Correspondence: R. John Parkes, School of Earth and Ocean Sciences, Cardiff University, Main Building, Park Place, Cardiff, Wales CF10 3AT, UK. Tel.: +44 29 208 70058; fax: +44 29 208 74326; e-mail: ParkesRJ@cardiff.ac.uk
Stable isotope probing of prokaryotic DNA was used to determine active prokaryotes using 13C-labelled substrates (glucose, acetate, CO2) in sediment slurries from different biogeochemical zones of the Severn Estuary, UK. Multiple, low concentrations (5 × 100 μM) of 13C-substrate additions and short-term incubations (7 days) were used to minimize changes in the prokaryotic community, while achieving significant 13C-incorporation. Analysis demonstrated clear metabolic activity within all slurries, although neither the net sulphate removal nor CH4 production occurred in the anaerobic sulphate reduction and methanogenesis zone slurries. Some similarities occurred in the prokaryotic populations that developed in different sediment slurries, particularly in the aerobic and dysaerobic zone slurries with 13C-glucose, which were dominated by Gammaproteobacteria and Marine Group 1 Archaea, whereas both anaerobic sediment slurries incubated with 13C-acetate showed incorporation into Epsilonproteobacteria and other bacteria, with the sulphate reduction zone slurry also showing 13C-acetate utilization by Miscellaneous Crenarchaeotic Group Archaea. The lower potential energy methanogenesis zone slurries were the only conditions where no 13C-incorporation into Archaea occurred, despite Bacteria being labelled; this was surprising because Archaea have been suggested to be adapted to low-energy conditions. Overall, our results highlight that uncultured prokaryotes play important ecological roles in tidal sediments of the Severn Estuary, providing new metabolic information for novel groups of Archaea and suggesting broader metabolisms for largely uncultivated Bacteria.
Tidal flat sediments occur at the land–sea interface in tropical and temperate regions and are among the most productive coastal marine ecosystems (Alongi, 1998). They receive organic matter and nutrient input from both land and sea, and as a result, are often characterized by intense heterotrophic and photoautotrophic activity (Poremba et al., 1999). As a consequence, high microbial activity in the upper sediment layers generates steep geochemical gradients and distinct biogeochemical zones. These biogeochemical zones enable the description of early diagenetic reactions in sediments and represent the prokaryotic degradation of organic carbon using successively less energy-yielding terminal electron acceptors (Froelich et al., 1979; Canfield & Thamdrup, 2009). Biogeochemical zones have been used to describe a wide variety of marine sedimentary environments where the supply of labile organic matter exceeds diffusion of oxygen into the sediment (e.g. Thomsen et al., 2001; Mortimer et al., 2002; Parkes et al., 2005). The depth range of each zone is determined from a characteristic sequence of chemical changes in the sediment pore water (Jørgensen, 1983). In estuarine sediments, typically oxygen is depleted within the top few millimetres and anoxic conditions prevail beneath this depth. Although nitrate, iron and manganese are then the initial important anaerobic electron acceptors for the degradation of organic matter, it is dissimilatory sulphate reduction that is the predominant organic matter anaerobic degradation process in marine sediments (Jørgensen, 1982), followed by methanogenesis at depth when sulphate becomes depleted (Oremland & Taylor, 1978).
The Severn Estuary is a nutrient-rich and tidally dynamic coastal ecosystem located between England and Wales. It is a large estuary with extensive intertidal mud-flats, sand-flats, rocky platforms and islands. The estuary's classic funnel shape, unique in the United Kingdom, is a factor causing the Severn Estuary to have the second highest tidal range in the world (∼15 m, but lower than the Bay of Fundy, Canada; Archer & Hubbard, 2003) and also one of the largest intertidal zones in the United Kingdom. However, a few reports suggest that the Severn Estuary as a whole supports a rather impoverished flora and fauna, and high turbidity means that phytoplankton productivity is generally lower than expected due to limited light penetration (Joint, 1984). Despite the low annual primary production, the Severn Estuary has a high nutrient input (Morris, 1984) from the many rivers that drain into the estuary from a large area of England and Wales, which presumably influences the sedimentary microbial processes within this system. However, very little is known about the prokaryotic diversity and activity within these tidal sediments. One study of the intertidal mudflat at Aust Warth (Wellsbury et al., 1996) demonstrated that anaerobic sulphate reduction was the dominant degradation process within these sediments, accounting for 60% of the total organic matter degradation, with methanogenesis occurring at much lower rates (<1%) in subsurface sediments. However, evidence was also shown that the top ∼8 cm of sediment may have been recently disturbed or deposited due to the large tidal impact encountered in the Severn Estuary. The effect of this continual disturbance and oxygenation on the anaerobic prokaryotic community is unknown.
Culture-independent approaches based on 16S rRNA gene sequencing and phospholipid fatty acid (PLFA) biomarker profiles have been widely used to characterize microbial populations in marine sediments, including those from estuaries (e.g. Parkes et al., 1993; Purdy et al., 2001; Roussel et al., 2009a). Stable-isotope probing (SIP) has extended this approach by determining 13C substrate incorporation/utilization into specific biomarkers, hence enabling a direct link between substrate utilization and specific prokaryotes (Boschker et al., 1998; Radajewski et al., 2000; Neufeld et al., 2007b). Using this approach, we have previously characterized acetate-, glucose- and pyruvate-utilizing bacteria in an established sulphate-reducing sedimentary prokaryotic community (Webster et al., 2006b). In the present study, we have used SIP with environmentally relevant substrates (13C-glucose, 13C-acetate and 13CO2) to investigate the active prokaryotic community in Severn Estuary sediments from different biogeochemical depth zones, without prior substrate enrichment, to identify prokaryotes in addition to terminal-oxidizers. These prokaryotes may be particularly important in Severn Estuary sediments as tidal sediment disturbance may hinder the establishment of communities dominated by anaerobic terminal-oxidizing Bacteria and Archaea.
Materials and methods
Desulfobacter sp. DSM 2035 was grown on DSM medium 195 supplemented with either 12C-sodium acetate or 13C-sodium acetate (1,2-13C2, 99%; CK Gas Products Ltd) to provide known 13C-labelled and unlabelled DNA for use as markers in density gradient ultracentrifugation (Webster et al., 2006b).
Sediment slurry microcosms
Estuarine sediment cores (diameter 10 cm, depth 1 m) were collected at low tide from tidal flats of the Severn Estuary, Woodhill Bay, Portishead, UK (51o29′30.94″N, 2o46′28.91″W), during two expeditions in early October 2004. The first exploratory visit was to identify the site and sample sediments for pore water chemical analysis and the second visit was to collect sediment samples for SIP experiments. All sediment samples were transported back to the laboratory for immediate processing.
Sediments from different depths and biogeochemical (metabolic) zones with potentially different prokaryotic populations were used to establish replicate sediment slurries (see Fig. 1): aerobic zone slurry A, an aerobic sediment slurry made from sediment from the top 3 cm including the aerobic zone (0 to ∼0.5 cm); dysaerobic (Raiswell & Canfield, 1998) zone slurry B, an anaerobic sediment slurry comprising of sediment from a regularly disturbed (Wellsbury et al., 1996) anoxic, but not sulphidic, sediment zone from 3 to 20 cm depth, presumably containing the nitrate, manganese and iron reduction zones [Canfield & Thamdrup, 2009; recent data from fresh cores (June 2009) show active manganese and iron reduction occurring at this site in the top 15 cm, E.G. Roussel, M. Olivier, R.J. Parkes & H. Sass, unpublished data] and part of the sulphate reduction zone; sulphate reduction zone slurry C, an anaerobic sediment slurry made from only black sulphide-rich sediments from the active sulphate reduction zone (sulphate concentrations decrease from 20 to ∼15 mM) and overlapping methanogenesis zone (20–40 cm); and methanogenesis zone slurries D and E, two anaerobic sediment slurries both comprising sediment from the methanogenesis zone (40–70 cm, containing the highest methane concentrations) and the bottom of the overlapping sulphate reduction zone, where rates of sulphate removal decrease.
Sediment (250 mL) was added to 750 mL of oxic or anoxic (reduced with 1 mM sodium sulphide) mineral salts medium (Wellsbury et al., 1994). The anoxic media used for the sulphate reduction zone sediment slurry C also contained 18 mM sodium sulphate. All slurries were contained within modified 2-L screw-capped bottles (except aerobic slurry A, which was closed with a foam bung) fitted with a shoulder opening for gas input and a three-way stopcock at the base for slurry sampling. The gas headspace in the anoxic sediment slurries (slurries B, C, D and E) was replaced with oxygen-free nitrogen. All slurries were incubated at 25oC in the dark on an orbital shaker (150 r.p.m.).
Labelled 13C-substrates (100 μM) were added to the slurries at time zero and subsequently each day for the first 4 days (total 5 × 100 μM), after which slurries were left to incubate for a further 10 days (total 14 days). Slurry samples (50 mL) were taken at 0, 0.5, 1, 2, 3, 4, 5, 7 and 14 days, and stored at −0oC until required. The substrates used were 13C-glucose (U-13C6, 99%, CK Gas Products Ltd), slurries A and B; 13C-acetate (1,2-13C2, 99%, CK Gas Products Ltd), slurries C and D; and 13C-CO2 (U-13C, 99%; CK Gas Products Ltd), slurry E.
Chemical analysis of sediment pore water
Fifty millilitres of sediment was sampled from all slurries at regular intervals into 50-mL volume centrifuge tubes flushed with oxygen-free nitrogen. Sediment samples were then centrifuged for 15 min at 2300 g in a Hettich Rotanta 460R centrifuge at 10oC. The pore water supernatant was then removed and analysed, while the remaining sediment pellet was stored at −80oC for subsequent molecular analysis. Sulphate, nitrate, volatile fatty acid (VFA) concentrations and other anions were determined using an ICS-2000 ion chromatography system with an AS50 autosampler (Dionex UK Ltd) fitted with two Ionpac AS15 columns in series, and an anion self-regenerating suppressor (ASRS-ULTRA II 4-mm) in combination with a DS6 heated conductivity cell (Dionex UK Ltd) under the conditions described previously (Webster et al., 2009). Ammonium and other cations were analysed using a DX-120 ion chromatography system with an AS40 autosampler (Dionex UK Ltd) fitted with an Ionpac CS16 and a cation self-regenerating suppressor (CSRS-300 4 mm) in combination with a DS4-1 heated conductivity cell (Dionex UK Ltd), and using 25 mM methanesulphonic acid as an eluent.
For measuring methane, 2 cm3 of sediment was added to 20 mL 10% (w/v) KCl in 50-mL volume gas-tight serum bottles and headspace gas analysed by GC using a modified Perkin Elmer/Arnel Clarus 500 Natural Gas Analyser fitted with a flame ionization detector and a thermal conductivity detector. Headspace gas from sediment slurries was analysed by GC directly.
Genomic DNA was extracted from pure cultures of Desulfobacter sp. DSM 2035 using the FastDNA SPIN Kit (MP Biomedicals). Sediment DNA was extracted from either 1 g sediment or 5 × 1 g 13C-enriched sediment slurry using the FastDNA SPIN kit for Soil (MP Biomedicals), with modifications as described by Webster et al. (2003). DNA extracts were visualized by standard agarose gel electrophoresis, and the DNA was quantified against Hyperladder I DNA marker (Bioline) using the Gene Genius Bio Imaging System (Syngene).
CsCl density gradient ultracentrifugation
The CsCl density gradient ultracentrifugation and DNA fractionation into 12C- and 13C-DNA conditions were as described by Webster et al. (2006b). DNA (∼5 μg) from the 13C-enriched sediment slurries were fractionated, and 12C- and 13C-DNA fractions were removed from the CsCl gradient alongside a ‘marker’ tube containing Desulfobacter sp DSM 2035 12C- and 13C-DNA as a visual guide (Webster et al., 2006b). 12C- and 13C-DNA fractions were cleaned to remove ethidium bromide and CsCl using molecular-grade water- (Severn Biotech) saturated n-butanol, followed by dialysis with Microcon YM-100 filters (Millipore Corporation). Purified DNA was eluted in 40 μL sterile molecular-grade water and stored at −80oC until required.
16S rRNA gene PCR-denaturing gradient gel electrophoresis (DGGE) analysis
PCR amplification, cloning and phylogenetic analysis
Bacteria and Archaea 16S rRNA genes were amplified from 12C- and 13C-DNA with the PCR primers 27F/907R and 109F/958R, respectively, under the conditions described in Webster et al. (2006a). Dissimilatory sulphite reductase (dsrA) gene sequences were also amplified using DSR1F and DSR4R (Wagner et al., 1998).
Five PCR reactions from each sample were pooled, cleaned and concentrated using Microcon YM 100 spin filters (Millipore Corporation) and eluted in 40 μL sterile-distilled water. Pooled PCR products were quantified and ligated into pGEM T-easy vector, and transformed into Escherichia coli JM109 competent cells (Promega Corporation). Clones containing the correct insert, after checking by amplification with M13 primers, were sequenced using an ABI 3130xl Genetic Analyzer (Applied Biosystems).
Sequence chromatographs were analysed using the chromas lite software version 2.01 (http://www.technelysium.com.au/chromas.html). Partial sequences and their closest relatives were identified by NCBI blastn (http://www.ncbi.nlm.nih.gov/). All nucleotide sequences were aligned using clustalx (Thompson et al., 1997) with sequences retrieved from the database. Alignments were edited manually using BioEdit Sequence Alignment Editor version 5.0.9 (Hall, 1999) and regions of ambiguous alignment were removed. Phylogenetic trees were constructed using neighbour-joining with the Jukes and Cantor correction algorithm in mega 4 (Tamura et al., 2007).
The new sequences reported here have been submitted to the EMBL database under accession numbers FN424302–FN424337 for 16S rRNA gene sequences and FN424338–FN424349 for dsrA gene sequences.
Sampling site geochemistry
The sediments at the Severn Estuary sampling site were homogeneous silty-clay (Mclaren et al., 1993) down to the bottom of the core (100 cm). Methane and sulphate profiles were measured throughout the core. Sulphate concentrations at the surface were 29 mM and decreased with depth to ∼10 mM at 60 cm below the sediment surface (Fig. 1a). In contrast, methane concentrations increased with depth from zero in the surface 10 cm to concentrations of ∼16–25 μmol L−1 sediment below 50 cm. Acetate (Fig. 1a) and other VFAs (lactate and formate; data not shown) were consistently present in the sediment pore water. Acetate concentrations decreased with depth (103 to 63 μM acetate), while both lactate and formate varied in concentration between 8–34 μM lactate and 90–127 μM formate. Nitrate (10 μM) was only detected in the sediment surface (data not shown). Samples from a recent core taken in June 2009 had an in situ temperature of 22oC down to 10 cm and a porosity of 51–77%.
Bacterial 16S rRNA gene PCR-DGGE analysis of DNA extracted from sediments representative of each of the four zones used to prepare the slurries clearly demonstrated that each zone had a different bacterial community structure (Fig. 1b), indicative of different geochemical conditions at each depth.
Sediment slurry geochemistry and prokaryotic activity
Pore water from all sediment slurries incubated with 13C-substrates was analysed at specific time points throughout the SIP experiment.
Aerobic zone sediment slurry A
During and after the repeated addition of 13C-glucose (5 × 100 μM) to the aerobic zone sediment slurry A, no increases in VFAs occurred, suggesting complete oxidation of glucose; acetate slowly decreased from ∼30 μM at 1 day to 19 μM at 14 days, lactate remained relatively constant between 36 and 40 μM and formate decreased with time (6 to <0.2 μM), with a slight peak in concentration (14 μM) at 3 days. Sulphate concentrations remained constant throughout the incubation at ∼4.5 mM. Nitrite and nitrate concentrations were below detection until 14 days, when the values were 175 and 300 μM, respectively, coupled with a steady decrease in ammonia from ∼1 mM at 24 h to below detection at 14 days, indicative of aerobic ammonia oxidation.
Dysaerobic zone sediment slurry B
In contrast to slurry A, addition of the same concentration of 13C-glucose to the anoxic sediment slurry B resulted in a clear increase in acetate (175 μM) after 1 day, suggesting fermentation of glucose. This acetate was rapidly utilized, decreasing to <27 μM by 4 days. Lactate and formate remained low or below detection (0–7 μM), with the exception that the lactate concentration peaked at 4 days (∼36 μM). Sulphate remained constant at ∼6 mM.
Sulphate reduction zone sediment slurry C
Repeated 100 μM additions of 13C-acetate to the sulphate reduction zone sediment slurry C caused a steady increase in acetate from time zero (57 μM) to 3 days (460 μM), and then by 4 days, acetate had been rapidly utilized, decreasing to ∼50 μM. After 4 days, acetate utilization was slower, with concentrations decreasing to only ∼30 μM by 14 days. Formate concentrations remained low (0–7.2 μM) throughout the incubation and lactate was undetectable until 4 days (∼45 μM) and then remained constant until 14 days. Sulphate concentrations remained relatively stable throughout the incubation (fluctuating between 22 and 23 mM), which suggests that no or very little net sulphate reduction occurred during the experiment.
Methanogenesis zone sediment slurries D and E
In contrast to slurry C, the repeated addition of 13C-acetate to the methanogenesis zone sediment slurry D demonstrated that acetate was being rapidly removed from the slurry as increased acetate concentrations were only observed at 1 day (78 μM), after which acetate remained low and subsequent 13C-acetate additions were not detected. Acetate concentrations fluctuated between 22 and 29 μM over the remaining 14-day incubation period. The acetate concentration in the 13CO2-amended methanogenesis zone sediment slurry E increased to 289 μM after 1 day and then rapidly decreased to 24 μM by 2 days, remaining relatively constant thereafter until 14 days, suggesting initially possible acetogenesis, followed by acetate utilization or removal. In both methanogenesis zone sediment slurries, formate was low (0–7.2 μM). Lactate increased from zero at 24 h to ∼40 μM at 2 days and remained relatively constant for 14 days in the 13C-acetate slurry D and was below detection in the 13CO2 slurry E at most time points, with the exception of a peak at 3–4 days (52 and 36 μM). In both sediment slurries D and E, sulphate slightly increased from 4 to ∼6 mM and no significant increase in methane above the background occurred during the 14-day incubation.
Interestingly, acetate concentrations detected in all the slurries after the 14-day incubation were mostly similar to those in the original sediment pore water, suggesting that acetate may have reached a steady state (Fig. 1a).
Changes in the prokaryotic community structure of sediment slurries over time as assessed by 16S rRNA gene PCR-DGGE
Changes in the archaeal and bacterial community structures of the sediment slurries were monitored throughout the incubations with 13C-substrates using PCR-DGGE analysis of 16S rRNA genes. During the 14-day incubation, some clear changes in the bacterial 16S rRNA gene DGGE profiles were observed for all 13C-substrates with time compared with the time zero community structure, before 13C-substrate addition. Figure 2a, for example, clearly demonstrates the enrichment of new bacterial species (marked with an asterisk) in slurry C, 1 day after addition of 13C-acetate, which, when sequenced, belonged to the Arcobacter cluster of the Epsilonproteobacteria. Archaeal communities were more difficult to interpret because banding patterns varied slightly at different time points, with some bands disappearing and reappearing with time (see Fig. 2b for example). It seems likely that there was little change in the overall archaeal population with time and the observed banding pattern fluctuations are probably due to stochastic amplification within the PCR, possibly due to the low numbers of archaeal cells within the sediment. Although, between 0.5- and 1-day incubation, with all substrates, there was some evidence of a change in the band intensity for some bands compared with time zero, suggesting the stimulation and growth of some archaeal species. For example, Fig. 2b shows the stimulation of Archaea (marked with asterisk) belonging to the Marine Benthic Group D (MBG-D) in sediment slurry C amended with 13C-acetate.
CsCl density gradient ultracentrifugation for separation of 12C- and 13C-labelled sediment DNA
Separation of 13C-DNA from the unlabelled 12C-DNA was carried out by CsCl-density gradient ultracentrifugation on DNA extracted from the 13C-substrate sediment slurries after 7 days of incubation (Webster et al., 2006b). This incubation time was selected as previously we determined that it provided sufficient incorporation of 13C from low concentrations of substrate into sediment prokaryotic DNA to enable separation of 13C-DNA from the bulk 12C-DNA. DNA extracted from all 13C-labelled sediment slurries after centrifugation in a CsCl–ethidium bromide density gradient was visible as two faint diffuse bands (Fig. 3a). This suggested that the sediment DNA was comprised of unlabelled, partially labelled and fully 13C-labelled DNA, similar to that observed in other SIP experiments (e.g. Morris et al., 2002; Hutchens et al., 2004; Webster et al., 2006b). Two DNA fractions (12C- and 13C-DNA fractions labelled 1 and 2, respectively; Fig. 3a) were collected from the CsCl gradients guided by a ‘marker’ density gradient, containing Desulfobacter sp. 12C- and 13C-DNA.
Molecular analysis of 12C- and 13C-DNA fractions
PCR-DGGE analysis of bacterial and archaeal 16S rRNA genes
Bacterial and archaeal 16S rRNA genes from the 12C- and 13C-DNA fractions from each of the 13C-labelled sediment slurries were amplified by PCR and analysed by DGGE (Figs 3 and 4).
Aerobic zone sediment slurry A
Analysis of the 13C-glucose-amended slurry A (Fig. 3b) showed that the 13C-DNA fraction was a subset of the total bacterial community (12C-DNA fraction), suggesting that the 13C-DNA DGGE profile represented the bacterial population able to utilize 13C-glucose and/or glucose degradation products. Sequencing of two of the most intensely stained DGGE bands from the 13C-DNA fraction revealed that some of the sediment bacteria able to incorporate glucose under oxic conditions belonged to the Gammaproteobacteria and were related to Vibrio species and Idiomarina baltica (both with 97% sequence similarity; Table 1).
Table 1. Bacterial 16S rRNA gene sequence matches to excised DGGE bands from sediment slurries incubated for 7 days with different 13C-substrates
Sediment slurry (DNA fraction)
Nearest match by blastn search (accession number)
% Sequence similarity (alignment length, bp)
Isolation environment of nearest sequence match
Bold DGGE band names highlight sequences retrieved from 13C-DNA fractions (see Fig. 3).
In contrast, the DGGE pattern for the archaeal 13C-DNA fraction (Fig. 4a) was very similar to the 12C-DNA pattern, suggesting that all Archaea identified in the aerobic sediment slurry A by PCR-DGGE were able to incorporate 13C-glucose or its degradation products. All dominant archaeal bands that were excised and sequenced belonged to the largely uncultivated Archaea group Marine Group 1 (MG1) and had 97–99% sequence similarity to archaeal 16S rRNA gene sequences retrieved from a marine sponge, seawater and marine sediments (Table 2).
Table 2. Archaeal 16S rRNA gene sequence matches to excised DGGE bands from sediment slurries incubated for 7 days with different 13C-substrates
Sediment slurry (DNA fraction)
Nearest match by blastn search (accession number)
% Sequence similarity (alignment length, bp)
Isolation environment of nearest sequence match
Bold DGGE band names highlight sequences retrieved from 13C-DNA fractions (see Fig. 4).
PCR-DGGE analysis of the bacterial 16S rRNA genes in the 13C-DNA fraction from the 13C-glucose-amended slurry B incubated under anaerobic conditions showed that the 13C-DNA bacterial community structure was different from the 12C-DNA bacterial population. However, unlike the aerobic slurry A (Fig. 3b), the 13C-DNA bacterial population in slurry B (Fig. 3c) was more complex and dominated by a number of distinct DGGE bands, suggesting that under anaerobic conditions, a more diverse population of sediment bacteria are able to utilize glucose and/or its degradation products. Sequencing a number of DGGE bands revealed that 13C was incorporated by bacteria within the phyla Gamma- and Epsilonproteobacteria. Gammaproteobacterial sequences were related to Vibrio species (97–99% sequence similarity) and Marinobacter aquaeolei (97% sequence similarity), while the sequence (band Fb6) belonging to the Epsilonproteobacteria had 99% sequence similarity to clone MZ-53.NAT from a coastal near-surface sediment (Table 1) and was also closely related (99% sequence similarity) to the nitrogen-fixing Arcobacter nitrofigilis isolated from a salt marsh plant root (McClung et al., 1983) and Arcobacter sp. strain NA105 isolated from a tidal flat sediment of the Wadden Sea (Freese et al., 2008).
Similar to slurry A, all Archaea sequences identified in the anoxic slurry B, which were able to incorporate 13C from glucose, belonged to the MG1 (Fig. 4b; Table 2). However, in contrast to slurry A, the MG1 within slurry B were an active subset of the total archaeal population. All MG1 sequences were closely related (97–100%) to clone sequences identified previously in other marine environments and to sequences identified in the aerobic slurry A.
Sulphate reduction zone sediment slurry C
Analysis of the bacterial 16S rRNA genes from the 12C- and 13C-DNA fractions from the sulphate reduction zone slurry C amended with 13C-acetate showed a considerable similarity in their DGGE profiles (Fig. 3d), suggesting that all members of the dominant bacterial community were able to incorporate acetate under these conditions. Both 12C- and 13C-DNA DGGE profiles were dominated by one brightly stained band (bands Sb1 and Sb2) related to the seawater clone VH-FL6-38 (97–98% sequence similarity) within the Arcobacter cluster of the Epsilonproteobacteria. In addition, a small number of other bands in the 13C-DNA DGGE profile (e.g. band Sb3) were brighter than their corresponding bands in the 12C-DNA profile. Band Sb3 was most similar to sequences belonging to the Deltaproteobacteria order Desulfuromonadales and was 96% similar to the Fe(III)- and Mn(IV)-reducing bacterium, Geoalkalibacter subterraneus (Greene et al., 2009).
Archaeal sequences shown to be active and incorporating 13C-acetate or metabolites within sediment slurry C were a discrete subset of the 12C diversity and these belonged to members of the C3 (Inagaki et al., 2006) subgroup of the uncultivated Miscellaneous Crenarchaeotic Group (MCG; Inagaki et al., 2003), whereas DGGE bands excised and sequenced from the DGGE profile of the 12C-DNA fraction demonstrated that the total archaeal community within this slurry also contained members of the MBG-D in addition to MCG (identified by band position; see Fig. 4c).
Methanogenesis zone sediment slurries D and E
Sediment slurries from the methanogenesis zone (Fig. 1) were individually incubated with the substrates 13C-acetate (slurry D) and 13CO2 and (slurry E). PCR-DGGE analysis of bacterial 16S rRNA genes demonstrated that a large diversity of bacterial species were able to use 13CO2 or metabolites (Fig. 3f) in slurry E, and a more specific community was able to utilize 13C-acetate (Fig. 3e) in slurry D. Interestingly, a different subset of Bacteria seemed to be able to utilize acetate under low-sulphate conditions (∼4 mM) than those in the high-sulphate sediment slurry C (∼22 mM), with the exception of band Mb1. Band Mb1 was identical to the Epsilonproteobacteria (Arcobacter-like) sequence identified in slurry C (Table 1). It should be noted that because there was a high diversity of bacteria able to incorporate 13C from 13CO2, represented by many faint DGGE bands (Fig. 3f), it was difficult to excise DGGE bands and obtain good sequence information, and therefore no sequence data are presented for slurry E.
However, in contrast to Bacteria, Archaea (neither methanogens nor uncultivated lineages) were not detected in the 13C-DNA fraction of either slurry D or E, suggesting that Archaea identified in the 12C-DNA fractions (Fig. 4d and e) were not active and/or were unable to incorporate 13CO2 or 13C-acetate under the conditions or the incubation time used in this experiment. Nevertheless, presumed active methanogenic Archaea were readily identified in these sediments. For example, methanogen mcrA functional genes were amplified from Severn Estuary DNA extracted from the same sediment zone, Euryarchaeaota 16S rRNA genes were detected in slurries D and E (data not shown) and methanogens have been enriched from the same site using a range of substrates including acetate and H2/CO2 (A.J. Watkins, H. Sass & R.J. Parkes, unpublished data).
Phylogenetic analysis of bacterial and archaeal 16S rRNA genes, and dsrA genes
Because of the importance of sulphur cycling and sulphate reduction within marine sediments in general (Jørgensen, 1982; Muyzer & Stams, 2008) and that documented in a previous study on Severn Estuary sediments (Wellsbury et al., 1996), further investigation of sediment slurry C (predominantly from a zone associated with sulphate reduction; Fig. 1) was carried out. Gene libraries (Bacteria and Archaea 16S rRNA genes, and dsrA genes) were constructed from the 12C- and 13C-DNA fractions.
Analysis of the bacterial 16S rRNA gene libraries (Fig. 5) did not identify any sequences related to known sulphate-reducing bacteria (SRB), although sequences related to other sulphur cycling bacteria were present. For example, in the 13C-DNA library (n=29), a large number of sequences fell within the Arcobacter cluster of the Epsilonproteobacteria (52%), some of which are known to oxidize sulphide to sulphur (Telang et al., 1999; Gevertz et al., 2000; Wirsen et al., 2002), and 31% fell within the Deltaproteobacteria order Desulfuromonadales, which contains known sulphur- and metal-reducing bacteria. Representatives of these dominant groups of sequences (e.g. clones 13CSRZ-B1/13CSRZ-B2 and 13CSRZ-B17/13CSRZ-B19; Fig. 5) were also closely related (96–97% and 93–100% sequence similarity) to sequences identified by PCR-DGGE (Fig. 3d; Table 1). In addition, other sequences included members of the sulphur-oxidizing Epsilonproteobacteria family Thiovulgaceae (10%; Campbell et al., 2006) as well as members of the Bacteroidetes and a novel group related to Chlorobi.
The majority of 13C-DNA archaeal 16S RNA gene sequences (95%, n=20; Fig. 6) from slurry C fell within the Crenarchaeota MCG, with 65% grouping within the Marine Benthic Group C (MBG-C) and 30% within the C3 (Fig. 6a). Additionally, one sequence belonged to the Euryarchaeota group MBG-D (Fig. 6b). Such a limited diversity of archaeal phylotypes that utilized 13C-acetate and were active within sediment slurry C is interesting considering that a much higher diversity of archaeal sequences (n=23) was obtained from the 12C-DNA. For example, the archaeal community identified from the 12C-DNA also contained sequences belonging to a number of other clusters within the MCG along with members of the MG1 (Crenarchaeota Group 1.1a), Marine Hydrothermal Vent Group and the Crenarchaeota Group 1.1b. Similar to the bacterial 16S rRNA gene library, some archaeal sequences identified by PCR cloning of the 13C-DNA were closely related to sequences detected by PCR-DGGE (e.g. clone 13CSRZ-A15 was 91–94% similar to bands Sa1–Sa3; Figs 4c and 6; Table 2).
Phylogenetic analysis of 10 partial sequences from the 13C-DNA fraction of sediment slurry C revealed a very low diversity of novel dsrA genes (data not shown). All 10 novel dsrA gene sequences were similar to each other (98–100% sequence similarity) and were related (82–83% sequence similarity) to a clone sequence (clone DSR-W) retrieved from a deep-sea hydrothermal sediment from the Rainbow (Mid-Atlantic Ridge) vent field (Nercessian et al., 2005), whereas 10 sequences analysed from the 12C-DNA fraction showed a much higher level of diversity and included sequences that were related (84–90% sequence similarity) to known SRB within the family Desulfobacteraceae and to other novel dsrA genes (93–99% sequence similarity) previously identified in estuarine (Joulian et al., 2001; Leloup et al., 2006), hydrothermal (Dhillon et al., 2003) and salt marsh sediments (Bahr et al., 2005), as well as a sulphate-reducing sediment slurry (Webster et al., 2006b).
The repeated addition of low concentrations of 13C-substrates (5 × 100 μM) to marine sediment slurries (Severn Estuary tidal flat), without the use of prior enrichment of active prokaryotes as used previously in SIP of sediment slurries by Webster et al. (2006b), resulted in clearly detectable 13C-incorporation. This was despite using relatively short-term incubations (up to 7 days) under a variety of prokaryotic metabolic conditions and with sediment slurries from different biogeochemical zones. Analysis of prokaryotic 16S rRNA genes in all sediment slurries demonstrated rapid enrichment (within 1 day) of the original community after incubation with 13C-substrates, with limited changes in the total prokaryotic diversity. Therefore, the repeated addition of low concentrations of 13C-substrates for up to 4 days, followed by further 10 days of incubation caused very little changes in the prokaryotic diversity, and hence the prokaryotes detected by SIP should reflect the active prokaryotic community of the sediment zone under the conditions used. The results of geochemical analysis of the slurries demonstrated that the desired conditions were achieved in some cases [e.g. lack of fermentation products and an increase in nitrite and nitrate in the aerobic zone slurry A with 13C-glucose; fermentation products in the anaerobic (dysaerobic zone) slurry B with 13C-glucose and no sulphate removal]. However, in the sulphate reduction zone sediment slurry C and methanogenesis zone slurries D and E no net sulphate removal or methane production occurred, demonstrating that in these slurries, terminal-oxidizing processes did not dominate and that activities could have been similar to those occurring in situ (Wellsbury et al., 1996) and/or that any sulphide or methane produced may have been reoxidized. This, however, provided an opportunity to detect nonterminal-oxidizing prokaryotes that are active under the incubation conditions and/or those prokaryotes involved when sedimentary zones are impacted by sediment disturbance, preventing the dominance of anaerobic terminal-oxidizing prokaryotes, which is common in the dynamic Severn Estuary (Yallop & Paterson, 1994).
Glucose utilization in marine sediment slurries
The addition of 13C-glucose to both aerobic (slurry A) and anaerobic (slurry B) sediment slurries showed that similar sediment bacteria, Vibrio species, were able to incorporate 13C-glucose and/or its degradation products under both these conditions. It is well documented that Vibrio species are facultative anaerobes that are often isolated from marine environments (Freese et al., 2009) and readily incorporate glucose under both anaerobic and aerobic conditions (Alonso & Pernthaler, 2005). It has been suggested that because of the facultative nature of Vibrio species, they are perfectly adapted to survive in the oxic–anoxic zones of tidal sediments (Alonso & Pernthaler, 2005). The demonstration in this study that these bacteria can be stimulated quickly and incorporate added glucose carbon under both oxic and anoxic sedimentary conditions supports this suggestion. Interestingly, a 16S rRNA gene phylotype was also detected in the aerobic sediment slurry A, which was similar to I. baltica, a strict aerobe that can grow poorly on glucose, but can grow better on acetate (Brettar et al., 2003; Martínez-Cánovas et al., 2004). Therefore, it is likely that a bacterium related to Idiomarina species is incorporating 13C from glucose/or metabolites and conducting a similar metabolism in slurry A. Similarly, under anaerobic conditions (dysaerobic zone slurry B), one sequence was also obtained that was related to Marinobacter, a ubiquitous marine bacterial genus, capable of anaerobic growth. Some pure cultures of Marinobacter species are known to utilize glucose under these conditions (e.g. Marinobacter salsuginis; Antunes et al., 2007) and other species are known to grow on glucose fermentation products such as acetate and lactate (Sass et al., 2001; González & Whitman, 2006).
The dominance of Gammaproteobacteria within the bacterial population of these two 13C-glucose-amended sediment slurries (aerobic zone slurry A and dysaerobic zone slurry B) is consistent with this subphyla often being dominant in the surface of coastal and tidal sediments (Wilms et al., 2006; Edlund et al., 2008; Kim et al., 2008). However, the dominance of the archaeal group MG1 within the active archaeal population from these slurries is somewhat surprising because there are no reports of this group of largely uncultured Archaea being able to utilize glucose, despite reports that MG1 Archaea are phylogenetically diverse and ubiquitous in marine sediments and the overlying water column (Vetriani et al., 1999; Francis et al., 2005; Teske & Sørensen, 2008; Roussel et al., 2009b). The only cultured representative of the MG1 is an aerobic, chemolithoautotrophic ammonia-oxidizing archaeon (Könneke et al., 2005), although the ability of some marine Crenarchaeota to actively take up amino acids (Ouverney & Fuhrman, 2000; Herndl et al., 2005) and evidence that the carbon isotopic composition of MG1 lipids shows some degree of organic carbon assimilation suggest that some members of MG1 have a heterotrophic or a mixotrophic metabolism (Ingalls et al., 2006). Alternatively, rapid assimilation of 13C-glucose by Gammaproteobacteria within these sediment slurries could be providing metabolites, including 13CO2, which might be fixed by autotrophic MG1 species, particularly those in the aerobic slurry A, where there was evidence of ammonia oxidation.
Acetate utilization in anaerobic marine sediment slurries
As sulphate reduction is a very important process within marine sediments (Muyzer & Stams, 2008), including Severn Estuary sediments (Wellsbury et al., 1996), a greater focus in terms of molecular analysis was on the sulphate reduction zone sediment slurry C incubated with 13C-acetate, an important substrate for sulphate reduction in marine sediments (Parkes et al., 1989). Consistent with the short incubations times and the absence of net sulphate removal, no incorporation of 13C-acetate into known terminal-oxidizing SRB was detected. This may be expected as even in active sulphate-reducing sediment zones, SRB only represent a relatively small proportion of the total bacterial population (e.g. up to 11%, Mußmann et al., 2005; average of 13%, Leloup et al., 2009) and are sometimes not even detected by molecular approaches (Parkes et al., 2005). However, they can be detected by SIP in sediment slurries pre-enriched for sulphate reduction (Webster et al., 2006b), but under these conditions, the pre-enrichment results in considerable changes in the prokaryotic community. In our nonenriched sulphate reduction zone slurry C, SRB were only detected using the more selective and sensitive dsrA functional gene, including sequences related to known SRB in the 12C-DNA fraction and some novel dsrA sequences within the active 13C-DNA fraction. In addition, other sulphur cycle prokaryotes were also found to have incorporated 13C-acetate. For example, sequences belonging to the Deltaproteobacteria order Desulfuromonadales were detected in the 13C-DNA by both DGGE and 16S RNA gene library analysis. Pure cultures of the genus Desulfuromonas, such as Desulfuromonas acetoxidans and Desulfuromonas palmitatis are known sulphur reducers that are capable of oxidizing acetate and using it as a sole source of carbon. It is intriguing that a large number of sequences that belong to the epsilonproteobacterial Arcobacter and Thiovulgaceae clusters were also found (Fig. 5). Epsilonproteobacteria have been increasingly recognized as important bacteria involved in sulphur-dependent biogeochemical cycles and are globally ubiquitous in marine and terrestrial environments (Campbell et al., 2006). For example, deep-sea hydrothermal fields (Nakagawa et al., 2005), sulphidic caves (Porter & Engel, 2008) and symbiotic associations (Urakawa et al., 2005) have epsilonproteobacterial populations. Representative cultured members of the Arcobacter cluster and the Thiovulgaceae are often chemolithoautotrophic and oxidize sulphide to sulphur under microaerophilic conditions and/or with nitrate as the electron acceptor (Gevertz et al., 2000; Wirsen et al., 2002; Kodama & Watanabe, 2004). Our results suggest that within the sulphate reduction zone sediment slurry C sulphur cycling may be occurring between populations of novel sulphate reducers, novel sulphur/sulphide-oxidizing Epsilonproteobacteria, sulphur-reducing Deltaproteobacteria and other bacteria similar to the sulphur cycling occurring in sulfureta (Postgate, 1979) and defined mixed cultures (Biebl & Pfennig, 1978; Telang et al., 1999). This is consistent with no net sulphate removal occurring in this slurry.
Many cultured sulphur-oxidizing members of the Arcobacter cluster do not utilize acetate under microaerophilic or anaerobic conditions (Gevertz et al., 2000; Wirsen et al., 2002), and therefore, it seems unlikely that the Arcobacter-like sequences detected in this study have the same metabolic restrictions as these previously cultured species. This is because the Arcobacter species in sediment slurry C seem to be the dominant members of the community (by 13C-DNA 16S rRNA gene libraries and PCR-DGGE) and that they are quickly stimulated during incubation (within 1 day; Fig. 2a), suggesting that they are actively utilizing 13C-acetate and not relying on 13CO2 produced from other bacteria within the sediment slurry. It is possible that these uncultured Arcobacter group members are able to utilize acetate for sulphur, metal or nitrate reduction and/or incorporate acetate carbon for cell synthesis. For example, some strains of the nitrogen-fixing bacterium A. nitrofigilis are able to utilize acetate and grow anaerobically with nitrate (McClung et al., 1983), and sequences belonging to uncultured members of the Arcobacter group have been implicated in manganese reduction in Black Sea shelf sediments coupled with oxidation of acetate (Thamdrup et al., 2000). Interestingly, low concentrations of dissolved Fe and Mn were shown to increase with time (3–7 days of incubation) in slurry C (data not shown), indicating that very low rates of metal reduction had occurred. However, significant Arcobacter biomass had already been observed by 1 day (see Fig. 2a) and therefore the detected metal reduction within the slurry may have been carried out by other prokaryotes, such as members of the Desulfuromonadales (Lovley, 1993).
A range of metabolisms within this slurry may be associated with the physical mixing of the original sediments, bringing previously spatially separated electron donors and acceptors into close proximity, facilitating their utilization by fast-growing, acetate-oxidizing/incorporating Arcobacter species. In addition, the closely related Arcobacter species strain NA105, isolated from tidal flat sediments of the Wadden Sea (Freese et al., 2008), can reduce dimethylsulphoxide and trimethylamine oxide (TMAO) (H. Sass, unpublished data), and TMAO and dimethylsulphoxide reductases are also found in the Arcobacter butzleri genome (Miller et al., 2007). Moreover, recently, several SIP studies indicate that syntrophic acetate oxidation in some environments (soil and lake sediment) can be carried out by nonacetogenic bacteria such as Geobacter, Syntrophus, other Deltaproteobacteria, Betaproteobacteria and Nitrospira when electron acceptors are limited (Chauhan & Ogram, 2006; Schwarz et al., 2007).
As was observed in the aerobic (slurry A) and dysaerobic (slurry B) zone sediment slurries with 13C-glucose, the active archaeal population determined by analysis of the 13C-DNA from the sulphate reduction zone sediment slurry C comprised of sequences derived entirely from uncultivated groups of Archaea. Both PCR-DGGE and PCR-cloning of archaeal 16S rRNA genes showed that the active Archaea able to utilize 13C-acetate belonged to members of the diverse MCG. The MCG are one of the predominant archaeal groups in 16S rRNA gene libraries of deep subsurface sediments (Fry et al., 2008; Teske & Sørensen, 2008), although they have also been found in other environments including terrestrial (Chandler et al., 1998) and coastal marine surface sediments (Roussel et al., 2009a). The term ‘miscellaneous’ within the name reflects the diverse habitat range and phylogenetic diversity of sequences that make up the MCG (e.g. see Fig. 6a). Given the MCG's substantial sequence diversity, the identification of distinct MCG subgroups and their dominance in marine sediments, it is not unreasonable to suggest that some members of this group can incorporate 13C-acetate under anaerobic conditions, considering the importance of acetate as an anaerobic substrate. In addition, carbon isotopic signatures of archaeal cells and polar lipids from MCG-dominated subsurface sediments suggest that these archaeal populations are able to utilize buried organic carbon (Biddle et al., 2006), and hence, that most members of the MCG are heterotrophic, an inference directly supported by our results. Interestingly, although MCG sequences were abundant in the 13C-DNA 16S rRNA gene libraries from the sulphate reduction zone slurry, no sequences belonging to MG1 Archaea were detected in this fraction despite their presence in the 12C-DNA library (Fig. 6a). This suggests that MG1 are not able to utilize 13C-acetate carbon or were not active within the sulphate reduction zone sediment slurry C.
As with sediment slurry C, no terminal-oxidizing prokaryotes, in this case methanogens, were identified in the methanogenesis zone sediment slurries D and E by 16S rRNA gene analysis incubated with 13C-acetate or 13CO2, respectively. However, a diverse range of Bacteria were rapidly stimulated (within 1 day) and had incorporated 13C by 7 days (Fig. 3e and f). Also, like sediment slurry C with 13C-acetate, Arcobacter-related sequences were present in the methanogenesis zone sediment slurry D, but interestingly, they did not dominate the 13C-DNA DGGE profile. This suggests that Arcobacter were less active in the methanogenesis zone sediment slurry, possibly as their electron acceptors were less abundant in deeper sediments and/or sulphur cycling was less prevalent in this low-sulphate sediment slurry. Interestingly, as the lower potential energy conditions of these methanogenesis zone slurries were the only conditions where no 13C incorporation into Archaea occurred, this may suggest that Archaea were not as active under these conditions compared with the other, potentially higher energy slurry conditions (slurries A, B and C). This was surprising because Archaea have been suggested to be low-energy specialists (Valentine, 2007). However, it would be interesting to know whether 13C-incorporation would have occurred with increased incubation times and/or under methane-producing conditions.
Implications of DNA-SIP for the study of sedimentary processes
This study shows that DNA-SIP is a very useful tool to study uncultivated groups of prokaryotes involved in different biogeochemical sedimentary processes, without the need for prior enrichment of specific prokaryotic populations. However, under these conditions, it is difficult to detect terminal-oxidizing SRB and methanogens, probably due to their low cell numbers and activities, unless more selective functional genes are used. It should also be noted that PCR-DGGE and analysis of only a limited number of 16S rRNA gene clones were used in this study and, therefore, it is probable that other less abundant prokaryotes including terminal-oxidizers would have been detected with more extensive sampling (Quince et al., 2008), for example through use of high-throughput sequencing technologies (Tringe & Hugenholtz, 2008) such as pyrosequencing (Huber et al., 2007).
Multiple, low concentrations of 13C-substrate additions and relatively short incubations (7 days) seemed successful in restricting changes in the prokaryotic community, while achieving significant 13C incorporation (confirmed by 13C analysis of PLFA; J. Rinna, G. Webster, A.J. Weightman & R.J. Parkes, unpublished data). This is important because longer incubation times are often cited as a disadvantage of DNA-SIP because of cross-feeding from original 13C-substrates and changes in the community structure (Neufeld et al., 2007b). Some studies have, therefore, used very short incubation times and reported 13C incorporation into DNA and SSU rRNA of some sedimentary bacteria within 1–2 h (Gallagher et al., 2005; MacGregor et al., 2006). However, such very short incubations may not have resulted in 13C-labelling of some of the uncultured Bacteria and Archaea detected in this study. The use of 15N-labelling techniques as used for slow-growing anaerobic methane-oxidizing communities (Krüger et al., 2008), in combination with DNA-SIP (Buckley et al., 2007) and/or RNA-SIP (Manefield et al., 2002), may further improve our approach. Also, in DNA-SIP, it has been shown to be difficult to completely separate 12C- and 13C-DNA fractions with confidence (Neufeld et al., 2007a). In the present study, the absence of archaeal PCR products in the 13C-DNA fractions of the methanogenesis zone sediment slurries D and E (Fig. 4d and e) demonstrates that there was no smearing of 12C-archaeal DNA into the 13C-DNA fraction, and hence, that our procedures enabled good 13C incorporation and 12C- and 13C-DNA separation.
This study demonstrates that although it may be a difficult challenge to use DNA-SIP to understand complex prokaryotic communities and processes in marine sediments, it is a very useful tool to identify the activities of uncultured groups of sediment Archaea and Bacteria, the substrates that they incorporate and the conditions under which they are active. Our results highlight that several groups of uncultured prokaryotes play important ecological roles in carbon and sulphur cycling of tidal sediments of the Severn Estuary, specifically providing new metabolic information for uncultured groups of Archaea (e.g. MG1, MCG) and suggesting broader metabolisms for Bacteria with limited cultured representatives (e.g. Arcobacter species).
G.W. and J.R. were funded by the NERC Marine and Freshwater Microbial Biodiversity programme research grant numbers NER/T/S/2000/636 and 2002/00593. The work was also supported by the European Union contract number EVK3-CT-1999-00017, and NERC NE/F018983/1 and NE/F00477X/1. The authors would like to thank Mr Stephen Hope (Cardiff University) for technical support with DNA sequencing, Professor Richard Evershed and Dr Richard Pancost (University of Bristol) for advice and kind use of their analytical facilities and Dr Henrik Sass for helpful comments during the preparation of this manuscript.