OUP user menu

Methane assimilation and trophic interactions with marine Methylomicrobium in deep-water coral reef sediment off the coast of Norway

Sigmund Jensen, Josh D. Neufeld, Nils-Kåre Birkeland, Martin Hovland, John Colin Murrell
DOI: http://dx.doi.org/10.1111/j.1574-6941.2008.00575.x 320-330 First published online: 1 November 2008

Abstract

Deep-water coral reefs are seafloor environments with diverse biological communities surrounded by cold permanent darkness. Sources of energy and carbon for the nourishment of these reefs are presently unclear. We investigated one aspect of the food web using DNA stable-isotope probing (DNA-SIP). Sediment from beneath a Lophelia pertusa reef off the coast of Norway was incubated until assimilation of 5 μmol 13CH4 g−1 wet weight occurred. Extracted DNA was separated into ‘light’ and ‘heavy’ fractions for analysis of labelling. Bacterial community fingerprinting of PCR-amplified 16S rRNA gene fragments revealed two predominant 13C-specific bands. Sequencing of these bands indicated that carbon from 13CH4 had been assimilated by a Methylomicrobium and an uncultivated member of the Gammaproteobacteria. Cloning and sequencing of 16S rRNA genes from the heavy DNA, in addition to genes encoding particulate methane monooxygenase and methanol dehydrogenase, all linked Methylomicrobium with methane metabolism. Putative cross-feeders were affiliated with Methylophaga (Gammaproteobacteria), Hyphomicrobium (Alphaproteobacteria) and previously unrecognized methylotrophs of the Gammaproteobacteria, Alphaproteobacteria, Deferribacteres and Bacteroidetes. This first marine methane SIP study provides evidence for the presence of methylotrophs that participate in sediment food webs associated with deep-water coral reefs.

Keywords
  • stable isotope probing
  • methane
  • trophic level
  • phylogeny
  • deep-water coral reef
  • Methylomicrobium spp

Introduction

Deep-water corals build reefs below the photic zone of fjords and continental shelves around the world. These reefs attract fish and benthic animals as they represent oases of life on the sea floor (Fosså, 2002; Roberts et al., 2006; Hovland, 2008). Recent molecular biological studies of deep-water coral-associated microbial communities have revealed the presence of a wide diversity of Proteobacteria (Alphaproteobacteria, Gammaproteobacteria and Deltaproteobacteria), Acidobacteria, Nitrospira, Firmicutes, Bacteroidetes and additionally Planctomycetes, Verrucomicrobia, Actinobacteria and Crenarchaea in ambient reef water and sediment (Penn et al., 2006; Yakimov et al., 2006). Little is known about the activities and the trophic interactions of these microorganisms, and their relevance for the well being of deep-water coral reefs is presently unclear. Findings indicate that particle transport (food) via currents can optimize deep-water coral reef nutrition (Thiem et al., 2006). Deep-water coral reefs are sited in cold permanent darkness, in stark contrast to the sunlit tropical coral reefs, which are directly supported by fresh nutrients provided on a daily basis by the photosynthetic algae zooxanthella.

Replacing light-driven primary production, chemoautotrophic microorganisms have been hypothesized to stimulate life in deep-water coral reefs by utilizing micro-seepage and pockmark fluid flow (Hovland & Risk, 2003; Roberts et al., 2006). The methane content of the open ocean seawater can be as low as c. 1 nM, but sediments and seeps have elevated levels of methane (micro- to millimolar; Reeburgh, 2007). Methane, together with hydrogen sulphide, is a major constituent of marine seepage. The anaerobic oxidation of methane in the seabed (Boetius et al., 2000) produces hydrogen sulphide, which together with any remaining methane may seep into the oxic seafloor environment. Aerobic methane-oxidizing bacteria can consume this methane (Hanson & Hanson, 1996) and the application of molecular biology techniques have facilitated the specific detection and quantification of these methanotrophs in environmental samples (Dumont & Murrell, 2005a; McDonald et al., 2008).

All extant methanotrophic bacteria belong to the Proteobacteria and include lineages of the gamma (type I) and alpha (type II) subdivisions, except for the recently discovered lineage of methanotophic Verrucomicrobia (Dunfield et al., 2007; Pol et al., 2007; Islam et al., 2008). Marine methanotrophs contribute to the attenuation of methane flux to the atmosphere and have been found to be associated with estuaries (McDonald et al., 2005), the open ocean (Sieburth et al., 1987; Holmes et al., 1995; Hayashi et al., 2007), seeps (Inagaki et al., 2004; Tavormina et al., 2008), gas hydrates (Yan et al., 2006), hydrothermal vents (Elsaied et al., 2004; Nercessian et al., 2005a) and mud volcanoes (Losekann et al., 2007). Methanotrophs at hydrothermal vents and seeps can supply invertebrates with methane-derived carbon and energy either via grazing (Deangelis et al., 1991) or via endosymbiosis (Stewart et al., 2005). Methanotrophs may also contribute to primary fixed carbon in the ground-water ecosystem (Hutchens et al., 2004), peat bogs (Raghoebarsing et al., 2005) and rice soil (Murase & Frenzel, 2007).

The specific detection of methanotrophs is enabled by PCR amplification of pmoA and mmoX genes, which encode subunits of the two known variants of methane monooxygenase (MMO) in aerobic methanotrophs. The phylogeny of the pmoA gene is largely congruent with the universal prokaryotic marker 16S rRNA gene (Holmes et al., 1995; Kolb et al., 2003). A broader range of methylotrophs can be targeted via mxaF, the gene which encode the large subunit of methanol dehydrogenase that catalyses the second step in aerobic methane oxidation. Recent methodological advances have enabled a distinction to be made between the detection of methanotrophs and methylotrophs that may be present in the environment but not necessarily active, and those that are actively involved in substrate metabolism. The advent of stable-isotope probing (SIP; Radajewski et al., 2000) allows for a culture-independent link to be made between taxonomy and function and has been applied to the study of active C1-consuming microorganisms in a variety of terrestrial and aquatic environments (Dumont & Murrell, 2005b).

In the present study, methane was explored as a carbon source for fuelling bacterial growth in the surface sediment beneath a deep-water coral reef at Haltenpipe off the coast of Norway. The objective was to use DNA-SIP by adding 13C-labelled methane, to follow substrate depletion, and to retrieve and characterize 13C-labelled DNA. These first marine methane SIP results, involving fingerprinting and multiple clone library analyses of ‘heavy’ DNA, all indicated a role for the genus Methylomicrobium (Bowman et al., 1995) in the assimilation of methane-carbon in this deep-water coral reef sediment.

Materials and methods

Reef sediment sampling

A sediment sample was collected in November 2006 during a remotely operated vehicle (ROV) dive to coral reef ‘E’, c. 50 km off the mid-Norwegian coastline (Hovland, 2008). The site is located c. 600 m due south of the ‘Haltenpipe Reef Cluster’ (63°54′42″N, 07°52′30″E). At 280 m depth, the ROV scooped sediment (upper 10 cm) and interstitial water into a 6-cm-diameter tube, which was sealed and brought back to the surface. During transport to the laboratory, the sediment was stored at 5 °C. Sediment pH was measured directly in the slurry and organic matter content was estimated by loss on ignition after heating the dry sample to 550 °C for 5 h.

Experimental strategy, activity and 13CH4 labelling

Within 2 days of sampling, 10-g wet weight (ww) sediment aliquots were added to autoclaved 120-mL serum flasks and mixed with 10 mL autoclaved and 0.2-μm-filtered seawater containing 20% (v/v) distilled water to minimize precipitation. Flasks were capped with red butyl rubber stoppers and sealed with aluminium crimp caps. Twenty millilitres of 99 atom%13CH4 (Isotec) was added by syringe to each flask. Flasks were incubated in the dark at 10 °C on a reciprocal shaker operated at 110 r.p.m. Triplicate flasks with added 2 mL of 12CH4 (to maintain sensitivity) were incubated alongside and monitored repeatedly for headspace methane disappearance as described previously (Jensen & Olsen, 1998). Briefly, headspace samples were analysed with a thermal conductivity detector (150 °C) in a Hewlett Packard HP 6890 gas chromatograph equipped with a 6 FT Haysep R column (80/100 mesh, oven temperature 35 °C) and with He as the carrier gas (flow rate 60 mL min−1). Following 13CH4 incubation, flasks were removed at selected time points and frozen at −70 °C before Fast Spin DNA extraction according to the manufacturer's protocol (Qbiogene).

Preparation of DNA-SIP gradients

Approximately 5 μg of DNA was added to caesium chloride (CsCl) solutions for isopycnic ultracentrifugation, gradient fractionation and DNA recovery as described previously (Neufeld et al., 2007b). Briefly, DNA was mixed with CsCl to a final density of 1.725 g mL−1 and centrifuged at 177 000 g in a Vti 65.2 rotor (Beckman) at 20 °C for 40 h. Gradients were fractionated into twelve 425-μL fractions and DNA was precipitated with glycogen and a polyethylene glycol solution. Final DNA pellets were suspended in 20 μL of nuclease-free water and aliquots were run on a 1% (w/v) agarose gel.

Bacterial community fingerprinting using PCR-denaturing gradient gel electrophoresis (DGGE)

A variable region (V3) of the 16S rRNA gene was PCR-amplified with the ‘universal’ bacterial primer set 318fGC/518r (Muyzer et al., 1993) and primer sets designed specifically for type I (IF/IR) and type II (IIF/IIR) methanotrophs (Chen et al., 2007) to produce amplicons for DGGE analysis. PCR amplifications were performed in 0.5-mL microcentrifuge tubes with the reagents supplied with Taq polymerase (Fermentas) at a Mg2+ concentration of 1.5 mM, 1–100 ng of template, 0.5 μM of each primer, 0.2 mM each dNTP and bovine serum albumin supplied at a concentration of 20 μg mL−1. Reactions were carried out in a Tetrad thermal cycler (BioRad) with an initial denaturation step of 95 °C for 5 min; 30 cycles of 94 °C for 1 min, annealing at 55 °C (318fGC/518r) or 60 °C (IF/IR, IIF/IIR) for 1 min and 72 °C for 1 min; followed by a final extension of 7 min at 72 °C (Muyzer et al., 1993; Chen et al., 2007). DGGE was performed in 10% (w/v) acrylamide with 30–70% denaturant (V3) and in 8% (w/v) acrylamide with 40–70% denaturant (IF/IR, IIF/IIR), using a DCode apparatus (BioRad) and the protocol of Leckie (2004). The gels (0.75 mm) were run for 14 h at 60 °C and 85 V, stained with SYBR Green (Invitrogen) and scanned with an FLA-5000 scanner (Fujifilm). Bands of interest were excised, eluted in water overnight, reamplified using PCR, checked again for correct electrophoretic mobility and purity, and sequenced. Sequencing was performed using a selected PCR primer and the Applied Biosystems chemistry (BigDye v3.1) and sequencers (373A and ABI 3700).

Detection of active populations using PCR and cloning

Previous results with this SIP protocol demonstrated that labelled 13C-DNA is associated with fractions 7 and 8 (c. 1.725 g mL−1; Neufeld et al., 2007a, b). Because fraction 7 was the heaviest fraction to contain sufficiently labelled DNA for agarose gel visualization, this fraction was chosen for subsequent PCR-based analyses. Aliquots of fraction 7 were PCR-amplified using the 16S rRNA gene primer set 27f/1492r (Lane, 1991), the particulate MMO pmoA gene primer set A189f/mb661r (Dumont & Murrell, 2005a), the soluble MMO mmoX gene primer set 206f/886r (Dumont & Murrell, 2005a) and the methanol dehydrogenase mxaF gene primer set 1003f/1555r (Neufeld et al., 2007a). The PCR was performed as described above with annealing conditions for mxaF of 55 °C for 1 min, for pmoA of 55 °C for 1.5 min and for the mmoX and 16S rRNA genes of 60 °C for 1 min (Dumont & Murrell, 2005a; Neufeld et al., 2007a). Methylococcus capsulatus Bath and Methylosinus trichosporium OB3b DNA served as positive controls (30 ng). Reaction products were quantified and examined for reaction specificity in 2% (w/v) agarose gels. PCR fragments were ligated into the TOPO vector pCR 2.1 according to the manufacturer's protocol (Invitrogen). Amplicons of 16S rRNA and pmoA genes were ligated directly; the mxaF gene had to be reamplified (60 PCR cycles in total) before excising the c. 500-bp band, purifying the amplicons with a QIAquick column (Qiagen) and ligation into the TOPO vector. Transformation was into chemically competent Escherichia coli hosts maintained on Luria–Bertani plates with kanamycin (25 μg mL−1) and X-gal (0.8 mg per plate). Correctly sized M13 amplicons were digested overnight at 37 °C using the endonucleases EcoRI and RsaI. Restriction patterns were categorized into operational taxonomic units (OTUs) and representative OTUs were sequenced. Nearest relatives of the sequences were identified using blastn and blastx (Altschul et al., 1990). The blastx algorithm also served to translate the protein encoding gene fragments. Phylotype abundance was referred to as the library fraction (representative clones/clones in library) × 100%.

Phylogenetic analysis

Sequences of the 16S rRNA genes were inspected for chimeras using the chimera check program (http://rdp8.cme.msu.edu/cgis/chimera.cgi). Sequences of the pmoA and mxaF were inspected for chimeras by searching for large regions of unexpected nucleotide changes when compared with sequences from reference organisms (Radajewski et al., 2002). No potential chimeras were detected. Alignments were performed in arb (http://www.arb-home.de) using the automatic alignment tool. The 16S rRNA gene sequences were corrected manually according to the secondary structure of the 16S rRNA molecule, while the PmoA- and MxaF-derived peptide sequences were manually verified. Filters were generated that omitted alignment positions of sequence ambiguity (N for DNA and X for proteins) and for instances in which sequence data were not available for all sequences. Phylogenetic relationships were calculated in ARB using slightly modified default parameters (see figure legends). The topology resulting from the maximum-likelihood, neighbour-joining and maximum-parsimony treeing methods was compared and the confidence of branch points was determined using a strict consensus rule applied to the maximum-likelihood trees (Radajewski et al., 2002).

GenBank accession numbers of the reported sequences

Gene sequences recovered and analysed from this Haltenpipe reef have been deposited in GenBank under accession numbers EU595514EU595515 (pmoA), EU595516,EU595519EU595520 (mxaF), EU595522EU595538 (16S rRNA gene) and EU595539EU595543 (16S rRNA gene of DGGE bands 1–5).

Results and discussion

Characteristics of site, sample and DNA recovery

Observations performed during the dive revealed a deep-water coral reef dominated by the coral Lophelia pertusa and the bivalve Acesta excavata in an offshore environment at 280 m depth, 5–8 °C and a current of <50 cm s−1. The collected sediment had a fine-grained and grey clay-like texture, a pH of 6.7 and an organic matter content of 5.8% g−1 dw. Triplicate incubations of 2%12CH4 (v/v in headspace) and sediment provided surrogate methane assimilation rates for the 13CH4-incubated flasks of c. 5 μmol 13CH4 assimilated g−1 ww sediment within 34 days. A direct headspace measurement of the 13CH4-incubated sample was taken 1 day before DNA extraction (day 48) and this confirmed active methane assimilation, with an estimated consumption of 13 μmol 13CH4 g−1 ww. Nucleic acid extraction from 1.5 g ww of the sediment after 49 days of incubation yielded c. 7.5 μg of high-molecular-mass DNA. This indicates the presence of c. 3.0 × 109 cells g−1 ww using a predicted 1.6–2.4 fg DNA per cell (Bakken & Olsen, 1989). Following the SIP incubation, extracted DNA was added to a CsCl solution and subjected to ultracentrifugation to separate labelled 13C-DNA from unlabelled background 12C-DNA. Agarose gel analysis of aliquots of the fractionated gradient demonstrated that the 13C-DNA had been eluted as a smear below the predominant original and native 12C-DNA (Fig. 1).

1

Agarose gel electrophoresis of aliquots of fractionated ultracentrifuge gradients from the sediment SIP incubation. The fractions are ordered from the bottom (Fraction 1, ‘heavy’) to the top (Fraction 12, ‘light’). DNA with a buoyant density of fraction 7 (c. 1.725 g mL−1) and below is expected to be heavily enriched with 13C. No DNA is visibly discernible below fraction 7 because any DNA at this high density was at a concentration below the detection limit of our agarose gel electrophoresis conditions. Bands shown for the 1-kb marker (M) range from 12216 to 1018 bp.

DGGE profiling of native and 13C-labelled DNA fractions

DGGE profiling of 16S rRNA gene fragments PCR amplified from the native sample indicated high microbial diversity in this deep-water coral reef sediment (Fig. 2). Repeated DGGE profiling during SIP labelling of microorganisms in this sediment revealed the emergence of three distinct DGGE bands after 13 days of 13CH4 exposure (Fig. 2). During the following 36 days of labelling, one of these bands was maintained to be accompanied by another distinct DGGE band by day 49 (Fig. 2). The SIP methodology may cause perturbations to microbial communities and result in the rapid labelling of initially inactive community members (Hutchens et al., 2004; Cébron et al., 2007). Prolonged incubation was, however, required for sufficient 13C labelling of this sediment community. DGGE profiling of 16S rRNA gene fragments PCR amplified from the SIP labelled (49 days) and fractionated DNA (Fig. 3) confirmed the high native diversity (fractions 10 and 11) and the shift, which was visible as two predominant phylotypes associated with methane uptake (fractions 5–8). Sequence analysis of DGGE bands from fraction 7 identified Methylomicrobium ML1 and clone 3B with c. 93% homology to 16S rRNA genes of Methylophaga (Fig. 3). The ML1 bacterium was also represented in two of the three DGGE bands visible in the heavy fraction IF/IR profile generated with PCR directed towards type I methanotrophs (Chen et al., 2007), while the third band belonged to a Methylobacter-related clone, HM19 (data not shown). Sequencing of major bands from the heavy fraction IIF/IIR profile generated with PCR directed towards type II methanotrophs (Chen et al., 2007) detected marine Alphaproteobacteria with no affiliation to extant methanotrophs (data not shown).

2

DGGE of PCR-amplified 16S rRNA gene fragments from the SIP incubated deep-water coral reef sediment prior to fractionation of DNA, at time points: 0 days (lane 1), 3 days (lane 2), 13 days (lane 3) and 49 days (lane 4). Sequenced bands were affiliated with an uncultured gammaproteobacterium clone 3B at 97% identity (○), an uncultured Piscirickettsiaceae clone DS021 at 97% identity (◻) and a Bacteriovorax sp. NZAH13 at 90% identity (◊). The marker (M) is a ladder prepared from cloned bands from community fingerprints.

3

DGGE of PCR-amplified 16S rRNA gene fragments from DNA of SIP fractions 5–11 of the deep-water coral reef sediment. Sequenced bands were affiliated with an uncultured gammaproteobacterium clone 3B at 97% identity (○) and Methylomicrobium sp. ML1 at 99% identity (●). The marker (M) is a ladder prepared from cloned bands from community fingerprints.

Clone library analysis of the 13C-labelled DNA

Further PCR amplification and cloning of genes diagnostic for bacterial taxonomy (16S rRNA, 71 clones), methane uptake (pmoA; 42 clones) and methane processing (mxaF; 48 clones) from 13C-labelled DNA (fraction 7) resulted in a large portion of 16S rRNA gene clones (52%), mxaF clones (85%) and pmoA clones (100%) linked to a marine Methylomicrobium japanense strain NI (Kalyuzhnaya et al., 2008). Methylomicrobium japanense was isolated from marine mud off the Japanese coast off Hiroshima and provisionally named Methylomicrobium sp. strain NI by Fuse (1998). In the 16S rRNA gene clone library (Table 1), strain NI is represented by two OTUs, the major phylotype hpA02ssu (36 clones) and a nearly identical hpF10ssu (one clone). The 16S rRNA gene of the Methylomicrobium identified in reef sediment and the NI strain are >99.5% identical and the genes share significant identity with the Methylomicrobium pelagicum isolated from the Sargasso Sea (Sieburth et al., 1987; Bowman et al., 1995). Methylomicrobium pelagicum has been lost from culture collections (Kalyuzhnaya et al., 2008) and no other cultivated methanotrophs were found to be closely related to these marine Methylomicrobium representatives. The other cultured relatives are halophilic isolates from surface sediments of soda lakes in Russia and Kenya (Fig. 4). Included herein are three unpublished isolates (FM3, E3, ML1) and the three recognized species Methylomicrobium buryatense (Kaluzhnaya et al., 2001), Methylomicrobium kenyense and Methylomicrobium alcaliphilum (Kalyuzhnaya et al., 2008). Most relatives of the reef Methylomicrobium are, however, uncultivated and were observed in experiments with Russian Transbaikal soda lakes (Lin et al., 2004) (Fig. 4). The few sequences currently available for marine methanotrophs may provide an explanation for the relatively frequent affiliation with soda lake methanotrophs. Interestingly, 16S rRNA gene sequences from the nonsaline soda lake Suduntuiskii Torom (Lin et al., 2004) were more distantly related (Fig. 4). The reef-associated Methylomicrobium is also phylogenetically distinct from Methylomicrobium spp. affiliated with sediment clones from the brackish Newport Bay Estuary (McDonald et al., 2005) and the freshwater Lake Washington (Nercessian et al., 2005b), which clustered with Methylomicrobium album and Methylomicrobium agile (Fig. 4; Kalyuzhnaya et al., 2008).

View this table:
1

Categorization of 16S rRNA, mxaF and pmoA genes from DNA-SIP fraction 7 containing 13C-DNA of the deep-water coral reef sediment

PhylotypeDivision (subdivison)Nearest taxon (% identity) origin and accession numberPhysiology/enzymeAbundance
hpA02ssu●′′Proteobacteria (gamma)Methylomicrobium sp. NI (99%) marine sediment D89279Methanotroph37 clones
hpA06ssu○′′Proteobacteria (gamma)Uncultured (96%) marine sediment AY375062Methylotroph4 clones
hpE10ssu○Proteobacteria (gamma)Uncultured (93%) marine plankton DQ009128Methylotroph1 clone
hpC04ssuProteobacteria (gamma)Uncultured (97%) ridge flank crustal fluid DQ513017Heterotroph2 clones
hpD12ssuProteobacteria (alpha)Thalassospira sp. MACL12B (99%) saline lake EF198251Heterotroph4 clones
hpF08ssuProteobacteria (alpha)Uncultured (97%) Haliclona simulansEU350873Methylotroph2 clones
hpD01ssuProteobacteria (alpha)Uncultured (90%) marine plankton DQ300657Heterotroph1 clone
hpA12ssuProteobacteria (alpha)Uncultured (93%) anoxic lake sediment AM086142Heterotroph1 clone
hpA01ssuBacteroidetes (Flavobacteria)Uncultured (96%) marine biofilm EF215723Heterotroph?7 clones
hpA05ssuBacteroidetes (Flavobacteria)Uncultured (93%) marine biofilm EF491373Heterotroph3 clones
hpA09ssuBacteroidetes (Flavobacteria)Uncultured (94%) mangrove soil DQ811912Heterotroph3 clones
hpE12ssuBacteroidetes (Flavobacteria)Uncultured (95%) marine sediment DQ395037Heterotroph2 clones
hpE11ssuBacteroidetes (Flavobacteria)Uncultured (94%) marine sediment AB015545Heterotroph1 clone
hpE03ssuBacteroidetes (unclear)Uncultured (91%) mangrove sediment EF061974Unclear physiology1 clone
hpF11ssuDeferribacteresUncultured (98%) marine sediment EF459950Versatile anaerobe2 clones
hpG01mxaFProteobacteria (gamma)Methylomicrobium sp. NI (99%) marine sediment -MDH41 clones
hpG02mxaFProteobacteria (alpha)Fulvimarina pelagi HTCC2506 (87%) marine EAU40815PQQ dehydrogenase1 clone
hpA05mxaFProteobacteria (alpha)Hyphomicrobium sp. P 768 (97%) Y08082MDH2 clones
hpD02pmoA′′Proteobacteria (gamma)Uncultured (100%) soda lake AY236081pMMO42 clones
  • Phylotypes marked with ○/● match the dominant DGGE bands of this fraction and ′′ represent two OTUs with >99% sequence identity. Reef sediment clone nomenclature includes site, prefix and gene.

4

Maximum-likelihood tree of 16S rRNA genes detected in 13C-DNA fraction 7. The reef sediment clones are highlighted and identified by site, prefix, gene and abundance (in parentheses). Clones marked with ○/● match the dominant DGGE bands in Fig. 3. The tree was constructed with a filter of 570 aligned nucleotide positions, excluding ambiguities and missing data. Branch points have been collapsed to support all treeing methods used (maximum-likelihood with arb and axml, neighbour-joining with arb and the Jukes–Cantor model, maximum-parsimony with arb and dnapars). Aligned partial sequence AY831464 was inserted into each tree using a parsimony tool and the same filter in arb. The 16S rRNA gene of Archaeoglobus fulgidus was used as the outgroup. The scale bar represents 10% sequence divergence.

To our knowledge, these 16S rRNA pmoA and mxaF genes represent the first methane SIP sequences recovered from marine sediment. The deep-water coral reef sediment pmoA library (Table 1) contained only one phylotype (hpD02pmoA), affiliated with isolates and clones from the Transbaikal soda lakes (Kaluzhnaya et al., 2001; Lin et al., 2004) and the Japanese Methylomicrobium NI strain (Fuse et al., 1998). The D02 sequence also affiliated with pmoA sequences from sediment of alkaline and hypersaline Mono Lake in California (Lin et al., 2005) (Fig. 5). The mxaF gene library (Table 1) contained three authenticated phylotypes. Clone hpG01mxaF was numerically dominant and affiliated with mxaF of the Japanese NI strain (unpublished) and a soil Methylobacter 5FB (Knief & Dunfield, 2005). Also, affiliated with clone G01 (Fig. 5) were uncultivated methanotrophs associated with roots of Italian rice plants (Horz et al., 2001), Chinese alkaline coal mines, Romanian Movile cave clones (Hutchens et al., 2004), Irish landfill cover soil clones (Chen et al., 2007) and a Japanese methane seep (Inagaki et al., 2004). Comparing these results with the methane DNA-SIP study of Lin (2004), in which many active Methylomicrobium- and Methylobacter-related methanotrophs were detected in samples from the Lake Gorbunka sediment, suggests that type I methanotrophs may also be responsible for the majority of methane oxidized in the deep-water coral reef sediment at Haltenpipe. The contribution of previously detected marine methanotrophs cannot, however, be ruled out (Holmes et al., 1995; Elsaied et al., 2004; Inagaki et al., 2004; Nercessian et al., 2005a; Yan et al., 2006; Hayashi et al., 2007; Losekann et al., 2007; Tavormina et al., 2008). The high methane concentration and the long incubation time used in this experiment may have ‘masked’ the activity of additional uncultivated methanotrophs in this SIP incubation. Subsequent research will ideally confirm the activity and predominance of Methylomicrobium spp. at Haltenpipe and other deep-water coral reefs. We were unable to detect mmoX gene sequences in this reef sediment, although Methylomicrobium strain NI has been shown to contain sMMO (Fuse et al., 1998). Heavy DNA from the SIP experiment did not yield a mmoX PCR product, even after a reamplification of the amplification products from an initial PCR. Others have also failed to detect mmoX genes in marine samples (Holmes et al., 1995; Inagaki et al., 2004) and it is known that most type I methanotrophs lack the sMMO enzyme (Hanson & Hanson, 1996). Of the 16 marine methanotrophs isolated by Fuse (1998) only Methylomicrobium strain NI screened positive for sMMO.

5

Maximum-likelihood trees of derived PmoA (a) and MxaF (b) sequences detected in 13C-DNA fraction 7. The reef sediment clones are highlighted and identified by site, prefix, gene and abundance (in parentheses). The trees were constructed with a filter of 121 (PmoA) and 178 (MxaF) aligned amino acid positions, excluding ambiguities and missing data. Branch points have been collapsed to support all treeing methods used (maximum-likelihood with arb protein_ml and the pam model, neighbour-joining with arb and the pam model, maximum-parsimony with arb and protpars). Amino acids of the homologous subunit of ammonium monooxygenase from Nitrosococcus oceani (a) and ethanol dehydrogenase from Pseudomonas aeruginosa (b) were used as outgroups. The scale bar represents 10% sequence divergence.

Evidence for a methane-driven food web

The remaining 16S rRNA gene clones (48%) and mxaF clones (15%) were not affiliated with bacteria recognized as methanotrophs, implying contamination from 12C-DNA or cross-feeding of metabolites associated with 13CH4 oxidation (Table 1, Fig. 3). Contamination by 12C-DNA was evident based on DGGE fingerprints (Fig. 3) although known methylotrophs (five clones) were represented by Methylophaga-affiliated sequences detected by 16S rRNA genes in DGGE (Figs 2 and 3) and the clone library (Table 1). Methylophaga and related uncultivated Gammaproteobacteria may be consuming methane-derived carbon via methanol. Evidence for this includes marine DNA-SIP incubations with methanol and methylamine; both incubations resulted in the identification of Methylophaga as active methylotrophs in a marine UK surface water sample (Neufeld et al., 2007a). Apart from clones 24-7_D07/D10/E04 (Neufeld et al., 2007a), representing uncultivated Gammaproteobacteria spp., the closest cultured relative of the Methylophaga-like sequences discovered in this SIP study appears to be the dimethylsulphide-consuming DMS048 bacterium isolated by Schäfer (2007) from a UK tidal rock pool at Coral Beach (Fig. 4). The finding of 16S rRNA genes affiliated with Colwellia and Thalassospira was unexpected (six clones). Likewise, it is difficult to explain the existence of 13C-assimilating Alphaproteobacteria (four clones) and the Deferribacteres (two clones), unaffiliated with extant methylotrophs. If not due to contaminating 12C-DNA, these 12 sequences suggest the presence of novel methylotrophs, cross-feeding and involvement of a food chain, reflecting the findings of Murase & Frenzel (2007) with rice soil. As previously suggested, aerobic methanotrophs may actively convert CH4 into complex organic compounds and help to sustain the microbial diversity of the system (Hutchens et al., 2004). Surprisingly, the present study detected no mxaF gene sequences of the Methylophaga. Facultative C1 consumers of the mxaF library were phylotype hpA05mxaF (two clones) affiliated with a peat soil clone LOM13.7 (Morris et al., 2002) and a Hyphomicrobium strain P768 (Fesefeldt & Gliesche, 1997), and the phylotype hpG02mxaF (one clone) affiliated with Fulvimarina pelagi (Cho & Giovannoni, 2003), Sinorhizobium and Methylobacterium spp. (Table 1, Fig. 5). These Alphaproteobacteria are facultative methylotrophs, growing on methanol and a wide range of sugars, organic acids and amino acids. The last four mxaF clones were false positives and their derived amino acid sequences were affiliated with enzymes of Planctomycetes (arylsulphatase), Oligotropha (peptidase) and a serine protease.

A significant proportion of the heavy DNA clones were related to previously unrecognized methylotrophs, and as many as 17 of these 16S rRNA gene sequences belonged to the phylum Bacteroidetes. Diverse communities of Bacteroidetes are common in marine sediments that degrade complex organic materials (Musat et al., 2006), and Bacteriodetes were detected in a 16S rRNA clone library (<90% identical sequences) prepared from native deep-water coral reef sediment (Jensen et al., 2008). The seven Bacteriodetes-affiliated sequences represented by the phylotype hpA01ssu is the largest nonmethanotrophic clade in this study. More importantly, Bacteroidetes has been detected in methane SIP experiments with forest soil, field soil and peat soil (Morris et al., 2002; Radajewski et al., 2002; Cébron et al., 2007; Héry et al., 2008). Morris (2002) and Héry (2008) implicated Bacteroidetes involved with C1-cycling. Radajewski (2002) even suggested that Bacteroidetes might be directly involved with methane uptake.

This marine DNA-SIP study has provided the first molecular biological evidence for microbial methane consumption in the sediment beneath a deep-water coral reef. Methane uptake measurements in combination with analysis of extracted 13C-DNA indicated involvement of a geographically widespread and marine pMMO-driven Methylomicrobium sp. Concurrent activities of heterotrophic bacteria in the sediment of this investigated Haltenpipe site suggest a role for methane in the food webs of deep-water coral reefs.

Acknowledgements

We thank the crew of the Normand Tonjer for collecting the sediment, as well as sponge samples from the deep-water coral reef at Haltenpipe. Special thanks to Jens Vee who brought these samples ashore. Hiroyuki Fuse kindly provided the mxaF gene sequence of Methylomicrobium japanense strain NI before publication. This work was funded by the Norwegian Academy of Science and Statoil (VISTA project 6146).

Footnotes

  • Editor: Gary King

References

View Abstract