Molecular Analysis of Deep Subsurface Microbial Communities in Nankai Trough Sediments (ODP Leg 190,Site 1176)

FEMS Microbiology Ecology 45 (2003) 115^125 www.fems-microbiology.org

Molecular analysis of deep subsurface microbial communities in Nankai Trough sediments (ODP Leg 190,Site 1176)

Konstantinos Ar. Kormas a;Ã,David C. Smith b,Virginia Edgcomb a,Andreas Teske a;1

a Biology Department, Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA b Graduate School of Oceanography, University of Rhode Island, Narragansett, RI 02882, USA

Received 19 December 2002; received in revised form 5 April 2003; accepted 14 April 2003

First published online 4 June 2003

Abstract

The prokaryotic community inhabiting the deep subsurface sediments in the Forearc Basin of the Nankai Trough southeast of Japan (ODP Site 1176) was analyzed by 16S rDNA sequencing. Sediment samples from 1.15,51.05,98.50 and 193.96 m below sea floor (mbsf) harbored highly diverse bacterial communities. The most frequently retrieved clones included members of the Green non-sulfur bacteria whose closest relatives come from deep subsurface environments,a new epsilon-proteobacterial phylotype,and representatives of a cluster of closely related bacterial sequences from hydrocarbon- and methane-rich sediments around the world. Archaeal clones were limited to members of the genus Thermococcus,and were only obtained from the two deepest samples. ß 2003 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.

Keywords: Bacterium; Archaea; 16S rRNA; Marine; Deep subsurface; Ocean Drilling Program; JOIDES resolution; Leg 190; Site 1176; Nankai Trough

1. Introduction of the pure culture isolations and characterizations from deep marine sediments have focused on sulfate-reducing The deep subsurface biospheres of the world’s oceans bacteria from the Japan Sea [6,7] and the Cascadia Margin have only recently become a main focus of scienti¢c re- [8]. Diverse heterotrophic proteobacteria,Gram-positive search. Although approximately 70% of the Earth’s sur- bacteria and members of the Cytophaga^Flavobacte- face is marine,little is known about the microbiology of rium^Bacteroides (CFB) phylum have also been enriched the underlying sediments. Recent estimates of global bac- from deep sediments of the Nankai Trough southeast of terial biomass indicate that a large fraction of bacterial Japan [9]. Likewise,extensive cultivation e¡orts on sam- biomass resides in the deep subsurface,most of which is ples from terrestrial subsurface environments have yielded in the marine deep subsurface [1,2]. predominantly members of the alpha-,beta-,and gamma- Although rates of metabolic processes in the deep-sea proteobacteria,the CFB phylum and of the high G+C and subsurface have been estimated [2^4],the microorganisms low G+C Gram-positive bacteria [10]. To date,almost all remain largely unidenti¢ed. Cultivation surveys and most cultured bacteria from the deep subsurface fall into well- probable number quanti¢cations have detected fermenta- known bacterial genera with many cultured species and tive,nitrate-reducing and sulfate-reducing bacteria as well strains from near-surface habitats. as methanogenic archaea ([5] and references therein). Most A very di¡erent picture emerges from culture-independent surveys using polymerase chain reaction (PCR),cloning and sequencing of rRNA genes and functional genes. From the ¢rst sequencing studies of deep subsurface sedi- * Corresponding author. Present address: Department of Zoology ^ ments [11,12],it has become clear that these microbial Marine Biology,School of Biology,University of Athens,157 84 Pane- ecosystems harbor diverse bacteria and archaea that are pistimiopoli,Athens,Greece. Tel.: +30 (210) 727-4721; Fax: +30 (210) in many cases only distantly related to cultured isolates 727-4604. [13^15]. This 16S rRNA survey describes the prokaryotic E-mail address: [email protected] (K.A. Kormas). biodiversity of non-hydrothermal deep subsurface sedi- 1 Present address: Department of Marine Sciences,Venable Hall 12-1, ments from 1^194 m below sea £oor (mbsf) at a water CB 3300,University of North Carolina,Chapel Hill,NC 27599,USA. depth of 3017 m in the Nankai Trough southeast of Japan.

FEMSEC 1525 9-7-03 116 K.A. Kormas et al. / FEMS Microbiology Ecology 45 (2003) 115^125

2. Materials and methods mbsf,the sediments are approximately 800 000 years old [19]. The sediments consist predominantly of nannofossil- 2.1. Site description and sampling rich hemipelagic mud interlayered with volcanic ash,and sand-to-silt turbidites showing local soft-sediment folding All sediment samples for this study were obtained dur- and disruption. Diatoms become scarce in the lower part ing Ocean Drilling Program (ODP) Leg 190,from Hole of the unit,but nannofossils are abundant throughout. 1176A in the Nankai Trough,at a depth of 3017 m below This sedimentary sequence is characterized by hemipelagic sea level in June 2000 (Fig. 1). Sediment samples were settling,occasional turbidity currents,and submarine mud retrieved by advanced piston coring (APC),except for £ows,together with air falls of volcanic ash. Porosities the sample from 194 mbsf that was retrieved by extended decrease gradually with depth,from V65^70% at the core barrel (XCB) drilling [16,17]. Using cutting wires, mudline to 55^60% at 220 mbsf,with a steeper decrease whole-round samples were obtained from core sections by 3^5% around 200 mbsf,and remain relatively constant immediately after core retrieval. Subcore samples of ap- between 40 and 45% at 300 mbsf. Temperature increased proximately 706^1178 ml were taken from sediment layers linearly from 1.4‡C at 1.15 mbsf to 12.2‡C at 193.94 mbsf. at 1.15,51.05,98.50 and 193.96 mbsf,and were stored at Sulfate concentrations decreased with depth,and reached 380‡C on the ship until transport to the home laboratory zero at approximately 20 mbsf. Sulfate concentrations de- on dry ice. creased linearly with depth over the upper 10 m,followed Site 1176 is located in the upper slope basin within the by a local maximum at 20 mbsf. At the base of Hole large thrust-slice zone in the Nankai Trough accretionary 1176A (325^393 mbsf),sulfate concentrations increased prism,which is created by the subduction of the sediment- again to V17 mM,almost 60% of the seawater value. rich Philippine plate below the Eurasian plate at a rate of Except for sulfate,pore £uid composition at these depths 4 cm/year [18]. All samples analyzed in the present study remained near that of seawater,suggesting slow,ongoing belong to the same quaternary sediment unit in the upper bacterial sulfate reduction. Ammonium produced by mi- slope basin that covers the accretionary prism. At 200 crobially mediated decomposition of organic matter in-

Fig. 1. Location of ODP Site 1176 in relation to other ODP sites on the transect covering the Nankai Trough southeast of Japan.

FEMSEC 1525 9-7-03 K.A. Kormas et al. / FEMS Microbiology Ecology 45 (2003) 115^125 117 creased in concentration with depth,having a broad max- primers BAC-8F (5P-AGAGTTTGATCCTGGCTCAG- imum of V5 mM between V20 and 200 mbsf and peaked 3P) and BAC-907R (5P-CCCGTCAATTCCTTTGAGT- at V100 mbsf. Total organic carbon (TOC) ranged from TT-3P) were used. Each PCR run consisted of a 1-min 0.30 to 1.73 wt% in the ¢rst 200 mbsf (lowest at 62.48 and pre-PCR hold at 94‡C,followed by 30 cycles consisting highest at 135.20 mbsf). Headspace gas concentrations of of a 45-s denaturation step at 94‡C,a 45-s annealing step methane were very low in the ¢rst 11.9 mbsf within the at 52.5‡C,a 2-min elongation step at 72‡C,and at the end sulfate reduction zone (7.3^11.7 ppm),increased to 3000^ of the 30 cycles,a ¢nal 7-min ¢nishing step at 72‡C. All 30000 ppm below the sulfate reduction zone,and returned PCR ingredients were prepared with twice-autoclaved ul- to generally less than 10 ppm in the deepest portion of the tra pure water,and stringent anticontamination controls core,below 325 mbsf. Methane was the most dominant were used during PCR preparation. The PCR products hydrocarbon measured throughout Hole 1176A. In gener- were checked on a 1% low-melt agarose gel,bands were al,bacterial cell densities declined rapidly from the sur- excised,and PCR products were extracted with the Prep- face. Microbial abundances,determined by acridine or- A-Gene DNA gel puri¢cation kit (Bio-Rad,Hercules,CA, ange direct counts [2,5],decreased by one order of USA) following the manufacturer’s protocol. The puri¢ed magnitude over the sampling interval from 9.33U107 cells PCR products were A-tailed to improve cloning e⁄ciency per cm3 sediment at 1.19 mbsf to 1.19U107 at 193.94 by mixing approximately 50 ng of puri¢ed PCR product mbsf,followed by a further decrease to 1.11 U106 per with 5 Wlof10UPCR bu¡er (200 mM Tris (pH 8.55),100 3 cm in the two deepest samples at 344 and 363.5 mbsf. mM (NH4)2SO4,20 mM MgCl 2,300 mM KCl,0.1% (wt/ The most intense microbially mediated reactions occurred vol PCR bu¡er) gelatin,0.5% (vol/vol PCR bu¡er) NP- in the top 100 mbsf of the section [19]. 40),5 Wl deoxynucleoside triphosphate (2 mM) and 1 U of Taq DNA polymerase (Promega,Madison,WI,USA). 2.2. DNA isolation, 16S rRNA gene ampli¢cation, cloning The mixture was incubated at 72‡C for 10 min,extracted and sequencing with phenol^chloroform^isoamyl alcohol (25:24:1) and centrifuged for 5 min. The aqueous top layer was then Whole-round core samples were stored at 380‡C until transferred to a fresh microcentrifuge tube. The A-tailed processing. The outer layers of the deep-frozen cores were PCR product was precipitated with 10% (vol/vol aqueous pared away aseptically in order to remove sediment that layer) 5 M KCl and two volumes of 100% ethanol,washed had been in contact with the drilling £uid. DNA was ex- with 70% ethanol and resuspended in 4 Wl of distilled tracted from approximately 1.5 g of sediment from the water. The PCR products were cloned using the TOPO interior of each core using the UltraClean soil DNA iso- XL PCR cloning kit (Invitrogen Corporation,Carlsbad, lation kit (MoBio Laboratories Inc.,Solana Beach,CA, CA,USA) and transformation by electroporation accord- USA) according to the manufacturer’s protocol,and dis- ing to the manufacturer’s speci¢cations. For each sediment solved in 25 Wl PCR water. The DNA was diluted with layer,24^39 clones were analyzed. Of all the clones ana- PCR water before PCR ampli¢cation,to overcome persis- lyzed,on average,27% were false positive and were extent PCR inhibition problems. All samples for archaeal cluded from further analysis. PCR were diluted 1:1000; samples for bacterial DNA am- Sequence data were obtained on an ABI3700 (Applied pli¢cation from the surface (1 mbsf) were diluted 1:10, Biosystems Inc.,Foster City,CA,USA) capillary sequenc- and samples from the other layers (51,98,194 mbsf) ing apparatus using the BigDye Terminator v 3.0 kit (Ap- were diluted 1:100. For PCR ampli¢cation,1 Wl of the plied Biosystems Inc.,Foster City,CA,USA) with the diluted DNA preparation was used. The DNA input for primers M13F (5P-GTAAAACGACGGCCAG-3P) and the archaeal PCR corresponded to the DNA content of M13R (5P-CAGGAAACAGCTATGAC-3P) and a cycle 6U1035 g sediment ([1 Wl/25Wl]Udilution factor 0.001U sequencing protocol according to the manufacturer’s pro- 1.5 g sediment = 6U1035 g sediment); the DNA input tocol. For each individual clone,forward and reverse for the 1:100 and 1:10 diluted bacterial PCR corre- reads were assembled using Sequencher 3.0 (Gene Codes 3 3 sponded to 6U10 4 and 6U10 3 g sediment,respectively. Corporation,Ann Arbor,MI,USA). The 5 P and 3P ends With nested PCR,a 900-bp portion of the 16S rRNA of each PCR product were covered by one sequence read genes was ampli¢ed using nested PCR on a Gene-Amp each,and the central 0.7- to 0.8-kb portion was covered PCR Cycler (Perkin-Elmer,Foster City,CA,USA). For on both strands. Sequences were checked for chimeras the ¢rst ampli¢cation,the archaeal primers ARC-8F using the CHIMERA_CHECK function of the Ribosomal (5P-TCCGGTTGATCCTGCC-3P) and ARC-1390R (5P- Database Project II [20]. GACGGGCGGTGTGTGCAA-3P) and bacterial BAC- 8F (5P-AGAGTTTGATCCTGGCTCAG-3P) and BAC- 2.3. Phylogenetic analysis 1390R (5P-TGTACACACCGCCCGTC-3P) were used. For the second ampli¢cation,the archaeal primers ARC- Sequence data were compiled using the ARB software 8F (5P-TCCGGTTGATCCTGCC-3P) and ARC-927R (5P- (www.arb-home.de) and aligned with sequences obtained CCCGCCAATTCCTTTAAGTTT-3P) and the bacterial from the ARB and GenBank databases,using the ARB

FEMSEC 1525 9-7-03 118 K.A. Kormas et al. / FEMS Microbiology Ecology 45 (2003) 115^125

FastAligner utility,and followed by manual aligning ac- notation of a 100% complete microbial community survey, cording to secondary structure. Analyses were performed the abundances of clone groups are given in absolute num- using minimum evolution and parsimony methods imple- bers,not as percentages. mented in PAUP* version 4.08b [21]. Heuristic searches The highly diverse bacterial 16S rDNA clone libraries under minimum evolution criteria used 1000 random-ad- were dominated by proteobacterial clones (35 clones; in- dition replicates per data set,each followed by tree bisec- cluding 16 epsilon- and 12 gamma-proteobacterial clones), tion^reconnection topological rearrangements. Models for members of the Green non-sulfur (GNS) bacteria,and a use in minimum evolution analyses of the archaeal and bacterial lineage without cultured representatives (Fig. 2). bacterial data sets were selected using the likelihood ratio The sequences in the epsilon-proteobacteria consisted of test implemented in Modeltest v.3.0 [22]. For both data essentially identical clones of a phylotype that was related sets,all models tested that were simpler than the general to sulfur- and iron-reducing Sulfurospirillum species [23], time reversible (GTR)+invariant sites (I)+gamma shape and the sulfur oxidizer Arcobacter sul¢dicus [24]. They parameter (G) model had a signi¢cantly poorer ¢t to the di¡er from the epsilon-proteobacterial group I sequences data for the reference topology (based on a neighbor-join- (related to Alvinella pompejana and Rimicaris exoculata ing tree according to Jukes^Cantor). Therefore,minimum symbionts) that are frequently found at the sediment^ evolution distance trees were inferred using a GTR model water interface in deep-sea sediments,hydrothermal sedi- with six classes of substitutions,unequal base frequencies, ments and cold seep sediments [25,26]. Sequences of the proportion of invariant sites,and the evolutionary rate of GNS division were found eight times. Their closest rela- the remaining portions of sites di¡ering according to a tives were sequences from terrestrial and marine subsur- gamma distribution (with shape parameter alpha). Model face sediments including hydrate-containing sediments of parameters estimated for minimum evolution analyses for the Nankai Trough [27,13],and the dehalogenating,an- the archaeal and bacterial data sets were: alpha of 0.620 aerobic bacterium Dehalococcoides ethenogens [28]. Eleven and 0.818,and proportion of invariable sites of 0.618 and clones belonged to a bacterial lineage that predominantly 0.194,respectively. Bootstrapping under minimum evolu- harbors sequences from hydrocarbon- and methane-rich tion and parsimony criteria was done with 200 replicates sediments from around the world,cold seeps,and enrich- for both the archaeal and bacterial data sets. New sequen- ment cultures on alkanes and aromatic compounds. With ces have GenBank numbers AY191324^AY191357 and the exception of Nankai subsurface clones 1 mbsf-15 and AY191321^AY191323 for bacteria and archaea,respec- 51 mbsf-15,the members of this lineage form a well-sup- tively. ported cluster of closely related strains [29,30,31,13,26]. Fewer clones from other phylogenetic groups were found. Four clones grouped within the alpha-proteobacte- 3. Results rial genera Hyphomicrobium and Pedomicrobium,methyl- otrophic and heterotrophic bacteria who divide by hyphae 3.1. 16S rDNA from Nankai Trough sediments formation (Fig. 2). Three clones belonged to the cyanobacteria and chloroplasts. Lineages that were represented Only low amounts of DNA could be extracted from by single clones included the delta-proteobacteria,the OP8 Nankai Trough sediments,using 1.5 g sediment/sample. candidate subdivision,the CFB phylum,and the Plancto- 5-Wl aliquots of the 25-Wl DNA extracts could not be vi- myces phylum (Fig. 2). sualized with ethidium bromide staining,indicating that In marked contrast to these subsurface clones that are the DNA concentration was below 1^2 ng Wl31,or 5^10 only distantly related to cultured bacterial species,almost ng per sediment sample. After the ¢rst round of 16S all beta- and gamma-proteobacterial sequences in the rDNA-targeted PCR ampli¢cation,no visible PCR prod- Nankai clone libraries closely match the 16S rDNA se- uct could be observed. Visible 900-bp PCR products ap- quences of Escherichia coli, Pseudomonas stutzeri, Vario- peared only after the second (nested) PCR,and were ab- vorax paradoxus, Acinetobacter junii,and near-identical sent in the negative controls for DNA blind preparation sequences (MT9,MT5,MT11,MT6) that have been ob- and PCR. tained by cloning from negative controls or from environmental samples with low DNA content [32]. Several of 3.2. Bacterial sequences these sequences were obtained multiple times from di¡erent sediment layers,except the near-surface sediment ( Ta- The sequences of 61 bacterial and 80 archaeal clones ble 1). were determined. As a caveat,the absolute numbers of clones and their depth distribution patterns cannot be in- 3.3. Depth distribution terpreted as indicators of the abundance of di¡erent bacterial groups in nature; also,the frequency of rarely re- Di¡erent lineages of bacterial 16S rRNA sequences fol- trieved clones indicates that the microbial community was lowed di¡erent patterns in depth distribution. Some bac- not sampled to exhaustion. To avoid the misleading con- terial groups were only recovered from deep sediment

FEMSEC 1525 9-7-03 K.A. Kormas et al. / FEMS Microbiology Ecology 45 (2003) 115^125 119

Fig. 2. 16S rRNA-based minimum evolution distance tree showing the phylogenetic relationships of bacterial Nankai subsurface clones from ODP Hole 1176A to cultured bacterial species and related environmental clones,mostly from deep subsurface sediments and deep-sea sediments. The tree is bas ed on E. coli positions 28^906 of the 16S rRNA gene. Bootstrap support values (%) are given at nodes for minimum evolution distance (¢rst),and maximum parsimony (second) criteria. * indicates value 6 50. layers. The epsilon-proteobacterial clones were obtained 1 m surface layer of the sediment column; such as all of from the 51,98.5 and 194 mbsf samples,but not from the alpha-proteobacterial clones,the OP8 candidate divi- the 1 mbsf samples (Table 1). The cyanobacteria- and sion clone,and most of the clones in the hydrocarbon- chloroplast-related sequences,the delta-proteobacterial associated uncultured bacterial lineage. The sequences in clone,and the CFB clone were obtained from the 194 the GNS phylum were found at all depths except 194 mbsf sample. Other sequences are associated with the mbsf.

FEMSEC 1525 9-7-03 120 K.A. Kormas et al. / FEMS Microbiology Ecology 45 (2003) 115^125

3.4. Archaeal 16S rDNA sequences soils. The relatively well-studied anoxygenic phototrophic bacteria that became the namesake of the GNS bacterial Archaeal sequences were found only at 98.5 and 194 phylum are limited to one subdivision only. The phyloge- mbsf,not in the surface layer or at 51 mbsf. All archaeal netic diversity within the GNS bacteria indicates consid- sequences from Hole 1176A belonged to a single phylo- erable physiological and ecophysiological diversity,as evi- type within the Thermococcales (Fig. 3); sequences of denced by the physiological dissimilarity of the few methanogens or other archaeal groups were not recovered. cultured species,and the unusual range of their extreme The archaeal sequences formed a monophyletic branch habitats. with Thermococcus strains MZ1,MZ2,MZ5 and MZ13 The representatives of the well-supported clone cluster isolated from sul¢de chimney samples,and with strain (99 and 90% bootstrap support under minimum evolution MZ3 isolated from alvinellid worms (MZ3) at the Main and maximum parsimony criteria) whose members are Endeavor and High Rise ¢elds on the Juan de Fuca Ridge generally found in hydrocarbon- and methane-rich sedi- [33]. ments were retrieved mostly in the surface sediment layer of the Nankai Trough samples,except for two clones from 51 mbsf (Fig. 2). These bacteria may require a more or- 4. Discussion ganic carbon-rich environment than the deeper sediment layers of Site 1176 can o¡er. Representatives of this group 4.1. Bacterial diversity and distribution patterns in the are isolated either from near-surface sediment,or from sediment column methane-rich subsurface samples where methane hydrates are present,for example clone AT425 EubA5 [31] and The best candidates for specialized subsurface bacteria clone MB-B2-103 [13]. at Nankai Site 1176 are the members of the GNS bacterial The most frequently retrieved cluster of near-identical phylum that have been repeatedly found in both the ter- sequences,the epsilon-proteobacterial clones (represented restrial and marine subsurface [27,34,13]. In deeply buried in Fig. 2 by clone 194 mbsf-12),has to our knowledge not sediments in Eastern Mediterranean organic-rich sapro- been found in deep subsurface environments before. Its pels,70% of all bacterial clones were members of the phylogenetic position suggests similarities to the faculta- GNS phylum; signi¢cantly fewer clones of this group tively anaerobic,heterotrophic and fermentative genus were found in organic-poor sediments between the sapro- Sulfurospirillum,which can use a diverse spectrum of elec- 3 pel layers [34]. Apparently,these bacteria require organic tron acceptors (O2 in low concentrations; NO3 ,S‡, 23 carbon sources,but their substrate preferences are un- S2O3 ). The species Sulfurospirillum barnesii can also re- known. The closest cultured relatives of these subsurface spire with Fe(III) and arsenate [23]. The epsilon-proteo- clones within the GNS bacteria are several strains of the bacterial subsurface clones are also related to the micro- reductively dehalogenating bacterium D. ethenogens. This aerophilic,sul¢de-oxidizing bacterium Candidatus A. sul- bacterium gains energy by oxidation of hydrogen with ¢dicus. This autotrophic bacterium does not ¢x carbon via concomitant dechlorination of tetrachloroethene,but re- the Calvin cycle,but may instead use the reverse TCA quires extracts of mixed microbial enrichments for growth cycle [24]. Thus,the closest cultured relatives of this sub- [28]. The phylogeny of the GNS phylum groups the cur- surface clone group show divergent physiologies. rently known Dehalococcoides sequences and related sub- The alpha-proteobacterial sequences were found exclu- surface clones (Fig. 2) into one of at least four major sively in the surface samples. These Hyphomicrobium- and subdivisions within the GNS phylum [35]. Other subdivi- Pedomicrobium-related sequences may represent organo- sions include clones from jet fuel-contaminated aquifers, trophic sediment bacteria which grow in situ near the sedi- hot geothermal springs,and subsurface sediments and ment surface. As an alternative possibility,they could be derived from the overlying water column,as Hyphomi- Table 1 crobium and Pedomicrobium spp. are found as epibionts Bacterial clones obtained multiple times from di¡erent sediment layers on settling diatoms [36]. Representative clone 1 mbsf 51 mbsf 98.5 mbsf 194 mbsf Total Clones that were isolated only in a single instance are 194 mbsf-1 (E. coli)123not necessarily insigni¢cant. Near-identical sequences have 194 mbsf-14 (E. coli)213been found in other molecular surveys of the Nankai deep 194 mbsf-18 11 2subsurface environment and of marine sediments around (Aquaspirillum) Japan. This study has revealed four cases of such similar- 98 mbsf-11 123ities (Fig. 2). The deeply-branching delta-proteobacterial (Acinetobacter) 194 mbsf-23 (epsilon) 4 7 5 16 clone 194 mbsf-22 is closely related to Nankai clone 1 mbsf-3z16 (HMAC)a 22MB-C2-152,a member of a well-supported cluster from 1 mbsf-3z3 (HMAC)a 22Nankai sediments at 297 mbsf below the hydrate stability 1 mbsf-4 (HMAC)a 31 4 zone,located ca. 350 km from Site 1176 [13]. The alpha- aHydrocarbon- and methane-associated cluster. proteobacterial surface clone 1 mbsf-1 matches a clone

FEMSEC 1525 9-7-03 K.A. Kormas et al. / FEMS Microbiology Ecology 45 (2003) 115^125 121

Fig. 3. 16S rRNA-based minimum evolution distance tree showing the phylogenetic relationships of archaeal Nankai subsurface clones from ODP Hole 1176A to cultured species and strains,and environmental clones within the genera Thermococcus and Pyrococcus. The tree is based on E. coli positions 24^906 of the 16S rRNA gene. Bootstrap support values (%) are given at nodes for minimum evolution distance (¢rst) and parsimony (second) criteria. * indicates value 6 50. obtained by reverse transcription of rRNA from cold seep clone BD2-16 from Suruga Bay at 1520 m depth [38],and sediments on the Sanriku Escarpment of the Japan Trench of Nankai clone MB-C2-147,one of two clones from Nan- at 5343 m depth,indicating the occurrence of Pedomi- kai subsurface sediments at 297 mbsf below the hydrate crobium and Hyphomicrobium spp. at the surface of these stability zone [13]. Within the OP8 candidate subdivision, deep-sea sediments [37]. The Planctomyces-a⁄liated Nan- the sur¢cial Nankai clone 1 mbsf-6 is the closest relative to kai clone 51 mbsf-20 is the closest relative of the deep-sea clone NKB18 from Nankai Trough surface sediments at

FEMSEC 1525 9-7-03 122 K.A. Kormas et al. / FEMS Microbiology Ecology 45 (2003) 115^125

3843 m depth [39] and of reverse-transcribed clone CS8.15 identi¢ed more directly by preparing control clone libra- from the Sanriku Escarpment cold seeps [37]. Apparently, ries from seawater-based drilling £uid,from the outer rim these rarely retrieved deep-sea clones from di¡erent sites of the core samples,and from the PCR-negative DNA match and substantiate each other,and indicate the pres- isolation controls [32]. ence of a common deep-sea bacterial £ora that can be found in several di¡erent deep-sea sediment habitats and 4.3. Archaeal diversity at di¡erent depths,although it does not dominate individual clone libraries. All 80 archaeal clones formed a tight cluster of nearly identical sequences with 1^2% mismatches (in the sequenc- 4.2. Contamination potential ing error range) randomly distributed over the 16S rRNA gene,including universally conserved regions. Three rep- Sediment samples with low DNA content are particu- resentative sequences from this highly redundant clone larly sensitive to ampli¢cation of contaminant DNA. The library were included in the phylogenetic tree (Fig. 3). gamma-proteobacterial clones obtained in this study are All retrieved archaeal sequences belonged to the genus good candidates for contaminants,based on their close Thermococcus,the most common genus of archaeal hyper- similarity to published 16S rRNA clones from negative thermophiles at hydrothermal vent sites,and a reliable DNA preparation controls [32]. The most frequently re- microbial marker for hydrothermal activity [25,42]. Special trieved clones are E. coli sequences that were obtained attention was paid to compare the Nankai clones to Ther- seven times after PCR ampli¢cation and cloning from mococcus clones from subsurface locations [43],or to the two deepest deep sediment samples (98 and 194 Thermococcus isolates from vent plumes and warm vent mbsf). These sequences could have originated from the £uids that could have carried subsurface strains to the E. coli strain that was used for cloning the PCR products. surface [44,33]. The Nankai subsurface clones formed a The 98 and 194 mbsf layers had reduced cell numbers (1.5 well-supported cluster (bootstrap support above 90%) and 1.2U107 g31 respectively [19] and therefore a lower with Thermococcus strains MZ1,MZ2,MZ3,MZ5 and DNA content compared to the surface layer,where the MZ13 (Fig. 3). These strains have been isolated from sul- E. coli contaminant was not detected. A further potential ¢dic vent chimneys and alvinellid worm tubes at the Juan source of contamination lies in the laboratory water sup- de Fuca Ridge,indicating a trans-Paci¢c occurrence pat- ply. Acinetobacter spp., Alcaligenes xylosoxidans and Pseu- tern of closely related Thermococcus strains. However,the domonas spp. are common constituents of bio¢lms in Nankai clones were not speci¢cally related to other sub- stainless-steel water mains [40]; P. stutzeri strains can surface clones or isolates. Phylogenetically distinct Ther- even be isolated from ultra pure water with deionization mococcus isolates (MZ4,8,9,10,11,12,and 14) from columns [41]. The clones related to P. stutzeri, A. junii, warm vent £uids at Juan de Fuca presumably represent Aquaspirillum delicatum, A. xylosoxidans and V. paradoxus strains from a speci¢c subsea£oor habitat,the zone where were probably derived from dead cell remnants or their hot hydrothermal vent £uids mix with entrained seawater DNA in the water supply. The water that was used for in highly porous pillow basalts and lava sheets [33].A preparing reagents,for DNA isolation and PCR reactions, similar subsurface origin has been inferred for Thermococ- was ultra ¢ltered using a MilliQ system,and autoclaved cus strains Gorda 1^6,isolated from hydrothermal vent twice before use; it is therefore unlikely to contain living plumes at the North Gorda Ridge [44]. The Thermococcus cells,but could have harbored traces of their genomic subsurface clones (pPCA7.21 and pPCA12.6) from rela- DNA. tively shallow,non-hydrothermal subsurface samples in A di¡erent kind of contamination,the intrusion of sea- the Philippine Basin,possibly originated at southeast water and pelagic microorganisms and their DNA into Asian vent sites,and could have been dispersed and sub- sediment samples,is inherent in the drilling process that sequently deposited by sedimentation through the water uses seawater as a drilling £uid. The deepest sample,194 column [43]. mbsf,was retrieved using XCB drilling,which is more Members of the genus Thermococcus are hyperthermo- prone to seawater contamination than the drilling tech- philes that require a minimum growth temperature be- nique used for other sampling depths,APC [16,17,42]. tween 48 and 73‡C [45]. Since the in situ temperatures in The three clones related to chloroplasts of green plants ODP Hole 1176A ranged from approximately 1 to 12‡C, and eukaryotic green algae that were found only in this the Thermococcus populations in the Nankai Trough sedi- sample were most likely derived from near-surface sea- ments examined here are presumably not metabolically water. The alternative scenario,burial and long-term pres- active in situ. They either represent inactive relic cells ervation of photosynthetic cells or their DNA in 200 m from past hydrothermal activity at Site 1176,or cells deep,800 000 years old quaternary sediments,is unlikely. that were recently introduced through interstitial £uid These identi¢cations of contaminants are based on com- £ow from hydrothermally active sites,for example from parisons with published studies,or on consistency argu- the nearby Nankai subduction zone and the accretionary ments. Sampling and laboratory contaminations can be prism. Thermococcus clones in non-hydrothermal sedi-

FEMSEC 1525 9-7-03 K.A. Kormas et al. / FEMS Microbiology Ecology 45 (2003) 115^125 123 ments were interpreted as paleontological markers of distance. Bacterial groups whose members are consistently archaeal populations that were introduced from distant, found in deep subsurface sediments and on the surface of geothermally active regions millions of years ago,and deep-sea sediments (the GNS bacteria,the hydrocarbon- buried and preserved in the sediment column [43]. These associated bacterial cluster,the OP8 division) point to ¢ndings raise the question whether thermophilic microor- common denominators in microbial community composi- ganisms can survive long-term deposition and burial in tion and function in these environments. The archaeal cold marine sediments to the extent that their cell integrity data set shows the existence of microbial populations is not damaged and their genomic DNA remains pre- that are unlikely to be active and growing in situ,but served. So far,thermophilic bacteria that have been found may have been introduced by lateral subsurface £ow. in high numbers in cold marine and terrestrial sediments The deep subsurface is not a hermetically sealed and static belong to spore-forming bacterial genera,the anaerobic environment that encloses its microbial populations in per- sulfate-reducing genus Desulfotomaculum [46] and the petuity,but instead is subject to exchange processes,in aerobic genus Geobacillus [47]. The long-term survival of particular with the deep-sea sediment surface. spore-forming bacteria in cold environments is consistent with the observation that thermophile-containing samples from hot environments can be maintained for long periods Acknowledgements at low temperatures without loss of viability [48]. Similar long-term survival may apply for the archaeal genera This research used samples and data provided by the Thermococcus and Pyrococcus; cultures can be maintained Ocean Drilling Program (ODP). ODP is sponsored by in liquid culture at 4‡C for years,and tolerate even oxygen the U.S. National Science Foundation (NSF) and partic- exposure at cool temperatures to some extent [49]. Possi- ipating countries under management of Joint Oceano- bly,the Thermococcus populations in the Nankai subsur- graphic Institutions (JOI),Inc. Funding for this research face sediments have been introduced by deep,lateral sea- was provided by the NASA Astrobiology Institute (Sub- water in£ux that replenishes pore water sulfate levels surface Biospheres) and National Science Foundation below 325 mbsf in the sediment column of Hole 1176A (Lexen Grant OCE-9978310). D.C.S. was also supported [19]. Horizontal £uid £ow through continuous channels, by the Joint Oceanographic Institutions/US Science Sup- from the accretionary prism seawards to the deformation port Program. The authors thank Barry Cragg,Yuki Mu- front and centered around the de¤collement depth,has been rakami,Stefan Sievert and Steven D’Hondt for their help inferred for the Nankai system by comparisons of chloride during this project. This is Woods Hole Oceanographic pro¢les at ODP Sites 808,1174,and 1173,on both sides of Institution contribution No. 10914. the deformation front [50]; £uid velocities were estimated as ca. 13 cm/year [50]. The Thermococcus clones were obtained from the meth- References ane-harboring sediment layers,where methanogenic archaea would have been expected. This conspicuous ab- [1] Whitman,W.B.,Coleman,D.C. and Wiebe,W.J. (1998) Prokary- sence of methanogen-a⁄liated sequences in these deep otes: the unseen majority. Proc. Natl. Acad. Sci. USA 95,6578^ subsurface 16S rDNA clone libraries is not unprecedented. 6583. A sequencing survey of methane hydrate-rich sediments in [2] Parkes,R.J.,Cragg,B.A. and Wellsbury,P. (2000) Recent studies on bacterial populations and processes in subsea£oor sediments: A re- the Cascadia Margin did not detect methanogens with view. Hydrogeol. J. 8,11^28. general archaeal 16S rDNA primers,but found cold-water [3] D’Hondt,S.,Rutherford,S. and Spivack,A.J. (2002) Metabolic ac- Crenarchaeota otherwise described from seawater [51]. tivity of subsurface life in deep-sea sediments. Science 295,2067^ Only PCR ampli¢cation of the highly conserved methyl 2070. coenzyme M reductase gene (mcrA),a speci¢c key gene [4] Wellsbury,P.,Mather,I. and Parkes,R.J. (2002) Geomicrobiology of deep,low organic carbon sediments in the Woodlark Basin,Paci¢c of methanogenesis,indicated the presence of methanogen- Ocean. FEMS Microbiol. Ecol. 42,59^70. ic archaea [51]. Such results could re£ect di¡erences in the [5] Cragg,B.A.,Parkes,R.J.,Fry,J.C.,Weightman,A.J.,Rochelle,P.A. proportions of methanogens and other archaea in the sedi- and Maxwell,J.R. (1996) Bacterial populations and processes in sediments. ments containing gas hydrates (ODP Leg 146: Cascadia Margin). Earth Planet. Sci. Lett. 139,497^507. [6] Parkes,R.J.,Cragg,B.A.,Bale,S.J.,Goodman,K. and Fry,J.C. 4.4. Conclusions (1995) A combined ecological and physiological approach to studying sulphate reduction within deep marine sediment layers. J. Microbiol. The Nankai subsurface sediments harbor bacterial Methods 23,235^249. clones without close cultured relatives,indicating that [7] Bale,S.J.,Goodman,K.,Rochelle,P.A.,Marchesi,J.R.,Fry,J.C., our knowledge of deep subsurface bacterial populations Weightman,A.J. and Parkes,R.J. (1997) Desulfovibrio profundus sp. nov.,a novel barophilic sulphate-reducing bacterium from deep sedi- and their physiologies is inadequate at best. Inferences ment layers in the Japan Sea. Int. J. Syst. Bacteriol. 47,515^521. based on comparisons with cultured species become [8] Barnes,S.P.,Bradbrook,S.D.,Cragg,B.A.,Marchesi,J.R.,Weight- more and more unreliable with increasing phylogenetic man,A.J.,Fry,J.C. and Parkes,R.J. (1998) Isolation of sulfate-re-

FEMSEC 1525 9-7-03 124 K.A. Kormas et al. / FEMS Microbiology Ecology 45 (2003) 115^125

ducing bacteria from deep sediment layers of the Paci¢c Ocean. Geo- that produces ¢lamentous sulfur. Appl. Environ. Microbiol. 68,316^ microbiol. J. 15,67^83. 325. [9] To⁄n,L. and Prieur,D. (2002) Cultivable microbial populations in [25] Jeanthon,C. (2000) Molecular ecology of hydrothermal vent micro- deep marine sediments (abstract). Geochim. Cosmochim. Acta 66, bial communities. Antonie van Leeuwenhoek 77,117^133. A778. [26] Teske,A.,Hinrichs,K.-U.,Edgcomb,V.,de Vera Gomez,A.,Ky- [10] Balkwill,D.L.,Reeves,R.H.,Drake,G.R.,Reeves,J.Y.,Crocker, sela,D.,Sylva,S.P.,Sogin,M.L. and Jannasch,H.W. (2002) Micro- F.H.,King,M.B. and Boone,D.R. (1997) Phylogenetic character- bial diversity of hydrothermal sediments in the Guyamas Basin: evi- ization of bacteria in the subsurface microbial culture collection. dence for anaerobic methanotrophic communities. Appl. Environ. FEMS Microbiol. Rev. 20,201^216. Microbiol. 68,1994^2007. [11] Rochelle,P.A.,Fry,J.C.,Parker,R.J. and Weightman,A.J. (1992) [27] Chandler,D.P.,Brockman,F.J.,Bailey,T.J. and Fredrickson,J.K. DNA extraction for 16S ribosomal-RNA gene analysis to determine (1998) Phylogenetic diversity of Archaea and Bacteria in a deep sub- genetic diversity in deep sediment communities. FEMS Microbiol. surface paleosol. Microb. Ecol. 36,37^50. Lett. 100,59^65. [28] Maymo¤-Gatell,X.,Chien,Y.-T.,Gossett,J.M. and Zinder,S.H. [12] Rochelle,P.A.,Cragg,B.A.,Fry,J.C.,Parkes,R.J. and Weightman, (1997) Isolation of a bacterium that reductively dechlorinates tetra- A.J. (1994) E¡ect of sample handling on estimation of bacterial di- chloroethene to ethene. Science 276,1568^1571. versity in marine sediments by 16S rRNA gene sequence analysis. [29] Phelps,C.D.,Kerkhof,L.J. and Young,L.Y. (1998) Molecular char- FEMS Microbiol. Ecol. 15,215^226. acterization of a sulfate-reducing consortium which mineralizes ben- [13] Reed,D.W.,Fujita,Y.,Delwiche,M.E.,Blackwelder,D.B.,Sheri- zene. FEMS Microbiol. Ecol. 27,269^279. dan,P.P.,Uchida,T. and Colwell,F.S. (2002) Microbial commun- [30] Li,L.,Kato,C. and Horikoshi,K. (1999) Microbial diversity in ities from methane hydrate-bearing deep marine sediments in a Fore- sediments collected from the deepest cold-seep area,the Japan arc basin. Appl. Environ. Microbiol. 68,3759^3770. Trench. Mar. Biotechnol. 1,391^400. [14] Chapelle,F.H.,O’Neill,K.,Bradley,P.M.,Methe ¤,B.A.,Ciufo,S.A., [31] Lanoil,B.D.,Sassen,R.,La Duc,M.T.,Sweet,S.T. and Nealson, Knobel,L.L. and Lovley,D.R. (2002) A hydrogen-based subsurface K.H. (2001) Bacteria and Archaea physically associated with gulf of microbial community dominated by methanogens. Nature 415,312^ Mexico gas hydrates. Appl. Environ. Microbiol. 67,5143^5153. 315. [32] Tanner,M.,Goebel,B.M.,Dojka,M.A. and Pace,N.R. (1998) Spe- [15] Marchesi,J.R.,Weightman,A.J.,Cragg,B.A.,Parkes,J.R. and Fry, ci¢c ribosomal DNA sequences from diverse environmental settings J.C. (2001) Methanogen and bacterial biodiversity and distribution in correlate with experimental contaminants. Appl. Environ. Microbiol. deep gas hydrate sediments from the Cascadia Margin as revealed by 64,3110^3113. 16S rRNA molecular analysis. FEMS Microbiol. Ecol. 34,221^228. [33] Summit,M. and Baross,J.A. (2001) A novel microbial habitat in the [16] Smith,D.C.,Spivack,A.J.,Fisk,M.R.,Haveman,S.A.,Staudigel, mid-ocean ridge subsea£oor. Proc. Natl. Acad. Sci. USA 98,2158^ H. and Leg 185 Shipboard Scienti¢c Party (2000) Methods for Quan- 2163. tifying Potential Microbial Contamination During Deep Ocean Cor- [34] Coolen,M.J.L.,Cypionka,H.,Sass,A.M.,Sass,H. and Overmann, ing. ODP Technical Note 28 [Online]. http://www-odp.tamu.edu/pub- J. (2002) Ongoing modi¢cation of Mediterranean pleistocene sapro- lications/tnotes/tn28/INDEX.HTM. pels mediated by prokaryotes. Science 296,2407^2410. [17] Smith,D.C.,Spivack,A.J.,Fisk,M.R.,Haveman,S.A.,Staudigel, [35] Hugenholtz,P.,Goebel,B.M. and Pace,N.R. (1998) Impact of cul- H. and ODP Leg 185 Shipboard Scienti¢c Party (2000) Tracer-based ture-independent studies on the emerging phylogenetic view of bac- Estimates of Drilling-Induced Microbial Contamination of Deep Sea terial diversity. J. Bacteriol. 180,4765^4774. Crust. Geomicrobiol. J. 17,207^219. [36] Gebers,R. (1981) Enrichment,isolation,and emended description of [18] Seno,T.,Stein,S. and Gripp,A.E. (1993) A model for the motion of Pedomicrobium ferrugineum Aristovskaya and Pedomicrobium manga- the Philippine sea plate consistent with Nuvel-1 and geological data. nicum Aristovskaya. Int. J. Syst. Bacteriol. 31,302^316. J. Geophys. Res. Sol. Earth 98,17941^17948. [37] Inagaki,F.,Sakihama,Y.,Inoue,A.,Kato,C. and Horikoshi,K. [19] Moore,G.F.,Taira,A.,Klaus,A.,Becker,L.,Boeckel,B.,Cragg, (2002) Molecular phylogenetic analyses of reverse-transcribed bacte- B.A.,Dean,A.,Fergusson,C.L.,Henry,P.,Hirano,S.,Hisamitsu, rial rRNA obtained from deep-sea cold seep sediments. Environ. T.,Hunze,S.,Kastner,M.,Maltman,A.J.,Morgan,J.K.,Muraka- Microbiol. 4,277^286. mi,Y.,Sa¡er,D.M.,Sa ¤nchez-Go¤mez,M.,Screaton,E.J.,Smith, [38] Li,L.,Kato,C. and Horikoshi,K. (1999) Bacterial diversity in deep- D.C.,Spivack,A.J.,Steurer,J.,Tobin,H.J.,Ujiie,K.,Underwood, sea sediments from di¡erent depths. Biodiv. Conserv. 8,659^677. M.B. and Moyra Wilson,M. (2001) New insights into deformation [39] Li,L.,Guenzennec,J.,Nichols,P.,Henry,P.,Yanagibayashi,M. and £uid £ow processes in the Nankai Trough accretionary prism: and Kato,C. (1999) Microbial diversity in Nankai Trough Sediments Results of Ocean Drilling Program Leg 190. G3 2,10.129/ at a depth of 3,843 m. J. Oceanogr. 55, 635^642. 2001GC000166. [40] Percival,S.L.,Knapp,J.S.,Edyvean,R. and Wales,D.S. (1998) [20] Maidak,B.L.,Cole,J.R.,Lilburn,T.G.,Parker Jr.,C.T.,Saxman, Bio¢lm development on stainless steel in mains water. Water Res. P.R.,Farris,R.J.,Garrity,G.M.,Olsen,G.J.,Schmidt,T.M. and 32,243^253. Tiedje,J.M. (2001) The RDP-II (Ribosomal Database Project). Nu- [41] Matsuda,N.,Agui,W.,Tougou,T.,Sakai,H.,Ogino,K. and Abe, cleic Acids Res. 29,173^174. M. (1996) Gram-negative bacteria viable in ultra pure water: iden- [21] Swo¡ord,D.L. (2000) PAUP*. Phylogenetic analysis using parsi- ti¢cation of bacteria isolated from ultra pure water and e¡ect of mony (and other methods),version 4,CD-ROM. Sinauer Associates, temperature on their behaviour. Coll. Surf. B Biointerfaces 5,279^ Sunderland,MA. 289. [22] Posada,D. and Crandall,K.A. (1998) MODELTEST: testing the [42] Kelley,S.,Baross,J.A. and Delaney,J.R. (2002) Volcanoes,£uids model of DNA substitution. Bioinformatics 14,817^818. and life at Mid-Ocean Ridge spreading centers. Annu. Rev. Earth [23] Stolz,J.F.,Ellis,D.J.,Blum,J.S.,Ahmann,D.,Lovley,D.R. and Planet. Sci. 30,385^491. Oremland,R.S. (1999) Sulfurospirillum barnesii sp. nov. and Sulfuro- [43] Inagaki,F.,Takai,K.,Komatsu,T.,Kanamatsu,T.,Fujiioka,K. spirillum arsenophilum sp. nov.,new members of the Sulfurospirillum and Horikoshi,K. (2001) Archaeology of Archaea: geomicrobiolog- clade of the epsilon Proteobacteria. Int. J. Syst. Bacteriol. 49,1177^ ical record of Pleistocene thermal events concealed in a deep-sea 1180. subsea£oor environment. Extremophiles 5,385^392. [24] Wirsen,C.O.,Sievert,S.M.,Cavanaugh,C.M.,Molyneaux,S.J.,Ah- [44] Summit,M. and Baross,J.A. (1998) Thermophilic subsea£oor micro- mad,A.,Taylor,L.T.,De Long,E.F. and Taylor,C.D. (2002) Char- organisms from the 1996 North Gorda Ridge Eruption. Deep-Sea acterization of an autotrophic sul¢de-oxidizing marine Arcobacter sp. Res. II 45,2751^2766.

FEMSEC 1525 9-7-03 K.A. Kormas et al. / FEMS Microbiology Ecology 45 (2003) 115^125 125

[45] Zillig,W. and Reysenbach,A.L. (2001) Thermococci class. nov. In: [49] Jannasch,H.W.,Wirsen,C.O.,Molyneaux,S.J. and Langworthy, Bergey’s Manual of Systematic Bacteriology (Boone,D.R. and Cas- T.A. (1992) Comparative physiological studies on hyperthermophilic tenholz,R.W.,Eds.),Vol. 1,pp. 341^348. Springer,New York. archaea isolated from deep-sea hot vents with emphasis on Pyrococ- [46] Isaksen,M.F.,Bak,F. and Jrgensen,B.B. (1994) Thermophilic sul- cus strain GB-D. Appl. Environ. Microbiol. 58,3472^3481. fate-reducing bacteria in cold marine sediment. FEMS Microbiol. [50] Spivack,A.J.,Kastner,M. and Ransom,B. (2002) Elemental and Ecol. 14,1^8. isotopic chloride geochemistry and £uid £ow in the Nankai Trough. [47] Marchant,R.,Banat,I.M.,Rahman,T.J. and Berzano,M. (2002) Geophys. Res. Lett. 29,10.1029/2001GL014122. The frequency and characteristics of highly thermophilic bacteria in [51] Bidle,K.A.,Kastner,M. and Bartlett,D.H. (1999) A phylogenetic cool soil environments. Environ. Microbiol. 4,595^602. analysis of microbial communities associated with methane hydrate [48] Wiegel,J. (1986) Methods for isolation and study of thermophiles. containing marine £uids and sediments in the Cascadia margin (ODP In: Thermophiles: General,Molecular and Applied Microbiology site 829B). FEMS Microbiol. Lett. 177,101^108. (Brock,T.D.,Ed.),pp. 17^37. Wiley,New York.

FEMSEC 1525 9-7-03