Biogeosciences Biogeochemistry of a low-activity cold seep in the Larsen B area , western Weddell Sea , Antarctica

First videographic indication of an Antarctic cold seep ecosystem was recently obtained from the collapsed Larsen B ice shelf, western Weddell Sea (Domack et al., 2005). Within the framework of the R/V Polarstern expedition ANTXXIII-8, we revisited this area for geochemical, microbiological and further videographical examinations. During two dives with ROV Cherokee (MARUM, Bremen), several bivalve shell agglomerations of the seep-associated, chemosynthetic clamCalyptogenasp. were found in the trough of the Crane and Evans glacier. The absence of living clam specimens indicates that the flux of sulphide and hence the seepage activity is diminished at present. This impression was further substantiated by our geochemical observations. Concentrations of thermogenic methane were moderately elevated with 2 μM in surface sediments of a clam patch, increasing up to 9 μM at a sediment depth of about 1 m in the bottom sections of the sediment cores. This correlated with a moderate decrease in sulphate from about 28 mM at the surface down to 23.4 mM, an increase in sulphide to up to 1.43 mM and elevated rates of the anaerobic oxidation of methane (AOM) of up to 600 pmol cm −3 d−1 at about 1 m below the seafloor. Molecular analyses indicate that methanotrophic archaea related to ANME-3 are the most likely candidates mediating AOM in sediments of the Larsen B seep. Correspondence to: H. Niemann (helge.niemann@unibas.ch)


Introduction
Ocean research of the last decade has provided evidence for a variety of fascinating ecosystems associated with fluid, gas and mud escape structures.These so-called cold seeps are often colonized by thiotrophic bacterial mats, chemosynthetic fauna and associated animals (Jørgensen and Boetius, 2007).Cold seep sediments also harbour diverse microbial populations that degrade hydrocarbons anaerobically along fluid and gas escape pathways.Key agents of the anaerobic oxidation of methane (AOM) with sulphate are consortia of methanotrophic archaea (ANME) and sulphate-reducing bacteria (SRB) (Knittel and Boetius, 2009).These consortia effectively reduce the efflux of methane to the hydrosphere (Hinrichs and Boetius, 2002;Reeburgh, 2007), and are the source for sulphide which fuels the chemosynthetic seep fauna (Jørgensen and Boetius, 2007).So far, three archaeal groups of marine, anaerobic methanotrophs (ANME-1, -2 and -3) related to the methanogenic Methanosarcinales and Methanomicrobiales were discovered (Hinrichs et al., 1999;Boetius et al., 2000;Orphan et al., 2001Orphan et al., , 2002;;Knittel et al., 2005;Niemann et al., 2006b;Lösekann et al., 2007).The (presumed) sulphate-reducing partner bacteria are members of the Desulfosarcina/Desulfococcus (ANME-1 and -2) or the Desulfobulbus branch (ANME-3) of the δ-proteobacteria (Boetius et al., 2000;Orphan et al., 2001;Knittel et al., 2003;Niemann et al., 2006b;Lösekann et al., 2007).In addition to the sulphate-dependent mode of AOM, recent studies demonstrated that alternative oxidants are utilised as terminal electron acceptors during AOM.Denitrifying bacteria of the candidate division "NC10" were found to mediate AOM with nitrate or nitrite in a bioreactor with sediments from a fresh water canal systems (Raghoebarsing et al., 2006;Ettwig et al., 2008).Also, incubations of cold seep sediments from the Eel River Basin augmented with oxidised iron and manganese showed an increase in AOM activity (Beal et al., 2009).
Cold seep systems were discovered at almost all continental margins, from Arctic to tropical latitudes.No marine cold seep ecosystems were known in Antarctica until the recent discovery of a potential methane seep in the trough of the Evans and Crane glacier in the Larsen B area by Domack and co-workers (2005).After the collapse of the Larsen B ice shelf at the South-Eastern margin of the Antarctic Peninsula in 2002 (Rott et al., 1996;Shepherd et al., 2003;Rack and Rott, 2004), Domack et al. (2005) conducted a videographic survey of the seafloor in the now open water embayment of Larsen B in March 2005.At the time of observation, 50 to 70% of the explored seafloor (5540 m 2 ) were covered by some whitish material, which the authors interpreted as sulphide-oxidising microbial mats.Such mats are typical for highly active seeps where AOM leads to high sulphide concentration in near-surface sediment horizons (Boetius et al., 2000;Treude et al., 2003;Niemann et al., 2006b).Furthermore, the video-footage showed features that were interpreted as bubbles emanating from the seafloor and traces of mudflows.The authors also observed aggregations of large bivalve shells, which were most likely extinct populations of vesicomyid clams (see Fig. 2d in Domack et al., 2005).Just as the sulphide-oxidising bacterial mats, vesicomyid clams are indicative for highly reduced, sulphidic environments, because the known species harbour sulphide-oxidising symbionts in their gills tissue and depend on their chemosynthetic production (Fisher, 1990;Childress and Fisher, 1992;Goffredi and Barry, 2002).
In January 2007, we revisited the Larsen B area for videographic and geochemical examinations within the framework of the R/V Polarstern cruise ANTXXIII-8, an expedition contributing to the "Census for Antarctic Marine Life" (Gutt, 2008).To improve the biogeography of deep-water chemosynthetic ecosystems at a global scale, and to understand the processes driving these ecosystems, we aimed at estimating the extent of seep activity in the Larsen B area, and to identify the microbial community as well as megafauna associated with methane seepage and chemosynthetic production.

Bathymetric mapping and seafloor observations
Bathymetric mapping of the Crane and Evans glacier trough was performed with R/V Polarstern's multibeam echosounder system Hydrosweep DS2 (Atlas Hydrographic) operated in a 90 • aperture angle and 59 hard beam mode.
Recorded raw data were edited with the CARIS HIPS and SIPS and subsequently mapped using the Generic Mapping Tool (GMT) software packages.Visual observations of the seafloor were carried out with the ROV Cherokee (MARUM, University of Bremen) and a video-guided multi grab sampler (Gerdes et al., 1992).The video footage was used to select positions for sediment sampling (photographs as well as metadata to videos recorded can be found in the PANGAEA data base -http://doi.pangaea.de/10.1594/PANGAEA.702077).

Sample collection and storage
Sediment samples were collected with a video-guided multiple corer (MUC) and a gravity corer (GC) at positions of clam beds, the surrounding area within the glacier trough and from a reference station outside the trough (Table 1, Fig. 1).Immediately upon corer retrieval, sediment samples were transferred into a cold room (2 • C) and subsampled for concentration measurements of hydrocarbon gases, pore water constituents (sulphate, sulphide), AOM and sulphate reduction (SR) rate measurements and the analysis of microbial diversity.Sediment subsampling for gases and DNA was performed with cut-off syringes (GC and MUC cores) and for AOM and SR rates with glass tubes (GC cores) and acrylic core liners (MUC cores) according to previously described methods (Treude et al., 2003;Niemann et al., 2006a).Pore water was extracted with rhizons connected to syringes through holes drilled into the core liners (Seeberg-Elverfeldt et al., 2005).A vertical sampling resolution of 2 cm was chosen for MUC-cores, and of 10 to 20 cm for GCs.AOM and SR rates were determined from 3 (MUC-cores) or 5 replicated (GCs) incubations, whereas limitations of station time and/or sampling material did not allow to sample replicates for the analyses of pore water constituents, gases and DNA.

Methane and ethane concentrations
Hydrocarbons were analysed by static headspace gas chromatography according to previously described principles (James and Martin, 1952;Harley et al., 1958;Kolb and Ettre, 2005).In short, 3 ml of sediment were fixed in 7 ml of saturated NaCl solution in a 20 ml gas-tight sealed glass vial.Shortly after the cruise (<3 month), methane and ethane concentrations were determined from the headspace by gas chromatography and flame ionisation detection.The gas chromatograph (Agilent 6890N) was equipped with an OPTIMA-5 capillary column (5 µM film thickness; 0.32 mm ID, 50 m length) and a split/splitless injector (operated in splitless mode).The carrier gas was helium at a flow rate of 3 ml min −1 .Initial oven temperature was set to 45 • C held for 4 min, subsequently increased at 15 • C min −1 to 155 • C, held for 2 min, then raised at 25 • C min −1 to 240 • C and finally held for 7 min.The chromatography system was calibrated for a concentration range of 1 to 100 µM methane and 0.1 to 10 µM ethane (final concentration in the sediment).Concentrations of hydrocarbons with more than 2 carbon atoms (propane, butane etc.) were below detection/quantification limit.

Sulphate and sulphide concentrations
Sulphate and sulphide concentrations were analysed according to modified methods from Small et al. (1979)  and Popowsky (1949), respectively, as described in detail elsewhere (Grasshoff et al., 1983).Briefly, for sulphate concentration measurements, 0.4 ml of filtered pore water was diluted with MilliQ water (1:50, v/v) and the sample was stored at −20 • C till further analysis.Shortly after the cruise (<2 month), sulphate concentrations were determined by ion chromatography (IC) using a Metrosep Anion Dual2 column (4.6×70 mm) in a Metrohm 761 Compact IC equipped with a 732 IC Detector.Na 2 CO 3 /NaHCO 3 (3.2mM/1 mM, respectively) was used as a solvent at a flow rate of 0.7 ml min −1 .For sulphide concentration measurements, 1.5 ml of pore water was fixed in 0.6 ml zinc acetate (ZnAC) solution (20%, w/v) and stored at 4 • C till further analysis.According to the Methylene-Blue method, sulphide was determined photometrically at 660 nm using a Kontron Uvikon 922 photometer.

Ex situ AOM and SR rate measurements
AOM and SR rates were determined by radiotracer incubations using 14 C-methane and 35 S-sulphate, respectively, as described previously (Jørgensen, 1978;Treude et al., 2003;Niemann et al., 2005).In short, aqueous tracer solutions (40 µl 14 CH 4 , 5 kBq or 5 µl 35 SO 2− 4 , 40 kBq) were injected into butyl rubber sealed glass tubes (GC core subsamples) or in 1 cm intervals into small push cores (MUC core subsamples).The incubations were carried out for 24 h at 1 • C. Subsequently, AOM and SR rate samples were fixed in 25 ml NaOH (2.5%, w/v) and 20 ml ZnAC solution (20%, w/v), respectively.The analyses of radioactive substrates and products were performed as described previously (Fossing and Jørgensen, 1989;Treude et al., 2003).AOM and SR rates were calculated according to the following formulas: Here, 14 CH 4 , 14 CO 2 , 35 SO 2− 4 and TRI 35 S are the activities (Bq) of methane, carbon dioxide, sulphate and total reduced inorganic sulphur species, respectively.[CH 4 ] is the average methane concentration between the beginning and end of the incubation and [SO 2− 4 ] is the sulphate concentration at the beginning of the incubation.Concentrations were corrected for porosity.t denotes incubation time.Aerobic methane oxidation (MOx) was determined in the same way as AOM.In both processes, the oxidized 14 C-methane is trapped as 14 CO 2 but the availability of oxygen determines which process occurs (Niemann et al., 2006b).

DNA extraction and clone library construction
Total DNA was extracted from ca. 30 g of wet sediment recovered from the upper margin of the sulphate-methane transition zone (SMTZ) of core 706-4 (110 cm below seafloor -b.s.f.) according to the method described by Zhou et al. (1996).PCR amplification of 16S rRNA genes, sequencing and phylogenetic analysis of archaeal and bacterial 16S rRNA genes was performed according to previous works (Knittel et al., 2003(Knittel et al., , 2005)).Sequences were checked for chimeras using RDP programme Chimera Check (version 2.7) (http://rdp8.cme.msu.edu/cgis/chimera.cgi?su= SSU).The sequence data reported here will be published in the EMBL, GenBank and DDBJ nucleotide sequence databases under the accession numbers FN429775 to FN429813.

Seafloor observations
Two years after the discovery of a potential cold seep habitat in the trough of the Evans and Crane glacier (Domack et al., 2005), we revisited the Larsen B area (Fig. 1a) for further videographic and the first biogeochemical examinations.The bathymetry in the Larsen B embayment was caused by glacier erosion forming a trough in an East-West direction with several depressions deeper than 800 m water depth.In the easternmost depression (Fig. 1b), we examined the seafloor during two ROV (stations 706-1, S-N; 706-2, W-E) and multi grab transects (stations 706-3, N-S; 709-6, E-W) (Fig. 1c), of which the South-North ROV transect was previously surveyed with a video sledge by Domack et al. (2005) in March 2005.The total area monitored in 2007 was ca.5500 m 2 (assuming a 2 m wide field of vision).However, we could not detect any whitish material (putative bacterial mats) in the wider surrounding of the clam shell accumulations.The seafloor was mainly covered by pelagic sediments and phytodetritus, giving the seabed a green-greyish appearance (Fig. 2a).A survey with R/V Polarstern's fisheries echosounder (Simrad TM EK60, frequencies: 38, 70, 120, 200 kHz, beam width: 7 • , 1-2 kts survey speed, Sauter et al., 2006) could not detect gas bubbles in the water column (data not shown).However, our observations could confirm the presence of bivalve agglomerations of which we found about 10 dense beds and 3 more scattered aggregations of approximately 0.5 m in diameter (Figs.1c, 2b, c).The bi-valve shells belonged to the vesicomyid clam genus Calyptogena and were on average about 8 cm long (Fig. 2d).Living specimens of this chemosynthetic genus provide evidence for sulphidic conditions in near surface sediments (Fisher, 1990;Goffredi and Barry, 2002;Sahling et al., 2002Sahling et al., , 2008)), which are typically found at active cold seeps.Nevertheless, videographic observations (Fig. 2b, c) and samples recovered with a TV-guided MUC and multigrab only revealed empty shells.Calyptogena sp.occurs at seep habitats with low to medium methane fluxes, since it can reach into subseafloor sulphide zones down to 20 cm b.s.f.(Sahling et al., 2005).Living clams tend to accumulate in patches of 0.5-5 m, and are oriented vertically in the sediments (Olu-Le Roy et al., 2007).However, the images from the video survey obtained by Domack et al. (2005) show accumulations of horizontally oriented and therefore most likely dead specimen, so that it remains unknown when sediments were sulphidic enough to support growth of these chemosynthetic bivalves.Furthermore, in 2007, we did not observe any whitish precipitates nor recorded signs of gas ebullition as described by Domack

Geochemical observations and seep activity
During the expedition ANTXXIII-8, one TV-guided MUC (core 709-3) was launched directly above a scattered aggregation of clams, one GC (core 706-4) was recovered from a position where most dense clam beds had been observed and two GCs were recovered about 1.4 km to the west (core 711-4) and 2.7 km to the south (core 711-5) of the ROV transect intersection, respectively (Table 1, Fig. 1b, c).To compare the geochemical data from these cores (inferred to be associated with methane seepage) to the background biogeochemistry, one additional GC (core 702-6) was recovered from a reference station 73 km to the south-east of the glacier trough (Fig. 1a).Surface sediments within a scattered aggregation of clam shells (core 709-3, Fig. 3a) showed slightly elevated methane concentrations (<2 µM) in comparison to background levels (<1 µM throughout core 702-6, Fig. 6a).Furthermore, moderately elevated methane concentrations in sediments deeper than 30 cm b.s.f. were found in the GCs that were recovered from the glacial trough.Highest concentrations were found at the clam shell patches with up to 9.1 µM at depth below 95 cm b.s.f. in core 706-4 (Fig. 4a).
With distance to the shell aggregations, methane concentrations tended to decrease indicating a decrease in seep activity.In core 711-4, a maximum of 7.5 µM was found at 90 cm b.s.f.(Fig. 5a) and only 4 µM could be found with depth below 70 cm b.s.f. in core 711-5 (Fig. 5b).In the core with most methane (core 706-4), a moderate decrease in sulphate down to 23.4 mM and an increase of sulphide up to 1.43 mM below 100 cm b.s.f. was measured.Also MUC core 709-3 showed a decrease in sulphate down to 22.4 mM at 22 cm b.s.f.(Fig. 3b).However, sulphide concentrations were not elevated in this core (Fig. 3c).Cold seeps and hydrothermal vents are often characterised by an advective upward flow of chloride-depleted geofluids, which can be traced by a conspicuous decrease in chloride concentrations with sediment depth (von Damm et al., 1995;Schulz and Zabel, 2006;Hensen et al., 2007).Chloride concentrations in core 706-4 were uniform with about 520 mM and only pore waters in the bottom section were slightly depleted with a concentration of 492 mM (data not shown).This indicates that the advection of geofluids is very low and that diffusion is the dominant mass transport mechanism of pore water solutes at Larsen B; at least in the upper metre of sediments.According to Fick's first law of diffusion, the methane, sulphate and sulphide gradients at station 706-4 (0.0003, 0.1, 0.04 µmol cm −4 , respectively) together with the diffusion coefficients (193, 117, 239 cm 2 yr −1 , respectively; porosity = 80%, T =2  (ca.0.00025 and 0.0001 µmol cm −4 translating to 0.38 and 0.2 mmol m −2 yr −1 at station 711-4 and 711-5, respectively) further indicating a decrease in seep activity with distance to the clam patches.Unfortunately, a layer of glacial gravel at about 100 cm b.s.f.prevented the GCs to penetrate deeper into the sea floor.Hence, the cores sampled included only the upper boundary of the SMTZ and fluxes of methane, sulphate and sulphide could be higher in deeper sediment layers.
In contrast to seep-influenced sediments, no apparent geochemical gradients could be detected at the reference station (core 702-6).Here, methane concentrations were uniform and <1 µM (Fig. 6a) and sulphide concentrations were below detection limit (Fig. 6b).Sulphate (Fig. 6b) and chloride concentrations (data not shown) were roughly 28 and 520 mM, respectively, in the lower section of the core (down to 140 cm b.s.f.).
Methane consumption in surface sediments of the clam patch (600 pmol cm −3 d −1 , core 709-3, Fig. 3d) was not coupled to SR.Although we did not measure oxygen and nitrate concentrations, the beige colour of the upper 8 cm of sediments in core 709-3 indicates the nitrate penetration depth.Accordingly, the oxygen penetration depth was probably a few centimeters above 8 cm b.s.f.Similar oxygen penetration depths have been reported previously from Weddell Sea sediments between 500 and 2000 m water depth (Rutgers Van der Loeff et al., 1990).Hence, the detected methane consumption in surface sediments was most likely due to aerobic oxidation of methane (MOx).Furthermore, elevated rates of SR just beneath the MOx peak in core 709-3 (Fig. 3d) and at 33.5 cm b.s.f. in core 706-4 (Fig. 4d) did not correlate with methane oxidation and were thus probably fuelled by buried  detritus.However, in the bottom of core 706-4, we could also detect elevated, correlating rates of AOM and SR of about 600 pmol cm −3 d −1 (Fig. 4d).seep system, similar to some cold seep systems in the Gulf of Cadiz which do not host thiotrophic mats or chemosynthetic bivalves, but sporadically populations of siboglinid tubeworms (Niemann et al., 2006a).In comparison, methane concentrations in surface sediments at active seeps supporting thiotrophic microbial mats or dense clam populations (e.g., Haakon Mosby Mud Volcano, the Gulf of Mexico, Hydrate Ridge) are in the millimolar range, and sulphate is consumed in the first few centimetres of the sediment, leading to high sulphide concentrations >4 mM (Boetius et al., 2000;Torres et al., 2002;Joye et al., 2004;de Beer et al., 2006;Niemann et al., 2006b).In these systems, the rates of AOM and SR are orders of magnitude higher and may exceed 1 µmol cm −3 d −1 in surface sediment horizons often amounting to methane, sulphate and sulphide fluxes of some moles m −2 yr −1 (Boetius et al., 2000;Treude et al., 2003;Joye et al., 2004;Niemann et al., 2006b).Yet, the presence of Calyptogena sp.shells provides evidence for a higher seep activity in the past.In sediments below clam shell aggregations, micro to millimolar sulphide concentrations were detected only in sediment horizons below 50 cm b.s.f.(Fig. 4c) and the sulphide flux was about 0.07 mol m −2 yr −1 .However, the maximum depth that Calyptogena sp. can reach is 20 cm b.s.f.(Sahling et al., 2005, and references therein), and the lowest sulphide flux supporting Calyptogena sp.found so far is 0.3 mol m −2 yr −1 (Juniper et al., 1992;Grehan and Juniper, 1996;Sahling et al., 2005).Thus, from the time the Calyptogena populations at the Larsen B shelf were alive, the sulphide front has moved ca.30 cm deeper into the sediment corresponding to a decrease of the sulphide flux by about 4-fold.

Origin and composition of light hydrocarbons
In addition to methane, elevated ethane concentrations of up to 0.6 µM (core 711-4) were found in deeper sediment horizons (Figs. 4a,5a,b).Interestingly, ethane and methane concentrations correlated.Comparable to previous findings (Niemann et al., 2006a), this indicates concurrent transport from a single source in the deeper subsurface and subsequent anaerobic consumption of higher hydrocarbons in shallow sediments possibly mediated by SRB (Kniemeyer et al., 2007).Ethane contributed 3.6, 7.2 and 2.5% to the analysed hydrocarbon gases in the bottom sections of cores 706-4, 711-4 and 711-5, respectively.Other gases were not analyzed and/or below detection limit.No ethane was found in sediments of the reference station (Fig. 6a).A distinction of thermogenic and microbial methane is usually based on the source gas composition in combination with their stable isotope ratios (δ 13 C, δD) because abiotic and biotic processes influence the molecular and isotopic composition of light hydrocarbons during production, migration and consumption in the seabed (Bernard et al., 1976;James and Burns, 1984;Larter and di Primio, 2005;Stadnitskaia et al., 2006).AOM for instance leads to ethane enrichment relative to methane and a 13 C-enrichment of the residual methane (Whiticar, 1999;Hinrichs and Boetius, 2002).Although, we did not sample the gas source, the composition of light hydrocarbons at the Larsen B cold seep shows exceptionally low C 1 /C 2 ratios (15 to 29) with δ 13 C-CH 4 values (data not shown) of −32.0‰V-PDB (110 cm b.s.f.) and −37.6‰V-PDB (120 cm b.s.f.) in core 706-4.This indicates a thermogenic origin of methane at the study site, but probably also 13 C enrichment of the reservoir gas caused by its anaerobic consumption in shallow sediments.A thermogenic origin of light hydrocarbons agrees well with the geological setting of the Larsen B area.The bedrock beneath the trough consists of Mesozoic marine shale, which is a potential source rock for hydrocarbons (Domack et al., 2005).The geological setting of the Larsen B area may even support an abiotic, hydrothermal methane production (Lollar et al., 1993;Foustoukos and Seyfried, 2004).As part of a back arc basin, it is associated with the volcanic arc on the Antarctic Peninsula and a subduction zone in the Drake Passage (Anderson, 1999, and references therein;Hathway, 2000).

Identity of AOM communities
The composition of the microbial assemblage in the AOM horizon below a clam shell aggregation (core 706-4; 110 cm b.s.f.) was studied by 16S rRNA gene analysis.A total of 118 archaeal and 71 bacterial 16S rRNA gene sequences were analysed (Table 2).In the following, relative proportions of archaeal and bacterial groups are defined as "% of total archaeal sequences" or "% of total bacterial sequences".Among the archaeal sequences, 10 OTUs (operational taxonomic unit; 98% sequence similarity) were detected.The largest archaeal clone group (54.2%) was affiliated with uncultivated Methanomicrobiales.Cultivated members of this group are methanogens (Garcia et al., 2006).The second largest group (32.2%) was related to the Methanosarcinales and included sequences of uncultivated archaea related to ANME-3 and other AOM-associated archaea (AAA).Since neither ANME-1 nor ANME-2 sequences could be retrieved, ANME-3 archaea are most likely the dominant anaerobic methanotrophs at Larsen B. The seep site investigated here is the only AOM habitat described so far where no ANME-1 and ANME-2 sequences were found.Usually, ANME-1 and/or ANME-2 sequences are present in cold seep sediments and comprise the dominant group of anaerobic methanotrophs, while ANME-3 has been detected only rarely in these systems (Knittel et al., 2005;Omoregie et al., 2008Omoregie et al., , 2009;;Knittel and Boetius, 2009;Roussel et al., 2009).So far, only one other cold seep, the Arctic Haakon Mosby Mud Volcano, has been found where ANME-3 is the dominant ANME type (Niemann et al., 2006b;Lösekann et al., 2007).This raises the question whether ANME-3 is best Other potential anaerobic methanotrophs could also be archaea belonging to the AAA cluster (Knittel and Boetius, 2009).This group comprises many sequences from terrestrial and limnic sources, but also from a deep-sea sulphide chimney (Schrenk et al., 2003).However, the biogeochemical functioning of these archaea is not well constrained at present (Ettwig et al., 2008).Sequences of the Thermoplasmatales were also found (13.6%).Their functioning is also unclear, but they have previously been found to co-occur with ANMEs (Hinrichs et al., 1999;Orphan et al., 2001;Inagaki et al., 2004;Girguis et al., 2005;Knittel et al., 2005;Wegener et al., 2008).Interestingly, no Crenarchaeota were found although they are common members of seep communities (Knittel et al., 2005;Lösekann et al., 2007).
The bacterial diversity was higher than the archaeal diversity.71 sequences could be assigned to 29 OTUs of which 17 belong to the Deltaproteobacteria (57.8%).Interestingly, we could not find the known sulphate reducing partner, Desulfobulbus, which forms a consortium with ANME-3 at the Haakon Mosby Mud Volcano (Niemann et al., 2006b;Lösekann et al., 2007).This may indicate that ANME-3 mediates AOM with another partner at the Larsen B seep.Diverse syntrophic partnerships have also been shown for ANME-2 archaea (Pernthaler et al., 2008).However, we cannot exclude undersampling of the bacterial sequences with the applied set of methods.We detected one sequence (1.4%) of the Desulfosarcina/Desulfococcus Seep Cluster I, presumably the partner bacteria of ANME-1 and -2, as well as 3 other sequences (4.2%) of the Desulfosarcina/Desulfococcus group.Also, clones belonging to Desulfobacterium indolicum/Desulfatirhabdium butyrativorans (2.8%) and Desulfobacterium anilini (4.2%), which are related to the Desulfosarcina/Desulfococcus group were found.The function of these SRB at Larsen B remains speculative but some could be involved in the degradation of higher hydrocarbons.Desulfobacterium anilini relatives are known to degrade aromatic hydrocarbons (Widdel, 2009) and one of the clones without a close cultivated relative (clone Bac3) was affiliated with a butane degrading strain (strain BuS5, clone Butane12-GMe, Kniemeyer et al., 2007).The most abundant bacterial sequences (23.9%) were distantly or closely affiliated with Desulfocapsa spp.The cultivated species Desulfocapsa thiozyomogenes and Desulfocapsa sulfoexigens grow well by disproportionation of sulphur (in the presence of a sulphide scavenger), thiosulphate and sulphite (Janssen et al., 1996; for a review see Rabus et al., 2006).In addition, Desulfocapsa thiozymogenes can grow by reduction of sulphate to sulphide.However, as with the Desulfosarcina/Desulfococcus relatives, the functioning of Desulfocapsa spp. at Larsen B cannot be determined at present.

Conclusions
Based on videographic evidence, it was recently proposed that the former ice shelf area at Larsen B hosts the first cold seep system discovered in Antarctic waters (Domack et al., 2005).During R/V Polarstern cruise ANTXXIII-8, bivalve shells could be sampled and were identified as Calyptogena sp., a genus of chemosynthetic bivalves, which typically populates sulphidic sediments associated with intermediate to low methane seepage.However, neither living clam populations, thiotrophic bacterial mats or gas ebullition were observed, which indicates a reduction of seepage activity.Biogeochemical analyses of the seabed showed an elevated AOM activity below 1 m depth in the seafloor, and that methane, sulphide and sulphate transport was dominantly diffusive.The methane source for the Larsen B seep is thermogenic (and possibly hydrothermal), based on the hydrocarbon composition containing considerable amounts of ethane, and a stable carbon isotope signature of methane in the range of −32 to −37‰ (V-PDB).The AOM zone hosted members of the ANME-3 clade of anaerobic methanotrophs as well as of a new group of archaea associated mostly with terrestrial AOM, and showed a high diversity of sulphatereducing bacteria.

Fig. 2 .
Fig. 2. Images showing (a) the sea floor typically encountered during video surveys in the trough of the Crane and Evans glacier and (b) a dense bed of Calyptogena sp.shells at the sea floor.Elasipod holothurians and drop stones are visible on images (a) and (b).(c) A close up of the clam bed shows only empty shells.Shell agglomerations were photographed on the South-North transect (Photographs: J. Gutt and W. Dimmler, ©AWI/Marum, University of Bremen).(d) Calyptogena sp.shells recovered during R/V Polarstern expedition ANTXXIII-8.

Fig. 3 .
Fig. 3MOx SR Fig. 4 Fig. 4. Concentration profiles of (a) methane and ethane, (b) sulphate, (c) sulphide as well as rates of anaerobic oxidation of methane (AOM) and sulphate reduction (SR) in sediments of gravity core 706-4 which was recovered from a position where many dense beds of clam shells were observed during videographic examinations (see Fig. 1c for position).

Table 1 .
Sediment cores recovered during expedition ANTXXIII-8 in the Larsen B area.The analyses conducted for each core are indicated.
H.Niemann et al.:Biogeochemistry of a low-activity cold seep in the Larsen B area adapted to the ice-cold temperature conditions of Arctic or Antarctic waters or if additional factors select for this particular group in these environments.