Methyl (Alkyl)-Coenzyme M Reductases: Nickel F-430-Containing Enzymes Involved in Anaerobic Methane Formation and in Anaerobic Oxidation of Methane or of Short Chain Alkanes.

Methyl-coenzyme M reductase (MCR) catalyzes the methane-forming step in methanogenic archaea. The active enzyme harbors the nickel(I) hydrocorphin coenzyme F-430 as a prosthetic group and catalyzes the reversible reduction of methyl-coenzyme M (CH3-S-CoM) with coenzyme B (HS-CoM) to methane and CoM-S-S-CoB. MCR is also involved in anaerobic methane oxidation in reverse of methanogenesis and most probably in the anaerobic oxidation of ethane, propane, and butane. The challenging question is how the unreactive CH3-S thioether bond in methyl-coenzyme M and the even more unreactive C-H bond in methane and the other hydrocarbons are anaerobically cleaved. A key to the answer is the negative redox potential (Eo') of the Ni(II)F-430/Ni(I)F-430 couple below -600 mV and the radical nature of Ni(I)F-430. However, the negative one-electron redox potential is also the Achilles heel of MCR; it makes the nickel enzyme one of the most O2-sensitive enzymes known to date. Even under physiological conditions, the Ni(I) in MCR is oxidized to the Ni(II) or Ni(III) states, e.g., when in the cells the redox potential (E') of the CoM-S-S-CoB/HS-CoM and HS-CoB couple (Eo' = -140 mV) gets too high. Methanogens therefore harbor an enzyme system for the reactivation of inactivated MCR in an ATP-dependent reduction reaction. Purification of active MCR in the Ni(I) oxidation state is very challenging and has been achieved in only a few laboratories. This perspective reviews the function, structure, and properties of MCR, what is known and not known about the catalytic mechanism, how the inactive enzyme is reactivated, and what remains to be discovered.

Methyl-coenzyme M reductase (MCR), which is composed of the subunits McrA (∼65 kDa), McrB (∼45 kDa), and McrG (∼35 kDa), was discovered in the laboratory of Ralph Wolfe at the University of Illinois in the 1970s when studying methane formation in cell extracts of Methanobacterium thermoautotrophicum strain ΔH (renamed Methanothermobacter thermoautotrophicus), an archaeon that grows with H2 and CO2 as sole carbon and energy sources, forming methane. The Urbana group found that methyl-coenzyme M (CH3–S-CoM) is reduced in cell extracts by H2 in an ATP-dependent reaction[1−3] involving the three components A, B, and C,[4,5] of which component C turned out to be the methyl-coenzyme M reductase (MCR) per se[6−8] and component B to be 7-mercaptoheptanoylthreonine phosphate[9,10] that was later named coenzyme B (HS-CoB). Subsequently, it was shown that besides methane also the heterodisulfide CoM-S–S-CoB is formed (Figure ).[11,12] Final evidence that CoM-S-S-CoM is a product was provided by Hedderich et al.,[13] 1989 in Marburg: the enzyme heterodisulfide reductase (HdrABC) was discovered that specifically catalyzes the reduction of CoM-S–S-CoB to HS-CoM and HS-CoB. Most of the enzymes involved in the biosynthesis of coenzyme M and coenzyme B are unique to methanogenic archaea.[14]

Figure 1

Reaction catalyzed by methyl-coenzyme M reductase that in the active form has a greenish color derived from Ni(I)F-430. Ni(I)F430 has a λmax = 382 nm and a weaker maximum at 754 nm.[15] In the inactive state, MCR is yellow from Ni(II)F-430 that has an absorbance maximum at 430 nm. When MCR catalyzes the reduction of methyl-coenzyme M with coenzyme B in 100% D2O, DCH3 and some D2CH2 are formed. When MCR catalyzes CH4 oxidation to methyl-coenzyme M in 100% D2O, only CH3–S-CoM is formed. It follows that the intermediates in the catalytic cycle do not exchange hydrogen with deuterium of the solvent. The hydrogen or deuterium atoms involved in the reaction are carried into the active site only via the thiol group of coenzyme B or, in the reverse reaction, via methane.[16] The standard free energy change (ΔGo) associated with ethane formation from ethyl-coenzyme M and coenzyme B is near −20 kJ/mol and thus by −10 kJ/mol less exergonic than ΔGo = −30 kJ/mol associated with methane formation from methyl-coenzyme M and coenzyme B (see the paragraph on “Reversibility”).

In 1979, the growth of methanogens was found by Schönheit et al.[17] in Marburg to be dependent on nickel. This finding led in 1980 to the discovery of nickel in F-430.[18,19] The nickel-containing, nonfluorescent compound, yellow in its Ni(II) and greenish in its Ni(I) oxidation states, was shown to be the prosthetic group of MCR[20] and to be a nickel tetrapyrrole.[21−24] The structure of the nickel hydrocorphin coenzyme F-430 (Abstract graphic; R = H) was elucidated in collaboration with the group of Albert Eschenmoser at the ETH Zürich.[25−28] (For the history, see ref (29).) Alternative structural proposals[30] were not confirmed.[31] The nickel in the prosthetic group has to be in the 1+ oxidation state for the enzyme to be active.[32,33] The genes and enzymes involved in the biosynthesis of coenzyme F-430 have been elucidated only recently.[34,35]

The first 20 years of MCR research were hampered by the fact that the specific activity of this enzyme in cell extracts was less than 0.1% of that expected from in vivo specific rates of methane formation. This led to many misleading results. It took to the early 1990s until it was found in Marburg how to obtain and purify active MCR with its F-430 to almost 100% in the active, Ni(I) oxidation state.[32,33] Only from then on, real biochemical investigations were possible.

In 2010, we could finally demonstrate that purified MCR from Methanothermobacter marburgensis (formerly Methanobacterium thermoautotrophicum strain Marburg) also catalyzes the back reaction, the oxidation of methane to methyl-coenzyme M, at rates sufficient to account for observed rates of anaerobic methane oxidation in vivo (Figure ).[36] This was important to show since the beginning of the 2000s evidence had accumulated that MCR is present in anaerobic methanotrophic archaea and that MCR therefore could be involved in the anaerobic oxidation of methane (AOM).[37,38] The anaerobic monofunctionalization of methane sparked the interest of biochemists and chemists in the catalytic mechanism of MCR because the homolytic cleavage of the first hydrogen of methane (439 kJ/mol) is very endergonic and by 18 kJ/mol more endergonic than of the second one.[39]

The “Biochemistry of Methyl-Coenzyme M Reductase” has recently been reviewed by Stephen Ragsdale at the University of Michigan, Ann Arbor, in the book “The Biological Chemistry of Nickel”.[40] The book chapter focuses on experimental results obtained in the Ragsdale laboratory with respect to the catalytic mechanism of MCR. The present perspective has a broader scope. It aims at providing the reader with the understanding of why and where our present knowledge of MCR is incomplete.

The perspective begins with each a section on the role of MCR in the global carbon cycle, on the phylogeny of MCR and on M. marburgensis as a model organism for the MCR study. Sections on activity and the oxidation state of nickel in MCR follow as well as sections on localization, isoenzymes, and structural features. The discussion of the catalytic properties and the proposed catalytic mechanisms is followed by that of the individual steps of the catalytic cycle. A section dealing with proton inventories highlights how important it would be to know the solvent isotope effect (H2O/D2O) more precisely. A section on half-of-the sites reactivity brings together structural and catalytic features. Finally, it is summarized what is known about the reactivation mechanism.

Role of MCR in the Global Carbon Cycle

Methane is an important intermediate in the global carbon cycle (Figure ).[41,42] About 1 Gt methane per year is formed in anoxic environments from CO2 and H2, acetate, methylamines, and methanol by methanogenic archaea involving MCR.[42] An estimated amount of 0.1 Gt methane diffuses per year from deeply buried methane deposits into the anoxic biosphere. From the 1.1 Gt methane per year, about 0.1 Gt per year is oxidized with sulfate, Fe(III), Mn(IV), and nitrate to CO2[43−45] involving MCR[46,47] (Figure ). About 1 Gt methane per year diffuses into oxic environments, where about 60% is oxidized with O2 to CO2 by aerobic bacteria involving two types of methane monooxygenases (MMO), the iron enzyme sMMO (“s” for soluble), and the copper enzyme pMMO (“p” for particulate).[48] The other 40% escapes into the atmosphere, where methane is photo-oxidized to CO2 involving OH radicals.[49] The concentration of methane in the atmosphere has more than doubled within the last 200 years (from 700 to 1875 ppb), which is of concern since methane is a potent greenhouse gas.[50]

Figure 2

Role of methyl-coenzyme M reductase (MCR) in the global carbon cycle. GPP, gross primary production. Terrestrial GPP ≈ 500 Gt CO2/year.[51] Marine GPP ≈ 300 Gt C/year.[52] Methyl-X, methanol, methylthiol, methylamines, choline, and betaine. 3-PGA, 3-phosphoglycerate. Electron acceptors for anaerobic methane oxidation are highlighted in orange. For a review on the anaerobic oxidation of ethane, propane, and butane, see Singh et al., 2017.[53]

MCR is not only involved in anaerobic methane formation and anaerobic oxidation of methane (AOM) but also appears to be also involved in the anaerobic oxidation to CO2 of short chain alkanes such as butane, propane,[54] and ethane[55] by archaea (Figure ). In the metagenome of ca. Syntrophoarchaeum butanivorans and ca. S. caldarius, which were both enriched on butane plus sulfate, and in the metagenomes of ca. Agroarchaeum ethanivorans, which was enriched on ethane plus sulfate, all of the genes required to synthesize active MCR were not only found but also found to be expressed. Most importantly, the cells were shown to contain the respective alky-S-CoM intermediate, indicating that MCR in these archaea indeed catalyzes the oxidation of these alkanes. Upon incubation with methane, the cells did not form methyl-coenzyme M, suggesting that the MCR might be able to discriminate against methane.[54,55]

Due to the low specific activity of MCR in catalyzing methane oxidation, the enzyme has to be present in methanotrophic archaea in very high concentrations (>10% of the cytoplasmic protein).[38] After RuBisCo,[56] MCR is estimated to be one of the most abundant enzymes on Earth. In anoxic sediments catalyzing AOM, the presence of F-430 can easily be detected and quantified.[57−59] MCR could even be purified and crystallized from microbial mats catalyzing AOM.[38,60]

MCR Phylogeny

Until 2003, MCR was considered to be present only in the five orders of methanogens then known: Methanobacteriales, Methanococcales, Methanomicrobiales, Methanopyrales, and Methanosarcinales.[61] From the year 2003 and on, using the mcrA gene as a genetic marker (Figure .),[62,63] the list of archaea containing MCR extended to archaea catalyzing the anaerobic oxidation of methane (ANME).[37] In 2007, the candidate order ANME was split up in the candidate’s orders of ANME-1, ANME-2, and ANME-3.[64] In 2008, the MCR-containing order of Methanocellales (uncultured archaeal group rice cluster I) was inaugurated.[65] In 2014, the mcr genes were found in methanogenic archaea of the new order Methanomassiliicoccales that are closely related to the Thermoplasmatales[66,67] in 2016 in archaea that can oxidize butane and that are distantly related to ANME-1,[54] in 2017 in methanogenic archaea closely related to Haloarchaea,[68] and in 2019 in archaea that can oxidize ethane with sulfate and that are distantly related to the Methanosarcinales.[55] (See also ref (69).)

Figure 3

Phylogenetic relation of McrA amino acid sequences present in Euryarchaeota (black, red, and orange branches) and in Bathyarchaeota (blue branches). The phylogenetic tree was constructed based on a maximum likelihood algorithm considering more than 450 amino acid positions. The scale bar indicates the number of amino acid substitutions per site. Bootstrap values higher than 90% are indicated by filled circles on the corresponding branch. Black branches, sequences from methanogens and methanotrophic archaea. Red branches, sequences from ca. Syntrophoarchaeum;[54] SBU, ca. S. butanivorans; SCAL, ca. S. caldarius. Orange branch, sequence from ca. Argoarchaeum ethanivorans.[55] Blue branches indicate Bathyarchaeota-related sequences.[70] The identifiers refer to the locus tag of the gene sequences in the draft genomes. MrtA is synonymous to McrA isoenzymes II in the corresponding groups. The figure and legend are taken from Laso-Peres et al., 2016,[54] (with permission of Nature-Springer), and updated by the orange branch with the branching order taken from Chen et al., 2019.[55]

All of these MCR-containing orders mentioned above belong to the kingdom of Euryarchaeota (Figure , black, red, and orange branches). But recent findings indicate that MCR is not restricted to this kingdom because metagenomes of archaea belonging to the Bathyarchaeota were found to contain the genes for a functional MCR (Figure , blue branches).[70−72] However, outside the domain of archaea, mcr genes have not yet been found.

M. Marburgensis: The Model Organism Used for the Study of MCR

Most of what is known about the biochemistry of MCR comes from studies of isoenzyme I from M. marburgensis that thrives on a completely mineral salts medium with 80% H2 and 20% CO2 as a sole energy source. Its energy metabolism is outlined in Figure . M. marburgensis was isolated in 1978 by Georg Fuchs from the municipal anaerobic digestion plant in Marburg.[73] It grows optimally at 65 °C with a doubling time of below 2 h to a cell concentration of 2.5 g dry mass per liter[74] and contains MCR up to 10% of the cellular proteins From such grown cells, the nickel enzyme is relatively easy to purify.[33]

Figure 4

Energy metabolism of Methanothermobacter marburgensis. MFR, methanofuran. H4MPT, tetrahydromethanopterin. F420, coenzyme F420. HS-CoM, coenzyme M. HS-CoB, coenzyme B. Fd, ferredoxin. MCR, methyl-coenzyme M reductase. Yellow dot, heterodisulfide reductase- hydrogenase complex (HdrABC-MvhADG) that couples the exergonic reduction of CoM-S–S-CoB with H2 with the endergonic reduction of ferredoxin with H2 via flavin-based electron bifurcation.[42,75−78] The redox potentials, in blue, are given under standard conditions.

Inactivation of MCR When in the Cells the Redox Potential of the COM–S–S-COB/HS-COM plus HS-COB Couple Becomes Too High

To obtain active MCR, cultures of M. marburgensis grown on 80% H2 and 20% CO2 must be exposed to 100% H2 before harvest. Only then the cells contain mainly active MCR that exhibits the electron paramagnetic resonance (EPR) signal designated as MCR-red1.[32,79] To the contrary, when the cells are exposed to 80% N2/20% CO2 prior to harvest, MCR is inactive and exhibits the EPR signal MCR-ox1.[33,79] When the cells are directly harvested, the MCR is inactive and EPR-silent. An explanation for the observed effects can probably be found in the differences in intracellular concentrations of CH3–S-CoM, HS-CoM, HS-CoB, and CoM-S–S-CoB. The redox potential Eo′ of the CoM-S–S-CoB/HS-CoM + HS-CoB couple under standard conditions is −140 mV.[80]

When only H2 rather than CO2 is present, then high intracellular concentrations of HS-CoM and HS-CoB and essentially zero concentration of CoM-S–S-CoB are to be expected; in the absence of CO2, methyl-coenzyme M cannot be regenerated (Figure ); the redox potential of the CoM-S–S-CoB/HS-CoM + HS-CoB couple under these conditions will be much more negative than −140 mV. Vice versa, when the cells are gassed with CO2 in the absence of H2 before harvest, the concentrations of HS-CoM and HS-CoB will be essentially zero and that of CoM-S–S-CoB very high; in the absence of H2, the heterodisulfide CoM-S–S-CoB cannot be reduced (Figure ); the redox potential of the CoM-S–S-CoB/HS-CoM + HS-CoB couple under these conditions will be substantially more positive than −140 mV. Finally, when the cells are directly harvested, there is an intermediary situation with a redox potential of the CoM-S–S-CoB/HS-CoM + HS-CoB couple probably near −140 mV. It thus appears that MCR in M. marburgensis cells is only in its active state when the redox potential of the CoM-S–S-CoB/HS-CoM + HS-CoB couple is considerably more negative than the standard redox potential of −140 mV.

Oxidation States of Nickel in MCR

F-430 in MCR can be in the Ni(I), Ni(II), and Ni(III) oxidation state, of which the Ni(I) state (MCR-red) and Ni(III) state (MCR-ox) show characteristic EPR spectra, as first revealed in 1988 by Simon Albracht at the University of Amsterdam.[79] The Ni(II) state is EPR-silent (MCR-silent). The three states can also be distinguished by UV–visible spectroscopy.[81−83]

Active MCR in the Red1 State

M. marburgensis cells growing exponentially at 65 °C on 80% H2/20% CO2 with a doubling time of 2 h form methane at a specific rate of 10 μmol min–1 mg protein–1. Extracts of such grown cells are therefore expected to catalyze methane formation from methyl-coenzyme M and coenzyme B at a specific activity of at least 10 μmol min–1 mg protein–1. For many years, however, only specific activities below 0.1 μmol min–1 mg protein–1 were observed. The breakthrough came from the finding that cell suspensions of M. marburgensis exhibited a strong MCR-red1 signal when incubated under 100% H2[79] and that extracts of such H2-reduced cells catalyzed methyl-coenzyme M reduction with coenzyme B at specific rates near 2 μmol min–1 mg protein–1.[32] After 10-fold purification in the presence of the competitive inhibitor coenzyme M to stabilize the activity, the purified enzyme-catalyzed methane formation from methyl-coenzyme M and coenzyme B at a specific activity of about 20 μmol mn–1 mg protein–1. The purified MCR I had a greenish color with a maximum at 386 nm and exhibited the axial MRC-red1 signal (g1 = 2.25; g2 = 2.07; g3 = 2.06) with a spin concentration near 0.2. The EPR and UV–visible spectra of MCR-red1 compared amazingly well with those exhibited by the penta-methyl ester of Ni(I)F-430 (Ni(I)F-430M).[84] It was therefore concluded that in MCR-red1 the nickel is in the 1+ oxidation state. Upon inactivation of active MCR with chloroform, the spin concentration and the specific activity decreased in parallel, from which it was concluded that nickel in MCR has to be in the 1+ oxidation state to be active.[32]

The one-electron, reversible reduction potential (Em′) of Ni(II)F-430M/Ni(I)F-430 couple in dimethylformamide was found to be −504 mV versus the hydrogen electrode (0.0 V)[15,84] and that of F-430 in water to be −650 mV.[85] Oxidation of MCR-red1 with ferricyanide yielded a one-electron Nernst plot with Em = −440 mV, but the oxidation was irreversible.[82] The redox potential of the MCR-bound F-430 is probably considerably lower than those of the free F-430 M in acetonitrile due to the difference in coordination, four coordinates in free F-430M, five coordinates in MCR-red1a, and six coordinates in MCR-silent (see below).

Inactive MCR in the Ox1 State

After growth of M. marburgensis on 80% H2/20% CO2, the cultures were gassed with 80% N2/20% CO2 before cooling and harvesting; the cell extracts exhibited a strong MCR-ox1 EPR signal (g1 = 2.2310; g2 = 2.1667; g3 = 2.1532)[82] and showed essentially no MCR activity.[33,79] Instead of gassing the cultures with 80% N2/20% CO2, it is also possible to induce the MCR-ox1 signal by adding 20 mmol sodium sulfide per liter culture before the harvest; this was shown for M. marburgensis and Methanosarcina thermophila.[86] The ox1 EPR signal of inactive MCR was, as the red1 EPR signal of active MCR, quenched by O2, chloroform, or nitric oxide, but only at much higher concentrations and lower rates.[87] MCR in the ox1 state is therefore relatively easy to purify to apparent homogeneity with almost 100% of its Ni in the ox1 state. Upon reduction of MCR-ox1 in the presence of methyl-coenzyme M with titanium(III) citrate at pH 9, inactive MCR-ox1 was almost 100% converted to active MCR-red1. Specific activities of up to 100 μmol min–1 mg protein–1 were reached.[33] The requirement of the strong reductant Ti(III) for the conversion of MCR-ox1 to MCR-red1 was interpreted to indicate that in MCR-ox1 the nickel is in the 3+ oxidation state.[33]

This interpretation was subsequently challenged by the group of Steven Ragsdale, which proposed that, as in the red1 state, the nickel in MCR-ox1 is in the 1+ oxidation state.[88] One valid argument was that the EPR spectrum of the pentamethylester of Ni(III)F-430 in aprotic solvent[89] was completely different from that of MCR in the ox1 state.[33] The one-electron reversible oxidation potential of F-430 M in dimethylformamide was measured to be +625 mV.[89] In view of the high oxidation potential, four-coordinate or axially weakly coordinated forms of Ni(III)F-430 were considered unlikely to be of biological importance.[15] A counter argument was that the axial coordination of a strong donor like a thiolate or a CH3– anion to Ni(III)F-430 in MCR could change the spectra and lower the oxidation potential sufficiently.[89] Indeed, the spectra of MCR in the ox state, in which the Ni(III) in F-430 is either thiolated (CoM-S-Ni(III)F-430)[90] or methylated (methyl-Ni(III)F-430)[91,92] are, as expected, quite different from Ni(III)F-430 M in acetonitrile.[89] Another valid argument was that MCR-silent could be converted to the ox1 state by irradiation under cryo-conditions.[93]

To explain the Ti(III) reduction results, it was postulated that upon reduction of MCR in the ox1 state to the red1 state, the hydrocorphin ring of F-430 rather than the nickel was reduced by two electrons and experimental evidence for this was provided.[87,94] However, later these experimental findings were reevaluated.[95−98] There is now general agreement that in the ox1 state the thiol sulfur of coenzyme M forms a bond to N(III), in which the Ni(III) thiolate and Ni(II) thiyl radical are in equilibrium.[81,97]

Inactive MCR in the Ox2 and Ox3 States

Addition of Na2SO3 to MCR in the red2 state (see below) was found to induce the light-sensitive EPR signal designated as MCR-ox2 (g1 = 2.2263; g2 = 2.1425; g3 = 2.1285),[82] and exposure of MCR in the red2 state to O2 was found to induce the light-sensitive EPR signal MCR-ox3a (g1 = 2.2170; g2 = 2.1400; g3 = 2.1340).[82] Induction of the two ox signals is irreversible in the sense that removal of Na2SO3 or O2 did not restore the red2 state. The light sensitivity discriminates MCR-ox2 and MCR-ox3 from MCR-ox1, which is not light-sensitive.[82] The nickel in MCR-ox2 and MCR-ox3 is, as in MCR-ox1, most likely also in the 3+ oxidation state although this has not been analyzed in detail.

Cytoplasmic Location and Isoenyzmes

Cytoplasmic Location

Although MCR is found primarily in the soluble cell fraction, whole-cell immune-labeling experiments with Methanosarcina(99) and Methanothermobacter(100,101) have indicated that the enzyme is membrane associated. (See also ref (102).) Localization close to the cytoplasmic membrane was confirmed recently.[103] However, none of the genes associated with MCR synthesis are predicted to code for proteins that have a membrane anchor. Localization close to the membrane is therefore probably by another mechanism that remains to be elucidated.

Isoenzymes

M. marburgensis contains two MCR isoenzymes that are differently expressed in different growth phases: MCR I is synthesized when, during growth on H2 and CO2, the H2 supply is limiting at high cell concentrations and MCR II when H2 is repleted.[104−106] The two isoenzymes can be separated by chromatography on anion exchange resins; the one that eluted first was designated as MCR II and the one that eluted second MCR I. At the gene level, the two isoenzymes are referred to as mrt (MCR II) and mcr (MCR I).[106] The two isoenzymes, which both exhibit a ternary complex kinetic mechanism, were found to mainly differ in their KM for methyl-coenzyme M and coenzyme B and their pH optimum (see below).[107] Almost all of the experiments to resolve the catalytic mechanism were performed with MCR I because this isoenzyme is present in the highest concentrations in the archaeon at the end of exponential growth when the H2 supply becomes limiting and the cells are harvested.

M. marburgensis is not the only methanogen that contains two MCR isoenzymes. Also many other members of the Methanobacteriales, many members of the Methanococcales, and a few members of the Methanomicrobiales contain two MCR isoenzymes. Phylogenetic analyses on the basis of all three subunits grouped MCRs from Methanobacteriales and Methanococcales into three distinct types: (i) MCRs found only in Methanobacteriales; (ii) MCRs found in Methanobacteriales and Methanococcales; and MCR found only in Methanococcales. The first and second types include the MCR isoenzymes I and II, respectively, from M. marburgensis.[61] The three types are structurally highly similar with respect to the overall fold and active site architecture including the structure and binding mode of F-430 and the substrates. However, the MCRs significantly differ regarding the electrostatic surface potentials (which is why they can be separated by anion exchange chromatography), the loop architectures, and in particular the C-terminal end of their McrG subunit that interact with the McrA and McrB subunits in a different manner, which could affect the kinetic properties.[61]

Members of the Methanosarcinales, Methanomassiliicoccales, Methanocellales, and the ANME 1–3 groups appear to contain only one type of MCR, but in the archaea that catalyzes the oxidation of butane four sets of mcr genes were found.[54]

Structural Features

Subunit Composition

MCR I is composed of three different subunits α (McrA), β (McrB), and γ (McrG), each being present twice. Per mol mass of up to 300 kDa, the enzyme contains 2 mol of the nickel hydrocorphin F-430 tightly but not covalently bound. The genes for the three subunits were found in the transcription units mcrBGCDA (isoenzyme I) and mrtBGDA (isoenzyme II).[108−110] McrC and McrD are not required for MCR activity[111] but apparently are required for post-translational assembly (McrD)[112] and activation of the enzyme (McrC).[113]

MCR from acetate-grown Methanosarcina thermophila was found to be an α1β1γ1 trimer with a native molecular mass of between 132 000 and 141 000 Da and to contain 1 F-430 bound.[114] MCR from M. thermophila thus appears to differ from all other investigated MCRs, which are α2β2γ2 hexamers. Because of the intertwined hexameric structure (see below), a dissociation of the hexamer into two trimers is difficult to envisage. It would be therefore very interesting to obtain a crystal structure of this MCR.

Coenzyme B Bound to an UDP-Disaccharide?

Evidence was published in beginning of the 1990s that showed coenzyme B (Figure ) is covalently bound to an UDP-disaccharide through a carboxylic-phosphoric anhydride linkage and that this conjugated coenzyme B might actually be the real electron donor for methyl-coenzyme M reduction to methane.[115,116] Since the cell extracts used to test this proposed substrate had only specific activities of methane formation in the nmol min–1 mg protein–1 range, the validity of the results is difficult to judge. In crystal structures of MCR with coenzyme B bound, such a carbohydrate moiety was carefully looked for but never observed (see below).

Crystal Structures of MCR from Methanogens

In collaboration with Ulrich Ermler at the Max Planck Institute for Biophysics in Frankfurt, Seigo Shima in Marburg obtained in 1997, two crystal structures of inactive MCR I from M. marburgensis. The crystal structure of MCR-ox1-silent with coenzyme M and coenzyme B bound (Figure A) was resolved to 1.45 Å and the one of MCR-silent with CoM-S–S-CoB bound (Figure B) to 2 Å[117] and later refined to 1.16 and 1.8 Å, respectively.[118] A third structure, which MCR-red1-silent refined to 1.8 Å, was without any substrates or products bound.[118]

Figure 5

Schematic drawing of the active site of inactive methyl-coenzyme M reductase (MCR) with F-430 in the Ni(II) oxidation state based on crystal structures up to 1.1 Å resolution of isoenzymes I from M. marburgensis. (A) MCR-ox1-silent, MCR in complex with coenzyme M (CoM-SH) and coenzyme B (CoB-SH). (B), MCR-silent, MCR with the heterodisulfide CoM-S–S-CoB bound. Sulfur in green, carbon yellow, nickel in blue, and oxygen in red. From Mahlert et al., 2002.[83]

The three structures confirmed that the different subunits are arranged in an α2,β2,γ2 configuration and revealed two identical interconnected F-430-harboring active sites separated by 50 Å. The two F-430s are embedded deeply between the subunits α,α′,β,γ and α′,α,β′,γ′. Glutamine α′147 provides the lower axial ligand to Ni(II) in F-430 that is mainly bound to the subunits α, β, and γ, and glutamine α147 provides the lower axial ligand to Ni(II) in F-430 that is mainly bound to the subunits α′, β′, and γ′. Each active site is accessible from the surface only through a 50 Å long channel, which is 25 Å in diameter at the protein surface and narrows to 8 Å for the last 16 Å, which leads to a pocket at the base where F-430 binds. Through this channel, methyl-coenzyme M reaches F-430 and coordinates there axially with its thioether sulfur to the nickel from the front side. The channel is tightly locked after binding of the second substrate coenzyme B.[117,118] All attempts to obtain a crystal structure of active MCR failed until now because the Ni(I)F-430 in the active site always autoxidized to Ni(II) during the crystallization procedure even when performed in an anaerobic chamber containing O2-free gas (95% N2 + 5% H2).

Crystal structures were later also published by the Ragsdale group. MCR-red1-silent structures with coenzyme B analogues (CoB4SH to Cob8SH) were obtained that allowed to map the 50 Å long channel.[119] A structure of MCR, in which the Ni(I) of F-430 is methylated by methyl iodide, forming a methyl-Ni(III)bond, showed no significant differences in the conformation of the active site as compared to the three MCR-silent structures, in which the Ni of F-430 is in the Ni(II) oxidation state.[120]

Comparison of the crystal structures of MCR from M. marburgensis with MCRs from the phylogenetically distant Methanosarcina barkeri, Methanopyrus kandleri,[121]Methanotorris formicicus, and Methanothermococcus thermolithotrophicus(61) revealed highly similar overall structures and virtually an identical active site architecture, reflecting the chemical challenging mechanism of methane formation. Pronounced differences were found, however, at the protein surface, loop geometries and electrostatic properties, which also involve the entrance of the active site funnel.[61]

Structure of MCR from Nonmethanogens

Also an X-ray structure of the 280 kDa heterohexameric MCR complex from an ANME-1 archaeon was obtained.[60] The enzyme complex was crystallized uniquely from a protein ensemble purified from consortia of microorganisms collected with a submersible from a Black Sea mat catalyzing AOM with sulfate.[38] Crystals grown from the heterogeneous sample diffracted to 2.1 Å resolution and consisted of a single ANME-1 MCR population, demonstrating the strong selective power of crystallization. The structure revealed ANME-1 MCR in complex with coenzyme M and coenzyme B, indicating identical substrates for MCR from methanotrophic and methanogenic archaea. Differences between the highly similar structures of ANME-1 MCR and methanogenic MCR include a F-430 structural modification with a methylthio group (R= S-CH3 in the Abstract graphic),[122] a cysteine-rich patch in McrA and a somewhat altered post-translational amino acid modification pattern. However, the active site thioglycine and 1-N-methylhistidine were conserved (see below). The differences may tune the enzymes for their functions in different biological contexts.[60] In support of this possibility is the finding that recombinant Methanosarcina acetivorans containing mcr genes from an ANME-1 organism were more efficient in oxidizing methane with Fe(III) than the wild type.[123] Note, however, that MCR from ANME-2 and ANME-3 archaea contain normal coenzyme F-430.[122]

A screen for F-430 variants in cell extracts of various methanogens and ANME organisms revealed the presence of multiple modified F-430 structures. A total of nine modified F-430 structures were identified. It was not determined, whether these F-430 variants are associated with MCR or whether they may have another function is still to be determined.[124]

Unfortunately, only enrichment cultures of butane or ethane-oxidizing archaea are available to date. Therefore, all that is known of MCR from these archaea comes from metagenomes studies, by which the primary structure of the MCRs can be deduced. From homology modeling from the primary structures using the crystal structures of MCRs from methanogens, it was found that the MCR from butanotrophic and ethanotrophic archaea have overall three-dimensional structures very similar to those of MCRs from methanogenic archaea.[54,55] Differences were observed in the amino acid sequence of the catalytic pocket. However, these differences do not yet explain why MCR from ANME archaea appear to specifically catalyze the oxidation of methane, whereas the MCR from butane- or ethane-oxidizing archaea appear to specifically catalyze the oxidation of butane or ethane, respectively.[54,55]

Post-Translational Modifications

The subunit McrA of MCR I from M. marburgensis was shown to contain five post-translational modifications of active site amino acids: a thioglycine, 1-N-methyl-histidine, S-methyl-cysteine, 5-methylarginine, and 2-methy-glutamine.[117] The methyl group in the modified amino acids was shown to be derived from the methyl group of methionine most probably via S-adenosylmethionine.[125] Recently a sixth novel post-translationally modified amino acid, namely, didehydroaspartate, was found adjacent to the thioglycine as revealed by mass spectrometry and high-resolution X-ray crystallography. Upon chemical reduction, the didehydroaspartate residue was converted into aspartate.[61]

Most of the six modifications were found also in MCRs from other methanogens.[60,61,121,126] Especially the thioglycine and the 1-N-methylhistidine appear to be highly conserved, the latter probably because this modification is required for correct coenzyme B binding.[117] MCR III from M. formicicus lacks the S-methylcysteine but possesses in the α subunit a 6-hydroxytryptophane. Instead of a 2-methyl-glutamine there is a glutamine in McrA from M. barkeri and where there is a didehydroaspartate there is an aspartate in McrA from Methanothermobacter wolfei.(61)

In Methanosarcina, the genes responsible for the post-translational modification of the glycine to thioglycine[127−129] and for the conversion of the active site arginine to 5-methylarginine[130] were identified and individually deleted. The deletions were found to negatively affect growth of the methanogens, however, only under stress conditions. A role of the two highly conserved post-translational modifications in stabilizing the protein’s secondary structure near the active site was proposed.[127,130] However, the specific activity and kinetic properties of the enzyme variants were not determined. Whether the variant enzyme still shows full activity cannot be deduced from the in vivo growth experiments, as has been proposed, because it is not known whether in vivo the reduction of methyl-coenzyme M to methane via MCR was growth-rate limiting, which it most probably was not. Indeed, at the low-energy substrate concentrations prevailing in the natural habitats of methanogens, their energy metabolism is generally growth-rate limiting.[131] However, this is not the case at the much higher methanogenic substrate concentrations used in the laboratory to grow the archaea in batch cultures. This is evidenced by the fact that the growth rate and growth yield of Methanosarcina species in mineral salts media are increased when, for example, yeast extract and peptone are added, indicating that biosynthesis rather than the energy metabolism was growth-rate limiting. An interpretation of the mutation results thus has to await comparative measurements of the enzyme’s catalytic efficiency (kcat/KM) and activation energy (temperature dependence of activity) and determination of whether the variant MCR still shows half-of-the-sites reactivity (see below).

Post-Translational Assembly

When the mcrBGCDA genes from the thermophile Methanothermococcus okinawensis were expressed in the mesophile Methanococcus maripaludis, the purified recombinant rMCRok was found to be post-translationally modified correctly and to contain F-430 and McrD,[112] which in Methanococcus species is associated with the McrABG complex.[132] Subunits of the native M. maripaludis (MCRmar) were largely absent in the purified enzyme, even though MCRmar was produced in the cells, suggesting that the recombinant enzyme was formed by an assembly of cotranscribed subunits. Strong support for this hypothesis was obtained by expressing in M. maripaludis a chimeric operon, comprising the His-tagged mcrA from M. maripaludis and mcrBDCG from M. okinawensis. Indeed, the His-tagged MCR was found to contain M. maripaludis McrA and M. okinawensis McrBDG.[112]

Catalytic Properties and Inhibitors

The amino acid sequences of McrA, McrB, and McrG of isoenzymes I from M. marburgensis, which belongs to the Methanobacteriales, are significantly different from those of MCRs from other methanogens and methanotrophic archaea (Figure ). Even the two isoenzymes of M. marburgensis share only about 70% sequence identity.[61] It is well-known that the substrate specificity and catalytic efficiency of phylogenetically closely related enzymes can differ. These properties are reliably known only for MCR I from M. marburgensis. It could therefore well be that MCR even from closely related methanogens differs significantly in catalytic properties.

Purified MCR I and MCR II from M. marburgensis in the active red1 state are difficult to obtain because in the presence of trace amounts O2, which are difficult to exclude completely during purification, the red1 EPR signal is quenched. The Ni(I)F-430 in the active site is somewhat protected from oxidation in the presence of the MCR substrate methyl-coenzyme M or the substrate analogue coenzyme M.[133] However, even in the presence of methyl-coenzyme M or coenzyme M, loss of activity in the presence of O2 is still rapid, which is why purifications have to be performed in anaerobic chambers containing O2-free gas (95% N2 + 5% H2).

MCR I-red1, obtained from purified MCR-ox1 by reduction with Ti(III) in the presence of methyl-coenzyme M, catalyzed at 60 °C, the reduction of methyl-coenzyme M (apparent KM = 0.7 ± 0.2 mM) with coenzyme B (apparent KM = 0.2 ± 0.1 mM) to methane and CoM-S–S-CoB at a specific activity of up to 100 μmol min–1 mg protein–1 (apparent Vmax).[33,107] MCR II-red1 differed in its kinetic properties from MCR I-red1: the apparent KM for methyl-coenzyme M was 1.4 ± 0.2 mM and that for coenzyme B was 0.5 ± 0.2 mM. The pH optimum of MCR II was 7.5–8.0 as compared to 7.0–7.5 in the case of MCR I.[33,107]

When MCR I-red1 was purified in the presence of coenzyme M, the apparent KM for methyl-coenzyme M was 5 mM rather than 0.7 mM and the apparent Vmax was considerably lower than 100 μmol min–1 mg protein–1, most probably because the preparations still contained coenzyme M that is an efficient reversible MCR inhibitor, inhibition being competitive to methyl-coenzyme M.[133]

Substrate Specificity

MCR I purified in the presence of coenzyme M also catalyzed the reduction of ethyl-coenzyme M (apparent KM = 20 mM), albeit at a specific activity of only 0.1 μmol min–1 mg protein–1. The enzyme did not catalyze the reduction of propyl-coenzyme M, of allyl-coenzyme M and of trifluoromethyl-coenzyme M.[133] Besides coenzyme B (7-mercaptoheptanoyl-threonine phosphate), also 6-mercaptohexanoyl-threonine phosphate (HSCo6B) could serve as an electron donor; however, the rates of methyl-coenzyme M reduction with HSCo6B were less than 1% of the rates with HSCo6B.[134,135] The slow substrate CoB6–SH has been a valuable analogue in helping decipher the MCR catalytic mechanism.[136]

Stereospecificity

The reduction of ethyl-coenzyme M proceeds with inversion of stereoconfiguration as determined with an isotopically chiral form of ethyl-coenzyme M in cell extracts of Methanosarcina barkeri.[137] Studies with purified enzyme from M. marburgensis, in which the incorporation of deuterium from D2O into the ethyl group of ethyl-coenzyme M was followed, are consistent with the inversion of stereoconfiguration.[138]

Cell extracts of M. thermoautotrophicus catalyzed, beside the reduction of methyl-coenzyme M and ethyl-coenzyme M with H2, also the reduction of monofluoromethyl-coenzyme M, difluoromethyl-coenzyme M, methyl-selenocoenzyme M, and 3-methylthiopropionate. The relative rates observed are difficult to interpret because the specific activities even with methyl-coenzyme M were very low, and these low specific activities were only obtained in the presence of ATP, which was required to activate the two isoenzymes in the cell extracts.[139−141]

Reversibility

MCR I was shown to catalyze the back reaction, the oxidation of methane, with a specific activity of near 10 nmol min–1 mg protein–1 at 60 °C and 1 bar 13CH4 (corresponding to about 1 mM methane at 60 °C), which is only about 0.01% of the specific activity of the forward reaction.[36] The rate of methane oxidation increased with the methane concentration linearly up to 2 mM, the highest concentration reached experimentally, indicating that the apparent KM for methane is way above 2 mM. The kinetic properties of MCR from M. marburgensis are compatible with MCR being the enzyme that in the very slowly growing ANME organisms catalyzes the anaerobic oxidation of methane.[36]

The specific rate of methane oxidation (10 nmol min–1 mg protein–1) (V–1) is low relative to that of methane formation (100 μmol min–1 mg protein–1) (V+1): V+1/V–1 = 104. The Haldane equation,[142] which relates the free energy change (ΔGo) of −30 kJ/mol associated with the MCR-catalyzed reaction (Figure ) to the reaction’s equilibrium constant, predicts that the catalytic efficiency (Vmax/KM) of the forward reaction (methane formation) divided by the catalytic efficiency of the back reaction (methane oxidation) should be somewhat above 105. Since the Vmax and KM for methane oxidation are not yet known, it cannot be decided whether V+1/V–1 = 104 is a good match. Also, it has to be considered that ΔGo = −30 kJ/mol is only known with some uncertainty.

ΔGo = 30 kJ/mol was calculated from the ΔGo of three reactions, namely, (i) methanol reduction with H2 to methane (ΔGo = −112.5 kJ/mol),[131] (ii) CoM-S–S-CoB reduction with H2 to HS-CoM and HS-CoB (ΔGo = −53 kJ/mol), and (iii) methyl-S-CoM formation from methanol and coenzyme M (ΔGo = −26.5 kJ/mol).[143] (Methanol, methyl-S-CoM, HS-CoM, HS-CoB and CoM-S-S-CoB in aqueous solution at 1 mol per kg and H2 and methane in the gaseous state at 1 atm pressure). In turn, the ΔGo for the CoM-S–S-CoB reduction reaction was calculated from the redox potential Eo′ = −140 mV of the CoM-S–S-CoB/HS-CoM + HS-CoB couple.[80] Using these values, the free energy change associated with the MCR-catalyzed reaction is −33 kJ/mol. Before 2003, the redox potential of the CoM-S–S-CoB/HS-CoM + HS-CoB couple had been assumed to be near −200 mV, leading to a calculated free energy change of the MCR-catalyzed reaction of −45 kJ/mol.[144] From ΔGo = −45 kJ/mol, a Keq = k+1/k–1 of almost 108 is calculated.

ΔGo associated with ethane formation from ethanol and H2 (−88.3 kJ/mol) is by 24 kJ/mol less exergonic than ΔGo associated with methane formation from methanol and H2 (−112.5 kJ/mol).[131] However, ΔGo associated with ethane formation from ethyl-S-CoM and HS-CoB is not by 24 kJ/mol less exergonic than ΔGo of methyl-S-CoM reduction to methane with HS-CoM (ΔGo = −33 kJ/mol) because ethyl-S-CoM formation from ethanol and HS-CoM is expected to be less exergonic than methyl-S-CoM formation from methanol and HS-CoM (how much was not found in the literature). Based on the bond strength difference H-CH3 (438.9 kJ/ml) versus H-CH2CH3 (423.0 kJ/mol) the ΔGo of ethane formation from ethyl-S-CoM and HS-CoB is estimated to be near −20 kJ/mol.

Reversible Inhibitors

Many MCR substrate analogues and homologues have been tested to probe the accessibility and reactivity of the Ni(I) in F-430 of MCR I.[91,119,133] Coenzyme M (app Ki = 4 mM), propyl-coenzyme M (app Ki = 2 mM), 2-azidoethane-coenzyme M (app Ki = 1 μM), and allyl-coenzyme M (app Ki = 0.1 mM) were found to be reversible inhibitors, inhibition being competitive to methyl-coenzyme M.[111,133] Reversible inhibition was also observed in the presence of diverse coenzyme B homologues and analogues.[119,134,135]

Irreversible Inhibitors

In 1978, 2-bromoethanesulfonate (BES) was found to be an effective inhibitor of methanogenesis,[139,140] which today is still widely used in ecological studies of the methane cycle. BES (app Ki = 2 μM) competes reversibly with methyl-coenzyme M in binding within the active site of MCR, but once it is bound, it alkylates the Ni(I) to a labile alkyl-Ni(III) derivative that collapses, yielding ethylene,[133,145,146] Also the inhibition of MCR with 3-bromopropanesulfonate (BPS) is irreversible. BPS (app Ki = 0.1 μM),[111] like BES, competes reversibly with methyl-coenzyme M in binding within the active site, but once it is bound, it alkylates the Ni(I) to a stable alkyl-Ni(III) derivative.[147,148] A stable alkyl-Ni(III) is also generated by brominated carboxylic acids with carbon chain length 5–16.[91] C1- and C2-polychlorinated hydrocarbons[149] are also irreversible inhibitors that all can easily enter the active site and oxidize the N(I) there.

An inhibitor used in ecological studies is 3-nitrooxypropanol (3-NOP) that oxidizes the N(I) in the MCR active site.[150] The nitroester is reduced by the Ni(I) in the active site of MCR to 1,3-propanediol and nitrite; the latter also reacts with Ni(I) in the active site, which is why 3-NOP is classified as a dual warhead inhibitor. 3-NOP is presently tested as an inhibitor of methanogenesis in the rumen of cattle and sheep.[151] 3-NOP has the advantage to not be charged and to thus be able to permeate into the methanogenic cells without requiring a transporter. E.g., in the case of BES, only those methanogens are sensitive to this inhibitor, which require coenzyme M as a vitamin and have a transport system for coenzyme M that also transports BES into the cells. Although BPS (app Ki = 0.1 μM) is a much better inhibitor of MCR than BES (app Ki = 2 μM) in vitro, it does not affect the enzyme in vivo because the coenzyme M transporter does not appear to transport BPS.

Whereas nitrite inactivates MCR at very low concentrations, nitrate is without any effect on MCR.[150] Nitrocompounds such as nitroethane, 2-nitroalcohol, and 2-nitro-l-propanol at mM concentrations inhibit methanogenesis in vivo,[152] whether they attack MCR is not known. There are a number of other compounds that inhibit methanogenesis at a low concentration in vivo, which might target MCR. Of these, ethylene (ethene) is the most interesting[153] because if the Ni(I) in MCR is able to add to the double bond of the alkene then this would be a sign of a strong one-electron reactivity of the Ni(I): Alkenes are known to react with free radicals and nucleophilic metals.[154]

Coenzyme B-Dependent Inactivation

The rate of inactivation by BES and BPS is stimulated by coenzyme B. At very low concentrations of BES, inactivation is dependent on the presence of coenzyme B. Also several fold stimulated by coenzyme B is the irreversible inhibition of MCR by cyano-coenzyme M, seleno-coenzyme M, and trifluoromethyl-coenzyme M.[133] These findings are of great importance for the catalytic mechanism because it shows that, upon coenzyme B binding, the nucleophilicity of Ni(I) in F-430 is considerably increased.

Proposed Catalytic Mechanisms

Kinetic analyses revealed the two MCR isoenzymes from M. marburgensis to have a ternary complex kinetic mechanism (most probably an ordered bibi ternary complex mechanism) rather than a ping pong mechanism: Double reciprocal plots of initial rates versus the concentration of either one of the two substrates at different constant concentrations of the other substrate were linear and intersected on the abscissa to the left of the 1/v axis.[107] Thus, both substrates, methyl-coenzyme M and coenzyme B, have to bind after another to MCR before the first product, presumably methane, is released. From the MCR structure, it is not evident how methane gets out of or in the active site: a convincing gas channel has not been identified.

SN2 Proposal

The mechanism long favored by the author of this Perspective assumed methyl-Ni(III)F-430 and methyl-Ni(II)F-430 as intermediates in the MCR catalytic cycle, the methyl-Ni(III) being formed by a nucleophilic attack by Ni(I) of the methyl group of methyl-coenzyme M in an SN2 reaction.[117,118] The proposal was based on the finding that 2 mol pentamethylester of Ni(I)F-430 (Ni(I)F-430M) reacts with 1 mol electrophilic methyl donors to 1 mol methane and 2 mol Ni(II)F-430 M[155] and that paramagnetic methyl-Ni(II)F-430 M can be trapped as likely an intermediate.[156] Ni(I)F-430 M is thus a nucleophile similar to cob(I)alamin with the difference that methyl-Ni(III)F-430 M is much easier to reduce to methyl-Ni(II)F-430 than methyl-cob(III)alamin to methyl-cob(II)alamin: the F-430-catalyzed reduction of methyl chloride with Ti(III) is almost 50 times more rapid than the cobalamin-catalyzed reaction and does not generate ethane as a side product.[157,158] The SN2 proposal was also based on the property of cobalamin-dependent methionine synthase to catalyze the formation of methyl-cob(III)alamin from methionine, which is also a methylthioether (ΔG0′ = +30 kJ/mol).[159,160] An important mechanistic argument was that both reactions proceed with inversion of stereoconfiguration of the methyl group[144]

Experimental support for the SN2 mechanism came from the laboratory of Steven Ragsdale and of Evert Duin reporting that Ni(I)F-430 in MCR can be alkylated with methyl bromide or methyl iodide to methyl-Ni(III)F-430,[91,161] the structure of which was elucidated by X-ray absorption spectroscopy (XAS)[162] and X-ray structure analysis.[120] The methyl-nickel complex in MCR was found to react with coenzyme M to methyl-coenzyme M and, e.g., with CoB8–SH, to Cob8–S-CH3, regenerating the MCR-red1 state[163] and with Ti(III) to methane and MCR-red1.[91,92] Observed radicals were interpreted to indicate that the methyl-Ni(III) bond is homolyzed, generating a methyl radical and Ni(II)F-430.[146] By transient kinetic methods, the decay of the active Ni(I) state coupled to the formation and subsequent decay of alkyl-Ni(III) and organic radical intermediates were observed at catalytically competent rates. These observations were considered to provide support for a MCR catalytic mechanism that involves methyl-Ni(III) and an organic radical as catalytic intermediates.[164]

However, the formation of the proposed methyl-Ni(III) intermediate from methyl-coenzyme M and MCR-red1 was never observed. This is in agreement with predictions from hybrid density function theory (DFT) computations performed in the laboratory of Per Siegbahn at the Stockholm University.[165−169] The computations predict that alkylation of the Ni(I)F-430 with the methyl thioether is too endergonic to allow the SN2 reaction as part of the catalytic cycle. Also, a proposal involving protonation of the thioether’s sulfur of the CH3–S-CoM substrate, in analogy to cobalamin-dependent methyltransfer reactions, which should facilitate methyl-Ni(III)F-430 formation, was disfavored by DFT computations since the substrate has a much smaller proton affinity than the F-430 cofactor.[168]

Oxidative Addition Proposals

In collaboration with the group of Bernhard Jaun at the ETH Zürich, the Thauer laboratory could demonstrate that in 100% D2O during methyl-coenzyme M reduction with coenzyme B to DCH3 a significant amount of CH2D–S-CoM was formed.[170,171] This finding allowed proposing that a σ-alkane-nickel complex could be formed by oxidative addition as an intermediate. However, based on DFT computations, this mechanism was predicted to be unfavorable for the MCR reaction, due to the large endothermicity for the formation of the ternary intermediate with side-on C–S (for CH3–S-CoM) or CH (for methane) coordination of nickel.[168] In a different version of an oxidative addition mechanism,[172] the reaction is initiated by protonation of one of the ligands: nitrogen or Ni(I) in F-430. DFT computations indicated that also this mechanism is most probably energetically inaccessible for the MCR reaction.[168]

Siegbahn Proposal

On the basis of DFT computations using a truncated model of the active site derived from the crystal structure of the inactive Ni(II)MCR-ox1-silent form, the Siegbahn group proposed an alternative catalytic mechanism, in which the Ni(I) of F-430 in the active site attacks the thioether sulfur rather than the methyl group of coenzyme M, yielding coenzyme M with its thiolate sulfur bound to Ni(II)F-430 and a transient methyl radical that subsequently reacts with the thiol hydrogen of coenzyme B, generating methane and a coenzyme B thiyl radical (Figure ).[165−167]

Figure 6

Transition state 1 according to the Siegbahn proposal.[167] Two bonds are broken (C–S and S–H, drawn blue), and two bonds are formed (Ni–S and C–H, drawn red) in a single step. F-430 is depicted as “Ni”; R mimics CoM, and R′ mimics CoB. OAr depicts phenol groups used in the model for the tyrosine residues. From Scheller et al., 2013.[16]

The group of Stephen Ragsdale recently provided experimental evidence for the transient methyl radical proposal by showing in single turnover experiments that the proposed high spin, EPR-silent Ni(II)thiolate is a kinetically competent intermediate.[136] Therefore, currently the “Siegbahn mechanism” appears to be the most acceptable one for MCR, which is why in the following this mechanism will be the basis for the discussion of the experimental results. Note that a catalytic mechanism cannot be proven but only disproven, and a mechanism has to explain all obtained experimental results, which at present the transient methyl radical proposal does not, as will be outlined in the next section.

Individual Steps of the Catalytic Cycle

The catalytic cycle begins with active MCR-red1a (“a” for absence of substrate) (Figure ). Methyl-coenzyme M binds (step 1) followed by coenzyme B binding (step 2). The rate-determining endothermic step is most probably the reductive cleavage of the C–S bond of methyl-coenzyme M (step 3). In this first transition state, a planar methyl radical is positioned between the Ni(II) thiolate and the thiol hydrogen of coenzyme B, from which the methyl radical abstracts the hydrogen generating methane and a coenzyme B thiyl radical (step 4). Methane is assumed to be released in step 5 and the coenzyme thiyl radical reacts with the Ni(II) thiolate to generate the disulfide anion radical in an endothermic reaction (second transition state) (steps 5 and 6). Finally, the disulfide anion radical transfers its surplus electron to Ni(II) regenerating the Ni(I) oxidation state and forming CoM-S–S-CoB (step 7), which leaves the enzyme regenerating the active enzyme without substrates bound (step 8). Note, it could also well be that methane is the last to leave having to wait until the heterodisulfide has diffused out making the way free.

Figure 7

Schematic drawing of the catalytic cycle of methyl-coenzyme M reductase (MCR). The Ni(I) oxidation states are highlighted in greenish color, the Ni(II) oxidation states in yellow, and in the Ni(III) oxidation states in orange. TC, ternary complex. TS, transition state. IM, intermediate. *, increased reactivity of Ni(I)F-430 as a result of coenzyme B binding. k+1 = rate constant of the forward reaction (methane formation) (in black); k–1 = rate constant of the back reaction (methane oxidation) (in dark blue); k–2 = rate constant of the formation of CH3–S-CoM from sterically contained methane (in red). The Ni(I) oxidation states exhibit characteristic EPR spectra: MCR-red1a, MCR in the absence of substrates; MCR-red1m, MCR in the presence of methyl-coenzyme M; MCR-red1c, MCR in the presence of coenzyme M; MCR-red1/2, MCR in the presence of coenzyme M and coenzyme B. In MCR-red1/2, part of the enzyme is in a Ni(III) hydride state (Figure ). In the MCR-ox1 state, the thiol sulfur of coenzyme M is bound to the Ni(III) of F-430 (Figure ). For details, see the text.

Figure 9

EPR spectra of methyl-coenzyme M reductase (MCR) with coenzyme M bound (MCR-red1c) and after additional binding of coenzyme B (MCR-red1/2). At saturating coenzyme B concentrations, the MCR-red1/2 spectrum is composed of 50% of the EPR spectrum MCR-red1cc, which is very similar to the EPR spectrum MCR-red1c and 50% of the MCR-red2, which in turn is composed of the axial MCR-red2a spectrum and the rhombic MCR-red2r spectrum. The active site nickel coordinations of MCR-red1cc, MCR-red2a, and MCR-red2r deduced from the spectra rather than the spectra are shown. Note that for MCR-red2a an alternative structure has been proposed, in which the acidic CoM-SH proton is making a hydrogen bond to the Ni ion.[168] The color of MCR in the red1/2 state is greenish orange and that of MCR in the red1 cm3 state is greenish.[83,170,176] MCRs with nickel in the 3+ oxidation state have been shown to have a more orange than yellow color.[82] MCR-red2r is expected to have a greenish color similar to that of other MCRs with nickel in the 1+ oxidation state.

Figure 12

Reductive activation of inactive MCR-silent and MCR-ox1 to MCR-red1a. The active site nickel coordination deduced from the EPR spectra is shown. For coordination of CoM-S–S-CoB via the sulfonate group of CoM-SH in MCR-silent, see Figure B. A2 is a 60 kDa protein with two ATP-binding sites and A3a a 700 kDa nonhomogeneous iron–sulfur flavoprotein complex composed of several different proteins. The electrons can be provided by dithiothreitol.

Where indicated in Figure , MCR-red1c, MCR-silent, and MCR-ox1 branch off from the catalytic cycle. In the case of MCR-red1c, this is when the methanogenic cells are exposed to high H2 and relatively low CO2 concentrations (high intracellular HS-CoM and low CoM-S–S-CoB concentrations). In the case of MCR-silent and MCR-ox1, this is when the cells are exposed to low H2 and relatively high CO2 concentrations (low intracellular HS-CoM and high CoM-S–S-CoB concentrations).

Active and inactive MCR states in the catalytic cycle are based on steady state and presteady state kinetics, spectroscopic analyses, isotope effect measurements, and DFT computations. The EPR and isotope effect studies in the author’s laboratory were performed in close collaboration with the groups of Bernhard Jaun at the ETH Zürich and Evert Duin, now at Auburn University, Alabama.

Active MCR without Substrates

The catalytic cycle starts with MCR without any substrates bound. The enzyme has a greenish color (λmax = 385 nm) and exhibits an axial Ni(I)-derived EPR signal (g1 = 2.252; g2 = 2.070; g3 = 2.061) designated as MCR-red1a (“a” for absent of substrates) derived from five coordinated Ni(I) in F-430 with an open upper nickel coordination site.[162,173] Double integration of the signal of the fully active enzyme revealed that both active sites contain F-430 in the reduced form, the spin concentration per Ni being in most cases approximately 0.9. The red1 signal shows a superhyperfine splitting due to the interaction of the electron of Ni(I) with the nuclear spin of the four nitrogens of the tetrapyrrolic ring system. The superhyperfine splitting is clearly resolved when the active enzyme is in the absence or presence of its substrates.[135] In fully active MCR (100 U/mg protein), the Ni(I) spin concentration is 1.0.[33]

Binding of the First Substrate

The first substrate to bind to MCR in the red1a state is methyl-coenzyme M (Figure , reaction 1). Binding results in the conversion of the Ni(I) EPR signal red1a into the EPR signal MCR-red1m (“m” for methyl-coenzyme M) (g1 = 2.2515; g2 = 2.0730; g3 = 2.0635)[82] and in a small conformational change increasing the apparent affinity of the enzyme for its second substrate coenzyme B, thus guarantying that methyl-coenzyme M binds first. This was first deduced from structural studies indicating that binding of coenzyme M, an inhibitor competitive to methyl-coenzyme M, induces specific conformational changes ensuring that methyl-coenzyme M enters the substrate channel prior to coenzyme B as required by the active site geometry.[118] The Ragsdale laboratory could show that MCR can bind both substrates independently; however, only one binary complex, that with methyl-coenzyme M, is productive, while the one with coenzyme B is inhibitory: binding of methyl-coenzyme M to the inhibitory MCR-coenzyme B complex is highly disfavored (Kd = 56 mM) relative to the binding of coenzyme B to the productive MCR-methyl-coenzyme M complex (Kd = 79 μM).[174]

Using continuous-wave and pulse electron paramagnetic resonance spectroscopy in combination with selective isotope labeling (13C and 2H) of CH3–S-CoM, it was shown that CH3–S-CoM binds in the active site of MCR such that its thioether sulfur is weakly coordinated to the Ni(I) of F-430 with a long bond length near 3.94 Å.[175] The complex is stable until the second substrate, coenzyme B, binds. Results from EPR spectroscopy, along with quantum mechanical calculations, were used to characterize the electronic and geometric structure of the complex (Figure ).[175]

Figure 8

Electron nuclear double resonance (ENDOR) structure of methyl-coenzyme M bound to Ni(I) in the active site of methyl-coenzyme M reductase, exhibiting the EPR signal MCR-red1m. The Ni–S bond length is 3.94 Å, compared with a Ni–S bond length of 2.48 Å in the proposed transition state (Figure ). From Hinderberger et al., 2008.[175]

In Figure , the S-methyl group of methyl-coenzyme M points away from the nickel and has a large degree of structural freedom, which can be deduced from the observation of broad line widths in electron nuclear double resonance (ENDOR) spectra. In view of the indirect evidence for a major structural change in the active site upon the binding of HS-CoB (see below) and the substantial degree of freedom found for the S-methyl group, the question of the binding geometry of CH3–S-CoM with respect to the nickel center in the transition state TS1 of the bond breaking step remains open. Interesting in this respect is that propyl-coenzyme M (app Ki = 2 mM) and allyl-coenzyme M (app Ki = 0.1 mM), which both inhibit MCR activity competitive to methyl-coenzyme M, induce the same red1 EPR signal as methyl-coenzyme M without being reduced.[133]

X-band hyperfine sublevel correlation (HYSCORE) spectra recorded at the echo maximum of the field-swept EPR spectrum for MCR-red1a and MCR-red1m revealed intense cross peaks at ca. 2.7 and 3.3 MHz and 3.3 and 2.7 MHz, representing the two double-quantum signals from a weakly coupled 14N nucleus. The intense 14N peaks were assigned to the NH2 of the glutamine residue coordinate to the nickel ion from below via the oxygen (Figure ).[175]

Binding of the Second Substrate

Coenzyme B can only effectively bind to MCR in the red1 state after methyl-coenzyme M is bound (reaction 2). How the enzyme can enforce a strictly ordered ternary complex mechanism[174] was described above. The reaction starts only when coenzyme B is added to the binary complex most probably because upon coenzyme B binding, via a conformational change in the nickel environment (see below), the nucleophilicity of the Ni(I) is increased such that it can react with methyl-coenzyme M. The latter is evidenced by the finding that the rate of Ni(I) quenching by 2-bromo-ethanesulfonate (BES) and other irreversible inhibitors (see above) is enhanced several fold upon addition of coenzyme B.[133] To indicate this, Ni(I) in the ternary complex in Figure is drawn somewhat out of plane with an asterisk.

When the substrate analogue and competitive inhibitor coenzyme M rather than methyl-coenzyme M is bound to the active enzyme, which in vivo is the case when in the methanogenic cells the HS-CoM and HS-CoB concentrations are high and that of CoM-S–S-CoB is low, then there is a pronounced effect on the red signal that changes from MCR-red1c (g1 = 2.2500; g2 = 2.0710; g3 = 2.0605)[82] to MCR-red1/2 (g1 = 2.2880; g2 = 2.2348; g3 = 2.1790).[82] The red1c and red1/2 forms are not intermediates in the catalytic cycle but are enzymatically active forms because they react back to MCR-red1a as soon as coenzyme M is methylated to methyl-coenzyme M again.

The red1/2 signal is composed of 50% of the EPR signal MCR-red1cc (cc for coenzyme M and coenzyme B) and 50% of the two EPR species called MCR-red2a (“a” here for axial) and MCR-red2r (r for rhombic) (Figure ).[83,170,176] The MCR-red1-type red1cc signal differs only slightly from the red1c signal. MCR-red2a has been assigned to a Ni(III) hydride complex,[170] or contrarily to a complex, in which the acidic CoM-SH proton is making a hydrogen bond to the Ni ion.[168] MCR-red2r has a Ni(I)-S coordination from HS-CoM, rhombic principal g values that are unusually high, and a very asymmetric spin density distribution on the four hydrpyrrolic nitrogens of the macrocycle F-430.[177,178]

The two MCR-red2 signals (Figure ) can also be induced by the S-methyl- and the S-trifluoromethyl analogues of coenzyme B.[179]19F ENDOR data for MCR-red2a and MCR-red2r induced by S-CF3-coenzyme B show that upon binding of the coenzyme B analogue, the end of the 7-thioheptanoyl chain of coenzyme B moves closer to the nickel center of F-430 by more than 2 Å as compared to its position in both the MCR-red1cc form[177,178] and the X-ray structure of the inactive Ni(II) MCR-ox1-silent form,[118] in which the sulfur of the heptanoyl arm of coenzyme B remains 8 Å from the nickel. This possible distance reduction of up to 2 Å was not captured in the DFT computations.[167,168]

The finding that MCR-red1c is able to undergo such a large conformational change upon binding of coenzyme B can explain how binding of the second substrate can force methyl-coenzyme M and Ni(I) of the prosthetic group to interact in the ternary enzyme complex.[177,178,180] The results also help explain the dramatic change in the coordination environment induced in the transition from MCR-red1 to MCR-red2 forms (Figure ) and opens the possibility that nickel 1+ coordination geometries other than square planar, tetragonal pyramidal, or elongated octahedral might occur in intermediates of the catalytic cycle.

Formation of the First Transition State

Based on hybrid DFT computations,[165−167] the Ni(I) in the ternary complex attacks the thioether sulfur of methyl-coenzyme M generating a transient methyl-radical and a high spin Ni(II) thiolate that is hydrogen bonded by the hydroxyl groups of two conserved tyrosines in the active site (shown in Figure but not in Figure ). A significant contribution to the transition state is also the axial coordination of the rear glutamine oxygen ligand to nickel that is a permanent ligand but is assumed to coordinate more tightly in the Ni(II) and Ni(III) oxidation states.

At first sight, the modeling of a free methyl radical in the transition state did not appear to conform to the finding that ethyl-coenzyme M reduction to ethane proceeds with inversion of configuration of the ethyl group. However, the distance of the coenzyme B thiol hydrogen to the methyl carbon radical and the low energy barrier for H abstraction make a concerted mechanism possible so that the methyl radical has no time to rotate around its y-axis before abstracting the hydrogen from the thiol group of coenzyme B and by that generating the coenzyme B thiyl radical and methane (Figures and 7).[165,166] The activation barrier for this first transition state (TS1) was calculated to be near 15 kcal/mol and for the second transition state (TS2) (see below) to be near 10 kcal/mol, considering entropy and dispersion.[167] The higher activation barrier for the first transition state implies that the first transition state is rate-limiting. The calculated transition state is supported by measured isotope effects using 13CH3–S-CoM or CD3–S-CoM as substrates.[16] The finding of a substantial 12C/13C isotope effect (1.04 ± 0.01) indicates that either cleavage of the C–S bond in methyl-coenzyme M or the formation of the C–H bond in methane occur in the rate-limiting step. The finding of an H/D secondary isotope effect of 1.19 ± 0.01 per D in the methyl group of CD3–S-CoM indicates that it is the C–S bond cleavage that is the rate-limiting step because a larger H/D primary isotope effect would be expected if formation of the C–H bond was the main contributor of the 12C/13C isotope effect. A secondary isotope effect of 1.19 per D is consistent with a geometrical change of the methyl group from tetrahedral to trigonal planar upon going to the transition state of the rate-limiting step and with an almost free methyl radical in the highest transition state.[16] The secondary isotope effect appears not to be consistent with the formation of methyl-Ni(III) in the rate-determining step.[46]

The proposed transition state 1 (Figure ) is also backed up by transient kinetic, spectroscopic, and computational approaches to study the reaction between active Ni(I) enzyme and its substrates by the Ragsdale group[136] as mentioned already above. Under single turnover conditions, CoM-S–Ni(II)F-430 was identified as an intermediate in the catalytic cycle when measured at 18 °C with CoB6SH as an electron donor. The Ni(I) decay rate (0.35 s–1) matched the methane formation rate determined with 14CH3–S-CoM (0.31 s–1) and the rate of CoM-S–Ni(II)F-430 formation (0.35 s–1) determined via MCD. Only after more than 100 s, the spectrum of active MCR-red1 returned again.[136] Why it took so long still remains to be explained. Concomitant with the Ni(I) decay, the formation of low concentrations of various radicals was observed but none of these could be assigned to the coenzyme B thiyl radical.[40] However, thiyl radicals can have a broad EPR signal that can therefore easily be overlooked.[181]

Interesting for the catalytic mechanism of methyl-coenzyme M reduction is that trifluoromethyl-coenzyme M in contrast to difluoromethyl- and monofluoromethyl-coenzyme M is not a substrate of MCR; rather it is a coenzyme B-dependent irreversible inhibitor (app Ki = 6 mM).[133] “Differences in the S–C bond strength do not explain these results, since the DFT calculation shows that the S–C bond strengths are approximately equal for the different substrates. The major qualitative difference found between CHF2–S-CoM and CF3–S-CoM is the stereoinversion barrier. For CF3, a stereoinversion barrier of 24.3 kcal/mol was obtained as compared to the value of only 6.1 kcal/mol for CHF2. In the case of the natural substrate methyl-coenzyme M, the stereoinversion barrier is obviously equal to zero, since the methyl radical is planar”.[165,166]

Sterically Contained Methane as an Intermediate

When methyl-coenzyme M reduction with coenzyme B is performed in 100% D2O both DCH3 and CH2D–S-CoM are formed, the rate of the CH3D formation (k+1) (forward reaction 5) being 25 times more rapid than the rate (k–2) of CH2D–S-CoM formation (Figure ; back reaction 4).[16,171] Since under the assay conditions the concentration of the product methane in solution was near zero, the rate of free DCH3 oxidation to methyl-coenzyme M was essentially also zero. However, even when the methane concentration is allowed to build up, e.g., to 1 mM, the rate of methane formation from methyl-coenzyme M (k+1) is 104 times more rapid than the rate of free methane oxidation to methyl-coenzyme M (k–1). The rate (k–2) of CH2D–S-CoM formation in D2O from CH3–S-CoM thus indicates that the methane generated after the first transition state is sterically hindered from leaving the active site and can thus react backward.[16,171] This is even more pronounced when ethane is formed from ethyl-coenzyme M in D2O. In this case, the k+1 to k–2 ratio was found to be 1:100, indicating that ethane was almost completely trapped within the active site. Labeling with deuterium of ethyl-coenzyme M during MCR-mediated reduction to ethane in D2O was both in the α- and β-positions. This indicates that the trapped ethane was able to freely rotate in all directions before reacting back to ethyl-coenzyme M.[138,171]

As mentioned above, binding of propyl-coenzyme M and of allyl-coenzyme M to active MCR induces an identical change in the MCR-red1a EPR spectrum as methyl-coenzyme M despite the fact that the methyl-coenzyme M analogues are not noticeably reduced to propane and propene, respectively. When propyl- and allyl-coenzyme M conversion with coenzyme B was tested in D2O, no deuterium exchange into the propyl and allyl group was observed,[16,138] indicating that propane or propene were not formed because they were trapped in the active site. The propyl and allyl radicals are much more stable than the methyl radical. The lack of exchange of deuterium into the propyl group of propyl-coenzyme M is of interest with respect to the observation that in some archaea MCR appears to be involved in the anaerobic oxidation of propane.[54]

An alternative mechanism to explain the deuterium exchange into the methyl group of coenzyme M is that methane binds to nickel as a σ-complex[171] (Figure ). Between the two mechanisms (methane physically prevented from leaving the active site for steric reasons versus methane covalently bound to nickel as a σ-complex) cannot be decided because the second transition state (TS2) is lower in energy and thus cannot be probed by kinetic isotope effects.[16] Although DFT calculations have made this oxidative addition mechanism unlikely,[168] it should not be completely excluded until it is understood what “sterically contained CH4” really is and how the measured primary and secondary kinetic isotope effects possibly affect the outcome of the DFT computations.[168] Also, the exact determination of the solvent kinetic isotope effect (H2O/D2O) is awaited (see below).

Figure 10

Isotopic exchange of deuterium from D2O into the methyl group of methyl-coenzyme M, consistent with formation of a σ-alkane-nickel complex as an intermediate. For clarity, the intermediates are drawn with the non-natural substrate ethyl-coenzyme M labeled with one 13C (drawn in red) in a deuterated medium. The ligand of coenzyme F-430 is shown schematically as bold lines. The bend in the Ni(III) intermediate symbolizes two cis coordination sites and a distorted equatorial macrocycle. The legend and figure are from Scheller et al., 2010 (reproduced with permission of Wiley-VCH Verlag GmbH & Co. KG).[171] (See also refs (16) and (138)).

Formation of the Second Transition State

Methane formation catalyzed by MCR involves at least one intermediate (sterically contained methane or possibly σ-complex) and thus two transition states (Figure ). Partitioning of the intermediate between the forward reaction and the reversal to methyl-coenzyme M is temperature-dependent with the highest forward commitment (25:1) found at a temperature of 60 °C. The commitment at 4 °C was 1.1 to 1. The temperature dependence in itself constitutes independent evidence for at least two transition states.[16] From the 25:1 ratio, the second transition state is predicted to lie at 60 °C about 10 kJ/mol lower in energy than the first one. To the contrary, in ethane formation from ethyl-coenzyme M, it is the second transition state that was found to be energetically higher by more than 10 kJ/mol at 60 °C.[138,171] As indicated above the activation barrier obtained from DFT for TS1 was near 15 kcal/mol (near 60 kJ/mol) and for TS2 it was near 10 kcal/mol (near 40 kJ/mol). Transition state 2 is therefore calculated to lie 20 kJ/mol lower than transition state 1.[167] From the observed deuterium-labeling rates of methyl-coenzyme M and methane, when the reaction was performed in D2O, the difference between the two transition states was calculated to be only about 10 kJ/mol.[16] The two methods thus yielded significantly different results, which is not surprising since the DFT calculations did not consider conformational changes and the existence of the sterically contained methane as an intermediate. Also, potentially valuable quantum mechanical calculations of the isotope effect, although important, are still missing.[168]

In transition state II, the heterodisulfide anion radical is in complex with Ni(II)F-430. Disulfide anion radicals have redox potentials more negative than −1 V,[182] enabling the disulfide anion radical to transfer its electron to Ni(II)F-430 with a redox potential more negative than −600 mV.

Enzyme Product Complex

The Ni(I) CoM-S–S-CoB complex is not very stable[113] probably because it is in equilibrium with the Ni(II) CoM-S–S-CoB radical anion complex (transition state II) (Figure ), when the product CoM-S–S-CoB builds up. The disulfide anion radical is, because of its negative one-electron redox potential below −1 V,[182] even more sensitive to oxidation than Ni(I) in Mcr-red1, oxidation yielding inactive MCR-silent with the structure shown in Figure B. Thus, when M. marburgensis growing on 80% H2 and 20% CO2 is harvested, the cells contain almost only MCR-silent because the low amount of H2 dissolved in the medium (<0.2 μmol/mL) is consumed within minutes after gassing has ceased, resulting in an intracellular increase in CoM-S–S-CoB that is no longer reduced (see Figure ). The electron acceptor for the one-electron oxidation reaction is not yet known.

When cells are gassed with 80% N2 and 20% CO2 before the harvest, which removes all of the H2 within seconds and which induces the formation of MCR-ox1,[79] then the increased intracellular CoM-S–S-CoB concentration probably pushes the back reaction beyond transition state II (Figure ). This would allow the coenzyme B thiyl radical (near 0 V),[182] in the absence of methane, to slowly extract an electron from CoM–S-Ni(II)F-430, generating coenzyme B and MCR-ox1, in which –S-CoM is bound via its thiolate sulfur to Ni(III)F-430. The structure of MCR-ox1 has been determined by NMR and XAS measurements to be very similar to that of CoM-S Ni(II) shown in Figure A.[90]

In vivo the MCR-ox1 signal can also be induced by the addition to cultures of M. marburgensis of sodium sulfide (20 mmol/L), which at this concentration completely inhibits methanogenesis.[86]In vitro MCR-ox1 was found to be formed from MCR-red1/2 upon polysulfide addition[82] rather than by sulfide. Both findings can be explained assuming that a coenzyme M persulfide is formed[183] from either coenzyme M and polysulfide or CoM-S–S-CoB and sulfide. In the presence of coenzyme B, the persulfide would react with Ni(I)F-430 to S–-Ni(III) + HS-CoM and/or CoM–S-Ni(III) + H2S. The finding that 35S was incorporated from sulfide into MCR-ox1 and released again upon reduction to Ni(I)F-430 supports this possibility.[86]

The formation of inactive MCR-silent and MCR-ox1 (Figure ), when within methanogenic cells the CoM-S–S-CoB concentration builds up, leads to an inactivation of the enzyme under conditions, where electrons are not available for CO2 reduction. When more electrons become available again, then the inactive MCR is ready[86] to be reduced to the active form in an ATP-dependent reaction (see the Reactivation of Inactive MCR section).

Methane Oxidation to Methyl-Coenzyme M

According to the principle of microscopic reversibility, the reaction mechanism is identical in both directions of the catalyzed reaction.[46,167] Because the equilibrium of the MCR-catalyzed reaction is on the methane side under standard conditions (ΔGo = −30 kJ/mol for methane formation), MCR is predicted to catalyze the back reaction at only a very low specific activity (see the Catalytic Properties section). Indeed, methane oxidation to methyl-coenzyme M in H2O as catalyzed by MCR-red1from M. marburgensis proceeds at a specific rate of only about 0.01% of the specific rate of methyl-coenzyme M reduction to methane (100 μmol min–1 mg protein–1). The initial rate of methane oxidation at 1 bar CH4 and 60 °C was found to be 11.4 nmol min–1 mg protein–1.[36] The ratio of the back reaction to the forward reaction (100 μmol min–1 mg protein–1) is thus near 10–4 (V–1/V+1 = 10–4) (Figure ).

As already mentioned, when methane oxidation to methyl-coenzyme M was measured in 100% D2O, only CH3–S-CoM was formed, indicating the absence of H/D exchange from the medium into the ternary complexes (see the legend in Figure ). This made it possible to study isotope effects for methane activation because the methyl-coenzyme M formed from methane does not undergo isotopic exchange and can therefore be analyzed in order to determine isotope effects. Methane activation was found to proceed with a primary isotope effect of 2.44 ± 0.22 for the C–H versus the C–D bond breakage and with a secondary isotope effect corresponding to 1.17 ± 0.05 per D.[16] The primary isotope effect is quite significant, reflecting that in the methane activating direction a thiyl radical abstracts a hydrogen atom from methane, which is a very endothermal reaction (ΔHo ≈ 70 kJ/mol) and which is not observed in nonenzymatic reactions because follow-up reactions of the thiyl radicals such as dimerization are much faster in solution. However, in the active site of MCR, a thiyl radical might persist for the time needed to react with methane.[16] The secondary isotope effect on methane activation (1.17 ± 0.05 per D) is compatible with methyl radical formation from methane (respectively, a methyl radical-like transition state).[16,167]

Alkane oxidation via an attack of the alkane by a thiyl radical, as proposed for MCR-catalyzed methane oxidation (Figure ), is unprecedented in organic chemistry but not in biochemistry.[53,184] In organic chemistry, anoxic alkane C–H bond activation is generally achieved by interaction of the alkane with a transition metal center (e.g., ref (185)). As indicated above, the involvement of a σ-alkane-nickel complex as an intermediate in methane oxidation to methyl-coenzyme M[171] cannot yet be completely ruled out. This and other mechanisms of methane oxidation catalyzed by MCR involving organometallic intermediates have been reviewed by Scheller et al., 2017.[46]

Solvent Kinetic Isotope Effect (H2O/D2O), Proton Inventory Studies, and Clumped Isotopologues

Solvent Kinetic Isotope Effect

In 100% D2O, CH3–S-CoM was reduced to DCH3 with low amounts of D2CH2 accumulating at the end. The apparent KM and apparent Vmax were only slightly different from those determined in 100% H2O. From the differences, a primary solvent kinetic isotope effect (KIE) between 1 and 1.6 was estimated. The reason for the large error bar is that determinations of kcat over KM with the assay employed (discontinuous measurement of methane via gas chromatography; low enzyme concentration) is not very precise, considering that even traces of O2 negatively affect kcat of MCR.[16]

Proton Inventory Studies

In the case of the MCR-catalyzed reaction, a proton or deuterium ends up in the product dependent on whether the reaction is performed in H2O or D2O (Figure ). Therefore, the isotope distribution in the product (CH4 versus CH3D) can be studied as the function of the deuterium content of water (proton inventory technique),[186] yielding the apparent fractionation factor Φapp that is a composite of the equilibrium fractionation factor Φ of the hydrogen-donating group before the rate-limiting step and the solvent kinetic isotope effect KIE = Φ/Φapp. The equilibrium fractionation factor Φ for HS-CoB in distilled water at 25 °C was determined by NMR spectroscopy to be 0.42, increasing somewhat (7%) between 4 and 50 °C.[187] The apparent fractionation factor Φapp for methane formation from methyl-coenzyme M was found to be 0.27 (Figure , red curve). From ΦHS-CoB = 0.42 and Φapp = 0.27, a solvent KIE = 1.55 is calculated, which is within the range of the solvent KIE estimated to lie between 1.0 and 1.6 from kcat/KM measurements. The proton inventory study is thus consistent with coenzyme B being the hydrogen-donating group before the rate-limiting step as assumed in Figure . If, for example, the OH groups (Φ = 1.0) of the two conserved tyrosines in the active site would be the hydrogen donor before the rate-limiting step, then the solvent KIE effect would have to be 3.7, which clearly was not found.

Figure 11

Proton inventory for methyl-coenzyme M reduction with coenzyme B to methane in H2O/D2O containing increasing amounts of deuterium as catalyzed by MCR I of M. marburgensis. The reaction was performed in the presence of Ti(III) and traces of cob(II)alamin to regenerate coenzyme B from CoM-S–S-CoB. The figure was taken with permission from the Ph.D. Thesis of S. Scheller[188] (Figure ) and slightly modified. When coenzyme B was limiting (not regenerated by reduction of CoM-S–S-CoB), then the apparent fractionation factor was found to be 0.297.[188]

However, one has to consider that the solvent KIE is near 1.0 rather than near 1.6, which would also be in the range determined experimentally. In this case, the found apparent fractionation factor of 0.27 is only explainable if the group providing the hydrogen in the rate-limiting step has an equilibrium fractionation factor smaller than 0.3, which would exclude coenzyme B with Φ = 0.42 as being the hydrogen-donating group before the rate-limiting step (KIE = Φ/Φapp). Metal hydrides such as CpW(CO)2(PMe3)H can have equilibrium fractionation factors as small as 0.2.[189,190] A solvent kinetic isotope effect near 1 would thus favor a Ni hydride as an intermediate (Figure ) and exclude coenzyme B as the hydrogen-donating group before the rate-limiting step as assumed in Figure . It will therefore be of utmost importance in future studies to determine the solvent kinetic isotope effect of methane formation from methyl-coenzyme M with a higher accuracy than was until now possible.

There is one conclusion from the proton inventory results that appears to be straightforward: the results are not in favor of the formation of a methyl-Ni(III) intermediate in the rate-limiting step because the hydrogen-donating group would provide the hydrogen after the rate-limiting step, resulting in an apparent fractionation factor of 1, which was definitely not found. Both in the case of NH and OH groups, Φ = 1.

Clumped Isotopologues

Measurements of clumped isotopologues of methane (13CH3D and/or 12CH2D2) are being applied as tools to constrain the source of methane in a variety of environments.[191,192] Δ13CH3D values are a composite of temperature-dependent equilibrium isotope effects and nonequilibrium kinetic isotope effects. Low Δ13CH3D values characteristic for methanogens growing under laboratory conditions have been suggested to be produced during enzymatic reactions common in all methanogenic pathways, such as the reduction of methyl-coenzyme M.[193] Determination of Δ13CH3D with purified MCR could not only help interpret the in vivo results but also contribute to the understanding of the catalytic MCR mechanism.

Half-of-the-Sites Reactivity

As mentioned above, when MCR I from in the red1a state is supplemented with coenzyme M and coenzyme B, the axial red1c EPR signal is partially converted into the rhombic MCR-red2r and axial MCR-red2a signals (Figure ). The conversion is in a temperature-dependent equilibrium. The equilibrium constant (Keq = [MCR-red2]/[MCR-red1]) was 0.1 at 0 °C and increased to approximately 1 at 20 °C. Above 20 °C, Keq remained constant. This was shown both by EPR spectroscopy and by UV–vis spectroscopy.[135] Thermodynamics predict for an endothermic reaction that the equilibrium constant increases with temperature until Keq = ∞. The finding that the increase in MCR-red2 with an increasing temperature leveled off, when approximately 50% of the red1 signal present initially was converted to the red2 signals (Keq = 1), was therefore unexpected. It can be explained, however, considering that each MCR molecule harbors two structurally interacting active sites and postulating that in the presence of coenzyme M and coenzyme B only one of the two active sites can be converted from the MCR-red1 state into the two MCR-red2 states. The temperature dependence thus indicates that in the presence of coenzyme M and coenzyme B the two active sites in one MCR molecule are in two different states, suggesting that active MCR shows half-of-the sites reactivity.[135]

The reduction of methyl-coenzyme M with coenzyme B catalyzed by MCR takes place in a hydrophobic pocket, from which water is excluded. The two substrates thus have to be stripped of water when entering the active site, and after reaction, the product CoM-S–S-CoB has to be expelled into the water phase. This is most probably achieved by a conformational change, which is driven by one of the exergonic steps in the catalytic cycle. The two active sites could operate independently or coupled. The finding of half-of-the-sites reactivity for MCR is in favor of the two active sites being coupled. The intertwined hexameric structure of MCR, in which the lower axial ligand to nickel in the one active site is provided by the glutamine of the α subunits that mainly forms the other active site and, vice versa, is optimally suited for such a mechanism, considering also that the lower axial acylamide ligand to nickel oscillates between loosely bound N(I) states and tightly bound N(II) states during the catalytic cycle (Figure ). Thus, coupling of the endergonic and the exergonic steps of the catalytic cycle in the two active sites could be envisioned as a strategy employed by MCR to lower the activation energy of the rate-limiting step in analogy to dual stroke motors.[135]

The active site thioglycine in the α subunit (McrA) has been considered to be involved in the half-of-the-sites reactivity because the thioglycine peptide bond facilitates a cis–trans isomerization of the peptide bond, connecting the thioglycine to the next amino acid.[135] Exchange of the thioglycine to glycine yielded an MCR that was still active as deduced from the fact that the mutated cells grew normally when not stressed.[127] Whether the variant enzyme still shows full activity cannot be deduced from this in vivo experiment as has been outlined above (see the Post-Translational Modifications).

ATP-Dependent Reductive Reactivation of MCR

When the nickel in MCR gets oxidized to Ni(II), either as the result of a high intracellular CoM-S–S-CoB/HS-CoM + HS-CoB ratio or by, for example, O2 or chloroform,[149] the enzyme is inactivated but can be reactivated again in an ATP-dependent enzyme-catalyzed reaction. In 1976, it was reported that three components, A, B, and C, of Methanobacterium cell extracts and ATP are required for methyl-coenzyme M reduction with H2 to methane.[4,5] Component A of the MCR system was later resolved into three protein fractions, A1, A2, and A3; the presence of all three of which was required besides coenzyme B (component B) and catalytic amounts of ATP for MCR (component C) to catalyze methyl-coenzyme M reduction with H2 to methane.[194−196] Component A3 was finally resolved into two components, A3a and A3b. Only components A2 and A3a, coenzyme B, MCR, and ATP were required when T(III) rather than H2 was the electron donor.[197] A2 was found to be an ATP-binding protein[198] and A3a to be an iron–sulfur protein.[197] Based on these studies, it was proposed in 1988 that the function of A2 and A3a is to catalyze the reduction of nickel in F-430 of MCR from the inactive Ni(II) to the active Ni(I) oxidation state in an ATP-dependent reaction.[199] All of these activation studies were hampered by the fact that only specific activities of MCR of about 0.1 μmol min–1 mg protein–1 were maximally reached.

A breakthrough came from the finding that activated MCR is instable in the presence of its product CoM-S–S-CoB[113] (see the Enzyme Product Complex section). Taking this into account, inactive MCR-silent or MCR-ox1 were incubated in the presence of A2, A3a, ATP, and dithiothreitol as an electron donor in the absence of methyl-coenzyme M and coenzyme B that upon MCR activation would be converted to CoM-S–S-CoB. Under these conditions, MCR-ox1 was activated to 100% and MCR-silent to 65% as revealed by EPR spectroscopy (Figure ).

The colorless A2 protein has a molecular mass of near 60 kDa and possesses two ATP-binding domains.[198] A2 appears to be specific for the MCR of the methanogen, in which it is synthesized. In M. marburgensis, there is only one gene for A2 although the organism contains two MCR isoenzymes. In other methanogens, there can be two copies of the A2 gene and only one set of genes for MCR. The brown A3a protein is a nonhomogeneous 700 kDa multienzyme complex that includes the mcrC gene product (see the Subunit Composition section), an Fe-protein homologue, an iron–sulfur flavoprotein, polyferredoxin, acetyl-CoA synthase/decarbonylase, and protein components involved in heterodisulfide reductase coupled electron bifurcation.[113] The details still have to be worked out.

Recently it was found that inactive MCR in cells of M. marburgensis is fully activated when the cells are incubated in the presence of CO, which in the cells is oxidized to CO2 with ferredoxin as an electron acceptor in a carbon monoxide dehydrogenase-catalyzed reaction.[200] Partial activation (to 50 nmol min–1 mg protein–1 ≙ 0.05%) of purified MCR from Methanosarcina thermophila by CO in a ferredoxin- and carbon monoxide dehydrogenase-dependent reaction had previously been reported.[114] Incubation of the M. marburgensis cells with H2 (Eo′ = −414 mV) also yielded an active enzyme but much more slowly than with CO (Eo′ = −520 mV) in agreement with the assumption that the reduction of the nickel in MCR (Eo′ = ←600 mV) with CO at high concentrations is thermodynamically possible but not with H2. The reduction of the Ni in MCR by H2 becomes exergonic only when coupled to the hydrolysis of ATP.

In hydrogenotrophic methanogens, 0.5 mol ATP is formed per mol CO2 reduced to methane (Figure ).[42] If MCR has to be reactivated, let us say after 100 turnovers, and if for the reactivation 1 ATP is required, then the ATP gain is predicted to decrease from 0.5 mol ATP/mol methane to 0.49 mol ATP/mol methane. Such a decrease in growth yield makes the methanogens not less viable.[74] However, without the reactivation mechanism, these microorganisms would not be able to recover from conditions not favorable for methyl-coenzyme M reductase and methanogenesis.

Light-Dependent Activation of MCR

In 1991, it was shown by Olson and Wolfe[201] that inactive purified MCR was partially activated by exposure to light above 400 nm. Components necessary for light activation and methanogenesis were coenzyme B, methyl-coenzyme M, and titanium(III) citrate rather than ATP, A2, and A3. Photoactivation was suppressed by inhibitors of methanogenesis such as 2-bromoethanesulfonate, CoB6–SH, and dithionite. Although the specific activity was very low (35 nmol of CH4 per h per mg of protein), the experiments clearly showed that there is a light-sensitive component involved in the activation of MCR.

Conclusion and Outlook

MCR was discovered almost 50 years ago in methanogens and has been intensively studied ever since then. In the beginning of the 2000s, MCR was also found in methane-oxidizing anaerobic archaea.[37,38] Recent reports indicate that MCR may be also involved in anaerobic butane oxidation[54] and in anaerobic ethane oxidation,[55] as evidenced by the finding of MCR and butyl-S-CoM or ethyl-S-CoM in the respective archaea. Butane-oxidizing archaea did not oxidize methane or ethane, and ethane-oxidizing archaea did not oxidize methane or butane, which, based on the crystal structure and substrate specificity of MCR from methanogenic archaea, is very surprising. How these alkyl-coenzyme M reductases can prevent methane from being oxidized to methyl-coenzyme M and methyl-coenzyme M from being reduced to methane is not understood. The catalytic mechanism proposed for MCR from methanogens in Figure postulates a “sterically contained methane” as an intermediate in order to explain the incorporation of deuterium into the methyl group of methyl-coenzyme M when MCR catalyzes methane formation from methyl-coenzyme M in D2O. The exchange is also consistent with a σ-alkane-nickel complex as an intermediate but formation of such a σ-complex is not favored by DFT calculations. When MCR I from M. marburgensis catalyzes ethane formation from ethyl-coenzyme M in D2O, the exchange rate of deuterium into the ethyl group even exceeds the rate of ethane formation.[138,171] There thus appears to be a yet unknown gate that determines substrate specificity. The described catalytic mechanism also cannot explain, why the two active sites of the α2β2γ2 hexamer are not only structurally but also mechanistically coupled (half-of-the-sites reactivity).[135] It appears from mutation studies published in the last two years that the unique, highly conserved post-translational modifications, thioglycine and 5-methylarginine, in the active site of MCRs are essential only under stress conditions.[127,130] One of the assumptions made in the interpretation of the in vivo genetic results is that MCR catalyzes the growth-rate-limiting step, which is only correct under growth conditions prevailing in the natural environment of the methanogens. Under batch culture laboratory conditions, frequently, the rate of anabolism rather than that of catabolism limits growth rate. Measurements of the catalytic efficiency (kcat/KM) and of the activation energy of the mutated MCR are therefore required to elucidate the function of the post-translational modifications. The 700 kDa enzyme complex A3a that catalyzes together with the chaperone A2, the ATP-dependent reduction of inactive MCR, to active enzyme has not yet been characterized.[113] The ATP costs of MCR reactivation during growth are not known. In conclusion, there is still a lot to be resolved before we start understanding how MCR and its activation function.