Research Article

The radical mechanism of biological methane synthesis by methyl-coenzyme M reductase

See allHide authors and affiliations

Science  20 May 2016:
Vol. 352, Issue 6288, pp. 953-958
DOI: 10.1126/science.aaf0616

A radical route to making methane

Microorganisms are the main drivers of Earth's methane cycle. The enzyme ultimately responsible for biological methane production has an ambiguous mechanism because it involves difficult-to-isolate reaction intermediates. Wongnate et al. used stopped-flow and rapid freeze-quench experiments to trap a methyl radical in the active site of methyl-coenzyme M reductase (see the Perspective by Lawton and Rosenzweig). Spectroscopy demonstrated that cofactor F430 contained Ni(II), consistent with computational results. The final step of methanogenesis thus proceeds through Ni(II)-thiolate and methyl radical intermediates rather than an organometallic methyl-Ni(III) mechanism.

Science, this issue p. 953;, see also p. 892

Abstract

Methyl-coenzyme M reductase, the rate-limiting enzyme in methanogenesis and anaerobic methane oxidation, is responsible for the biological production of more than 1 billion tons of methane per year. The mechanism of methane synthesis is thought to involve either methyl-nickel(III) or methyl radical/Ni(II)-thiolate intermediates. We employed transient kinetic, spectroscopic, and computational approaches to study the reaction between the active Ni(I) enzyme and substrates. Consistent with the methyl radical–based mechanism, there was no evidence for a methyl-Ni(III) species; furthermore, magnetic circular dichroism spectroscopy identified the Ni(II)-thiolate intermediate. Temperature-dependent transient kinetics also closely matched density functional theory predictions of the methyl radical mechanism. Identifying the key intermediate in methanogenesis provides fundamental insights to develop better catalysts for producing and activating an important fuel and potent greenhouse gas.

Methanogenic archaea produce more than 90% of Earth’s atmospheric methane (1), totaling more than 1 billion tons of methane per year globally (2). Furthermore, methanogens living in microbial communities containing sulfate- or nitrate-reducing bacteria are responsible for the annual anaerobic oxidation of 0.1 billion tons of methane (36). The enzyme that catalyzes the chemical step of methane synthesis (Eq. 1) or oxidation (the reverse of Eq. 1) is methyl-coenzyme M reductase (MCR), which contains a nickel hydrocorphinate F430 at its active site (4, 79). This reaction involves conversion of the methyl donor, methyl-coenzyme M (methyl-SCoM), and the electron donor, coenzyme B (CoBSH, N-7-mercaptoheptanoylthreonine phosphate) (10), to methane and the mixed disulfide CoBS-SCoM (11) (Eq. 1). The substrates bind inside a deep substrate channel with CoBSH nearer to the surface, stretching toward methyl-SCoM, which is close to F430 (12).

Methyl-SCoM + CoBSH → CH4 + CoBS-SCoM ΔG0′ = –30 kJ/mol (1)

The mechanism of methane formation is not fully resolved, mainly because intermediates in the catalytic cycle have not been identified. Uncovering the MCR mechanism is critical because of the important biogeochemical and environmental roles of this enzyme in generating (and metabolizing) a Janus-like compound that serves as a key energy source and is a potent greenhouse gas. Furthermore, the chemical principles underlying both synthesis and activation of methane may inform the development of catalysts that mimic the structure and/or function of the key enzymatic intermediate(s) or transition state(s). The two proposed mechanisms for how methane is generated differ in whether the first step involves an organometallic methyl-Ni(III) [mechanism I (1315)] or a methyl radical intermediate [mechanism II (16)] (Fig. 1). In both mechanisms, the nickel center of F430 must be in the Ni(I) oxidation state for the enzyme to initiate catalysis (17, 18).

Fig. 1 Initial steps in three mechanisms of MCR catalysis.

Mechanism I involves nucleophilic attack of Ni(I)-MCRred1 on the methyl group of methyl-SCoM to generate a methyl-Ni(III) intermediate (34). This mechanism is similar to that of B12-dependent methyltransferases (48), which generate a methyl-cob(III)alamin intermediate. In mechanism II, Ni(I) attack on the sulfur atom of methyl-SCoM promotes the homolytic cleavage of the methyl-sulfur bond to produce a methyl radical (•CH3) and a Ni(II)-thiolate. Mechanism III involves nucleophilic attack of Ni(I) on the sulfur of methyl-SCoM to form a highly reactive methyl anion and Ni(III)-SCoM (MCRox1).

Support for mechanism I is based on experiments using F430 model complexes (19, 20), enzymatic studies involving isotope exchange (21), and the reaction of the active form of MCR (MCRred1) with various activated alkyl donors such as alkyl halides (2224). These substrate analogs react with Ni(I) to generate alkyl-Ni(III) species that undergo reduction to the alkane (as in the forward direction of Eq. 1) or conversion to thioethers—e.g., methyl-SCoM upon reaction with organic thiolates like CoM (as in the reverse reaction) (2224). Mechanism II is supported by density functional theory (DFT) computations in which it was argued that formation of the methyl-Ni(III) intermediate is not energetically feasible, being endoergic by 91 kJ/mol (with an activation free energy of 94 kJ/mol), whereas the formation of a methyl radical and Ni(II)-thiolate is exoergic by 10 kJ/mol (with an activation free energy of 63 kJ/mol) (16, 2527).

A third mechanism is also possible in which nucleophilic attack of Ni(I) on methyl-SCoM generates a Ni(III)-SCoM species and, formally, an anionic methyl group that undergoes protonation to generate methane (mechanism III) (Fig. 1). A Ni(III)-thiolate species known as MCRox1 is otherwise formed when growing cells are exposed to sodium sulfide (18) or to an oxidizing gas mixture (80% N2/20% CO2) (28). MCRox1 is also called the “ready” state of the enzyme because it can be activated to the active MCRred1 state (17, 18). Both the methyl-Ni(III) (23, 29) and the Ni(III)-thiolate (MCRox1) (30) states have been generated in high yield, are relatively stable, and exhibit distinctive electron paramagnetic resonance (EPR) spectra. Actually, spectroscopic and computational studies indicate that MCRox1 is best described as a high-spin Ni(II)-thiyl radical in resonance with a Ni(III)-thiolate species (30, 31). Conversely, the Ni(II)-MCRox1-silent state is EPR-silent. The MCRox1, MCRox1-silent, and MCRred1 states also display distinct magnetic circular dichroism (MCD) spectra (31, 32). Thus, performing rapid mixing experiments and monitoring the accumulation of an intermediate exhibiting the spectroscopic features of the methyl-Ni(III), MCRox1-silent, or MCRox1 states associated with decay of MCRred1 should provide unambiguous evidence supporting one of the three mechanisms. However, only minor spectroscopic changes are observed when MCR reacts with methyl-SCoM and the natural substrate CoBSH (33).

We performed transient kinetic, spectroscopic [ultraviolet-visible (UV-Vis), EPR, and MCD], and computational studies of the first step in the MCR catalytic mechanism to trap and identify the key intermediates that differ between mechanisms I and II. MCR contained a sufficiently high amount (70 to 80%) of the active Ni(I)-MCRred1 state to monitor changes in its spectroscopic properties during the reaction and identify intermediates. We rapidly mixed MCR with methyl-SCoM and CoB6SH, containing a hexanoyl instead of heptanoyl side chain, which sufficiently slows down the first step in the MCR reaction (34, 35) to allow accumulation and detection of the first intermediate in the MCR mechanism.

Rapid kinetic studies rule out methyl-Ni(III) and trap the MCRox1-silent intermediate

We performed stopped-flow studies by rapidly mixing a solution containing MCR and methyl-SCoM with the slow substrate CoB6SH (Fig. 2A). We tracked the reaction at 385 nm to follow Ni(I) decay and at 420 nm to measure the rate at which the Ni(II) or Ni(III) intermediate forms. Although the steady-state and presteady-state rate constants are slower by factors of 1000 and 440 with CoB6SH than with CoBSH, no spectroscopic changes are observed upon addition of methyl-SCoM alone; in fact, even for a single turnover, both substrates must be present before any reaction can occur (34). This strongly suggests that with the slow (CoB6SH) substrate, MCR employs the same strict ternary-complex mechanism as with the native (CoBSH) substrate (33, 34). The spectroscopic features at both 385 and 420 nm exhibited monophasic kinetics, with 60% of the starting MCRred1 state undergoing conversion at a rate constant of 0.35 ± 0.01 s−1. Two additional slow phases [observed rate constant 2 (kobs2) = 0.05 ± 0.01 s−1 and kobs3 = 0.008 ± 0.001 s−1] with greatly reduced amplitudes are observed that also occur in control reactions lacking substrate, indicating that these phases correspond to the noncatalytic Ni(I) oxidation and are not relevant to the catalytic mechanism. Over a longer time frame, the spectrum of active MCRred1 returns (kobs = 0.011 ± 0.001 min−1), validating that MCR remains active during these spectroscopic transformations with methyl-SCoM and CoB6SH, as recently shown for the reaction with CoBSH (33).

Fig. 2 Rapid kinetic studies of the reaction of the MCR:methyl-SCoM complex with CoB6SH.

(A) Stopped-flow. Kinetic traces of the reaction of a premixed solution containing MCRred1 (20 μM, after mixing) and methyl-SCoM (20 μM, after mixing) with CoB6SH (500 μM, after mixing) in 50 mM Tris-HCl, pH 7.6. The reactions were performed under anaerobic conditions using the stopped-flow spectrophotometer at 18°C and monitored by following the decay of Ni(I) at 385 nm (blue line) and the formation of Ni(II)/Ni(III) at 420 nm (red line). The reaction showed monophasic kinetics with a rate constant of 0.35 ± 0.01 s−1. (B) Rapid chemical-quench. Reactions of a premixed solution containing equimolar MCRred1 (20 μM, after mixing) and [14C]methyl-SCoM (20 μM, after mixing) with CoB6SH (500 μM, after mixing) were quenched with 0.2 M perchloric acid at various times using the rapid chemical-quench apparatus. Volatile methane product was lost from the solution, and the percentage conversion of [14C]methyl-SCoM was determined by comparing the remaining concentration of [14C]methyl-SCoM to the initial concentration. Plotting the percentage conversion versus time yielded a single-exponential curve with a rate constant of 0.31 ± 0.04 s−1. The vertical brackets at each point indicate the standard deviation of the measurement. (C and D) RFQ EPR. A solution containing MCRred1 (48 μM) and methyl-SCoM (600 μM) was reacted with CoB6SH (120 μM) and freeze-quenched at various times using an RFQ apparatus. Representative time-dependent EPR spectra are shown on (C). The inset shows the g ~ 2.2 region near the S-shaped feature of MCRox1. The percentage decay of MCRred1 (blue) and formation of MCRox1 (red) were determined by comparing their doubly integrated signal intensities at various quenching times to their initial intensities. The data were plotted and fit to single-exponential curves in (D). The MCRred1 signal decayed by 90% during the first phase of the reaction with a rate constant of 0.53 ± 0.25 s−1, whereas the MCRox1-like signal increased by ~3% (relative to the initial MCRred1), with a rate constant of 0.69 ± 0.24 s−1. A radical formed with a rate constant of 0.52 ± 0.32 s−1 and reached ~7% of the initial MCRred1 (see fig. S1). A vertical line at each point indicates the standard deviation of each measurement.

To further study the conversion of methyl-SCoM to methane, we used the rapid chemical-quench method under conditions similar to those of the stopped-flow experiments. A solution containing MCRred1 (20 μM, after mixing) and [14C]methyl-SCoM (20 μM, after mixing) was rapidly mixed with CoB6SH (500 μM, after mixing), and incubated for various time points between 0.62 and 90 s. The reaction was quenched by mixing with 0.2 M perchloric acid and analyzed by liquid scintillation counting. The amount of remaining [14C]methyl-SCoM was plotted versus time (Fig. 2B) and fit to a single-exponential curve, revealing a limiting rate constant of 0.31 ± 0.04 s−1. The results demonstrated that conversion of the methyl group of methyl-SCoM to methane is limited by the same rate constant (~0.30 s−1) as conversion of Ni(I) to Ni(II)/Ni(III) in the stopped-flow experiment. These results support mechanism II (or III), because the reactive methyl radical (and methyl anion) would have very transient existence and would immediately abstract a hydrogen atom or proton, respectively, from CoBSH to generate methane with the same rate constant as that of Ni(I) decay. However, because methyl-Ni(III) is relatively stable, methane formation by mechanism I requires another step and, thus, would occur more slowly than Ni(I) decay. For example, the alkyl-Ni(III) state of MCR reacts slowly with thiolates. The methyl-Ni(III) state of MCR reacts with CoMSH to generate methyl-SCoM and MCRred1 at a rate constant of 0.04 s−1 (at 25°C) (23). Similarly, the alkyl-Ni(III) state generated from bromopropanesulfonate or various brominated acids react slowly (kmax ~ 0.005 s−1) with small thiolates and even slower with CoBSH (and analogs) to generate MCRred1 and the thioether (22, 24, 36).

To identify and characterize any EPR-detectable intermediates formed during the MCR reaction, we performed rapid freeze-quench (RFQ) EPR experiments under similar conditions as those for the stopped-flow and rapid chemical-quench experiments. We observed a dominant (~90%) decrease in intensity of the characteristic Ni(I) EPR spectrum of MCRred1 at g values of 2.249, 2.084, and 2.061 (Fig. 2C). Comparable results are observed in two similar FQ-EPR experiments (fig. S1). The decay curve fits to a single-exponential equation with a limiting rate constant of 0.53 ± 0.25 s−1 (circles in Fig. 2D). We did not observe any EPR-active species [e.g., methyl-Ni(III)] that accumulate to an amplitude similar to that of MCRred1 decay (~43 μM), suggesting that the Ni-based product of this reaction is a Ni(II) EPR-silent species. However, two other EPR signals formed at low levels during the time course of Ni(I) decay (insets, Fig. 2C and fig. S1). A rhombic signal (g = 2.212, 2.183, and 2.150) identical to that of MCRox1 (30), present at very low levels in the initial sample, slightly increased in intensity to 3% of the initial MCRred1 according to a rate constant of 0.69 ± 0.24 s−1. A radical-type species (g ~ 2.0), observed earlier (35), formed with a rate constant of 0.52 ± 0.32 s−1 to an amplitude that reached 6 to 7% of the initial Ni(I)-MCRred1 signal (fig. S1). The low level of accumulation of this radical species may be due to the difficulty of observing thiyl radicals due to their short lifetimes and large spin-orbit coupling with the sulfur atom (37). Regardless, a sulfur radical(s) is predicted to be an intermediate in all three mechanisms and, thus, is not diagnostic of which one is correct.

The above results suggest that, when active MCRred1 is incubated with methyl-SCoM and CoB6SH, it converts nearly quantitatively to an EPR-silent Ni(II)-thiolate species, consistent with the predictions of mechanism II. Therefore, we turned our attention to a spectroscopic method that could positively identify that intermediate. The MCD spectra of MCRred1, MCRox1, and MCRox1-silent are distinct (31). Figure S2 shows the MCD data of MCRred1 and MCRox1-silent, prepared for comparison. The temperature dependence of the data showed that both species are paramagnetic, with MCRred1 and MCRox1-silent having ground states of S = ½ and S = 1, respectively. MCRred1 has a characteristic negative feature (at 21,620 cm−1) that can be used to monitor the reaction and decay of this species. The kinetic studies indicated that the first intermediate in methanogenesis forms at a rate constant of 0.35 s−1 (t1/2 = 2 s) and remains for at least 70 s; thus, to directly observe this intermediate by MCD, we mixed a solution containing MCRred1 and methyl-SCoM with CoB6SH, rapidly froze this mixture in liquid nitrogen within ~10 s, and performed MCD experiments of the samples prepared before and after reaction with substrates. The dominant changes in the MCD spectrum indicate an almost quantitative conversion of MCRred1 to a species nearly identical to MCRox1-silent (Fig. 3). For example, the negative band at 21,620 cm−1 disappeared, direct evidence of the consumption of MCRred1, as all of the characteristic bands associated with MCRox1-silent appear, positively identifying that Ni(II) species as the reaction product. Although there is some difference in absolute intensity between the data sets (which is likely due to problems with depolarization of the circular polarized light by the frozen glass MCD sample), the spectra are almost identical in band shape—i.e., the number of features and their relative intensities—confirming formation of MCRox1-silent as the major reaction product. The CD spectrometer also records a single-channel voltage curve that can be converted into a low-resolution UV-Vis absorption spectrum. These data demonstrate that the MCD samples were initially in the Ni(I) state and underwent quantitative conversion after reaction (fig. S2D). Altogether, our spectroscopic results provide direct evidence that the EPR-silent Ni(II)-thiolate intermediate proposed in mechanism II is the key intermediate in methanogenesis.

Fig. 3 MCD studies of the reaction of the MCR:methyl-SCoM complex with CoB6SH.

MCD spectra were taken at 2 K of MCRred1 before and after the reaction with methyl-SCoM and CoB6SH. Samples were prepared in 50 mM GPT buffer [(50 mM glycine, 50 mM phosphate, and 50 mM Tris), pH 7.6, containing 0.05 mM Ti(III) citrate] with 50% glycerol for MCRred1 and 73% glycerol for the reaction with CoB6SH.

Mechanisms I, II, and III propose the formation of distinct intermediates. We found no evidence for a methyl-Ni(III) species in our RFQ EPR studies. This [and other alkyl-Ni(III)] species is relatively stable when bound to MCR (23, 38) and should have been observed if it had formed. Moreover, there was no evidence for the MCRred2a or MCRred2r states (g values of 2.273, 2.077, 2.073 and 2.288, 2.231, 2.175, respectively) (39) during the transient kinetic reaction, strongly indicating that neither of these species, assigned as a Ni(III)-hydride or a Ni(I)-thiolate, respectively, serve as an intermediate in the MCR mechanism. A recent computational study suggested that the Ni(III)-H should be reassigned to a catalytically inactive species in which the proton of CoMSH binds near the Ni(I) of MCRred1 (27). Regardless, the data do not support a mechanism involving a methyl-Ni(III), Ni(III)-hydride or side-on C-S coordination to the Ni (39).

We further explored whether the minor MCRox1 signal could have any catalytic relevance for methane formation, perhaps as a parallel pathway via mechanism III (fig. S3). If MCRox1 forms with a rate constant of 0.5 s−1 and accumulates to 3% of the initial amount of MCRred1, it must decay to another Ni-based intermediate with a rate constant (k2) of ~ 15 s−1 (fig. S4). However, the concentration of MCRox1 did not change for at least 70 s, strongly suggesting that the slight increase in the MCRox1 EPR signal occurred by an off-pathway process that is unrelated to catalysis.

Computational studies rule out the Ni(III)-thiolate MCRox1-like intermediate

DFT calculations were undertaken to assess the relative energetics of the initial steps in mechanisms I, II and III involving methyl-Ni(III), Ni(II)-MCRox1-silent, or the Ni(III)-MCRox1 species, respectively (fig. S5). We used truncated forms of the F430 cofactor, CoBSH, and CoM and performed these computations as previously described (2527, 40). The formation of a MCRox1-silent-like intermediate via the homolytic process corresponding to mechanism II is thermodynamically favored (ΔG0 = –3.5 kJ/mol) and has an activation energy (ΔG) of 72 kJ/mol, whereas generation of the methyl-Ni(III) state (by mechanism I) is highly endoergic (ΔG0 = 100 kJ/mol) with a ΔG of 102 kJ/mol.

We were unable to identify an MCRox1-like state, specifically a F430-Ni(III)-SCoM/CoBS intermediate, from direct DFT calculations; however, performing a wave function optimization (41) imposing a –1 e charge on the CoBS fragment, we found this MCRox1-like state residing at 136.0 kJ/mol above the MCRox1-silent baseline. A similar computational strategy applied to identify an anionic transition state connecting MCRred1 to MCRox1 failed. Nevertheless, starting from the homolytic transition state of mechanism II, we identified an excited state with a strong anionic CH3 character, residing at 134.6 kJ/mol above the MCRred1 baseline. The inability to optimize any transition state corresponding to MCRox1 and the corresponding methyl anion could be due to the reduced size of the model; however, the large estimated energetic difference between the MCRox1 and MCRox1-silent states is unlikely to be substantially affected by increasing the model size.

Additional calculations suggested that the MCRox1-like state is inaccessible because of the very positive reduction potential of the Ni(III)-SCoM intermediate species, E0′ = 1.4 V versus the normal hydrogen electrode (NHE) (see the supplementary materials). Indeed, with a calculated redox potential for the CoBS•/CoBS couple in the enzyme cavity of E0′ = 0.0 V versus NHE, Ni(III)-SCoM would promptly oxidize CoBS during turnover. Taken together, these computational findings, along with the kinetic simulations described above, rule out the possibility of a MCRox1-like intermediate (and of mechanism III) in methane synthesis.

Temperature-dependence studies map the MCR reaction profile

One way to discriminate among different mechanisms is to compare experimental activation energy barriers with the transition-state barriers obtained by DFT computations. To obtain the experimental thermodynamic values, a premixed solution containing equimolar MCRred1 and methyl-SCoM was rapidly mixed with a saturating concentration of CoB7SH or CoB6SH at various temperatures between 10° and 50°C, and the reaction was monitored by stopped-flow spectrophotometry (Fig. 4 and fig. S6).

Fig. 4 Effect of temperature on the presteady-state reaction of the MCR:methyl-SCoM complex with CoB6SH and CoB7SH.

Reactions of a premixed solution containing equimolar MCRred1 (20 μM, after mixing) and methyl-SCoM (20 μM, after mixing) with the saturating concentration of CoB7SH (1 mM, after mixing) (circle blue) or CoB6SH (500 μM, after mixing) (diamond red) at various temperature (10° to 50°C) were investigated using stopped-flow spectrophotometry (A). (B) is the Arrhenius plot and (C) is the Eyring plot for formation of a Ni(II)-thiolate showing a transition temperature (arrow). A vertical line at each point indicates a standard deviation of the measurement. The thermodynamic values (Ea, ΔH, and ΔS) and kobs from rapid kinetics are summarized in Table 1.

For the reactions with both CoB7SH and CoB6SH, kobs increased as the temperature increased (Fig. 4A), giving a biphasic Arrhenius plot with a transition temperature at ~30°C (Fig. 4B). For CoB7SH, the activation energies (Ea) above and below the transition temperature were calculated to be 51 and 84 kJ/mol, respectively (Table 1), indicating that there is a temperature-dependent structural transition that alters the MCR mechanism. The Eyring plot (Fig. 4C) also showed a similar nonlinear response with a transition temperature at the same temperature. The enthalpy of activation (ΔH) and the entropy of activation (ΔS) above the transition temperature are 48 kJ/mol and –56 J/(mol K), respectively (Table 1). Between 30° and 50°C, the Gibbs energy of activation (ΔG) increased due to a more negative entropy of activation (ΔS). The ΔH and ΔS values below the transition temperature were 82 kJ/mol and 54 J/(mol K), respectively.

Table 1 Activation barriers and observed rate constants for the first step in the MCR reaction.

Activation barriers and rate constants were determined by presteady-state kinetic experiments (Fig. 4). These experiments were performed at the transition temperature, 30°C.

View this table:

The experimental values of Ea (51 kJ/mol), ΔH (48 kJ/mol), and calculated ΔG at 30°C (65 kJ/mol) were congruent with previous (25) and current DFT-calculated transition-state barriers for the formation of a Ni(II)-thiolate (about 63 kJ/mol and 71 kJ/mol, respectively) via mechanism II, given that only the enzyme’s active site and truncated substrates were considered in the DFT model. In contrast, the DFT-calculated transition barrier (ΔG) for formation of the methyl-Ni(III) intermediate was 102 kJ/mol, which is much higher than the experimental value. In addition to the high kinetic barrier, formation of methyl-Ni(III) from MCRred1 is computed to be highly unfavorable thermodynamically (ΔG0 = 100 kJ/mol). Thus, our temperature-dependent studies strongly support the formation of the Ni(II)-MCRox1-silent-like species as the key intermediate in the MCR reaction, consistent with our spectroscopic data.

Similar temperature-dependence studies measured the thermodynamic values for the reaction with CoB6SH. As with CoB7SH, the Arrhenius and Eyring plots were biphasic with a transition temperature at ~29°C. The Ea values above and below the transition temperature were 47 and 127 kJ/mol, respectively (Fig. 4B). The Eyring analysis resulted in ΔH of 45 and 125 kJ/mol and ΔS of –90 and 176 J/mol.K, above and below the transition temperature, respectively (Fig. 4C). Between 30° and 50°C, the Ea (47 kJ/mol) and ΔH (45 kJ/mol) for the CoB6SH reaction are in the same range as those for CoB7SH (Ea = 51 kJ/mol and ΔH = 48 kJ/mol). These results indicate that the same type of temperature-dependent transition occurs in the MCR reaction with CoB6SH as with the natural CoB7SH substrate. The ΔS value between 30° and 50°C for the reaction with CoB6SH was 1.6 times lower than that for CoB7SH, suggesting that the slow substrate has a higher degree of freedom in the MCR active site than the longer natural substrate. As expected, the ΔG at 30°C for the reaction of CoB6SH is 7 kJ/mol higher than that for CoB7SH (72 kJ/mol versus 65 kJ/mol), in agreement with the kobs at 30°C for CoB6SH being lower by a factor of 15 than that with CoB7SH (2 s−1 versus 31 s−1). These results strongly indicate that MCR employs the same rate-limiting chemical step in its reaction mechanism, whether the second substrate is CoB6SH or CoB7SH. Furthermore, it appears that the major reason for the factor of 1000 slower rate with CoB6SH is the increased entropy of the slowly reacting substrate, CoB6SH, at the active site and in the substrate channel relative to that of CoB7SH.

Molecular dynamics simulations highlight the presence of temperature-dependent structural changes at the active site

Long time-scale molecular dynamics simulations over a wide range of temperatures (from room temperature to 50°C) were performed to identify residues at the active site that could be responsible for the observed biphasic temperature dependence of the kinetics (Table 1 and Fig. 4). The simulations suggested that above 25°C, Tyr333, which is one of the two tyrosine residues that is hydrogen-bonded to the thioether sulfur of methyl-CoM, moves away from CoM, establishing a weak hydrogen bond with Ser399 and allowing methyl-SCoM to approach the Ni center of F430 (Fig. 5). This conformational change lengthens the Tyr-OH-(S)CoM hydrogen bond distance from 2.5 ± 0.4 Å (25°C) to 3.1 ± 0.5 Å at around 30°C as the CoM(S)-Ni average distance decreases from 3.7 ± 0.2 Å to 3.2 ± 0.3 Å (table S1 and Fig. 5). Shortening the CoM(S)-Ni distance is expected to facilitate cleavage of the CoM(S)-CH3 bond. None of the other hydrogen bond interactions at the active site changed as the temperature increased (table S1 and fig. S7).

Fig. 5 Hydrogen bonding interactions among MCR active site residues.

Red sticks indicate hydrogen bonds at 25°C. The dashed line indicates the weak hydrogen bond between Ser399 and Tyr333 above 30°C. Residues are numbered according to MCR from Methanothermobacter marburgensis. See also table S1.

Implications

In the first step of mechanism II (Fig. 6), attack of Ni(I) on the sulfur of methyl-SCoM leads to homolytic cleavage of the C-S bond and generation of a methyl radical and a Ni(II)-thiolate, known as MCRox1-silent. Carbon (1.04) and secondary deuterium (1.19) isotope effect studies (42, 43) indicate that the transition state for the rate-limiting C-S bond cleavage involves a trigonal planar carbon, consistent with formation of a methyl radical in the first step in methane synthesis. Because anaerobic methane oxidation occurs by direct reversal of its synthesis (3), the methyl radical will also be formed in the transition state for the final step in reverse methanogenesis. The second step involves H-atom abstraction from CoBSH, generating methane and the CoBS• radical, which in the third step combines with the Ni-bound thiolate of CoM to generate a Ni(II)-disulfide anion radical, poised for one-electron transfer to Ni(II) to generate Ni(I)-MCRred1 and the heterodisulfide (CoBSSCoM) product. The final mechanistic step involves dissociation of the heterodisulfide to allow the ordered binding of methyl-SCoM and CoBSH and initiate the next catalytic cycle.

Fig. 6 Proposed steps of mechanism II.

In the first step, Ni(I) attack on the sulfur of methyl-SCoM leads to homolytic cleavage of the C-S bond and generation of a methyl radical and a Ni(II)-thiolate (MCRox1-silent). Next, H-atom abstraction from CoBSH generates methane and the CoBS• radical, which in the third step combines with the Ni-bound thiolate of CoM to generate the Ni(II)-disulfide anion radical. Then, one-electron transfer to Ni(II) generates MCRred1 and the heterodisulfide (CoBSSCoM) product, which dissociates leading to ordered binding of methyl-SCoM and CoBSH and initiation of the next catalytic cycle.

Understanding the mechanism of these reactions has large implications for developing technologies to catalytically generate and activate methane (44, 45). The latter process is one of the most challenging chemical reactions because a catalyst must selectively break the first carbon-hydrogen bond of methane at a huge free-energy cost (438.9 kJ/mol) without cleaving any of the remaining weaker carbon-hydrogen bonds. Exploitation of methane for energy and chemistry is timely because natural gas reserves are predicted to increase by >40% in the United States over the next 30 years (46) and because anthropogenic sources of methane contribute a large proportion (20%) of the world’s annual greenhouse gas warming potential (48).

Supplementary Materials

www.sciencemag.org/content/352/6288/953/suppl/DC1

Materials and Methods

Figs. S1 to S8

Tables S1 to S3

References (4979)

References and Notes

Acknowledgments: Data presented here are available through the Ragsdale LabGuru portal (https://my.labguru.com/knowledge/projects/361/milestones/711). This work was supported by U.S. Department of Energy (DOE), Office of Science, Office of Basic Energy Sciences, under award DE-FG02-08ER15931 and by U.S. DOE, Advanced Research Projects Agency – Energy, under award number DE-AR0000426. Computer resources were provided by the W. R. Wiley Environmental Molecular Sciences Laboratory, a DOE Office of Science User Facility located at Pacific Northwest National Laboratory and sponsored by DOE’s Office of Biological and Environmental Research. Computer resources were also provided by the National Energy Research Computing Center at the Lawrence Berkeley National Laboratory. Author contributions: T.W. generated the data, performed the data analysis, and prepared Figs. 2, A and B, and 4, and Table 1; and performed the data analysis and generated Fig. 2, C and D, and figs. S1, S4, and S6. D. Sliwa prepared CoBSH substrates and enzyme; generated the data and performed preliminary data analysis for Fig. 2, C and D, and fig. S1; and prepared the samples for the MCD measurements described in Fig. 3 and fig. S2. B.G. and S.R. performed and interpreted the computational studies described in Fig. 5; figs. S5, S7, S8, and S9; and tables S1 to S3. D. Smith performed initial computational experiments on the MCR mechanisms. M.W. and N.L. performed, analyzed, and interpreted the MCD spectroscopic experiments and prepared Fig. 3 and fig. S2. S.W.R. conceptualized and refined the research idea; prepared Figs. 1 and 6 and fig. S3; and coordinated the various collaborations and guided analysis and interpretation of the biochemical experiments. All authors were involved to varying degrees in writing and/or editing the manuscript.
View Abstract

Navigate This Article