How Low Can You Go: Methane Production of Methanobacterium congolense at Low CO₂ Concentrations

Introduction

Methanogenic archaea play a key role in the production of biogas from anaerobic digesters (AD), yielding a product gas with 50–75% CH₄ and 25–50% CO₂ (Plugge, 2017). Methanogenic archaea here produce CH₄ from either H₂/CO₂ (hydrogenotrophic methanogensis, 4H₂ + CO₂ → CH₄ + 2H₂O) or acetate (acetoclastic methanogensis, CH₃COOH → CH₄ + CO₂). Hydrogenotrophic methanogens are ubiquitous in natural anaerobic environments other than engineered AD systems, e.g., the gastrointestinal tracts, flooded soils, and anoxic lake and marine sediments (Whitman et al., 2014). In anaerobic environments, H₂ is an intermediate produced by fermentative and syntrophic bacteria, where it undergoes rapid turnover and its concentration is extremely low (Lin et al., 2012). Use of H₂ as an electron donor is however not restricted to hydrogenotrophic methanogens, but other anaerobic microorganisms, e.g., sulfate reducers and acetogens compete for available H₂ with methanogens in anoxic environments (Robinson and Tiedje, 1984; Cordruwisch et al., 1988; Kotsyurbenko et al., 2001). Therefore, many studies have been committed to the understanding of H₂ uptake kinetics of hydrogenotrophic methanogens through either pure cultures or the whole microbial community in environmental samples (e.g., Conrad, 1999; Kotsyurbenko et al., 2001; Eecke et al., 2012, 2013).

In an AD, dissolved H₂ concentration is usually low [0.5–3 μM, (Frigon and Guiot, 1995)]. Low H₂ concentration limits methane production through hydrogenotrophic methanogenesis, which has been verified by many studies devoted to biogas upgrading by injecting H₂ directly into the AD (Luo and Angelidaki, 2012; Agneessens et al., 2017). Above studies showed that the addition of H₂ to ADs greatly increases methane concentration in the biogas, while decreasing CO₂ concentration—a process known as biomethanation.

Through biomethanation the CH₄ concentration is increased to as high as natural gas quality (>95%), and thus this process dramatically alters the standard operational conditions in AD because CO₂ concentration is correspondingly reduced to lower than 5%. Such low CO₂ concentration is rarely seen in natural anaerobic environments where methanogens are present, so it is not clear whether such low CO₂ concentration affects the activities of hydrogenotrophic methanogens. However, CO₂ is known to be an important substrate for hydrogenotrophic methanogens, as it serves as both electron acceptor for energy production and (sole) carbon source for biosynthesis through the Wood-Ljungdahl pathway (Berg, 2011; Borrel et al., 2016). To the best of our knowledge, there is limited knowledge about CO₂ uptake kinetics of methanogens in literature. Nevertheless, understanding CO₂ uptake kinetics of methanogens could consequently be crucial when dealing with the concept of biomethanation, which aims at upgrading the CH₄ concentrations in biogas to >95% by consuming CO₂ to as low concentration as possible.

Previous work gives some insights about limitation of CO₂ consumption rate and methanogenic rate at low CO₂ concentrations during biomethanation (Luo et al., 2012; Garcia-Robledo et al., 2016; Agneessens et al., 2017). Here it was shown that H₂ consumptions rate decreased when headspace CO₂ concentrations was lower than 12% during H₂ pulse injection batch experiments in bioreactors (Agneessens et al., 2017), while an inhibition of H₂ consumption rate was found when CO₂ concentration was below 6% in a methanogenic manure samples (Garcia-Robledo et al., 2016). These results thus indicate that CO₂ uptake rate and methanogenic rate is limited at low CO₂ concentrations with great implications for the limits of the biomethanation technology. However, the exact impact of CO₂ on the methanogenic activity was not clearly depicted by those studies, as they were conducted on complex microbial communities that include both CO₂ consumers and producers—including homoacetogens that compete with methanogens for H₂ and CO₂. Therefore, a thorough understanding of methanogenic reaction kinetics at low CO₂ concentrations seems necessary in order to find out under which conditions hydrogenotrophic methanogenic rates will be reduced or even inhibited during biomethanation. This will enable us to optimize the efficiency of biomethanation.

In this study, we present the first trial to examine CO₂ uptake kinetics of a hydrogenotrophic methanogen, Methanobacterium congolense, by studying its CH₄ production and CO₂ consumption rates at low CO₂ concentrations with surplus of H₂. We chose to study a model organism from the genus Methanobacterium because methanogens of this genus were found to increase substantially after pulse H₂ injections in reactors with mesophilic sludge (Agneessens et al., 2017). M. congolense was used as test strain because it is a mesophilic methanogen originally isolated from a mesophilic anaerobic digester, and also because it solely utilizes H₂ and CO₂ as substrates for growth and methane production (Cuzin et al., 2001).

Materials and Methods

Strains and Culture Medium

Type strain Methanobacterium congolense (DSM7095) was purchased from Leibniz-Institut DSMZ-Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ). The methanogen was cultivated with a H₂:CO₂ (4:1) gas mixture at 37°C in versatile medium for methanogenic archaea, based on the medium developed by Khelaifia et al. (2013), but adding only acetate, formate and NaHCO₃ as carbon source.

For the kinetic studies of CO₂ uptake by M. congolense, the versatile medium was prepared (pH 7) without any carbon sources, by omitting the addition of acetate, formate, and NaHCO₃ (hereinafter referred to as mineral medium). CO₂ gas injected into the headspace at the beginning of incubation was thus the sole carbon source and electron acceptor for M. congolense during all incubations. To minimize pH effects by high amount of CO₂, a version of mineral medium with higher buffering capacity, 10 times higher K₂HPO₄ (5.0 g/L) and 5 times higher KH₂PO₄ (2.5 g/L) than that of the mineral medium, was employed for cultures with high CO₂ concentration (HBC mineral medium, pH 7). M. congolense performed normal growth in either medium or HBC mineral medium as in the versatile medium.

Experimental Setup

Two series of batch culture experiments were carried out for determination of CO₂ uptake kinetics of M. congolense: long-term batch culture lasted for about 1 week and short-term batch culture lasted for around 3 h (see details below). For both batch cultures, M. congolense was cultivated in 330 mL serum bottles filled with 150 mL sterile medium and sealed with butyl rubber stoppers. Cultures were transferred several times in either mineral medium or HBC mineral medium before batch culture experiments, in order to make certain that they were adapted to the medium.

Long-Term Batch Culture

A series of 330 mL serum bottles filled with 150 mL mineral medium were prepared aseptically for the long-term batch culture. Firstly, the bottles were flushed thoroughly with sterile H₂ gas (0.22 μm filtered) through butyl rubber septa and headspace pressure was kept at around 2 atmospheres eventually. Different volumes of sterile CO₂ (0.22 μm-filter filtered), 2, 4, 8, 12, 16, 20, 24, 28, 32, 36, and 40 mL, were subsequently injected into the headspace to reach different CO₂ partial pressure in the bottles. The mole fraction of H₂:CO₂ was higher than 4:1 in all bottles, so H₂ was in excess. The bottles were incubated overnight at 37°C with continuous shaking (90 rpm) to allow the partitioning of CO₂ between the gas and liquid phases and the dissolution of CO₂ in water to be at equilibrium before the experiment started. Following overnight incubation, headspace pressure in each bottle was measured and gas composition was determined by gas chromatography (GC). Based on pressure and gas composition, CO₂ partial pressure in the headspace (pCO₂) were calculated and used to make a series of standard curves for describing the carbonate system in the bottles with sterile medium (details in Standard Curves).

Each of the long-term bottles was inoculated with 8 mL inoculum from an exponential growth phase culture of M. congolense. The bottles were incubated at 37°C and shaken at 90 rpm for about 1 week. During incubation, headspace pressure was monitored and 1 mL headspace was taken for gas composition analysis at regular intervals. 2 mL of medium was sampled aseptically once a day for optical density at 600 nm (OD600) and pH measurement.

Short-Term Batch Culture Experiment

M. congolense was incubated in either mineral medium or HBC medium with 4:1 H₂:CO₂ gases in the headspace. When the methanogens reached late exponential growth phase, as indicated by OD600, the short-term batch cultures were flushed thoroughly with sterile H₂ gas (0.22 μm filtered) through a needle submerged in the liquid phase. The final pressure of H₂ in the headspace was kept at around 2 atmospheres to ensure that H₂ was in excess. Different volumes of sterile CO₂ gas was subsequently injected into the headspace: 2, 4, 8, 12, 16, 20, 24, 32, and 40 mL for bottles with mineral medium and 4, 12, 14, 20, 28, 32, 40, 60, 80, and 90 mL for bottles with HBC mineral medium. Some of the above setup was repeated to check data reproducibility.

Following CO₂ gas injection, the bottles were incubated in a rotary incubator (37°C, 90 rpm) for about 1 h to allow the partitioning of CO₂ between the gas and mineral medium and the dissolution of CO₂ in the medium to reach equilibrium. The headspace pressure and gas composition (CO₂ and CH₄) were monitored with an interval of 15–25 min for 1.5–2 h. At the beginning and end of the experiment, 2 mL liquid were removed aseptically from the bottle for OD600 and pH measurement and 2 mL for determination of cell abundance using quantitative PCR (qPCR). OD600 stayed nearly constant during the experiment. pH at the beginning of the experiment varied with CO₂ amount injected to the bottle. The maximum pH variation was observed with 40 mL CO₂ injected into bottles with mineral medium, where pH dropped to 6.44, compared to pH 7 before adding any CO₂. pH dropped to 6.63 with 90 mL CO₂ injected into HBC mineral medium bottles.

Analytical Measurements

The pressure of headspace was monitored by gas pressure sensor during the incubation. Headspace gas composition (CO₂ and CH₄) was determined immediately on a gas chromatograph equipped with a thermal conductivity detector (Shimadzu-2014) and a stainless steel column packed with Poropaq Q column. The carrier gas was helium. OD600 was measured on a Genesys 10 UV-VIS spectrophotometer (ThermoFisher, USA). pH was measured using a pH meter B-71X (Horiba, Kyoto, Japan).

Cell Abundance Estimation

Liquid samples for cell abundance were flash frozen with liquid nitrogen and stored at −20°C until analysis. DNA was extracted by using FastDNA kit (MP Biomedicals, LLC) and quantitative PCR was executed for quantification of cell abundance by using archaeal 16S rRNA primer pair–arc806F and arc915r-mod (Chen et al., 2017). Cell abundances of M. congolense were estimated by dividing the 16S gene copies with a factor of 3, since M. congolense's genome harbors three 16S rRNA gene copies (Tejerizo et al., 2017).

Carbonate System Calculation

Standard Curves

Following the measurements in Long-Term Batch Culture, standard curves were generated with sterile medium for carbonate system calculation. Under gas-liquid equilibrium conditions, CO₂ gas injected into the 330 mL bottles with 150 mL sterile mineral medium at the starting time ultimately split into two fractions, CO₂ gas in the headspace [CO_2(g)] and dissolved inorganic carbon (DIC) in the medium. The latter was composed of three species (dissolved CO₂ gas–CO_2(aq), ${\text{HCO}}_3^{-}$ and ${\text{CO}}_3^{2-}$).

To generate standard curves, pCO₂ values at equilibrium [pCO_2(eq)] was firstly measured in each bottle receiving different amounts of CO₂ (∑CO₂). Then pCO_2(eq) was plotted against ∑CO₂, and the resulting curve was used for the calculation of total CO₂ in each bottle during the experiments. Furthermore, [DIC] at equilibrium [[DIC]_(eq)] were calculated by subtracting CO_2(g) from ∑CO₂, and both pCO_2(eq) and ∑CO₂ were plotted with [DIC]_(eq) to generate standard curves for deduction of [DIC] during experiments.

Standard curves for carbonate system in the HBC mineral medium were generated in a similar fashion.

Carbonate System Calculation During Incubation

During the microbial growth, we assumed that the buffering capacity in the medium changed very little because the uptake of phosphate was minor compared to the amount of phosphate present in the medium. Thus, the carbonate system in bottles with methanogens behaved similarly as in sterile medium without microbial activity. This was supported by the fact that, when CO₂ was nearly completely consumed by methanogens at the end of long-term batch culture experiments, medium pH returned to nearly 7, which was the initial sterile medium pH before CO₂ injection.

Based on the above assumption, ∑CO₂ in the bottle and [DIC] in the medium during the incubation were estimated from measured pCO₂ by using the standard curves generated in Standard Curves. Concentrations of three inorganic carbon species in the medium (CO_2(aq), ${\text{HCO}}_3^{-}$ and ${\text{CO}}_3^{2-}$) were estimated by software CO₂SYS (Pierrot et al., 2006) by providing pCO₂ and [DIC]. In this study, carbonate is negligible because the pH range was about 6.44–7, so [DIC] ≈ [CO_2(aq)] + [${\text{HCO}}_3^{-}$]. Both CO_2(aq) and ${\text{HCO}}_3^{-}$ are bioavailable carbonate species, and they are shown to be utilized by methanogens in different steps in methanogenesis and carbon fixation (Ferry, 2013). Moreover, the enzyme carbonate anhydrase, which can actively transform ${\text{HCO}}_3^{-}$ to CO₂ gas or vice versa, is found to be ubiquitous in the culturable methanogens isolated so far, including M. congolense. Therefore, DIC is used as the main parameter for studying the CO₂ uptake kinetics here.

Calculation of Kinetic Parameters

Short-Term Batch Culture Experiment

∑CO₂ consumption and CH₄ production kinetics were estimated from specific ∑CO₂ consumption rates (V_∑CO2) and specific CH₄ production rates (V_CH4) in a range of [DIC] in the medium.

[DIC] values were estimated from the CO₂ gas volumes injected in the bottles by fitting them to the standard curve of [DIC] and initial CO₂ amount injected as mentioned in Standard Curves.

In each incubation, V_∑CO2 was calculated by linear fitting of the ∑CO₂ concentration as a function of time. The ∑CO₂ concentrations in the incubations were estimated as indicated in Standard Curves. Similarly, V_CH4 was estimated by linear fitting of the methane concentration in the headspace as function of time in each incubation. We assumed here that the dissolved CH₄ was negligible due to its low solubility in water.

Long-Term Batch Culture

During the long-term incubation, specific ∑CO₂ consumption rates (V_∑CO2) were calculated by dividing ∑CO₂ consumed between two sampling points with time and normalized with cell abundance. Specific CH₄ production rates (V_CH4) were calculated in a same manner. Since V_∑CO2 and V_CH4 changed with time during incubation, only the maximum V_∑CO2 (r_∑CO2) and V_CH4 (r_CH4) from each bottle were taken for kinetic analysis.

The growth yields (Y_biomass) were calculated at the end of incubation when CO₂ was nearly consumed. We assumed that the CO₂ fraction, which was not converted to CH₄, was assimilated into cell biomass. Thus,

Here, CH_4(end) refers to total CH₄ amount produced by the end of incubation; [∑CO_2(initial)] refers to the ∑CO₂ amount injected at the beginning of the experiment; [∑CO_2(end)] refers to the ∑CO₂ amount left in the bottle at the end of the incubation.

During the exponential growth phase, cell numbers were plotted against the volume of ∑CO₂ consumed and CH₄ produced, respectively. The best-fit linear slopes were taken as the growth yields with respect to CO₂ consumed and CH₄ produced (Y_CO2 and Y_CH4).

Modeling of Kinetics

In summary, we estimated V_∑CO2 and V_CH4 from three different experiment setups: (1) short-term incubation with versatile mineral medium; (2) short-term incubation with HBC mineral medium; (3) long-term incubation with versatile mineral medium (r_∑CO2 and r_CH4).

We used the Michaelis-Menten equation to represent the effect of [DIC] on the performance of methanogen's ∑CO₂ consumption rate and CH₄ production rate, with the consideration of [DIC] threshold [DIC]*:

Where: V_max−CO2 and V_max−CH4 are the maximum specific ∑CO₂ consumption rate and CH₄ production rate, respectively; K_m−CO2 and K_m−CH4 are the [DIC] giving one-half the maximum specific ∑CO₂ consumption rate and CH₄ production rate respectively.

By the end of the batch culture experiments, headspace CO₂ was depleted to a nearly constant partial pressure without further consumption, even though H₂ was still in surplus. This constant CO₂ partial pressure remained unchanged over a period even longer than the period of active CO₂ uptake. This indicates that there is a CO₂ and/or [DIC] threshold for M. congolense. The [DIC]* concentration at which no active CO₂ uptake activity could be detected was found to be 44.4 ± 0.4 μM of DIC (N = 7) in this study.

Results

Growth of M. congolense in Long-Term Batch Cultures

Table 1 summarizes cell growth rates, maximum ∑CO₂ consumption (r_∑CO2) and CH₄ (r_CH4) production rates and growth yield of M. congolense incubated under different initial ∑CO₂ amount and surplus H₂. Growth rates, r_CH4 and r_∑CO2 increased linearly with added ∑CO₂. Growth rate increased from <0.02 d⁻¹ with 2 mL CO₂ (~0.09 mmol, [DIC] = 0.44 mM) injected to about 1.12 d⁻¹ with 32 mL CO₂ (~1.43 mmol, [DIC] = 5.13 mM) injected. Lowest r_CH4 and r_∑CO2 (0.26 and 0.21 f mol cell⁻¹ d⁻¹, respectively) were found in the bottle receiving least amount of CO₂ (2 mL). The bottle receiving 36 mL CO₂ showed the highest r_CH4 of about 108.99 f mol cell⁻¹ d⁻¹, while the bottle receiving 40 mL CO₂ showed the highest r_∑CO2 of about 80.90 f mol⁻¹ cell d⁻¹. The growth yield, Y_biomass-carbon assimilated for anabolism with respect to total CO₂ assimilated, was in the range of 11–26%, but it did not show a clear trend with ∑CO₂ amount injected. Cell specific growth yields with respect to ∑CO₂ consumption (Y_CO2) and CH₄ production (Y_CH4), which were estimated from data acquired during the exponential growth phase, were nearly equivalent in all conditions. Y_CO2 and Y_CH4 were consistent at around 0.8–1.8 × 10¹⁴ cells per mole gas consumed or produced, except for the low CO₂ conditions (2–8 mL) where cell growth was not observed during incubation. Y_CO2 and Y_CH4 estimated in the bottles receiving <8 mL CO₂ can be biased due to slow growth, therefore the yields were not used here.

Table 1. Growth parameters of M. congolense under different initial CO₂ amounts injected into the bottles.

In this long-term batch culture experiment, both r_CH4 and r_∑CO2 were of first order kinetics and showed a linear correlation with [DIC] instead of showing Michaelis-Menten kinetics. This implies that the saturation concentration of [DIC] for M. congolense is higher than 6 mM, which was the highest concentration tested in this study (Table 1). Unfortunately, attempts to further increase [DIC] by injecting more CO₂ gas (60 mL) with the same setup led to a pH drop from 7 to <6.4 and substantial reduction of methanogenic activity (data not shown).

Kinetics of ∑CO₂ Consumption and CH₄ Production

Figure 1 gives an example of how V_∑CO2 and V_CH4 were estimated in short-term batch culture experiments. The CO₂ gas in the headspace started to dissolve in the medium right after its injection, and cells started to utilize CO₂ and produce CH₄. CH₄ production rates were low but kept increasing in the first 1–1.5 h during CO₂ gas dissolution until CO₂ equilibrium between gas and liquid phase reached. After the CO₂ gas-liquid equilibrium, CH₄ production rate was stable for a few hours (data not shown). To minimize the influence of increasing cell abundance due to growth during experiment, specific CH₄ production rates were retrieved within 1–1.5 h after reaching CO₂ gas-liquid equilibrium. In CO₂-depleted medium M. congolense produced CH₄ at extremely low rates in the presence of H₂ (data not shown).

Figure 1. An example for determination of methane production rate and DIC consumption rate in a short-term batch culture.

Figures 2A,B show the estimation of V_∑CO2 and V_CH4 from three independent experimental setups, respectively: (1) short-term incubation with versatile mineral medium; (2) short-term incubation with HBC mineral medium; (3) long-term incubation with versatile mineral medium. V_∑CO2 and V_CH4 were both dependent on available [DIC] and followed Michaelis-Menten kinetics. Maximum specific CO₂ consumption rate (V_max−CO2) was estimated to be about 69.2 f mol CO₂ cell⁻¹ d⁻¹ and half-saturation concentration of DIC (K_m−CO2) was about 2.5 mM. Maximum specific CH₄ production rate of M. congolense, V_max−CH4, was about 48.8 f mol CH₄ cell⁻¹ d⁻¹ and half-saturation concentration of DIC (K_m−CH4) was estimated to be about 2.2 mM. Reaction speeds of 80% V_max−CH4 could be reached at 9 mM [DIC]. Extrapolation of the fitted curve in Figure 2B shows that DIC concentrations needed for 90% V_max−CH4 and 100% V_max−CH4 would be 22 mM and 100 mM. As these values are outside the range of [DIC] that could be tested in the present setup without inferring changes in media pH, these concentrations could not be verified experimentally and should thus be interpreted with great care.

Figure 2. K_m and V_m of DIC consumption (A) and methanogenic activity (B) for M. congolense, determined from both short-term and long-term batch cultures by Michaelis-Menten fitting. (C), Correlation between CO₂ consumption rate (V_CO2) and methane production rate (V_CH4) in short-term batch cultures with both versatile mineral medium and HBC mineral medium (N = 25).

Figure 2C shows that there was good correlation between V_∑CO2 and V_CH4 for cells in late exponential growth phase in the short-term batch culture, where V_∑CO2 was about 1.3 times higher than V_CH4.

Estimation of Methanogenic Activities Under Various CO₂ Headspace Concentrations

Biomethanation aims at decreasing pCO₂ in the off-gas, which will directly influence [DIC] in the liquid (slurry) in AD. The [DIC] is furthermore dependent on slurry pH as this determines the CO₂/${\text{HCO}}_3^{-}$ partitioning. Figure 3A shows the profile of [DIC] with pH in liquid (salinity = 0) under headspace CO₂ concentration of 2, 5, 25, and 50%, assuming that the headspace pressure was 1 atmosphere and in equilibrium with slurry and using the growth pH range for M. congolense (5.9–8.2) (Cuzin et al., 2001) for calculation. It can be seen that increased pCO₂ in headspace refers to higher [DIC] in the slurry and [DIC] further increases with pH. It can be calculated that [DIC] is tens of mM in the slurry in a normal mesophilic AD, where CO₂ gas composes 25–50% of the biogas (Plugge, 2017). The lowest [DIC] is found at the lowest pH tested (5.9), about 9 mM, which can support the methanogenic rate of 80% of V_max−CH4.

Figure 3. Carbonate systems, methanogenic activity, and pH in mesophilic AD. (A), Calculated DIC concentration in sludge in a mesophilic AD with headspace CO₂ concentration of 2, 5, 25, and 50%; pH was set to be within the growth range for M. congolense. (B,C), pCO₂ at pH range of 5.9–8.2 in an mesophilic AD, when [DIC] = 2.2 and 6 mM, respectively. Cell specific methanogenic rate for M. congolense is estimated to be 50 and 80% of V_m−CH4 at [DIC] of 2.2 and 9 mM, respectively, following the kinetic modeling from Figure 2B.

For biomethanation, it is critical to know how much pCO₂ can be lowered through upgrading CO₂ to CH₄ but without affecting methanogenic activity. In Figures 3B,C, we model the lowest pCO₂ required to maintain 2.2 and 9 mM [DIC] at different pH so as to maintain 50 and 80% of V_max−CH4 according to the kinetic modeling. It is found that CO₂ can be lowered to 5% and 2% at pH >6.2 and pH >6.8, respectively, with a methanogenic rate of 50% V_max−CH4 (Figure 3B). Correspondingly, with a methanogenic rate of 80% V_max−CH4, CO₂ can be lowered to 5% and 2%, respectively, at pH >7.1 and pH >7.5 (Figure 3C).

Discussion

Using batch-culture experiments, we provided the first estimation of CO₂/DIC uptake kinetics of an autotrophic hydrogenotrophic methanogen, M. congolense. We found that the affinity for DIC was dramatically lower than that of H₂, the other reactant involved in hydrogenotrophic methanogenesis. With a K_m of 2.2–2.5 mM, the affinity for DIC was shown to be a few tens to thousands times lower than the K_m of H₂, 0.44–66 μM, as previously reported for other hydrogenotrophic methanogens (Kotsyurbenko et al., 2001; Karadagli and Rittmann, 2007). Likewise, DIC threshold of 44.4 μM, at which concentration the methanogenic activity could no longer be detected for M. congolense, was also hundreds to thousands times higher than reported H₂ thresholds of 6–70 nM, observed for methanogens (Lin et al., 2012). Such high K_m and threshold of DIC might be related to the CO₂ fixation pathway used by methanogens. M. congolense utilizes the Wood-Ljungdahl pathway for CO₂ fixation, which has previously been shown to have the highest K_m of DIC among the six autotrophic inorganic carbon assimilation pathways (Raven et al., 2012). From an evolutionary perspective, poor affinity for DIC is in accordance with the ubiquitous distribution of methanogens in habitats with high CO₂ concentrations, such as anaerobic digesters, animal guts and sediments. Maximum CH₄ production rate (V_max−CH4, 48.8 f mol CH₄ cell⁻¹ d⁻¹) of M. congolense, estimated from the kinetic model here is comparable to methanogenic rates found for methanogens from other complex environments incubated at a comparable temperature (35°C): 108-135 f mol CH₄ cell⁻¹ d⁻¹ in anaerobic reactors (Li and Noike, 1992) and 31.5 f mol CH₄ cell⁻¹ d⁻¹ in lake sediments (Lay et al., 1996).

Addition of different amounts of CO₂ gas at the beginning of the batch culture experiments caused the decrease of media pH due to the dissolution of CO₂ gas into the medium, where pH decreased more in bottles receiving higher amounts of CO₂. A further challenge was that the CO₂ concentration changed constantly during the incubation, due to continuous CO₂ consumption. Therefore, we controlled the amount of CO₂ added to keep pH within an optimal range for M. congolense (pH 5.9–8.2) (Cuzin et al., 2001), so that growth and methanogenic activity of M. congolense were not affected during our trials. A previous study on an obligate hydrogenotrophic autotrophic methanogen, Methanocaldococcus strain JH146, showed that pH did not affect methanogenic activity when it was within the range for optimal growth (Eecke et al., 2013). In our experiments, addition of 2–40 mL CO₂ lowered pH from 7 to 6.44–6.94 in the mineral medium, which has lower buffering capacity but methanogenic rates kept increasing with CO₂ amount and reached a rate of ~44 f mol CH₄ cell⁻¹ d⁻¹ with 40 mL CO₂ (pH = 6.44). Moreover, specific ∑CO₂ consumption rates and CH₄ production rates determined from all three different experimental setups here were in good agreement with respect to [DIC], regardless of buffering capacity or incubation time (Figure 2). Thus, pH seems to have little impact on the rates within the experimental range. However, addition of excessive CO₂ beyond the buffering capacity was shown to greatly inhibit methanogenic activity: addition of 60 mL CO₂ into the mineral medium decreased pH to 6.01 and resulted in a dramatic reduction of methanogenic activity to ~12 f mol CH₄ cell⁻¹ d⁻¹.

Our study also shows that DIC concentration influences the microbial growth rate. This is revealed by a clear reduction of growth rate from 1.12 d⁻¹ at 5.13 mM [DIC] to <0.02 d⁻¹ at the lowest DIC concentrations tested (0.44 mM) in the long-term batch culture experiment (Table 1). A previous study showed that the mixotrophic methanogen, Methanosarcina barkeri, has very slow growth and low methanogenic rate when incubated with only H₂ but lacking CO₂ (Weimer and Zeikus, 1978).

Here we showed that growth of M. congolense was limited when [DIC] was lower than 1.6 mM, although methanogenesis still continued at low rates. Methanogens fix CO₂ autotrophically into biomass through the Wood-Ljungdahl pathway, with which its methanogenesis pathway is associated (Berg, 2011). Therefore, whether the lowered growth rate at low [DIC] was due to reduced assimilation of carbon for biomass formation, or due to a reduced energy generation from methanogenesis remains unknown. Nonetheless, cell growth yield at the end of long-term incubation seems to be consistent for all DIC concentration tested: the methanogen converts approximately four moles of CO₂ into CH₄ for each mole of CO₂ incorporated into biomass (Table 1). Similar fraction of CO₂ was assimilated into biomass by late exponential growth phase cells during our short-term batch incubation, as ∑CO₂ consumption rate was about 1.3 times higher than methane production rate for all bottles (Figure 2C). The ratio of CO₂ used in dissimilatory and assimilatory metabolisms were thus independent of the [DIC] concentration.

As pCO₂ concentrations in anaerobic digesters is often high (25–50%), our data showed that inorganic carbon availability might not limit methanogenic activity of M. congolense under standard operating conditions of an anaerobic digester. Fermentation processes in the sludge will furthermore supply CO₂ to the hydrogenotrophic methanogens and hereby decrease the likelihood of CO₂ limitation under standard conditions. This study is of primary importance for biomethanation, a power-to-gas technology used for biogas upgrading to increase CH₄ concentration in the produced biogas through reduction of the CO₂ concentration. Knowledge on methanogen's CO₂/DIC kinetics is relevant as low CO₂ concentrations (<2%) are required to fulfill criteria for injection of upgraded biogas to the natural gas grid. If M. congolense is treated as a representative of hydrogenotrophic methanogens, the biogas can be upgraded to >98% CH₄ (<2% CO₂) at 80% V_max−CH4 when slurry pH in the reactor is >7.5 (Figure 3C) and 50% V_max−CH4 at pH >6.8 (Figure 3B). Our result suggests that the availability of bioavailable inorganic carbon under low CO₂ concentration might not greatly decrease methanogenic activity during biogas upgrading, but other factors, such as pH, might have greater impact on methanogenic rates.

A decrease in methanogenic activities to 50% of the maximum was shown in previous studies at different CO₂ concentrations: 2.9% CO₂ (59 mM [DIC] in methanogenic manure samples, Garcia-Robledo et al., 2016) and 10% CO₂ (257 mM [DIC] in anaerobic digestate, Agneessens et al., 2017). These [DIC] concentrations were higher than the K_m values reported here for M. congolense, which would indicate that the organisms in these studies either had lower affinities for [DIC]/CO₂ than M. congolense, or that their methanogenic activities were inhibited by other factors like NH₃ or pH. The pH levels of 8.2 (Garcia-Robledo et al., 2016) and 8.3 (Agneessens et al., 2017), were close to the value of 8.5 reported to be inhibitory to the biomethanation process (Angelidaki et al., 2018). As bicarbonate is the dominant buffering system in anaerobic slurries, it is often difficult to separate effects by high pH from effects by low CO₂ concentrations here, as these are inversely related. Through the pure culture study on M. congolense reported here, it was possible to separate the direct pH effect from low concentrations of CO₂ and hereby elucidate microbial physiological limitations to the process of biomethanation of a methanogenic type strain. The results are applicable to both in situ methanation, where H₂ and CO₂ are converted by methanogens in the main reactor, and separate ex situ reactor harboring specialized methanogenic communities.

Conclusions

Although CO₂ affinity of M. congolense is many times higher than H₂ affinity, CO₂ concentrations will only become severely limiting for biomethanation at very low [DIC] concentrations. Experiments were only carried out on a single methanogenic strain here and further testing of other methanogens will reveal if they elicit a similar affinity for CO₂.

Author Contributions

XC performed the experiments and drafted the manuscript. All authors designed the study, interpreted the data and wrote the manuscript.

Funding

This study was funded by the Innovation Fund Denmark, project Electrogas (grant number 4106-333 00017B).

Conflict of Interest Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

We acknowledge Felix Beulig and Kai Finster from the Section of Microbiology, Department of Bioscience, Aarhus University for giving instructive suggestions on methanogen cultivation techniques and thank their section for providing lab facilities to perform DNA extraction and quantitative PCR. We would also like to thank technicians, Janni Ankerstjerne Sørensen and Britt Amby Malthesen, from Department of Engineering, Aarhus University for their help in setting up the gas chromatographic equipment. We would furthermore like to thank Mads Borgbjerg Jensen and Alastair James Ward from the Department of Engineering, Aarhus University, for their constructive feedback to the manuscript.

References