Variations in Archaeal and Bacterial Diversity Associated with the Sulfate-Methane Transition Zone in Continental Margin Sediments (Santa Barbara Basin, California)

Site description and sampling.In June 2005, an expedition on the R/V New Horizon collected a series of sediment cores within the SBB (34°13.61′N, 119°59.42′W). At this location, the water depth was 587 m, and the bottom water oxygen content at the time of sampling was determined to be 0.2 μM (4). Sediment was collected by either gravity coring or multicoring for pore water, isotope, mineralogical, and microbiological analyses. One gravity core (GC2) with geochemical properties representative of the majority of cores analyzed was selected for more detailed molecular analysis of the microbial community. Gravity core GC2 (length, 186 cm) was fully processed within 3 h of arrival on deck. Subsamples (∼5 g) were taken from the center portion of the core using a cutoff 5-ml sterile syringe and immediately flash frozen in liquid nitrogen for nucleic acid extraction. Five sediment horizons were targeted in the study; the samples included one sample from the nominal sulfate reduction zone (50 cm), one sample from immediately above the SMTZ (103 cm), two samples from within the SMTZ (125 cm and 139 cm), and one sample from below the SMTZ (163 cm).

Pore water chemistry (sulfate and dissolved inorganic carbon).The pore water sulfate and dissolved inorganic carbon contents were determined as previously described (4). Multicores with an intact sediment-water interface were processed, and these cores served to link gravity cores to the sediment-water interface. Total CO₂ (TCO₂) ([H₂CO₃] + [HCO₃⁻] + [CO₃²⁻]) was analyzed with a Coulemetrics coulometer as previously described (4), and pore water sulfate was analyzed using the turbidimetric method (59).

Nucleic acid extraction and purification.Total DNA was extracted from sediment samples (0.5 g) using a Powersoil DNA extraction kit (Mo Bio Laboratories, Carlsbad, CA). The Mo Bio protocol was modified by initially heating the sample at 65°C twice for 5 min, followed by bead beating using a Fastprep machine (Bio101-Thermo Electron Corp., Gormley, Ontario, Canada) set at a speed of 5.0 for 45 s. The extracts from two independent extractions were combined and cleaned by cesium chloride density gradient centrifugation as previously described (48). The washed and purified DNA was recovered in Tris-EDTA buffer using a Microcon membrane device (YM-100 or Amicon Ultra-4; Millipore, Billerica, MA).

Bacterial and archaeal 16S rRNA gene clone libraries from sediment horizons corresponding to depths above (50 cm and 103 cm), within (125 cm and 139 cm), and below (163 cm) the SMTZ were constructed. Parallel with the rRNA gene analysis, metabolic gene clone libraries for the fragment encoding the alpha and beta subunits of dissimilatory sulfite reductase (dsrAB) (65) and for the fragment encoding the alpha subunit of methyl coenzyme M reductase (mcrA) were constructed (Table 1).

Archaeal and bacterial 16S rRNA gene, mcrA, and dsrAB library construction.PCR mixtures (25 μl) were prepared as follows: 1× Hotmaster PCR buffer with 1.5 mM MgCl₂, 0.2 mM of deoxynucleoside triphosphates, 0.1 μg of each primer, and 0.25 μl of Hotmaster Taq polymerase (Eppendorf AG, Hamburg, Germany).

Archaeal 16S rRNA genes were PCR amplified using archaeon-specific primers AR-8F (5′-TCCGGTTGATCCTGCC-3′) and AR-958R (5′-YCCGGCGTTGAMTCCAATT-3′), and bacterial libraries were constructed using bacterium-specific forward primer BAC-27F (5′-AGAGTTTGATCCTGGCTGAG-3′) and universal reverse primer U-1492R (5′-GGTTACCTTGTTACGACTT-3′) (10, 33). PCR amplifications were performed in a Eppendorf Mastercycler using an initial denaturing step of 95°C for 2 min, followed by 30 cycles (archaea) or 27 cycles (bacteria) of 94°C for 1 min, 54°C for 1 min, and 72°C for 1 min and then an elongation step of 72°C for 6 min. PCR mixtures started with 0.1, 1, and 2 μl template were pooled to minimize PCR artifacts associated with differences in template abundance (52).

A 480-bp fragment of mcrA was amplified with primers MCR-F (5′-GGTGGTGTMGGATTCACACAR-3′) and MCR-R (5′-TTCATTGCRTAGTTWGGRTAG-3′) (41) using an initial denaturing step of 95°C for 3 min, followed by 40 (163-cm and 125-cm samples) or 45 cycles (139-cm sample) of 94°C for 1 min, 50°C for 1 min, and 72°C for 1 min and then a final elongation step of 72°C for 6 min. mcrA amplification was attempted for depths of 103 cm and below. The 103-cm horizon did not yield an mcrA amplicon.

A 1.9-kb fragment of dsrAB was amplified using a mixture of primers, DSR1F mix and DSR4R mix (65). To minimize significant nonspecific product formation observed using standard PCR, a touchdown PCR procedure was performed with the annealing temperature decreasing (1°C every two cycles) from 61 to 54°C for the first 15 cycles, which was followed by 22 cycles of 94°C for 1 min, 54°C for 1 min, and 72°C for 3 min and then a final elongation step of 72°C for 6 min. This touchdown procedure significantly reduced the products with multiple sizes observed with standard PCR and generated a robust 1.9-kb PCR product.

Sequence and phylogenetic analysis.PCR products of the correct size were cloned using either the pCR4-TOPO (Invitrogen, Carlsbad, CA) or pGEM-T Easy (Promega, Madison, WI) system according to the manufacturer's instructions. The amplified inserts were further analyzed by using restriction fragment length polymorphism (RFLP) with either HaeIII (for 16S rRNA genes and dsrAB) or RsaI (for mcrA) (New England Biolabs, Ipswich, MA) to identify unique RFLP patterns for sequencing. One or two unique clones from each RFLP pattern were selected. Unique clones from each library were bidirectionally sequenced using a CEQ 8800 capillary sequencer according to the DTCS protocol (Beckman Coulter, Fullerton, CA). For most samples, vector-targeted primers T7 and M13R were used. Inserts more than 1 kb long (i.e., bacterial 16S rRNA gene and dsrAB libraries) were additionally sequenced using 16S rRNA-targeted primer 515F (5′-GTGCCAGCMGCCGCGGTAA-3′) and internal dsrAB-targeted primers 1FI (5′-CAGGAYGARCTBCAYCG-3′) and 1R1 (5′-CCCTGGGTRTGNAYRAT-3′) (2, 12). For phylogenetic analysis, only the dsrA subunit was used, resulting in a bidirectional sequence combining a vector-targeted primer with an internal dsr primer.

Sequence assembly was performed using Sequencher 4.5 software (Gene Codes, Ann Arbor, MI). The closest relatives of the retrieved sequences in the GenBank database were identified using BLASTN (1). Potential chimeric sequences were identified using CHIMERA_CHECK (8) and Bellerophon (22). For the 16S rRNA gene, sequence data were compiled by using the ARB software package (40) and were initially aligned using the ARB Fast Aligner utility. The resulting alignments were manually verified using known secondary-structure regions. For phylogenetic analysis, the alignments were exported to PAUP*4.0 (58). Bacterial 16S rRNA gene phylogenies were constructed using MEGA 4 (60). Tree reconstruction was performed with the distance matrix and maximum parsimony algorithms. For 16S rRNA genes, trees based on 519 and 1,251 unambiguously aligned positions were constructed for archaeal and bacterial 16S rRNA genes, respectively. mcrA and dsrA trees were based on 147 and 164 translated amino acid characters.

Statistical methods.Correspondence analysis (19) was applied to SBB 16S rRNA clone libraries and libraries for similar marine sedimentary environments obtained from the literature (see Table S1 in the supplemental material) in order to discern patterns of variation among major phylogenetic groups for different geochemical environments. Clone libraries were characterized by previously described geochemistry data or by sediment description and were normalized to describe variation between 10 dominant groups (see Table S1 in the supplemental material). Data sets that contained fewer than 30 environmental clones or in which >50% of the phylotypes fell outside these divisions were excluded (see Table S1 in the supplemental material). Correspondence analysis was implemented using DECORANA software (20) and provides a visual representation of the relationship between samples (i.e., each library) and species (unique phylogenetic groups) along ordination axes calculated to maximize correlation between samples and species scores (20). Neighbor-joining phylogenetic trees constructed for representative SBB 16S rRNA sequences and additional sequences from related environments were used in UniFrac analysis (38). UniFrac derives an environmental distance matrix from the portion of a phylogenetic tree that may be uniquely ascribed to a specific environment (clone library). We used both unweighted (38) and weighted (39) UniFrac to describe the similarity between clone libraries derived from common environmental conditions, as assigned above. Weighted UniFrac uses clonal abundance of specific sequences to adjust the environmental distance calculated with the unweighted algorithm based on each environment's unique portion of the phylogenetic tree.

Nucleotide sequence accession numbers.The nucleotide sequences of the rRNA gene clones have been deposited in the GenBank database under accession no. EU181461 to EU181514 and FJ455875 to FJ455963, the nucleotide sequences of the mcrA clones have been deposited in the GenBank database under accession no. FJ456011 to FJ456019, and the dsrA sequences have been deposited in the GenBank database under accession no. FJ455964 to FJ456010.

Geochemical analyses.Pore water profiles acquired from SBB gravity cores indicate that the SMTZ was approximately 140 cm beneath the sediment-water interface. The observed TCO₂ concentration gradient was linear (R² = 0.91) from 30 cm below the sediment-water interface down to about 120 cm to 155 cm, where a change in slope corresponding to the SMTZ was documented (Fig. 1). The sulfate concentration gradient was also linear through this portion of the core, and there was a change in the slope at the SMTZ. Models of these concentration gradients suggest that reactions involving sulfate and CO₂ occur within the SMTZ but cannot be explained by AOM exclusively (4; W. M. Berelson and F. Sansone, unpublished data). Based on linear gradients and as determined by the methodology described by Berelson et al. (4), the diffusive flux of methane to the sulfate-methane interface (0.5 ± 0.05 mmol/m² day⁻¹) is less than the change in the TCO₂ flux occurring at the horizon (0.7 ± 0.1 mmol/m² day⁻¹); thus, methane oxidation is probably not the only source of TCO₂ added to this horizon, and we hypothesize that there may be an additional, localized source of TCO₂ generated by sulfate oxidation of organic carbon specifically at the SMTZ.

• Open in new tab
• Download powerpoint

FIG. 1.

Pore water profiles for SO₄, CH₄, and TCO₂ for the SBB. The trend lines indicate approximate TCO₂ and sulfate profiles combined from multicore data preserving the water-sediment interface and gravity core sampling to greater depths. The change in slope of the TCO₂ gradient at ∼125 cm suggests that CO₂ production occurs at depth. Methane measurements were obtained using gravity core GC5 (•), sulfate measurements were obtained using adjacent multicores MC5 and MC13 (×) and gravity core GC2 (▪), and TCO₂ measurements were obtained using multicore MC13 (+) and gravity core GC2 (▴).

Patterns of species richness in SBB.Microbial diversity was assessed by RFLP screening of 503 bacterial and 395 archaeal clones from the five discrete depth horizons from core GC-2 (see Fig. S1 in the supplemental material). Shannon-Wiener and Simpson's diversity indices indicate that there is decreasing archaeal diversity with increasing depth in SBB sediment, while the recovered bacterial 16S rRNA diversity appears to increase from the uppermost 50 cm to the 125-cm horizon in the upper part of the SMTZ and then declines at depths below 139 cm (Table 1). In all but the 50-cm depth horizon, the bacterial species richness was greater than the archaeal species richness. Changes in diversity indices with depth were not consistent between archaeal and bacterial clone libraries. However, for each domain the 125-cm upper SMTZ horizon appeared to represent a local maximum for both diversity indices.

Microbial diversity overlying the SMTZ.The major groups recovered from the 50-cm and 103-cm horizons above the SMTZ were the Planctomycetes, green nonsulfur bacteria (GNS) (Chloroflexi), and candidate division OP8 (Fig. 2). Plantomycetes was the dominant bacterial phylum above the transition zone (Table 2). The majority of Planctomycetes sequences fall into the uncultured WPS-1 clade (14), which is common in environments with significant concentrations of organic matter, including marine continental margins (Fig. 2C) (14). Candidate division OP8 clones constitute ∼22% of the 50- and 103-cm bacterial libraries, and the percentage then declines with depth. This group is a minor component of marine sediments, but it has been detected in similar methane-impacted environments, such as the Guaymas Basin (61), the Gulf of Mexico (37), and the Peru and Cascadia Margins (23), as well as in iron- and sulfate-reducing zones of a hydrocarbon-contaminated aquifer (13). Gammaproteobacteria and Betaproteobacteria sequences were recovered exclusively from sediments overlying the SMTZ (Table 2 and Fig. 2B). Deltaproteobacteria sequences were surprisingly absent from the 50-cm library and represented only 6% of the bacterial diversity recovered from the 103-cm sample. These low-abundance phylotypes were most closely related to putative sulfate-reducing phylotypes recovered from the underlying SMTZ.

• Open in new tab
• Download powerpoint

• Open in new tab
• Download powerpoint

FIG. 2.

(A) Bacterial 16S rRNA neighbor-joining distance tree for representative SBB sequences (indicated by bold type). The values at the nodes are bootstrap values based on distance for 2,000 replicates/bootstrap values based on parsimony for 800 replicates. Values of <70 are not shown. (B and C) Expanded phylogeny of the bacterial 16S rRNA neighbor-joining distance tree for representative SBB sequences (indicated by bold type) for the Proteobacteria subdivision (B) and Planctomycetes (C).

The archaeal diversity at 50 cm was dominated (58%) by euryarchaeotal marine benthic group D (MBGD) (64), related to the Thermoplasmatales (Fig. 3). The close relatives of the MBGD phylotypes included sequences recovered from other methane-impacted marine sediments (21, 23, 24, 30), but they are not found exclusively in these habitats. Adjacent to and within the SMTZ, at 103 cm and 139 cm, Crenarchaeota marine benthic group B (MBGB) replaced MBGD as the dominant archaeal group, representing 71% and 61% of the clones, respectively.

• Open in new tab
• Download powerpoint

FIG. 3.

Archaeal 16S rRNA neighbor-joining distance tree for representative SBB sequences (indicated by bold type). The values at the nodes are bootstrap values based on distance for 5,000 replicates/bootstrap values based on parsimony for 200 replicates. Values of <70 are not shown. MHVG, marine hydrothermal vent group.

Microbial diversity within the SMTZ.The most abundant bacterial phylotypes from the SMTZ were affiliated with the Eel-1 group within the Deltaproteobacteria, a cluster of putative sulfate-reducing bacteria first described from a methane seep in the Eel River Basin (48) and distantly related to Desulfobacterium anilini (91% similarity). On average, the Eel-1 clade accounts for ∼30% of both the 125-cm and 139-cm clone libraries (59% and 80% of the recovered Deltaproteobacteria phylotypes, respectively) (Table 2 and Fig. 3). The second most abundant clade of Deltaproteobacteria was related to the putative syntrophic Desulfosarcina-Desulfococcus clade (DSS) commonly recovered from seafloor methane seeps (30, 48) (Table 2). In the SBB, DSS phylotypes were present within the SMTZ (∼14% of the total), as well as the sediment above and below this zone. Planctomycetes-affiliated phylotypes related to the WPS-1 group were enriched in one of the SMTZ horizons; they represented 30% of the 139-cm bacterial library but only 8% of the 125-cm bacterial assemblage.

Sequences belonging to candidate division JS1 represent only ∼6% of the bacterial libraries. However, this group appears to be enriched within the SMTZ. Originally identified in Japan Sea sediments, members of candidate division JS1 have also been commonly recovered from methane hydrate-associated marine sediments, such as sediments from the Nankai Trough, Hydrate Ridge, and the Peru Margin (23, 30, 45). Additionally, sequences affiliated with the Actinobacteria were recovered with low abundance, but only from the SMTZ and the immediately overlying horizon (Fig. 2). Inferences based on the environmental distribution and cultured relatives suggest that these organisms are heterotrophic and may be adapted to sulfur- and/or methane-rich environments (7).

In methane-containing horizons within and below the SMTZ, sequences closely related to uncultured ANME-1 archaea were recovered. These phylotypes were not present in the two sulfate-reducing horizons above the SMTZ (50 and 103 cm), which is consistent with their role in AOM. ANME-1-affiliated phylotypes accounted for 14% of the archaeal diversity within the SMTZ (125 and 139 cm), and the percentage increased to 24% in the sulfate-depleted sediments below this zone (163 cm). The ANME-1-affiliated sequences from all three depths formed two clades (95% similarity) falling in the ANME-1a group, which includes sequences from the Eel River Basin and Gulf of Mexico methane seeps (15, 37) (Fig. 3).

Microbial diversity in sediments underlying the SMTZ.Members of the Chloroflexi, also referred to as the GNS, were the dominant bacteria beneath the SMTZ (163 cm) and represented 40% of the recovered bacterial diversity, compared to 6 to 14% in the horizons above the transition zone. GNS-affiliated sequences were quite diverse (75% overall similarity). Sequences from the 50-cm and 103-cm horizons group with the T78 clade, which is closely related to the dominant phylotype reported for Mediterranean sapropels (9). Other sequences that were recovered from all depths but 50 cm and are dominant within and below the SMTZ group broadly with cultured strains of Dehalococcoides. Sequences associated with candidate division OP1 exhibited a distribution similar to that of the GNS group; the greatest abundance was recovered from the 163-cm library (17%), and these sequences were restricted primarily to sediments outside the SMTZ (50 cm, 103 cm, and 163 cm) (Table 2).

Diversity of dissimilatory sulfite reductase and methyl coenzyme M reductase across the SMTZ.In order to further characterize the diversity of sulfate-reducing microorganisms associated with the SMTZ, dissimilatory sulfite reductase (dsr) genes were analyzed for the four lower horizons of the SBB core. Phylogenetic analysis of the most abundant dsr sequences confirmed the presence of sulfate-reducing Deltaproteobacteria (Fig. 4), with several sequences—exclusive to the SMTZ—clustering with Desulfobacterium anilini (group IV [28] and cluster D [35]). The sequences most commonly recovered from the 103-cm and 125-cm dsr libraries fall in a clade composed of uncultured Deltaproteobacteria and cluster with sequences from Black Sea and South China Sea sediments (DSS relatives, group I [28], and cluster B [35]). Additional dsrA diversity recovered from the SBB included deeply branching sequences related to dsrA genes reported from the Black Sea (clusters I, H, G, and F [35]).

• Open in new tab
• Download powerpoint

FIG. 4.

Neighbor-joining distance tree based on 114 informative characters of translated partial dissimilatory sulfite reductase (dsrA) amino acid sequences from the SBB. The values at the nodes are bootstrap values based on distance for 2,000 replicates/bootstrap values based on parsimony for 100 replicates. Values of <50 are not shown. The group names are those used by Kaneko et al. (28). The numbers in the sequence designations correspond to the depths from which the sequences were recovered.

The diversity and distribution of methanogens and methanotrophic archaea were examined using methyl coenzyme M reductase gene (mcrA) analysis. mcrA genes were successfully amplified from all sample depths below 125 cm but were not recovered from sediments above the SMTZ. mcrA phylotypes from the SBB exhibited limited diversity (92% similarity at the amino acid level), and all of them clustered within the previously described mcrA “group a,” which was assigned to the uncultured ANME-1 archaea (15, 26, 37) (see Fig. S2 in the supplemental material).

Statistical comparison of diversity for methane-influenced marine sediments.As characterized by correspondence analysis (19), marine sediments harbor a diverse bacterial assemblage within which communities associated with different geochemical horizons and sediment depths (i.e., advective methane seeps and mud volcanoes, diffusion-controlled SMTZ, organic matter- and clay-rich sediments) overlap considerably (Fig. 5; see Fig. S3 in the supplemental material). Site-specific variations in physicochemical conditions, biogeographic influences, and a lack of standardized criteria for defining SMTZ horizons in the literature complicate the application of microbial diversity and abundance methods to statistically identify the SMTZ, the sulfate-replete horizon above the SMTZ, and the sulfate-depleted, methane-rich sediment below the SMTZ. Differences in the bacterial communities associated with these distinct geochemical horizons in the SBB are apparent; however, a statistically identifiable difference between the SMTZ and a lower methanogenic horizon was not observed, with both communities grouping within the 95% confidence interval defining the SMTZ-associated bacterial diversity (Fig. 5).

• Open in new tab
• Download powerpoint

FIG. 5.

Correspondence analysis (19) of marine sedimentary bacterial communities assigned to specific redox zones (see Table S1 in the supplemental material). SMTZ-related communities overlap with near-seafloor seep assemblages and communities in other organic matter-rich sediments with similar abundances of major phylogenetic groups and exhibit enrichment of Deltaproteobacteria and, to a lesser extent, Planctomycetes. Detrended correspondence analysis (20) and principal component analysis resulted in similar distributions of sediment communities. Other significant axes of variation in the correspondence analysis exhibit similar environmental groups (see Fig. S3 in the supplemental material). Open symbols indicate environments that are >50 cm below the seafloor, and filled symbols indicate near-seafloor habitats (0 to 30 cm below the seafloor). Data for the following redox zones and lithological classifications are shown: ash; clay; SO₄; NO₃; CH₄; seafloor seep, vents, and mud volcano (Methane Seep); and SMTZ. alpha, beta, gamma, delta, and epsilon indicate the corresponding subdivisions of the Proteobacteria. The ellipses indicate 95% confidence intervals. The ellipse for SMTZ environments is shaded.

Despite the coarse resolution afforded by correspondence analysis, trends in bacterial diversity between different geochemical and lithological sedimentary environments are apparent. For example, while there is substantial overlap in the bacterial groups present in near-seafloor methane seeps and vents and diffusion-controlled deeper SMTZ horizons, the frequent enrichment of Epsilonproteobacteria in near-seafloor seeps along with the rare occurrence of this proteobacterial lineage at depth appears to be a distinguishing feature of these two methane-influenced habitats. Candidate division JS1 phylotypes are often associated with hydrate-bearing sediments (66) and may play a role in diffusion-based SMTZ horizons as well. Overall, the SMTZ-associated horizons cluster together as a subgroup in a broader range of marine sedimentary environments classified by high organic matter content (i.e., continental margin sediments) (Fig. 5; see Fig. S3 in the supplemental material).

In our analysis, SMTZ horizons additionally exhibited enrichment of Deltaproteobacteria, consistent with dissimilatory sulfate reduction, and showed some enrichment of Planctomycetes-related phylotypes. Candidate division OP8 and Betaproteobacteria phylotypes also appear to be relatively abundant in some methane-impacted environments, but they were not enriched in the SBB SMTZ. The clustering of bacterial diversity within specific geochemical environments was similar when different axes of variation derived from correspondence analysis were used (see Fig. S3 and Table S2 in the supplemental material), as well as when detrended correspondence analysis and principal component analysis were used.

Greater resolution of the variation in microbial assemblages associated with the transition through the SMTZ was achieved using the UniFrac metric (38). Applying UniFrac to the SBB sequences alone, we observed a close relationship between the SBB SMTZ horizons to the exclusion of sediments above and below, using the unweighted and weighted algorithms with our bacterial 16S rRNA data set (Fig. 6). The relationship between SMTZ horizons persisted within the SBB Deltaproteobacteria (see Fig. S4 in the supplemental material). This observation suggests that the SMTZ horizons are broadly similar in terms of species richness and phylotype identity, predominantly within the Deltaproteobacteria but extending to other bacterial phyla as well. Compared to data for similar environments described previously (18, 37), SMTZ-related bacterial diversities from independent studies group together, again primarily due to close relationships within the Deltaproteobacteria (see Fig. S4 in the supplemental material). Similar clustering was observed for the archaea in the SBB; however, the clustering order was not maintained using the weighted algorithm, largely due to the abundance of MBGD sequences at 125 cm in contrast to the dominance of MBGB at 139 cm and the presence of ANME-1 in the 163-cm horizon (Fig. 6).

• Open in new tab
• Download powerpoint

FIG. 6.

UniFrac cladograms for 16S rRNA archaeal and bacterial community composition across the SMTZ (as assessed by using environmental clone libraries). The Jacknife values at the nodes are based on 100 replicates. Values of <40 are not shown. (A) Unweighted UniFrac. (B) Weighted UniFrac.

Phyla characteristic of the SMTZ in the SBB.Within the SBB, members of the ANME-1a lineage appear to be the dominant methanotrophic archaeal group, based on analysis of 16S rRNA, recovered from both sediment horizons analyzed for the geochemically defined SMTZ. The presence of the ANME-1 group supports previous observations for advective methane seep environments indicating potential adaptation by this group to deeper sediments with reduced sulfate levels (21, 31, 48). However, recent studies suggest that selective pressures determining the distribution of the ANME groups are likely controlled by other ecological or physicochemical factors in addition to sulfate and methane concentrations (32, 37, 49). While data are limited, there appears to be no consistent trend for the ANME diversity recovered from diffusively controlled SMTZs. For example, organic matter-rich sediments of Skagerrak, Denmark, appeared to select for members of the ANME-2 and ANME-3 groups within and below the SMTZ (49). The potential for distinct ANME groups to carry out AOM at disparate localities complicates the use of these groups for broadly characterizing SMTZ microbial diversity by statistical analyses.

Bacterial diversity, largely dominated by Deltaproteobacteria, was significantly correlated for the two SMTZ horizons, suggesting that there is selective colonization of the SMTZ by related bacterial phylotypes (Fig. 6; see Fig. S4 in the supplemental material). Deltaproteobacteria sequences affiliated with sulfate-reducing bacteria were enriched within both SMTZ horizons relative to sediment horizons above and below this zone. Sulfate-reducing microorganisms that may be involved in sulfate-mediated AOM and/or organic carbon remineralization in the SMTZ included phylotypes previously reported for near-seafloor methane seeps, including members of the cosmopolitan DSS group and the Desulfobacterium-affiliated Eel-1 group (48). The facultatively syntrophic DSS clade has been described for diverse marine sedimentary environments (36, 43, 53, 54), and members of this group were recovered from all horizons except the uppermost horizon in the SBB. Likewise, the distribution and enrichment of the Desulfosarcina-related dsrA sequence, closely related to sequences recovered from Black Sea sediments near the SMTZ (35), further support the common occurrence of this SBB lineage in diverse marine sedimentary environments.

A comparison of methane-associated environments where members of the Eel-1 group have been recovered indicates cooccurrence and a possible relationship with members of the methanotrophic ANME-1 group (16, 21, 37, 48). The significant enrichment of this group within the SBB SMTZ relative to surrounding sediment horizons suggests that members of the Eel-1 clade may be linked to AOM and/or involved in sulfate-dependent organic carbon remineralization hypothesized to be enhanced within the transition zone. While the physiology of the uncultured Eel-1 clade has not been described, related Desulfobacterales isolates and Eel-1-containing enrichment cultures are capable of degrading aromatic hydrocarbons (17, 51), suggesting that members of the Eel-1 group may utilize complex sources of carbon associated with the SMTZ. As Eel-1 members have not been cultured, the associated dsrAB sequence is currently not known. However, the distributions of Eel-1 phylotypes and Desulfobacterium-related group IV dsrA suggest a possible link based on phylogenetic inference.

In addition to the Deltaproteobacteria-affiliated dsrA clades in the SBB, there were also a number of deeply branching dsr clades within and surrounding the SMTZ, implying that there may be additional, as-yet-unidentified microorganisms with the capacity for sulfate reduction that were not readily identified or recovered in the parallel 16S rRNA survey. Many of these dsrA phylotypes clustered with sequences recovered from similar environments, in particular sequences from the Nankai Trough (28) and Black Sea (35). Some phylotypes, like the Syntrophobacter-affiliated clade, likely represent additional diversity within the Deltaproteobacteria, while other deeply branching clades were represented only by environmental sequences and may be associated with novel groups of sulfate-reducing bacteria. While horizontal gene transfer may obscure the phylogenetic correlation between 16S rRNA and dsr (29), the substantial diversity of uncultured 16S rRNA candidate divisions detected in the SBB and related methane-impacted sedimentary environments indicates that one or more of these uncultured lineages may be capable of dissimilatory sulfate reduction.

Defining a global community signature for the SMTZ by microbial phylogeny.Statistical evaluation of the microbial diversity recovered from the SBB revealed microbial groups common to the SMTZ whose clonal abundance changes across this geochemically defined horizon, including the Planctomycetes, candidate division JS1, Actinobacteria, Crenarchaeota MBGB, and Thermoplasmatales-related Euryarchaeota (Table 2). These groups contribute to a general SMTZ signature (Fig. 6) but are not necessarily associated with AOM. Correspondence analysis of a broad range of organic matter-rich and methane-impacted marine sedimentary environments in addition to the SBB revealed a possible relationship between Deltaproteobacteria, GNS, and Planctomycetes in SMTZ-related environments (Fig. 5; see Fig. S3 in the supplemental material). In addition to the predicted enrichment of sulfate-reducing Deltaproteobacteria within the SMTZ, the implied relationship between members of the Planctomycetes, GNS, and, in some settings, candidate division JS1 and Betaproteobacteria in the SMTZ is less clear. The Planctomycetes-affiliated WPS-1 clade was abundant in the lower 139-cm SMTZ horizon but was not confined exclusively to the redox interface in the SBB. The apparent affinity of this clade for organic matter-rich, anaerobic environments suggests that the WPS-1 clade may be an important group of heterotrophs within the SMTZ. The relative percentage of candidate division JS1 (previously identified as OP9 related [67]) increased slightly within the SBB SMTZ. The apparent relationship of the organisms in this clade with hydrate-bearing deep subseafloor habitats (23) and their ability to anaerobically metabolize organic carbon (67, 68) suggest that this group may also contribute to enhanced carbon cycling within the SMTZ.

SMTZs frequently exhibit enrichment of one or more of the methanotrophic ANME groups; however, exceptions have been noted, opening the possibility that additional lineages (e.g., MBGB) may be involved in methane oxidation (5, 57). The distribution of archaeal diversity in our study supports the observation of MBGB enrichment in association with methane and active AOM assemblages (23). In contrast, the miscellaneous crenarchaeotal group (MCG) exhibited a negative relationship with the SBB SMTZ. Similar patterns of spatial segregation between the MCG and MBGB have been documented for other methane-influenced marine sediments (23, 57), suggesting that unique adaptive traits and/or selective pressures may influence the distribution of the uncultured MBGB and MCG in these subseafloor habitats.

Geochemical modeling predicts that complex and integrated microbial processes involving anaerobic oxidation of methane and organic carbon mineralization are stimulated at the SMTZ. Through combined molecular analysis of microbial assemblages associated with SMTZ horizons around the globe, we are beginning to develop an understanding of the patterns of diversity associated with this important redox interface. The application of rigorous statistical tests to produce a unified overview of the common groups of microorganisms associated with the SMTZ within continental margin sediments is complicated by poor representation in the public databases, inconsistencies in defining and sampling the transition zone in published studies, and possible biogeographic differences. Nonetheless, characterization of key microbial groups, such as the Planctomycetes, candidate division JS1, and the Deltaproteobacteria commonly inhabiting the SMTZ, in this and other studies is valuable and provides an essential framework for follow-up studies in which select lineages from the SMTZ can be studied in detail using quantitative measures.