Tyrosine 89 accelerates Co-carbon bond homolysis in methylmalonyl-CoA mutase

Article abstract of DOI:10.1021/ja029420+

The contribution of the active-site residue, Y89, to the trillion-fold acceleration of Co-carbon bond homolysis rate in the methylmalonyl-CoA mutase-catalyzed reaction has been evaluated by site-directed mutagenesis. Conversion of Y89 to phenylalanine or alanine results in a 10³-fold diminution of k_cat and suppression of the overall kinetic isotope effect. The spectrum of the enzyme under steady-state conditions reveals the presence of AdoCbl but no cob(II)alamin. Together, these results are consistent with homolysis becoming completely rate determining in the forward direction in the two mutants and points to the role of Y89 as a molecular wedge in accelerating Co-carbon bond cleavage.

Full text of DOI:10.1021/ja029420+

Published on Web 04/09/2003
Tyrosine 89 Accelerates Co-Carbon Bond Homolysis in
Methylmalonyl-CoA Mutase
Monica D. Vlasie and Ruma Banerjee*
Contribution from the Biochemistry Department, UniVersity of Nebraska,
Lincoln, Nebraska 68588-0664
Received November 20, 2002; E-mail: rbanerjee1@unl.edu
Abstract: The contribution of the active-site residue, Y89, to the trillion-fold acceleration of Co-carbon
bond homolysis rate in the methylmalonyl-CoA mutase-catalyzed reaction has been evaluated by site-
directed mutagenesis. Conversion of Y89 to phenylalanine or alanine results in a 103-fold diminution of kcat
and suppression of the overall kinetic isotope effect. The spectrum of the enzyme under steady-state
conditions reveals the presence of AdoCbl but no cob(II)alamin. Together, these results are consistent
with homolysis becoming completely rate determining in the forward direction in the two mutants and points
to the role of Y89 as a molecular wedge in accelerating Co-carbon bond cleavage.
The cofactor, AdoCbl or coenzyme B12, functions as a radical
generator in enzymes that catalyze 1,2-rearrangement reac-
tions.1,2 A remarkable aspect of the enzyme-catalyzed cleavage
of the Co-carbon bond of the cofactor to generate radicals is
the ca. trillion-fold rate acceleration compared to the uncatalyzed
reaction,3 and the strategies used by these enzymes have been
a subject of ongoing debate. A member of this family of
enzymes, methylmalonyl-CoA mutase, catalyzes the isomer-
ization of methylmalonyl-CoA to succinyl-CoA (Figure 1) and
is found in both bacteria and mammals. The first step common
to all mutases, homolysis of the Co-carbon bond, is observed
upon addition of substrate and leads to the formation of cob-
(II)alamin and the deoxyadenosyl radical (Figure 1, step I). This
is followed by hydrogen atom (H-atom) abstraction from the
substrate, methylmalonyl-CoA, by the deoxyadenosyl radical
to generate a reactive primary substrate radical (step ii).
Rearrangement, by a mechanism that is at present poorly
understood, generates a secondary product-centered radical (step
iii). Finally, a second H-atom transfer from deoxyadenosine
yields product (step iv), and recombination of the cofactor-
centered radical pair (step v) completes a turnover cycle.
Evidence for the intermediacy of radical species in the reaction
has been obtained from stopped-flow spectroscopy which reveals
the presence of cob(II)alamin4,5 and from EPR spectroscopy
which demonstrates the presence of a biradical intermediate
comprising cob(II)alamin and an organic radical.6,7
The wild-type enzyme controls the reactivity of the cofactor
by kinetic coupling of the first two chemical steps: homolysis
(Figure 1, step i) and abstraction of a hydrogen atom from the
Figure 1. Postulated reaction mechanism of methylmalonyl-CoA mutase.
M-CoA and S-CoA denote methylmalonyl-CoA and succinyl-CoA, respec-
tively.
(1) Halpern, J. Science 1985, 227, 869-875.
(2) Banerjee, R. Biochemistry 2001, 40, 6191-8.
(3) Finke, R. G.; Hay, B. P. Inorg. Chem. 1984, 23, 3041-3043.
(4) Padmakumar, R.; Padmakumar, R.; Banerjee, R. Biochemistry 1997, 36,
3713-3718.
substrate (step ii).4 This renders the homolysis step isotope
sensitive, and large deuterium isotope effects have been
measured (49.9 at 5 °C) with [CD3]-labeled substrate, which
indicates the involvement of quantum mechanical tunneling in
this H-atom transfer step.8 X-ray crystallography studies have
(5) Chowdhury, S.; Banerjee, R. Biochemistry 2000, 39, 7998-8006.
(6) Zhao, Y.; Such, P.; Retey, J. Angew. Chem., Int. Ed. Engl. 1992, 31, 215-
216.
(7) Padmakumar, R.; Banerjee, R. J. Biol. Chem. 1995, 270, 9295-9300.
9
10.1021/ja029420+ CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 5431-5435
5431

A R T I C L E S
Vlasie and Banerjee
Figure 2. Movement of Y89 in the active site of methylmalonyl-CoA mutase induced by binding of substrate. (A) Structure of active-site residues in the
absence of substrate taken from PDB file 3REQ with partial density for the adenosine moiety (in red) of AdoCbl. (B) Structure of active-site residues
following binding of substrate (only a fragment up to the thioester bond is shown here) taken from PDB file 4REQ. Significant motion of Y89 is observed.
revealed a large conformational change that accompanies
substrate binding and results in a TIM barrel snapping in around
the CoA tail.9 The active-site residue that undergoes the most
significant motion in the process is Y89, which moves from its
position above and roughly parallel to the adenine ring to one
that destroys the adenine-binding site and establishes new
hydrogen-bonding contacts with the substrate and the cofactor
(Figure 2). This observation has led to the hypothesis that Y89
plays a key role in labilizing the Co-carbon bond by physically
wedging into the space occupied by deoxyadenosine.9
enzyme,11 it was concluded that loss of the hydroxyl group
resulted in an increase in the rearrangement barrier between
substrate and product radicals. However, the effect of this
mutation on the Co-carbon homolysis step was not examined.
In this study, we have examined the role of Y89 in
accelerating the Co-carbon bond homolysis reaction by creating
a conservative mutation, F89, and by creating a cavity at that
position with the A89 mutation. Both mutants have small effects
on the Km for substrate (3- to 7-fold) but profound effects on
the homolysis and the kinetically coupled substrate radical
generation step and exhibit ∼103-fold lower kcat value in the
forward direction. The absence of an overall isotope effect in
conversion of methylmalonyl-CoA to succinyl-CoA is consistent
with a change in the rate-determining step from product release
in wild-type enzyme to Co-carbon bond homolysis preceding
the first isotope-sensitive step. We estimate that ∼103 to 104 of
the total 1012-fold acceleration of the homolysis rate may be
derived from the contributions of a single residue, Y89, to this
reaction.
The effects of the conservative substitution, Y89F, have been
partially evaluated by Leadlay and co-workers who also reported
the crystal structure of this mutant enzyme.10 This structure was
obtained in the presence of the nonhydrolyzable substrate,
succinylcarbadethia-CoA, and revealed that it was essentially
superimposable on that of the wild-type enzyme except at the
locus of the alteration. However, the electron density on the
aromatic ring of F89 was broadened, indicating increased
mobility in the absence of the hydrogen-bonding interaction with
the substrate that is observed with the wild-type Y89 residue.
Methods
The Y89F substitution was found to lower kcat in the reverse
direction, i.e., conversion of succinyl-CoA to methylmalonyl-
CoA, by 580-fold but had little effect on the Km value.10 The
overall deuterium isotope effect in the reverse direction was
similar in the mutant (HV/DV ) 5.2 ( 0.8) and wild-type
(HV/DV ) 3.4 ( 0.5) enzymes within the limits of experimental
error. On the basis of the directional sensitivity of the fraction
of tritium partitioning from AdoCbl to substrate versus product,
which is in marked contrast to the behavior of the wild-type
Chemicals. AdoCbl and methylmalonlyl-CoA were purchased from
Sigma. Radioactive [14C]-CH3-malonyl-CoA (56.4 Ci/mol) was pur-
chased from New England Nuclear. All other chemicals were reagent
grade commercial products and were used without further purification.
Construction of Site-Specific Mutants. Site-directed mutants were
created using the Quick Change strategy (Stratagene) and the following
primers for Y89F (forward: CCCTGGACGATTCGCCAGTTCGC-
CGGTTTCTCCACGGC) and Y89A (forward: CCCTGGACGAT-
TCGCCAGGCCGCCGGTTT CTCCACGGC). The mutagenic codons
specifying F and A are italicized. The reverse primers had the
complementary sequences.
(8) Chowdhury, S.; Banerjee, R. J. Am. Chem. Soc. 2000, 122, 5417-5418.
(9) Mancia, F.; Evans, P. Structure 1998, 6, 711-720.
(10) Thoma¨, N. H.; Meier, T. W.; Evans, P. R.; Leadlay, P. F. Biochemistry
1998, 37, 14386-14393.
(11) Meier, T. W.; Thoma¨, N. H.; Leadlay, P. F. Biochemistry 1996, 35, 11791-
11796.
9
5432 J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003

Acceleration of Co−Carbon Bond Homolysis by Y89
A R T I C L E S
Table 1. Comparison of Kinetic Parameters of Y89 Mutants and
Wild-Type Methylmalonyl-CoA
mutase
wild typea
Y89F
Y89A
k
cat (s-1 at 37 °C)
120
0.125 ( 0.02
0.135 ( 0.03
357 ( 83
0.137 ( 0.02
0.29 ( 0.05
926 ( 153
Kd (AdoCbl), µM
Km(M-CoA), µM
HV/DV
0.17 ( 0.01
133 ( 37
5 ( 0.6
1.2 ( 0.09
0.96 ( 0.12
a These parameters are taken from ref 17. The radiolabel assay containing
an (R,S) mixture of methylmalonyl-CoA was employed as described
previously.12
Enzyme Expression and Purification. The mutant enzymes were
purified through the step preceding reconstitution with cofactor using
a modified procedure described previously for isolation of wild-type
enzyme.7
Enzyme Assays. Specific activity of the mutase was determined in
the radiolabeled assay at 37 °C as described previously.12 One unit of
activity catalyzes the formation of 1 µmol of succinyl-CoA min-1 at
37 °C. The concentration of mutant enzymes was increased 1000-fold
with respect to the wild-type enzyme in the standard assay. Kinetic
parameters for the two mutants were determined by increasing the
duration of the fixed timed assay from 3 to 5 min in the presence of
varying concentrations of [14C]-methylmalonyl-CoA (100 to 5000 µM).
Protein concentration was determined using the Bradford reagent
(BioRad) and bovine serum albumin as a standard.
Determination of Equilibrium Binding Constants for AdoCbl by
Fluorescence Spectroscopy. Addition of the cofactor to apo-methyl-
malonyl-CoA mutase results in a decrease in fluorescence emission at
340 nm and has been used to determine the equilibrium dissociation
constant for the wild-type enzyme.13 The enzyme concentration used
in this experiment was 0.8 µM in 50 mM potassium phosphate buffer,
pH 7.5. Successive aliquots (2-5 µL) of a stock 50 µM AdoCbl solution
prepared in the same buffer were added followed by incubation at
4 °C for 30 min prior to measurement of the fluorescence emission.
Determination of Equilibrium Binding Constant by UV-Visible
Absorption Spectroscopy. Binding of AdoCbl to methylmalonyl-CoA
mutase was followed spectrophotometrically as previously described.13
The enzyme concentration used in this experiment was 30 µM, and
total AdoCbl concentration added was varied from 1.5 to 40 µM. The
Kd values obtained from the equilibrium fluorescence experiments were
used to calculate the free ligand concentrations. The experimentally
obtained values for ∆A562 nm were plotted versus the concentration of
free AdoCbl, L, and the data were fitted using eq 1.
Figure 3. Comparison of wild-type and Y89F enzyme spectra under steady-
state conditions. UV-visible spectra of holo-Y89F (55 µM in AdoCbl
concentration, thin solid line) in 50 mM potassium phosphate buffer, pH
7.5 at 37 °C. Addition of (R,S)-methylmalonyl-CoA to a final concentration
of 2 mM, prepared in the same buffer, did not show any significant change
in the spectra of enzyme (dashed line) over a 2-h incubation. The spectrum
of wild-type enzyme under steady-state turnover conditions (thick line) is
from ref 4. The arrow at 470 nm indicates formation of cob(II)alamin in
wild-type enzyme.
possibilities, homolysis (step i) and rearrangement (step iii), the
former appears to be significantly impacted since the spectrum
of the enzyme under steady-state conditions reveals only the
presence of AdoCbl (Figure 3). This is in contrast to wild-type
enzyme in which the ratio of AdoCbl:cob(II)alamin is ∼4:1
under steady-state conditions.4 If an increase in the rearrange-
ment barrier were chiefly responsible for suppression of the
isotope effect, the enzyme would be predicted to accumulate
in the cob(II)alamin state.
Presteady-State Kinetics. In principle, the absence of
observable cob(II)alamin under steady-state turnover conditions
could result from a completely rate-determining product-release
step which would result in accumulation of the cofactor in the
AdoCbl state. However, the presteady-state formation of cob-
(II)alamin would not be affected under these conditions. To
distinguish between these possibilities, we attempted to measure
the homolysis reaction under presteady-state conditions with
the F89 mutant as described previously for the wild-type
enzyme.5 However, we were unable to measure this rate constant
since the spectroscopic changes accompanying formation of cob-
(II)alamin were not observed in the presence of either protiated
or deuterated substrate, consistent with the spectral observations
under steady-state conditions. These results indicate that the
barrier or equilibrium for the homolysis step have been
dramatically altered by loss of the hydroxyl group in the F89
mutant and support the conclusion that homolysis rather than
the rearrangement step represents the kinetic bottleneck in the
forward direction. In wild-type enzyme, the rate-limiting step
is believed to be product release, leading to the predominance
of AdoCbl in the steady-state spectrum of the enzyme. The
absence of detectable levels of cob(II)alamin under presteady-
state conditions in the Y89F mutant is also consistent with an
earlier step along the reaction coordinate, limiting conversion
of substrate to product.
Lb ) E0L/(Kd + L) + cL
(1)
where cL is the contribution from a low-affinity nonspecific binding
site.
Results and Discussion
Steady-State Kinetic Properties of Mutant Proteins. The
mutant proteins were expressed at levels comparable to the wild-
type protein, and were purified using the same procedure.7 A
comparison of the steady-state kinetic parameters (Table 1)
reveals significant differences. Thus, kcat for the conversion of
methylmalonyl-CoA to succinyl-CoA is diminished ∼103-fold,
whereas the Km for substrate and Km-app for cofactor are only
modestly affected. Significantly, the overall reaction in the
forward direction is insensitive to isotopic substitution, sug-
gesting that a step preceding one of the two isotope-sensitive
steps (ii and iv in Figure 1) is rate-determining. Of the two
In the wild-type enzyme, khom is >10-fold faster than kcat.4
Thus, an upper limit for the net effect of the Y89F mutant on
homolysis is an ∼104-fold deceleration, corresponding to 5.4
(12) Taoka, S.; Padmakumar, R.; Lai, M.-t.; Liu, H.-w.; Banerjee, R. J. Biol.
Chem. 1994, 269, 31630-31634.
(13) Chowdhury, S.; Banerjee, R. Biochemistry 1999, 38, 15287-94.
9
J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003 5433

A R T I C L E S
Vlasie and Banerjee
kcal mol-1, which could be attributed to loss of a single
hydrogen-bonding group, i.e., the OH in Y89.
Stereoselectivity of Mutant Enzymes. The bulk and position
of the Y89 residue have been postulated to be important in the
stereoselectivity for the (R)-isomer exhibited by the wild-type
enzyme.14 A plausible explanation for the higher Km for substrate
displayed by the Y89 mutants is loss of stereochemical dis-
crimination in binding and possibly, in substrate utilization. The
equilibrium concentrations of substrate, (R,S)-methylmalonyl-
CoA, and product, succinyl-CoA, were determined following
HPLC separation of the two.15 The Keq for the mutase reaction
is reported to be ∼20 in favor of succinyl-CoA when starting
with (R)-methylmalonyl-CoA.16 In both mutants and the wild-
type enzyme, succinyl-CoA and the unreacted (S)-isomer of
methylmalonyl-CoA were present at ca. equal concentrations
(not shown). Thus, these results preclude utilization of (S)-
methylmalonyl-CoA by the Y89 mutants as substrate but do
not rule out its binding.
Oxygen Stability of Mutant Enzymes. Mutations in the
active-site residue, H244, which is in hydrogen-bonding contact
with the substrate (Figure 2), lead to acute oxygen sensitivity
of the reaction due to increased interception of the cob(II)alamin
intermediate.17,18 In contrast, mutations at Y89 lead to a highly
stable enzyme whose activity is insensitive to oxygen and which
is very resistant to photolysis even when irradiated with a
tungsten lamp in the presence or absence of substrate (not
shown). Qualitatively, the latter result is consistent with
decreased accumulation of the cob(II)alamin intermediate,
leading to reduced susceptibility of the reaction to oxidative
side reactions.
Qualitative Free Energy Profile. The consequences of
mutating the active-site Y89 residue on the energetics of the
reaction are discussed within the framework of a qualitative
free energy profile shown in Figure 4. In wild-type enzyme,
the overall deuterium isotope effect is suppressed, and AdoCbl
and cob(II)alamin are observed in a 4:1 ratio under steady-state
conditions. Furthermore, the homolysis rate is fast (∼10-fold
greater than kcat). Taken together, the predominance of AdoCbl
is consistent with accumulation of the enzyme in the E‚P
complex and with product release being rate limiting.
Tritium-partitioning experiments furnish two additional pieces
of information.10,11 First, the fraction of tritium released from
AdoCbl to product versus that from substrate is independent of
the direction in which the reaction is monitored, and the isotope
partitions 75:25 in favor of succinyl-CoA over methylmalonyl-
CoA with wild-type enzyme. This indicates that the substrate
and product radicals are in rapid equilibrium and that the
rearrangement barrier is low and not rate limiting. Second the
tritium isotope effect, monitored by the rate of appearance of
tritium in succinyl-CoA from 5′-tritiated AdoCbl when the
reaction is initiated with methylmalonyl-CoA, reports on the
isotope effect on the second H-atom transfer step and is
influenced by the rate of product release, i.e., the subsequent
Figure 4. Qualitative free energy profiles of wild-type (A) and Y89F (B)
forms of methylmalonyl-CoA mutase. The energy barriers that are isotope
sensitive are depicted by the dotted lines, and the isotope effect is shown
on the transition state for clarity although it actually affects the ground
state. For wild-type enzyme (A), the barrier for product release is shown
as being high in both forward and reverse directions. These barriers have
been arbitrarily set at the same height in the mutant and wild-type enzyme
and the relative increase in the other barriers in the mutant is shown. Ado•,
S•, and P• refer to the deoxyadenosyl-, substrate-, and product radicals
respectively, and the cob(II)alamin radical that is present along with these
organic radicals is not shown for clarity.
step.11 For wild-type enzyme, the tritium isotope effect on
H-atom transfer from deoxyadenosine to product radical is
suppressed (Hk/Tk ) 3.2),19 consistent with product release being
rate limiting. The tritium isotope effect on the H-atom transfer
from deoxyadenosine to the substrate radical (i.e. when the
reaction is initiated with succinyl-CoA) has not been measured.
However, the deuterium isotope effect on cob(II)alamin forma-
tion has been measured in the forward direction (i.e. on the
first H-atom transfer step) under presteady-state conditions and
found to be anomalously high, suggesting quantum mechanical
tunneling.8
The Y89F mutant displays several kinetic features that
distinguish it from wild-type enzyme and reveals that loss of
the hydroxyl group in the tyrosine residue has a marked
influence on the energetics of the reaction. First, while the
overall deuterium isotope effect is suppressed in the forward
direction, it is comparable to that exhibited by the wild-type
enzyme in the reverse direction. Second, cob(II)alamin does not
accumulate at detectable levels under presteady-state or steady-
state conditions. Third, the partitioning of tritium from AdoCbl
to substrate and product is sensitive to the direction in which
the reaction is initiated.10 When methylmalonyl-CoA is em-
(14) Mancia, F.; Smith, G. A.; Evans, P. R. Biochemistry 1999, 38, 7999-
8005.
(15) Padmakumar, R.; Padmakumar, R.; Banerjee, R. V. Methods Enzymol. 1997,
279, 220-224.
(16) Kellermeyer, R. W.; Allen, S. H. G.; Stjernholm, R.; Wood, H. G. J. Biol.
Chem. 1964, 239, 2562-2569.
(17) Maiti, N.; Widjaja, L.; Banerjee, R. J. Biol. Chem. 1999, 274, 32733-
(19) The tritium kinetic isotope effects reported in ref 10 were overestimated
since an equivalence factor of 2 rather than 3 was employed. The derivation
of this correction factor is discussed in Chih, H. W.; Marsh, E. N.
Biochemistry 2001, 40, 13060-13067.
32737.
(18) Thoma, N. H.; Evans, P. R.; Leadlay, P. F. Biochemistry 2000, 39, 9213-
21.
9
5434 J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003

Acceleration of Co−Carbon Bond Homolysis by Y89
A R T I C L E S
ployed as substrate, the tritium partitions 60:40 in favor of
succinyl-CoA whereas the ratio is 90:10 in favor of succinyl-
CoA when the latter is employed to initiate the reaction. Fourth,
the tritium isotope effect on the conversion of product radical
to product is 19.10,19
between substrate (or product) and deoxyadenosine, a consid-
eration that becomes particularly important in tunneling and in
promoting homolysis by kinetic coupling.
Loss of the hydrogen bond could differentially affect precise
positioning of the product versus substrate radical leading to
increase in barrier V versus III in Figure 4B. Furthermore, the
relative stabilities of the substrate and product radicals that are
â and R to the carboxylate, respectively, may be influenced by
loss of this hydrogen-bonding interaction. Thus, the product
radical is likely to be preferentially destabilized in the Y89F
mutant as shown in Figure 4B. However, it is important to note
that the substrate carboxylate is engaged in additional electro-
static interactions via a two-pronged hydrogen-bonding contact
with the guanidino group of R207 in the active site. Thus,
neutralization of the negative charge on the carboxylate is
unlikely to be governed primarily by Y89. This view is
supported by a recent computational study on the effect of the
Y89F mutant on the rearrangement reaction.20 Loss of the
hydrogen-bonding interaction is predicted to increase the
isomerization barrier from methylmalonyl-CoA to succinyl-CoA
by 1 kcal mol-1. The hydrogen-bond-accepting oxygen of the
substrate carboxyl is affected primarily and displays an increased
charge in the mutant.20
These kinetic properties can be interpreted in the context of
the free energy profile shown in Figure 4B. Increase in the
homolysis barrier (II) results in suppression of the deuterium
isotope effect in the forward direction and the failure to observe
cob(II)alamin under presteady-state and steady-state conditions.
The altered tritium partitioning ratios reveal an increase in the
rearrangement barrier (IV) relative to the wild-type enzyme,
while the sensitivity of the fractionation to the direction of the
reaction indicates that the barriers connecting deoxyadenosine
and substrate (III) or product (V) radicals have been differen-
tially altered. Thus, the barrier between the product radical and
deoxyadenosine (V) is increased relative to that between the
substrate radical and deoxyadenosine (III), leading to less
favorable partitioning toward product when the reaction is
initiated with methylmalonyl-CoA. This is also consistent with
the increase in the magnitude of the tritium isotope effect for
the H-atom transfer between deoxyadenosine and product
radical.
Suppression of the overall deuterium isotope effect on Vmax
in the forward but not reverse direction is explained by
differences in the balance of energetic barriers encountered
during progression of the reaction in the two directions. Thus,
in the forward direction, the highest barrier, Co-carbon bond
homolysis (II), suppresses the isotope effect, whereas in the
reverse direction, homolysis (VI) and product release (I)
contribute to its partial suppression.
Conclusions
Movement of Y89 propelled by occupation of the substrate-
binding site, plays a pivotal role in initiating radical chemistry
by cleavage of the Co-carbon bond in methylmalonyl-CoA
mutase. Tritium partitioning studies, in which isotope transfer
from the 5′ position of AdoCbl to either substrate or product
was monitored, revealed an altered and increased barrier to
rearrangement in the Y89F mutant.10 In addition, together with
the large tritium kinetic isotope effect on H-atom transfer step
between deoxyadenosine and the product radical in the mutant
versus a highly suppressed one in wild-type enzyme,10,11 they
suggest a change in the relative barriers connecting deoxyadeno-
sine and substrate versus product radicals.10 In combination with
the present study, which reveals that the cob(II)alamin inter-
mediate is not visible under presteady-state or steady-state
conditions and that the overall isotope effect in the forward
direction is suppressed, the major role of the aromatic wedge
residue, Y89, is ascribed to accelerating Co-carbon bond
homolysis driven by substrate-binding energy.
Can the multiple energetic perturbations resulting from loss
of a single hydrogen-bonding side chain in the active site be
rationalized on chemical grounds? Structural insights into the
active site of methylmalonyl-CoA mutase in the presence and
absence of substrate are helpful in guiding our speculations here.
The crystal structures reveal significant repositioning of Y89
accompanying substrate binding that has two discernible effects.
First, the apparent destruction of the deoxyadenosine binding
site in the E‚S complex appears to result from motion of Y89
and could explain the contribution to the Co-carbon bond
homolysis step. The deoxyadenosine binding site is not well
determined in the structures of methylmalonyl-CoA mutase, and
it is not known if there are direct interactions between Y89 and
deoxyadenosine. In principle, Y89 may play a role in stabilizing
the deoxyadenosyl radical and thereby promote the homolysis
reaction. Second, the hydroxyl group of Y89 is engaged in
hydrogen-bonding interactions with the substrate carboxylate
and with a cofactor side chain (Figure 2). These interactions
may be important in precise positioning of the substrate and in
modulating the distance between the H-atom transfer sites, i.e.,
Acknowledgment. This work was supported by the National
Institutes of Health (DK45776). R.B. is an Established Inves-
tigator of the American Heart Association.
JA029420+
(20) Loferer, M. J.; Webb, B. M.; Grant, G. H.; Liedl, K. R. J. Am. Chem. Soc.
2003, 125, 1072-1078.
9
J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003 5435