Direct Interspecies Electron Transfer between Geobacter metallireducens and Methanosarcina barkeri

Abstract

Direct interspecies electron transfer (DIET) is potentially an effective form of syntrophy in methanogenic communities, but little is known about the diversity of methanogens capable of DIET. The ability of Methanosarcina barkeri to participate in DIET was evaluated in coculture with Geobacter metallireducens. Cocultures formed aggregates that shared electrons via DIET during the stoichiometric conversion of ethanol to methane. Cocultures could not be initiated with a pilin-deficient G. metallireducens strain, suggesting that long-range electron transfer along pili was important for DIET. Amendments of granular activated carbon permitted the pilin-deficient G. metallireducens isolates to share electrons with M. barkeri, demonstrating that this conductive material could substitute for pili in promoting DIET. When M. barkeri was grown in coculture with the H₂-producing Pelobacter carbinolicus, incapable of DIET, M. barkeri utilized H₂ as an electron donor but metabolized little of the acetate that P. carbinolicus produced. This suggested that H₂, but not electrons derived from DIET, inhibited acetate metabolism. P. carbinolicus-M. barkeri cocultures did not aggregate, demonstrating that, unlike DIET, close physical contact was not necessary for interspecies H₂ transfer. M. barkeri is the second methanogen found to accept electrons via DIET and the first methanogen known to be capable of using either H₂ or electrons derived from DIET for CO₂ reduction. Furthermore, M. barkeri is genetically tractable, making it a model organism for elucidating mechanisms by which methanogens make biological electrical connections with other cells.

INTRODUCTION

Effective interspecies electron transfer is key to the efficient functioning of methanogenic communities (1–4). Promoting interspecies electron transfer to methanogens enhances the anaerobic digestion of wastes, and appropriate models of the pathways for interspecies electron transfer are necessary in order to predictively model the response of methanogenic communities to environmental change.

The best-known strategy for interspecies electron transfer is H₂ interspecies transfer (HIT), in which electron-donating microorganisms reduce protons to H₂ and methanogens oxidize the H₂ with the reduction of carbon dioxide to methane (1). In some instances, formate substitutes for H₂ as the electron carrier (2–6). The abundance of H₂/formate-consuming methanogens in anaerobic soils (7, 8) and sediments (9, 10) as well as some anaerobic digesters (11) suggests that HIT plays an important role in methane production in those environments. HIT has been documented in studies with defined cocultures of H₂-donating microorganisms and H₂-consuming methanogens, and the physiology and biochemistry of both H₂ production and H₂ consumption are relatively well understood (1, 12, 13).

Direct interspecies electron transfer (DIET) is an alternative to HIT. Methanosaeta (Methanotrix) harundinacea, which is representative of the Methanosaeta species, which are responsible for a substantial portion of the global methane production (14), directly accepted electrons from Geobacter metallireducens for the reduction of carbon dioxide to methane in defined cocultures (15). DIET in these cocultures required that G. metallireducens produce pili (15), that have a metal-like conductivity (16, 17), but further details of the interspecies electrical connections are as yet unknown. Multiple lines of evidence suggested that Methanosaeta species were also participating in DIET in anaerobic digesters treating brewery wastes (15, 18). Geobacter species were abundant in the digester granules, which possessed a metal-like conductivity (18) similar to the metal-like conductivity of Geobacter pili (16–18). The potential importance of conductive pili in electron transfer to Methanosaeta species is consistent with the clear importance of conductive pili in DIET in cocultures of G. metallireducens and G. sulfurreducens, cooperating to oxidize ethanol with the reduction of fumarate (19–22). The electrical connections that M. harundinacea employs for DIET are unknown, and elucidating them may be technically difficult because a system for genetic manipulation of this organism has not yet been identified.

Another genus of acetoclastic methanogens, Methanosarcina, is often abundant in methanogenic soils and sediments (23–25), coal mines (26, 27), landfills (28), and anaerobic digesters (29, 30). Several studies suggested that Methanosarcina species could accept electrons from nonbiological extracellular surfaces (31, 32). For example, M. barkeri attached onto electrically conductive granular activated carbon (GAC), which served as a mediator for electron transfer between M. barkeri and G. metallireducens (32). Conductive iron minerals were proposed to function as mediators between Geobacter and Methanosarcina species in a similar manner, based on the abundance of these genera in mineral-amended enrichment cultures (31). Here we provide evidence that M. barkeri does not require a conductive mediator for DIET because it is capable of forging direct biological electrical connections to receive electrons from G. metallireducens.

MATERIALS AND METHODS

Strains and conditions for cultivation.

Methanosarcina barkeri (DSM 800) and Pelobacter carbinolicus (DSM 2380) were purchased from the DSMZ culture collection. Geobacter metallireducens strain ATCC 53774 as well as gene deletion strains in which either the gene for PilA, the structural pilin protein (Gmet 1399) gene (33), or the Gmet 1868 gene, encoding a putative outer surface c-type cytochrome (34), was deleted were obtained from our laboratory culture collection. Strict anaerobic culturing procedures with either pressure tubes or serum bottles sealed with thick butyl rubber stoppers were used throughout.

M. barkeri was grown on 30 mM acetate in a modification of DSM 120 medium, which was determined to also support good growth of G. metallireducens. The DSM 120 modifications were as follows: 0.5 mM sulfide, 1 mM cysteine, 0.002 g/liter CaCl₂·2H₂O, 1 g/liter NaCl, no yeast extract, no Casitone, no resazurin, and 2 g/liter NaHCO₃.

G. metallireducens wild-type and mutant strains were routinely cultured on a ferric citrate medium amended with 10 mM acetate (33). However, prior to initiating cocultures with M. barkeri, the G. metallireducens strains were adapted for at least three transfers (5 to 10% inoculum) in ferric citrate medium in which the acetate was replaced with 20 mM ethanol.

P. carbinolicus was routinely grown on 10 mM acetoin, as previously described (20).

The coculture inocula were obtained from cultures that were in late exponential or early stationary phase, as determined by monitoring methane production (M. barkeri), Fe(II) production (G. metallireducens), or the optical density at 600 nm (OD₆₀₀) (P. carbinolicus).

The G. metallireducens-M. barkeri cocultures were initiated with a 0.5-ml inoculum of the ethanol-adapted G. metallireducens and a 0.5-ml inoculum of an acetate-grown M. barkeri culture added to 9 ml of the modified DSM 120 medium described above, but with the acetate replaced with 20 mM ethanol as the electron donor. CO₂ or HCO₃⁻ was the only potential electron acceptor. After the coculture was established, it was routinely transferred into 45 ml of medium with a 10% inoculum. Coculture growth was also tested in the absence of sulfide and cysteine, but for long-term incubations, both sulfide and cysteine are required as sulfur sources for M. barkeri (35–37) and were therefore added consistently to the coculture medium. When noted, GAC was added as previously described (32). Similar to GAC-free cocultures, 9 ml medium with GAC was inoculated with 0.5 ml of a G. metallireducens strain culture and 0.5 ml of M. barkeri culture.

M. barkeri grew in the same medium as P. carbinolicus, but P. carbinolicus could not grow in DSM 120 medium or the modified version of it. Therefore, the P. carbinolicus-M. barkeri cocultures were initiated with a 5% inoculum of acetoin-grown P. carbinolicus and a 5% inoculum of acetate-grown M. barkeri in the previously described P. carbinolicus medium (20). After establishing the coculture, P. carbinolicus and M. barkeri were routinely transferred together using 10% inocula.

Analytics.

Headspace methane (0.5 ml) was sampled at regular intervals using strict anaerobic sampling techniques (38), injected on a Shimadzu 8A gas chromatograph (GC) equipped with an 80/100 Hayasep Q column heated at 110°C. The injector port and flame ionization detector (FID) were both set at 200°C (15). To determine the concentration of organic acids and ethanol, 200 μl of culture medium was sampled, diluted, and placed in tightly sealed vials specially designed for high-performance liquid chromatography (HPLC) analysis (for fatty acids) and GC analysis (for ethanol). Organic acid samples were analyzed immediately or stored at 4°C for a maximum of 1 week. HPLC analyses were performed as previously described (39) by separating the organic acids on an Aminex NPX-87H column using 8 mM H₂SO₄ as the eluent. Organic acids were detected at 210 nm by a UV detector. Under these conditions, the following organic acids could be monitored: acetate, formate, succinate, fumarate, lactate, propionate, and butyrate. The detection limit for all was ca. 5 μM. In all our samples, only acetate was detected. Gas chromatography analyses of ethanol were performed on a PerkinElmer Clarus600 GC equipped with a headspace automatic sampler and an FID. Separation of ethanol was obtained on an Elite 5 (PerkinElmer) column (30-m length, 0.25-mm inner diameter) using He as the carrier gas and the following separation parameters: 50°C for 1 min, a ramp of 12°C per minute to reach 200°C, and 1.5 min at 200°C. The injector and detector temperatures were set at 200°C and 300°C, respectively. The chromatography data from GC and HPLC analyses were analyzed with an integrated TotalChrom data analysis system.

The increase in aggregate biomass over time was determined as previously described (40) by harvesting (15 min, 3,600 × g) 50 ml of three M. barkeri-G. metallireducens cocultures at different growth stages: during the initial growth phase (day 5) and in stationary phase (day 28). Aggregates were washed in isotonic buffer and freeze-dried for 48 h on a Labconco lyophilizer at −50°C. The weighted dry biomass was solubilized in 0.5% SDS by steaming. The total protein in the steamed biomass was determined using the Pierce bicinchoninic acid (BCA) protein assay kit according to the manufacturer's instructions. Bovine serum albumin was used as a protein standard.

Microscopy.

To evaluate cell-to-cell contact, we checked unfixed cells immediately after removing them from a culture tube and visualized the cultures with phase-contrast microscopy on a Leica Axioplann microscope. The presence of the methanogens was verified by fluorescence at 420 nm and identification of their specific coccus shape. To better visualize how cells were distributed within the aggregates, we performed fluorescence in situ hybridization. Cells were sampled in mid-log phase, fixed in premixed glutaraldehyde (0.5%) and paraformaldehyde (1%), rinsed in 50 mM PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid)] buffer, dehydrated (30 min at 4°C in an ethanol series at 30%, 50%, 70%, 80%, and 100%), and preembedded in a 50:50 mix of ethanol (100%) and butyl-methyl-methacrylate (BMM). Ultimately, aggregates were embedded in BMM resin and polymerized on a Spectorlinker cross-link XL-1000 instrument at room temperature overnight. Sections of the aggregates were obtained by microtome and stained with Cy3 probes (red) targeting Methanosarcina species (MS1414, 5′-CTCACCCATACCTCACTCGGG-3′) (41) and Cy5 probes (green) specific for the Geobacter cluster (Geo3ABC, 5′-CCGCAACACCTRGTWCTCATC-3′) (42) or Cy5 (green)-labeled probes specific for P. carbinolicus (PCARB1, 5′-GCCTATTCGACCACGATA-3′) (42) depending on the coculture. The labeled cocultures were visualized using a confocal laser microscope as previously described (20).

RESULTS AND DISCUSSION

Growth of M. barkeri via DIET with G. metallireducens.

Methanosarcina barkeri was cocultured with Geobacter metallireducens in a medium with ethanol as the sole electron donor in order to determine whether it was capable of DIET. Multiple lines of evidence have indicated that G. metallireducens is unable to release the electrons derived from the oxidation of ethanol to acetate as either H₂ or formate but can directly transfer these electrons to microorganisms such as G. sulfurreducens or Methanosaeta harundinacea, which are capable of DIET (15, 19–21). After a lag of ca. 39 days, cocultures of G. metallireducens and M. barkeri started metabolizing ethanol to methane (Fig. 1a). Neither G. metallireducens nor M. barkeri metabolized ethanol in pure culture (Fig. 1b). With continued transfer of the coculture, the initial long lag in ethanol metabolism was eliminated, and ethanol began to be metabolized to methane in less than 7 days (Fig. 1c).

FIG 1.

Syntrophic growth of M. barkeri with G. metallireducens. (a) Metabolism of ethanol when the cocultures were initiated with a 5% (vol/vol) inoculum of each species. (b) Lack of ethanol metabolism with single species. (c) Decreased lag after 25 transfers (10% inoculum) of the cocultures. All cultures were grown in triplicate (n = 3). A difference in the y axis scale can be noted between panels a and b and panel c. This is because of differences in the total incubation volumes: 10 ml to initiate the cocultures (a, b) and 50 ml for later incubations (c).

Acetate only transiently accumulated in the cocultures (Fig. 1a and c). The total ethanol consumed (mean ± standard deviation, 0.95 ± 0.01 mmol; n = 4) resulted in 1.59 ± 0.04 mmol methane, consistent with the expected production of 1.5 mol of methane for each mole of ethanol consumed according to the following equation: 2 C₂H₆O + H₂O → 3 CH₄ + CO₂.

This stoichiometric conversion of ethanol to methane requires that M. barkeri not only metabolize the acetate that G. metallireducens produces from ethanol (2 C₂H₆O + 2 H₂O → 2 C₂H₄O₂ + 8 H⁺ + 8 e⁻ via the reaction 2 C₂H₄O₂ → 2 CH₄ + 2 CO₂) but also use the electrons released from ethanol metabolism to reduce carbon dioxide to methane: 8 H⁺ + 8 e⁻ + CO₂ → CH₄ + 2 H₂O. Thus, the high electron recovery of electrons available from ethanol in methane and the fact that G. metallireducens is incapable of producing H₂ during ethanol metabolism (22) strongly suggested that M. barkeri was accepting electrons for carbon dioxide reduction via DIET.

The fact that the coculture could be continually transferred (10% inoculum) indicated that a small fraction of the carbon and electrons derived from ethanol must also be consumed for biomass production. Biomass was examined by sacrificing triplicate cocultures for protein analysis at day 5, when methane started increasing, and at day 28, when methane production was complete. Protein increased from 233 ± 62 μg ml⁻¹ (mean ± standard deviation) to 861 ± 115 μg ml⁻¹, verifying coculture growth.

Cells involved in DIET need physical connections. G. metallireducens and M. barkeri cocultures formed aggregates of ca. 100- to 200-μm diameter (Fig. 2a) in which Methanosarcina and Geobacter were intertwined (Fig. 2b).

FIG 2.

Coculture composition. (a) Phase-contrast micrograph of an aggregate of M. barkeri (cocci) and G. metallireducens (rods). (b) Scanning laser confocal microscope image of coculture aggregate after in situ hybridization, which targeted Methanosarcina cells with a red probe (Cy3) and G. metallireducens cells with a green probe (Cy5). Images are representative of triplicate samples taken during the mid-exponential growth phase of the cocultures.

Electrical connections via outer surface proteins and granular activated carbon.

Previous studies have demonstrated that deleting pilA, which encodes the structural protein of the conductive pili of Geobacter species, eliminates the ability of G. metallireducens to participate in DIET (15, 21). Cocultures initiated with a pilA-deficient strain (33) of G. metallireducens failed to metabolize ethanol or produce methane (Fig. 3a). Furthermore, cocultures initiated with a strain of G. metallireducens in which a gene (Gmet 1868 gene) for an outer surface cytochrome required for Fe(III) oxide reduction was deleted (34) also failed to metabolize ethanol to methane (Fig. 3b).

FIG 3.

Mutant studies. (a) Lack of ethanol metabolism and methane production in a coculture initiated with a pilA-deficient G. metallireducens strain. (b) Methane production in cocultures initiated with wild-type G. metallireducens or a strain deficient in the Gmet 1868 gene, which encodes a predicted extracellular cytochrome. (c) GAC stimulation of methane production in cocultures initiated with wild-type G. metallireducens and enabling methane production in cocultures initiated with either the PilA- or cytochrome (Gmet 1868)-deficient strains of G. metallireducens. All incubations were set up with a 5% (vol/vol) inoculum of each coculture partner.

However, cocultures initiated with either the pilA-deficient or Gmet 1868-deficient strains of G. metallireducens did effectively produce methane in the presence of granular activated carbon (GAC) (Fig. 3c). This is consistent with previous studies that demonstrated that GAC, which is electrically conductive, served as an electrical conduit for DIET between G. metallireducens and G. sulfurreducens, permitting the two species to share electrons even when the coculture was initiated with strains of G. sulfurreducens that could not produce pili or the pilin-associated c-type cytochrome OmcS (32). As previously noted (32), GAC amendments also stimulated the conversion of ethanol to methane in cocultures initiated with wild-type G. metallireducens and M. barkeri (Fig. 3c), substantially reducing the 39-day lag phase observed in the absence of GAC (Fig. 1a). Methane production of the cocultures initiated with the pilA-deficient or Gmet 1868-deficient strains of G. metallireducens was slower than that of the wild-type cocultures amended with GAC, but even in the absence of pili or the outer surface c-type cytochrome, methane production rates in the presence of GAC required a <6-day lag period to initiate ethanol metabolism (Fig. 3b). Previous studies demonstrated that in cocultures of G. metallireducens with either M. barkeri or G. sulfurreducens, both syntrophic partners attached to the GAC surface, but the individual cells were too far separated for biological electrical connections between the species (32). These results suggest that GAC can substitute for pili and/or outer surface cytochromes as the electrical conduit between the electron-donating G. metallireducens and the electron-accepting M. barkeri.

Growth of M. barkeri via HIT with P. carbinolicus.

Like G. metallireducens, P. carbinolicus can grow on ethanol in the presence of a syntrophic partner, but it relies on HIT rather than DIET (20, 43). P. carbinolicus and M. barkeri grew in coculture in medium with ethanol as the sole electron donor, but unlike G. metallireducens-M. barkeri cocultures, there was a steady accumulation of acetate (Fig. 4 left panel). This was probably due to the fact that acetate metabolism in M. barkeri is repressed when H₂ is available (44–46).

FIG 4.

Syntrophic growth in cocultures of P. carbinolicus and M. barkeri. (Left) Syntrophic growth in the third transfer of the coculture. (Center) Phase-contrast image illustrating the lack of aggregation of P. carbinolicus (rods) and M. barkeri (cocci). (Right) Scanning laser confocal microscope image of coculture after in situ hybridization, which targeted Methanosarcina cells with a red probe (Cy3) and P. carbinolicus cells with a green probe (Cy5). Images are representative of triplicate samples taken during the mid-exponential growth phase of the cocultures.

P. carbinolicus metabolizes ethanol to acetate and H₂ according to the following equation (20): 2 C₂H₆O + 2 H₂O → 2 C₂H₄O₂ + 4 H₂. The metabolism of ethanol to acetate and H₂ is exergonic only if M. barkeri can oxidize H₂ with the reduction of carbon dioxide as follows: 4 H₂ + CO₂ → CH₄ + H₂O.

The stoichiometry of ethanol consumed (mean ± standard deviation, 0.91 ± 0.07 mmol; n = 4), methane produced (0.41 ± 0.07 mmol), and acetate accumulated (0.66 ± 0.07 mmol) was consistent with methane being produced almost exclusively from H₂ oxidation with the reduction of carbon dioxide, assuming that some of the acetate produced was diverted to biomass production. A similar partial oxidation of ethanol with the accumulation of acetate was observed in the ethanol-metabolizing cocultures of P. carbinolicus and G. sulfurreducens, which were dependent on HIT (20), whereas acetate did not accumulate in G. metallireducens-G. sulfurreducens cocultures that relied on DIET (20).

In contrast to the G. metallireducens-M. barkeri cocultures (Fig. 2a and b), the P. carbinolicus-M. barkeri cocultures did not form multispecies aggregates (Fig. 4b and c). In a similar manner, P. carbinolicus-G. sulfurreducens cocultures did not aggregate (20), whereas G. metallireducens-G. sulfurreducens cocultures did (19). These results are consistent with the need for direct cell-to-cell contact for DIET, but not for HIT.

Implications.

The results demonstrate that M. barkeri is able to participate in DIET. It is only the second methanogen found to have this capability. Methanosarcina species are important contributors to methane production in some anaerobic digesters (29) as well as soils and sediments (7–10). Thus, the potential for DIET should be considered when attempting to promote the growth of Methanosarcina in digesters or when modeling their activity in soils and sediments.

In previous pure-culture studies, Methanosaeta harundinacea, which is incapable of utilizing H₂ as an electron donor, was found to be capable of DIET, but the two known H₂-utilizing methanogens tested, Methanospirillum hungatei and Methanobacterium formicicum, were not (15). Thus, M. barkeri is more versatile than other methanogens that have been evaluated for DIET, with the capacity to either accept electrons in the form of H₂ from syntropic partners like Desulfovibrio (46) and Pelobacter (this study) or to participate in DIET.

In the studies on DIET with M. harundinacea, acetate was available as an additional substrate for methane production (15), as it was in the M. barkeri cocultures reported here. However, studies with cocultures of G. metallireducens and a strain of G. sulfurreducens unable to oxidize acetate demonstrated that electrons derived from DIET could support fumarate reduction and growth of G. sulfurreducens in the absence of acetate as an additional electron donor (22). Similar studies should be feasible with an M. barkeri mutant unable to metabolize acetate (47) to determine whether carbon dioxide reduction to methane with electrons derived from DIET provides sufficient energy to support cell maintenance and growth. Other than the obvious importance of DIET in some anaerobic digesters treating brewery wastes (15, 18), the prevalence of DIET in other methanogenic communities is unknown. The fact that at least two major genera of methanogens have evolved the capacity for DIET suggests that there are conditions in some soils and sediments in which DIET confers a selective advantage.

The ability of GAC to promote methanogenic DIET has important implications for the design of anaerobic digesters. Previous studies that have reported stimulation in methane production when carbon cloth was introduced into digesters have attributed this response to the cloth retaining methanogens (48–50). However, the possibility that these conductive materials might promote DIET should also be considered.

Although the importance of electrically conductive pili and outer surface cytochromes in extracellular electron exchange, including DIET, is becoming increasingly apparent for Geobacter species (51, 52), the potential extracellular electron contacts that might permit methanogens to accept electrons via DIET are less clear. Methanosaeta and Methanosarcina are the only genera of methanogens with membrane-bound cytochromes (53, 54) that could potentially play a role in extracellular electron exchange. The availability of tools for genetic manipulation of M. barkeri (55) suggests that it may be the ideal candidate for functional analysis of DIET mechanisms in methanogens.