Microbial diversity in Cenozoic sediments recovered from the Lomonosov Ridge in the Central Arctic Basin

Authors

  • Stephanie R. Forschner,

    1. Department of Biomedical and Pharmaceutical Sciences, College of Pharmacy, University of Rhode Island, 41 Lower College Road, Kingston, RI 02881, USA.
    Search for more papers by this author
  • Roberta Sheffer,

    1. Department of Biomedical and Pharmaceutical Sciences, College of Pharmacy, University of Rhode Island, 41 Lower College Road, Kingston, RI 02881, USA.
    Search for more papers by this author
  • David C. Rowley,

    1. Department of Biomedical and Pharmaceutical Sciences, College of Pharmacy, University of Rhode Island, 41 Lower College Road, Kingston, RI 02881, USA.
    Search for more papers by this author
  • David C. Smith

    Corresponding author
    1. Graduate School of Oceanography, University of Rhode Island, 215 South Ferry Road, Narragansett, RI 02882, USA.
    Search for more papers by this author

*E-mail dcsmith@gso.uri.edu; Tel. (+1) 401 874 6172; Fax (+1) 401 874 6240.

Summary

The current understanding of microbes inhabiting deeply buried marine sediments is based largely on samples collected from continental shelves in tropical and temperate latitudes. The geographical range of marine subsurface coring was expanded during the Integrated Ocean Drilling Program Arctic Coring Expedition (IODP ACEX). This expedition to the ice-covered central Arctic Ocean successfully cored the entire 428 m sediment stack on the Lomonosov Ridge during August and September 2004. The recovered cores vary from siliciclastic sediment low in organic carbon (< 0.2%) to organic rich (∼3%) black sediments that rapidly accumulated in the early middle Eocene. Three geochemical environments were characterized based on chemical analyses of porewater: an upper ammonium oxidation zone, a carbonate dissolution zone and a deep (> 200 m below sea floor) sulfate reduction zone. The diversity of microbes within each zone was assessed using 16S rRNA phylogenetic markers. Bacterial 16S rRNA genes were successfully amplified from each of the biogeochemical zones, while archaea was only amplified from the deep sulfate reduction zone. The microbial communities at each zone are phylogenetically different and are most closely related to those from other deep subsurface environments.

Introduction

The subsurface biosphere is the largest and most widespread reservoir of living biomass and yet remains one of the least understood. Evidence for active microbial communities existing in sediments to depths of ∼800 m below seafloor (mbsf) (Wellsbury et al., 2002) have led to estimates of the total microbial biomass in the subseafloor habitat to comprise ∼10% of the total biomass on Earth (Parkes et al., 2000). The depth to which microorganisms appear to be metabolically active has recently been extended to ∼1600 mbsf in sediments that were deposited ∼111 million years ago (Ma) (Roussel et al., 2008). The composition of the microbial communities residing in these deeply buried marine sediments has been investigated by 16S rRNA gene sequencing (Kormas et al., 2003; D'Hondt et al., 2004; Sørensen et al., 2004; Inagaki et al., 2006; Teske and Sørensen, 2008) and fluorescent in situ hybridization studies (Mauclaire et al., 2004; Schippers et al., 2005; Biddle et al., 2006). 16S rRNA clone libraries constructed with environmental DNA from deeply buried marine microbial communities have yielded some common results (reviewed in Smith and D'Hondt, 2006). Archaea are often represented by members of the Deep Sea Archaeal Group, Miscellenous Crenarchaeal Group (MCG), Marine Crenarchaeotic Group I, Marine Benthic Group and the South African Gold Mine Euryarchaeotic Group (reviewed in Teske and Sørensen, 2008). Bacterial lineages that are well represented in the clone libraries include Proteobacteria, Bacteroidetes, Planctomycetes, Chloroflexi, and the JS1 division. It has been demonstrated that sites that are widely separated geographically but share geochemical characteristics harbour similar microbial communities (e.g. hydrate bearing sediments: Inagaki et al., 2006).

Much of our understanding of the extent of microbial biomass and community composition is a result of coring in tropical and temperate regions of the globe. With greater than 90% sea surface ice coverage, the Arctic Ocean basin is the least studied of all ocean basins due to its inaccessibility. Consequently, the underlying sediment microbial communities have remained unexplored. In 2004, the Integrated Ocean Drilling Program Arctic Coring Expedition (IODP Expedition 302) successfully cored the entire sediment stack (428 m) on the Lomonosov Ridge that separates the American and Urasian Basins in the central Arctic. This ridge broke away from the Eurasian continental margin ∼57 Ma ago and subsided ∼1200 m below the surface of the sea. The ridge has been accumulating sediments during this period with the exception of two hiatuses (Backman et al., 2008). The ridge is now located ∼250 km from the North Pole. The primary impetus for coring this ridge was to reconstruct the paleoclimatic record of the Arctic region (Moran et al., 2006).

On the basis of pore water analysis, three main geochemical units were defined for the sediment column: an upper ammonium oxidation zone, a carbonate dissolution zone and a deep (> 200 mbsf) sulfate reduction zone (Expedition 302 Scientists, 2006). The porewater chemistry reflects the productivity in the region over geological time scales. The sediment in the deep sulfate reduction zone were deposited in the Paleogene when the productivity of the region, then warm and ice-free, was estimated to be on the order of 50–100 gC m−2 year−1 and export production was high (Knies et al., 2008). This contrasts with the other two zones that were deposited during the Neogene, a period when the Arctic was ice-covered and experienced relatively low production (< 20 gC m−2  year−1) (Knies et al., 2008).

Molecular analyses of subseafloor microbial communities have proven to be technically challenging due to low biomass and the presence of co-extracted PCR inhibitors. Coupled with limited sample volumes, these challenges have spawned much research on efficient methods to obtain high quality DNA and PCR products from subsurface sediments. Some solutions have included the addition of poly A to block the binding of released DNA to sediment particles (Webster et al., 2003), diluting DNA 10- to 1000-fold to overcome PCR inhibition (Kormas et al., 2003), and gel purification (Sørensen et al., 2004).

This study capitalized on the unique sediment cores obtained by IODP ACEX to investigate microbial communities residing in deep subsurface habitats below the polar ice cap. Multiple approaches were taken to assess the biodiversity from each of the three identified geochemical environments. A particularly successful strategy used whole-genome amplification of environmental DNA followed by PCR amplification (WGA-PCR) to obtain otherwise elusive amplicons.

Results

Extraction and purification of DNA from deep subsurface sediments

One sediment sample from each of the three geochemical zones was chosen for phylogenetic studies (Table 1). The sample at 55 mbsf was obtained from the upper ammonium oxidation zone while samples 103 and 242 mbsf were acquired from the carbonate dissolution and deep sulfate reduction zones respectively. Great efforts were taken to optimize the DNA extraction and purification procedures due to the irreplaceable and limited nature of the samples, as well as the anticipated difficulties in extracting amplifiable DNA from sediments (Webster et al., 2003). Of several commercially available soil extractions kits tested, FastDNA SPIN Kit for Soils (Qbiogene) yielded 2–4 times more DNA using the modifications described in the methods (Webster et al., 2003). The extraction protocol resulted in the co-extraction of other UV absorbing molecules, as determined by 260 nm/230 nm UV absorbance ratios averaging 0.5, indicating the possible presence of PCR inhibiting compounds. It was anticipated that the poor purity of the DNA would compromise subsequent PCR reactions. Therefore, the genomic material was purified using ChromaSpin TE-100 columns (Clontech), a size-exclusion gel-filtration technique, resulting in the acquisition of significantly higher quality DNA with 260 nm/280 nm and 260 nm/230 nm ratios around ∼2.

Table 1.  Physical descriptions and cell counts of sediments from the three depths.
Depth (mbsf)Sediment age (Ma)aCell counts (per cm3)bTotal organic carbon (%)cLithologyTemperature (°C)d
553.82.3 × 107 ± 0.60.26Silty clay1.44
1037.12.7 × 107 ± 0.90.27Silty clay2.90
242461.2 × 107 ± 0.12.93Diatom ooze7.14

Amplification of 16S rRNA genes

PCR amplification of 16S rRNA genes was attempted with the purified genomic materials using universal bacterial and archaeal primers (Table 2). The anticipated products were only observed using bacterial primers with the 55 and 103 mbsf samples (Table 3). Due to a lack of products from the deepest sample and limited quantities of the obtained environmental DNA, non-specific WGA was undertaken to increase the copy number of the targeted genes. Whole-genome amplification has been previously used to enable the amplification of microbial DNA from low biomass and contaminated soils (Gonzalez et al., 2005; Abulencia et al., 2006). Subsequent PCR reactions with universal archaeal and bacterial primers on the WGA materials resulted in amplification of bacterial products from all three samples and archaeal amplicons from the 242 mbsf genetic material (Table 3).

Table 2.  Primers used in this study for amplification and sequencing of the bacterial and archaeal 16S rRNA genes.
Amplification targetPrimer nameSequence (5′−3′)Reference
  1. References indicate the studies when the primers were used in combination and the PCR conditions employed, not necessarily the original design of the primers.

Bacterial 16S rRNA gene27FAGA GTT TGA TCC TGG CTC AGLane (1991)
1392RACG GGC GGT GTG TRC 
Bacterial 16S rRNA geneBAC 907RCCC GTC AAT TCC TTT GAG TTTKormas et al. (2003)
Archaeal 16S rRNA geneArch21FTTC CGG TTG ATC CYG CCG GADeLong (1992)
Arch958RYCC GGC GTT GAM TCC AAT T 
Archaeal 16S rRNA geneARC-8FTCC GGT TGA TCC TGC CTeske et al. (2002)
1492RGGC TAC CTT GTT ACG ACT T 
Archaeal 16S rRNA gene109FACK GCT CAG TAA CAC GTGrosskopf et al. (1998)
915RGTG CTC CCC CGC CAA TTC CT 
Archaeal 16S rRNA genePRA46FYTA AGC CAT GCR AGTOvreas et al. (1997)
PREA1100RYGG GTC TCG CTC GTT RCC 
Archaeal 16S rRNA geneArch21FTTC CGG TTG ATC CYG CCG GAKormas et al. (2003)
ARC-1390RGAC GGG CGG TGT GTG CAA 
Table 3.  Results from PCR experiments with various primers.
PrimersPCR amplification
55 mbsf103 mbsf242 mbsf
PCRWGA-PCRPCRWGA-PCRPCRWGA-PCR
  • *

    Positive amplification (+), no amplification (−), products subsequently cloned ().

Bacterial 16S      
 27F/1392R+*+*+*+*+*
Archaeal 16S      
 Arch21F/Arch958R+
 Arch21F/ARC-1390R   +*
 ARC-8F/1492R   +*
 109F/915R   +
 PRA46F/PREA1100R   

Archaeal detection bias as a result of primer choice is a concern in studies of deep subsurface microbial communities (Teske and Sørensen, 2008). To better sample the full diversity of archaea and to limit the effects of primer bias due to mismatch, PCR amplification was attempted with multiple combinations of archaeal primers (Table 2). None of the primer sets produced any evidence of archaea in sample 55 or 103. PCR products of the anticipated size were obtained in sample 242 using primers shown to be more successful for detecting methanogenic archaea (Banning et al., 2005) (Table 3).

Assessment of microbial biodiversity

A total of seven clone libraries were constructed for the three samples (Table 3). Bacterial libraries included 16S PCR products using the native genomic material from 55 and 103 mbsf, as well as the WGA-PCR products from all three sites. The archaeal libraries were prepared using WGA-PCR products from the 242 mbsf sample with primers 8F and 1492R, and also using primers 21F and 1390R. Restriction fragment length polymorphism (RFLP) analysis revealed multiple unique phylotypes in each library. The bacterial clone libraries from the PCR with no prior WGA resulted in two unique phylotypes from 16 clones for sample 55 mbsf and 3 from 16 clones in sample 103 mbsf. The WGA-PCR clone libraries resulted in 88 clones with 6 unique phylotypes, 118 clones with 4 phylotypes, and 76 clones with 8 phylotypes for samples 55 mbsf, 103 mbsf and 242 mbsf respectively. One phylotype from sample 103 mbsf was recovered independently in both the WGA-PCR library and the non-preamplified PCR library. The archaeal 16S clone libraries using primers 21F and 1390R resulted in five unique phylotypes from a total of 165 clones analysed. Primers 8F and 1492R yielded 3 phylotypes from 49 clones with two being unique from those found with primers 21F and 1390R. Rarefaction curves generated from the RFLP analysis suggested that most of the easily accessible diversity was studied (Fig. 1). Because the rarefaction curves did not reach saturation, more exhaustive sampling of clone libraries would likely have revealed additional phylotypes. In particular, the two libraries generated by direct PCR amplification of purified environmental DNA are expected to possess higher diversity than what was sequenced. The highest bacterial diversity was observed in sample 242 mbsf and interestingly was the only depth where archaea were found (Table 4). Unique phylotypes were sequenced and submitted to the GenBank database under accession numbers EU915212EU915240.

Figure 1.

Rarefaction curves for bacterial and archaeal clone libraries. Legend names beginning with ‘B’ indicate bacterial clone libraries while that with ‘A’ indicate archaeal clone libraries.

Table 4.  Estimates of ecological diversity for the clone libraries.
SampleNumber of clonesNumber of phylotypesSimpsonShannonEvenness
  1. Sample names beginning with ‘B’ indicate the bacterial libraries while that with ‘A’ indicate the archaeal libraries.

B 55mbsf WGA-PCR8860.5991.1750.656
B 55mbsf PCR1620.2330.3770.544
B 103mbsf WGA-PCR11840.0990.2450.177
B 103mbsf PCR1630.6080.9470.862
B 242mbsf WGA-PCR7680.7831.6760.806
A 242mbsf WGA-PCR 21F/1390R16550.3400.7090.441
A 242mbsf WGA-PCR 8F/1492R4920.0800.1710.246

Phylogeny of arctic subsurface microbial communities

The archaeal 16S rRNA clones all align within the group Crenarchaeota (Fig. 2), and more specifically within the MCG. Two clones, 242mbsf_1_5E and 242mbsf_1_3C, cluster most closely with other deep subsurface sediment clones from a wide geographical distribution in the Pacific Ocean, including the Peruvian continental shelf and the Cascadia Margin. The neighbours originate from sediments collected at depths of 6.55–442 mbsf and another reported clone is from crustal fluids. Three other clones (242mbsf_1_1C, 242mbsf_1_1A and 242mbsf_1_1D) cluster most closely with uncultured crenarchaeota from a paleosol 188 m below ground surface and estuarine sediment in Brazil. The two remaining clones, both unique clones from the ARC-8F/1492R library (242mbsf_3_2A and 242mbsf_3_3E), form a clade with uncultured archaea from hydrothermal fluids and subsurface groundwater.

Figure 2.

Miscellaneous Crenarchaeote Group neighbour-joining phylogenetic tree showing representative archaeal 16S rRNA gene sequences from sediments retrieved at 242 mbsf on the Lomonosov Ridge. Clones are named by Depth_clone library_index number. The number in parenthesis indicates the number of clones for the represented phylotype. Bootstrap values were obtained from 1000 analyses.

The bacterial 16S rRNA gene clones formed several unique clades and group distinctly by depth (Fig. 3). Clones that cluster within the Chloroflexi were found exclusively with the deepest sample, and these five clones align most closely with those from other deep subsurface sediments. Representatives from the phyla Bacteroidetes and γ-Proteobacteria also emerged from the 242 mbsf sample. The highest phyla diversity was encountered from DNA obtained at 55 mbsf, which yielded eight unique clones spread across five phyla. Those in the γ-Proteobacteria and Firmicutes align with cultured representatives from deep sea sediments, while clones from within the β-Proteobacteria, Bacteroidetes and Actinobacteria show close relationships with isolates and clones from glacial and Antarctic sediments. Less bacterial diversity was encountered at 103 mbsf, with six clones distributed across four phyla. Those in Firmicutes and Actinobacteria relate to marine sediment clones while the two in Bacteroidetes cluster with uncultured representatives from tundra meadow soil and river sediment.

Figure 3.

Neighbour-joining phylogenetic tree showing representative bacterial 16S rRNA gene sequences amplified from subsurface sediments collected on the Lomonosov Ridge. Clones are named by Depth_clone library_index number. The number in parenthesis indicates the number of clones for the represented phylotype. Bootstrap values were obtained from 1000 analyses.

Discussion

Comprehensive studies of life in deeply buried sediments are primarily hindered by the cost and the logistics of sampling this environment. Both of these concerns are particularly exacerbated while coring in the ice-covered Arctic. The Integrated Ocean Drilling Program and its predecessor programs have provided unique access to deeply buried marine sediments for microbiological analyses. Early insights into the microbial communities residing in this habitat indicate that clone libraries are commonly dominated by similar lineages. This suggests that these organisms are particularly well adapted to conditions in the subseafloor independent of geographical location. These generalizations will continue to be tested as new areas are explored and new techniques are brought to bear on this subject.

Once deep sediments are in hand, technical challenges such as low biomass and the co-extraction of PCR inhibitors have hindered culture-independent investigations of microbial communities. In this study, purification of extracted DNA by size exclusion gel filtration enabled successful PCR amplification of two samples (55 and 103 mbsf), and ultimately revealed biodiversity not seen with other methods. The sample from the deepest sediments was still recalcitrant to PCR reactions following the purification step, perhaps because the higher organic content prevented sufficient separation of DNA and PCR inhibitors. The WGA-PCR approach helped solve both issues, and ultimately allowed detection of most of the novel sequences. The WGA methods have been previously applied to environmental studies. Gonzalez and colleagues (2005) showed that the Φ29 DNA polymerase used in the WGA method is less sensitive to soil components that inhibit Taq polymerase. The enzyme has proofreading activity and can amplify DNA up to 100 kb. In the process of providing larger quantities of DNA, the method also effectively dilutes inhibitors prior to subsequent PCR methods. Abulencia and colleagues (2006) used WGA to reveal higher microbial biodiversity in low biomass sediments contaminated with heavy metals. However, WGA methods can introduce bias. While all sequences are expected to be present in the WGA material, certain bacteria genomes are preferentially amplified, meaning that the DNA of some organisms are underrepresented in the product (Abulencia et al., 2006). Bias was encountered in this study as only one sequence (103mbsf_2_2C) overlapped between clone libraries constructed with and without the WGA step. Thus, diversity studies conducted using both techniques can be complementary.

Amplification of archaeal sequences was only accomplished with the deepest samples. No indication of archaea was observed at 55 or 103 mbsf with the various approaches used. These upper sediments are characterized by low organic carbon concentrations and silicialastic sediments (Moran et al., 2006). It is possible that the limited organic carbon content at these depths is linked with the observed low bacterial diversity and apparent absence of archaea. Clone libraries revealed an archaeal community at 242 mbsf where TOC is approximately 2.93% as compared with ∼0.26% in the other two depths (Stein, 2007). These deeper biosilica-rich sediments dated to 45 Ma on a geological timescale when the Arctic Ocean was a warm ‘greenhouse’ (Moran et al., 2006), and are characterized as a sulfate reduction zone as determined by alkalinity of the porewater (Expedition 302 Scientists, 2006).

The archaeal 16S rRNA gene clone library from sample 242 was comprised entirely of crenarchaeota. In fact, all clones are phylogenetically placed into the MCG, and neighbour most closely to sequences obtained from sulfate-methane transition zones (Biddle et al., 2006), sulfate reduction zones (Sørensen and Teske, 2006) and sediments with high concentrations of methane and methane hydrate (Inagaki et al., 2006). The MCG has wide distribution throughout the marine subsurface environment (Coolen et al., 2002; Parkes et al., 2005) and has shown dominance in hydrate-free, organic rich sites off Peru, below hydrate stability zones from the Cascadia Margin and volcanic ash layers in the sea of Okhotsk (Inagaki et al., 2003). While the biochemical activities of these uncultured crenarcheaota remain unknown, their continual discovery in similar geochemical habitats suggests a physiological adaptation.

Bacterial 16S rRNA gene clone libraries demonstrate stratification with depth and include diversity across several phyla. Sediments retrieved from 55 mbsf contained the highest phyla diversity, and 50% of the 16S rRNA sequences show close phylogenetic relationships with bacteria from glacial and frozen sediments, suggesting that these are possibly cold adapted. These include a Sphingobacterium sp. from Antarctic sediment, an uncultured clone from glacier ice and an actinobacteria isolate from glacier soil. Two sequences obtained from 103 mbsf fall into the phylum Bacteroidetes and cluster with environmental clones from frozen terrestrial soil and freshwater sediment. The sediments from these two depths are estimated to be between 3.2 and 10 Ma and show evidence of seasonal ice cover and ice rafted debris (Moran et al., 2006). Thus, sea ice could be a possible ancestral source of these terrestrial and freshwater lineages. The bacterial clone library from 242 mbsf is dominated by Chloroflexi, which was not represented in the shallower samples. This deepest sample is characterized by organic rich sediments from approximately 45 Ma, just prior to the transition of the Arctic Ocean from greenhouse to icehouse. Chloroflexi were found to be a dominant 16S rRNA phylotype in several hydrate free deep subsurface sediments at the Peru Margin (Inagaki et al., 2006) and the Nankai trough (Reed et al., 2002). The closest related cultured organism is Dehalococcoides ethenogenes, known for its ability to reductively dechlorinate chloroethenes to ethane (Maymo-Gatell et al., 1997).

According to rarefaction curves generated by RFLP analysis, the most easily accessed diversity within the WGA-PCR clone libraries appears to be expressed in the sequenced clones whereas the PCR clone libraries were undersampled. In addition, RFLP patterns serve as proxies for full-length sequences, and therefore can conceal additional diversity. More exhaustive sampling and sequencing of the clone libraries would likely reveal harder to access phylotypes. Biases in DNA extraction, WGA and PCR amplification, and cloning may also contribute to underestimations of true diversity in the natural assemblages.

The observed phylotypes differed with depth and may be a reflection of the geochemical environment. Both bacterial and archaeal phylotypes found in the high organic carbon sediments cluster with clones from high organic carbon sediments in the Peru and Cascadia Margins (Inagaki et al., 2006; Teske and Sørensen, 2008). Therefore, it may be the geochemistry and not geography that is driving the community structures. In addition, a few clones show a phylogenetic relationship to surface glacial and polar samples. However, the microbial community of the arctic subsurface sediments is more closely related to other deep subsurface environments than to surface arctic sediments (Ravenschlag et al., 2001), suggesting that the deep sedimentary subsurface of the ocean is a unique habitat.

Experimental procedure

Sample collection

Sediment samples were collected during the Integrated Ocean Drilling Program Arctic Coring Expedition (IODP Expedition 302) during August and September of 2004. All samples used for this study were recovered from Hole M0002A (87°55.271′N; 139°21.901′E) using the extended core barrel onboard the drillship Vidar Viking. The drill site is located ∼250 km from the North Pole on the Lomonosov Ridge with a water depth of 1209 m. The cores were subsampled on deck by inserting sterile cut-off syringes into the centre of the core. The subcores were stored inside the syringes at −80°C until DNA extractions were conducted onshore. Three main geochemical units were defined by the porewater chemistry: an upper ammonium oxidation zone, a carbonate dissolution zone and a deep sulfate reduction zone (Expedition 302 Scientists, 2006). One sample from each unit was chosen for cloning and sequencing [302-M0002A-12x-1 (55 mbsf), 302-M0002A-24x-1 (103 mbsf) and 302-M0002A-56x-1 (242 mbsf].

DNA extractions and purification

Environmental DNA (eDNA) was extracted from 1 g of sediment from each depth using the FastDNA® SPIN Kit for Soil (Qbiogene, Irvine, CA). The manufacturer's protocol was followed along with the following modifications. Two hundred micrograms of polyadenylic acid was added to tubes prior to lysis in order to prevent eDNA from binding to sediment (Webster et al., 2003). Mechanical lysis of the cells was accomplished by vigorous vortexing for 30 min (MoBio, vortex adapter). An increased elution volume of 100 μl was employed to ensure complete re-suspension of the silica for efficient DNA elution. DNA was next purified by size-exclusion chromatography using ChromaSpin™ TE-100 columns (Clontech, Mountain View, CA) to remove co-extracted PCR inhibitors, such as fulvic acids. After prepping the column according to manufacturer's guidelines, eDNA was adsorbed onto the resin and then eluted with five 500 μl aliquots of TE wash (Mincer et al., 2005). Each aliquot was analysed using a NanoDrop Spectrophotometer (NanoDrop, Wilmington, DE) to identify fractions containing the purified eDNA. Samples were stored at −20°C.

DNA amplification

Whole-genome amplification of the eDNA was performed by multiple displacement amplification with Φ29 DNA polymerase using a REPLI-g® Midi kit (Qiagen, Valencia, CA) as per the manufacturer's instructions. Reactions were carried out at 30°C for 16 h using 5 μl of DNA template. Subsequent 16S rRNA gene amplifications were carried out using 20 ng of the whole-genome amplicon in a 50 μl PCR reaction mixture containing 25 μl of Qiagen Taq MasterMix (Qiagen, Valencia, CA) and 0.1 μm of primer. Multiple sets of bacterial and archaeal specific 16S rRNA gene primers were used (Table 2). An initial 5 min denaturing step at 94°C was followed by 30 cycles of 30 s at 94°C, 90 s at 60°C or 52.5°C (bacterial or archaeal respectively) and 120 s at 72°C and a final 10 min extension step at 72°C. The products were analysed by electrophoresis on a 1% agarose gel imaged on a Typhoon™ 9410 variable mode imager (GE Healthcare Bio-Sciences, Piscataway, NJ). All amplification experiments contained a negative control of molecular biology grade water and showed no amplification. The amplicon was cleaned up for cloning using the QIAquick PCR purification kit (Qiagen, Valencia, CA) following manufacturer's guidelines and then quantified using a NanoDrop spectrophotometer.

Environmental clone libraries

16S amplicons were cloned into the pGEM®-T Easy Vector System (Promega, Madison, WI) at a 1:1 insert to vector ratio as per the manufacturer's protocol. The clones containing an insert, as indicated by their white colour, were picked and cultivated overnight in 96-well plates containing LB broth (10 g of NaCl, 5 g of yeast extract, 10 g of tryptone per 1 l deionized water). A total of 384 clones from each ligation were picked and the first 192 were grown in duplicate. Frozen stocks of the overnight cultures were generated by supplementing with glycercol (20% total volume) and storing at −80°C. Plasmids were extracted using TurboFilter 96 and QIAprep 96 Turbo miniprep kit on the QIAvac 96 following the methods recommended by Qiagen (Valencia, CA). Plasmid inserts were PCR amplified and then dereplicated using RFLP analysis. PCR amplicons (∼500 ng) from each clone were digested with HaeIII, AluI and RsaI (3 units of each) in a 10 μl reaction volume for 1 h at 37°C. The reaction products were separated by electrophoresis on a 2.5% agarose gel and then visualized using a Typhoon™ 9410. The RFLP fingerprints were analysed with Kodak Digital Science™ 1D software and rarefaction curves were built based on clone-to-phylotype ratios. To analyse diversity, the Simpson, Shannon and evenness indices were calculated (as described in Liao et al., 2007). Clones representing unique phylotypes were sequenced using the 16S primers or vector primers (T7, SP6). Sequencing was performed on a Beckman Coulter CEQ 8000 (Beckman Coulter, Fullerton, CA) at the University of Rhode Island's Genomics and Sequencing Center (Kingston, RI). The resulting sequences were trimmed and contigs were built using Sequencher (Gene Codes, Ann Arbor, MI). Sequences were aligned and phylogenic trees built using ARB, an open source phylogenetic software package (Ludwig et al., 2004). Bootstrap analysis of the trees was performed using Paup* 4.0 (Swofford, 2003).

Acknowledgements

This research was supported by NOAA Grant NA04OAR4600193 to D.C.R. and D.C.S. and an NSF-IODP grant to D.C.S. It was further made possible by the use of Rhode Island INBRE Research Core Facilities supported jointly by NCRR/NIH Grant No. P20 RR016457 and the Network institutions. This research used samples and data provided by the Integrated Ocean Drilling Program.

Ancillary