Community Composition and Ultrastructure of a Nitrate-Dependent Anaerobic Methane-Oxidizing Enrichment Culture

ABSTRACT

Methane is a very potent greenhouse gas and can be oxidized aerobically or anaerobically through microbe-mediated processes, thus decreasing methane emissions in the atmosphere. Using a complementary array of methods, including phylogenetic analysis, physiological experiments, and light and electron microscopy techniques (including electron tomography), we investigated the community composition and ultrastructure of a continuous bioreactor enrichment culture, in which anaerobic oxidation of methane (AOM) was coupled to nitrate reduction. A membrane bioreactor was seeded with AOM biomass and continuously fed with excess methane. After 150 days, the bioreactor reached a daily consumption of 10 mmol nitrate · liter⁻¹ · day⁻¹. The biomass consisted of aggregates that were dominated by nitrate-dependent anaerobic methane-oxidizing “Candidatus Methanoperedens”-like archaea (40%) and nitrite-dependent anaerobic methane-oxidizing “Candidatus Methylomirabilis”-like bacteria (50%). The “Ca. Methanoperedens” spp. were identified by fluorescence in situ hybridization and immunogold localization of the methyl-coenzyme M reductase (Mcr) enzyme, which was located in the cytoplasm. The “Ca. Methanoperedens” sp. aggregates consisted of slightly irregular coccoid cells (∼1.5-μm diameter) which produced extruding tubular structures and putative cell-to-cell contacts among each other. “Ca. Methylomirabilis” sp. bacteria exhibited the polygonal cell shape typical of this genus. In AOM archaea and bacteria, cytochrome c proteins were localized in the cytoplasm and periplasm, respectively, by cytochrome staining. Our results indicate that AOM bacteria and archaea might work closely together in the process of anaerobic methane oxidation, as the bacteria depend on the archaea for nitrite. Future studies will be aimed at elucidating the function of the cell-to-cell interactions in nitrate-dependent AOM.

IMPORTANCE Microorganisms performing nitrate- and nitrite-dependent anaerobic methane oxidation are important in both natural and man-made ecosystems, such as wastewater treatment plants. In both systems, complex microbial interactions take place that are largely unknown. Revealing these microbial interactions would enable us to understand how the oxidation of the important greenhouse gas methane occurs in nature and pave the way for the application of these microbes in wastewater treatment plants. Here, we elucidated the microbial composition, ultrastructure, and physiology of a nitrate-dependent AOM community of archaea and bacteria and describe the cell plan of “Ca. Methanoperedens”-like methanotrophic archaea.

INTRODUCTION

Anthropogenic influence on the carbon and nitrogen cycles has vastly increased since the industrial revolution, resulting in a significant contribution to greenhouse gas emissions and global warming (1–3). Methane is a very potent greenhouse gas, with a global warming potential 25 to 30 times higher than that of carbon dioxide on a scale of 100 years (4). Methane can be oxidized aerobically or anaerobically (5–10) through microbe-mediated processes. Over the last decade, the significance of anaerobic oxidation of methane (AOM) in decreasing methane release to the atmosphere became clear (8), and the microorganisms responsible for AOM were identified (11). Sulfate, nitrate, iron, and manganese can be used as electron acceptors for AOM (6, 9, 12–15). AOM coupled to nitrate reduction was first described in a coculture of anaerobic methanotrophic (ANME) archaea and bacteria of the NC10 phylum (6). Later studies showed that both the bacteria and archaea individually were capable of AOM (7, 16).

“Candidatus Methylomirabilis oxyfera,” the only cultured representative of the NC10 phylum, anaerobically oxidizes methane to carbon dioxide through a postulated intra-aerobic pathway (17). In these bacteria, nitrite is first reduced to nitric oxide, which is then proposed to be dismutated into oxygen and nitrogen (17). The produced oxygen is used to activate methane by a methane monooxygenase, following the canonical aerobic methane oxidation pathway (17). An ultrastructural study of M. oxyfera revealed that this species is a polygonal rod with sharp cell ridges along the cell length (18).

AOM coupled to nitrate reduction to nitrite was shown to be performed by the archaeon “Candidatus Methanoperedens nitroreducens” (16). Recently, a “Ca. Methanoperedens”-like archaeon, belonging to ANME-2d, was found to reduce nitrate to ammonium via nitrite and to couple methane oxidation to iron(III) and Mn(IV) reduction (12). ANME archaea are related to methanogenic Euryarchaeota (19) and metabolize methane through the hypothesized reverse methanogenesis pathway (16), in which methyl-coenzyme M reductase (Mcr) initiates the first step of methane oxidation.

Since nitrite and nitrate coexist in the interface of oxic and anoxic environments, it was speculated that “Ca. Methylomirabilis”-like bacteria metabolize most of the nitrite produced by “Ca. Methanoperedens”-like archaea, thereby alleviating potential toxic effects of nitrite (12). The capability of this combination of AOM microorganisms to metabolize polluting nitrogenous substrates may find application in more sustainable wastewater treatment plants (20).

As only a few studies have investigated the community composition of nitrate AOM cultures and none have addressed the cell plan of the “Ca. Methanoperedens”-like archaea, we studied the microbial community composition and the ultrastructure of a nitrate-reducing AOM enrichment culture and its key players. We used light and electron microscopy techniques, physiological experiments, and phylogenetic analysis to characterize the microbial community and provide insight into their metabolic interactions.

RESULTS

Enrichment of a methane-oxidizing enrichment culture.

A membrane bioreactor was started with 1 liter of biomass (51 mg/liter protein) and continuously fed with excess methane. The nitrate concentration in the influent was increased from 0.5 to 40 mM according to the reactor performance in 150 days (Fig. 1). During the first 50 days, nitrate and nitrite accumulation occurred intermittently. After 150 days, the bioreactor reached a daily consumption of 10 mmol nitrate · liter⁻¹ · day⁻¹ at a hydraulic retention time of 4 days. The hydraulic retention time and increases in the concentration of the supplied substrates were adjusted according to the conversion efficiency of the bioreactor; when there was no detectable nitrate, either substrate concentration was increased or hydraulic retention time was decreased. An ammonium accumulation was observed ranging from 0.5 to 1 mM, which corresponded to, at most, 10% of the nitrate concentration in the influent. After 1 year, the protein concentration in the reactor was 2.2 mg/liter.

FIG 1.

Concentrations (mM) of ammonium, nitrite, and nitrate over time (in days) in a continuous anaerobic methane-oxidizing membrane bioreactor. The broken line represents the concentration of nitrate in the medium. Closed squares, closed circles, and open diamonds represent the concentrations of nitrate, nitrite, and ammonium in the reactor, respectively.

Phylogenetic analysis.

To investigate the microbial composition of the enrichment culture, we extracted the total DNA from the bioreactor using both the cetyltrimethylammonium bromide (CTAB)-based extraction method (21) and the PowerSoil kit (Mo Bio, Carlsbad, CA, USA). The amount of reads obtained with the CTAB method after trimming (1,531,534 reads) was higher than that obtained using the PowerSoil kit (527,918 reads). With the PowerSoil method, 93% of the reads were bacterial and only 7% were archaeal, whereas with the CTAB method, only 2% of the reads were archaeal (Table 1). Mapping the sequence reads to the Silva 16S rRNA database indicated two dominant microorganisms in the enrichment culture: NC10 bacteria (mainly “Ca. Methylomirabilis” spp., with 44.4% for CTAB and 34% for PowerSoil kit) was the most represented bacterial phylum, and “Ca. Methanoperedens”-like archaea were the only archaea detected in the enrichment culture. To determine the other phyla within the community, we set a threshold at ≥1%. Proteobacteria (15% for CTAB and 19% for PowerSoil kit), Planctomycetes (15% for CTAB and 18% for PowerSoil kit), Chlorobi (7.6% for CTAB and 7% for PowerSoil kit), Chloroflexi (4.7% for CTAB and 3% for PowerSoil kit), Bacteroidetes (1.8% for CTAB and 2.1% for PowerSoil kit), and Acidobacteria (2.7% for CTAB and 1.8% for PowerSoil kit) were detected.

TABLE 1.

Abundances of the different microbial communities at the phylum level in the AOM bioreactor

Phylum	Abundance by DNA extraction method (%)
	CTAB	PowerSoil kit
ANME-2d	2.2	7.6
Acidobacteria	2.7	1.8
Bacteroidetes	1.8	2.1
Chloroflexi	4.7	3
Chlorobi	7.6	7
NC10	44.4	34
Planctomycetes	15	18
Proteobacteria	15	19
Other	6.7	7

Near-complete sequences of two “Ca. Methylomirabilis” strains and one “Ca. Methanoperedens”-like species were identified. Based on coverage, the two “Ca. Methylomirabilis” strains were present in about equal amounts, with one being most closely related to the original “Ca. Methylomirabilis oxyfera” (6) and the other a new “Ca. Methylomirabilis” species (96% identity for the 16S rRNA gene). The “Ca. Methanoperedens”-like species was very closely related to a strain described before (12, 22) and showed 95% identity to the 16S rRNA gene of “Candidatus Methanoperedens nitroreducens” (16).

Community (ultra)structure.

The enrichment culture was examined using light and electron microscopy-based approaches to investigate the ultrastructure of the community and of the most abundant microorganisms. First, fluorescence in situ hybridization (FISH) was performed by combining probes targeting “Ca. Methylomirabilis”-like bacteria (DAMOBACT-0193), “Ca. Methanoperedens”-like archaea (DAMOARCH-0641), most Bacteria (EUB 338 [23], EUB 338 II and III [24]), and most Archaea (S-D-Arch-0915-a-A-20). Based on the FISH analysis, the community was dominated by “Ca. Methylomirabilis”-like bacteria (approximately 50%; Fig. 2A in cyan) and “Ca. Methanoperedens”-like archaea (approximately 40%; Fig. 2A in red). Bacteria not belonging to “Ca. Methylomirabilis” were also detected and comprised about 10% of the total community, whereas no archaea other than “Ca. Methanoperedens”-like archaea were detected. “Ca. Methylomirabilis”-like bacteria and “Ca. Methanoperedens”-like archaea occurred together in aggregates, in which “Ca. Methanoperedens”-like cells formed a cauliflower-shaped aggregate and “Ca. Methylomirabilis”-like bacteria surrounded the “Ca. Methanoperedens”-like cells (Fig. 2B and C).

FIG 2.

Fluorescence in situ hybridization of biomass from the AOM enrichment culture. Epifluorescent micrographs taken with ×40 (A and B) and ×63 (C) magnification after hybridization with probes DAMOBACT-0193 (dye) for “Ca. Methylomirabilis”-like bacteria; DAMOARCH-0641 (dye) for “Ca. Methanoperedens”-like archaea; EUB 338, EUB338 II, and EUB338 III (dye) for most bacteria; and S-D-Arch-0915-a-A-20 (dye) for most archaea. (A) “Ca. Methanoperedens”-like archaea are depicted in red, “Ca. Methylomirabilis”-like bacteria are in cyan, and other bacteria are in blue. (B and C) “Ca. Methanoperedens”-like archaea are depicted in orange, and “Ca. Methylomirabilis”-like bacteria are in green. Scale bar = 20 μm.

Electron microscopy of cryo-fixed, freeze-substituted, and resin-embedded cells was used to investigate the ultrastructure of the microbial population (Fig. 3, and see Fig. S1 in the supplemental material). All cells in the community were embedded in a matrix of extracellular polymeric substances (EPS). “Ca. Methylomirabilis” sp. bacteria could be recognized by their polygonal cell shape and surrounded the compact “Ca. Methanoperedens” sp. aggregates, in line with the observation of the fluorescence light microscopy. The other microorganisms present in the community showed a number of different ultrastructures, as shown in Fig. S1.

FIG 3.

Transmission electron micrographs of high-pressure frozen, freeze-substituted, resin-embedded, and thin-sectioned cells from the AOM enrichment culture. (A) Overview of the enrichment culture showing aggregates of “Ca. Methanoperedens” sp. archaea at the top and “Ca. Methylomirabilis” sp. bacteria at the bottom. (B) “Ca. Methylomirabilis” sp. bacterium. (C) “Ca. Methanoperedens” sp. cells showing putative cell-to-cell contacts (arrows) and tubules (arrowhead). Scale bar = 5 μm (A), 200 nm (B and C).

Cell biology of “Ca. Methanoperedens”-like archaea.

Methyl-coenzyme M reductase (Mcr), which is the methane-activating enzyme in archaeal methanotrophs (25), was used to identify the “Ca. Methanoperedens” sp. archaea in the enrichment culture. The immunogold localization of Mcr resulted in an abundant and specific labeling of the cytoplasm of “Ca. Methanoperedens” sp. cells, with very low background labeling (Fig. 4). “Ca. Methanoperedens” spp. were slightly irregular coccoid cells (approximately 1.5 μm diameter) occurring in aggregates. In the aggregate, the cells were surrounded by a thick EPS matrix, which often distorted the round cell shape. No flagella or other appendages were observed. “Ca. Methanoperedens” sp. archaea contained an electron-dense, low-contrast, and compact cytoplasm with evenly distributed ribosomes and no intracellular structures. Putative cell-to-cell contacts were observed between adjoining “Ca. Methanoperedens” sp. cells in the aggregates (Fig. 3C). Since “Ca. Methanoperedens”-like archaea were recently shown to encode a large amount of multiheme c-type cytochromes (12), we performed a cytochrome staining on “Ca. Methylomirabilis” sp. and “Ca. Methanoperedens” sp. cells to localize their cytochromes. Figure 5A and C show “Ca. Methanoperedens” sp. archaea and “Ca. Methylomirabilis” sp. bacteria, respectively, in which cytochromes were stained with 3,3′ diaminobenzidine (DAB), compared to cells in Fig. 5B and D, in which potassium cyanate was used to inhibit the cytochrome staining (negative control). Compared to the negative control, the cytoplasm of cytochrome-stained “Ca. Methanoperedens” sp. archaea appeared darker, especially in close proximity to the cytoplasmic membrane. No increased electron density was observed in the EPS that embedded the “Ca. Methanoperedens” sp. cells compared to the negative control. In “Ca. Methylomirabilis” sp. cells, the cytochrome staining was confined to the periplasmic area, with no staining in the cytoplasm or in the surrounding EPS (Fig. 5C).

FIG 4.

Transmission electron micrographs of high-pressure frozen, freeze-substituted, resin-embedded and thin-sectioned “Ca. Methanoperedens” sp. cells showing immunogold localization of methyl-coenzyme M reductase (Mcr). (A) Mcr was labeled with protein A-gold and localized in the cytoplasm of “Ca. Methanoperedens” sp. archaea (see arrowhead for example of a gold label); (B) negative control showing no labeling. Scale bar = 200 nm.

FIG 5.

Transmission electron micrographs showing the cytochrome staining on “Ca. Methanoperedens” sp. archaea (A and B) and “Ca. Methylomirabilis” sp. bacteria (C and D). Panels B and D are negative controls of panels A and C, respectively. The cytochrome staining stains the cytoplasm in “Ca. Methanoperedens” sp. archaea (A) and is most intense in close proximity with the cytoplasmic membrane. In “Ca. Methylomirabilis” sp. bacteria (C), the periplasm is stained. Scale bar = 200 nm.

To gain further insight into the cell biology of “Ca. Methanoperedens” sp. archaea and to investigate the putative cell-to-cell contacts observed in resin-embedded cells, we used electron tomography (Fig. 6 and Movie S1 in the supplemental material). We observed that the cell walls of three out of five imaged thick sections of “Ca. Methanoperedens” sp. cells came in contact (Fig. 6C and D) and at times seemed to fuse together. In addition, these archaea produced tubular structures (32.8 ± 6.5 nm wide, calculated on 28 tubular structures on thin sections). In some cases, we observed continuity between the membrane of the archaeon and of the newly formed tubule. The tubular structures extruded in the surrounding EPS (Fig. 6E and F).

FIG 6.

Snapshot of electron tomogram (A) and model (B) of “Ca. Methanoperedens” sp. cells. The cell walls of “Ca. Methanoperedens” sp. archaea (blue) appear to contact each other in putative cell-to-cell contacts (white arrowheads in B and C). Tubular structures (purple) originate or fuse together with a “Ca. Methanoperedens” sp. cells (yellow arrowheads in B and E). (C and D) Close-up images showing the putative cell-to-cell contacts. (E and F) Close-up images of the extruding tubule. Scale bar = 50 nm (A), 25 nm (C and E).

DISCUSSION

Methane is an important greenhouse gas, and anaerobic methane oxidation by methanotrophs may significantly reduce methane emission into the atmosphere (11). To get more insight into the process of nitrate-dependent methane oxidation, we studied the community composition and cell biology of an enrichment culture dominated by “Ca. Methanoperedens”-like archaea and “Ca. Methylomirabilis”-like bacteria.

Phylogenetic analysis.

To investigate the community composition of the enrichment culture, we performed a phylogenetic analysis which focused on the 16S rRNA genes. The archaeal population consisted only of “Ca. Methanoperedens” sp. archaea. The bacterial population was heterogeneous but dominated by “Ca. Methylomirabilis” sp. bacteria, followed by Planctomycetes, Proteobacteria, Chlorobi, Chloroflexi, Acidobacteria, and Bacteroidetes. The 16S rRNA gene analysis revealed that two “Ca. Methylomirabilis” sp. strains were present in the enrichment culture. No ultrastructural differences were observed between the two “Ca. Methylomirabilis” sp. strains. To reduce possible methodological bias, two different DNA extraction methods were used. Even though the two methods identified the same most abundant microorganisms, they yielded small amounts of archaeal DNA, severely underestimating the amount of archaeal cells in the enrichment culture. In the 16S rRNA gene analyses, we found only 2.2% and 7.6% archaeal reads from CTAB and PowerSoil kit DNA extraction, respectively, which was much lower than the 40% observed with FISH analysis. DNA extraction methods that include mechanical lysis (such as the PowerSoil kit) have been shown to yield more (archaeal) DNA in the cases of environmental samples (26) and to provide less biased archaeal community results (27). Probably due to the cell walls of archaea being more rigid than those of bacteria (28) and the presence of a thick matrix surrounding the cells (8), the archaeal population might be underestimated in community studies. However, even in the case of the PowerSoil kit, we obtained much less archaeal DNA than expected from FISH analysis. From this, we conclude that even harsher DNA extraction methods are needed (with more intense mechanical shearing) to obtain more representative results for the archaeal abundance. On the other hand, when using even harsher DNA extraction methods, there is a high risk of bacterial DNA shearing (29). Surveys using standard DNA extraction methods also might grossly underestimate the abundances of these archaea in natural and man-made ecosystems. Next to the bias introduced by the DNA extraction method, the presence of multiple 16S rRNA gene copies per genome in certain microorganisms can also give rise to a misrepresentation of the actual abundance upon phylogenetic analysis.

Nitrogen conversions.

In order to determine the relative contribution of the methane-oxidizing microorganisms to nitrogen conversion in our enrichment culture, we monitored nitrate, nitrite, and ammonium concentrations in the continuous membrane bioreactor. After a 150-day adaptation period to the new growth conditions, the enrichment culture reached a nitrate reduction rate of 10 mmol · liter⁻¹ · day⁻¹, which was 20 times higher than the rate of the enrichment culture described previously (12). Approximately 10% of the reduced nitrate was recovered as ammonium, indicating that anaerobic ammonium-oxidizing bacteria were not present in the bioreactor, which was in line with the phylogenetic analysis of the culture. Ammonium accumulation was in agreement with previous results showing that “Ca. Methanoperedens”-like archaea reduce nitrate to ammonium via nitrite, most likely achieved by the concerted action of nitrate reductase (encoded by NarGH) and the ammonium-producing nitrite reductase (encoded by NrfAH) proteins (22). Still, when nitrate was supplied continuously, under apparent steady-state conditions, nitrite was not detected, unlike the transient accumulation observed previously when nitrate was added in pulses (12). This suggested that nitrate reduction to nitrite occurred at a higher rate than nitrite reduction to ammonium. The remaining 90% of the reduced nitrate was most likely converted via nitrite to nitrogen gas through the activity of nitrite-dependent “Ca. Methylomirabilis” sp. bacteria. This could be an indication that under these growth conditions, “Ca. Methylomirabilis” sp. bacteria were more efficient in converting nitrite than “Ca. Methanoperedens” sp. archaea.

Cell biology.

In addition to molecular and physiological studies, the microbial population was also investigated using light and electron microscopy, in particular focusing on the ultrastructure of the “Ca. Methanoperedens” sp. archaea. FISH and thin sections of cryo-fixed, freeze-substituted, and resin-embedded samples showed that the archaeal cells formed compact aggregates, and the bacteria surrounded the archaea.

Identification of “Ca. Methanoperedens”-like archaea.

To identify the “Ca. Methanoperedens”-like cells within the microbial community, we performed an immunogold labeling on resin-embedded sections to target Mcr. We used an anti-Mcr antibody developed against Mcr purified from the methanogen Methanosarcina barkeri (30) to target Mcr in “Ca. Methanoperedens”-like cells. The similarity of Mcr between M. barkeri and ANME archaea (31) has been shown to allow the specific targeting of ANME-2 cells (30). In agreement with previous observations (31), the Mcr-specific labeling was localized in the cytoplasm of “Ca. Methanoperedens”-like cells. As with other ANME-2 archaea (31), “Ca. Methanoperedens” sp. archaea appeared as slightly irregular coccoid cells of about 1.5 μm in diameter, growing in cauliflower-shaped aggregates and embedded in a thick matrix. Similar cell morphologies and aggregate structures have been described for the methanogens Methanosarcina mazei (32), Methanosarcina thermophila (33) and M. barkeri (34), of the order Methanosarcinales.

“Ca. Methanoperedens”-like archaeal ultrastructure.

Tubular structures and putative direct cell-to-cell contacts were observed between “Ca. Methanoperedens” sp. cells. The putative cell-to-cell contacts appeared as protrusions of the cell walls of two adjoining “Ca. Methanoperedens” sp. cells. In some cases, the two cell walls were intact at the interaction area between two cells, and in other cases, the cell walls appeared blurry and it was difficult to assess whether they were still intact or fused together. If cytoplasm sharing occurred between “Ca. Methanoperedens” sp. cells, these putative connections could serve as cytoplasmic bridges to transfer (genetic) material, as it was shown, for example, for the archaeon Haloferax volcanii (35). In thin sections, we also observed that “Ca. Methanoperedens” sp. cells were surrounded by vesicle-like structures. However, electron tomography revealed that these structures were in fact tubules still connected to the cell of origin or fused to another “Ca. Methanoperedens” sp. cell. Tubule formation is known to occur in bacteria and archaea (36, 37). Even though their role is often unclear, these structures are likely to play a fundamental role in microbial communities, for example, for their possible role in vesicle and network formation (38), DNA transfer (36), metabolite exchange (39), and electron transfer (40). The tubules produced by “Ca. Methanoperedens” sp. archaea had a diameter comparable to those observed in the extremophile archaeon Pyrodictium, for which a role in cell-to-cell communication and network formation has been hypothesized. Unfortunately, due to the recent discovery of “Ca. Methanoperedens”-like archaea and the lack of a pure culture, a detailed investigation of the function and nature of these putative cell-to-cell contacts is very challenging.

Cytochrome localization.

Genomes of ANME microorganisms, particularly “Ca. Methanoperedens”-like archaea, were recently shown to encode a high number of c-type cytochromes (12, 16, 22, 41). In previous studies (13, 42), ANME-1 and -2 archaea were suggested to promote electron transfer in a methanotrophic consortium with sulfate-reducing bacteria. In the light of these findings, we performed cytochrome staining on the enrichment culture. Cytochrome staining was most electron dense in close proximity to the cytoplasmic membrane of “Ca. Methanoperedens” sp. archaea, indicating that these areas were rich in cytochromes. This is the same area in which most of the c-type cytochromes of “Ca. Methanoperedens” sp. archaea are predicted to be located based on genome analysis (22). The cytochrome staining of “Ca. Methanoperedens” sp. cells did not show staining of the matrix surrounding the archaea. Therefore, unlike what has been observed for other types of ANME archaea (13, 42), we did not find indications that “Ca. Methanoperedens”-mediated extracellular electron transfer was occurring in our enrichment culture. In “Ca. Methylomirabilis” sp. bacteria, the cytochrome staining highlighted the periplasm. This is in line with the suggestion that cytochrome c proteins of “Ca. Methylomirabilis”-like bacteria (NirS, MxaF, and Hao) are transported to the periplasm after cleavage of their signal peptides (17).

Subpopulation of the enrichment culture.

The presence of a heterogeneous bacterial community was confirmed by observations of different morphologies in resin-embedded samples for electron microscopy (EM). In some cases, it was possible to speculate which morphology belonged to which phylum. For example, long filamentous bacteria (Fig. S1C) were described in the Proteobacteria, Chloroflexi, and Bacteroidetes phyla. Since filamentous Proteobacteria and Bacteroidetes are known to be less active or not active under anoxic conditions (43–45), the filamentous bacteria present in our enrichment culture might belong to the phylum Chloroflexi, which can thrive in anoxic environments (46, 47). Because Chloroflexi feed on organic compounds, it is possible that these are provided by the other members of the microbial community. Cells in Fig. S1B show a condensed nucleoid and a relatively large ribosome-free area between two lipid bilayers that is reminiscent of Planctomycetes (48). For other relevant bacterial phyla that emerged from the phylogenetic analysis of our enrichment culture, it is not possible to associate a likely morphology due to limited ultrastructural studies of these phyla or the heterogeneity within a particular phylum.

Concluding remarks.

In nitrate-dependent AOM communities, multiple microbial species interact with each other, but their relationships are complex and unknown. Here, we studied the composition and ultrastructure of a nitrate-dependent AOM community. In this continuous bioreactor system, bacteria and archaea reduce nitrate to nitrogen gas via their concerted activity. Although we observed putative cell-to-cell connections among individual “Ca. Methanoperedens” sp. cells, no direct cell-to-cell contacts were observed between bacteria and archaea. The heterogeneity of the system and the novelty of the microbes involved in the nitrate- and nitrite-dependent AOM processes leave plenty of room for further physiological and community studies. A better understanding of the physical and metabolic interactions taking place in these communities would shed light on how methane is oxidized in nitrate-rich environments and help design wastewater treatment plants where these microorganisms might find an application in biological methane removal.

MATERIALS AND METHODS

Enrichment conditions.

To study a microbial population capable of nitrate-dependent anaerobic methane oxidation, a membrane bioreactor (MBR; working volume, 2 liters) (Applikon Biotechnology BV, Applisens, Schiedam, the Netherlands) was seeded with biomass (1 liter, containing 51 mg/liter protein) from a culture consisting of about 40% “Ca. Methanoperedens”-like archaea and 40% “Ca. Methylomirabilis”-like bacteria (12). To supply methane and maintain anoxic conditions, the reactor was flushed continuously with CH₄-CO₂ (95/5%, 20 ml · min⁻¹). In addition, the medium was flushed continuously with Ar-CO₂ (95:5). The MBR was stirred at 100 rpm, and the temperature was maintained at 30°C with a water jacket. Mineral medium (adapted from reference 49) containing nitrate (0.5 to 40 mM according to consumption rates) was added continuously to the reactor at a flow rate of 20 ml · h⁻¹ using a peristaltic pump equipped with Neoprene tubing (Cole-Parmer, IL, USA). The medium (not sterilized and not filtered) contained, per liter: 1 g of KHCO₃, 0.05 g of KH₂PO₄, 0.3 g of CaCl₂·2H₂O, 0.2 g of MgSO₄·7H₂O, 0.5 ml of acidic trace element solution, and 0.2 ml of alkaline trace element solution. The acidic (100 mM HCl) trace element solution contained, per liter: 2.085 g of FeSO₄·7H₂O, 0.068 g of ZnSO₄·7H₂O, 0.12 g of CoCl₂·6H₂O, 0.5 g of MnCl₂·4H₂O, 0.32 g of CuSO₄, 0.095 g of NiCl₂·6H₂O, and 0.014 g of H₃BO₃. The alkaline (10 mM NaOH) trace element solution contained, per liter: 0.067 g of SeO₂, 0.050 g of Na₂WO₄·2H₂O, and 0.242 g of Na₂MoO₄.

Analytical methods.

Ammonium and nitrite samples taken at regular intervals were analyzed by colorimetric assays using spectrophotometry. Ammonium was determined at a 420-nm wavelength after a reaction with ortho-phthalaldehyde, with a detection limit of 500 μM, as previously described (50). Nitrite was determined at a 520-nm wavelength by the sulfanilamide reaction (Griess reaction), with a detection limit of 50 μM (51). Nitrate was measured by conversion into nitric oxide at 95°C using a saturated solution of VCl₃ in HCl (52). Nitric oxide was then measured using a nitric oxide analyzer (NOA280i; GE Analytical Instruments, Manchester, UK).

Phylogenetic analysis.

DNA was extracted from a 4-ml biomass of the enrichment culture using the CTAB-based extraction method (21) and the PowerSoil kit (Mo Bio, Carlsbad, CA, USA). DNA quality was checked by agarose gel electrophoresis (0.8%) gels stained with ethidium bromide. The DNA concentration was determined using a NanoDrop spectrophotometer (Thermo Scientific). Genomic DNA (100 ng) from each extraction method was sheared (6 cycles, with 1 min on, 1 min off) using sonication (Bioruptor; Diagenode, Liege, Belgium). Library preparation was performed using the Ion Plus fragment library kit (Thermo Scientific), according to the manufacturer's instructions. Size selection was performed using E-Gel SizeSelect agarose gel (Life Technologies). The size-selected libraries were then amplified using a OneTouch 400-bp kit and sequenced on the Ion Torrent PGM using the Ion PGM 400-bp sequencing kit and one Ion 318 version 2 Chip (Thermo Scientific). The sequencing resulted in 1,665,605 reads from the CTAB extraction method and 619,523 reads from the PowerSoil kit. The reads were imported into CLC Genomics Workbench (version 10; CLC bio, Aarhus, Denmark) and quality trimmed using standard settings, and reads shorter than 100 bp were discarded. This resulted in a data set with 1,531,534 reads from the CTAB extraction method and 527,918 reads from the PowerSoil kit. For 16S rRNA gene analysis, the trimmed reads were mapped (global alignment; mismatch penalty, 2; indel penalty, 3; 70% identity over 50% of the read length) to a 16S rRNA gene reference sequence database (Silva short subunit [SSU] Ref NR 119, 534,968 sequences) (https://www.arb-silva.de/documentation/release-119/). The 9,630 mapped reads were extracted and in a second step subjected to a BLAST search against the same database using an E value of 10⁻⁶. The resulting 1,122 sequences were aligned using the SINA aligner (53) and imported into an ARB version of the Silva SSU Ref NR 119 database (54). Sequences were added to the existing tree using the quick-add function and position variability filter bacteria implemented in ARB. For phylogenetic analysis, the 9,630 mapped reads were assembled using CLC with a word size of 15, a bubble size of 273, and a minimum contig length of 500 bp. The resulting contigs were compared with BLAST against the Silva SSU Ref NR 128 database to identify the 16S rRNA genes of both “Ca. Methanoperedens”-like archaea and “Ca. Methylomirabilis”-like bacteria. Alignments with related species and phylogenetic tree constructions were performed using MEGA6 (55).

FISH.

After 40 days from the start of the culture, 1.5 ml of biomass was harvested from the enrichment culture and centrifuged, and the pellet was washed twice with 1 ml of phosphate-buffered saline (PBS; 130 mM NaCl and 10 mM phosphate buffer [pH 7.4]). The samples were fixed with 900 μl of paraformaldehyde for 3 h at 4°C. Fluorescence in situ hybridization (FISH) was performed as previously described (7), using 20% formamide stringency. To facilitate the attachment of the granular biomass to the slides, silane-coated slides were used. The following oligonucleotide probes were used: DAMOBACT-0193 (CGC TCG CCC CCT TTG GTC), specific for “Ca. Methylomirabilis”-like bacteria; DAMOARCH-0641 (GGT CCC AAG CCT ACC AGT), specific for “Ca. Methanoperedens”-like archaea; EUB 338 (S-D-Bact-0338-a-A-18) (23), EUB 338 II (S-D-Bact_0338-b-A-18) (24), and EUB 338 III (S-D-Bact-0338-c-A-18) (24) for most bacteria; and S-D-Arch-0915-a-A-20 for most archaea. Images were collected with a Zeiss Axioplan 2 epifluorescence microscope equipped with a charge-coupled-device (CCD) camera, together with the AxioVision software package (Zeiss, Germany).

Cryo fixation, freeze substitution, Lowicryl embedding, sectioning, and poststaining.

After 60 days from the start of the culture, cells were placed into the 100-μm-deep cavity of a platelet (3-mm diameter, 0.1 to 0.2 mm depth; Leica Microsystems), closed with the flat side of a lecithin-coated platelet (3 mm diameter, 0.3 mm depth), and cryoimmobilized by high-pressure freezing (Leica HPM100). The platelets were stored in liquid nitrogen.

For Lowicryl embedding, frozen samples were freeze-substituted in 0.2% uranyl acetate in anhydrous acetone. The substitution started at −90°C for 48 h, was brought to −70°C at 2°C per h and kept at −70°C for 12 h, and was brought to −50°C at 2°C per h and kept at −50°C for 12 h in a freeze substitution unit (AFS2; Leica Microsystems, Vienna, Austria). To remove uranyl acetate, the samples were washed twice with 100% acetone for 30 min at −50°C. Keeping the temperature stable at −50°C, the sample was infiltrated with a dilution series of Lowicryl (10%, 25%, 50%, and 75%) in acetone. Each step was 1 h long. Three final infiltration phases were performed with 100% Lowicryl: first for 1.5 h, overnight, and then for 2 h. Polymerization of the resin was obtained by irradiating the sample with UV light for 96 h, after which the temperature was brought to 0°C in 24 h at 2.1°C per h. UV light was switched off, and the temperature was brought to 20°C in 5 h at 4°C per h. Ultrathin sections of 55 nm were cut using a Leica UCT microtome (Leica Microsystems) and collected on carbon-Formvar-coated 100-mesh hexagonal square copper grids (reference number G2100C; Agar Scientific). The sections were poststained with 2% uranyl acetate, washed with Milli-Q water, and investigated at 60 kV in a JEOL JEM-1010 transmission electron microscope (TEM) (Tokyo, Japan).

Cell extract, PAGE, and immunoblotting.

A sample of 100 ml was harvested from the enrichment culture and centrifuged for 15 min at 4°C and 10,000 × g in a Sorvall centrifuge (Sorvall Lynx 4000). The pellet (about 3 ml) was resuspended in 10 ml of 40 mM potassium phosphate buffer at pH 7. To disrupt the microbial aggregates, the sample was mildly pottered on ice. Cells were passed through a French press operated at 138 MPa in three passages. A protease inhibitor cocktail (Pierce protease inhibitor, mini tablets, EDTS-free; Thermo Scientific) was added, and the sample was incubated at 4°C for 5 min. To remove intact cells and debris, the crude extract was centrifuged in a Sorvall centrifuge for 15 min at 10,000 × g and 4°C. The resulting supernatant was the cell extract.

The cell extract was boiled for 7 min in SDS sample buffer (158 mM Tris-HCl buffer [pH 7] containing 5% β-mercaptoethanol, 2.6% SDS, and 16% glycerol), and 24 μg of protein (determined using 2-D Quant kit; GE Healthcare) per lane was loaded onto 4 to 15% Criterion TGX precast gels (Bio-Rad) for polyacrylamide gel electrophoresis (PAGE), according to the manufacturer's instructions. After PAGE separation, the proteins were transferred from the gel onto a Trans-blot Turbo, midi format 0.2-μm-pore-size nitrocellulose transfer membrane (Bio-Rad) with the Turbo blotter system (Bio-Rad), according to the manufacturer's instructions. The blotting was performed at 2.5 Å and 25 V for 7 min. Dried blots were stored at 4°C.

Prior to starting the immunoblot protocol, the blots were kept at room temperature (RT) for 30 min and then incubated in Milli-Q water for an additional 30 min. Blocking was performed for 1 h in 5% skimmed milk powder (Frema Reform instant skimmed milk powder) in 10 mM TBS (10 mM Tris-HCl, 137 mM NaCl, 2.7 mM KCl [pH 7.4]). The blots were then incubated for 60 min in anti-methyl-coenzyme M reductase (Mcr) antibody (obtained from Milucka et al. [30]) diluted 500-fold in blocking buffer. The negative control was only incubated in blocking buffer. The blots were then washed three times for 10 min in TBS containing 0.05% Tween and incubated for 60 min in monoclonal mouse anti-rabbit IgG alkaline phosphatase conjugate (Sigma-Aldrich) diluted 150,000-fold in blocking buffer. The blots were washed two times for 10 min in TBS containing 0.05% Tween and two times for 10 min in TBS. Finally, blots were incubated with a 5-bromo-4-chloro-3-indolylphosphate–nitroblue tetrazolium (BCIP-NBT) liquid substrate system (Sigma-Aldrich) for 5 min and rinsed for 10 min in Milli-Q water.

Immunogold localization of Mcr in “Ca. Methanoperedens” sp. archaea.

Ultrathin sections (55 nm) of the Lowicryl-embedded samples from the enrichment culture were collected on 50-mesh copper grids with a carbon-coated Formvar support film (FC200Cu; Agar Scientific). Grids containing sections were rinsed for 10 min in 0.1 M PHEM buffer pH 6.9 [60 mM piperazine-N,N′-bis(2-ethanesulfonic acid) (PIPES), 25 mM HEPES, 10 mM EGTA, 2 mM MgCl₂] and blocked for 20 min in 0.5% BSA-c (Aurion) in 0.1 M PHEM buffer (pH 6.9). Grids were incubated for 60 min at room temperature, with the primary antibody targeting Mcr (30) diluted 1:300 in blocking buffer. Negative controls were incubated for 60 min in blocking buffer without primary antibody. After this incubation, the grids were washed for 10 min in 0.1% BSA-c in 0.1 M PHEM buffer (pH 6.9) and incubated for 30 min with secondary antibody, with protein A coupled to 10-nm gold particles (PAG-10; Cell Microscopy Core [CMC], University Medical Center [UMC] Utrecht), diluted 70-fold in blocking buffer. The grids were then washed again first in 0.1% BSA-c in 0.1 M PHEM buffer (pH 6.9) for 5 min and then in 0.1 M PHEM buffer (pH 6.9) for 5 min. To fix the labeling, the grids were incubated for 5 min in 1% glutaraldehyde in PHEM buffer (pH 6.9) and subsequently washed for 10 min in Milli-Q water. Poststaining was performed by incubation for 1 min on drops of Reynolds lead citrate, after which the grids were quickly washed with three drops of Milli-Q water and allowed to air dry. The sections were investigated at 60 kV in a JEOL JEM-1010 TEM (Tokyo, Japan).

Cytochrome staining.

The cytochrome peroxidase reaction was modified from a method described earlier (56). Cells from the enrichment culture were fixed for 30 min in 100 mM cacodylate buffer (pH 7.2) containing 1.5% glutaraldehyde and 4% formaldehyde and washed with 150 mM sucrose in 100 mM cacodylate buffer (pH 7.2). Cells were incubated for 25 min with 2.5 mM 3,3′ diaminobenzidine (DAB) and 0.02% H₂O₂ in 100 mM cacodylate buffer (pH 6.5) and washed with cold 100 mM cacodylate buffer (pH 7.2) containing 150 mM sucrose. Negative controls were preincubated for 10 min with 100 mM potassium cyanide in 100 mM cacodylate buffer (pH 6.5) and then incubated with DAB and H₂O₂ in the presence of potassium cyanide. Cells were postfixed for 90 min in 2% osmium tetroxide in Milli-Q water at 4°C, washed with Milli-Q water, and embedded in 2% low-melting-point agarose. Cells were dehydrated in a graded ethanol series (25, 50, 70, 80, 90, and 96% ethanol) and then in 100% propylene oxide and subsequently embedded in Epon resin. For embedding, cells were infiltrated overnight in 1:1 propylene oxide-Epon, for 8 h in 1:2 propylene oxide-Epon, and then overnight in pure Epon. Fresh Epon was used for final polymerization for 48 h at 70°C. Ultrathin sections were investigated unstained using a JEOL (Tokyo, Japan) JEM-1010 TEM operated at 60 kV.

Electron tomography on “Ca. Methanoperedens” sp. cells.

Semithin sections (250 μm) of the Lowicryl-embedded samples from the “Ca. Methylomirabilis”-“Ca. Methanoperedens” enrichment culture were collected on 50-mesh copper grids with a carbon-coated Formvar support film (FC200Cu; Van Loenen Instruments). Sections were poststained with 4% uranyl acetate in Milli-Q water for 30 min and Reynolds lead citrate for 2 min. Protein A coupled to 10-nm gold (PAG-10; CMC, UMC Utrecht) was applied to both sides of the sections to act as a fiducial marker during tilt-series acquisition and reconstruction. The regions of interest were selected to be aggregates of “Ca. Methanoperedens”-like cells. Dual-axis tilt series (from −60 to +60° tilt angle) were recorded on a JEOL JEM-2100 microscope operating at 200 kV, using SerialEM for automated image acquisition (57). The tilt series were reconstructed using the IMOD package (58), and tomograms were generated using both the weighted backprojection and simultaneous iterative reconstruction technique (SIRT) algorithms.

Accession number(s).

Sequences were submitted to the European Nucleotide Archive (ENA) with accession numbers ERS1810550 and ERS1810551.