Controls on the isotopic composition of microbial methane

Microbial methane production (methanogenesis) is responsible for more than half of the annual emissions of this major greenhouse gas to the atmosphere. Although the stable isotopic composition of methane is often used to characterize its sources and sinks, strictly empirical descriptions of the isotopic signature of methanogenesis currently limit these attempts. We developed a metabolic-isotopic model of methanogenesis by carbon dioxide reduction, which predicts carbon and hydrogen isotopic fractionations, and clumped isotopologue distributions, as functions of the cell’s environment. We mechanistically explain multiple isotopic patterns in laboratory and natural settings and show that these patterns constrain the in situ energetics of methanogenesis. Combining our model with data from environments in which methanogenic activity is energy-limited, we provide predictions for the biomass-specific methanogenesis rates and the associated isotopic effects.

If, for example, 13

Sensitivity analysis
We conducted a sensitivity analysis to the tunable model parameters by examining the effect of a 3-fold increase or decrease in the value of these parameters on the reversibility of the enzymaticallycatalyzed reactions (fig. S13) and on 13 ε CO 2 -CH 4 (S14) and 2 ε CH 4 -H 2 O (S15). We find that the values of 13 ε CO 2 -CH 4 and 2 ε CH 4 -H 2 O are sensitive only to some of the model parameters, but that the overall trajectories of the isotopic fractionation dependence on ∆G net are preserved. We identified three types of sensitivity of 13 ε CO 2 -CH 4 and 2 ε CH 4 -H 2 O to some of the model parameters, which are typical to three ∆G net ranges: (i) At small-negative ∆G net ( −60 kJ mol −1 ) the Mtr-and Mcr-catalyzed reactions depart from equilibrium, and consequently, the combination of these reactions' equilibrium and kinetic fractionation factors (EFFs and KFFs, respectively) determines the magnitude of 13 ε CO 2 -CH 4 over this ∆G net range. Thus, changes to parameters that affect the departure from equilibrium of these reactions manifest as a change in the maximal 13  reactions (R ∆G 0 r , see table S1), and the initial concentration of HS-CoB and CoM-S-S-CoB (C i CoB). Over the same ∆G net range, changes in some parameters drive a shift of the 13 ε CO 2 -CH 4 and 2 ε CH 4 -H 2 O trajectories along the ∆G net axis, with minimal effect on the peak 13 ε CO 2 -CH 4 and 2 ε CH 4 -H 2 O values (as in panels 19, 20, and 29 in fig. S14 and panels 13, 19, 26, 29, 35, and 36 in fig. S15). This is mostly evident in K M and V + values of the Hdr-catalyzed reaction, which pyrolyzed CH 4 directly to H 2 , which was then introduced into an IRMS. The latter approach is considered more accurate due to a lower risk of contamination during sample processing and lesser exposure to humid air. This is one reason for which we chose to use the Oku16 over the Yos08 data.
A second reason is that the Oku16 data also include measurements of both carbon and hydrogen isotopes over a larger range of ∆G net values.
There are considerable differences between ∆ 13 CH 3 D values in hydrogenotrophic methanogens with and without membrane-embedded electron carriers (e.g., methanophenazine), the latter of which are the focus of our model. While 2 ε CH 4 -H 2 O is of a similar range for these two groups, ∆ 13 CH 3 D for methanogens with methanophenazines is between −6 and 0‰, whereas for methanogens without methanophenazines the observed range is −2 to 3‰. The laboratory culture data collected to date is insufficient to determine whether this difference is statistically significant, though there may be a physiological basis for it, as methanogens with methanophenazines have distinct metabolic characteristics that may affect the dynamics of departure from equilibrium.

Cell-specific methanogenesis rates
There are currently limited data on cell-specific methanogenesis rates in natural environments.
We bridge this gap by comparing compiled bulk methanogenesis rates (bMR) and estimates of cell density. bMR values were obtained from the results of either ratiotracer experiments or reaction-diffusion models, both of which carry uncertainties. In radiotracer experiments, the rate of methanogenesis is assumed to be equal to the rate of CO 2 reduction to methane, but in fact the measured rate of CO 2 reduction serves as an upper limit on methanogenesis rates. If the reaction is close to equilibrium, then the net rate of methanogenesis will be lower than the radiotracer-based estimate, possibly by orders of magnitude if the reversibility between methane and CO 2 is higher than 0.9. Moreover, some of the radiotracer experiments are conducted under conditions that may favor higher methanogenesis rates (e.g., increased partial pressure of H 2 in the headspace), resulting in overestimation of the in-situ rates. Models of in-situ methanogenesis rates provide an estimate for the net methanogenesis rates based on the concentration and isotopic gradients of methane and DIC, but carry uncertainties due to the choice of model parameters, such as the net fractionation of carbon and hydrogen isotopes associated with methanogenesis.
Where not measured directly, cell densities were estimated based on their dependence on depth (90,91) and assuming that of these cells 12% are Archaea in open-ocean sites and 40% in ocean margin sites (92), and that 50% of Archaea are methanogens (91). † 14 C tracer measurements. § Tracer measurements were done for both gross methane production and consumption, and the results are shown as a net rate, which is calculated as the difference between the measured forward and backward rates. ¶ Direct cell counts. ‡ Rhone River pro-delta.
$ Gulf of Lion shelf.