SEARCH

SEARCH BY CITATION

Keywords:

  • Bacteria;
  • candidate division JS1;
  • tidal sediments;
  • bacterial community composition;
  • marine subsurface;
  • deep subseafloor biosphere

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The bacterial candidate division JS1 dominates a number of 16S rRNA gene libraries from deep subseafloor sediments, yet its distribution in shallow, subsurface sediments has still to be fully documented. Sediment cores (down to 5.5 m) from Wadden Sea tidal flats (Neuharlingersieler Nacken and Gröninger Plate) were screened for JS1 16S rRNA genes using targeted PCR-denaturing gradient gel electrophoresis (DGGE), which also detects some other important Bacteria. Bacterial subpopulations at both sites were dominated by Gammaproteobacteria in the upper sediment layers (down to 2 m) and in deeper layers by members of the Chloroflexi. The deeper layers of Neuharlingersieler Nacken consisted of grey mud with low sulphate (0.1–10 mM), elevated total organic carbon (TOC) (∼1–2%) and JS1 sequences were abundant. In contrast, the deeper sandy layers of Gröninger Plate, despite also having reduced sulphate concentrations, had lower TOC (<0.6%) with few detectable JS1 sequences. Results indicated that JS1 prefers muddy, shallow, subsurface sediments with reduced sulphate, whereas Chloroflexi may out-compete JS1 in shallow, sandy, subsurface sediments. Bacterial population changes at both sites (∼2 m) were confirmed by cluster analysis of DGGE profiles, which correlated with increased recalcitrance of the organic matter. This study extends the biogeographical range of JS1. The presence of JS1 and Chloroflexi in Wadden Sea sediments demonstrates that subsurface tidal flats contain similar prokaryotic populations to those found in the deeper subseafloor biosphere.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Nucleic acid sequences obtained directly from the environment provide the only information available for about 99% of the prokaryotes in most natural ecosystems, including marine sediments (Amann et al., 1995; Schloss & Handelsmann, 2003). Hugenholtz et al. (1998) constructed a tree that reflected the phylogenetic diversity of the Bacteria based on all 16S rRNA gene sequences in the GenBank database and identified 36 divisions or putative divisions. Less than a decade later, it has been estimated that there are now more than 80 divisions and at least half of them are composed entirely of uncultured bacteria (Schloss & Handelsmann, 2004; DeSantis et al., 2006). One candidate division with no cultivated representatives is the candidate division JS1 (Webster et al., 2004), which takes its name from the Japan Sea where the first sequences were identified in deep subsurface sediments (900 m water depth) down to 78 m below the seafloor (Rochelle et al., 1994). Since then, JS1 sequences have consistently been found in deep marine and subsurface sediments (e.g. Li et al., 1999; Inagaki et al., 2003; Newberry et al., 2004; Webster et al., 2006a; Parkes et al., 2007). It is now believed that members of the candidate division JS1 are widespread within the deep subseafloor biosphere, and that they constitute a deep marine subsurface bacterial lineage (Teske, 2006). However, in addition to marine subsurface sediments, JS1 have also been found in a number of other anoxic sedimentary environments, such as methane hydrate bearing sediments (Reed et al., 2002; Inagaki et al., 2006), marine and terrestrial mud volcanoes (Alain et al., 2006; Niemann et al., 2006; J. Sas, G. Webster, A.J. Weightman & R.J. Parkes, unpublished data), near-surface hydrothermal marine sediments (Teske et al., 2002), near-surface brackish and coastal sediments (Webster et al., 2004), and benzene-degrading (Phelps et al., 1998) and acetate-utilizing (Webster et al., 2006b) sulphate-reducing sediment enrichments. Despite such a wide and cosmopolitan distribution of JS1 within marine sediments, it is hard to infer any biogeochemical role for JS1.

Recently, JS1 and other uncultured deep biosphere-related bacteria were also identified in deeper sediment layers from the tidal flats of the German Wadden Sea (Wilms et al., 2006a). Because tidal mud flats are highly productive coastal marine systems with generally higher prokaryotic biomass and diversity (Kim et al., 2004; Mussmann et al., 2005; Stevens et al., 2005) than sediments from the deep subseafloor biosphere (Parkes et al., 2000) and are relatively accessible, they may provide a convenient model to study and understand the ecological role of uncultured prokaryotic lineages within the subsurface environment.

The aim of this investigation was to undertake a bacterial survey of the tidal flats from the Wadden Sea down to around 5.5 m to document the depth distribution of JS1 and compare this with sediment lithology and geochemical data. The study used the nonspecific JS1 targeted 16S rRNA gene PCR-denaturing gradient gel electrophoresis (DGGE) screening method, developed for identifying JS1 in subseafloor sediments (Webster et al., 2004, 2006a), comprehensively for the first time in a sedimentary environment that supports a complex and diverse bacterial population.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Site description

Sediment cores were taken from tidal flats located close to the island of Spiekeroog in the East Frisian Wadden Sea on the German North Sea coast in June 2002. Two sites were sampled, Neuharlingersieler Nacken (53°43.270′N, 7°43.718′E; core length 4.5 m) and Gröninger Plate (53°43.638′N, 7°45.960′E; core length 5.5 m) as described elsewhere (Köpke et al., 2005; Wilms et al., 2006a) and sediment samples were stored at −20°C until analysed. Sediment physical and geochemical parameters for the cores were measured previously and methods described elsewhere (Rütters et al., 2002; Chang et al., 2006; Wilms et al., 2007). Statistical analysis of data (Runs test and anova) was carried out using minitab release 14 (Minitab Inc.).

DNA extraction and purification

Total community genomic DNA was extracted from all sediment samples (Neuharlingersieler Nacken, 0.5–450 cm depth; Gröninger Plate, 0.5–520 cm depth) using the FastDNA Spin Kit for Soil (QBiogene) with modification as described in Webster et al. (2003) between 10 November 2004 and 22 February 2005. DNA extracts were aliquoted and stored at −80°C until required for PCR amplification. Additionally, a negative control DNA extraction (where no sediment was added to the extraction reagents) was also carried out and analysed by nested PCR.

PCR conditions

Candidate division JS1 (and some other bacterial) 16S rRNA gene sequences were amplified by PCR from sediment DNA using the JS1 targeted primer pair 63F-665R (Webster et al., 2004). All JS1 PCR products were then reamplified with the general bacterial primer set 357FGC-518R for DGGE analysis as described elsewhere (Muyzer et al., 1993; Webster et al., 2004). As reported previously (Webster et al., 2004), nested PCR-DGGE with the JS1 targeted primers can result in non-JS1 sequences being amplified, such as those belonging to some members of the Gammaproteobacteria (see ‘Discussion’). Analysis of the 665R with oligolocator version 1.0 (http://www.bioinformatics-toolkit.org) and the closest blastn 16S rRNA gene sequence (90–100% sequence identity to non-JS1 Wadden Sea phylotypes) shown in Table 1 demonstrated that 665R could potentially tolerate one to five base mismatches with Gammaproteobacteria, one mismatch with Deltaproteobacteria and one mismatch with Chloroflexi.

Table 1.   Closest 16S rRNA gene sequence matches to excised DGGE bands using the NCBI blastn search and RDP sequence match tools
DGGE band identifier*Alignment length (bp)PhylaClosest match by blastn search (accession number)Sequence similarity (%)Environment of nearest environmental clone matchNearest described relative in RDP (accession number)Sequence similarity (%)
  • *

    NN, Neuharlingersieler Nacken; GP, Gröninger Plate (Figs 1 and 2).

  • JS1, candidate division JS1.

NN1170GammaproteobacteriaClone JH10_C96 (AY568842)95Intertidal mud flatThialkalivibrio denitrificansT (AF126545)90
NN2169GammaproteobacteriaClone PDA-OTU12 (AY700599)99Associated with the coral Pocillopora damicornisThiobacillus prosperus DSM 5130 (AY034139)96
NN3172GammaproteobacteriaClone SIMO-1680 (AY711046)98Salt marshThiobacillus prosperus DSM 5130 (AY034139)96
NN4, NN10164–174GammaproteobacteriaClone SIMO-2306 (AY711672)96–97Salt marshThiobacillus prosperus DSM 5130 (AY034139)92–93
NN5, NN9167–170GammaproteobacteriaClone MBMPE52 (AJ567564)94Deep-sea sedimentMarichromatium purpuratumT DSM 1591 (AJ224439)88–90
NN6171GammaproteobacteriaMarinobacter aquaeolei isolate OC-11 (AY669171)98Marine sedimentMarinobacter aquaeolei isolate OC-11 (AY669171)98
NN7160GammaproteobacteriaClone SIMO-660 (AY712197)95Salt marshThialkalivibrio thiocyanodenitrificans strain ARhD (AY360060)89
NN8170GammaproteobacteriaClone SIMO-2233 (AY711599)98Salt marshThiobacillus prosperus DSM 5130 (AY034139)94
NN11, NN14162–168JS1Clone ODP1176A6H_16_B (AY191353)98Subseafloor sediment
NN12, NN18161–170JS1Clone OHKB2.83 (AB094826)96–99Subseafloor sediment
NN13135ChloroflexiClone a2b029 (AF419669)98Hydrothermal surface sedimentDehalococcoides sp. BHI80-52 (AJ431247)91
NN15157JS1DGGE band NANK-B-GW-1 (AJ585415)98Subseafloor sediment
NN16162JS1Clone MA-A2-104 (AY093459)96Methane hydrate bearing subseafloor sediment
NN17140JS1Clone GoM GC234 602E (AY211673)96Gas hydrate mound sediment
NN19, NN27142–148ChloroflexiClone FS118-10B-02 (AY704393)99–100Oceanic crustDehalococcoides sp. BHI80-52 (AJ431247)88–92
NN20172JS1Clone Amsterdam-2B-61 (AY592418)99Submarine mud volcano sediment
NN21120DeltaproteobacteriaClone 42-B28 (AJ867588)100Subseafloor sedimentDesulfovibrio baarsii DSM 2075 (AF418174)92
NN22119DeltaproteobacteriaClone ODP1230B19.07 (AB177163)97Methane hydrate bearing subseafloor sedimentDesulfovibrio baarsii DSM 2075 (AF418174)92
NN25, NN26147GammaproteobacteriaClone SBseep6 (AY456981)97–98Marine hydrocarbon seep sedimentsThiobacillus prosperus DSM 5130 (AY034139)94–95
GP1166GammaproteobacteriaClone SIMO-921 (AY710455)97Salt marshThiobacillus prosperus DSM 5130 (AY034139)94
GP2, GP11110–117GammaproteobacteriaClone AT-s3-25 (AY225639)90–94Hydrothermal sedimentThialkalivibrio nitratisT ALJ12 (AF126547)89–91
GP3166GammaproteobacteriaClone SIMO-2136 (AY711502)96Salt marshMarichromatium purpuratumT DSM 1591 (AJ224439)93
GP4171GammaproteobacteriaMarinobacter aquaeolei isolate OC-11 (AY669171)95Marine sedimentMarinobacter aquaeolei isolate OC-11 (AY669171)95
GP5, GP7165–166GammaproteobacteriaClone SIMO-2233 (AY711599)96Salt marshThiobacillus prosperus DSM 5130 (AY034139)93–96
GP6147GammaproteobacteriaClone SBseep6 (AY456981)97Marine hydrocarbon seep sedimentsThiobacillus prosperus DSM 5130 (AY034139)95
GP9166GammaproteobacteriaClone BS1-0-106 (AY254927)98Tidal sedimentThiorhodococcus mannitophagusT strain WST (AJ971090)95
GP10166GammaproteobacteriaClone BS1-0-106 (AY254927)95Tidal sedimentThiococcus sp. AT2206 (AJ401213)93
GP12, GP14, GP16, GP19, GP21129–151ChloroflexiClone FS118-10B-02 (AY704393)99–100Oceanic crustDehalococcoides sp. BHI80-52 (AJ431247)86–93
GP13, GP18, GP20146–151ChloroflexiClone JH10_C69 (AY568821)98–100Intertidal mud flatDehalococcoides sp. BHI80-52 (AJ431247)87
GP15125ChloroflexiClone Milano-WF1B-16 (AY592859)95Mud volcano microbial matDehalococcoides sp. BHI80-52 (AJ431247)87
GP17143ChloroflexiClone GIF1 (AF407193)93Monochlorobenzene-contaminated groundwater treatment reactorDehalococcoides sp. BHI80-52 (AJ431247)88
GP22170JS1Clone ODP1176A6H_16_B (AY191353)98Subseafloor sediment
GP25140JS1Clone GoM GC234 602E (AY211673)95Gas hydrate mound sediment

To restrict contamination to a minimum, PCR set-up was carried out under aseptic conditions using autoclaved and/or UV-treated plasticware and pipettes, and only sterile nuclease-free molecular grade water (Severn Biotech Ltd). Positive (representative clone DNA) and negative (molecular grade water) controls were used in all PCR amplifications, and undertaken in a DNA Engine Dyad Thermal Cycler gradient block (MJ Research). The 16S rRNA gene PCR mixture for both PCR reactions contained (total 50 μL, molecular grade water) 0.4 pmol μL−1 of primers (MWG-Biotech), 1 μL of sediment DNA template (1/10 dilution), 1 × reaction buffer (Bioline), 1.5 mM MgCl2, 1.5 U Biotaq DNA polymerase (Bioline), 0.25 mM each dNTP (Bioline). Bovine serum albumin (10 μg; Promega) was also added to the primary PCR reaction mix only.

DGGE analysis

DGGE was carried out as described by Webster et al. (2002) with some modifications. PCR products (c. 100 ng of each PCR product) were separated using a DCode Universal Mutation Detection System (Bio-Rad Laboratories) and 1-mm-thick (16 mm × 16 cm glass plates) 8% (w/v) polyacrylamide gels (Acrylogel 2.6 solution, acrylamide : N,N′-methylenebisacrylamide; 37 : 1; BDH Laboratory Supplies) with a gradient of denaturant between 30% and 60%. A denaturing gradient consisting of 100% denaturant is defined as 7 M urea with 40% (v/v) formamide. Gels were poured with the aid of a 50 mL volume Gradient Mixer (Fisher Scientific) and prepared with 1 × TAE buffer (pH 8; 40 mM Tris base, 20 mM acetic acid, 1 mM EDTA). Electrophoresis was carried out at 200 V for 5 h (with an initial 10 min at 80 V) at 60°C in 1 × TAE buffer. Polyacrylamide gels were stained with SYBRGold nucleic acid gel stain (Molecular Probes) for 30 min and viewed under UV. Gel images were captured with a Gene Genius Bio Imaging System (Syngene). DGGE profiles were analysed as described elsewhere (Fry et al., 2006) and agglomerative cluster analysis was carried out using the software package community analysis package version 3.1 (Pisces Conservation Ltd). Dendograms were constructed by average linkage and average distance method.

Recovery and sequence analysis of DGGE bands

Individual DGGE bands were excised and washed in sterile nuclease-free molecular grade water (Severn Biotech Ltd) for 10 min. Bands were then air-dried and crushed in 10–20 μL molecular grade water and incubated overnight at 4°C. The supernatant (1 μL) was used as template DNA in a PCR using primers 357F-518R as described above. The PCR products were reanalysed by DGGE to confirm recovery of excised bands. PCR products of excised DGGE bands were purified by dialysis with molecular grade water using a Microcon YM-50 centrifugal filter device (Millipore Corporation) and DNA yield quantified by 1.2% (w/v) agarose gel electrophoresis and DNA quantification marker Hyperladder I (Bioline). Sequencing reactions were performed with primer 518R and an ABI PRISM 3100-Genetic Analyzer (Applied Biosystems). Gene sequence chromatographs were analysed using the chromas lite software package version 2.01 (http://www.technelysium.com.au/chromas_lite.html) and the resulting sequences' closest relatives identified by NCBI nucleotide–nucleotide blast (blastn) (http://www.ncbi.nlm.nih.gov/) and the Ribosomal Database Project II (RDP-II) sequence match software (http://rdp.cme.msu.edu/). All 16S rRNA gene sequences were deposited in the database under the accession numbers AM492275AM492320.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Bacterial community composition as assessed by JS1 targeted PCR-DGGE

High molecular weight DNA (>10 kb) was readily extracted from all sediment depths at both Wadden Sea sites, but the amount of extractable DNA decreased with depth (∼650–40 ng DNA g−1 sediment). Contamination by exogenous DNA is of particular concern when prokaryotic genes are amplified by PCR with highly conserved 16S rRNA gene primers from low DNA templates such as subsurface sediments (Reed et al., 2002; Webster et al., 2003), and hence stringent anticontamination procedures were used in this study. Negative control DNA extracts and PCR-negative controls amplified by nested PCR showed no visible PCR products, demonstrating that all amplified bacterial 16S rRNA genes obtained in this study were derived from sediment DNA and not from contamination of laboratory reagents.

Bacterial 16S rRNA gene PCR products were obtained from all sample depths at both Wadden Sea sites by nested PCR with the candidate division JS1 targeted PCR primers followed by general bacterial primers. Changes in the composition of JS1 targeted bacterial 16S rRNA genes were assessed by DGGE (Figs 1a and 2a). DGGE profiles of the upper sediment layers, at both Wadden Sea sites, were more complex than from the deeper subsurface layers; below 2 m for Neuharlingersieler Nacken and 2.2 m for Gröninger Plate. Similar differences in DGGE profile complexity with depth were also obtained for the total bacterial community using PCR products obtained with the general bacterial 16S rRNA gene primers 357F and 518R (Muyzer et al., 1993) directly on sediments of Neuharlingersieler Nacken (data not shown).

image

Figure 1.  DGGE analysis of bacterial 16S rRNA genes from Neuharlingersieler Nacken derived from different sediment depths (0.5–450 cm). (a) DGGE of PCR products amplified by nested PCR with primers 63F-665R and 357FGC-518R. Lanes: M, DGGE marker (Webster et al., 2003); +, candidate division JS1 positive control clone Nank-B7 (Newberry et al., 2004). Labelled DGGE bands represent bands that were excised and sequenced (see Table 1 for full description, grey arrows, candidate division JS1; black arrows, Chloroflexi; white arrows, Gammaproteobacteria; striped arrows, Deltaproteobacteria). (b) Cluster analysis of DGGE band pattern shown in (a).

Download figure to PowerPoint

image

Figure 2.  DGGE analysis of bacterial 16S rRNA genes from Gröninger Plate derived from different sediment depths (0.5–520 cm). (a) DGGE of PCR products amplified by nested PCR with primers 63F-665R and 357FGC-518R. Lanes: M, DGGE marker (Webster et al., 2003); +, candidate division JS1 positive control clone Nank-B7 (Newberry et al., 2004). Labelled DGGE bands represent bands that were excised and sequenced (see Table 1 for full description, grey arrows, candidate division JS1; black arrows, Chloroflexi; white arrows, Gammaproteobacteria). (b) Cluster analysis of DGGE band pattern shown in (a).

Download figure to PowerPoint

Identification of major phylotypes from DGGE bands

Sequencing excised DGGE bands identified the dominant members of the targeted bacterial community (Table 1). At Neuharlingersieler Nacken, 15 (70%) of the 20 distinguishable band positions were excised and sequenced giving rise to 19 different sequence phylotypes. Sequenced phylotypes belonged to the bacterial groups, candidate division JS1 (31%), Chloroflexi (11%), Gammaproteobacteria (47%) and Deltaproteobacteria (11%). By comparison, at Gröninger Plate, 12 (67%) of the 18 band positions identified were sequenced and the resulting 14 phylotypes belonged to the bacterial groups, candidate division JS1 (15%), Chloroflexi (28%) and Gammaproteobacteria (57%). This observed discrepancy between the number of band positions and the number of different sequenced phylotypes is a result of the known limitations of the PCR-DGGE method. It is well documented that different sequences may comigrate to the same position in a DGGE gradient or that comigration of similar PCR fragments can be a problem for retrieving clean sequences from individual bands (Muyzer & Smalla, 1998).

Extensive sequencing of DGGE bands established that, overall, bands with equivalent positions in different lanes corresponded with the same bacterial groups. However, at Neuharlingersieler Nacken (Fig. 1), the bands NN25 and NN26 identified as Gammaproteobacteria were at a similar position to bands NN12, NN15 and NN18 belonging to the candidate division JS1. It can be proposed that the unsequenced bands at this position from depths 0.5 to 200 cm belonged to the Gammaproteobacteria and the bands from deeper depths >200 cm are JS1, because JS1 were only identified below 2 m. However, definitive assignment of bands would require further sequencing or identification by hybridization with group-specific rRNA gene probes (Teske et al., 1998; Webster et al., 2002).

The number of JS1 phylotypes (five band positions representing six sequence phylotypes) in Neuharlingersieler Nacken subsurface sediments demonstrates a greater 16S rRNA gene sequence diversity, in terms of DGGE band positions, than observed in the authors' other studies (Webster et al., 2004, 2006a). This increase in JS1 DGGE phylotypes may also suggest that the phylogenetic divergence (sequence difference) within the candidate division JS1 could be greater than previously observed (Webster et al., 2004) because DGGE band sequences showed >7.5% sequence difference over the hyper variable V3-region of 16S rRNA gene analysed (corresponding to Escherichia coli 16S rRNA nucleotide positions 341–534; Brosius et al., 1981).

Cluster analysis of DGGE profiles

The bacterial community profiles of the two sites were compared by cluster analysis. The resulting dendograms of DGGE patterns from both sites showed two distinct clusters (Figs 1b and 2b). These clusters of similar community composition seem to reflect sample depth. For example, at Neuharlingersieler Nacken one cluster (cluster NNI) consists of sample depths 0.5–200 cm and the other 220–450 cm (cluster NNII; Fig. 1b). Similar results were also observed at Gröninger Plate with one cluster (cluster GPI) comprising depths 0.5–180 cm and the other consisting of 220–520 cm (cluster GPII; Fig. 2b). However, at Gröninger Plate sample depth 220 cm branched separately from the other community profiles within the same cluster. A closer observation of the 220 cm DGGE profile (Fig. 2a) shows that the bacterial community at this depth shares similarities (bands present) with both the shallower depths and also those from the depths below 220 cm.

Sampling site lithology and geochemistry

The sediments at the two Wadden Sea sampling sites differed with respect to both lithology and geochemistry (Fig. 3). At Neuharlingersieler Nacken, sand dominated the upper layers interspersed with thin layers of black mud (down to 160 cm), followed by a porous layer of sand and shells (ca. 160–240 cm). Below 240 cm, the sediments consisted of more homogeneous grey mud with organic remnants (>70% mud; grain size fraction <63 μm). The Gröninger Plate site was dominated by sand and mixed tidal flat sediments intercalated with thin layers of mud and shells throughout the core (Fig. 3).

image

Figure 3.  Depth profile of lithology (M, mud; S, sand; Sh, shells; simplified from Chang et al., 2006) and selected geochemical data (total organic carbon (TOC), dissolved organic carbon (DOC), carbon isotope (δ13C) composition of organic matter and pore water sulphate concentrations) for tidal sediments from two sites of the German Wadden Sea, (a) Neuharlingersieler Nacken and (b) Gröninger Plate. A summary of the dominant bacterial populations at each depth are also indicated.

Download figure to PowerPoint

Oxygen was only detected in the sediments at both sites in the upper 3 mm (Köpke et al., 2005) and pore water sulphate concentrations decreased with depth down to around 140–180 cm below the surface (Fig. 3). Both sites also had increasing sulphate concentrations at depth but whereas at Gröninger Plate, the sulphate steadily increased in deeper layers, Neuharlingersieler Nacken sulphate concentrations peaked at c. 240 cm and then declined to around zero (Fig. 3). Methane measured at Neuharlingersieler Nacken was present throughout the anoxic layers of the sediment with highest values in the two sulphate-minimum zones (100–200 cm and below 400 cm; Wilms et al., 2007). Constant pore water chloride concentrations with depth (Köpke et al., 2005), indicated that the changes in pore water were not a result of an influx of groundwater but possibly from an influx of sulphate-rich water from a tidal creek (Wilms et al., 2007).

At Neuharlingersieler Nacken, the total organic carbon (TOC) was found to be low in the sand layers and increased with increasing mud content (up to 2% TOC). The grey mud in the lower part of the core showed elevated TOC contents (∼1–2%) due to the presence of roots, plant fragments and peat particles. Dissolved organic carbon (DOC) increased with depth from c. 22 to 100 p.p.m. in pore water (Fig. 3a). The δ13C values of the bulk organic matter were −22.6±1.1‰ in the upper 2 m sediment layers, changing to c. −24.4±0.4‰ in the deeper sediments of Neuharlingersieler Nacken. The TOC at Gröninger Plate was <0.6% in the sand and mixed sediments with peaks in layers at around 120 and 340 cm (c. 1.4% and 1% TOC, respectively; Fig. 3b). DOC increased from the surface and then remained relatively constant below 50 cm (c. 20–40 p.p.m.). The δ13C values of the bulk organic matter at Gröninger Plate ranges from c. −22.1±0.7‰ in the upper 2 m to c. −24.7±0.5‰ in the subsurface layers. Statistical analysis of the δ13C data using the Runs test showed that the data at both sites could be divided into two groups according to depth (divided at the 2 m) and were significantly nonrandom (Neuharlingersieler Nacken, P<0.001; Gröninger Plate, P<0.0001), and in both cases the mean δ13C values for each group were significantly different at the 2 m depth (anova, P<0.0001).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The distribution of a specific, subsurface bacterial population was investigated in two tidal flat sediments of the German Wadden Sea by screening with the candidate division JS1 targeted 16S rRNA gene PCR-DGGE (Webster et al., 2004). The specific bacterial communities present at both sites were dominated by Gammaproteobacteria in the upper sediment layers (down to 2 m) and in deeper layers by the so far, largely uncultivated phylum, Chloroflexi (Fig. 3). However, in Neuharlingersieler Nacken subsurface sediments (below 2 m), sequences of the candidate division JS1 were also a major component. However, the deeper subsurface sediments (below 220 cm) at Gröninger Plate only had a few detectable JS1 sequences, which were within specific sediment layers (Fig. 3). Similar dominant bacterial populations (Gammaproteobacteria and Chloroflexi) were also reported by Wilms et al. (2006a), using general bacterial 16S rRNA gene primers (357F and 907R), on the same sediments. However, JS1 were not detected with general bacterial primers in sediments of Neuharlingersieler Nacken and was only observed in some deeper layers at site Gröninger Plate (Wilms et al., 2006a). The absence of JS1 in the previous study may be due to differences in PCR primer specificity (Baker et al., 2003) and/or that JS1 have a relatively low abundance compared with other taxa in these sediments and are not observed by general bacterial 16S RNA gene DGGE (Wilms et al., 2006a).

Previously, nested PCR-DGGE with the JS1 targeted primers had resulted in a few non-JS1 sequences being amplified from other sedimentary environments (Webster et al., 2004, 2006a). Non-JS1 sequences previously amplified belonged to the same phyla as detected in this study. For example, a small number of Chloroflexi and Deltaproteobacteria phylotypes were also amplified from Peru Margin subseafloor sediments and Gammaprotobacteria from a UK salt marsh (Webster et al., 2004, 2006a). Analysis of the JS1 targeted primer 665R using rdp-ii probe match software demonstrated that it matched 0.35% of relevant nontarget database sequences including Chloroflexi (22%), Proteobacteria (15%), Firmicutes (16%), Deferribacteres (2%), Actinobacteria (1%), Acidobacteria (2%) and RDP unclassified bacteria (41%). These unclassified bacteria were mainly sequences affiliated with the candidate division JS1, Chloroflexi and candidate division OP8. Proteobacterial sequences hit by 665R were Betaproteobacteria (1%), Deltaproteobacteria (34%; unclassified phylotypes) and Gammaproteobacteria (56%; families Ectothiorhodospiraceae, Xanthomonadaceae and Halomonadaceae) phylotypes. The majority of Gammaproteobacteria identified in this study were distantly related (88–96% sequence identity) to sequences affiliated with the families Ectothiorhodospiraceae and Chromatiaceae (e.g. Thialkalivibrio, Marichromatium, Thiorhodococcus and Thiococcus; see Table 1) and the chemolithotrophic Thiobacillus prosperus (Huber & Stetter, 1989).

Recently it has been shown that the microbial diversity in the world's oceans is much greater than previously estimated (Sogin et al., 2006) based on conventional molecular techniques and it is likely to be the same in tidal sediments. The use of selective bacterial PCR, which targets specific bacterial populations that may be low in abundance, such as the candidate division JS1, has been beneficial in this study because it has clearly identified JS1 in sediments of Neuharlingersieler Nacken that were not detected using culturing methods or general bacterial PCR in other studies (Köpke et al., 2005; Wilms et al., 2006a, b). The identification of candidate division JS1 within tidal flat subsurface sediments of the Wadden Sea adds further information on the biogeographical distribution of these organisms in the environment and strongly suggests that these organisms play a role in anoxic sedimentary systems. To date, the metabolism of this uncultured bacterial group has only been inferred from the environments where it has been found. For example, JS1 have been identified in sulphate-reducing consortia able to mineralize benzene (Phelps et al., 1998), in methane hydrate-bearing sediments (Reed et al., 2002; Inagaki et al., 2006), associated with anaerobic methanotrophic communities in sediments with high sulphate reduction rates (Teske et al., 2002; Dhillon et al., 2003), and in methane- and hydrocarbon-discharging, brine-soaked, mud breccias in a terrestrial mud volcano (Alain et al., 2006).

However, in a recent stable-isotope probing study (Webster et al., 2006b) using Tamar Estuary (UK) sediment slurries incubated under anoxic sulphate-reducing conditions with low sulphate (c. 2 mM), it was observed that JS1 was active and able to utilize 13C-glucose (or glucose metabolites) and 13C-acetate. In the present study JS1 was more prevalent in the deeper mud layers of Neuharlingersieler Nacken at sediment depths with 0.1–10 mM sulphate, presence of methane, elevated TOC (c. 1–2%) and DOC (c. 30–100 p.p.m.; Fig. 3) and with more refractory organic matter. However, the sandy Gröninger Plate site, which had fewer detectable JS1 DGGE bands, also had sulphate at depth (c. 6–12 mM), but it had much lower concentrations of TOC and DOC. Interestingly, the two depths where JS1 were identified (280 and 400 cm) at Gröninger Plate are both close to narrow mud layers (Fig. 3) and either side of a subsurface TOC peak (c. 1.2%; 330–350 cm).

Although from this study the ecological role of JS1 in subsurface sediments cannot be further defined, the results do reinforce previous observations that members of JS1 inhabit strictly anoxic organic-rich (although probably poor quality recalcitrant carbon) environments. Recently, it has been suggested that JS1 may prefer organic-rich sedimentary habitats with high concentrations of methane hydrate (Inagaki et al., 2006). However, our research adds to previous findings that JS1 is not restricted to gas hydrate-containing sediments (Webster et al., 2004). Because low sulphate concentrations are present in the deeper layers at both Wadden Sea sites, JS1 may prefer low-sulphate habitats, and given its dominance in deeper mud layers of Neuharlingersieler Nacken, JS1 may be adapted to surviving in sediments with reasonable concentrations of TOC, small grain size and associated small pore size. For example, it has been reported that small pores restrict bacterial movement, reduce activity, limit nutrient transport, slow cell growth rate and reduce bacterial diversity (Rebata-Landa & Santamarina, 2006). This and the refractory type of organic carbon available could account for their distribution and sometimes dominance in a number of subsurface sediments.

Uncultured members of the Chloroflexi, which are also often found to dominate organic-rich subseafloor sediments (Coolen et al., 2002; Parkes et al., 2005; Inagaki et al., 2006; Webster et al., 2006a), were identified in the subsurface sediments of both the Wadden Sea sites analysed. This may suggest that Chloroflexi distribution is not affected by the same lithological conditions as JS1 and may be able to out-compete JS1 in sandy sediments with a larger grain and pore size. Because JS1 and Chloroflexi are often the dominant bacterial members of subsurface sediments they have now been described as bacteria that typify this unique extreme environment (Teske, 2006). These results support previous findings (Wilms et al., 2006a) that subsurface tidal flat sediments sustain similar prokaryotic populations as those found in the deep subseafloor biosphere. It is interesting to note that at both Wadden Sea sites there was a distinct change in the specific bacterial community structure at around the 2 m depth, which was clearly demonstrated by cluster analysis of DGGE profiles (Figs 1 and 2) from Gammaproteobacteria to typical deep biosphere-related Bacteria (JS1 and/or Chloroflexi). Similarly at the same depth there was also a significant change in the quality of organic matter as indicated by lighter δ13C values in the subsurface sediments. Light isotopic values are indicative of more refractory organic matter of terrestrial origin, which in this environment is mainly derived from eroded fossil peat (Volkman et al., 2000). Recalcitrant carbon, changes in dominant bacterial species to typical uncultured subsurface phyla coupled with a significant decrease in total cell numbers (Köpke et al., 2005; Wilms et al., 2007) may suggest that this depth could mark the beginning of conditions similar to the deep biosphere in tidal flat sediments.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

G.W. was funded by the Natural Environment Research Council (NERC) Marine and Freshwater Microbial Biodiversity thematic programme grant numbers NER/T/S/2000/636 and 2002/00593. E.F. and J.K. were funded by the Deutsche Forschungsgemeinschaft (DFG) BioGeochemistry of Tidal Flats research project (FOR432). The authors would like to thank Mr Stephen Hope for technical assistance with sequencing and Professor John Fry for help with statistical analysis.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References