ARTICLE IN PRESS

Microbiological Research 164 (2009) 624—633

0944-5013/$ - sdoi:10.1016/j.

�CorrespondE-mail addr

jiao@xmu.edu.

www.elsevier.de/micres

Vertical distribution and phylogenetic compositionof bacteria in the Eastern Tropical NorthPacific Ocean

Ying Maa,b, Yonghui Zenga, Nianzhi Jiaoa,�, Yang Shia, Ning Honga

aState Key Laboratory of Marine Environmental Science, Xiamen University, Xiamen 361005, ChinabFisheries College, Jimei University, Xiamen 361021, China

Received 26 October 2007; received in revised form 20 December 2007; accepted 6 January 2008

KEYWORDSBacteria;Vertical distribution;Phylogeneticcomposition;16S rRNA gene;Pacific Ocean

ee front matter & 2008micres.2008.01.001

ing author. Tel./Fax: +8esses: maying@jmu.educn (N. Jiao).

SummaryThe vertical community structure of bacteria along a depth profile in the EasternTropical North Pacific Ocean (131N, 1041W) was studied by flow cytometrymeasurement and 16S rRNA gene clone libraries analysis. Picoeukaryotes andSynechococcus peaked at 30m and decreased sharply below 50m, while Prochlor-ococcus peaked at both 30 and 100m layers and disappeared below 200m.Heterotrophic bacteria peaked above shallow thermocline and decreased alongthe depth profile. Sequences of total 322 clones from four clone libraries (10, 100,1000, and 3000m) clustered into nine major lineages. g-Proteobacteria dominatedall the depths and occupied almost the whole bacterial community at the 3000m.a-Proteobacteria was abundant throughout the water column except near the seabottom, and d-Proteobacteria peaked at the 1000m depth. Cyanobacteria wereprimarily limited to the photic zone, and the genetic diversity of Prochlorococcusshowed a good correlation with niche adaptation. The appearance of theCytophaga–Flexibacter–Bacteroides (CFB) group did not show a clear relationshipwith depth. Actinobacteria were found both in the photic zone and in deep water.Planctomyetes, Acidobacteria, and Verrucomicrobia were present as minor groupsand more dominant in the deeper layers of water.& 2008 Elsevier GmbH. All rights reserved.

Elsevier GmbH. All rights rese

6 592 2187869..cn (Y. Ma),

Introduction

Bacteria are thought to be important compo-nents of aquatic ecosystems and play a critical rolein the global carbon cycle (Cole et al. 1988).

rved.

ARTICLE IN PRESS

Vertical distribution and phylogenetic composition of bacteria in the Eastern Tropical North Pacific Ocean 625

Although they are all unicellular and similar in cellsize, bacteria are composed of different taxonomicgroups with potentially different phenotypic prop-erties, physiological activities, and ecologicalfunctions. The connection between overall bacter-ial community structure and environmental vari-ables is critical for understanding the function ofvarious bacteria within marine carbon and nutrientcycles. In aquatic systems, one of the most relevantgradients affecting the community structure isdepth in water columns. Molecular methods target-ing 16S rRNA genes have revealed some details ofthe bacterial phylogenetic diversity from differentdepths in the northeast Pacific, the subtropical andtemperate Atlantic Oceans (Delong et al. 2006;Gallagher et al. 2004), and in the Antarctic PolarFront (Lopez-Garcıa et al. 2001). These resultsindicated that the deep sea contains numerousnovel and widespread major prokaryotic lineages.Using rRNA hybridization and gene cloning andsequencing, Garcıa-Martınez et al. (2002) foundthat Alteromonas macleodii-like microorganismsare present in high proportions in deep waters,suggesting that some specific bacteria might playan important ecological role in deep waters. Theprogress, however, is far from complete. A recentlyreported study has shown that bacterial commu-nities of deep water masses of the North Atlanticare far more complex than previously reported forany microbial environment (Sogin et al. 2006).

Compared with the oligotrophic North Atlanticand the subtropical gyre in the western Pacific, theEastern Tropical North Pacific (ETNP) exhibits ahigher surface nutrient level, chlorophyll concen-trations, and higher biological productivity (Schlit-zer 2004). The eutrophic environment mightsupport a bacteria community different from theoligotrophic oceans. In the present study, weexamined four seawater samples obtained fromdifferent depths in the ETNP. 16S rRNA gene clonelibraries were constructed and a total of 332sequences were obtained and subject to a detailedanalysis, with the aim to better characterize thedistribution pattern of bacterial community alongthe depth profile.

Materials and methods

Sample collection

Water samples for flow cytometry analysis werecollected at 1, 10, 30, 50, 75, 100, 125, 150, 175,200, 500, 1000, 1500, 3000, 3100, and 3150mdepths using Niskin bottles (Oceanic) in the ETNP

(131N, 1041W) in November 2003. Samples werefixed with glutaraldehyde (final concentration: 1%),quick-frozen in liquid nitrogen, and stored in afreezer at �20 1C for later analysis. Five literseawater samples collected from 10, 100, 1000,and 3000m (about 150m above the sea bottom)depths for clone libraries construction were filteredonto 47-mm diameter 0.2 mm-pore-size polysulfonefilters (PALL, Ann Arbor, MI, USA) at a pressureo0.03MP. Filtered samples were immediatelyfrozen and stored at �20 1C until DNA extraction.

Data for temperature, salinity, and chlorophyllwere from the Chinese JGOFS cruise project reportsof the same cruise.

Flow cytometry measurement

Glutaraldehyde-fixed samples were run on a FACSCalibur flow cytometer (Becton-Dickinson),equipped with an external quantitative sampleinjector (Harvard Apparatus PHD 2000). Procedureswere as described by Marine et al. (1997). Thethree autotrophs (Prochlorococcus, Synechococcus,and picoeukaryotes) and heterotrophic bacteriawere distinguished and enumerated according toJiao et al. (2002).

DNA extraction, PCR amplification, and clonelibrary construction

Total DNA was extracted and purified accordingto the previously described method (Gallagheret al. 2004) and was used for PCR amplification ofthe 16S rRNA genes. The universal bacterialprimers, 27F (50-AGAGTTTGATCATGGCTCAG-30)and 1492R (50-GGTACCTTGTTACGAC TT-30), wereused for PCR amplification, and the amplificationprocedures were the same as Bano and Hollibaugh(2002). The amplified products were gel-purifiedand ligated into the pMD18-T vector (TaKaRa,Dalian, China) and then transformed into compe-tent cells of Escherichia coli DH5a. The ampicillin-resistant clones were randomly picked andscreened for inserts by performing colony PCR withM13 primers (Invitrogen, Shanghai, China) for thevector.

DNA sequencing, statistical analysis of theclone libraries, and phylogenetic treeconstruction

A total of 337 clones with correct inserts(approximately 1500 bp) were directly sequenced.The sequencing was carried out on an ABI model377 automated DNA sequence analyzer (Applied

ARTICLE IN PRESS

Temperature (°C)

Temperature / Salinity / Chla / Cell abundance

–1

–10

–30

–50

–75

–100

–125

40200 1008060

–150

–175

–200

–500

–1000

–1500

–3000

–3100

–3150

Dep

th (

m)

Salinity (psu)Chla (μ g/L 0.01)

Euk (cell/ml 1000)Pro (cell/ml 1000)Syn (cell/ml 1000)Bac (cell/ml 1000)

Figure 1. Depth profiles of temperature, salinity, chlor-ophyll a, and picoplankton cell abundance in the Eastern

Y. Ma et al.626

Biosystems, Perkin-Elmer) using the primer 27F. Allthe nucleotide sequences were checked for puta-tive chimeras by the CHIMERA_CHECK (Maidaket al. 2001), and reliable sequences were comparedto known 16S rDNA sequences in the database byusing the BLASTN search (http://www.ncbi.nlm.nih.gov/BLAST/). Sequences were then assignedto major groups by using BLAST similarities andRDP Classifier (http://rdp.cme.msu.edu/classifier/classifier.jsp). Sequences were also grouped as opera-tional taxonomic units (OTUs) by 97% or greatersequence similarity in the DOTUR program (Schlossand Handelsman 2005). Diversity indices (Coverage,Dominance, Evenness, Shannon, and Simpson) werecalculated using the statistical program PAlaeonto-logical STatistics (PAST, ver. 1.34, http://folk.uio.no/ohammer/past). Representative sequences fromeach OTU were aligned using the software ClustalX(Thompson et al. 1997), and phylogenetic trees wereconstructed by the software MEGA3 using theneighbor-joining algorithm (Kumar et al. 2004).

Nucleotide sequence accession numbers

Clone sequences have been deposited in GenBankunder the accession numbers AY664161 to AY664249,AY726783 to AY726875, AY726876 to AY726979, andAY726980 to AY727030.

Tropical North Pacific Ocean (131N, 1041W). Euk: pico-eukaryotes; Pro: Prochlorococcus; Syn: Synechococcus;Bact: heterotrophic bacteria.

Results

Depth profile of hydrographic and biologicalvariables

The depth profiles of picoplankton cell abun-dance, temperature, salinity, and chlorophyll a ofthe water column are shown in Figure 1. Salinitywas stable throughout the water column, while athermocline was between 30 and 75m, separatingthe surface mixed layer at a temperature of 27 1Cfrom the water below 75m at a temperature of16 1C. In addition, a much deeper thermocline waspresent at depths between 200 and 1000m,marking the change from cold (411 1C) to verycold water (o5 1C, Figure 1). The chlorophyll aconcentration was high at depths above 100m(ranging from 0.14 to 0.28 mg/L), peaked at 30m(0.28 mg/L) and 100m (0.26 mg/L) depths, and thendeclined dramatically with depth to nearly underdetection limits (p0.003 mg/L) at depths below175m. Correspondingly, the picoplankton cellabundance showed a good depth-related as wellas temperature-related distribution pattern. Theautotrophic picoplankton (picoeukaryotes, Pro-

chlorococcus, and Synechococcus) peaked at 30m,decreased sharply below the first thermocline(except for the Prochlorococcus, it had anotherpeak at about 100m) and disappeared below 200m.The heterotrophic bacteria also peaked above thefirst thermocline and decreased sharply below 50mand then gradually with depth. They had anotherdramatic decline below the second thermocline butcould be detected throughout the water column.

Statistical analysis of 16S rDNA clonelibraries

A total of 337 clones with correct inserts weresequenced and the potential chimeras were re-moved. There were 322 reliable sequences (85 from10m, 88 from 100m, 99 from 1000m, and 50 from3000m) used for further analysis. These clonedsequences were designated as JL-ETNP-H (or I, R, S,Y, Z, E, and F) n, in which H and I represent clonesthat are derived from 10m depth, R and S from100m, Y and Z from 1000m, E and F from 3000m

ARTICLE IN PRESS

Vertical distribution and phylogenetic composition of bacteria in the Eastern Tropical North Pacific Ocean 627

depth, and n represents the number of the clone.Sequences were grouped into 44, 42, 59, and 23OTUs in each clone library, and the diversityindices, the Shannon and Simpson are both highestin the 1000m depth (3.74 and 0.96, respectively)and lowest in the 3000m depth (2.51 and 0.83,respectively), indicating that the overall diversitywas highest in 1000m and lowest in 3000m. Thehighest Dominance occurred in the 3000m (0.17),corresponding to its lowest Evenness (0.53) amongthe depth layers, indicating that the bacterialcommunity near the sea bottom was comprised bya few phylotypes. The second highest Dominancewas in the deep chlorophyll maxima of 100m depth(0.08), corresponding to its lower Evenness (0.58)than that of the other two depths (about 0.7 in the10 and 1000m), suggesting that the deep chlor-ophyll maxima layer might grow some specificphylotypes. However, the values of coverage ofthe clone libraries were not high (ranging from53.54% to 68.18%), indicating that the diversitymight have been underestimated.

Depth profile of the composition of 16S rDNAclone libraries at division and subdivisionlevels

The bacterial community structure was analyzedby the 16S rDNA clone library method in detail, andthe species-composite clone libraries are shown inFigure 2. All of our sequences fell into nine majorlineages of the bacterial domains: a (18%), g (58%),and d-Proteobacteria (10%); cyanobacteria (Pro-chlorococcus and Synechococcus, 6%); Cytophaga–

Flexibacter–Bacteroides group (CFB, 2%); andActinobacteria; Planctomycetes, Acidobacteria,and Verrucomicrobia as well as some clonesoriginating from eukaryotic chloroplast or plastids.The percentage of 16S rDNA sequences of each

Relative ab0%

–10

–100

–1000

–3000

EUK

Y-Proteobacteria

Verrucomicrobia

Prochlorococcus

δ-Proteobacteria

Acidobacteria

10% 20% 30% 40%

Dep

th (m

)

Figure 2. Percentage of abundance of the bacterial phylotypphylogenetic analysis. EUK: eukaryotic-like clones; C/F/B gro

group in the total picoplankton rDNA pool indicatedthat the g-Proteobacteria dominated all the depthsand occupied almost the whole picoplankton com-munity (84%) at 3000m, corresponding to thehighest Dominance of the depth layer. a-Proteo-bacteria was abundant (17–26%) throughout thewater column except for the sea bottom (4%).d-Proteobacteria peaked (27%) at the 1000m depthlayer, but occupied a minor component at shallowerwater (1–2%) and bottom of the sea (4%). Theabundance of autotrophic picoplanktons showedsimilar distribution patterns as seen through theFCM enumeration. The 16S rDNA sequences ofeukaryotic chloroplast or plastid were only de-tected in shallow water. Prochlorococcus weremore abundant at the 100m than that of the10m, while Synechococcus showed a contrarydistribution pattern. Synechococcus were moreabundant in the surface mixed layer, decreasedgradually with depth, and showed a deeperdistribution than that of Prochlorococcus.

A few high G+C Gram-positive bacteria (Actino-bacteria) were only detected at the middle twodepths, 100 and 1000m, and were mainly distrib-uted at 100m. The CFB group was found throughoutthe water column except at the 100m depth, andits abundance did not show a clear relationshipwith depth. Verrucomicrobia, Acidobacteria, andPlanctomycetes were also detected as minor groupsand tended to be more abundant at deeper waterlayers.

Phylogenetic analysis of each group

a-ProteobacteriaA total of 57 sequences (18%) grouped with the a

subdivision of the proteobacteria, dominated bySAR11 sequences (31 of 57). The SAR11 cluster,assigned to 22 OTUs, could be further subdivided

undance (%)

C/F/B groupActinobacteria

Planctomycete

Synechococcus α-Proteobacteria

50% 60% 70% 80% 90% 100%

es as determined by 16S rDNA clone library statistics andup: Cytophaga–Flexibacter–Bacteroides group.

ARTICLE IN PRESS

Y. Ma et al.628

into six subclusters (A–F, Figure 3a), and four ofwhich (A–D) showed different depth-related dis-tributions. All the clones in subcluster A were fromthe 10m depth, while all the clones in subclusters

Figures 3a–c. Neighbor-joining trees generated from alignmrepresentative references retrieved from Genbank. Some growith vertical height and horizontal width of the triangles rrespectively, within a given group. Accessions numbers are shin boldface and designated as H (or I, R, S, Y, Z, E, and F) n,depth, R and S from 100m, Y and Z from 1000m, E and F fromclones. Numbers in brackets that follow the accession numcorresponding libraries. Bootstrap values above 50 (100 iteranucleotide substitution percentage. (a) a-Proteobacteria.high-light-adapted and low-light-adapted genotypes of Proch

B–D were found at X100m depth layers and weremost closely related with clones obtained fromdeep waters of various environments. Sequences insubcluster E had no clear relationship to the

ents of 16S rDNA sequences from the ETNP samples andups in the trees were compressed using Mega 3.0 programepresenting OTUs abundance and degree of divergence,own in parentheses. Clones from this study are indicatedin which H and I represent clones that derived from 10m3000m depth, and n represents the number of differentbers indicate the occurrence frequency of the OTUs intions) are shown at each node. Scale bars represent the(b) g-Proteobacteria. (c) Others. HL and LL refer tolorococcus, respectively.

ARTICLE IN PRESS

Figures 3a–c. (Continued)

Vertical distribution and phylogenetic composition of bacteria in the Eastern Tropical North Pacific Ocean 629

sampling depth. They were found throughout thewater column and were most similar to clonesretrieved from different depths (Figure 3a). Two1000m clones (Y26 and Z38) affiliated withSAR11 cluster (93% bootstrap support) formeda separate subcluster F and clustered with clonesfrom different depths of Southern California off-shore.

The second most abundant cluster within thea-Proteobacteria was the SAR83 (a-3) cluster(Allgaier et al. 2003; Beja et al. 2002), they wereall retrieved in the upper layers of the ocean(p100m depth), and were clustered with theRhodovulum, Ruegeria, Roseobacter, and Sulfito-bacter genera and the SAR83 clone (Figure 3a). Twosequences (Y43 and H29) clustered tightly with thetypical aerobic anoxygenic phototrophic bacteriumErythrobacter litoralis, and composed the a-4subclass of proteobacteria (Figure 3a). Only oneclone (S40) was grouped into Brevundimonas genus.Three other clones (Z40, R55, and Z1) containedinserts that were closely related to only environ-mental sequences (Figure 3a).

c-Proteobacteria

The highest diversity of phylotypes was foundwithin the g-Proteobacteria. A total of 186 sequences(58%), classified into 55 OTUs, were clustered into theg-Proteobacteria. These belonged to seven generaand a few environmental sequences (Figure 3b).Clones affiliated with Alteromonas were dominantgroups (56% of all sequences identified as g-Proteo-bacteria) and were mainly composed of two sub-clusters (Alt-1 and 2). Sequences in Alt-1 wereobtained from different depth layers, while se-quences in subclusters Alt-2 could only be found inthe 10m depth. Alt-1 was the most frequentlyoccurring phylotype, it comprised 14 OTUs and 75clones and was closely affiliated with A. macleodiistrains and some attached bacteria. Alt-2 comprised7 OTUs and 29 clones and was most closely related tosome unclassified bacteria, including a biphenyl-degrading bacterium. Three additional Alteromonas-affiliated clones had no close relatives in the GenBankand could not group with any known sequences in thedatabase (Figure 3b).

ARTICLE IN PRESS

Figures 3a–c. (Continued)

Y. Ma et al.630

ARTICLE IN PRESS

Vertical distribution and phylogenetic composition of bacteria in the Eastern Tropical North Pacific Ocean 631

Other g-affiliated clones were grouped intothe Alcanivorax (27 clones), Pseudoalteromonas(10 clones), Halomonas (eight clones), Colwellia(three clones), Vibrio (two clones), and Oceanos-pirium (one clone), respectively (Figure 3b), andmost of these clones (94%) were obtained fromX100m depths. These clones were mainly relatedwith some functionally important bacteria, such asthe hydrocarbon- or petroleum-degrading bacteria,the alga-lysing bacteria, and some extremelyhalotolerant or psychrophilic bacteria (Figure 3b).

The remaining g-affiliated clones (27 clones, 15OTUs) were only related to unidentified sequences,including the OM182 clade and the SAR86 clustersequences. All of the clones were obtained atX100m depths and most of them were closelyrelated with clones derived from deeper-layerwaters, suggesting that a lot of novel bacteria notpreviously recognized exist in deep layers of theETNP Ocean.

d-Proteobacteria, Cyanobacteria, and others

A total of 33 clones (11%) classified into 17 OTUswere affiliated with the d-Proteobacteria, whichcomprised SAR324, SAR406, and Nitrospinasubclusters and two unidentified subclusters(Figure 3c). Subclusters SAR324 and Nitrospinawere predominant (containing 20 and 7 clones,respectively), while the others contained only oneto two clones, respectively (Figure 3c).

A total of 19 sequences (6% of all sequences weobtained), assigned to nine OTUs, were affiliatedwith cyanobacteria, in which eight sequences wereProchlorococcus, and others were Synechococcus.Clone I27 was obtained from 10m depth and wasmost similar to the high-light-adapted ecotype (HL)Prochlorococcus strain, while all the other Pro-chlorococcus clones were from 100m depth andwere most closely related to the low-light-adaptedecotype (LL) strains, which showed a good correla-tion between genetic diversity and ecotypic adap-tation (Figure 3c). Synechococcus-affiliated clonesalso showed a depth-related difference of distribu-tion. Clones derived from the 100m depth com-posed a separate subcluster (Syn-1) and could notbe grouped meaningfully with any known Synecho-coccus, suggesting that these clones representsome novel Syenchococcus not previously recog-nized. Clones derived from the 10m depth weresubdivided into two subclusters (Syn-2 and Syn-3)and were most closely related to strains isolatedfrom the Red Sea and the equatorial Pacific,respectively (Fuller et al. 2003).

Six clones (about 2%) were affiliated with the CFBgroup. These clones were obtained from different

depth layers and were closely related with somefunctionally important bacteria, such as the or-ganic matter-degrading bacterium Leeuwenhoe-kiella blandensis (Pinhassi et al. 2006), theamoebae-resisting bacterium Candidatus Amoebi-natus massiliae (Greub et al. 2004), and someuncultured CFB-related clones from differentdepths of various marine waters. The remaining13 clones were grouped into the Planctomycetales(five clones), the Actinobacteria (two clones), theVerrucomicrobium (two clones), and the Acidobac-teriales group (two clones, Figure 3c).

Discussion

Analysis of bacteria diversity and dominantgroup in the ETNP water column

Statistical analysis of the clone libraries showedthat the diversity of 1000m depth clone library washighest among the libraries. Further analysisrevealed that the highest diversity was mainlydue to high abundance of the d-Proteobacteria,which contributed 14 OTUs to the library, whileonly one to two OTUs to the other three libraries(Figures 2 and 3c). Another reason might lie in thepresence of the minorities of Verrucomicrobia andAcidobacteriales. They were detected only in thislibrary and contributed four OTUs to the diversity(Figures 2 and 3c). The 3000m depth was charac-tered by its lowest diversity and highest Dom-inance, and the Dominance was far higher (0.17)than the other depth libraries (0.04–0.08). Themost abundant phylotype in this depth occurred 19times and occupied 38% of the clone library, andwas closely related to the A. macleodii. Theorganisms within A. macleodii are prevalent indifferent oceanic regions and seem to be morefrequent in deep waters (Garcıa-Martınez et al.2002). This kind of organisms is thought to be astrategist that can grow rapidly when organicnutrients are readily available (Lopez-Lopez et al.2005). These characteristics provide A. macleodiiadvantages when competing with other bacteria inextreme environments, such as the deep sea in ourstudy. However, temperature was thought to be oneof the factors that might preclude A. macleodiisince these organisms were absent in cold latitudes(Garcıa-Martınez et al. 2002). Nevertheless in ourcase, A. macleodii-like clones were found to bepredominant down to the 3000m depth layer wherethe temperature was only 1.8 1C, though the totalbacterial abundance was very low there (Figure 1).

The second highest dominance occurred at the100m depth, and was due to the abundant clones

ARTICLE IN PRESS

Y. Ma et al.632

related to Alcanivorx spp. and also A. macleodii.They occupied 22.7% and 12.5%, respectively, in theclone library. Since Alcanivorax spp. uses oilhydrocarbons as their exclusive source of carbonand energy and is barely detectable in unpollutedenvironments (Schneiker et al. 2006), there prob-ably exists hydrocarbon pollution in this layer. Theabundant presence of A. macleodii confirmed itswidespread distribution in marine habitats.

Variation of different phyla along the depthprofile

Phylogenetic analysis of 16S rDNA clones from theETNP revealed significant diversity among the clonedsequences, with similarity to previously reportedsequences ranging from 86% to 99%. Most of thesequences showed depth-specific distribution alongthe vertical profile. For example, sequences in the a-3 (SAR83) phylotype tended to be limited to theupper ocean, while the a-4 (Erythrobacter)-affiliatedclones were less abundant but could distributedeeper than the a-3 (Figure 3a). Even in theubiquitous SAR11 cluster, there existed differenthighly depth-specific distributions (Figure 3a andField et al. 1997). As for the g-Proteobacteria, theAlteromonas dominated all depths, but the subclus-ter Alt-2 was mainly distributed in the 10m depth.Although the 16S rDNA sequence is much too stable toallow the identification of depth-related ecotypes, aprevious work carried out with housekeeping genesdid show that there exists a deep-sea ecotype ofA. macleodii (Lopez-Lopez et al. 2005). Clones relatedto the Alcanivorax peaked at the 100m depth, butwere absent or just a minor component at otherdepths. Halomonas were relatively more abundant atdeeper oceanic waters than the shallow water, andColwellia, Vibrio, and Oceanospirillum were onlyfound in certain samples, which indicated that thedistributions and abundance of many genera variedwith depth. Proteobacteria in the d-subdivision alsodisplay uneven distribution in aquatic environments.SAR324 is vertically stratified and peaked at certaindepths in the water columns of both the Atlantic andPacific oceans (Wright et al. 1997), which is generallyin agreement with our results (peaking at 1000m inthe ETNP, Figure 2). These suggest that microbialcommunities in the lower surface layers of this regionmay be functionally specialized.

The vertical patterns of Prochlorococcus and itsecotypes suggest causal relationships with environ-mental variables. Light and temperature are twoimportant variables shaping Prochlorococcus dis-tribution. As has been observed recently in theAtlantic Ocean (Johnson et al. 2006), the HL clade

MIT9312 dominating near the surface grew faster ata temperature of 25 1C, but could not grow attemperatures below 15 1C, while the LL cladeMIT9313 peaked at 100m, and was more abundantin waters between the 13 and 23 1C isotherms andwas near the limit of detection at temperatureabove 25 1C. Similarly, our ETNP clone within theMIT9312 clade appeared at 10m where the tem-perature was 28 1C, and disappeared at 100mwhere the temperature was 14 1C, while the cloneswithin MIT9313 clade were derived from the 100mdepth (13.8 1C) and were not found in the 10mdepth (28 1C) (Figures 1 and 3c).

Other important phyla also show uneven verticaldistribution, not only in this case but also in otherenvironments (Nold and Zwart 1998). Actinobacter-ia have been found both in the photic zone and indeep water, and seem to be more frequentlyrecovered in deep water (Delong et al. 2006), butour Actinobacteria-like clones seemed to be moreabundant in the photic zone (Figure 2). Verrucomi-crobia-affiliated sequences had been found inshallow waters (10–200m) of North Pacific Ocean(Delong et al. 2006), while our study extendedits distribution down to the 1000m depth layer(Figure 2). Members of Acidobacteria and Plancto-mycete were mainly recovered from deep waters(Figure 2), which was in agreement with previousstudies (Delong et al. 2006).

The repeated discovery of sequences belongingto different gene clusters with similar distributionsin the water column from different marine habitatssuggests that the vertical variable in the watercolumn is one of the most relevant gradientsaffecting the distribution of microorganisms. Thereason for microbial variations is not fully under-stood, but the stratified microbial communities andthe phylogenetic analysis described in this studywill be good supplements to the current under-standing of biogeographic distribution of bacter-ioplankton and will provide important foundationaldata for future studies on microbial functionality.

Acknowledgments

This work was supported by the NSFC Project nos.40632013, 40576063, 40521003, and the MOST2005AA635240. We thank the China Ocean Associa-tion for Mineral and Resource for their samplingopportunity.

References

Allgaier M, Uphoff H, Felske A, Wagner-Dobler I. Aerobicanoxygenic photosynthesis in Roseobacter clade

ARTICLE IN PRESS

Vertical distribution and phylogenetic composition of bacteria in the Eastern Tropical North Pacific Ocean 633

bacteria from diverse marine habitats. Appl EnvironMicrobiol 2003;69:5051–9.

Bano N, Hollibaugh JT. Phylogenetic composition ofbacterioplankton assemblages from the Arctic Ocean.Appl Environ Microbiol 2002;68:505–18.

Beja O, Suzuki MT, Heidelberg JF, Nelson WC, Preston CM,Hamada T, et al. Unsuspected diversity among marineaerobic anoxygenic phototrophs. Nature 2002;415:630–3.

Cole JJ, Findlay S, Pace ML. Bacterial production in freshand saltwater ecosystems: a cross-system overview.Mar Ecol Prog Ser 1988;43:1–10.

Delong EF, Preston CM, Mincer T, Rich V, Hallam SJ,Frigaard NU, et al. Community genomics amongstratified microbial assemblages in the Ocean’s inter-ior. Science 2006;311:496–503.

Field KG, Gordon D, Wright T, Rappe MS, Urbach E, VerginK, et al. Diversity and depth-specific distribution ofSAR11 cluster rRNA genes from marine planktonicbacteria. Appl Environ Microbiol 1997;63:63–70.

Fuller NJ, Marie D, Partensky F, Vaulot D, Post AF, ScanlanDJ. Clade-specific 16S ribosomal DNA oligonucleotidesreveal the predominance of a single marine Synecho-coccus clade throughout a stratified water column inthe Red Sea. Appl Environ Microbiol 2003;69:2430–43.

Gallagher JM, Carton MW, Eardly DF, Patching JW. Spatio-temporal variability and diversity of water columnprokaryotic communities in the eastern North Atlan-tic. FEMS Microbiol Ecol 2004;47:249–62.

Garcıa-Martınez J, Acinas SG, Massana R, Rodrıguez-Valera F. Prevalence and microdiversity of Alteromo-nas macleodii-like microorganisms in different ocea-nic regions. Environ Microbiol 2002;4:42–50.

Greub G, La Scola B, Raoult D. Amoebae-resistingbacteria isolated from human nasal swabs by amoebalcoculture. Emerging Infect Dis 2004;10:470–7.

Jiao NZ, Yang YH, Koshikawa H, Watanabe M. Influence ofhydrographic conditions on picoplankton distribution inthe East China Sea. Aquat Microb Ecol 2002;30:37–48.

Johnson ZI, Zinser ER, Coe A, McNulty NP, Woodward EMS,Chisholm SW. Niche partitioning among Prochlorococ-cus ecotypes along ocean-scale environmental gradi-ents. Science 2006;311:1737–40.

Kumar S, Tamura K, Nei M. MEGA 3: integrated software formolecular evolutionary genetics analysis and sequencealignment. Briefings Bioinformatics 2004;5:150–63.

Lopez-Garcıa P, Lopez-Lopez A, Moreira D, Rodrıguez-Valera F. Diversity of free-living prokaryotes from a

deep-sea site at the Antarctic Polar Front. FEMSMicrobiol Ecol 2001;33:219–32.

Lopez-Lopez A, Bartual SG, Stal L, Onyshchenko O,Rodrıguez-Valera F. Genetic analysis of housekeepinggenes reveals a deep-sea ecotype of Alteromonasmacleodii in the Mediterranean Sea. Environ Microbiol2005;7:649–59.

Maidak BL, Cole JR, Lilburn TG, Parker CT, Saxman PR,Farris RJ, et al. The RDP-II (Ribosomal DatabaseProject). Nucleic Acids Res 2001;29:173–4.

Marine D, Partensky F, Jacquet S, Vaulot D. Enumerationand cell-cycle analysis of natural populations ofmarine picoplankton by flow cytometry using thenucleic acid stain SYBR Green I. Appl Environ Microbiol1997;63:186–93.

Nold SC, Zwart G. Patterns and governing forces inaquatic microbial communities. Aquat Ecol 1998;32:17–35.

Pinhassi J, Bowman JP, Nedashkovskaya OI, Lekunberri I,Gomez-Consarnau L, Pedros-Alio C. Leeuwenhoekiellablandensis sp nov., a genome-sequenced marinemember of the family Flavobacteriaceae. Int J SystEvol Microbiol 2006;56:1489–93.

Schlitzer R. Export production in the equatorial and northPacific derived from dissolved oxygen, nutrient andcarbon data. J Oceanogr 2004;60:53–62.

Schloss PD, Handelsman J. Introducing DOTUR, a compu-ter program for defining operational taxonomic unitsand estimating species richness. Appl Environ Micro-biol 2005;71:1501–6.

Schneiker S, Martins dos Santos VA, Bartels D, Bekel T,Brecht M, Buhrmester J, et al. Genome sequence ofthe ubiquitous hydrocarbon-degrading marine bacter-ium Alcanivorax borkumensis. Nat Biotechnol 2006;24(8):997–1004.

Sogin ML, Morrison HG, Huber JA, Welch DM, Huse SM,Neal PR, et al. Microbial diversity in the deep sea andthe underexplored ‘‘rare biosphere’’. PNAS 2006;103:12115–20.

Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F,Higgins DG. The ClustalX windows interface: flexiblestrategies for multiple sequence alignment aided byquality analysis tools. Nucleic Acids Res 1997;24:4876–82.

Wright TD, Vergin KL, Boyd PW, Giovannoni SJ. A noveldelta-subdivision proteobacterial lineage from thelower ocean surface layer. Appl Environ Microbiol1997;63:1441–8.