Bacterial Community Composition in Different Sediments from the Eastern Mediterranean Sea: a...

MicrobialEcology

Bacterial Community Composition in Different Sediments fromthe Eastern Mediterranean Sea: a Comparison of Four 16SRibosomal DNA Clone Libraries

Paraskevi N. Polymenakou1,2, Stefan Bertilsson3, Anastasios Tselepides1

and Euripides G. Stephanou2

(1) Hellenic Center for Marine Research, Gournes Pediados, GR 71003 Heraklion, Crete, Greece(2) Environmental Chemical Processes Laboratory, Department of Chemistry, University of Crete, 71409 Heraklion, Greece(3) Limnology/Department of Ecology and Evolution, Evolutionary Biology Centre, Uppsala University, Norbyv. 20, SE-75236 Uppsala, Sweden

Received: 7 January 2005 / Accepted: 9 March 2005 / Online publication: 23 November 2005

Abstract

The regional variability of sediment bacterial communitycomposition and diversity was studied by comparativeanalysis of four large 16S ribosomal DNA (rDNA) clonelibraries from sediments in different regions of theEastern Mediterranean Sea (Thermaikos Gulf, CretanSea, and South lonian Sea). Amplified rDNA restrictionanalysis of 664 clones from the libraries indicate that therDNA richness and evenness was high: for example, anear-1:1 relationship among screened clones and numberof unique restriction patterns when up to 190 cloneswere screened for each library. Phylogenetic analysis of207 bacterial 16S rDNA sequences from the sedimentlibraries demonstrated that Gamma-, Delta-, andAlphaproteobacteria, Holophaga/Acidobacteria, Plancto-mycetales, Actinobacteria, Bacteroidetes, and Verruco-microbia were represented in all four libraries. A fewclones also grouped with the Betaproteobacteria, Nitro-spirae, Spirochaetales, Chlamydiae, Firmicutes, and can-didate division OPl 1. The abundance of sequencesaffiliated with Gammaproteobacteria was higher in li-braries from shallow sediments in the Thermaikos Gulf(30 m) and the Cretan Sea (100 m) compared to thedeeper South Ionian station (2790 m). Most sequences inthe four sediment libraries clustered with uncultured 16SrDNA phylotypes from marine habitats, and many ofthe closest matches were clones from hydrocarbon seeps,benzene-mineralizing consortia, sulfate reducers, sulkoxidizers, and ammonia oxidizers. LIBSHUFF statisticsof 16S rDNA gene sequences from the four librariesrevealed major differences, indicating either a very high

richness in the sediment bacterial communities orconsiderable variability in bacterial community compo-sition among regions, or both.

Introduction

It is notoriously difficult to isolate and cultivate bacterialpopulations representative of natural bacterial commu-nities and this limitation is believed to be even moresevere in environments characterized by extreme environ-mental conditions [9]. Hence, our current understandingof microbial biodiversity in marine environments re-lies to a large extent on the development and use ofculture-independent molecular methods that can provideinformation on the phylogeny and distribution of non-cultivable microorganisms [6, 27]. Sequence analysis ofpolymerase chain reaction (PCR)-amplified and cloned16S ribosomal RNA genes (16S rDNA) is a widely usedapproach to assess microbial diversity and communitycomposition in environmental samples [10, 63]. Thisapproach has been used to describe the composition ofbacterial communities in several marine sediments, butmost of these libraries contain relatively few clones (e.g.,[24, 31, 32, 61]) and, therefore, provide limited infor-mation on the composition and diversity of the usuallyvery complex natural microbial communities found insediments. On the other hand, these libraries haverevealed that sequences related to the Alpha-, Beta-,Gamma-, and Deltaproteobacteria, Gram-positives,Cytophaga–Flavobacterium–Bacteroidetes, Planctomyce-tales, Actinobacteria, as well as Verrucomicrobia likelyoccur at high frequencies in most sediments. More ex-tensive screening of a few sediment 16S rDNA cloneCorrespondence to: Stefan Bertilsson; E-mail: [email protected]

DOI: 10.1007/s00248-005-0005-6 & Volume 50, 447–462 (2005) & * Springer Science+Business Media, Inc. 2005 447

libraries have also demonstrated that additional bacterialgroups can be abundant; for example, the recently intro-duced phylum Gemmatimonadetes (previously known ascandidate division KS-B) [65], the new candidate divi-sion KS-A [35], and clones affiliated with the candidatedivisions OP 3, OP 8, and OP 11 [10, 56]. Thus, it can beexpected that the analysis of more such large clonelibraries will reveal other, hitherto undiscovered groupsand divisions.

Analyses of multiple, large, clone libraries with simi-lar techniques enable comparisons to be made amongbacterial communities in different samples as well asidentification of widely distributed sediment bacterialphylotypes. A comparison of bacterial phylotypes (98%similar 16S rDNA) in three separate libraries constructedfrom different horizons (0- to 0.4-, 1.5- to 2.5-, and 20- to21-cm depth) in an Antarctic continental shelf sedimentdemonstrated that 76 out of a total 496 phylotypes werepresent in all three libraries [10]. A similar comparison ofmedium-sized 16S rDNA clone libraries (69–130 clones)from five Antarctic lake sediments and a coastal sedimentfrom the same area also indicates that sediments werehighly dissimilar even if sediment libraries from geo-chemically similar lakes shared several common phylo-types [11]. Knowing the phylogeny and occurrence ofdifferent bacterial populations in sediments is a first steptoward understanding the biogeochemical function andecology of these largely uncultivated microbial commu-nities as well as the extent of genetic diversity that can befound in sediment bacteria across the globe [10]. How-ever, with the exception of some functionally definedbacterial groups (methanogens, methanotrophs, sulfate-reducing bacteria), our knowledge about the ecology andphysiology of sediment bacteria is fragmentary due tothe lack of cultured representatives for many groups aswell as the lack of extensive biogeographical studies ofsediment bacterial populations [10].

The eastern basin of the Mediterranean Sea isconsidered to be one of the world’s most oligotrophicareas and is characterized by an overall nutrient deficit[28, 60] and extremely low primary productivity [19, 44].As a result, minute amounts of organic matter reach thesea floor [15–17] and the growth and abundance ofbenthic bacteria is likely constrained by the input of freshorganic material from the pelagic zone [18]. Further-more, some regions of the Eastern Mediterranean Sea arehighly exposed to anthropogenic influence and thereforecontain many steep pollutant concentration gradients[23, 42]. Until now, few attempts have been made todescribe the diversity and composition of sedimentbacterial communities in the Eastern MediterraneanSea. The only available information has been acquiredfrom sites exposed to extreme physicochemical condi-tions, such as (1) a hydrothermal vent area near MilosIsland [12, 50, 51], (2) the microbial community

inhabiting the chemocline of the hypersaline anoxicUrania basin in the Eastern Mediterranean Sea [48],and (3) the late Pleistocene organic-rich sediments(sapropels) southeast of Crete (Greece) [14]. There isalso a general lack of information regarding bacterialcommunities in the Western region of the MediterraneanSea, where the only available data were retrieved from16S rDNA analysis of mesocosms with MediterraneanSea water [45, 49] and marine bacterioplankton collectedoutside the Spanish coast [1, 4, 40].

Using multiple fingerprinting techniques [phospho-lipid linked fatty acid (PLFA) analysis and 16S rDNAgenotyping by denaturing gradient gel electrophoresis andterminal restriction fragment length polymorphism(T-RFLP)], we observed marked differences in bacterialcommunity composition in surface sediments fromvarious regions of the Eastern Mediterranean Sea [42].Bacterial communities from sediment of the Northern,more productive regions of the Thermaikos Gulf weresignificantly separated from the oligotrophic regions ofthe Cretan, South Ionian, and Levantine seas. Further-more, communities of deep sediments (91494 m depth)were clearly separated from their shallow (G617 m)counterparts. We could also correlate the communitycomposition to environmental-state variables such assediment carbon content and chlorophyll, whereas petro-leum contamination was less important in this regard[42]. The present study complements this survey inproviding a detailed comparative 16S rDNA analysis offour of these sediment bacterial communities obtainedfrom various regions of the oligotrophic Eastern Medi-terranean Sea.

Materials and Methods

Sampling Sites and Geochemical Characteristics. Sedi-ment samples were collected from three regions in theEastern Mediterranean Sea (Thermaikos Gulf, CretanSea, and South Ionian Sea; Fig. 1). A Bowers andConnelly multiple corer (eight cores, i.d. 9.0 cm) [8]was used to collect undisturbed sediment samples fromthe Thermaikos Gulf in February 2002 and from theCretan Sea in July 2002. Sediment from the South IonianSea was collected in November 2001, using an USNELBoxcorer. All sampling was carried out on board the R/Vs Aegaeo and Philia. Mixed surface sediment samples(0–1 cm) from each of the four sampled sediments werestored frozen in sterile plastic vials for subsequentanalysis. Sediment chlorophyll a [33] and total organiccarbon concentration [25] were determined using aTurner TD-700 fluorometer and a PerkinElmer CHN2400 analyzer, respectively. Petroleum hydrocarbons(H/C) and n-alkanes were analyzed according to Gogouet al. [22], whereas PLFAs were analyzed as described byPolymenakou et al. [42].

448 P.N. POLYMENAKOU ET AL.: BACTERIAL DIVERSITY IN MEDITERRANEAN SEDIMENTS

DNA Extraction, PCR, Cloning, and RFLP Screen-

ing. Total DNA was extracted from sediments usingthe FastDNA-Spin Kit for Soil (Q-BIOgene, Carlsbad,CA, USA) as previously described [42]. Bacterial 16SrDNA genes were amplified from mixed genomic sam-ples by using PCR with the universal bacterial primer 27fmodified to match also Planctomycetales (50-AGRGTTT-GATCMTGGCTCAG-30) [62] and 1492r (50-GGY-TACCTTGTTACGACTT-30) [30]. PCR conditions weredesigned to minimize bias [43, 63]. For each sample,eight replicate PCR reactions of 30 mL were amplified ina Stratagene Robocycler with initial denaturation at 94-Cfor 3 min followed by 25 cycles of 1 min at 94-C, 1 minannealing at 55-C, 3 min primer extension at 72-C, and afinal extension at 72-C for 7 min. Each tube contained1–4 ng of target DNA, PCR buffer [10 mM Tris–HCl(pH 9), 50 mM KCl, 0.1% Triton X-100, and 2 mMMgCl2], 100 nM of each primer, 200 mM of each deoxy-ribonucleotide triphosphate and 0.25 U Taq DNA poly-merase (Invitrogen, Carlsbad, CA, USA). Products fromeach of the eight PCR reactions were used as templates1:10 (v/v) in duplicate three-cycle reconditioning PCRreactions (16 reactions total) to eliminate heteroduplexformation that may introduce artificial diversity in clonelibraries [57]. All PCR products were pooled andprecipitated with ethanol and sodium acetate [47] fol-lowed by gel purification using the Qiaquick PCRpurification kit (Qiagen, Valencia, CA, USA). Theconcentration of PCR products generated from thedifferent sediment samples was determined by directcomparison to a Low DNA Mass Ladder (Invitrogen)using 2% agarose gel electrophoresis, ethidium bromide

staining, and UV transillumination. For each samplingsite, 5–10 ng of PCR product was cloned into the pCR4-TOPO vector and transformed into One shot TOP10chemically competent cells of Escherichia coli using theTOPO TA Cloning kit (Version M) as recommended bythe manufacturer (Invitrogen). At least 200 positiveclones from each clone library (selected by blue andwhite screening) were transferred to 96-well plates andincubated overnight at 37-C in Luria–Bertani mediumcontaining 50 mg kanamycin mLj1. Aliquots of theindividual clones were (1) archived at j80-C in 7%dimethyl sulfoxide or (2) washed by pelletizing cells in a30-min centrifugation at 10,000 � g followed by super-natant removal by low-speed centrifugation (G500 rpm)of inverted plates. Pelletized cells were resuspended in30 mL sterile and UV-irradiated MQ-grade water. Cellswere lysed by heating at 98-C for 10 min followed byagitation. The lysates were used (1:10 v/v) as templates ina PCR amplification of the insert using external (vector)primers M13f-20 (50-GTAAAACGACGGCCAG-30) andM13r (50-CAGGAAACAGCTATGAC-30; Invitrogen) toavoid co-amplification of E. coli host-cell DNA. PCRamplification was carried out for 25 cycles as describedbefore. Positive transformants (clones carrying an insertof correct size) were identified by agarose gel electro-phoresis as described above. Aliquots (5 mL) of indi-vidual PCR products were digested with two four-cuttingrestriction enzymes (HhaI and HaeIII) for 16 h accordingto instructions supplied by the manufacturer (Invitro-gen). After inactivation of the enzymes (20 min at 85-C),fragments were sized by electrophoresis on a 2% agarosegel (2.5 h, 80 V, 10-C). Fragments were recorded usingethidium bromide staining and UV transillumination. A100-bp DNA ladder (Invitrogen) was used for determi-nation of fragment size. The resulting RFLP patternswere then used to classify clones into operational taxo-nomic units (OTUs).

Sequencing and Phylogenetic Analysis. A total of219 clones from randomly chosen OTUs were sequencedon an ABI 3700 96 capillary sequencer (Applied Bio-systems) using primer 27f [62] and the BigDye terminatorkit v.3.1 (Applied Biosystems). This generated high-quality reads of between 450 and 780 bases. Ten of thesequences were identified as likely chimeric moleculesusing the Chimera Check software included in the Ribo-somal Database Project II [36] and these were excludedfrom further analysis. The remaining 209 sequences werecompared to GenBank entries by using basic local align-ments tool (BLAST) [5] to obtain a preliminary phylo-genetic affiliation of the clones. Two sequences wereaffiliated with eukaryote organelles and were excludedfrom further analysis. The remaining 207 clones (63 fromstation Therm01, 45 from station Thenn30, 49 fromstation Cretal, and 50 from station S.Ionian), were used

Figure 1. Geographic location of sampled sediment stations in theEastern Mediterranean Sea.

P.N. POLYMENAKOU ET AL.: BACTERIAL DIVERSITY IN MEDITERRANEAN SEDIMENTS 449

for phylogenetic analysis. All sequences were imported tothe ARB software Version 2.5b [54] and aligned using theintegrated aligner tool and the fast aligner option followedby manual alignment of the sequences to closely relatedsequences in the ARB database. Phylogenetic trees wereconstructed in ARB using maximum likelihood [21]. Therobustness of tree topologies was confirmed by maximumparsimony analysis [54] with 100 bootstrap replications[20]. Novel clusters of uncultured sediment bacteria weredefined as monophyletic groups of 16S rDNA sequenceswith a minimum sequence similarity of 90%. Additionalcriteria were that the cluster should contain at least threesequences from a minimum of two libraries.

Analysis of Species Richness and Clone Library

Similarity. For each clone library, the RFLP-baseddistribution of clones in different OTUs was used to es-timate species richness using the Web-based Rarefactioncalculator software (http://www2.biology.ualberta.ca/jbrzusto/rarefact.php). Species richness was estimatedusing the nonparametric Chao estimator S*

1 = Sobs +(a2/2b), where Sobs is the number of 16S rDNA clonesobserved, a is the number of clones observed just once,and b is the number of clones observed twice [13]. Thestandard deviation (SD) was estimated using the equa-tion SD = b[(a/4b)4 + (a/b)3 + (a/2b)2]. To determinethe significance of differences between two clone libraries(e.g., X and Y), differences (DC) between Bhomologous^CX(D) and Bheterologous^ coverage curves CXY(D) werecalculated using the LIBSHUFF software (http://libshuff.mib.uga.edu/ [52]). The Bhomologous^ coverage ofclone library X is calculated using the equation CX =1 j (NX/n), where NX is the number of unique sequencesin the sample and n is the total number of sequences. Ina similar way, the Bheterologous^ coverage of clonelibrary X by a second clone library Y is defined as:CXY = 1 j (NXY/n), where NXY is the number ofsequences in clone library X not found in the secondclone library Y and n is the number of sequences in X.Both NX and NXY can be defined at different levels ofevolutionary distance (D), e.g., homology of the se-quenced 16S rDNA fragments, to generate a coveragecurve. If clone libraries are similar, then the coveragecurves CX(D) and CXY(D) are also expected to be similar.

The significance of DC is described by P, which iscalculated by randomly shuffling sequences (e.g., 999times) and estimating DC after each shuffling. Therandomized values plus the empirical value of DC areranked from largest to smallest, and then the P value isestimated to be r/(N + l), where r denotes the rank of theempirical value of DC [52].

Nucleotide Sequence Accession Numbers. The207 partial 16S rDNA sequences generated in the pres-ent study were deposited in GenBank under accessionnumbers AY533880-AY534086.

Results

Sediment Characteristics. Sediment sampling siteswere located at variable depths (30–2790 m) in various re<gions of the Eastern Mediterranean Sea (Tables 1 and 2,Fig. 1). A detailed description of the biogeochemicalcharacterization of the sediments has been presentedelsewhere [42]. Briefly, the depths of the samplingstations varied from 30–100 m for the Thermaikos Gulfand the Cretan Sea, whereas the sediment from theSouth Ionian Sea was collected from 2790 m depth.Chlorophyll a, total organic carbon, petroleum H/C, andn-alkane levels were highest in station Therm01 (7.68 mgg

_1, 1.47%, 17.86 mg g_1, and 3.31 mg g

_1, respectively),whereas the lowest levels were recorded in sedimentsfrom station Cretal (0.32 mg g

_1, 0.45%, 0.67 mg g_1, and

0.15 mg g_1, respectively). Stations Therm30 and

S.Ionian, were characterized by intermediate levels oforganic carbon (0.53 T 0.04 and 0.79 T 0.07%) andpetroleum H/C (1.5 and 1.9 mg g

_1), whereas the PLFAcontent was much higher in the station located in theThermaikos Gulf (Therm01; 2.26 mg g

_1) compared tothe other three sampling sites (Table 1).

Taxonomic Groups and Their Distribution. All fourbacterial 16S rDNA clone libraries were diverse andincluded sequences affiliated with most classes or phylapreviously detected in marine sediments (Table 3).Among the 207 bacterial 16S rDNA sequences used forphylogenetic analyses, 37% shared 86–92% sequencesimilarity with any cultured bacterial strain. The remain-

Table 1. Biogeochemical characteristics of the sampling stations

Therm01 Therm30 Creta1 S.Ionian

Region Thermaikos Gulf Thermaikos Gulf Cretan Sea South Ionian SeaStation character Shallow mesotrophic Shallow mesotrophic Shallow oligotrophic Deep oligotrophicChlorophyll a (2g gj1) 7.68 T 2.75 1.22 T 0.07 0.32 T 0.06 0.05 T 0.001Organic carbon (%) 1.47 T 0.12 0.53 T 0.04 0.45 T 0.06 0.79 T 0.07Petroleum H/C (2g gj1) 17.86 1.52 0.67 1.92n-Alkanes (2g gj1) 3.31 0.39 0.15 0.65PLFA (2g gj1) 2.26 0.22 0.62 0.13


ing clones were also distinct from any cultured bacterialspecies (G86% sequence similarity). Pairwise compari-sons of all sequences in the four libraries showed highheterogeneity in the libraries as only 28 sequences (14%)shared more than 92% similarity to any of the other

sequences or any sequence published in GenBank. Threenovel clusters were identified within the Gammaproteo-bacteria and the Planctomycetales (Fig. 2A, E). Phyloge-netic analysis of the partial 16S rDNA revealed thatsequences grouped mainly with Gamma-, Alpha-, and

Table 2. Comparison of total screened clones, number of phylotypes, estimated total sequence richness, and coverage for large 16SrDNA clone libraries for sediment bacteria

Sampling areaSedimentdepth (m)

No. ofsampling

sitesNo. ofclones Phylotypes

Restrictionenzymes

Speciesrichness

%Coverage Data source

This studyThermaikos

Gulf (Therm01)30 1 94 80 HhaI/HaeIII 440 18.2 This study

ThermaikosGulf (Therm30)

86 1 190 165 HhaI/HaeIII 958 17.2 This study

Cretan Sea (Creta1) 100 1 190 152 HhaI/HaeIII 478 31.8 This studySouth Ionian

Sea (S.Ionian)2790 1 190 171 HhaI/HaeIII 1306 13.1 This study

Previous studiesContinental Shelf,

Antarctica761 3 1046 496 – 4350 õ22.6–35.6 Bowman and

McCuaig [10]Eastern Antarctica,

Vestfolds HillsCoastal 6 555 202 NciIHaeIII/RsaI/HinfI – õ15–36 Bowman

et al. [11]Arctic Ocean,

SpitsbergenCoastal 1 353 140 HaeIII – 71.95 Ravenschlag

et al. [46]Deep Sea

sediments1159–6379 7 149 75 RsaI/MspI – – Li et al. [31, 32]

French Guiana,S. America

Coastal 1 96 63 HinfI/HaeIII/DdeI/HhaI – – Madridet al. [35]

Sagami Bay,Japan

1159 1 77 57 HhaI/RsaI/HaeIII õ126 45 Urakawaet al. [61]

Sagami Bay,Japan


Tokyo Bay,Japan


The sediment depths at the sampling site, number of different samples/sublibraries, RFLP screening procedure, Chao-1 species richness estimator, and thecoverage of the libraries are included. Coverage was estimated by dividing the number of unique phylotypes detected from the RFLP screening with the totalspecies richness estimate obtained using the Chao-1 procedure.

Table 3. 16S rDNA phylotype distribution in the sediment clone libraries Therm01, Therm30, Creta1, and S.Ionian and estimationsfor the combined clone libraries for the Eastern Mediterranean Sea sediments

Taxonomic groups

% Phylotypes in the clone libraries Estimations for all libraries

Therm01 Therm30 Creta1 S.Ionian Eastern Mediterranean

Beta–Gamma–Proteobacteria 31.7 28.9 36.7 20.0 29.5Deltaproteobacteria 23.8 17.8 12.2 12.0 16.9Acidobacteria 11.1 15.6 12.2 24.0 15.5Planctomycetales 7.9 13.3 10.2 14.0 11.1Alphaproteobacteria 3.2 8.9 8.2 4.0 5.8Bacteroidetes 4.8 4.4 2.0 2.0 3.4Actinobacteria 3.2 4.4 2.0 4.0 3.4Verrucomicrobia 0.0 2.2 6.1 4.0 2.9Chloroflexi 3.2 0.0 0.0 4.0 1.9Nitrospirae 0.0 0.0 6.1 2.0 1.9Chlamydiae 0.0 2.2 2.0 0.0 1.0Firmicutes 1.6 0.0 0.0 0.0 0.5Spirochaetales 1.6 0.0 0.0 0.0 0.5OP11 group 3.2 0.0 0.0 0.0 1.0Nonaffiliated groups 4.8 2.2 2.0 10.0 4.9


Deltaproteobacteria, Holophaga/Acidobacteria, Plancto-mycetales, Actinobacteria, Bacteroidetes, and Verruco-microbia, whereas Epsilonproteobacteria were absentfrom all the libraries. In addition, a few clones (fewerthan four per group) were affiliated with Betaproteo-bacteria, Nitrospirae, Chloroflexi, Spirochaetales, Chla-mydiae, Firmicutes, and the candidate division OP11.Finally, 10 clones could not be affiliated with any knownbacterial group (Table 3).

Gammaproteobacteria were frequently encounteredin all four clone libraries, representing 31.7, 28.9, 36.7,and 20% of the clones from Therm01, Therm30, Cretal,and S.Ionian libraries, respectively (Table 3). Fewsequences (G7%) were affiliated with established phy-logenetic groups containing cultured representatives,i.e., Thioalcalovibrio, Legionellales, Thiotrichales, Chro-matiaceae, Pseudoalteromonas, and Pseudomonadales(Fig. 2A). The remaining sequences grouped with clonesobtained using culture independent methods (Table 3,Fig. 2A). Sequences affiliated with Legionellales andPseudoalteromonas were only found in the Cretan Sealibrary, whereas Thioalcalovibrio and Chromatiaceae wereonly found in the Therm30 clone library. The libraryfrom the deeper sediment of the South Ionian Seadid not contain any sequences affiliated with establishedgroups containing cultured representatives (Fig. 2A).The sequence that grouped with Thioalcalovibriowas most closely related to the known sulfur oxidizers,Thioalcalovibrio denitrificans and Thioalcalovibrioversutus, whereas sequences affiliated with Legionellaleswere most closely related to sequences retrieved fromhydrocarbon seep sediments (Fig. 2A). The membersof Thiotrichales were most closely related to Beggiatoasp. A single clone grouped in the Betaproteobacteria(Fig. 2A).

Of the sequenced clones, 16.9% were affiliatedwith Deltaproteobacteria and grouped mainly in fourfamilies containing cultured representatives (i.e., Desul-fobulbaceae, Bdellovibrionaceae, Polyangiaceae, andDesulfobacteraceae (Table 3, Fig. 2B). Three closelyrelated sequences from library Therm01 were tentativelyidentified as Deltaproteobacteria but could not beaffiliated to any known group. The closest match inGenBank (94% similar) was the uncultured deltaproteo-bacterium clone Sva0103 originating from an Arcticsediment library [46]. One cluster (NB1) was observed

distinct from cultured species and was related toPolyangiaceae and Desulfobacteraceae (Fig. 2B). Sequen-ces affiliated with Desulfobulbaceae were most closelyrelated to hydrocarbon seep bacteria, and to Desulfocapsasulfexigens and Desulfofustis glycolicus. Several of thesequenced clones from all four sediments were mostclosely related to various sulfate-reducing bacteria. Therepresentatives of Desulfobacteraceae all originated fromthe Thermaikos Gulf and were closely related to eitherhydrocarbon seep sediment bacteria, bacteria forming abenzene-mineralizing consortium, and Desulfobacteriumindolicum (Fig. 2B).

Sequences affiliated with Alphaproteobacteria ac-counted for 5.8% of the total clones, but also variedlargely among the libraries (Table 3). Most clonesaffiliated with Alphaproteobacteria grouped in threefamilies (Fig. 2C). All sediments except the most con-taminated and productive station (Therm01; Table 1)contained sequences affiliated with Hyphomicrobiaceae,Rhodobacteraceae, and Methylocystaceae (Fig. 2C).

Sequences affiliated with Holophaga/Acidobacteriawere abundant in all four libraries (Table 3). Between 11and 16% of the sequences from the shallow stationsTherm01, Therm30, and Cretal were affiliated with thisgroup, whereas the corresponding number for the deepstation in South Ionian Sea was 24% (Table 3). Thisdivision was introduced as a new phylum in the domainBacteria in the late 1990s and is related to Planctomyce-tales and Chlamydiae [34].

Clones affiliated with the order Planctomycetales ac-counted for 11.1% of the total sequenced clones (7.9,13.3, 10.2, and 14% for libraries from Therm01,Therm30, Cretal, and S.Ionian, respectively; Table 3).The sequence divergence within the group was high anda new sediment cluster could be identified (Therm01/Cretal), with a minimum sequence similarity exceeding94% (Fig. 2E).

Only 3.4% of the total sequenced clones wereaffiliated with the phylum Bacteroidetes (Table 3). Fourdifferent phylotypes from the Thermaikos Gulf groupedwith the Flavobacteria and three of these were related tothe Antarctic sea ice bacterium clone SW17 (994%sequence similarity; Fig. 2F). One clone from the SouthIonian Sea grouped with a halophilic eubacterium EHBderived from a solar saltern and three phylotypesgrouped with Cytophaga clones obtained from the

Figure 2. Maximum likelihood 16S rDNA tree showing positions of phylotypes affiliated with (A) Betaproteobacteria andGammaproteobacteria, (B) Deltaproteobacteria, (C) Alphaproteobacteria, (D) Holophaga/Acidobacteria, (E) Planctomycetales, and (F)Chloroflexi, Firmicutes, Spirochaetales, Actinobacteria, Nitrospirae, Bacteroidetes, Chlamydiae, Verrucomicrobia, the candidate divisionOP11 from sediment clone libraries in the Eastern Mediterranean Sea. Partial 16S rDNA sequences obtained from the clone librariesTherm01, Therm30, Creta1, and S.Ionian and their closest matching entries in GenBank were included in the analysis. Bootstrap values atthe nodes (100 replications) were calculated using maximum parsimony. Values below 50% are not shown. The scale bar indicates 10%nucleotide change per 16S rDNA position. Sequences from cultured representatives are indicated by italics.


Figure 2. Continued.


deepest cold-seep area in the Japan Trench (accessionnos. AB015585 and AB015262) and from the Suruga Bay(accession no. AB015543) [31, 32].

Verrucomicrobia, Chloroflexi, Nitrospirae, Chlamy-diae, and Spirochaetales were minor components in eachof the four libraries (between 0.5 and 2.9%; Table 3).Clones affiliated with Verrucomicrobia were found in thelibraries from Therm30, Cretal, and S.Ionian, whereasthis phylum was absent in the polluted Therm01sediment. The clones were most closely related to 16SrDNA clones derived from an Arctic Ocean bacterio-plankton sample [7] (Fig. 2F). Sequences affiliated withthe phylum Nitrospirae were retrieved from the twomost hydrocarbon contaminated sediments (Cretal,S.Ionian) and grouped in the class Nitrospira (Fig. 2F).

Finally, two clones from the Thermaikos Gulf andthe Cretan Sea were affiliated with the phylum Chla-

mydiae, and a single clone obtained from the Thermai-kos Gulf was affiliated with the Spirochaetales (Fig. 2F).Few sequences from the four sediments grouped in thecandidate divisions hitherto only described usingculture-independent techniques, e.g., less than 1.0% ofthe total screened clones were affiliated with candidatedivision OP11 (Fig. 2F).

All libraries contained sequences affiliated with thephylum Actinobacteria (2.0–4.4%; high-GC Gram-positivebacteria). This group contained members of Acidimicro-biales and Actinomycetales (Fig. 2F). A single sequencefrom the Therm01 library was affiliated within the orderAcidimicrobiales, with the filamentous sludge bacte-rium Microthrix parvicella as its closest relative. Mostof the Actinobacteria sequences were affiliated withthe order Actinomycetales that also included a clone de-rived from an anoxic sediment located in the Arctic





Ocean (accession no. AJ241019) [46]. A single sequencefrom the contaminated Therm01 sediment grouped inthe phylum Firmicutes (low-GC Gram-positives) andwas most closely related to Bacillus benzeovorans [41](Fig. 2F).

Bacterial Diversity. The number (richness) andfrequency (evenness) of 16S rDNA-based phylotypes wereevaluated by RFLP analysis of between 94 and 190randomly selected clones from each of the four sedimentlibraries. We were only able to manually compare andclassify the RFLP patterns for single libraries (max. 190clones) since it was necessary to run out similar clonesBside by side^ to determine if they represented uniquepatterns. This pairwise screening was extremely time-consuming and due to the large size of the combinedlibrary (664 clones), no attempts were made to comparethe RFLP patterns between the different libraries. Speciesrichness based on RFLP banding patterns was thereforecalculated only for individual clone libraries. Eightydifferent OTUs (e.g., unique RFLP patterns) were iden-tified among the 94 screened clones from the Therm01library, whereas 165, 152, and 171 different OTUs were

determined from the 190 clones screened for each of thethree additional sediments (i.e., Therm30, Cretal, andS.Ionian; Table 1). The Chao-1 richness estimator washighest for the South Ionian Sea and station Therm30(1306 T 187 and 958 T 134 OTUs, respectively; Table 2),whereas lower values were obtained for Therm01 andCretal (440 T 88 and 478 T 60 OTUs, respectively; Table 2).However, all these estimates of total sequence richness areunderestimates since the screening of the libraries sugges-ted that our libraries only contained a minor fraction ofthe total bacterial 16S rDNAs and that nearly every newclone screened represented a novel RFLP pattern (Fig. 3).

In an attempt to determine the significance ofdifferences between the clone libraries based on availablesequence data, we applied LIBSHUFF statistics (Fig. 4). Acomparison of all libraries revealed that bacterial com-munity composition differed significantly between thefour sampling sites (P = 0.001 for each combination;Table 4; Fig. 4). More detailed information on differencesbetween clone libraries was obtained by examining thedistribution of (CX j CXY)2 as a function of evolutionarydistance (D). The coverage curves for representative pairsof clone libraries clearly show major differences also at





low levels of genetic distance (D 9 0.2; Fig. 4) and thesedifferences were even more obvious for sediments fromdifferent regions (e.g., Thermaikos Gulf and SouthIonian Sea; Fig. 4B). Due to the uncertainties associatedwith comparing RFLP banding patterns between the fourlibraries screened, we used previously published T-RFLPresults to compare the bacterial community compositionin the sediments [42]. The maximum similarity value forthe T-RFLP pattern was observed for the two adjacentstations within the Thermaikos Gulf (S = 0.612; Table 4),whereas these community fingerprints were least similarfor the Cretan and South Ionian seas (S = 0.384;Table 4), corroborating the results from the LIBSHUFFclone library comparisons.

Discussion

The observed differences in microbial community com-position among the sampled sediments are likely asso-ciated with their geographic location since differentwater masses are subject to contrasting environmentalinfluences. For example, station Therm01 is subject toanthropogenic influences from Thessaloniki harbor andalso receives a major input of riverine water from ahydroelectric dam construction [29]. The deeper SouthIonian Sea has high levels of petroleum hydrocarbonswith strong influences from the southern Adriatic andthe Levantine Sea [37].

Clones affiliated with Gammaproteobacteria oc-curred more frequently in the three shallow sedimentscompared to the deep, low-chlorophyll sediment of theSouth Ionian Sea. These more shallow sediments also hadmultiple clones affiliated with the novel cluster Therm/Cretal (Fig. 2A). Gammaproteobacteria were also thedominant bacterial group in the previously publishedclone libraries from Antarctic shelf sediments [10] andan Arctic sediment [46]. Sequences affiliated withDesulfobacteraceae were only observed in the two

libraries from the Thermaikos Gulf, possibly indicatingthat this family is mainly associated with the moreproductive and shallow sediments of this subbasin.

The libraries from the two sediments with thehighest petroleum hydrocarbon levels (Therm01 andS.Ionian) contained few alphaproteobacterial sequences(3.2 and 4% of total sequences, respectively), whereas the

Figure 3. Rarefaction analysis of 16S rDNAsequence heterogeneity in clone librariesfrom four geographically separated sedi-ments situated in the Eastern MediterraneanSea (Therm01, Therm30, Creta1, andS.Ionian). Total number of screened clonesare plotted against unique operationaltaxonomic units (OTUs) identified by RFLPscreening using two four-cutter restrictionenzymes (HhaI and HaeIII). Error bars(hidden in symbol) indicate the standarddeviation and the diagonal line representthe 1:1 relationship where each screenedclone is unique.

Figure 4. Results of selected LIBSHUFF comparisons of clonesfrom (A) Therm01 (X) to Thelm30 (Y ) and (B) Therm30 (X ) toS.Ionian (Y ) clone libraries. Homologous ()) and heterologous(&) coverage curves for 16S rDNA are presented. Solid linesindicate DC or (CX j CXY)2 for the original samples at eachvalue of evolutionary distance (D). Broken lines indicate the 950thvalue (or P = 0.05) of DC for the randomized samples. Theapplied LIBSHUFF software was created by Singleton et al. [52].


frequency of alphaproteobacterial sequences was muchhigher in libraries from the sediments with lowerpetroleum concentration (Creta1, 8.2%, and Therm30,8.9%). The lower abundance of alphaproteobacterialclones in these libraries could be associated with thepresence of various hydrocarbon contaminants or withdifferences in productivity of the water column. This isin agreement with Horner-Devine et al. [26] who showedthat Alphaproteobacteria richness is linked to productivitylevels in aquatic ecosystems.

In addition to the observed differences in microbialcommunity composition between sampling locations, allclone libraries were found to be highly diverse. Torsvikand co-workers [58] used DNA–DNA reassociationkinetics to assess the total community genome complex-ity in various environments and found that bacterialcommunities in pristine marine sediments can have atotal community genome complexity exceeding valuesobtained for complex microbial soil communities(õ11,400 compared to õ8800 genome equivalents). Inaddition to being highly diverse, sediment bacterialcommunities consist largely of uncultivated representa-tives [11]. Hence, culture-independent tools play apivotal role in the analysis of diversity and distributionof these organisms that often regulate major biogeo-chemical cycles as well as the degradation and transfor-mation of pollutants. Similar to other methods, based onthe use of PCR to enrich multiple alleles of specific genesfrom complex mixtures of genomes, 16S rDNA clonelibraries suffer from methodological constraints that mayskew the distribution of phylotypes in the library relativeto the community it was derived from [63]. Such biasmay result from stochastic events during the first cyclesof the PCR [64], preferential amplification of certainsequences [38], and a leveling effect in later stages of thePCR leading to a 1:1 bias in product ratios [55]. Hence,we cannot assume that the quantitative distribution ofphylotypes in the library is a direct quantitative repre-sentation of the community it was derived from even ifcare is taken to minimize the bias (e.g., replicate PCRreactions, few amplification cycles). This is also evident ifwe consider that the number of rDNA operons in singlecells may vary from 1 to 15 between different taxa [2, 3].

PCR-induced formation of chimeric molecules, hete-roduplexes, and mutations resulting from Taq error mayalso bias the library and introduce Bartificial^ diversity.This interference can be largely avoided by loweringthe number of amplification cycles and carrying outBreconditioning PCR^ [57]. Finally, rDNA heterogeneitywithin genomes can be as high as 1% for bacteria [2, 3].This variability between operons may obscure discrimi-nation of phylotypes that are more than 99% similarsince they may represent a single bacterial population.

In the present study we used RPLP screening with twofour-cutting restriction enzymes (HhaI and HaeIII) toassess the richness in the respective library. Such ascreening, with one to four restriction enzymes, is normallyused to screen large clone libraries before sequencing ofselected clones for phylogenetic analysis and it has beenshown that parallel use of four such restriction enzymescan discriminate among sequences that are more than 98%similar [39]. Hence, our richness estimates ranging from440 to 1306 unique phylotypes are minimum estimatesthat should be largely insensitive to the Bartificial^ micro-diversity potentially introduced in the PCR. The use ofvariable definitions of a phylotype or OTU in differentstudies of sediment clone libraries may also be part of thereason behind the variability in estimated species richnessand coverage among different sediment 16S rDNA libraries(Table 2). A previous attempt to rarefy bacterial 16S rDNAclone libraries from marine sediments by direct compari-son of sequence homology (98% similarity) has clearlyshown that sediment communities are highly diverse; forexample, libraries do not rarefy as more clones are screenedand almost every sequenced clone is unique in the re-spective library [10]. Richness estimates based on capture–recapture statistics indicate a combined phylotype richnessof 4.350 unique bacterial phylotypes from a single Antarc-tic sediment core and a coverage that ranges from 23 to36% (three sediment horizons). In contrast, Ravenschlagand co-workers [46] used RFLP screening with a singlefour-cutter restriction enzyme to delineate a phylotype andarrived at a coverage of 970% when 353 clones werescreened (Table 2). The high sequence richness in all foursediment libraries in the present study indicates thatthousands of clones need to be screened to rarefy any

Table 4. Comparisons of the 16S rDNA clone libraries Therm01, Therm30, Creta1, and S.Ionian

Stations

Therm30 Creta1 S.Ionian

LIBSHUFF[�C (P=0.001)] T-RFLP (S)



Therm01 4.510 0.612 4.510 0.565 4.510 0.409Therm30 12.811 0.503 12.811 0.481Creta1 11.574 0.384

Analysis was carried out using the LIBSHUFF software. Values of DC indicated that differences between homologous and heterologous coverage curves weresignificant for all crosswise comparisons at P G 0.05. Similarities (S) using mixed-community T-RFLP data from Polymenakou et al. [42] are also presentedfor comparison. The similarity S = 2J/(NA+ NB), where J is the number of common bands or terminal restriction fragments in samples A and B, and NA andNB are the total number of bands in sample A and B, respectively [42].


individual sediment clone library. This observation con-firms that existing 16S rDNA clone libraries have onlyscratched the surface of the enormous bacterial diversitycontained in sediments. Finally, not a single 16S rDNAsequence in our four sediment libraries was found in morethan one of the four libraries.

The significance of differences between clone librarieswas examined with LIBSHUFF statistics. Paired reciprocalcomparisons indicated that each of the libraries differssignificantly from the others, and this is in agreement withdirect interpretation of phylogenetic trees (Fig. 2). TheLIBSHUFF program is a good test of library overlap;however, Stach et al. [53] reported that it is not sensitiveto the phylogenetic grouping of taxa and, therefore,LIBSHUFF statistics may indicate identical communities(P = 1.00) when libraries are composed of closely relatedtaxa [53]). Despite methodological constraints regardingsample size in our study, the LIBSHUFF statistics scoutsvariation in differences between different pairs of clonelibraries. Hence, libraries from geographically remoteregions were found to be much more different than thoseoriginating from the same general area (e.g., ThermaikosGulf). This observation was also corroborated by anindependent analysis of microbial community composi-tion using T-RFLP analysis [42]. Bacterial communitycomposition was found to differ greatly among thesampling locations probably because of either substantialregional variability or extremely high overall richness andevenness in Mediterranean sediment bacterial communi-ties, or a combination of the two.

To conclude, we assessed the bacteria communitycomposition in different sediments of the Eastern Medi-terranean Sea. All four communities were highly diverse andthe estimated total sequence richness was found to becomparable to estimates for microorganisms inhabitingterrestrial ecosystems [59, 66]. The quantitative distributionof different taxonomic groups (phylum/class) was overallvery similar among the communities but notable differ-ences were observed between deep and shallow sediments.All four libraries were found to be significantly different,containing sequences affiliated with various environmentalclones from hydrocarbon seeps, sulfate reducers, sulfuroxidizers, and ammonia oxidizers. A large proportion ofthe obtained phylotypes represent bacteria that are onlydistantly related to any sequences in GenBank, implyingthat many more prokaryotic lineages await discovery ashigh-throughput tools for analyzing large clone libraries arebeing put to use in studies of the complex microbialcommunities that inhabit marine sediments.

Acknowledgments

The officers and the crew of the R/Vs Aegaeo and Philiaare acknowledged for their assistance during the sam-

pling. This work was supported by the Swedish ResearchCouncil (grant 2002-4580 to S.B.) and the SwedishResearch Council for Environment, Agricultural Sciencesand Special Planning (grant 2002-0291 to S.B.). Samplingcruises were supported by the commission of theEuropean Communities (through the Energy and Envi-ronment projects INTERPOL and ADIOS) and the GreekMinistry of Development (General Secretariat of Re-search and Technology).

References

1. Acinas, SG, Anton, J, Rodrıguez-Valera, F (1999) Diversity of free-living and attached bacteria in offshore western Mediterraneanwaters as depicted by analysis of genes encoding 16S rDNA. ApplEnviron Microbiol 65: 514–522

2. Acinas, SG, Klepac-Ceraj, V, Hunt, D, Pharino, C, Ceraj, I, Distel,D, Polz, MF (2004) Fine-scale phylogenetic architecture of a com-plex bacterial community. Nature 430: 551–554

3. Acinas, SG, Marcelino, LA, Klepac-Ceraj, V, Polz, MF (2004)Divergence and redundancy of 16S rDNA sequences in genomeswith multiple rrn operons. J Bacterial 186: 2629–2635

4. Acinas, SG, Rodrıgez-Valera, F, Pedros-Alio, C (1997) Spatial andtemporal variation in marine bacterioplankton diversity as shownby RFLP fingerprinting of PCR amplified 16S rDNA. FEMSMicrobiol Ecol 24: 27–40

5. Altschul, SF, Madden, TL, Schaffer, AA, Zhang, J, Zhang, Z, Miller,W, Lipman, D (1997) Gapped BLAST and PSI_blast: a new gen-eration of protein database search programs. Nucleic Acids Res 25:3389–3402

6. Amann, RI, Ludwig, W, Schleifer, KH (1995) Phylogeneticidentification and in situ detection of individual microbial cellswithout cultivation. Microbiol Rev 59: 143–169

7. Bano, N, Hollibaugh, JT (2002) Phylogenetic composition ofbacterioplankton assemblages from the Arctic Ocean. Appl En-viron Microbiol 68: 505–518

8. Barnett, PRP, Watson, J, Connelly, D (1984) A multiple corer fortaking virtually undisturbed sediment samples from shelf bathyaland abyssal sediments. Oceanol Acta 7: 399–408

9. Boivin-Jahns, V, Ruimy, R, Bianchi, A, Daumas, S, Christen, R(1996) Bacterial diversity in a deep-subsurface clay environment.Appl Environ Microbiol 62: 3405–3412

10. Bowman, JP, McCuaig, RD (2003) Biodiversity, communitystructural shifts, and biogeography of prokaryotes within Antarcticcontinental shelf sediment. Appl Environ Microbiol 69: 2463–2483

11. Bowman, JP, Rea, SM, McCammon, SA, McMeekin, TA (2000)Diversity and community structure within anoxic sediment frommarine salinity meromistic lakes and a coastal meromistic marinebasin, Versfold Hills, Eastern Antarctica. Environ Microbiol 2:227–237

12. Brinkhoff, T, Sievert, SM, Kuever, J, Muyzer, G (1999) Distribu-tion and diversity of sulfur-oxidizing Thiomicrospira spp. at ashallow-water hydrothermal vent in the Aegean Sea (Milos,Greece). Appl Environ Microbiol 65: 3843–3849

13. Chao, A (1984) Nonparametric estimation of the number ofclasses in a population. Stand J Stat 11: 265–270

14. Coolen, MJ, Cypionka, H, Sass, AM, Sass, H, Overmann, J (2002)Ongoing modification of Mediterranean Pleistocene sapropelsmediated by prokaryotes. Science 28: 2407–2410

15. Danovaro, R, Dell’Anno, A, Fabiano, M, Pusceddu, A, Tselepides,A (2001) Deep-sea ecosystem response to climate changes: theeastern Mediterranean case study. Trends Ecol Evol 16: 505–510

16. Danovaro, R, Dell’Anno, A, Pusceddu, A, Fabiano, M (1999)


Nucleic acid concentrations (DNA, RNA) in the continental anddeep-sea sediments of the eastern Mediterranean: relationship withseasonally varying organic inputs and bacterial dynamics. Deep-Sea Res Part 146: 1077–1094

17. Danovaro, R, Dinet, A, Duineveld, G, Tselepides, A (1999) Benthicresponse to particulate fluxes in different trophic environments: acomparison between the Gulf of Lions-Catalan Sea (Western-Mediterranean) and the Cretan Sea (Eastern-Mediterranean). ProgOceanogr 44: 287–312

18. Danovaro, R, Marrale, D, Dell’Anno, A, Della Croce, N,Tselepides, A, Fabiano, M (2000) Bacterial response to seasonalchanges in labile organic matter composition on the continentalshelf and bathyal sediments of the Cretan Sea. Prog Oceanogr 46:345–366

19. Dugdale, RC, Wilkerson, FR (1988) Nutrient sources and primaryproduction in the Eastern Mediterranean. Oceanol Acta 9: 178–184

20. Felsenstein, J (1985) Confidence limits on phylogenies: anapproach using the bootstrap. Evolution 39: 783–791

21. Felsenstein, J (1981) Evolutionary trees from DNA sequences: amaximum likelihood approach. J Mol Evol 17: 368–376

22. Gogou, A, Apostolaki, M, Stephanou, EG (1998) Determinationof organic molecular markers in marine aerosols and sediments: one-step flash chromatography compound class fractionation and capillarygas chromatographic analysis. J Chromatogr A 799: 215–231

23. Gogou, A, Bouloubassi, I, Stephanou, EG (2000) Marine organicgeochemistry of the Eastern Mediterranean: 1. Aliphatic andpolyaromatic hydrocarbons in Cretan Sea surficial sediments.Mar Chem 68: 265–282

24. Gray, JP, Herwig, RP (1996) Phylogenetic analysis of the bacterialcommunities in marine sediments. Appl Environ Microbiol 62:4049–4059

25. Hedges, JI, Stern, JH (1984) Carbon and nitrogen determinationsof carbonate-containing solids. Limnol Oceanogr 29: 657–663

26. Horner-Devine, MC, Leibold, MA, Smith, VH, Bohannan, BJM(2003) Bacterial diversity patterns along a gradient of primaryproductivity. Ecol Lett 6: 613–622

27. Hugenholtz, P, Goebel, BM, Pace, NR (1998) Impact of culture-independent studies on the emerging phylogenetic view of bac-terial diversity. J Bacteriol 180: 4765–4774

28. Ignatiades, L (1969) Annual cycles, species diversity and successionof phytoplankton in lower Saronikos Bat, Aegean Sea. Mar Biol 3:196–200

29. Karageorgis, AP, Anagnostou, CL (2001) Particulate matterspatial–temporal distribution and associated surface sedimentproperties: Thermaikos Gulf and Sporades Basin, NW AegeanSea. Cont Shelf Res 21: 2141–2153

30. Lane, DJ (1991) 16S/23S rDNA sequencing. In: Stackebrandt,Goodfellow, M, (Eds.) Nucleic Acid Techniques in BacterialSystematics. Wiley( Chichester, pp 115–175

31. Li, L, Kato, C, Horikoshi, K (1999) Bacterial diversity in deep-seasediments from different depths. Biodivers Conserv 8: 659–677

32. Li, L, Kato, C, Horikoshi, K (1999) Microbial diversity insediments collected from the deepest cold-seep area, the JapanTrench. Mar Biotechnol 1: 391–400

33. Lorenzen, C, Jeffrey, J (1980) Determination of chlorophyll in seawater. UNESCO Tech Pap Mar Sci 35: 1–20

34. Ludwig, W, Bauer, SH, Bauer, M, Held, I, Kirchhof, G, Schulze, R,Huber, I, Spring, S, Hartmann, A, Schleifer, KH (1997) Detectionand in situ identification of representatives of a widely distributednew bacterial phylum. FEMS Microbiol Lett 153: 181–190

35. Madrid, VM, Aller, JY, Aller, RC, Chistoserdov, AY (2001) Highprokaryote diversity and analysis of community structure in mo-bile mud deposits off French Guiana: identification of two newbacterial candidate divisions. FEMS Microbiol Ecol 37: 197–209

36. Maidak, BL, Cole, JR, Lilbum, TG, Parker, CT, Saxman Jr, P-R,Farris, RJ, Garrity, GM, Olsen, GJ, Schmidt, TM, Tiedje, JM (2001)

The RDP-II (Ribosomal Database Project). Nucleic Acids Res 29:173–174

37. Malanotte-Rizzoli, P, Manca, BB, Ribera d’Alcala, M, Theocharis,A, Bergamasco, A, Bregant, D, Budillon, G, Civitarese, G,Georgopoulos, D, Michelato, A, Sansone, E, Scarazzato, P,Souvermezoglou, E (1997) A synthesis of the Ionian Sea hydro-graphy, circulation and water mass pathways during POEM-Phase-I. Prog Oceanogr 39: 153–204

38. Morasch, B, Richnow, HH, Schink, B, Meckenstock, RU (2001)Stable hydrogen and carbon isotope fractionation during microbialtoluene degradation: mechanistic and environmental aspects. ApplEnviron Microbiol 67: 4842–4849

39. Moyer, CL (2001) Molecular phylogeny: applications and impli-cations for marine microbiology. In: Paul( JH (Ed.) Methods inMicrobiology. Marine Microbiology( Academic Press( San Diego,pp 375–394

40. Phillips, CJ, Smith, Z, Embley, TM, Prosser, JI (1999) Phylogeneticdifferences between particle-associated and planktonic ammonia-oxidizing bacteria of the b subdivision of the class Proteobacteria inthe Northwestern Mediterranean Sea. Appl Environ Microbiol 65:779–786

41. Pichinoty, F, Asselineau, J, Mandel, M (1984) Caracterisationbiochimique de Bacillus benzoevornns sp. nov., une nouvelle especefilamenteuse, engainee et mesophile, degradant divers acidesaromatiques et phenols. Ann Microbiol 135B: 209–217

42. Polymenakou, PN, Bertilsson, S, Tselepides, A, Stephanou, EG(2005) Links between geographic location, environmental factorsand microbial community composition in sediments of the EasternMediterranean Sea. Microb Ecol 49: 367–378

43. Polz, MF, Cavanaugh, CM (1998) Bias in template-to-product ratiosin multitemplate PCR. Appl Environ Microbiol 64: 3724–3730

44. Psarra, S, Tselepides, A, Ignatiades, L (2000) Primary productivityin the oligotrophic Cretan Sea (NE Mediterranean): seasonal andinterannual variability. Prog Oceanogr 46: 187–204

45. Pukall, R, Pauker, O, Buntefuß, D, Ulrichs, G, Lebaron, P,Bernard, L, Guindulain, T, Vives-Rego, J, Stackebrandt, E (1999)High sequence diversity of Altevomonas mancleodii-related clonedand cellular 16S rDNAs from a Mediterranean seawater mesocosmexperiment. FEMS Microbiol Ecol 28: 335–344

46. Ravenschlag, K, Sahm, K, Pemthaler, J, Amann, R (1999) Highbacterial diversity in permanently cold marine sediments. ApplEnviron Microbiol 65: 3982–3989

47. Sambrook, J, Russel, DW (2001) Molecular Cloning: A LaboratoryManual, 3rd ed. Cold Spring Harbor Laboratory( Cold SpringHarbor, NY

48. Sass, AM, Sass, H, Coolen, MJL, Cypionka, H, Overmann, J (2001)Microbial communities in the chemocline of a hypersaline deep-sea basin (Urania basin, Mediterranean Sea). Appl Environ Micro-biol 67: 5392–5402

49. Schafer, H, Servais, P, Muyzer, G (2000) Successional changes inthe genetic diversity of a marine bacterial assemblage duringconfinement. Arch Microbiol 173: 138–145

50. Sievert, SM, Brinkhoff, T, Muyzer, G, Ziebis, W, Kuever, J (1999)Spatial heterogeneity of bacterial populations along an environ-mental gradient at a shallow submarine hydrothermal vent nearMilos Island (Greece). Appl Environ Microbiol 65: 3834–3842

51. Sievert, SM, Kuever, J, Muyzer, G (2000) Identification of 16Sribosomal DNA-defined bacterial populations at a shallow sub-marine hydrothermal vent near Milos Island (Greece). Appl En-viron Microbiol 66: 3102–3109

52. Singleton, DR, Furlong, MA, Rathbun, SL, Whitman, WB (2001)Quantitative comparisons of 16S rDNA gene sequences librar-ies from environmental samples. Appl Environ Microbiol 67:4374–4376

53. Stach, JEM, Maldonado, LA, Masson, DG, Ward, AC, Goodfellow,M, Bull, AT (2003) Statistical approaches for estimating actino-


bacterial diversity in marine sediments. Appl Environ Microbiol69: 6189–6200

54. Strunk, O, Ludwig, W (1996) ARB: A Software Environment forSequence Data, 2.1.1. Department of Microbiology, TechnicalUniversity of Munich, Munich

55. Suzuki, M, Giovannoni, SJ (1996) Bias caused by templateannealing in the amplification mixtures of 16S rDNA genes byPCR. Appl Environ Microbiol 62: 625–630

56. Teske, A, Hinrichs, K-U, Edgcomb, V, de Vera Gomez, A, Kysela,D, Sylva, SP, Sogin, ML, Jannasch, HW (2002) Microbial diversityof hydrothermal sediments in the Guaymas basin: evidence foranaerobic methanotrophic communities. Appl Environ Microbiol68: 1994–2007

57. Thompson, JR, Marcelino, LA, Polz, MF (2002) Heteroduplexes inmixed-template amplifications: formation, consequences andelimination by Breconditioning PCR^. Nucleic Acids Res 30:2083–2088

58. Torsvik, V, Lise Daae, F, Sandaa, R-A, Øvreas, L (1998) Noveltechniques for analyzing microbial diversity in natural and per-turbed environments. J Biotechnol 64: 53–62

59. Torsvik, V, Ølvreas, L, Thingstand, TF (2002) Prokaryoticdiversity—magnitude, dynamics and controlling factors. Science296: 1064–1066

60. Tselepides, A, Zervakis, V, Polychronaki, T, Danovaro, R, Chronis,G (2000) Distribution of nutrients and particulate organic matterin relation to the prevailing hydrographic features of the CretanSea (NE Mediterranean). Prog Oceanogr 46: 113–142

61. Urakawa, H, Kita-Tsukamoto, K, Ohwada, K (1999) Microbialdiversity in marine sediments from Sagami bay and Tokyo bay,Japan, as determined by 16S rDNA gene analysis. Microbiology145: 3305–3315

62. Vergin, KL, Urbach, E, Stein, JL, DeLong, EF, Lanoil, BD,Giovannoni, SJ (1998) Screening of a fosmid library of marineenvironmental genomic DNA fragments reveals four clones relatedto members of the order Planctomycetales. Appl Environ Microbiol64: 3075–3078

63. von Wintzingerode, F, Gobel, UB, Stackebrandt, E (1997)Determination of microbial diversity in environmental samples:pitfalls of PCR-based rDNA analysis. FEMS Microbiol Ecol 21:213–229

64. Wagner, A, Blackstone, N, Cartwright, P, Dick, M, Misof, B, Snow,P, Wagner, GP, Bartels, J, Murtha, M, Pendleton, J (1994) Surveysof gene families using polymerase chain reaction: PCR selectionand PCR drift. Syst Biol 43: 250–261

65. Zhang, H, Sekiguchi, Y, Hanada, S, Hugenholtz, P, Kim, H,Kamagata, Y, Nakamura, K (2003) Gemmatimonas aurantica gen.nov., sp. nov., a Gram-negative, aerobic, polyphosphate-accumu-lating micro-organism, the first cultured representative of the newbacterial phylum Gemmatimonadetes phyl. nov. Int J Syst EvolMicrobiol 53: 1155–1163

66. Zhou, J, Xia, B, Treves, DS, Wu, L-Y, Marsh, TL, O’Neill, RV,Palumbo, AV, Tiedje, JM (2002) Spatial and resource factorsinfluencing high microbial diversity in soil. Appl Environ Micro-biol 68: 326–334


Date post:	19-Nov-2023
Category:	Documents
Upload:	crete
View:	0 times
Download:	0 times

Bacterial Community Composition in Different Sediments from the Eastern Mediterranean Sea: a...

Documents