for the experiment with an intermediate
[CoA] (Figure 3B) was intermediate of that observed in traces
A and C.
Figure 3. Stopped-flow reaction of the acetylated ACS/CODH intermediate
with CoA and subsequent reaction of the reductively activated state with
CH3-Co3+FeSP. (O), experimental data. In A, B, and C, the [CO] was 20
µM and the [CoA] was 1000, 100, and 10 µM, respectively. In (D) [CO]
was 40 µM, and [CoA] was 1000 µM. (s) Simulations using the model of
Figure 1 with kinetic parameters listed in Table 1.
The experimental traces exhibited two distinctive features.
First, the extent of reaction was similar in all three experiments,
suggesting that the equilibrium position of the ACS reaction
lies on the product side. Second, the rate of the apparent reaction
for methyl group transfer depends weakly on [CoA], in that the
order-of-magnitude changes in [CoA] only altered the rates of
reaction by factors of ∼2.
Figure 4. Acetyl-CoA synthesis. (O), Acetyl-CoA; (s) Simulation using
the model of Figure 1, kinetic parameters listed in Table 1, and experimental
concentrations. Error bars reflect the estimated uncertainties for each point.
Data collected for the acetyl group transfer reaction were
initially analyzed using a minimal kinetic scheme consisting
solely of reactions 2 and 4. Reaction 3 was neglected because
it should have reached equilibrium before the content of the
two syringes were mixed. This scheme failed to reproduce all
experimental traces simultaneously using a common set of
kinetic parameters (it allowed each trace to be fitted separately
and with a unique set of parameters). Even with reaction 3
included, fits were unacceptable. The model was modified by
assuming that CoA inhibited the reaction by binding to the Ni0
form, but fits were unimproved (data not shown). Reaction 4
was decomposed into two to three separate steps (reflecting
separate binding, reaction, and/or dissociation steps); this
improved fits marginally but not to our satisfaction. Finally,
we augmented reactions 2-4 with a reaction in which CO binds
to the Ni0 form, reaction 5. This model fit well to the data,
yielding the simulations shown as the solid lines in Figure 3A-
C, obtained using the kinetic parameters given in Table 1. Note
that the values for k+met and k-met are those obtained for the
methyl group transfer reaction under reducing conditions.
The solid line of Figure 3D is a simulation using the concentra-
tions for the experiment and the kinetic parameters listed in
Table 1.
Acetyl-CoA Synthesis. The equilibrium constant for the
synthesis of acetyl-CoA, KACS of reaction 1, equals the product
of the equilibrium constants for three constituent reactions as
given by the relationship KACS ) Kmet‚Kins‚KCoA. Our simulations
afforded values for the three equilibrium constants contributing
to this relationship (listed in Table 1), suggesting that KACS is
greater than ∼660 µM-1 (300‚3.3 µM-1‚0.66). We estimated
K
ACS experimentally as a means of testing our model and these
values. The concentration of acetyl-CoA produced at different
times during an acetyl-CoA synthase assay was determined by
HPLC as the reaction was allowed to continue to equilibrium.
As shown in Figure 4, the final amount generated (98.5 µM (
1.5 µM, starting with 100 µM of CH3-Co3+FeSP) indicates
that the reaction proceeds to completion within the uncertainty
of the measurement. The curve was simulated using a model
composed of the bimolecular reaction obtained from reaction 1
after assuming a fixed [CO] ) 100 µM (as was the case during
the experiment). Best-fit kinetic parameters were k+ ) 0.00086
The model and kinetic parameters obtained above predicted
significant inhibition of acetyl-CoA synthesis by CO binding
to enzyme. The inhibitory effect of CO was confirmed in trace
D of Figure 3, where [CO] ) 40 µM (and [CoA] ) 1000 µM).
9
J. AM. CHEM. SOC. VOL. 128, NO. 37, 2006 12335

A R T I C L E S
Tan et al.
µM-1 s-1 and k- ) 0.00039 µM-1 s-1, indicating that KACS
)
0.00086/0.00039/100 ) 0.02 µM-1. A value of KACS ) 0.05
µM-1 was obtained using experimental equilibrium concentra-
tions. Given the uncertainty of these measurements, combined
with the fact that the equilibrium position is far to the side of
the products, we can conclude only that the lower limit of KACS
is ∼0.02 µM-1. Thus, these estimates are entirely consistent
with the value of KACS calculated from the kinetic parameters
reported here.
Discussion
This is the first study to report on the kinetics of the CO
insertion and acetyl group transfer steps occurring within the
ACS catalytic mechanism. This achievement was obtained using
the stopped-flow kinetic method and by beginning with the
methylated intermediate form of the enzyme. Reactions were
monitored indirectly, using the methyl group transfer reaction
to/from the CoFeSP as a “reporter” reaction monitorable at both
390 and 450 nm (we chose to monitor at 390 nm for reasons of
increased sensitivity). In these reactions, ACS/CODH and
CoFeSP were always placed in different stopped-flow syringes,
while CO and CoA were strategically placed in either syringe
such that the reaction of interest would occur in a manner that
could be reported at 390 nm.
The complexity of these arrangements required that the
resulting traces be simulated computationally. In turn, this led
to the construction of a comprehensive kinetic model of catalysis
using chemical rate coefficients. Constructing such models is
uncommon in contrast to the steady-state analysis of an enzyme
mechanism using combinations of such kinetic parameters. The
rarity of microscopic kinetic models is due to the difficulty (often
the impossibility) of monitoring each step of a mechanism and
of globally fitting the kinetic traces obtained. If the number of
unknown parameters is excessive relative to the experimental
data available, such systems will be underdetermined and lack
predictive value. We have minimized this problem by obtaining
data sensitive to each step of catalysis and by restricting the
complexity of the model. Thus, we have not explicitly included
a step in which the R subunit undergoes an open:closed
conformational change30 nor have we taken into account the
presence of the residual catalytic activity which probably reflects
the direct binding of CO in solution to Nip.24,31-32 The kinetic
model developed here reflects the majority activity in which
CO serves as both substrate and inhibitor.
As illustrated in Figure 1, our model consists of four reactions,
including reactions 2-5, the first three of which combine to
afford the overall ACS reaction 1. The parameters obtained by
fitting to the experimental traces reported here indicate that the
equilibrium position of the methylation reaction 2 lies far to
the side of the products. This is consistent with the superior
nucleophilic tendency of Ni0 (or whatever species on which
ACS/CODH is attacking) relative to the Co1+ cobalamin. Given
the complexity inherent in the redox-dependent changes in
kinetic and thermodynamic parameters for the methyl group
transfer reaction, our model is restricted to catalytic processes
Figure 5. Simulation of acetyl-CoA activity under different reaction
conditions. (Upper Panel) [ACS/CODH] ) 0.5 µM (active); [CH3-Co3+
-
FeSP] ) 30 µM; [CoA] ) 1000 µM. Simulations were obtained assuming
these concentrations, the model of Figure 1 and the kinetic parameters of
Table 1, except for KD(CO) which was set to 2.8 µM. Velocities were
determined at 0.3 s reaction time. (Lower Panel) Simulated plot of the
fraction of ACS/CODH in the Ni:CO form vs time, using the parameters
of Table 1 without exceptions. (A) [ACS/CODH] ) 25 µM (active); [CH3-
Co3+FeSP] ) 30 µM; [CoA] ) 500 µM; [CO] ) 120 µM. (B) Same as A
except [ACS/CODH] ) 2.5 µM (active).
under reducing conditions. CO is involved in two reactions,
namely as a substrate in the CO insertion step 3 and as an
inhibitor in the CO binding step 5. According to these values,
when CO inserts, it binds over 10-fold more tightly than when
it inhibits. The equilibrium position of the acetyl-transfer
reaction 4 lies slightly to the side of the reactants, in contrast
to the ACS/CODH homologue from M. barkeri, for which the
equilibrium constant is 7 times greater.22
A major advantage of a comprehensive kinetic model of an
enzyme-catalyzed reaction is that it can be used to quantitatively
predict the behavior of the enzyme under any set of substrate,
product, and enzyme concentrations. There are two well-known
properties of ACS/CODH catalysis against which we wanted
to test the model developed here. The variation of ACS activity
with substrate [CO] is not simply the hyperbolic curve predicted
by the Michaelis-Menten equation; in the case of ACS/CODH,
initial velocity divided by total ACS/CODH concentration
increases to ∼100 min-1 as [CO] increases from 0 to ∼100
µM, but then it declines as [CO] increases to 980 µM (1 atm).24
A set of simulations was obtained at different [CO] using our
model. The concentrations of species used for the simulations
were those used experimentally to generate the curve just
described. The kinetic parameters used were those given in Table
1 except for KD(CO) which was set to 2.8 µM rather than 5.2
µM. The plot of the simulated initial rate of product formation
(υ0/[ACS/CODH]tot) vs [CO] revealed behavior reminiscent of
that observed experimentally (Figure 5A, upper panel). This
(30) Tan, X.; Bramlett, M. R.; Lindahl, P. A. J. Am. Chem. Soc. 2004, 126,
5954-5955.
(31) Tan, X.; Loke, H.-K.; Fitch, S.; Lindahl, P. A. J. Am. Chem. Soc. 2005,
127, 5833-5839.
(32) Tan, X.; Volbeda, A.; Fontecilla-Camps, J. C.; Lindahl, P. A. J. Biol. Inorg.
Chem. 2006, 11, 371-378.
9
12336 J. AM. CHEM. SOC. VOL. 128, NO. 37, 2006

Model of ACS Catalysis
A R T I C L E S
Table 2. Sensitivity of the Overall Initial Velocity to Relative
includes a maximal rate near ∼100 min-1 occurring at [CO] ≈
100 µM. The exact magnitude and position at which maximal
activity occurred could be adjusted by changing the value of
Changes in Individual Rate Coefficients
rate coefficient
(∆υ0/υ0)/(∆k/k)
K
D(CO). The steepness in the decline of activity with increasing
k+met
k-met
k+ins
0.339
-0.002
0.657
-0.0004
0.002
0.001
-0.332
0.317
[CO] beyond the optimal [CO] was less than that observed
experimentally, perhaps because CO inhibition may be coopera-
tive,24 an effect not currently included in our simple model.
A second well-established behavior of ACS/CODH catalysis
is that the Ared-CO state (which exhibits the NiFeC EPR
signal10) decays when enzyme prepared in this state (by
incubation in CO) is reacted with CH3-Co3+FeSP.11,16,33 Under
the conditions of their experiment, with CoA included to
complete the catalytic cycle, Seravalli et al. reported that the
signal decays within ∼1 s followed by a plateau region in which
∼33% of the spectral intensity remains over the course of 10
s.16 They took this plateau as evidence that the Ared-CO state
is a catalytic intermediate. The plateau was viewed as resulting
k-ins
k+CoA
k-CoA
k+CO
k-CO
react with Ni0 (in the manner of reaction 4) to form the
acetylated intermediate, Ni2+-*C(O)CH3, which would decar-
bonylate (reaction 3) to yield Ni2+-CH3 and *CO. Radiolabeled
*CO would then exchange with unlabeled CO and reform the
unlabeled product by proceeding in the reverse direction. In
contrast, mechanisms in which CO bound to the enzyme before
the methyl group bound would require that *CO dissociate after
the methyl group was transferred from Ni2+-*C(O)CH3.
However, in the absence of CoFeSP, the methyl group could
not transfer, and thus, no exchange could occur. This argument
would become less unambiguous if a small amount of contami-
nating CoFeSP were included in ACS/CODH preparations
(indeed, we often have a trace of the CoFeSP present in our
preparations). However, in the seminal paper where this
exchange reaction was discovered, Ragsdale and Wood argued
forcefully35 that the only protein contained in their exchange-
active preparations was ACS/CODH. Thus, we regard the
argument presented here as definitive in disqualifying mecha-
nisms in which CO bind ACS/CODH before the methyl group
binds.
Another advantage to developing a microscopic kinetic model
of an enzymatic process is the ability to identify the degree to
which each step in that process controls the overall rate of
reaction. To evaluate this issue, we fixed the reactant and
enzyme concentration at the values commonly used to assay
the enzyme (i.e., [CO] ) 100 µM; [CH3Co3+FeSP] ) 30 µM;
[CoA] ) 1000 µM, [ACS/CODH] ) 0.5 µM (active fraction))
and then calculated the percentage of the enzyme in each
intermediate state. Our results (Ni0, 0.2%; Ni2+-CH3, 95%;
Ni2+-C(O)CH3, 0.3%; and Ni:CO, 4.4%) indicate that the
overwhelming proportion of the enzyme is in the methylated
intermediate state. Next, we calculated the relative change in
reaction velocity with respect to a relative change in the rate
coefficient associated with each step. Our results, given in Table
2, indicate that changes in the rate coefficient associated with
the CO insertion step leads to the greatest relative change in
reaction velocity. Other reactions with a significant impact on
the overall rate include the forward methyl transfer rate and
both forward and reverse rates of CO binding. Since intermedi-
ates tend to accumulate just prior to the rate-determining step
of a multistep process, the accumulation of Ni2+-CH3 and the
sensitivity of k+met in affecting the rate both suggest that the
rate-determining step of ACS catalysis is CO insertion. This
step may in reality include a number of distinguishable substeps,
only one of which needs to be slow. The only requirement is
that these steps occur after the methylated intermediate forms
from a steady-state condition where the rate at which the Ared
-
CO state formed from the previous catalytic intermediate
matched that at which the Ared-CO state converted into the
subsequent intermediate. However, simulations obtained using
our model, where the inhibitory Ni:CO state was presumed to
reflect the Ared-CO state, also reproduced the essential features
of this behavior (Figure 5, lower panel). The duration and
flatness of the plateau region in our simulations could be
adjusted by changing the concentrations of ACS/CODH and
CoFeSP. Simulation A was obtained using the reported experi-
mental conditions, whereas simulation B was obtained by
adjusting the [ACS/CODH] to achieve a plateau region more
characteristic of that reported. According to our model, this
region is never truly flat in CO atmospheres, as it inevitably
curves upward as the reaction reaches equilibrium and as the
Ared-CO state increases in proportion.
The form of the enzyme labeled “Ni:CO” in our model has
not been physically characterized, and thus it may or may not
correspond to the Ared-CO state. Neither the site at which CO
binds nor the oxidation state of the Ni in the Ni:CO form are
established. All that is known about this kinetic state is that it
is generated by adding CO to ACS/COOH in the state present
after acetyl-CoA is released and before a methyl group binds
and that it is an inhibitory state of catalysis (i.e., it does not
lead to products) that competes with the substrate CH3-Co3+
-
FeSP for the same state of the enzyme. Within those constraints,
there are of course many possibilities. However, it is significant
that the Ni:CO state and the Ared-CO state have similar
dissociation constants (KD(CO) ≈ 5 µM for the Ni:CO state and
3 µM for the Ared-CO state34), as would be expected if the
two states were identical.
If the Ni:CO and Ared-CO states were identical, the ability
of our kinetic model to recapitulate the major aspects of ACS/
CODH catalytic behavior would represent yet another argument
against mechanisms in which the Ared-CO state is an interme-
diate of catalysis. During the review of this manuscript, a
reviewer’s comment prompted us to consider the mechanistic
implications of the fact that ACS/CODH, in the absence of any
other protein, catalyzes the exchange of the carbonyl group of
acetyl-CoA with free CO,35 reaction 7.
CH3 C(O)-CoA + CO a CH3C(O)-CoA + *CO
(7)
*
(33) Grahame, D. A.; Khangulov, S.; DeMoll, E. Biochemistry 1996, 35, 593-
600.
(34) Russell, W. K.; Lindahl, P. A. Biochemistry 1998, 37, 10016-10026.
According to the mechanism proposed here, acetyl-CoA would
(35) Ragsdale, S. W.; Wood, H. G. J. Biol. Chem. 1985, 260, 3970-3977.
9
J. AM. CHEM. SOC. VOL. 128, NO. 37, 2006 12337

A R T I C L E S
Tan et al.
and before the acetylated intermediate develops. Substeps might
include: (a) the dissociation of the Co1+FeSP from CH3-ACS/
CODH; (b) a conformational change in the R subunit which
might allow CO to migrate toward Nip through a tunnel; (c)
the binding of CO to the Ni2+-CH3 species just prior to
insertion; and (d) the insertion reaction itself. Efforts are
underway to formulate experiments that could provide evidence
for these substeps and identify which of them is slow, but we
would not be surprised if the slow substep corresponded to either
the dissociation of CoFeSP or the conformational change of
the R subunit (or to a concerted step involving both processes).
Acknowledgment. This work was supported by the National
Institutes of Health (GM46441) and the Robert A. Welch
Foundation (A-1170).
JA0627702
9
12338 J. AM. CHEM. SOC. VOL. 128, NO. 37, 2006