Introduction

Unlike the majority of seafloor environments, hydrothermal vents and cold seeps represent marine ecosystems that may be largely independent of the input of photosynthetic primary production as its primary energy source (e.g. Levin 2005; Stewart et al. 2005; Saito and Osako 2007). In situ chemosynthetically produced microbial biomass, either as free-living prokaryotes or in symbiotic relationships with fauna, can be a major, sometimes the only, source of energy for resident benthic fauna. Unravelling the trophic relevance of the photosynthetic versus chemosynthetic organic carbon is of primary importance to improving our understanding of the functioning of these ecosystems. While a large amount of information was acquired on the trophodynamics of the surface-dwelling megafauna (e.g. large tube worms, mussels and clams) that colonise these habitats, very limited information is available on the sediment-dwelling fauna (infauna). Data available from larger macrofauna inhabiting deep-sea seeps suggest a dependence on chemosynthetic derived carbon, typically stronger than that estimated for shallow-water seeps (Levin 2005). Furthermore, this nutritional link is primarily heterotrophic (feeding on free-living chemoauto- or heterotrophic bacteria) and it depends to a lesser extent on bacterial symbiotic relationships (Levin 2005). Trophoecological studies on the metazoan meiofauna are scarce, primarily due to the small size of the fauna and the inherent difficulty in obtaining sufficient biomass for subsequent analyses of the stable isotopes or fatty acids (FAs). Nematodes generally represent a major fraction of seep sediment infauna, and some of the highest marine nematode densities and biomasses were observed in cold seep sediments (e.g. Olu et al. 1997; Van Gaever et al. 2006); bulk tissue stable carbon and nitrogen isotope signatures have revealed that some cold seep nematodes utilise chemosynthetically derived carbon (Van Dover et al. 2003; Van Gaever et al. 2006). At the seep addressed in this study, Van Gaever et al. (2006) reported the strong dominance and proliferation of a single nematode species, Halomonhystera disjuncta, in the microbial mat sub-habitat. The striking high nematode biomass could not be attributed to photosynthetically produced carbon because of the low organic matter input. Moreover, in the seep centre, the nematodes were replaced by a species of copepods (morphologically very similar to Tisbe wilsoni) with bulk tissue δ13C values indicative of a strong trophic link with methane-derived carbon (δ13C ~ −51‰).

Meiobenthic (nematode) assemblages in non-seep or shallow seep sediments are more diverse and may be related to the presence and utilisation of both photosynthetically and chemosynthetically derived carbon (Levin et al. 2000; Levin 2005). However, the extent to which these meiobenthic diversity and biomass patterns are related to trophic resources remains unclear, just as the carbon flow pathways (symbiosis versus feeding on free living prokaryotes; Levin et al. 2000; Levin and Mendoza 2007).

In this study, we report on a detailed investigation of metazoan meiofauna trophoecology at the Håkon Mosby Mud Volcano (HMMV). HMMV is located at a water depth of 1,280 m on the Barents Sea and is the only mud volcano in a polar region that was mapped in detail by photo and video camera observations and studied in detail for its biogeochemical habitats and microbial communities (Niemann et al. 2006; de Beer et al. 2006; Jerosch et al. 2007). Significantly different meiobenthic densities were found at the different sub-habitats (seep centre, microbial mats, Siboglinidae field and outer rim); the anoxic and sulphidic sediments at the microbial mat habitat were inhabited by extremely high densities (>11,000 ind. 10 cm−2, 0–5 cm sediment depth) of H. disjuncta (originally referred to and described as Geomonhystera disjuncta; Van Gaever et al. 2006). The central, seep part of the volcano was characterised by high concentrations of methane and the lack of sulphide (de Beer et al. 2006). Moreover, here nematodes were virtually absent while copepods were abundant. Different sediment geochemistry across seep sub-habitats supporting distinct microbial communities (Niemann et al. 2006) may strongly regulate and explain meiofaunal distribution patterns in this deep-sea cold seep (Van Gaever et al. 2006). Moreover, the tolerance of the nematode to high sulphide levels could be the main reason for its dense, mono-specific population at the microbial mats, in contrast to the absent copepod species. However, these density distribution patterns as well as trophic relationships remain unresolved and in this study we attempt to unravel carbon sources as well as carbon flow pathways of the two dominant taxa using multiple molecular approaches and transmission electron microscopy (TEM).

Bacterial symbionts (ecto- or endosymbionts) can be either a source and pathway of chemosynthetically derived organic matter entering the higher trophic levels or play a role in detoxification in fauna inhabiting sulphide-rich sediments (e.g. Thiermann et al. 2000; Levin 2005). We attempted to locate and identify symbiotic bacteria in the nematode H. disjuncta, using TEM photography coupled with automated rRNA intergenic spacer analysis (ARISA), fluorescent in situ hybridisation (FISH) and catalysed reporter deposition fluorescent in situ hybridisation (CARD-FISH). Fatty acid (FA) analysis combined with stable isotope analysis was applied to unravel the origin of the diet of the dominant copepod from the volcano’s centre and the dominant nematode from the microbial mats. Fatty acid analyses were repeatedly used as source-specific indicators of organic matter, both in environmental and food web studies (e.g. Canuel et al. 1995; Kharlamenko et al. 1995; Phleger et al. 2005). This tool can be complemented by the analysis of stable carbon isotope signatures of the dominant FAs for further clarification of the food sources and trophic pathways (e.g. Abrajano et al. 1994; Pond et al. 1997; Lösekann et al. 2008). With methane at HMMV having a δ13C of ~−60‰ (Lein et al. 1999) and microbial mat sediment porewater dissolved inorganic carbon δ13C ~−26‰ (mass balance estimate of sediment-water slurry incubations, Van Gaever et al. 2006), in situ chemosynthetically produced carbon is expected to be reflected in more negative δ13Corg values than that of surface water photosynthetically produced carbon (e.g. Levin and Michener 2002; MacAvoy et al. 2002), although this can sometimes be masked or magnified depending on the carbon fixation pathway (e.g. Robinson and Cavanaugh 1995; Stewart et al. 2005).

Materials and methods

Sampling strategy

Sediment samples for meiofaunal analyses were collected at the HMMV during the VICKING cruise aboard the RV “PourquoiPas?” in April–May 2006. The flat central zone of the volcano is characterised by grey mud flows with a high geothermal gradient (Pimenov et al. 2000). Beyond this central plain, the seafloor has a complex topography of hills and depressions, and is covered by extensive microbial mats of chemoautotrophic bacteria, dominated by giant sulphur-oxidising Beggiatoa bacteria (Pimenov et al. 2000; Niemann et al. 2006). Samples were retrieved at both the volcano’s centre (72°0.28′N; 14°43.56′E; 1273 m) and the surrounding microbial mats (72°00.16′N; 14°43.95′E; 1258 m). A multiple corer (surface area of 30.2 cm²) was deployed at the centre site, while the microbial mats were sampled with push cores of the ROV Victor 6000 (surface area: 21.2 cm²).

Transmission electron microscopy and molecular techniques

Transmission electron microscopy and multiple molecular approaches were used to identify possible bacterial symbionts in the dominant nematode species. The nematodes from the microbial mat sediment for TEM were picked out directly onboard, cleared of adhering sediment particles and immediately fixed in a Karnovsky solution diluted 50% with distilled water at room temperature. Further processing in the laboratory was according to the protocol in Willems et al. (2005) and cross-sections of 70 nm thick were examined using a TEM 1010 Jeol, operating at 80 kV.

Bulk sediment samples from the microbial mat site for nematode FISH and CARD-FISH were stored onboard in 4% formaldehyde at room temperature. Details of FISH and CARD-FISH are given in the online appendix (see “Protocol FISH and CARD-FISH”). For FISH, isolated and cleaned nematodes were further processed in Petri dishes according to the methods described in Pernthaler et al. (2002) and Vandekerckhove et al. (2002), followed by epifluorescence microscopy (Amann et al. 1995) on a Zeiss Axioskop 2 microscope. The preparation of the specimens for CARD-FISH analysis was modified from Vandekerckhove et al. (2002) after initial extra cleaning and dissection of the nematodes to improve probe penetration (see details on-line). The nematodes were then fixed on glass slides adding low-gelling-point (0.2%) to avoid specimen loss and CARD-FISH applied following the protocol by Pernthaler et al. (2002). Visualisation of the sample was done using a light microscope (Zeiss Axioskop 2) equipped with epifluorescence and with a Bio-Rad MRC-1024 confocal laser scanning system equipped with a krypton–argon laser (Amann et al. 1995; Vandekerckhove et al. 2002).

For the fingerprinting analysis ARISA, nematode specimens were isolated from bulk sediment that was stored in a frozen state. DNA was extracted from H. disjuncta specimens with the protocol used by Floyd et al. (2002). Before DNA extraction, the nematodes were carefully washed to remove possible epibiontic bacteria and then cut to facilitate DNA extraction. The universal primer 16S-1392F (5′-GYACACACCGCCCGT-3′) and the bacterial primer 23S-125R (5′-GGGTTBCCCCATTCRG-3′) were used (Hewson and Fuhrman 2004). Negative and positive controls were used.

For Beggiatoa and fauna fatty acid composition and carbon isotope signatures, surface sediment collected from both the microbial mat and centre habitat was stored in a frozen state for laboratory isolation of target organisms. Given that FAs constitute a minor fraction of animal organic matter, a large number of these small-sized organisms would have to be isolated and cleared of adhering sediment particles for reliable analysis of meiofauna FA composition and compound specific stable isotope analysis. This can be very time-consuming or virtually impossible, especially in cases of fauna-impoverished deep-sea sediment. These restrictions were overcome by performing one-step small volume extractions coupled with large volume injections (30 μl; GC analyses commonly employ 1–2 μl injections). Here we employed a one-step method according to Masood et al. (2005) that was first tested for our application by comparing with results obtained with a modified large volume multi-step Bligh and Dyer (1959) extraction method of total fatty acids (Boschker and Middelburg 2002). We compared laboratory-cultured bacteria and pelagic diatoms and their blank extractions. The one-step method produced similar or superior FA yields (ratio of total [FA] in multi-step:one-step in diatom and bacteria extraction equalled 1.03 and 0.67, respectively) and in accordance with Masood et al. (2005), with no evidence of polyunsaturated fatty acid (PUFA) loss (e.g. FA 20:5ω3 had an occurrence of 6 and 9% in the multi-and one-step methods, respectively). Additionally, there was very limited contamination in the small volume one-step method because only traces of FA 16:0 and 18:0 equivalent to <1% of FA were found. The addition of the antioxidant (BHT) did not alter PUFA yield and was therefore not further applied.

Three replicates of the meiofauna (160 nematodes and 30 copepods each) and strands of Beggiatoa were handpicked with a needle, cleared of adhering particles and rinsed twice in 0.2 μm filtered Milli-Q before being transferred to 2.5 ml GC vials that were frozen at −80°C and subsequently freeze dried. Fatty acid extraction and preparation of methyl esters (FAME) were carried out according to Masood et al. (2005) with reagent volumes adapted for use in this 2.5 ml GC-vials using FAME C19:0 as internal standard to calculate the concentration of FAs. These individual samples were analysed separately for their FA compositions employing a large volume splitless injection method on a Thermo Finnigan Trace Ultra GC with the following configuration: large volume liner with glass wool, pre-column deactivated silica 5 m × 0.53 μm and analytical column SGE BPX-70 50 m × 0.32 mm × 0.25 μm. After these individual runs, the three replicates of each taxa were combined and evaporated to obtain sufficient material for analysis of FA carbon isotope signatures. The carbon isotopic composition of the dominant FAME was determined with a gas-chromatograph combustion interface isotope-ratio mass spectrometer (GC-c-IRMS; Boschker and Middelburg 2002). The identification of FAMES was based on comparison of retention times with authentic commercially available reference material and standards, confirmed by GC-MS analysis (de Goeij et al. 2008).

Fatty acid source designation was achieved using data on distinctive FAs for bacteria, algal carbon and higher organisms (e.g. Parrish et al. 2000; Boschker and Middelburg 2002; Volkman 2006). Although individual FAs biomarkers are rarely produced by a single class or organism, interpretations can be verified using FAs combinations or spectra (Dalsgaard et al. 2003), even more so when complemented with FA δ13C signatures. Multivariate analyses of FAs composition profiles (non-transformed percentage abundance) were performed using the program PRIMER Version 5 (Plymouth routines in multivariate ecological research). Similarities between profiles were investigated using the SIMPER (similarity of percentages) function and statistical differences were determined using the ANOSIM (analysis of similarities) function.

Results

The FAs composition profiles of each component (Beggiatoa, nematodes and copepods) showed a very high similarity among the three replicates (SIMPER, similarity > 96%). In accordance with FA profiles reported for Beggiatoa, Thioploca and other sulphide-oxidising bacteria (Zhang et al. 2005; and references therein), a few FAs dominated the compositional pattern of the free-living sulphide-oxidising bacteria that typify the microbial white mat sub-habitat of the HMMV cold seep (Fig. 1). The monounsaturated FAs (MUFAs) 16:1ω7 (69%) and 18:1ω7 (21%) and the saturated FA 16:0 (6%) together contributed ~96% of the FAs composition of the Beggiatoa (Fig. 1). The corresponding light-stable carbon isotope signatures (δ13C ~–50‰; Table 1) can be attributed to the chemoautotrophic synthesis of these FAs (Conway and Capuzzo 1991; Pranal et al. 1996; Zhang et al. 2005). Conforming with earlier conclusions (e.g. Zhang et al. 2005), these results indicate that 16:1ω7 and 18:1ω7 can be used as signature biomarkers for sulphide-oxidising bacteria in hydrogen sulphide-rich sediments such as the HMMV mat sub-habitat.

Fig. 1
figure 1

Fatty acid composition of Beggiatoa sp., the microbial mat nematode Halomonhystera disjuncta and the seep centre copepod Tisbe sp. collected at the Håkon Mosby Mud Volcano. Average percentage of total (n = 3) ± 1 SE

Full size image
Table 1 δ13Corg (‰) of the dominant fatty acids and bulk tissue
Full size table

Although the copepod from the seep centre also had a strong contribution of bacteria-synthesised FAs (Fig. 1), the FAs profiles were significantly different among the Beggiatoa, nematode and copepod (ANOSIM, global R = 1, P < 0.05). The difference between Beggiatoa and the nematode were caused by a higher percentage abundance of 16:1ω7 in Beggiatoa and relatively higher percentage contributions of 18:1ω7 and 18:0 in the nematode. The difference between the nematode and copepod is not only an overall more diverse set of FAs, but also the presence and dominance of ω3 long-chain PUFAs in the copepod (Fig. 1). More specifically, SIMPER analysis revealed that the primary differences are the relatively higher percentage abundance of 18:1ω7 and 16:1ω7 in the nematode (39 and 38% versus 14 and 13%, respectively) compared to 16:1ω8, 22:6ω3 and 20:5ω3 in the copepod (17, 9 and 7% versus 2, 0.7 and 0%, respectively; Fig. 1).

TEM sections revealed well-developed intestines and ovaries in this nematode and clearly no evidence of symbiontic cells (Fig. 2). The conclusion that bacteria are used as a food source and not hosted as symbionts is supported by the lack of bacteria in the nematode tissues as revealed either by FISH, CARD-FISH or ARISA (data not shown).

Fig. 2
figure 2

Transmission electron microscopy photograph of the cross-section of a female nematode Halomonhystera disjuncta. As in all cross-sections, there was no evidence of either ecto- or endosymbiotic bacteria. Indication on photograph of: a the intestine; b the ovary

Full size image

Discussion

The successful meiobenthic species in this sub-habitat, the nematode H. disjuncta, was characterised by a FAs profile indicative of a bacterial diet (Fig. 1): strong dominance of light δ13C MUFAs (16:1ω7 and 18:1ω7 together account for 80% of FAs carbon with a corresponding δ13C of –54‰; Table 1) and absence of ω3 and ω6 long-chain polyunsaturated FAs (PUFAs, commonly originating from photosynthetic plankton). Although MUFAs such as 16:1ω7 and 18:1ω7 can be derived from diatoms (e.g. Dijkman and Kromkamp 2006; Volkman 2006), their light δ13C values confirmed utilisation of microbial-mediated chemosynthetically derived carbon. As discussed above, this suggests that sulphide-oxidising bacteria constituted the main diet of H. disjuncta.

Other observations of nematodes feeding on Beggiatoa bacteria were previously recorded for Diplogaster bernenesis (Sudhaus and Rehfeld 1990), a well known species from freshwater, and Diplolaimelloides deconincki in a harbour site in Denmark (Bernhard and Fenchel 1995). In the latter study, a large number of aerobic and anaerobic ciliate species were found to ingest filamentous sulphur bacteria. Another close relationship between nematodes and Beggiatoa was described in wastewater treatment processes (Salvadó et al. 2004). It was suggested that nematodes play an important role in wastewater bio-oxidation by their grazing activity on the bacterial community (Woombs and Laybourn-Parry 1986). The most significant biotic relationship found by Salvadó et al. (2004) was between nematodes and filamentous bacteria (specifically the genus Beggiatoa), and to a lesser extent, nematodes with ciliates. Based on the biofilm thickness, they concluded that the nematodes also could have a positive influence on the microbial communities. The mechanical effect of nematodes on the biofilm structure would allow a better oxygen penetration and nutrient diffusion into the biofilm, thus, stimulating its growth. Beggiatoa appeared also to be attached to rock samples retrieved from a hydrothermal vent field located in the Azores archipelago (North Altantic; Cardigos et al. 2005). Two species, the nematode Enoplus meridonalis and the oligochaete Lumbricillus sp. were abundant in the thick Beggiatoa mats, although no direct trophic relationship with the bacteria was shown (Cardigos et al. 2005). It can be concluded that nematodes are often associated with Beggiatoa bacteria in sulphidic conditions in freshwater as well as marine habitats. Their close relationship may be due to environmental factors or trophic factors. At the HMMV, the high biomass and standing stock of H. disjuncta can be attributed to feeding on the free-living sulphide-oxidising bacteria, a trophic specialisation previously not seen for any other deep-sea meiofaunal seep species (Van Gaever et al. 2006).

Although photosynthetic plankton is normally the predominant source of 22:6ω3 and 20:5ω3 in the marine environment, long-chain PUFAs can also be synthesised by some bacteria (e.g. Delong and Yayanos 1986; Nichols 2003; Valentine and Valentine 2004; Lösekann et al. 2008). This assertion of bacterial origin is supported by the extremely light δ13C of all dominant FAs including the PUFAs normally attributed to phytoplankton (mean δ13C < −69‰; Table 1). Methane in HMMV has a δ13C of ~−60‰ (Niemann et al. 2006). This clearly indicates a greater dependence on methane in the seep centre sediments as a result of prevailing higher methane fluxes compared to the microbial mat sediments (Niemann et al. 2006). The predominance of methanotrophic input is further supported by the dominance of the methanotroph type I specific bacterial biomarker 16:1ω8c (e.g. Pranal et al. 1996; Kiyashko et al. 2004, see Fig. 1) with very light carbon isotope signatures (–81‰ δ13C; Table 1). In corroboration, this FA was also found to be dominant in the seep centre surface sediment and with a very light carbon isotope signature (δ13C –80‰; Niemann et al. 2006). Large difference in the nematode bulk tissue isotope values measured in 2003 (Van Gaever et al. 2006) and in 2006 (this study; Table 1) indicate that heterogeneity can be high and confirm strong temporal variation in HMMV fauna carbon isotope signatures (Lösekann et al. 2008). Unfortunately, copepod bulk tissue was measured only in 2003 (Table 1), but could have similarly been much lighter in 2006. This may explain the seemingly larger difference between FA and bulk tissue isotope signature compared to nematodes, otherwise generally in the order of 3–5‰ (e.g. Boschker and Middelburg 2002).

Overall, the results indicate that both meiobenthic species proliferate on chemosynthetically derived carbon, but also that their segregation is the result of their spatial and trophic specialisation. The high biomass and abundance reached by the nematode thriving on sulphide-oxidising bacteria in the microbial mat sub-habitat suggest a certain level of trophic preference, which is different from the one found for the seep centre copepod, which fed on methanotrophic biomass. Different microbial activities are associated with differences related to habitat-specific characteristics. For example, in the seep centre, rapidly rising sulphate-free fluids prevent sulphide formation and thus oxidation, but support methanotrophic habitats, whereas high sulphide concentrations in the microbial mat habitat are translated into elevated chemosynthetically derived microbial biomass (de Beer et al. 2006; Niemann et al. 2006). The utilisation of chemosynthetic food sources for meiofauna may vary among sub-habitats; this appears to be of major importance for clarifying the drivers of diversity distribution, community structure and ecosystem functioning in chemoautotrophic systems. Thus, the observed distribution patterns of the dominant meiofauna taxa (Van Gaever et al. 2006) may be explained by both habitat and trophic specialisation.