A mechanism from quantum chemical studies for methane formation in methanogenesis.

Abstract

The mechanism for methane formation in methyl-coenzyme M reductase (MCR) has been investigated using the B3LYP hybrid density functional method and chemical models consisting of 107 atoms. The experimental X-ray crystal structure of the enzyme in the inactive MCR(ox1)(-)(silent) state was used to set up the initial model structure. The calculations suggest a mechanism not previously proposed, in which the most remarkable feature is the formation of an essentially free methyl radical at the transition state. The reaction cycle suggested starts from a Michaelis complex with CoB and methyl-CoM coenzymes bound and with a squareplanar coordination of the Ni(I) center in the tetrapyrrole F(430) prosthetic group. In the rate-limiting step the methyl radical is released from methyl-CoM, induced by the attack of Ni(I) on the methyl-CoM thioether sulfur. In this step, the metal center is oxidized from Ni(I) to Ni(II). The resulting methyl radical is rapidly quenched by hydrogen-atom transfer from the CoB thiol group, yielding the methane molecule and the CoB radical. The estimated activation energy is around 20 kcal/mol, which includes a significant contribution from entropy due to the formation of the free methyl. The mechanism implies an inversion of configuration at the reactive carbon. The size of the inversion barrier is used to explain the fact that CF(3)-S-CoM is an inactive substrate. Heterodisulfide CoB-S-S-CoM formation is proposed in the final step in which nickel is reduced back to Ni(I). The suggested mechanism agrees well with experimental observations.

Full text

Page 1

A Mechanism from Quantum Chemical Studies for Methane
Formation in Methanogenesis
Vladimir Pelmenschikov,†Margareta R. A. Blomberg,†Per E. M. Siegbahn,*,†and
Robert H. Crabtree‡
Contribution from the Department of Physics, Stockholm UniVersity, Box 6730,
S-113 85 Stockholm, Sweden, and Chemistry Department, Yale UniVersity, P.O. Box 208107,
225 Prospect Street, New HaVen Connecticut 06520-8107
Received July 9, 2001. Revised Manuscript Received November 21, 2001
Abstract: The mechanism for methane formation in methyl-coenzyme M reductase (MCR) has been
investigated using the B3LYP hybrid density functional method and chemical models consisting of 107
atoms. The experimental X-ray crystal structure of the enzyme in the inactive MCRox1-silentstate was used
to set up the initial model structure. The calculations suggest a mechanism not previously proposed, in
which the most remarkable feature is the formation of an essentially free methyl radical at the transition
state. The reaction cycle suggested starts from a Michaelis complex with CoB and methyl-CoM coenzymes
bound and with a squareplanar coordination of the Ni(I) center in the tetrapyrrole F430prosthetic group. In
the rate-limiting step the methyl radical is released from methyl-CoM, induced by the attack of Ni(I) on the
methyl-CoM thioether sulfur. In this step, the metal center is oxidized from Ni(I) to Ni(II). The resulting
methyl radical is rapidly quenched by hydrogen-atom transfer from the CoB thiol group, yielding the methane
molecule and the CoB radical. The estimated activation energy is around 20 kcal/mol, which includes a
significant contribution from entropy due to the formation of the free methyl. The mechanism implies an
inversion of configuration at the reactive carbon. The size of the inversion barrier is used to explain the
fact that CF3-S-CoM is an inactive substrate. Heterodisulfide CoB-S-S-CoM formation is proposed in
the final step in which nickel is reduced back to Ni(I). The suggested mechanism agrees well with
experimental observations.
I. Introduction
Archaebacteria (Archaea) contain several nickel enzymes, a
metal otherwise rather rare in biochemistry.1,2In several cases,
such as CO dehydrogenase, the nickel is generally assumed to
form organometallic intermediatessones with a direct Ni-C
bond.1Methanogenic bacteria are a diverse subgroup of ar-
chaebacteria that use methane formation to provide energy for
the cell.3One such pathway, shown in eq 1, involves the release
of 31 kcal/mol,
In this case, the overall 8e-reduction of CO2occurs via four
2e-steps, the reduced carbon fragment being bound to a series
of coenzymes. This report is concerned only with the final step,
in which two coenzymes are involved: one, coenzyme M,
carries the methyl group that comes from CO2reduction and
the other, coenzyme B, is an aliphatic thiol. This last step,
represented by eq 2, is a topic of intense current interest.4,5
Catalyzed by the nickel-dependent protein, methylcoenzyme M
reductase (MCR), the reaction in eq 2 involves the release of
11 kcal/mol6,7and occurs in the oxidative part of the meth-
anogenic archaea energy metabolism. The heterodisulfide also
formed is subsequently hydrogenolyzed with H2 back to the
thiol forms of the separate cofactors by heterodisulfide reduc-
tase8,10in the reductive part of the cycle.
The key MCR enzyme has an R2?2γ2subunit structure and,
as normally isolated, contains two molecules of a nickel
porphyrinoid cofactor, denoted F430because of its absorption
maximum at 430 nm, along with two molecules each of
* To whom correspondence should be addressed. E-mail: ps@physto.se.
†Stockholm University.
‡Yale University.
(1) Ragsdale, S. W. Curr. Opin. Chem. Biol. 1998, 2, 208-215.
(2) Ermler, U.; Grabarse, W.; Shima, S.; Goubeard, M.; Thauer, R. K. Curr.
Opin. Struct. Biol. 1998, 8, 749-758.
(3) Thauer, R. K. Microbiology 1998, 144, 2377.
(4) Cammack, R. Nature 1997, 390, 443.
(5) Ferry, J. G. Science 1997, 278, 1413.
(6) Thauer, R. K.; Hedderich, R.; Fischer, R. In Methanogenesis; Ferry, J. G.,
Ed.; Chapman and Hall: New York and London, 1993; Chapter 2, p 209.
(7) Becker, D. F.; Ragsdale, S. W. Biochemistry 1998, 37, 2639.
(8) Heiden, S.; Hedderich, R.; Setzke, E.; Thauer, R. K. Eur. J. Biochem. 1993,
213, 529.
(9) Ermler, U.; Grabarse, W.; Shima, S.; Goubeard, M.; Thauer, R. K. Science
1997, 278, 1457.
(10) Setzke, E.; Hedderich, R.; Heiden, S.; Thauer, R. K. Eur. J. Biochem. 1994,
220, 139.
CO2+ 4H2f CH4+ 2H2O (1)
CoB-S-H + CH3-S-CoM f
CoB-S-S-CoM + CH4(2)
Published on Web 03/22/2002
10.1021/ja011664r CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 4039-4049 9 4039

Page 2

coenzymes B and M. The structures of these cofactors are shown
in Figure 1. See also Figure 2.
In the Ni(II) state, F430can be four- or six-coordinate, and
can be readily reduced to a Ni(I) form. Unlike true porphyrins,
which have a Ni(II) center and a reduced porphyrin, F430
contains authentic Ni(I), as shown by the EPR spectrum.11Ni(I)
is believed to be relevant to the mechanism because the enzyme
only becomes active when the resting Ni(II) state of F430 is
reduced to Ni(I), for example with Ti(III) salts.12
Methyl-coenzyme M (2-mercaptoethanesulfonic acid) has the
simplest cofactor structure known and is unique to the meth-
anogens. Coenzyme B (7-mercaptoheptanoylthreonine phos-
phate) has an aliphatic linker of six methylene units between
the phosphothreonine headgroup and the thiol group. Alteration
in the length of this linker is very deleterious to activity. Having
one less methylene gives only 1% of activity, and having one
more methylene abolishes activity completely.13
Some insight into the reason for the sensitivity to linker length
comes from the crystal structure of the protein from Methano-
bacterium thermoautotrophicum.9Two independent active sites
are located 50 Å apart. Each consists of a nonpolar pit 30 Å
deep and about 6 Å in diameter. F430is located at the bottom
of the pit. Hanging down from the top is coenzyme B, anchored
at the top by salt bridges involving surface residues and the
phosphothreonine headgroup that lies at and partially blocks
the mouth of the pit. The six methylenes of the linker allow the
CoB sulfur atom to hang 8.7 Å above the Ni atom of F430.
Between CoB and Ni lies the small CoM cofactor. In the
particular state studied, the CoM sulfur is bound to the Ni.
In view of the presence of Ni-C bonds in other Ni enzymes
and of Co-C bonds in many derivatives of the well-known and
somewhat related coenzyme B12, it was perhaps inevitable that
the mechanisms suggested for methanogenesis have all involved
formation of a Ni-CH3bond at some point, usually followed
by protonolysis to release methane.3One of the mechanisms
proposed in ref 3 is illustrated in Figure 3 and summarized in
the set of eqs 3-5. The first step is suggested to be a methyl
cation transfer from methyl-CoM to nickel, yielding a CH3-
Ni(III)F430
compound. It is further assumed that this hetero-
lytic cleavage of the S-methyl bond needs to be accompanied
by a proton transfer to the CoM-leaving group, and that a
possible source for the proton could be CoB. This first step
can then be summarized in eq 3:
+
Thus, apart from cleaving the S-CH3bond and forming the
Ni-C bond, this step also includes a charge separation. The
CH3-Ni(III)F430
step it is suggested to oxidize H-S-CoM yielding a cation
radical on CoM according to eq 4:
+compound is a strong oxidant, and in the next
CH3-Ni(II)F430is suggested to be spontaneously protonolyzed
to give CH4and Ni(II)F430
radical cation. The CoM thiyl radical and the CoB thiolate are
assumed to combine into CoB-S-•S-CoM-disulfide radical
anion, which reduces Ni(II), forming the final neutral products
+, using the proton of the CoM thiyl
(11) Pfaltz, A. In The Bioinorganic Chemistry of Nickel; Lancaster, J. R., Ed.;
VCH: Weinheim, 1988; Chapter 12.
(12) Goubeaud, M.; Schreiner, G.; Thauer, R. K. Eur. J. Biochem. 1997, 243,
110.
(13) Ellermann, J.; Hedderich, R.; Boecher, R.; Thauer, R. K. Eur. J. Biochem.
1988, 172, 669.
Figure 1. Structures of methyl-coenzyme M, coenzyme B, and F430
prosthetic group, required for methane formation in methanogens.
Figure 2. Active site region of methyl-coenzyme M reductase as found in
the X-ray structure of the enzyme in the MCRox1-silent state.9Peripheral
substituents of the F430prosthetic group are omitted for clarity.
CoB-S-H + CH3-S-CoM + Ni(I)F430f
CoB-S-+ H-S-CoM + CH3-Ni(III)F430
+
(3)
H-S-CoM + CH3-Ni(III)F430
+f
H-•S-CoM++ CH3-Ni(II)F430(4)
A R T I C L E SPelm enschikov et al.
4040 J. AM. CHEM. SOC.9VOL. 124, NO. 15, 2002

Page 3

and closing the reaction cycle. The last step thus can be given
as eq 5:
When methane and the disulfide have left the active-site cleft,
the enzyme is ready for the next cycle. As will be discussed
below, the present calculations indicate that the nickel-methyl
bond strength is too low to make this reaction scheme
energetically plausible.
In an alternative mechanism, also described in ref 3, methyl-
CoM is proposed to be activated by a CoB thiyl radical, such
that a homolytic cleavage of the methyl-S bond is favored,
directly yielding a CH3-Ni(II)F430compound. This mechanism
thus assumes the presence of a strong oxidant, capable of
oxidizing CoB into a thiyl radical. However, the only redox
center at the active site of methyl-coenzyme M reductase is
Ni(II)F430, but this center has been experimentally ruled out as
the oxidant.
The present contribution uses DFT methods to examine
possible mechanisms with the very unexpected result that Ni-C
bonded intermediates can be excluded and the proposed
mechanism involves a nickel-induced release of •CH3radical
from CH3-S-CoM that is immediately quenched by H-atom
transfer from CoB-S-H.
II. Computational Details
The calculations were performed in two steps. For each structure
considered, a full geometry optimization was performed using the hybrid
density functional B3LYP method.14,15In this first step, standard
double-? basis sets were used for all light elements. For nickel a
nonrelativistic Hay and Wadt16effective core potential (ECP) was used.
The valence basis set used in connection with this ECP is essentially
of double-? quality. The geometry optimizations were carried out with
either the GAUSSIAN program17,18using the lanl2dz basis or the Jaguar
4.0 program19using the lacVp basis. The GAUSSIAN program and
the lanl2dz basis was used also for the Hessian calculations, that is,
the second derivatives of the energy with respect to the nuclear
coordinates. Some restrictions, taken from the X-ray structure, were
superimposed on the geometry optimizations, as further described in
the text below. In a second step, the energy was evaluated for the
optimized geometries using larger basis sets of triple-? quality in the
valence region, and including a single set of polarization functions on
each atom. This final energy evaluation was performed at the B3LYP
level using the Jaguar 4.0 program and the lacV3p** basis. The inherent
accuracy of the B3LYP method can be estimated from benchmark tests,
in which the average error in the atomization energies for 55 small
first- and second-row molecules is found to be 2.2 kcal/mol.20For
transition metals there are no benchmarks due to the lack of accurate
experimental numbers but indications from normal metal-ligand bond
strengths are that the errors are slightly larger, 3-5 kcal/mol.21
The surrounding protein was treated using self-consistent reaction
field methods, where the cavity follows the shape of the molecular
system. For Jaguar, a Poisson-Boltzmann solver was used with a probe
radius of 1.40 Å corresponding to the water molecule, while for
Gaussian, the conductor-like polarized continuum model (CPCM, or
COSMO) method22was used, again with water as a probe. The dielectric
constant of the protein is the main empirical parameter of these methods,
and it was chosen to be equal to 4 in line with previous suggestions
for proteins. In agreement with previous findings the calculated
dielectric effects on the relative energies were found to be small for
reactions where the charge state of the model is constant.
The relative energies discussed below are those obtained using
the large lacV3p** basis set, while zero-point vibrational effects,
entropy effects, and dielectric effects are included only when specifically
stated.
(14) Becke, A. D. Phys. ReV. 1988, A38, 3098; Becke A. D. J. Chem. Phys.
1993, 98, 1372; Becke, A. D. J. Chem. Phys. 1993 98, 5648.
(15) Stevens, P. J.; Devlin, F. J.; Chablowski, C. F.; Frisch, M. J. J. Phys. Chem.
1994, 98, 11623.
(16) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
(17) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.; Johnson, B.
G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G. A.;
Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V.
G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94, Revision B.2; Gaussian
Inc.: Pittsburgh, PA, 1995.
(18) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millan, J. M.; Daniels, A. D.; Kudin, K.
N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon, M.;
Replogle, E. S.; Pople, J. A. Gaussian 98; Gaussian Inc.: Pittsburgh, PA,
1998.
(19) Jaguar 4.0, Schro ¨dinger, Inc., Portland, OR, 1991-2000.
(20) Bauschlicher, C. W., Jr.; Ricca, A.; Partridge, H.; Langhoff, S. R. In Recent
AdVances in Density Functional Methods, Part II; Chong, D. P., Ed.; World
Scientific Publishing Company: Singapore, 1997; p 165.
(21) Siegbahn, P. E. M.; Blomberg, M. R. A. Annu. ReV. Phys. Chem. 1999,
50, 221-249.
(22) Barone, V.; Cossi, M. J. Phys. Chem. 1998, 102, 1995-2000.
Figure 3. Sketch of a previously suggested mechanism for methyl-CoM
reductase.9The Ni-CH3bond formed at point 2 is subsequently protono-
lyzed to form methane when proceeding from 3 to 4.
CoB-S-+ H-•S-CoM++ CH3-Ni(II)F430f
CoB-S-S-CoM + CH4+ Ni(I)F430(5)
Mechanismfor Methane Form ation in MethanogenesisA R T I C L E S
J. AM. CHEM. SOC. 9 VOL. 124, NO. 15, 2002 4041

Page 4

III. Results and Disscusion
The first step in every theoretical study of an enzyme
mechanism is to choose a chemical model that describes the
active site reasonably well yet is not too large and therefore
impractical. In the present case useful models have to be selected
for the F430, CoM and CoB cofactors. Finally, important
additional hydrogen bonds have to be identified. These modeling
considerations are discussed in section III.a. below. The next
step in the present study was to calculate some key bond
strengths. In particular, from the experimentally suggested
mechanisms described above, it is clear that the Ni-CH3bond
strengths for both Ni(II) and Ni(III) are critical quantities. These
and other bond strengths are discussed in section III.b. The bond
strengths obtained combined with key experimental information,
suggests a quite different mechanism from the ones proposed
earlier. This new mechanism is discussed in section III.c. In
section III.d., the calculations using the largest model are
described and in section III.e. the implication of stereoinversion
at the reactive carbon are discussed. Finally, the electronic and
geometric structure of the F430cofactor are described.
III.a. Chemical Models. The present modeling study is based
on the experimentally determined structure of the methyl-
coenzyme reductase in the MCRox1-silent state. Two different
models for the F430tetrapyrrole cofactor were tested, see Figure
4. The largest one, F430
cofactor, a lactam ring joined with pyrrole B, and a six-
membered carbocyclic ring joined with pyrrole D. The propi-
onate substituents of the rings A, B, and C, and acetate
substituents of C and D, anchoring F430 by the carboxylate
groups to the protein, were replaced by hydrogens. The methyl
groups at A and B, and the acetamide substituent of A were
also omitted to reduce the size of the system and to make it
computationally feasible. The less extended model used at the
initial stage of our investigations is a Ni-chelating macrocycle
complex F430
cofactor is the most saturated tetrapyrrole known in nature. The
low degree of conjugation supports the use of less extended
models such as F430
original cofactor is retained. For the catalytic properties of the
nickel center, the difference between the F430
was found to be quite small, see further below in the text. The
F430prosthetic group is known to be extremely flexible, again
due to the absence of extensive conjugation. To better model
A, includes four pyrrole rings of the
B, where the pyrrole rings are broken. The F430
B, where the existing conjugation of the
A
and F430
B
models
the F430deformations, important for the heterodisulfide forma-
tion (not discussed in the present report), the F430
perhaps better than F430
which may be somewhat less flexible.
The most critical properties of methyl-coenzyme M that need
to be accurately modeled are the CH3-SCoM and the Ni-
SCoM bond strengths. The CH3-SCoM bond strength, as given
by reaction 7,
A
model is
B
turns out to be surprisingly insensitive to the size of the model.
The simplest possible model, CH3-S-CH3, and the full model
of this cofactor, including also the terminal sulfonate group that
anchors the cofactor to the polypeptide chain in the wall of the
pit, give very similar bond strengths of 70.1 and 69.8 kcal/mol,
respectively. When other substrates were investigated, see
section III.e., it was found that even the S-CHF2and S-CF3
bond strengths are very similar to that of methyl, and indepen-
dent of the size of the cofactor model. The bond strength
between sulfur and nickel, as given by reaction 8,
is more difficult to estimate. The reason is that it is difficult to
describe the hydrogen-bonding interactions in a balanced way
before and after the formation of the Ni-S bond. However,
when the smallest possible model of methyl-coenzyme M, CH3-
S-CH3, is used, a good approximation should be to neglect
the additional hydrogen bonding. In that case a Ni-S bond
strength of 38.6 kcal/mol is obtained using the F430
the nickel cofactor. In another model the entire coenzyme M
was used explicitly, including its sulfonate group. The sulfonate
was protonated since its negative charge should be balanced
by a salt bridge to the guanidium group of Arg120 and by two
hydrogen bonds to the peptide nitrogen of Tyr444 and His364,
as indicated by the X-ray structure.9The full model of coenzyme
M, gives a reaction energy for reaction 8 of 46.1 kcal/mol. A
major part of the difference from the small model is the
unbalanced hydrogen-bonding energy. For this reason, the most
realistic value for the Ni-S bond strength should come from
the use of the small coenzyme M model of 38.6 kcal/mol. Still,
in the final calculations of the mechanism described below,
methyl-coenzyme M was modeled explicitly, including the
sulfonate group, since this is inherently a better model, and the
above problem in describing reaction 8 does not occur for the
full reaction. In modeling coenzyme B, the heptanoyl arm can
be simplified to an ethanethiol CH3CH2-S-H, or even a
methanethiol, without significant loss of accuracy.
The Gln147 on the rear face of F430, which binds to the Ni
center with its side-chain oxygen with a bond distance of 2.3
Å in the X-ray structure, was modeled by an acetamide molecule
for the large F430
model, and by a formamide molecule for the
smaller F430
model. Among the other amino acids close to the
active site, only the two tyrosines, Tyr333 and Tyr367, were
included in view of the mechanism proposed. Interacting by
their hydroxyl groups with the coenzyme M sulfur, the tyrosines
were modeled by methanols, which should be quite sufficient
for describing just the hydrogen bonds to sulfur.
III.b. The Binding of Methyl to Nickel. A key part of the
previously suggested mechanisms, shown in Figure 3 and
described in the Introduction, is the binding of methyl to nickel.
B
model for
A
B
Figure 4. Two different nickel complexes used to model the F430cofactor.
Conjugated areas or double bonds of the tetrapyrrole are shown in bold.
CH3• + •S-CoM f CH3-S-CoM (7)
CoM-S• + Ni(I)F430f CoM-S-Ni(II)F430
(8)
A R T I C L E S Pelm enschikov et al.
4042 J. AM. CHEM. SOC.9VOL. 124, NO. 15, 2002

Page 5

Methyl is bound to nickel in the Ni(III) state at point 2 and in
the Ni(II) state at point 3. A first test of this mechanism is
therefore to calculate these bond strengths. A binding energy
of methyl to the Ni(I)F430
complex of only 24.9 kcal/mol was
obtained using the large basis set, following eq 9 with infinitely
separated reactants:
B
For the methyl binding to the positively charged Ni(II)F430
complex as given in eq 10:
B+
an even smaller energy of 18.0 kcal/mol was found. Therefore,
for the methyl transfer from methyl-CoM to Ni as given in
eq 11:
to be energetically possible, the binding energy of methyl to
sulfur in methyl-coenzyme M have to be not much greater than
25 kcal/mol. In other words, reaction 11 will be endothermic
by at least the amount that the S-CH3bond strength in methyl-
coenzyme M is larger than the Ni(II)-CH3bond strength. As
already mentioned in the previous section, the S-CH3 bond
strength in methyl-coenzyme M is quite insensitive to the
modeling of the cofactor. For the full cofactor the bond strength
is found to be 69.8 kcal/mol, while for the simple S(CH3)2model
it is 70.1 kcal/mol. Thus, the process in eq 11 would be
endothermic by about 45 kcal/mol.
The products of the second step in the previously proposed
mechanism (right side of eq 4, or point 3 in Figure 3) differ
from the products in eq 11 by a proton transfer between the
cofactors as given in eq 12:
For the minimal methyl models of CoM and CoB, the step in
eq 12 was found to be endothermic by 67.3 kcal/mol for the
infinitely separated compounds. The surrounding protein was
treated as described in section II. Including the Coulombic
attraction between the CoB-S-thiolate and •SH-CoM+thiyl
radical cation at 8.7 Å leads to a decrease of this very large
value down to 53.9 kcal/mol. Combining reactions 11 and 12,
this means that point 3 in Figure 3 is predicted to be as much
as 100 kcal/mol higher than the initial reactants in point 1 using
the best present models. Considering probable stabilizing effects
from the particular groups in the MCR active site cannot bring
this large value down to a kinetically feasible one. Unless
B3LYP has a very large error, very unlikely on the basis of
previous benchmark tests,23,24this result rules out the mechanism
shown in Figure 3. The capability of Ni(III) in F430to oxidize
the CoM thiol group appears to have been overestimated, since
reaction 4 of the CoM thiyl radical formation is highly
endothermic by 73 kcal/mol, including the protein-surrounding
effects and charge-separation estimates.
A few additional results obtained when the above bond
strengths were calculated have implications for the mechanism
and modeling of the methane formation reactions. First, testing
different spin-states for the Ni(II)-CH3complex, showed quite
clearly that the alkyl is predicted to have a triplet ground state.
This is not so surprising since for this complex a strong bond
to methyl (ionic or covalent) has to be formed perpendicular to
the tetrapyrrole ring. Singlet Ni(II) forms strong bonds only in
one plane, while triplet Ni(II) prefers a tetrahedral coordination
and is therefore more flexible in forming bonds both in the plane
of the tetrapyrrole ring and perpendicular to it. For the same
reason, the Ni(II)-S-CH3complex also has a triplet ground
state. Another useful result is that the smaller F430
results very similar to the larger F430
methyl for the Ni(II) state is 21.3 kcal/mol for the larger model
using the lacVp basis set, differing by only 0.9 kcal/mol from
the smaller model with the same basis set. Further comparisons
between the models, also showing very good agreement, will
be described below. These results show that the smaller model,
which leads to significantly faster calculations, can be confi-
dently used to explore different mechanisms.
III.c. A New Mechanism for Methane Formation. From
the results discussed in the previous section concerning the bond
strength between methyl and nickel, any mechanism requiring
a strong bond of this type can be ruled out. The alternative
mechanism suggested, where weaker Ni-CH3 bonds are al-
lowed, involve the presence of a strong oxidant capable of
oxidizing CoB-S-H into a thiyl radical prior to the methane
formation reaction. This would clearly introduce an additional
amount of energy sufficient to make the methane formation
possible. However, since there is no evidence for such an
oxidant, this type of mechanism remains highly speculative.
Furthermore, since Ni(I) is the experimentally known active
state, the thiyl radical is required to be stable in the presence of
Ni(I), which is also very unlikely. It would therefore be
advantageous at this stage if another mechanism could be
suggested that does not require the action of an additional strong
oxidant. On the basis of the present calculations and key
experimental information, such a mechanism will be discussed
in this section.
Since a transfer of the methyl from methyl-coenzyme M to
nickel is now ruled out, the remaining possibility is to transfer
the methyl group directly between the cofactors M and B. A
puzzle in this context is that the distance between the sulfurs
of these cofactors is very long, 6.2 Å, which appears to make
a concerted transfer of methyl between the cofactors unlikely.
It is also known that cofactor B is firmly bound by several strong
hydrogen bonds and is therefore relatively immobile. The
coenzyme M sulfur should also become fairly strongly bound
to nickel in the process of forming methane, which leads to a
rather rigid system where the S‚‚‚S distance probably cannot
be significantly shortened without energy loss.
As a first test of a mechanism where methyl is transferred
from methyl-coenzyme M to cofactor B, the energy for releasing
a methyl radical can be calculated following reaction 13:
B
model gives
A
model. The binding of
Cofactor B is thus left out of the model for the moment. A
reliable estimate of this reaction energy is difficult to obtain,
(23) Curtiss, L. A.; Raghavachari, K.; Redfern, R. C.; Pople, J. A. J. Chem.
Phys. 2000, 112, 7374-7383.
(24) Siegbahn, P. E. M.; Blomberg, M. R. A. Chem. ReV. 2000, 100, 421-437.
CH3-S-CoM + Ni(I)F430f
•CH3+ CoM-S-Ni(II)F430(13)
CH3• + Ni(I)F430f CH3-Ni(II)F430
(9)
CH3• + Ni(II)F430
+f CH3-Ni(III)F430
+
(10)
CH3-S-CoM + Ni(I)F430f
•S-CoM + CH3-Ni(II)F430(11)
CoB-S-H + •S-CoM f CoB-S-+ H-•S-CoM+(12)
Mechanismfor Methane Form ation in Methanogenesis A R T I C L E S
J. AM. CHEM. SOC. 9 VOL. 124, NO. 15, 2002 4043