Microbial community structure in methane hydrate-bearing sediments of freshwater Lake Baikal


Correspondence: Nikolai V. Ravin, Centre ‘Bioengineering’, Russian Academy of Sciences, Prosp. 60-let Oktyabrya, Bld.7-1, Moscow 117312, Russia. Tel.: +7 499 7833264; fax: +7 499 1357319; e-mail: nravin@biengi.ac.ru; nravin@mail.ru


Gas hydrates in marine sediments have been known for many years but recently hydrates were found in the sediments of Lake Baikal, the largest freshwater basin in the world. Marine gas hydrates are associated with complex microbial communities involved in methanogenesis, methane oxidation, sulfate reduction and other biotransformations. However, the contribution of microorganisms to the formation of gas hydrates remains poorly understood. We examined the microbial communities in the hydrate-bearing sediments and water column of Lake Baikal using pyrosequencing of 16S rRNA genes. Aerobic methanotrophic bacteria dominated the water sample collected at the lake floor in the hydrate-bearing site. The shallow sediments were dominated by Archaea. Methanogens of the orders Methanomicrobiales and Methanosarcinales were abundant, whereas representatives of archaeal lineages known to perform anaerobic oxidation of methane, as well as sulfate-reducing bacteria, were not found. Affiliation of archaea to methanogenic rather than methane-oxidizing lineages was supported by analysis of the sequences of the methyl coenzyme M reductase gene. The deeper sediments located at 85–90 cm depth close to the hydrate were dominated by Bacteria, mostly assigned to Chloroflexi, candidate division JS1 and Caldiserica. Overall, our results are consistent with the biological origin of methane hydrates in Lake Baikal.


The conversion of complex organic matter to methane is an essential process in the global carbon cycle. Organic matter may either be aerobically oxidized to CO2 or enter anoxic environments where it is decomposed by a consortium of anaerobic microbes. In aquatic environments, fermentative and acetogenic microorganisms mainly convert organic matter into acetate, hydrogen and CO2, which are substrates for methanogenesis. The CH4 gas escapes into the upper aerobic zones, where aerobic microbes oxidize it to CO2, completing the carbon cycle. The CH4 gas that escapes oxidation enters the atmosphere, contributing to the greenhouse effect.

Besides escape into the atmosphere in a gaseous state, methane may be sequestered in a solid state, as gas hydrates. Approximately 500–2500 gigatons of total methane carbon are stored as gas hydrate in deep marine sediments (Milkov, 2004) and most of this methane is produced biologically (Kvenvolden, 1995). Recent investigations revealed that diverse microbial communities are associated with methane hydrate-bearing marine sediments that were collected along the Nankai Trough (Reed et al., 2002; Kormas et al., 2003; Newberry et al., 2004), the Sea of Okhotsk (Inagaki et al., 2003), and the Peru and Cascadia Margins in the eastern Pacific Ocean (Bidle et al., 1999; Marchesi et al., 2001; Yoshioka et al., 2010).

Microbial processes may contribute not only to the formation of methane hydrates but also to their decomposition through the anaerobic oxidation of methane (AOM) performed by several archaeal lineages. It is estimated that AOM consumes most of the methane released in marine settings (Valentine & Reeburgh, 2000). The AOM is coupled to sulfate reduction (Nauhaus et al., 2002), although other electron acceptors, for example manganese and iron (Beal et al., 2009), can be used. The AOM linked to nitrate reduction is thought to be an important process in freshwater systems (Raghoebarsing et al., 2006). In marine sediments, the rates of AOM and sulfate reduction are highest near the sulfate–methane transition zone, where supplies of oxidants (sulfate) and reductants (methane) are concentrated (Reeburgh, 1980; Hoehler et al., 1994). Phylogenetic and organic geochemical data have identified a putative syntrophic consortium of anaerobic methanotrophic archaea and sulfate-reducing bacteria that mediate AOM (Hoehler et al., 1994; Boetius et al., 2000; Orcutt et al., 2005).

In freshwater environments to date, methane hydrates have only been found in Lake Baikal (De Batist et al., 2002), the largest freshwater basin in the world. Here, low water temperatures (3–4 °C) and high pressure conditions (maximum depth of 1642 m) promote hydrate formation. To date, six gas hydrate areas in the central and southern parts of Lake Baikal have been identified. The entire water column of Lake Baikal is oxygenated, and the water is characterized by low mineralization. Therefore, microbial processes related to the generation and decomposition of gas hydrates in Lake Baikal should be significantly different from those in marine environments.

The aim of the present work was to study the archaeal and bacterial communities associated with hydrate-containing sediments of Lake Baikal. We investigated the composition and metabolic potentials of the microbial communities of water and sediment samples using deep pyrosequencing of 16S rRNA genes and functional gene surveys.

Materials and methods

Study site and sampling

Water and sediment samples were collected in Central Baikal (St20 GC9; 52°52.8′N and 107°09.4′E) in August 2009 at the ‘Saint Petersburg’ hydrate site (De Batist et al., 2002). The depth of the water column in this area is 1370–1400 m. Core sampling was conducted onboard the R/V Vereshchagin using a 5-m gravity corer equipped with a plastic liner. The length of the core was 95 cm, and methane hydrates were visually identified in the core at 85–89 cm (Supporting Information, Fig. S1). Water samples were obtained with the help of a deepwater submersible ‘Mir’. A large metallic syringe was used for sampling directly above the gas hydrate field. The core was taken immediately to the laboratory, where it underwent various analyses.

Porewater analysis

Porewater analysis was generally performed every 2–3 cm, or every 5–10 cm in cores with an undisturbed structure using core sediment samples obtained from onboard the ship immediately after core collection. Within 3 h of core retrieval, the porewater samples were extracted from the sediments by centrifugation and stored at 4 °C. Measurements of porewater anion and cation concentrations were carried out according to previously described techniques (Zemskaya et al., 2010).

DNA extraction, and 16S rRNA and mcrA gene libraries

Aseptic samples for microbiological analysis were taken from the hydrate-bearing core (St20 GC9), placed in sterile sachets, and then stored in liquid nitrogen. Three samples were used to assess microbial diversity. The first one (B1) was a water sample collected from the lake floor, the second (B2) was from the upper layer of sediment (0–1 cm depth), and the third (B3) was collected from deeper sediments (85–90 cm depth), just at the methane hydrate layer. The water sample was immediately filtered through a white polycarbonate filter (pore size 0.22 μm; Millipore, Germany). Bacteria were washed off of the filters twice with TE buffer. Total DNA extraction was carried out according to Shubenkova et al. (2005).

The ‘universal’ primers were used for amplification of the V3 variable region of the 16S rRNA gene: U341F (5′-CCT ACG GGR SGC AGC AG) and U515R (5′-TTA CCG CGG CKG CTG VCA C). The PCR fragments were pyrosequenced on GS FLX (Roche); most of the reads covered the full length of the PCR fragment.

For construction of the mcrA gene library, the DNA fragments encoding the mcrA (methyl coenzyme M reductase subunit A) gene were selectively amplified using the primers MCRf (5′-TAY GAY CAR ATH TGG YT) and MCRr (5′-ACR TTC ATN GCR TAR TT) (Springer et al., 1995). The community DNA isolated from sample B2 was used as a template. The PCR fragments were cloned using the pGEM-T vector system (Promega), according to the manufacturer's specifications. The clone library was constructed by transforming Escherichia coli DH10B cells. Plasmid DNA from 20 clones were used as templates for sequencing on an ABI3730 sequencer (Applied Biosystems) using the primer M13f (5′-GTA AAA CGA CGG CCA G).

Data analysis

The 16S rRNA gene sequences determined by pyrosequencing were subjected to the standard filter for environmental pyrosequencing datasets and selected for tags that displayed perfect matches to the primers and contained no ambiguous nucleotides. The resulting datasets consisted of 29 668 reads for the B1 sample, 26 938 reads for the B2 sample, and 38 372 reads for the B3 sample. The 16S rRNA gene data were analyzed using the rpdclassifier program package (Cole et al., 2009). At first, the sequences were distributed between bacteria and archaea using the online rdpnaivebayesianrrnaclassifier version 2.0 (http://rdp.cme.msu.edu/index.jsp). Subsequently, bacterial and archaeal data sets were separately analyzed using the Pyrosequencing pipeline (http://pyro.cme.msu.edu/) of the RPD Classifier (Cole et al., 2009). Cluster analysis and selection of representative sequences for operational taxonomic units (OTUs) were performed by Complete linkage clustering and Dereplicate request functions of the rdpclassifier. Then, we assigned OTUs to taxonomic groups (i.e. bacterial and archaeal divisions) on the basis of blastn sequence similarity searches against the NCBI database. The taxonomic assignments were refined following the construction of phylogenetic trees consisting of representative sequences of the clusters and a set of 16S rRNA gene sequences of representatives of different archaeal and bacterial lineages. The sequences were aligned using clustalx (Thompson et al., 1997) and the neighbor-joining tree was computed by treecon (Van de Peer & De Wachter, 1994).

The bacterial and archaeal 16S rRNA gene sequences were subjected to additional filters for diversity estimates. First, the AmpliconNoise (Quince et al., 2011) was used to account for homopolymer-derived and PCR errors. Then, we removed all remaining singletons (unique sequences only occurring once), as suggested by Behnke et al. (2011). Subsequently, complete linkage clustering and rarefaction analysis were performed using the rpdclassifier program package (Cole et al., 2009).


Biogeochemical characterization

The sediment samples were collected from the ‘Saint Petersburg’ hydrate site using a gravity corer. Methane hydrates were found about 85–89 cm below the lake floor (Fig. S1). The sediments in the core obtained from the surface consisted of gray homogeneous aleuropelite. The chemical composition of the porewater was hydrocarbonate-calcium with a low level of mineralization (Fig. 1). In particular, the concentration of sulfate was very low at the sediment surface (0.15 mM) and quickly decreased with increasing depth. We failed to measure the methane content in the core because of the high loss of gas during core retrieval. The gas mainly consisted of methane, with a small amount of ethane.

Figure 1.

Chemical composition of porewater.

Relative abundance and diversity of Archaea and Bacteria detected using the 16S rRNA gene libraries

To analyze the microbial communities of hydrate-containing sediments and the near-lake-floor water column, we used an approach based on deep pyrosequencing of variable regions of 16S rRNA genes (Sogin et al., 2006). The 16S rRNA dataset for the water sample contained 29 668 sequences, of which 28 902 were assigned to bacterial rRNA and 336 to archaeal 16S rRNA gene. The remaining reads showed no similarity with the 16S rRNA gene sequences and were excluded from the analysis. Thus, bacteria represented the majority of microorganisms in the water column, and the fraction of archaea was only about 1%. The rarefaction analysis of the bacterial community in the water sample (Fig. S2) revealed that when a sequence dissimilarity of 0.03 or 0.05 was used to define the OTU, the rarefaction curves reached the plateau phase, suggesting that the entire microbial community was well covered. The diversity of the bacterial community (Table S1) was estimated to be about 450 species (cluster distance 0.03).

The microbial community of the surface sediment differed from that of the water column. Of a total of 26 938 pyrosequencing reads, 5576 were assigned to bacteria, 11 811 to archaea, and the remaining sequences were unrelated to 16S rRNA gene. Thus, the archaea dominated the B2 sample, representing about two-thirds of all microorganisms detected. The richness of the bacterial community was lower than in the water sample and was comparable to the richness of the archaeal component (Table S1). The deep sediment dataset contained 38 372 reads, of which 18 177 were assigned to bacterial 16S rRNA gene and 6800 to archaeal 16S rRNA gene. Thus, contrary to the upper sediment layer, bacteria formed more than 70% of the microbial community. The bacterial community of the deep sediment was the most complex, whereas the richness of the archaea was approximately the same as in the upper sediment level (Table S1).

The structure of the bacterial community inhabiting the hydrate-proximal water column

Taxonomic classification of the bacterial community was facilitated by the observation that the majority of representative cluster sequences showed more than a 95% homology to the 16S rRNA genes of cultivated microorganisms. Thus, the bacterial sequences were assigned to particular taxa using the online rdpnaivebayesianrrnaclassifier (Fig. 2; Table S2). Gammaproteobacteria were the major component of the community (50.5% of all bacterial clones) and most of them belonged to the family Methylococcacea (43% of all bacteria), containing aerobic methanotrophs (reviewed in Hanson & Hanson, 1996). Approximately 5.6% of all of the sequences were assigned to the family Chromatiaceae of the same class. The second most abundant group was the Betaproteobacteria (27%), mainly represented by aquatic organisms of the families Comamonadaceae (19.6%), Rhodocyclaceae and Methylophilaceae. About 5.7% of the bacterial sequences were affiliated with Deltaproteobacteria, although most of them were only distantly related to the cultured genera. About 7% of the bacterial 16S rRNA gene sequences were affiliated with Bacteroidetes, represented by the genera Flavobacterium, Paludibacter, and different Sphingobacteriaceae. Firmicutes, primary Clostridia, constituted about 2.1% of the microorganisms; and other phyla accounted for < 1% of the bacteria.

Figure 2.

Phylogenetic community structures based on 16S rRNA gene sequences. Composition of archaeal component of the water community is not shown.

The structures of bacterial and archaeal communities of the surface sediments

The bacterial component of the microbial community of the surface sediments comprised lineages found in both the water sample and sediment-specific groups (Fig. 2; Table S2). The former included Gammaproteobacteria (24.9% of bacterial sequences, mainly methanotrophic Methylobacter spp.), Betaproteobacteria (7.4%), Deltaproteobacteria (3.4%), and Bacteriodetes (8.1%). Surprisingly, the second most abundant bacterial group was cyanobacteria that are closely related to Synechococcus spp. (20.1%) – photosynthetic microorganisms that are typically found in the surface layer of water. On the contrary, cyanobacteria accounted for < 0.5% of the bacterial population in the water sample. Representatives of three other bacterial phyla were found in the sediment samples but not in the water sample – Caldiserica (8.3%), candidate division JS1 (2.7%), and Chloroflexi (2.0%). About 10% of the bacterial sequences were only distantly related to cultured organisms and were not classified, even at the phylum level. Notably, we found no sequences representing known sulfate-reducing bacteria in the sediment samples or in the water sample.

In contrast to the bacterial communities of the water and surface sediments, only a minority of archaeal sequences exhibited more than a 95% homology to 16S rRNA genes of cultivated archaea, and thus could be assigned to a particular genus; others belonged to novel taxonomic divisions. Thus, the taxonomic assignments were based on the construction of phylogenetic trees consisting of representative sequences of the clusters and a set of 16S rRNA gene sequences of representatives of different archaeal lineages (Fig. 2; Table S2). The most abundant groups of archaea in the surface sediments were methanogens. Both representatives of the euryarchaeal orders Methanomicrobiales (17.4%), comprising hydrogenotrophic methanogens, and Methanosarcinales (16.8%), also comprising acetate-consuming methanogens (reviewed in Liu & Whitman, 2008), were identified, whereas other methanogenic orders, including Methanobacteriales, Methanococcales and Methanopyrales, were missing. Group 1 methanogens (Methanomicrobiales) were affiliated to the genus Methanosphaerula (more than 95% sequence identity), whereas group 2 (Methanosarcinales) was represented by a lineage that is only distantly related to Methanosaeta (Fig. 3). Archaea closely related to the latter group (97–98% sequence identity, Table S3) have been found in different river and freshwater lake sediments, as well as in submarine permafrost sediments from the Laptev Sea in the Arctic (Koch et al., 2009). Notably, we found no organisms that clustered with known anaerobic methane oxidizers (ANME-1, 2, 3).

Figure 3.

Phylogenetic tree based on 16S rRNA gene sequences of methanogens found in this work and representatives of known methanogenic and ANME lineages (Table S4). The tree was rooted with the crenarchaeon Aerophyrum pernix and constructed using the neighbor-joining method. The numbers at the nodes represent levels of bootstrap support (100 replicates of the original dataset); only values above 50% are shown. The scale bar represents 10% sequence divergence.

About 6.9% of archaea were affiliated with the crenarchaeal Marine group I, being related to Nitrosopumilus spp. (93% sequence identity). These organisms are widespread throughout marine environments and in the soils, where they play an important ecological role as autotrophic aerobic ammonia oxidizers (Konneke et al., 2005; Leininger et al., 2006).

Other surface sediment archaea represented environmental lineages lacking cultured representatives. About 14% of the archaeal sequences were affiliated with the terrestrial miscellaneous euryarchaeotal group (TMEG). This lineage was originally based on clones from South African gold mines (Takai et al., 2001) but this group now includes clones from the terrestrial subsurface and from soils, marine sediments and freshwater lakes. Clones closely related to this group were found in petroleum-contaminated soils and aquatic sediment samples (Table S3). The crenarchaeal group MCG1 constitutes about 11.4% of archaeal sequences. The MCG1 archaea have a wide habitat range that includes terrestrial and marine, hot and cold, and surface and subsurface environments (Teske, 2006). In particular, closely related clones have been found worldwide in freshwater sediment samples (Table S3).

About one-third of the archaeal sequences were classified into eight separate crenarchaeal groups (named Baikal-1 to 8), forming distinct lineages on the archaeal phylogenetic tree. The largest group, Baikal-1, constituting 19.2% of all archaeal clones, is widespread, although only in freshwater sediments (Table S3). Other Baikal groups are more distantly related to the environmental clones, of which most were also identified in the sediment samples. Two lineages, Baikal-5 and Baikal-6, constituting a total of about 2% of the archaeal sequences, lacked any close relatives.

The structures of bacterial and archaeal communities of the deep sediments

Bacteria accounted for about 70% of the microorganisms inhabiting the deep, hydrate-proximal sediments (85–90 cm depth). Fewer than 5% of the bacterial and archaeal sequences exhibited more than a 95% homology to the 16S rRNA genes of the cultivated microorganisms, highlighting the unexplored nature of subsurface microbiomes.

The bacterial community of the deep sediments was significantly different from those of the water sample and the surface sediments (Fig. 2; Table S2) – the major groups were Chloroflexi (38% of bacterial sequences), JS1 and Caldiserica, while Gammaproteobacteria and Cyanobacteria were present in minor amounts. More than 10% of the bacterial sequences were only distantly related to cultured organisms and were not classified, even at the phylum level.

Most of the bacteria assigned to Chloroflexi formed lineages affiliated with Dehalococcoidetes (subphylum II of the Chloroflexi) but they were only distantly related to canonical Dehalococcoides spp. (91–96% sequence identity). Dehalococcoides are anaerobic bacteria dedicated to the transformation of various chlorinated organic compounds via reductive dechlorination (Maymo-Gatell et al., 1997). These organisms were detected in a number of anoxic freshwater and marine sediment samples (Kittelmann & Friedrich, 2008ab), and the clones closely related to Chloroflexi lineages identified in this work were found in similar environments (Table S3).

About 19% of the bacterial sequences were assigned to the candidate division JS1. Environmental clones representing this lineage have been found consistently in deep marine and subsurface sediments (Teske, 2006). However, they have also been found in a number of other anoxic sedimentary environments, such as methane hydrate-bearing sediments (Reed et al., 2002; Inagaki et al., 2006), and marine and terrestrial mud volcanoes (Alain et al., 2006). In particular, members of the JS1 division were key representatives at marine methane hydrate sites, where they may constitute up to 50% of the clone libraries (Inagaki et al., 2006). The closest relatives of Baikal JS1 bacteria were detected in different freshwater sediments (Table S3).

The phylum Caldiserica (formerly OP5) constituted about 8% of the bacterial community. This phylum was originally represented by the environmental clone sequences retrieved from hydrothermal sources (Hugenholtz et al., 1998), but later its representatives were found in different anoxic environments, including Arctic tundra soil (Liebner et al., 2008).

The archaeal community (Fig. 2; Table S2) was dominated by environmental uncultured lineages, the only exception being methanogens. The same two groups of methanogens were identified but their total share was lower than in the surface sediment community (10.3% vs. 34.2% of all archaea). In contrast to the surface sediments, Methanomicrobiales were the dominant methanogenic lineage (8.4%). Euryarchaeota were also represented by the uncultured groups TMEG (8.8%) and SAGMEG (3.4%). The latter group was originally discovered in the deep terrestrial subsurface in the South African goldmines (Takai et al., 2001) but later these archaea were found in deep marine sediments containing methane hydrates (Reed et al., 2002), in marine sediments in the Sea of Okhotsk (Inagaki et al., 2003), and in the Peru Margin (Inagaki et al., 2006). Baikal SAGMEG archaea are closely related to environmental clones isolated in terrestrial sediments (Table S3).

Crenarchaea of several MCG1-related lineages accounted for 24.2% of the archaeal community. They are closely related to the MCG1 archaea found in the surface sediment sample (B2). Of the eight crenarchaeal lineages found in the surface sediments, three (Baikal 1, 2 and 3) were also identified in the deep sediments. Together, they accounted for 41.5% of all archaea, the most abundant being the Baikal-1 lineage (33.2%).

Phylogeny of the mcrA genes

To support the phylogenetic affiliation of the 16S rRNA gene sequences to methanogenic rather than ANME groups, clone libraries targeting genes coding for a conserved region of the alpha subunit of the methyl coenzyme M reductase gene (mcrA) were constructed for the surface sediment sample. Indeed, the mcrA genes can substitute for 16S rRNA genes in the reconstruction of phylogenetic relationships among methanogens (Luton et al., 2002). The clones screened apparently fell into two groups, consisting of 18 and two sequences (Fig. 4). These mcrA lineages corresponded to two Baikal methanogen groups identified on the 16S rRNA gene tree (Fig. 3). The smaller group was clearly affiliated with McrA from Methanomicrobiales. The larger one formed a distinct lineage related to McrA from Methanosarcinales methanogens isolated from rice field soil (Lueders et al., 2001; Conrad et al., 2008) and Zoige wetlands of the Tibetan plateau (Zhang et al., 2008). In particular, ZC-1 methanogens, which account for about 30% of the archaea at the Zoige wetlands, are a cold-adapted Methanosarcinales lineage, utilizing diverse methanogenic substrates including acetate and hydrogen (Zhang et al., 2008).

Figure 4.

Phylogenetic tree based on the deduced amino acid sequences of mcrA genes of methanogens found in this work and representatives of known methanogenic and ANME lineages (Table S4). The tree was rooted with Methanopyrus kandleri and constructed using the neighbor-joining method. The numbers at the nodes represent levels of bootstrap support (100 replicates of the original dataset); only values above 50% are shown. The scale bar represents 10% sequence divergence.


The results of this study provide an insight into the identities of the dominant members of the microbial communities in freshwater methane hydrate-bearing sediments of Lake Baikal. Methanogenesis and aerobic methane oxidation appeared to be among the primary biogeochemical processes reflecting the compositions of microbial communities. In the water sample, about half of the 16S rRNA gene sequences were related to aerobic methanotrophic Gammaproteobacteria of the genus Methylobacter. Another subgroup of methylophilic bacteria, Betaproteobacteria of the genus Methylophilus, was present in minor amounts. Such a structure of the bacterial population is consistent with the processes of aerobic oxidation of methane escaping from the sediments into the water. Archaea accounted for only about 1% of the 16S rRNA gene sequences and were mostly represented by methanogens (Table S2) that could have originated from the sediments.

The structure of the microbial community of the surface sediments reflects a mixture of different metabolic processes at the oxygenated water/anoxic sediment interface. The bacterial community comprised both methanotrophs and aquatic Bacteroidetes bacteria that could have originated from the oxygenated water circulating between the water column in the lake and the pores in the upper levels of the sediment, and also from lineages frequently found in sediments (Chloroflexi, JS1, Caldiserica). It was quite unexpected to find a high fraction of Cyanobacteria in this sample. Microscopic observations revealed cyanobacterial cells exhibiting characteristic autofluorescence (Fig. S3). These observations suggest that the cyanobacteria may grow heterotrophically at this depth, although they may also have originated from the fecal matter of zooplankton.

Archaea accounted for about two-thirds of the microorganisms detected in the upper sediments. Euryarchaea of the orders Methanomicrobiales and Methanosarcinales constituted about one-third of the archaeal phylotypes. Both 16S rRNA and mcrA phylogenies support an affiliation of these microorganisms with methanogenic lineages, but not with known anaerobic methane oxidizers. These data are consistent with the freshwater nature of the environment, characterized by low sulfate concentrations resulting in the absence of both sulfate reduction, which otherwise could out-compete methanogenesis, and anaerobic methane oxidation, known to be coupled to sulfate reduction. Thus, unlike marine and other sulfate-rich environments, methanogenesis in Lake Baikal could occur even in the upper levels of the sediments. An ecologically important group of archaea that was exclusively found in the upper sediments is the Marine group I constituted about 7% of archaeal sequences. Cultivated organisms of this lineage are aerobic autotrophs that oxidize ammonia to nitrite. The rest of the archaeal community represented different creanarchaeal lineages lacking cultured representatives. Their metabolic capabilities are mostly unknown, with the exception of the MCG1 group, which is suggested (Teske & Sorensen, 2008) to comprise heterotrophic anaerobes that utilize and assimilate complex organic substrates.

The composition of the 16S rRNA clone library indicates that the microbial community of the deep, hydrate-proximal sediments is dominated by bacteria, although the fraction of archaea is significant. Chloroflexi, JS1 and Caldiserica were the major components of this community. The marine methane hydrate-bearing sediments were dominated by JS1, containing more than half of the sequences, whereas the phylum Chloroflexi was a dominant group in organic-rich sites lacking hydrates (Inagaki et al., 2006). Chloroflexi bacteria were distantly related to the Dehalococcoidetes group, performing reductive dechlorination. The presence of halogenated compounds is typically associated with industrial contamination, although natural sources are also known, for example marine algae (Häggblom et al., 2003). If the Chloroflexi found in Lake Baikal sediments were also dehalo-respiring bacteria, then they would depend on the H2 supply from fermentative microorganisms for their growth. As the Gibbs free energy of reductive dechlorination (from −130 to −180 kJ mol−1 per unit of chlorine removed) is higher than that generated by methanogenesis, the reductively dechlorinating bacteria could out-compete methanogens (Dolfing, 2003; Tas et al., 2010). Such competition may account for the lower fraction of methanogenic archaea in the deep sediment community relative to the surface community (10.3 vs. 34.2% of the archaeal sequences).

The microbial community of Lake Baikal hydrate-bearing sediments differs significantly from the hydrate-proximal communities in marine environments. Analyzing the sediment cores from the Peru and Cascadia Margins, Inagaki et al. (2006) found that prokaryotic communities in marine sediments are mainly composed of Bacteria, and only at one hydrate-containing site were Archaea found to make up 30% of the total microbial community near the seafloor. Similar to Lake Baikal, bacterial communities in the marine methane hydrate-bearing sediments were dominated by JS1 and Chloroflexi (Inagaki et al., 2006). The archaeal components of the communities were drastically different. In marine sediments the archaeal fraction decreased with depth, constituting < 0.01% of the total prokaryotic signature in all sites where the hydrate stability zone was reached (Inagaki et al., 2006). Archaeal communities in the hydrate-bearing sites primarily comprised the deep-sea archaeal group (DSAG; also known as Marine Benthic Group B). The DSAG archaea have been reported to be the dominant archaeal lineages in the hydrate zones in the Nankai Trough (Reed et al., 2002) and the Sea of Okhotsk (Inagaki et al., 2003). The DSAG archaea were generally abundant in the sulfate reduction zone above hydrates, suggesting a role in the processes of sulfate reduction and methane oxidation (Inagaki et al., 2006). Notably, Inagaki et al. (2006) only detected a few sequences of 16S rRNA and mcrA genes of known methanogens in the hydrate-bearing sediments, leaving open the question of the microbial source of methane.

Overall, our results are consistent with the microbiological origin of methane hydrates in Lake Baikal. The isotopic composition of the gas from the ‘Saint Petersburg’ hydrate field (δ13Cav. = −63.6‰, δD (CH4) = −282‰) also supports its microbiological origin (Egorov et al., 2010). Organic matter produced in the water column reached the lake floor and became buried in the sediments. These organics were processed by a consortium of different bacterial and archaeal heterotrophs and, in the absence of sulfate and other alternative electron acceptors, finally converted to methane. The methane generated was either sequestered in gas hydrates or discharged to the water column, where it was consumed by aerobic methanotrophic bacteria. As could be expected from the lack of sulfate reducers, we did not find ANME archaea in the sediments. Thus, Lake Baikal appears to be a simple model for studying the generation of natural gas hydrates.


This work was supported by the Ministry of Education and Science of Russia (contract 16.552.11.7035) and by the Program ‘Molecular and Cellular Biology’ of RAS.

Authors’ contribution

V.V.K. and A.V.M. equally contributed to this work.