Prokaryotic community structure and diversity in the sediments of an active submarine mud volcano (Kazan mud volcano, East Mediterranean Sea)


  • Maria G. Pachiadaki,

    1. Environmental Chemical Processes Laboratory, Department of Chemistry, University of Crete, Voutes-Heraklion, Greece
    2. Hellenic Centre for Marine Research, Institute of Oceanography, Anavissos, Greece
    Search for more papers by this author
  • Vasilios Lykousis,

    1. Hellenic Centre for Marine Research, Institute of Oceanography, Anavissos, Greece
    Search for more papers by this author
  • Euripides G. Stefanou,

    1. Environmental Chemical Processes Laboratory, Department of Chemistry, University of Crete, Voutes-Heraklion, Greece
    Search for more papers by this author
  • Konstantinos A. Kormas

    1. Department of Ichthyology and Aquatic Environment, School of Agricultural Sciences, University of Thessaly, Nea Ionia, Greece
    Search for more papers by this author

  • Editor: Riks Laanbroek

Correspondence: Konstantinos A. Kormas, Department of Ichthyology and Aquatic Environment, School of Agricultural Sciences, University of Thessaly, 384 46 Nea Ionia, Greece. Tel.: +30 242 109 3082; fax: +30 242 109 3157; e-mail:


We investigated 16S rRNA gene diversity at a high sediment depth resolution (every 5 cm, top 30 cm) in an active site of the Kazan mud volcano, East Mediterranean Sea. A total of 242 archaeal and 374 bacterial clones were analysed, which were attributed to 38 and 205 unique phylotypes, respectively (≥98% similarity). Most of the archaeal phylotypes were related to ANME-1, -2 and -3 members originating from habitats where anaerobic oxidation of methane (AOM) occurs, although they occurred in sediment layers with no apparent AOM (below the sulphate depletion depth). Proteobacteria were the most abundant and diverse bacterial group, with the Gammaproteobacteria dominating in most sediment layers and these were related to phylotypes involved in methane cycling. The Deltaproteobacteria included several of the sulphate-reducers related to AOM. The rest of the bacterial phylotypes belonged to 15 known phyla and three unaffiliated groups, with representatives from similar habitats. Diversity index H was in the range 0.56–1.73 and 1.47–3.82 for Archaea and Bacteria, respectively, revealing different depth patterns for the two groups. At 15 and 20 cm below the sea floor, the prokaryotic communities were highly similar, hosting AOM-specific Archaea and Bacteria. Our study revealed different dominant phyla in proximate sediment layers.


Mud volcanoes (MVs) result from the extrusion of fluid-rich mud flows and have long been known onshore and at sea, mainly within active tectonic belts, but also on passive margins with high sedimentation rates. Submarine MVs have become environments of intensive study in recent years. They play an intrinsic role in many processes such as the shaping of ocean margins, release of greenhouse gases and thus climate change, petroleum genesis and accumulation of gas hydrates, a potential energy resource (Milkov, 2000). MVs may erupt violently at regular or irregular time intervals or may emit mud, fluid and gases continuously. In the eastern Mediterranean Sea, MVs and seep areas were discovered during the late 1970s (Cita et al., 1981) and continuing research has revealed several such formations on the accretionary prism of the Hellenic Arc (Mediterranean Ridge) and within the Anaximander Mountains (Woodside et al., 1998; Lykousis et al., 2004). Kazan MV is an isolated, oval-shaped 0.6 × 0.9 km dome aligned in a N–S direction, with a height of 50 m, lying on the edge of a relatively flat plateau of 1750 m average depth (Lykousis et al., 2009); it has been regarded as rather inactive.

High seafloor methane fluxes are associated with MVs as well as with the accompanying cold vents and seeps (Charlou et al., 2003). The available gas provides energy for rich benthic communities, which include chemosynthetic symbiotic fauna (Olu-Le Roy et al., 2004). Moreover, highly diverse and productive microbial ecosystems based on chemosynthesis can be sustained by the carbon-rich fluids in MVs. Probably the most important biogeochemical process fuelling these communities at these locations is the anaerobic oxidation of methane (AOM) (Boetius & Suess, 2004). In marine habitats, the metabolic process of AOM is suggested to be a reversed methanogenesis coupled to the reduction of sulphate-involving methanotrophic archaea (ANME archaea) and sulphate-reducing bacteria (SRB) (Hallam et al., 2004).

Studies have provided evidence for the occurrence of AOM in Kazan carbonate crusts, based on carbon isotopic signatures, porewater chemistry and general molecular fingerprints (Aloisi et al., 2000; Werne et al., 2002; Bouloubassi et al., 2006; Heijs et al., 2006), as well as in specific sediment layers of the Kazan MV (Werne et al., 2004; Heijs et al., 2007; Kormas et al., 2008).

There is a scarcity of available data on the fine-scale distribution of prokaryotes in marine surface sediments. However, the co-occurrence patterns of microorganisms are the first step in defining potentially interacting organisms and interaction networks in highly complex systems and also help to define ecological niches for better characterization of ecological species or microbial ecotypes. In addition, the relationship between co-occurring communities with environmental factors helps to delineate species that are specific to such sediments (Fuhrman, 2009).

The structure of highly complex prokaryotic communities in marine sediments is dictated by rapid changes of physical and geochemical factors with sediment depth in hot-spot sites such as MVs. However, it remains unclear whether such factors shape the communities of Bacteria and Archaea in a similar manner. Here, we investigated whether: (1) the community structures of Bacteria and Archaea vary concomitantly; (2) the prokaryotic assemblages are dominated by one or several phylotypes with different relative abundances, implying a specific or an opportunistic community, respectively; and (3) which of these specific communities are involved in biogeochemical processes related to mud volcanism. We analysed the diversity of Archaea and Bacteria every 5 cm in the top 30 cm of sediment in an active site of the Kazan MV, East Mediterranean Sea.

Materials and methods

Sampling site

The data presented in this study were acquired from the Kazan MV in May 2003. Selected seismic reflection profiling and sediment sampling were carried out in order to better evaluate the most active sites bearing gas hydrates. The seafloor bathymetry/backscatter survey was carried out using a SEABEAM 2120 swath system installed aboard the R/V AEGAEO of the Hellenic Centre for Marine Research (see fig. 10 in Lykousis et al., 2009). The backscattered multibeam (backscattering intensity) map of the Kazan MV revealed a clear tail-like structure, prominently directed towards the south, suggesting the presence of a mud flow. The highest backscatter values corresponded to the summit of the mound, which was expected to be the more active site of the MV, while concentric to this, the backscatter variations seemed to correspond to variations in the slope of the volcano. Gas hydrates were recovered for the first time at Kazan MV during the cruise reported in this paper. The gas hydrate crystals appeared as small rice-like lumps and were quite regularly dispersed throughout the sediment matrix, even close to the sediment surface (inset to fig. 11 in Lykousis et al., 2009).

Gas-hydrate-containing sediment samples were collected from a site (35°25′55″N, 30°33′42″E) of the Kazan MV, where mud flow indicated recent activity, as described in Kormas et al. (2008). Prokaryotic diversity at 1, 5, 10, 25 and 30 cm below the sea floor (cm b.s.f.) was analysed. Data from 15 and 20 cm b.s.f. were retrieved from the same sediment core and are taken from Kormas et al. (2008).

DNA extraction, amplification and cloning

DNA was extracted from 1 g of wet sediment from each sample, using the UltraClean Soil DNA kit (MoBio Laboratories Inc.) following the manufacturer's protocol with minor modifications: bead beating was reduced from 10 to 5 min, and this step was immediately followed by three freeze–thaw cycles (−80 °C for 3 min and then immediately in a 65 °C water bath for 5 min) after addition of the inhibitor removal solution. Bacterial 16S rRNA genes were amplified using the bacterial primers B8f–B1492r (Teske et al., 2002). The PCR included an initial denaturation step at 94 °C for 1 min, followed by 27–31 cycles consisting of denaturation at 94 °C for 45 s, annealing at 52.5 °C for 45 s and elongation at 72 °C for 2 min; a final 7-min elongation step at 72 °C was added. The archaeal 16S rRNA gene was amplified using the primer combination A8f and A1492r (Teske et al., 2002). An initial denaturation step at 94 °C for 1 min was followed by 25–29 cycles consisting of denaturation at 94 °C for 45 s, annealing at 52.5 °C for 45 s and elongation at 72 °C for 2 min; a final 7-min elongation step at 72 °C was added. The number of cycles was determined for each sample after cycle optimization. PCRs were repeated with different cycle numbers, and the lowest number of cycles that yielded a positive signal was then used for cloning and sequencing in order to avoid differential representation of 16S rRNA genes with low and high copy numbers (Spiegelman et al., 2005). Eight tubes of PCR products were pooled for cleanup and cloning to reduce the bias of each individual reaction.

PCR products were visualized on a 1% agarose gel under UV light, bands were excised and PCR products were extracted using the Wizard SV Gel and PCR Clean-up kit (Promega Inc.) following the manufacturer's protocol. The PCR products were cloned using the TOPO TA Cloning® Kit for Sequencing (Invitrogen Corp.) using electrocompetent cells according to the manufacturer's specifications. For each sample and each gene, randomly picked clones with inserts of the expected length were analysed. Clones were grown in liquid Luria–Bertani medium with kanamycin and their plasmids were purified using the NucleoSpin Plasmid QuickPure kit (Macherey-Nagel GmbH & Co. KG, Germany) for DNA sequencing.

Sequencing and phylogenetic analysis

Sequence data were obtained by Macrogen Inc. (South Korea) using capillary electrophoresis and a BigDye Terminator kit (Applied Biosystems Inc.) with the primers M13F(-20) and M13R. Every sequence read was approximately 900 bp, and for each individual clone, forward and reverse reads were assembled. The sequences were screened for chimeras by comparing neighbour-joining trees made of the first and second halves of all sequences. The sequences that had different groupings in the first and second halves were then checked using the pintail program (Ashelford et al., 2005; bellerophon software from GreenGene (DeSantis et al., 2006) ( was also used to detect chimeric sequences. All putative chimeras were excluded from further analysis.

For the detection of the closest relatives, all sequences were compared with the blast function (Zhang et al., 2000; The sequences were automatically aligned against sequences from their closest relatives using the rdp alignment utility (Cole et al., 2003; and revised by manual removal of ambiguously aligned regions. Phylotypes or operational taxonomic units were defined as sequences showing ≥98% homology (Fields et al., 2006) to each other.

Phylogenetic trees were constructed by the neighbour-joining method using the Kimura two-parameter correction. Bootstrap analyses for 1000 replicates were performed to assign confidence levels to the tree topology using the mega4 software (Tamura et al., 2007). Sequences of unique phylotypes found in this study have GenBank numbers FJ712361FJ712398 for Archaea and FJ712399FJ712411, FJ712413FJ712428, FJ712432FJ712451, FJ712453FJ712462, FJ712466FJ712472, FJ712478FJ712483, FJ712489FJ712509, FJ712511FJ712514, FJ712516FJ712536, FJ712538FJ712555, FJ712557FJ712560, FJ712562FJ712568 and FJ712570FJ712621 for Bacteria.

Clone library coverage, diversity and similarity analyses

Clone coverage was calculated using the equation C=[1−(ni/N)] × 100, where ni is the number of phylotypes and N is the number of 16S rRNA gene sequences examined (Good, 1953; Kemp & Aller, 2004). The Shannon–Wiener index H was used as a diversity index and was calculated as follows: inline image, where the summation is over all phylotypes i and pi is the proportion of phylotypes relative to the sum of all phylotypes. The Pielou evenness index J was calculated as J=H/lnS, where S is the total number of phylotypes (Shannon & Weaver, 1949; Pielou, 1969).

The similarity among the microbial communities in all sediment layers was determined using the Morisita index of similarity (Wolda, 1981) at both the phylotype and the phylogenetic group level, using the equation




where nij is the individuals of phylotype i in sample j, nik is the number of individuals of phylotype i in sample k (Morisita, 1959) and S is the total number of phylotypes. The Morisita indices of similarity were further analysed by cluster analysis using the past program (Hammer et al., 2001).


The clone names used indicate the sample site and library identification. The names begin with the letters KZNMV (for Kazan MV), followed by a number showing the layer used to construct the library (0, 5, 10, 25, 30 cm b.s.f.). A or B indicates Archaea and Bacteria, respectively. Surface sampling was actually conducted in the 0–1-cm layer and, for simplicity, it is coded as 0 in the phylotype names and annotated as the ‘surface layer’ in the text.


A total of 242 archaeal 16S rRNA gene sequences were retrieved, and these were attributed to 38 phylotypes (see Supporting Information, Table S1). The vast majority (99.2%) of the archaeal sequences belonged to the Euryarchaeota (Fig. 1). Only one Crenarchaeota phylotype (KZNMV-10-A9, 4.1%) was found. This phylotype belonged to the marine benthic group B (MBG-B), whose members have been found in marine sediments and hydrothermal vent samples.

Figure 1.

 Phylogenetic tree of the PCR-amplified archaeal 16S rRNA gene phylotypes (c. 1400 bp) in the sediments of the Kazan MV, East Mediterranean Sea, based on the neighbour-joining method as determined by distance using Kimura's two-parameter correction. Each phylotype recovered (in bold letters) is named after the sediment depth origin. Numbers of identical (≥98% sequence similarity) phylotypes of the total number phylotypes found in this sediment depth are shown in parentheses. One thousand bootstrap analyses (distance) were conducted, and percentages >50% are indicated at nodes. Numbers in brackets are GenBank accession numbers. Scale bar represents 5% estimated distance.

The surface layer was dominated by ANME-2 (KZNMV-0-A18, KZNMV-0-A25, KZNMV-0-A40, 89.1%). KZNMV-0-A18 was the most abundant phylotype in this layer and clustered within the ANME-2c subgroup. At 10 cm b.s.f., ANME-2 phylotypes again dominated (KZNMV-10-A2, KZNMV-10-A1, KZNMV-10-A4, 89.8%). Here, the most abundant phylotype, KZNMV-10-A2, was affiliated to the ANME-2a,b subgroup.

A group of five highly similar (>98%) phylotypes (KZNMV-5-A1, KZNMV-30-A1, KZNMV-25-A6, KZNMV-0-A11, KZNMV-10-A1) belonged to the ANME-3 group. Phylotypes KZNMV-5-A1 and KZNMV-30-A1 dominated at 5 cm b.s.f. (41.8%) and 30 cm b.s.f. (33.3%), respectively. One single phylotype (KZNMV-0-A42) also belonged to the ANME-3 group.

At 25 cm b.s.f., the most abundant phylotype, KZNMV-25-A23 (45.5%), grouped within the Methanomicrobiales and was closely related (bootstrap value 100) to phylotype 4E12 from Guaymas Basin sediments. Members of the ANME-1 were also found. They occurred at 5 cm b.s.f (3.6%), 25 cm b.s.f (22.5%) and 30 cm b.s.f (25.0%).

Phylotypes KZNMV-5-A12 (21.8%), KZNMV-30-A14 (8.3%) and KZNMV-25-A58 (4.6%) converged (bootstrap value 100) with a monophyletic group of Methanosaeta spp. Six phylotypes with low abundance, from all except surface layers, belonged to the MBG-D and one from 5 cm b.s.f. was affiliated to GoM-Arc-I (Table 1, Fig. 1a). Finally, a singleton phylotype from 25 cm b.s.f. was related to the ‘Guaymas Euryarchaeotal Group’ (GEG).

Table 1.   Phylogenetic grouping and relative abundance (%) of the phylotypes found at the Kazan MV, East Mediterranean Sea
 0 cm b.s.f.5 cm b.s.f.10 cm b.s.f.15 cm b.s.f.*20 cm b.s.f.*25 cm b.s.f.30 cm b.s.f.
 GoM Arc I0.
 Methanosaeta-related Methanosarcinales0.
 Unaffiliated Methanosarcinales0.


A total of 374 bacterial 16S rRNA gene sequences were retrieved, and these represented 205 unique phylotypes (Table S1). Phylotypes from the five subdivisions of the Proteobacteria accounted for the majority of sequences in each clone library, but not all subdivisions were retrieved in each layer (Figs 2 and 3). The remaining phylotypes were affiliated to 10 known phyla [Acidobacteria, Actinobacteria, Bacteroidetes, Chlorobi, Chloroflexi, Deferribacteres, Firmicutes, Fusobacteria, Planctomycetes, Elusimicrobia or Termite Group 1 (TG1)], five candidate divisions (JS1, OP8, OP11, WS2, WS3) and three unaffiliated groups (Fig. 4).

Figure 2.

 Phylogenetic tree of the PCR-amplified Gammaproteobacteria– except from the outgroup sequence of the Firmicutes Fusibacter paucivorans– 16S rRNA gene phylotypes (c. 1400 bp) in the sediments of the Kazan MV, East Mediterranean Sea, based on the neighbour-joining method as determined by distance using Kimura's two-parameter correction. Each phylotype recovered (in bold letters) is named after the sediment depth origin. Numbers of identical (≥98% sequence similarity) phylotypes of the total number of phylotypes found in this sediment depth are shown in parentheses. One thousand bootstrap analyses (distance) were conducted, and percentages >50% are indicated at nodes. Numbers in brackets are GenBank accession numbers. Scale bar represents 1% estimated distance.

Figure 3.

 Phylogenetic tree of the PCR-amplified Alpha-, Beta-, Delta-, Gamma- and Epsilonproteobacteria 16S rRNA gene phylotypes (c. 1400 bp) in the sediments of the Kazan MV, East Mediterranean Sea, based on the neighbour-joining method as determined by distance using Kimura's two-parameter correction. Each phylotype recovered (in bold letters) is named after the sediment depth origin. Numbers of identical (≥98% sequence similarity) phylotypes of the total number of phylotypes found in this sediment depth are shown in parentheses. One thousand bootstrap analyses (distance) were conducted, and percentages >50% are indicated at nodes. Numbers in brackets are GenBank accession numbers. Scale bar represents 2% estimated distance.

Figure 4.

 Phylogenetic tree of PCR-amplified of bacterial (except all Proteobacteria) 16S rRNA gene phylotypes (c. 1400 bp) in the sediments of the Kazan MV, East Mediterranean Sea, based on the neighbour-joining method as determined by distance using Kimura's two-parameter correction. Each phylotype recovered (in bold letters) is named after the sediment depth origin. Numbers of identical (≥98% sequence similarity) phylotypes of the total number phylotypes found in this sediment depth are shown in parentheses. One thousand bootstrap analyses (distance) were conducted, and percentages >50% are indicated at nodes. Numbers in brackets are GenBank accession numbers. Scale bar represents 2% estimated distance.

The Gammaproteobacteria phylotypes (Fig. 2) dominated (51.5%) the surface layer, and 30 cm b.s.f. (40.0%) and 25 cm b.s.f. (32.7%) layers. The Gammaproteobacteria codominated at 10 cm b.s.f. (25.3%), together with Deltaproteobacteria. The most abundant phylotype in each library was KZNMV-0-B38 (8.8%), KZNMV-30-B2 (32.0%) and KZNMV-10-B1 (10.3%), respectively.

Phylotypes of the Deltaproteobacteria (Fig. 3) dominated the 5 cm b.s.f. layer (68.5%). The dominant phylotype, KZNMV-5-B4 (34.8%), clustered within an as yet uncharacterized clade of Desulfobulbaceae. Phylotypes affiliated with the Deltaproteobacteria also occurred at 10 cm b.s.f. (25.3%), 0 cm b.s.f. (20.6%) and 25 cm b.s.f. (14.5%). Most of these phylotypes fell within SRB groups. At 30 cm b.s.f., only 2.7% of the phylotypes belonged to the Deltaproteobacteria.

Betaproteobacteria (Fig. 3) were only observed at 30 cm b.s.f. Three phylotypes, KZNMV-30-B38, KZNMV-30-B57 and KZNMV-30-B78, comprised 5.3% of the phylotypes at this layer.

The Alphaproteobacteria (Fig. 3) exhibited their highest abundance (10.7%) at 30 cm b.s.f. Moreover, three of four KZNMV-30 phylotypes affiliated with the Alphaproteobacteria formed a distinct clade from the other phylotypes originating from the other layers. This clade included Methylobacterium sp., a known methylotroph. At 10 cm b.s.f., the percentage of Alphaproteobacteria phylotypes was 5.7%, at 25 cm b.s.f. was 3.6% and in the surface layer and at 5 cm b.s.f. was c. 1%.

Phylotypes of the Epsilonproteobacteria (Fig. 3) occurred at all, except for the 5 cm b.s.f. layer. At the surface, 25 and 30 cm b.s.f. layers, only ∼5% of Epsilonproteobacteria of each library were found, whereas phylotypes from 10 cm b.s.f. reached 11.5%. The most abundant phylotype of Epsilonproteobacteria in this layer was KZNMV-10-B1, which was also found at 0 cm b.s.f. (KZNMV-0-B32) and at 25 cm b.s.f. (KZNMV-25-B21) clustered within the Sulfurovum sp. clade.

Among nonproteobacterial phylotypes (Fig. 4), sequences belonging to candidate division JS1 were present in all bacterial 16S rRNA gene libraries constructed in this study. JS1 phylotypes from all examined depths represented a single phylotype and appeared to be elevated in the 5 cm b.s.f. layer (7.9%).

Members of the Chloroflexi were abundant in the 25 cm b.s.f. layer (21.8%). The dominant phylotype in this layer, KZNMV-25-1, was closely related (99% similarity) to sequences found in cold seep sediment from the Gulf of Mexico. Sequences of Chloroflexi were recovered from all examined depths, except the surface layer.

Sequences affiliated to the Firmicutes were found mostly in the 30 cm b.s.f. layer (12%), whereas they appeared in small numbers at the surface (1.5%) and 10 cm b.s.f. (1.1%) layers.

Similarity and diversity indices

The Morisita index of similarity was applied at the phylotype and phylogenetic group (i.e. phylum) level. The index for the bacterial communities was <0.30, except for the 15 and 20 cm b.s.f. layers, at the phylotype level (Fig. 5a). This indicated little overlap among the bacterial phylotypes across these samples, and revealed that each sediment layer harboured a unique community of Bacteria. However, at the phylum level (Fig. 5b), the Morisita index was, on several occasions, higher (c. >0.75). Clustering of the bacterial communities showed four distinct clusters: 1/30, 10/25, 15/20 and 5 cm b.s.f. (the last being close to the 15/20 cm b.s.f. cluster).

Figure 5.

 Cluster analysis based on the Morisita similarity index for the (a) bacterial and (c) archaeal 16S rRNA gene phylotypes, and (b) bacterial and archaeal (d) groups, from sediments of the Kazan MV, East Mediterranean Sea.

The Morisita clustering for Archaea exhibited a similar pattern (c. >0.20) at both the phylotype and the phylum level (Figs 5c and d). The surface layer was closely related to the 15/20 cm b.s.f. layers and distantly related (c. 0.5) to the 10 cm b.s.f. layer. Another cluster was formed from the 5/30 cm b.s.f. layers, while the 25 cm b.s.f. layer did not cluster with any of the other layers.

The Shannon–Wiener diversity index H at all examined depths was higher for Bacteria (1.47–3.82) than for Archaea (0.56–1.73) (Fig. 6). Changes in H with depth were not consistent between archaeal and bacterial clone libraries. The 15 and 20 cm b.s.f. layers appeared to represent a local minimum for both archaeal and bacterial diversity indices. Archaea exhibited the lowest H-value at the 15 cm b.s.f. layer, followed by the surface and 20 cm b.s.f. layers. At all the remaining layers, H was two to three times higher. Bacteria, on the other hand, showed the maximum H diversity at 10 cm b.s.f., but this was roughly comparable with the 1 and 25 cm b.s.f. layers. The Pielou evenness index J (Fig. 6) varied between 0.35 and 0.95 (average 0.71) for Archaea and between 0.76 and 0.96 (average 0.87) for Bacteria, and was higher for Archaea only in the 20 and 30 cm b.s.f. layers. These values of J indicate that the bacterial clone libraries showed a more even distribution of phylotypes than the archaeal libraries.

Figure 6.

 Shannon diversity index H and Pielou evenness index J for the bacterial and archaeal 16S rRNA gene clone libraries from sediments of the Kazan MV, East Mediterranean Sea.


We investigated the diversity of Bacteria and Archaea at a high sediment depth resolution in the Kazan MV, East Mediterranean Sea. Although we took all the necessary steps to minimize the nonquantitative effects of PCR (Bohannan & Hughes, 2003), the innate limitations of the PCR-based approach remain and thus our data refer to the apparent richness of phylotypes and the microorganisms that they represent. Coverage of the clone libraries according to Good's C estimator (Kemp & Aller, 2004) was more satisfactory for the Archaea (>0.8) than for the Bacteria (c. 0.4–0.8) (Fig. 7). This implies that the majority of the Archaea phylotypes have been identified, but for the Bacteria, only the most abundant and several of the low-abundance (i.e. singletons and doubletons) phylotypes have been retrieved. This is also reflected in the higher number of bacterial singletons than archaeal singletons (Table S1). Revealing such rare phylotypes in complex communities is important. Rare species are thought to grow slow (they would otherwise become common), but in habitats such as MVs and deep-sea sediments, all microbial growth is slow (Fuhrman, 2009). For example, although ANMEs metabolize methane quite quickly (Nauhaus et al., 2005) their growth is slow. Thus, the functional role of rare species in the Kazan MV might be more important than it is in other, more productive habitats such as the upper marine water column. Moreover, the coexistence of rare with abundant phylotypes can serve as a ‘seed-bank’ for future perturbations (Pedrós-Alió, 2006), for example after mud fluid eruptions.

Figure 7.

 Clone library coverage based on Good's C estimator of the prokaryotic 16S rRNA gene libraries from the Kazan MV, East Mediterranean Sea.

Community composition

The relative dominance of Euryarchaeota over Crenarchaeota in all layers has been reported for other gas-hydrate- or methane-bearing sediments as well (e.g. Inagaki et al., 2006; Parkes et al., 2007). The Euryarchaeota found were related to methane metabolism. In particular, ANME groups were retrieved at a high relative abundance (>50%) in all depth layers. At 25 cm b.s.f., the dominant phylotype (KZNMV-25-A23) was closely related to the H2-using methanogen Methanogenium marinum. Therefore, the archaeal communities of the top 30 cm in the Kazan MV are highly specialized and selected for life within gas-hydrate sediments. The fact that high concentrations of gas hydrate of the size of rice grains were present even close to the sediment surface can be interpreted as a very strong methane flux (Perissoratis, 2005), providing suitable substrate conditions for ANME.

The Deltaproteobacteria found were highly diverse (Fig. 3), even at the order level (Desulfobacterales, Synthrophobacterales, Bdellovibrionales and Myxococcales). Many phylotypes found at 5 cm b.s.f. fell into the cold seep-associated clades of SRB, SEEP-SRB1 (Desulfosarcina/Desulfococcus-related), SEEP-SRB3 (Desulfobulbus-related) and SEEP-SRB4 (Desulforhopalus-related). Members of SEEP-SRB1 are considered to be the sulphate-reducing partners of ANME-1 and ANME-2 (Knittel et al., 2003), while the ecological role of SEEP-SRB3 and SEEP-SRB4 remains unknown. The most abundant phylotype (KZNMV-5-B4, 50.8% of the Deltaproteobacteria phylotypes) was clustered within an as yet uncharacterized clade of Desulfobulbaceae, which is physically associated with ANME-2c consortia (Pernthaler et al., 2008). In addition, a few phylotypes were related to Desulfobacterium anilii and related environmental sequences. This group is known to include several cultivated and uncultivated hydrocarbon-degrading Bacteria capable of complete oxidation of various aromatic hydrocarbons (Harms et al., 1999).

Non-SRB sequences of the Deltaproteobacteria were also present at all depths. Phylotype KZNMV-10-B49 was related to Bdellovibrio sp. and Bacteriovorax sp. Both genera are aerobic, obligatory predators of Bacteria that forage on a wide variety of susceptible gram-negative microorganisms (Baer et al., 2000). Four phylotypes fell into the Myxococcales, a group often found in marine sediments (Ravenschlag et al., 1999; Wilms et al., 2006; Zhang et al., 2008). Members of this group are known nonobligate predators as well (Burnham et al., 1984). Finally, a singleton phylotype (KZNMV-5-B10) was related to Syntrophus aciditrophus, known to ferment benzoate and living often syntrophically with H2-consuming methanogens (Jackson et al., 1999; Becker et al., 2005).

Based on the archaeal and Deltaproteobacteria phylotypes found, the prokaryotic communities of the Kazan MV sediments consist of microorganisms involved in methane- and, presumably, sulphate-based biogeochemical processes, as is the case for the eastern Mediterranean MV cold seep sediments (Heijs et al., 2006, 2007; Kormas et al., 2008; Omoregie et al., 2008, 2009). The geochemical profiles (Kormas et al., 2008) support the notion that AOM is a metabolic process than takes place at 15 and 20 cm b.s.f., whereas earlier studies placed the AOM zone at the 7–20 cm b.s.f. (Werne et al., 2002, 2004; Haese et al., 2003).

The sequences of Gammaproteobacteria represented diverse organisms (Fig. 2), as is the case in several marine sediments (Li et al., 1999; Inagaki et al., 2003; Bowman et al., 2005). The dominance of this group has been observed in another case of gas-hydrate sediment and was attributed to the opportunistic life strategies of the Gammaproteobacteria (Jiang et al., 2007). Their ecophysiological role is also diverse, but several of the phylotypes recovered by this study were related to characterized sulphur and sulphide oxidizers or to environmental sequences originating from habitats where sulphur and sulphide oxidation occurred or was predicted to occur.

Phylotypes most closely aligned with psychrophilic Legionellales primarily isolated from deep-sea environments were also recovered. Other phylotypes belonging to families that are more physiologically constrained include the methane-oxidizing Methylococcaceae (Methylococcales) (e.g. KZNMV-0-B63, KZNMV-0-B33), which were found mainly at the top layer, as well as at 10 and 25 cm b.s.f., while sulphide-oxidizing Ectothiorhodospiraceae (Chromatiales) (KZNMV-0-B47 and KZNMV-0-B59) were only retrieved from the top layer. The rest of the phylotypes from the top, 10 and 25 cm b.s.f. layers belonged to the orders of Pseudomonadales, Enterobacterales, Legionellales and Thiotrichales or formed clusters that could not be firmly affiliated to any known taxonomic groups within the Gammaproteobacteria and included only uncultured representatives. These uncharacterized groups, however, were clustered with phylotypes recovered from cold seep environments, gas-hydrate-associated sediments, ocean crusts, marine sediments, characterized symbionts–epibionts of marine invertebrates and putative sulphur oxidizers, implying their involvement in the sulphur cycle.

The Epsilonproteobacteria phylotypes found in this work were related to phylotypes from deep-sea sediments, hydrothermal vents, cave waters and cold seeps. The only cultured isolate within this clade is Sulfurovum lithotrophicum, a sulphur-oxidizing chemolithoautotroph from hydrothermal sediments in a black smoker environment (Inagaki et al., 2004). Thus, Epsilonproteobacteria in the Kazan MV sediment were probably involved, together with Gammaproteobacteria, in sulphur cycling. Indeed, the composition of major elements in the sediments of the Anaximander MVs showed a clear enrichment in sulphur compared with the general composition of pelagic sediments of the East Mediterranean Sea (Perissoratis, 2005).

The site studied was characterized by the absence of overlying pelagic sediments, indicative of relatively recent mud flows, suggesting possible oxygen diffusion down to the sediment. At the bottom of the core (35–40 cm b.s.f.), larger hydrates were found. The possible outcropping of these hydrates could cause sediment resuspension and mixing, thus enhancing oxygen diffusion from the sediment surface and creation of microaerophilic conditions. This would also explain the presence and possible metabolic activity of aerobic, methylotrophic Alphaproteobacteria at the bottom (30 cm b.s.f.) layer. Moreover, methylotrophs form different kinds of resting stages such as cysts and endospores, which enables them to survive even long periods of anoxia or lack of methane (Whittenbury et al., 1970).

Among the nonproteobacterial phylotypes, sequences belonging to phylogenetic groups related to deep biosphere and gas hydrate sediments were retrieved. The same JS1 phylotype was present in all layers. 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 (Webster et al., 2004, 2006, 2007; Fry et al., 2006; Inagaki et al., 2006; Parkes et al., 2007), but also from older Kazan MV sediments (Heijs et al., 2008). Sequences of Chloroflexi were recovered from all examined depths, except the top one, and were present in greatest abundance in our clone library from the 25 cm b.s.f. layer. Chloroflexi occur frequently in hydrocarbon-rich sediments and the deep subsurface (Kormas et al., 2003; Inagaki et al., 2006).

Diversity patterns

Cluster analysis was performed to reveal community similarity patterns, at both the phylotype and the phylogenetic group levels, between the different sediment layers. The results from both phylotype and phylum approaches were in agreement for the Archaea (Fig. 5c and d). For the Bacteria (Fig. 5a and b), however, the different layers showed different community similarities under the two different approaches. When phylotypes, especially the most abundant and/or the most dominant ones, are affiliated to a large, functionally diverse group (e.g. Gammaproteobacteria), cluster analysis may lead to ‘false’ community similarities (e.g. between the surface and the 30 cm b.s.f. layers). This highlights the need for diversity comparisons at the phylotype rather than at the phylum level, and that higher taxon analysis is appropriate only in highly specific communities.

The use of the Shannon–Wiener diversity index H in prokaryotic communities seems to be applicable and realistically informative, when clone library coverage is sufficiently large (Hill et al., 2003; Kemp & Aller, 2004), and can be used to compare the diversity between different sites or even studies. In addition, the application of such a universal index, which is fairly easily applied in meta-analysis of published data, will allow reasonable comparisons between hot-spot and non-hot-spot sites. In our study, the minimal H-values for both the Bacteria and the Archaea at 15/20 cm b.s.f. was further evidence that a unique, specialized assemblage was established there, related to AOM according to the inferred ecophysiology of the phylotypes found. Another pattern revealed was that the H-values for Bacteria and Archaea were lower than for previous works from sediments of the same and other MVs (Amsterdam and Napoli) in the eastern Mediterranean (Heijs et al., 2008), even though in that study, the more coarse vertical selection of the extracted sediment may have affected the overall values, but were comparable with hydrate sediments from the Gulf of Mexico (Mills et al., 2005). Bacterial diversity was greater than archaeal diversity for all layers, which is generally the case for bacterial and archaeal libraries constructed from the same sampling location (Aller & Kemp, 2008). The higher bacterial vs. archaeal diversity seems to apply to hydrate-bearing sediments (Mills et al., 2005), cold seeps (Reed et al., 2006) and MVs (Heijs et al., 2008), and is possibly a general trend in methane-related environments (Lanoil et al., 2001; Aloisi et al., 2002; Teske et al., 2002; Nauhaus et al., 2005; Lloyd et al., 2006; Lösekann et al., 2007) based on the relative abundance of phylotypes retrieved from each domain. The different ecophysiological roles of Bacteria and Archaea might primarily involve the use of available energy sources and adaptation to energy stress (Valentine, 2007). For methane oxidizers, extensive environmental, laboratory and modelling studies indicate that their mode of growth yields only small amounts of energy and often occurs at exceedingly slow rates. Valentine (2007) contends that the distribution of catabolic pathways among the Archaea results directly from their adaptation to chronic energy stress and that distinctive mechanisms of energy conservation allow many Archaea to adapt readily to environments of differing energy availabilities. In the case of methanogens and methane oxidizers, dominance is achieved through metabolic exclusivity, whereby these organisms have evolved to exclude or outcompete bacteria by the use of unique catabolic pathways. Bacteria, by contrast, seem to focus on exploiting new or variable resources.

Deep-sea sediment geochemical and physical data are acquired at a fine spatial scale more often than biological samples. For marine MVs, the studies available are characterized by a general lack of data on the temporal changes of their hosted microbial communities. In the case of the Kazan MV, however, Heijs et al. (2008) analysed the prokaryotic diversity in samples taken in 1999 from a site c. 30 m away – but still in the active site of the MV – from the one we analysed. Heijs et al. (2008) used a coarser sampling depth resolution (0–6, 6–22 and 22–34 cm b.s.f.). In addition, the different sediment mass used for DNA extraction and different PCR conditions and primers limit the safety of such comparisons. Because of the scarcity of such data, however, we believe that a comparison of relative abundances is feasible, and after pooling our results according to the depth layers of 0–5, 10–20 and 25–30 cm b.s.f., the two studies revealed very different biodiversities, especially in the top two layers. For Archaea, the top layer was dominated by Halobacteriales-related phylotypes (Heijs et al., 2008), but our study showed that it was dominated by ANME-2. The middle layer showed an overlap in ANME-2 (the second most dominant group in Heijs et al., 2008), although with different phylotypes. The bottom layer was dominated by ANME-2 (Heijs et al., 2008) and by ANME-1 and ANME-3 (present study). The dominant Bacteria in the top two layers were different in the two studies (Actinobacteria, Chloroflexi vs. Gamma- and Deltaproteobacteria). The bottom layer, although containing different dominant phyla in the two studies, showed some overlap for the Deltaproteobacteria and Chloroflexi. Such variations, apart from methodological and microscale differences, could be attributed to differences in the prevailing conditions, for example due to eruptions. We suggest that our sampling strategy revealed a higher prokaryotic diversity and also showed that the distribution of prokaryotes explains the geochemical features found. Future comparisons will confirm the persistent or the ephemeral aspect of these communities.

In conclusion, the present study showed that Bacteria and Archaea communities differ even every 5 cm of the top 30 cm in an active site of the Kazan MV with recent mud flow. The archaeal communities showed a lower diversity than the bacterial communities, but were more closely related to ANMEs, while the Bacteria included AOM-related phylotypes. Only at 15 and 20 cm b.s.f. did the two communities show the typical AOM-related profile, known to occur in similar habitats elsewhere, consistent with the prevailing methane–sulphate conditions.


This research project is cofinanced by the EU-European Social Fund (75%) and the Greek Ministry of Development – GSRT (25%). Sampling was conducted during the European Commission project ANAXIMANDER (contract no. EVK3-CT-2002-00068). The officers and crew of the R/V AEGAEO are gratefully acknowledged for their contribution to fieldwork and sampling. We thank two anonymous reviewers and Virginia P. Edgcomb for their critical reading of the manuscript.