Diverse communities of active Bacteria and Archaea along oxygen gradients in coral reef sediments

12 Pages·2009·0.48 MB·English

CoralReefs(2009)28:15–26
DOI10.1007/s00338-008-0427-y
REPORT
Diverse communities of active Bacteria and Archaea along oxygen
gradients in coral reef sediments
A. Rusch Æ A. K. Hannides Æ E. Gaidos
Received:4June2008/Accepted:6September2008/Publishedonline:24September2008
(cid:1)Springer-Verlag2008
Abstract Microbial communities inhabiting highly per- organic matter degradation and nutrient recycling in coral
meable sediments of Checker Reef in Kaneohe Bay, reef ecosystems.
Hawaii, were characterized in relation to porewater geo-
chemistry (O , NO -, NO -, NH ?, phosphate). The Keywords Coral reef sediment (cid:1) DIN (cid:1)
2 3 2 4
physiologically active part of the population, assessed by Microbial community (cid:1) Oxygen
sequencing cDNA libraries of 16S rRNA amplicons, was
very diverse, with an estimated ribotype richness C1,380
in anoxic sediment. Quantitative analysis of community Introduction
structure by rRNA-targeted fluorescence in situ hybrid-
ization (FISH) indicated that the archaeal population Tropical coral reefs are among the most productive eco-
(9–18%)wasdominatedbymarineCrenarchaeota(5–9%). systemsandhometohighlydiversemarinebiota,including
Planctomycetaleswerethemostabundantgroupintheoxic a plethora of prokaryotic species. Bacterial and archaeal
and interfacial habitat (17–19%) but were a minority inhabitants of living coral have been investigated in the
(\5%) in anoxic reef sediment, where c-Proteobacteria context of symbiotic relationships (Rohwer et al. 2002;
were numerically dominant (18%). Another 9–14% of the Kellogg 2004; Wegley et al. 2004; Klaus et al. 2005) or
microbial benthos belonged to b-Proteobacteria, predomi- coral disease and mortality (Breitbart et al. 2005; Kline
nantly within the order Nitrosomonadales, many cultured et al. 2006; Ritchie 2006). Heterotrophic microbes in the
representativesofwhichareNH ?oxidizers.Theresultsof underlyingcarbonatereefframeworkandofreefsediments
4
thisstudycontributetothephylogeneticcharacterizationof areconsideredimportantintheremineralizationoforganic
benthic microbial communities that are important in matter(OM)(Rasheedet al.2003;Wildet al.2005)andin
the efficient recycling of nutrients that is crucial for sus-
taining high ecosystem productivity in extremely
CommunicatedbytheGeologyEditorDrBernardRiegl. oligotrophic tropical seawater (D’Elia and Wiebe 1990;
Capone et al. 1992; Tribble et al. 1994; Miyajima et al.
A.Rusch(&)(cid:1)E.Gaidos
2001; Rasheed et al. 2002).
DepartmentofGeologyandGeophysics,UniversityofHawaii
Many coral reef frameworks and their unconsolidated
atManoa,1680East–WestRoad,Honolulu,HI96822,USA
e-mail:[email protected] sediments are highly permeable structures, where waves,
currents,andirrigatingfaunacandriveadvectiveexchange
A.K.Hannides
between porewater and overlying seawater (Sansone et al.
DepartmentofOceanography,UniversityofHawaiiatManoa,
1988; Tribble et al. 1992; Alongi et al. 1996; Haberstroh
1000PopeRoad,Honolulu,HI96822,USA
and Sansone 1999). For example, the porewaters of
PresentAddress: Checker Reef, a fully submersed patch reef in Kaneohe
A.K.Hannides
Bay, Hawaii, exhibited a deep oxycline, and their depth-
DepartmentofFisheriesandMarineResearch,Ministry
integrated O content increased with wave height (Falter
ofAgriculture,NaturalResourcesandtheEnvironment, 2
101BethlehemStreet,1416Nicosia,Cyprus and Sansone 2000a, b). Heterogeneity of physically and
123
16 CoralReefs(2009)28:15–26
biologicallyinducedinterstitialwaterflowgeneratesshort- 15–25 cmdepth(HaberstrohandSansone1999;Falterand
termvariability,microzonation,andsteeplocalgradientsin Sansone 2000a, b).
porewatergeochemistry(Sansoneet al.1988;Alongiet al. On April 20, 2006, July 12, 2006 and May 21, 2007,
1996; Huettel et al. 1998; Precht et al. 2004). Efforts to porewater and sediment samples were collected from
characterize the benthic microbial communities are still in Checker Reef at 21(cid:2)26.60N, 157(cid:2)47.60W, using the well-
their infancy; a surface sediment of the Great Barrier Reef point sampling method as previously applied to reef sands
containedanestimated30–140operationaltaxonomicunits (Falter and Sansone 2000b; Sørensen et al. 2007). Briefly,
(OTUs)inBacteria,withmaximumdiversityat3 cmdepth apointedstainlesssteelrod(1 mlong,1 cmdiameter)and
(Hewson and Fuhrman 2006). The subsurface prokaryotic sleeve were driven into the sediment to the desired depth,
communities in geochemically distinct sediments of Moku at which the rod was removed, allowingporewater toseep
o Loe, an inhabited reef island in Kaneohe Bay, Hawaii, into the sleeve. The top of the sleeve was sealed with a
showed a significant horizontal and vertical heterogeneity butyl rubber stopper fitted with a port, through which
indenaturinggradientgelelectrophoresis(DGGE)banding porewater was drawn into a 60 ml syringe. Samples were
patterns and DNA-based clone library structure (Sørensen takenbetween 10and60 cmdepthandfromtheoverlying
et al. 2007). water. At each depth, the first 60 ml of porewater was
ThisstudyreportsthefirstanalysisofRNAincoralreef discarded, before a second portion of 60 ml was retrieved
sediments, representing the currently active fraction of the for immediate measurement of O , and another 60 ml was
2
benthic microbial population. RNA-based clone libraries stored in a polypropylene tube for transport, minimizing
wereconstructedtosurveytheidentitiesofphysiologically the headspace. The porewater volume withdrawn corre-
active Bacteria in Checker Reef sediments and to estimate sponded to that contained in a sediment sphere of radius
their phylogenetic diversity. Fluorescence in situ hybrid- 4.4–5.2 cm. Then the sleeve was driven 2–3 cm deeper
ization (FISH) was applied to quantify the microbial into the reef to retrieve a small sediment sample, which
community structure and its vertical heterogeneity. The was transferred from the bottom of the stake into a poly-
communities in oxic, ‘‘interfacial’’ (suboxic, microoxic or propylene tube. All samples were placed on ice during
temporarily oxic) and anoxic benthic reef habitats are transport to the laboratory.
compared in the context of the effects of O on microbial Within 3 h of sampling, porewater samples were syr-
2
community structure. Furthermore, putatively important inge-filtered (pore size 0.2 lm), split into aliquots for
microbial inhabitants are discussed with respect to their NH ?, NO -, NO - and dissolved inorganic phosphate
4 3 2
potential metabolic functions in the coral reef ecosystem. (DIP) analyses, and stored at -20(cid:2)C. From each sediment
sample, 2.0 g were placed in a 10 ml serum bottle, pre-
servedwith3 mlofformaldehyde/aceticacidsolution(2%
Materials and methods each in filtered seawater), and stored at 4(cid:2)C for later cell
counts. For FISH, duplicate 0.5 cm3 subsamples were put
Study site, sample collection and preservation in 2 ml screw-cap tubes and incubated with 1.5 ml para-
formaldehyde(3%insalinephosphatebuffer,PBS)at4(cid:2)C
Kaneohe Bay,Oahu,Hawaii (21(cid:2)280N,157(cid:2)480W)harbors for4 h. Aftercentrifugation(15,000 9 g,5 min,4(cid:2)C),the
anextensivesystemofpatchreefsat\2 mwaterdepththat fixative was removed, and the sediment was washed three
arecoveredwithpartiallylithifiedcarbonatesandandcoral times with 1.5 ml PBS each, then preserved in 1 ml PBS/
rubble (Tribble et al. 1990, 1992). These sediments are ethanol (50:50) and stored at -20(cid:2)C. The remaining sed-
poorly sorted (size fractions from cobble to clay), with iment volume was stored at -80(cid:2)C for later extraction of
median grain sizes in the 5–10 mm range, and character- nucleic acids and analysis of organic content.
ized by porosities of 0.3–0.5 and permeabilities of
10-8–10-10 m2 (Falter and Sansone 2000a; Hannides
2008). During most of the year, northeasterly trade winds Analyses of porewater solutes and OM content
drive surface waves acrossthe reefs, with wave heights up
to 0.5 m and frequencies predominantly between 0.10 and Immediatelyuponsampling,porewaterwasaddedtoaglass
0.18 s-1 (Tribble et al. 1992; Haberstroh and Sansone bottle until full, and its O concentration was measured by
2
1999; Falter and Sansone 2000a). In the sediment of an Orion 820 probe with a portable oxygen meter. Instru-
CheckerReef,wave-drivenexchangeofporewaterwiththe mentreadingswerecross-calibratedwithWinklertitrations
overlying water occurs within 2.1 d at 1 m depth (Tribble (Winkler 1888) over arange ofO concentrations.
2
et al. 1992). Depending on wave action and sediment Concentrations of NH ? in preserved porewater and
4
permeability, dissolved O penetrates 15–50 cm into the seawater samples were determined spectrophotometrically
2
sand, and concentrations of NO - and NO - peak at by the phenol–hypochlorite method (Solo´rzano 1969).
3 2
123
CoralReefs(2009)28:15–26 17
ConcentrationsofNO -andNO -inpreservedporewater suspension.Theextractwasstoredat4(cid:2)Cforlessthan24 h
3 2
and seawater samples were also measured spectrophoto- until staining and enumeration.
metrically(BendschneiderandRobinson1952;Morrisand Aliquots of cell suspension were stained with 40–60-di-
Riley1963).Whereadvisable(sulfidicsmellorhighNH ? amidino-2-phenylindole (DAPI), concentrated on
4
concentrations), zinc acetate was added to 1 mM to pre- polycarbonate membrane filters (pore size 0.2 lm), and
cipitate sulfide from NO - samples before passing them viewed under an epifluorescence microscope (Olympus
3
over the catalytic column. DIP was analyzed by the BX51). Cell abundances were determined from 12 count-
molybdenum blue complexation method using ascorbic ing grids each on duplicate filters.
acid as the reductant (Murphy and Riley 1962; Koroleff
1976). Whole-cell FISH
Sediment samples collected on July 12, 2006 were
weighedintopre-combustedceramiccrucibles,driedunder Fixed cells were extracted from the sediment as described
ventilationatroomtemperaturefor3 days,andreweighed. for cell counts, but without sonication. Aliquots of the
Thedrysampleswerethencombustedat550(cid:2)Cfor4 hand resulting cell suspension were concentrated on white
weighed again to determine their weight loss on ignition, polycarbonate membrane filters (0.2 lm pore size, Milli-
considered equivalent to total organic matter (TOM) con- pore), allowed to air dry, and stored in a petri dish at
tent (Santisteban et al. 2004). -20(cid:2)C until hybridization. Filter sectors were incubated
withCy3-labeledoligonucleotideprobes(ThermoElectron
Cell abundance GmbH) and, if applicable, with equimolar amounts of
unlabeled competitor oligonucleotide, for *2 h in an
Microbial cells were extracted from the formaldehyde- equilibrated humidity chamber at 46(cid:2)C (Manz et al. 1992;
preserved sediment samples by applying ultrasound in Snaidret al.1997).Informationontheprobes,theirtargets
combination with acetic acid as suggested by Wild et al. and optimal stringency applied is given in Table 1 and
(2006), with slight modifications. Briefly, samples were references therein. Excess reagent was removed with
placed on ice and exposed to five pulses (5 s on, 10 s off) washing buffer at 48(cid:2)C for 15 min. After rinsing with
of ultrasound (22.5 kHz) emitted by the immersed tip of a distilled water, the hybridized filter sectors were counter-
probe (power output 7–8 W, Fisher Scientific Sonic Dis- stained with DAPI, rinsed briefly with distilled water, air-
membrator 100). After settling of the sand grains, the dried in the dark, and mounted on glass slides in Citiflu-
supernatant was collected with a Pasteur pipette. The orTM AF87 (Citifluor Ltd.).
remaining sediment was shaken with 3 ml extractant (2% The slides were examined by epifluorescence micros-
acetic acid in filtered seawater), let stand until sand grains copy under a magnification of 1,3009. On each of 1–3
settledout,andsupernatantcollected,withfiverepetitions. filters, cells showing probe-conferred fluorescence and
Thissonication/washingprocedurewasrepeatedfourmore DAPI-stained cells in the same field of view were enu-
times, producing a combined supernatant of 105 ml cell meratedin12–24randomlychosencountinggrids.Inorder
Table1 Oligonucleotideprobesappliedinfluorescenceinsituhybridization(FISH)analysesofcarbonatesedimentsfromCheckerReef
Targetgroup Probe Stringency(%FA) Reference
Neg.control NON 35 Wallneretal.1993
Archaea Arch915 20 Loyetal.2002
MarineCrenarchaeota GI-554 20 Manzetal.1992
Bacteria(excl.anammoxclade) EUB338I–III 35 Daimsetal.1999
Planctomycetales(excl.anammoxclade) EUB338II 35 Daimsetal.1999
Anammoxclade Amx368 15 Schmidetal.2003
PhylumBacteroidetes(excl.Flavobacteria) CFB719 30 Welleretal.2000
Flavobacteria CFB563 20 Welleretal.2000
b-Proteobacteria BET42a,unlabeledGAM42a 35 Manzetal.1992
c-Proteobacteria GAM42a,unlabeledBET42a 35 Manzetal.1992
PhylumNitrospirae Ntspa712 50 Daimsetal.2001
Nitrosomonadales Nso190 55 Mobarryetal.1996
Withexceptionofthecompetitors,alloligonucleotideswere50-labeledwithCy3.Stringencyreferstotheconcentrationofformamide(FA)inthe
hybridizationbuffer
123
18 CoralReefs(2009)28:15–26
to compare the percentage of probe-detected cells to the competentEscherichiacoli(OneShot(cid:3)TOP10,Invitrogen).
percentage of false positives (NON 338), a t-test or a Transformantsweregrownovernightat37(cid:2)ConLBagar/
Behrens/Fishertestwasapplied,dependingontheoutcome ampicillin (130 mg l-1) plates. Plasmid preparation and
of an F-test for possible differences in variance (Sachs sequencing were conducted by the University of Hawaii’s
1997).AllFISHdataaregivenin%ofDAPI-stainedcells Center for Genomics, Proteomics and Bioinformatics.
and represent the difference between the probe-related In order to determine the phylogenetic affiliation of the
signal and the unspecific background signal. In compari- novel clones, their closest matches were sought in Gen-
sons of FISH data between different samples, the same Bank by BLASTN 2.2.18 (Altschul et al. 1997) and in
statistical method was applied. RDPRelease9.60bytheClassifier(Wanget al.2007)and
Sequence Match tools. Novel sequences have been
Extraction of RNA deposited into GenBank under accession numbers
EU121687 – EU121844.
FromthreesedimentsamplescollectedonJuly12,2006at To estimate the community’s ribotype richness from a
depths representing oxic, interfacial, and anoxic habitats, sample of clones, a flat rank distribution approach (Lunn
RNA was extracted by a combination of bead beating and et al.2004)wasappliedtothesingleton-onlylibraries.For
phenol/chloroform treatment as described by Biddle et al. librariesconsistingofsingletonsanddoubletons,onecould
(2006). Briefly, samples were exposed to four iterations of compute the Sˆ estimator of species richness (Chao
Chao1
increasingly harsh extraction by phenol/sodium dodecyl 1987), with corrective terms to account for undersampling
sulfate (SDS) and beating with 0.1 mm zirconia beads. bias (e.g., Kemp and Aller 2004; Chao 2005; Schloss and
After the combination of the aqueous extracts, RNA was Handelsman 2005). This approach, however, is valid only
transferred into chloroform, precipitated, washed with eth- where the singletons and doubletons represent the rare
anol, and dissolved in nuclease-free water. Two-thirds of species of the community, i.e., the majority of phylotypes
these extracts were immediately reverse transcribed into occurthreeormoretimes(Chao1987,2005),andtherefore
cDNA using the iScriptTM system (Biorad) with random its application to the Checker Reef clone libraries was
hexamerprimers,whereasone-thirdoftheRNAextractwas dismissed as misleading.
retainedforanegativecontrolindownstreamapplications.
BothRNAandcDNAwerestoredat-80(cid:2)Cuntilanalysis.
Results
Polymerase chain reaction (PCR)
Reef porewater geochemistry
PCRsofbacterial16SrRNA,usingcDNAtemplates,were
performed on a Biorad MyiQ(cid:3) thermocycler, with SYBR The depth profiles of O , NO -, NO -, NH?, and DIP
2 3 2 4
Green(cid:3)-based real-time monitoring of the amplification. concentrations in Checker Reef porewater samples are
The PCR reactions contained 1 lM primers Bac 341f and shown in Fig. 1. Dissolved O was present down to
2
Bac 521r (Sørensen et al. 2005), 3 mM Mg2?, 0.4 mM of 30–35 cm and 25–30 cm depth in April and July/May,
each dNTP, and 25 U ml-1 iTaqTM DNA polymerase (iQ respectively. Below this oxygenated layer, [NH ?] and
4
SYBR Green(cid:3) Supermix, Biorad). All reactions were [DIP] became considerable, and a sulfidic smell was
subject to initial denaturation at 95(cid:2)C for 3 min; then 30 noticed.PorewaterNO –wasdepleted0–5 cmdeeperthan
3
cycles of 10 s at 95(cid:2)C, 30 s at 58(cid:2)C and 30 s at 72(cid:2)C; O , and its maximum concentrations were recorded at
2
followed by 3–5 min of final extension at 72(cid:2)C. DNA of 25–30 cm depth (Fig. 1).
Pseudomonas chlororaphis ATCC 43928 was used for a For the following discussion, oxic habitats are defined
positive control. The correct length and purity of the PCR by[O ] C 0.5[O ] ,[NH ?]\3 lMandtheabsenceof
2 2 sat. 4
productswereverifiedbyagarosegelelectrophoresis.PCR dissolvedsulfide.Sulfidic,O -freeporewatercharacterized
2
products were stored at -20(cid:2)C over night before further anoxic benthic habitats. Interfacial habitats are defined by
processing into a clone library. [O ]\0.5[O ] ,peak[NO -],[NH ?]\10 lMandthe
2 2 sat. 3 4
absence of olfactorily detectable sulfide from the reef
Clone libraries porewater.
Organic material in the upper 60 cm of Checker Reef
Clone libraries were established from amplificates of bac- amounted to 3.2–4.0% (wet weight), without obvious var-
terial16SrRNAinthethreesedimentextracts.ThesePCR iationoverdepth.Theratioofdissolvedinorganicnitrogen
products were ligated into the pCR(cid:3)4TOPO vector, using (DIN) to DIP can be indicative of the quality of organic
the TOPO-TA Cloning kit (Invitrogen). The ligation material degraded in the sediment. In Checker Reef pore-
products were then used to transform chemically waters, [DIN]:[DIP] ranged from 8 to 42 in July 2006 and
123
CoralReefs(2009)28:15–26 19
Fig.1 ConcentrationsofO ,
2
NO -,NO -,NH ?and a b
3 2 4
dissolvedinorganicphosphate
(DIP)ininterstitialwaters
collectedfromCheckerReef.
(a)Depthprofilesof[O ]on
2
April20,2006(r),July12,
2006(d),andMay21,2007
(j);anapparentoffsetof50–
60lMinAprilislikelydueto
aircontaminationduring
sampling.(b)Porewater
concentrationsofNO -(s),
3
NO -(9),NH ?(D)onApril
2 4
20,2006;DIP(h)wasnot
measured.(c)July12,2006.(d)
May21,2007;[NO -]
2
amountedtolessthan1.5%of
[NO -]
3
c d
from 24 to 52 in May 2007, without consistent trend with
depth or relation to oxic, interfacial, and anoxic zones.
MicrobialcellabundancesareshowninFig. 2.Theywere
108–109 (g sed)-1, with larger numbers found in the top
10 cm and in the anoxic layer, while fewer cells occurred
between 15 cm and 30 cm depth (Fig. 2). Microbial cell
abundance was not significantly correlated with TOM or
withanyoneoftheporewatersolutesanalyzed.
Composition of the active microbial community
The physiologically active bacterial population of oxic,
interfacial, and anoxic benthic habitats as defined above
was surveyed by constructing clone libraries from envi-
ronmental RNA extracts. Table 2 presents the resulting
Fig.2 CellabundancesinCheckerReefsedimentsonApril20,2006:
inventory for Checker Reef in comparison to DNA-based
(r);July12,2006:(d,s);andMay21,2007:(j,h).Errorbarsrepresent
datafromnearbyMokuoLoeReef(Sørensenet al.2007).
thestandarddeviationof24cellcountsinthesamesample.Opensymbols
The bacterial ribotypes detected in Checker Reef were markwhereparallelsampleswereusedformolecularanalyses
distributed over a broad range of phyla. Members of
Chloroflexi, Acidobacteria, a-, c-, and d-Proteobacteria the dominant groups in DNA-based Moku o Loe clone
werewellrepresentedintheseclonelibraries;Bacteroidetes libraries,appearedtobeminorcontributors orabsent from
and unclassified neighbors of Cyanobacteria, while among the RNA pool of Checker Reef (Table 2). Seven ribotypes
123
20 CoralReefs(2009)28:15–26
Table2 Taxonomicaffiliationofbacterial16SrRNAphylotypesdetectedincDNAfromoxic,interfacial,andanoxicCheckerReefsediment
%ofallphylotypeswithinlibrary
CheckerReef(RNA) MokuoLoeReef(DNA)
Oxic Interfacial Anoxic Surface 35cmdepth Anoxic
Cyanobacteria 2 6 0 19 0 0
UnclassifiedneighborsofCyanobacteria 0 0 0 29 9 6
Chloroflexi 25 6 11 3 6 23
Bacteroidetes 0 2 0 16 4 8
Planctomycetes 4 6 2 10 13 6
a-Proteobacteria 6 14 5 6 3 6
b-Proteobacteria 0 2 0 0 7 0
c-Proteobacteria 18 19 18 16 22 8
d-Proteobacteria 4 11 18 0 12 21
Nitrospirae 0 2 5 0 3 4
Acidobacteria 33 22 5 0 13 4
Others 0 2 29 0 7 15
Unclassified 8 8 7
Percentagesrefertothetotalnumberofribotypesfoundintherespectivelibrary(oxic:51;interfacial:63;anoxic:44).Correspondingdatabased
onDNAfromMokuoLoeReefsediments(Sørensenetal.2007)aregivenforcomparison.BothreefsitesarelocatedinKaneoheBay,about0.7
nauticalmilesfromeachother.ThetotalnumbersofphylotypesreportedinSørensen’slibrarieswere31(surface,oxic),68(35cmdepth)and52
(75cmdepth,anoxic)
Table3 Richnessofoperationaltaxonomicunits(OTUs)inbacterial16SrRNAclonelibrariesfromCheckerReefsediment
Thresholdidentity Library Numberofsequencesoccurring Richness
Once Twice Thrice 4times 5times 6times 7times
C99.5% Oxic 49 2 0 0 0 0 0 51
Interfacial 59 4 0 0 0 0 0 63
Anoxic 44 0 0 0 0 0 0 44
Combined 152 6 0 0 0 0 0 158
C99.0% Oxic 42 4 1 0 0 0 0 47
Interfacial 54 3 1 1 0 0 0 59
Anoxic 35 1 1 1 0 0 0 38
Combined 129 9 3 2 0 0 0 143
C95% Oxic 33 5 1 0 0 0 1 39
Interfacial 41 4 3 1 1 0 0 50
Anoxic 24 5 0 1 0 1 0 31
Combined 88 12 7 2 2 1 1 113
Sequence identity thresholds of 99.5, 99.0 and 95% were used as operational definitions of ribotype, species and genus level, respectively.
Librarysizeswere53(oxic),67(interfacial),and44(anoxic)sequencedclones
(2, 3 and 2 from oxic, interfacial, and anoxic zones, unculturedbacteriumassociatedwithMontastreafranksiin
respectively,allbelongingtoChloroflexiorAcidobacteria) Bermuda(Rohwer et al. 2002).
detected in Checker Reef sediment were closely related For a quantitative description of diversity in the clone
(95–98%)tophylotypesfoundat 35 cmor75 cmdepthin libraries, Table 3 presents the frequency distribution of
Moku o Loe Reef (Sørensen et al. 2007). Another three sequenced clones that are unique (singletons), double, tri-
ribotypes from anoxic Checker Reef sediment, affiliated ple, or more clustered, when different threshold levels of
withd-Proteobacteria,werecloselyrelated(92–94%)toan sequence identity are applied. Sequences that show
123
CoralReefs(2009)28:15–26 21
C99.5% identity, i.e., do not differ within the limits of Quantitative community structure
analytical error, are termed ribotypes. Operational defini-
tions of species and genus level are C99% (van Passel The quantitative structure of the microbial communities in
et al. 2006) and C95% sequence identity (Schloss and oxic, interfacial, and anoxic Checker Reef sediments, as
Handelsman 2004), respectively. At the ribotype level, the determined by FISH with group-specific, rRNA-targeted
vast majority of sequences occurred only once, indicating oligonucleotide probes (Table 1), is presented inFig. 3. In
that the diversity of the active bacterial community by far July 2006, overall detection by the domain-specific probes
exceeded the size of the libraries. Applying a flat rank was 58–66% of DAPI-stained cells, indicating similar
distribution approach (Lunn et al. 2004) to the singleton- levels of general physiological activity in all three com-
only library from anoxic sediment yielded an estimate of munities. The relative abundance of bacteria did not
C1380 different bacterial ribotypes in that habitat. The change with depth (p = 0.05, t-test), whereas Archaea
richness, i.e., number of different OTUs, among the were more abundant in interfacial than in oxic and anoxic
sequenced clones was 86–88% and 70–74% of the library sediment (p B 0.01, Behrens/Fisher test, t-test). Two-
size at the species and genus level, respectively (Table 3). thirds of the archaeal cells were identified as non-ther-
With the exception of one doubleton at the species level, mophilic marine Crenarchaeota. Note that only 50% of
overlaps between the three libraries occurred at no lower known sequences in this group are detected by the
than genus level (Table 3). domain-specific probe Arch915, compared to 78%
Fig.3 Microbialcommunity
a b
structureinCheckerReef
sedimentsasdeterminedby
fluorescenceinsitu
hybridization(FISH)analysis.
(a)July12,2006:oxicsediment
(15cmdepth);(b)July12,
2006:interfacialsediment
(25cmdepth);(c)July12,
2006:anoxicsediment(50cm
depth);(d)May21,2007:
interfacialsediment(28cm
depth).Notethatanammox
bacteriawereonlyprobedforin
theMaysample
c d
marine Crenarchaeota other Archaea anammox bacteria
other Planctomycetales Bacteroidetes Flavobacteria
Nitrosomonadales other b-Proteobacteria g -Proteobacteria
Nitrospirae other Bacteria not detected
123
22 CoralReefs(2009)28:15–26
coverage by the group-specific probe GI-554 (RDP-II Discussion
database, version 8.1).
Planctomycetales were the most abundant bacterial Coral reef microbial benthos in oxic, interfacial, and
groupintheoxicandinterfacialhabitat,butcomprisedless anoxic habitats
than 5% of the community in anoxic sediment (Fig. 3).
Their poor representation in the Checker Reef clone Oxygen is involved in the mineralization of OM both as
libraries(Table 2)canbeexplainedbyprimerbias:only47 terminal electron acceptor of aerobic respiration and as
of the currently 5,200 published 16S rRNA sequence oxidant for the reduced products of anaerobic respiration,
fragments of planctomycetes contain the priming site such as NH ? and sulfides. Metabolic pathways and the
4
complementary to Bac341f, the forward primer applied in structure of the catalyzing microbial community are con-
the PCRs (RDP-II database, release 9.60). Another 9–14% strained by their redox environment. Porewaters of the
of the benthic microbial population in Checker Reef highly permeable sediment of Checker Reef contained O
2
belonged to the b-Proteobacteria, with a substantial share toadepthof*30 cm(Fig. 1)ormore(FalterandSansone
in the order Nitrosomonadales, especially in interfacial 2000a).IfO consumptionratesareashighasmeasuredin
2
sediment (Fig. 3), where they were more abundant than in various coral reef sediments (Sansone et al. 1990; Grenz
theoxicoranoxichabitat(p\0.01,t-test).Thisapparently et al.2003;Rasheedet al.2004;Wildet al.2004a;Werner
dominantgroupwasseverelyunderrepresentedintheclone et al.2006),thenaconcomitanthighrateofO transportis
2
libraries (Table 2), without obvious blame to primer bias. required. Oceanic swells and local waves in Kaneohe Bay
Members of the c-Proteobacteria were the most abundant drive interstitial flow in the upper 1 m of Checker Reef at
group(18%)inanoxicsediment;theymadeup7–8%ofthe vertical velocities of up to 216 cm d-1 (Tribble et al.
inhabitants of oxic and interfacial sediment (Fig. 3). The 1992), providing an efficient mechanism of O influx. The
2
relative abundance of c-Proteobacteria was positively cor- resulting deep oxic zone is a suitable habitat for a benthic
related with [NH ?] in the porewater (p\0.05, n = 4). population of obligately and facultatively aerobic Bacteria
4
Based on the 16S rRNA survey, c-Proteobacteria in oxic and Archaea.
andinterfacialsedimentwerepredominantlyaffiliatedwith Interfacial habitats are likely to favor a community of
the order Chromatiales, whereas Pseudomonadales aboun- microaerophilic, aerotolerant, and facultatively aerobic
ded in the anoxic habitat. Bacteroidetes (excl. species. Alternatively, a mosaic of oxic and anoxic
Flavobacteria) occurred at 4–6% of DAPI-stained cells in microenvironments may be established, so that the micro-
all three habitats, with slightly lower abundance in oxic bial population in macroscopic samples of sediment will
than in interfacial sediment (p\0.05, t-test). Even fewer represent a blend of organisms belonging to the strictly
Flavobacteria were detected. The phylum Nitrospirae was aerobic and strictly anaerobic communities. Between the
foundmostlyintheinterfacialsediment,constituting5%of clone libraries of bacterial 16S rRNA from Checker Reef,
the community there (Fig. 3). The relative abundance of with a combined size of 164 sequences, there were only
Nitrospirae was positively correlated with [NO -] in the one and three overlaps at species and genus level, respec-
3
reef porewater (p\0.05, n = 4). tively (Table 3). However, as these libraries are small
In interfacial reef sediment collected in May 2007, compared to the estimated diversity of the natural com-
overall detection by the domain-specific probes was 55% munities (e.g., anoxic sediment: 44 clones, estimated
of DAPI-stained cells (Fig. 3d); note that the anammox richness C1,380), overlaps could have been missed.
cladeisoutsidethephylogeneticcoverageoftheseprobes. The microbial community structure in oxic, interfacial,
Compared to the corresponding sample from July 2006 and anoxic benthic habitats of Checker Reef reflected the
(Fig. 3b),therelativeabundancesofmarineCrenarchaeota, O -related ecophysiology of some detected groups. Plan-
2
Flavobacteria, and c-Proteobacteria did not change ctomycetes, with the exception of anaerobic NH ?
4
(p = 0.10, t-test). Archaea were more abundant in May oxidizing (anammox) bacteria, have been characterized as
than in July (p\0.01, t-test), while the relative abun- obligate or facultative aerobes (Fuerst 1995; Schlesner
dances of all other groups probed for were higher in July et al. 2004). Correspondingly, they were highly abundant
than in May (p\0.05, Behrens/Fisher test, t-test). Poten- in oxic and interfacial reef sediments, but much less so in
tial anammox bacteria constituted 6% of DAPI-stained the anoxic sediment (Fig. 3). Cultured representatives of
cellsinMay2007(Fig. 3d)andtheywerenotquantifiedin Bacteroidetes appear exclusively anaerobic (Kirchman
the July samples. Note that the probe Amx368 (Table 1) 2002), and in Checker Reef, members of this group were
can also hybridize with the 16S rRNA ofseveral microbes detected in oxic sediments at lower abundance than in the
outside the anammox clade, mostly uncultured Plancto- oxic/anoxic transition zone (Fig. 3). Members of the phy-
mycetales and Archaea (SILVA database, Pruesse et al. lumNitrospiraewerealmostabsentfromtheanoxichabitat
2007). and reached significant abundance only in the interfacial
123
CoralReefs(2009)28:15–26 23
sediment of Checker Reef (Fig. 3). Nitrospira spp. are between Moku o Loe and Checker Reef, for example, in
describedastypicallyaerobicNO -oxidizers(Daimset al. human impact and hydrographic regime. Both inventories
2
2001; Koops and Pommerening-Ro¨ser 2001) and the provide valuable information on candidate groups or spe-
dominant members of that guild in many systems (Rev- cies to expect in the subsurface part of coral ecosystems.
sbech et al. 2006). The apparent correlation between Small suspended particles carried by interstitial water
porewater [NO -] and the abundance of Nitrospirae gives flows get trapped and efficiently degraded within the reef
3
supportive evidence of their role as NO – oxidizers in sediment(Rasheedet al.2004;Wildet al.2004a,b,2005),
2
Checker Reef sediments. Further O -related differences in presumablybyheterotrophicBacteriaandArchaea.Benthic
2
community structure were not resolved by the FISH anal- microorganisms occur at abundances of 108–109 cells g-1
yses, as ecophysiology often varies at taxonomic levels (Fig. 2; Hansen et al. 1987; Wild et al. 2006; Sørensen
lower than those targeted by the set of probes applied in et al.2007),butthenumberofheterotrophicallyactivecells
this study. While the data from the clone libraries were remains unknown. Most phylogenetic groups detected in
phylogenetically more finely resolved, they lack informa- the RNA-based clone libraries and FISH analyses include
tion about abundance and suffer from considerable heterotrophic members, and thus there is at least the
methodical bias (Table 2 vs. Fig. 3). potentialformicrobialOMdegradationtobeconsiderable.
Checker Reef sediments were relatively rich in OM com-
Potential metabolic roles of ribosome-rich reef pared to sandy reef sediments elsewhere (Alongi et al.
microbes 1996, Rasheed et al. 2002, 2003, Schrimm et al. 2004).
However, as TOM content was virtually invariant with
The gross level of physiological activity of microbial cells depth, i.e., across a broad range of redox conditions and
is reflected in their ribosome content (Kemp et al. 1993; TOM influx rates, it is likely to represent refractory mate-
KerkhofandWard1993),whichinturnisamajorlimiting rial, unavailable due to its chemical composition or to
factorfor occurrenceinRNA-basedclonelibraries andfor enclosure deep in the cavities and crevices of biogenic
detection by rRNA-targeted FISH probes. The combined carbonate grains. Almost equally high TOM contents of
detection of Archaea and Bacteria in the FISH analyses of 2.9–3.2% (wet weight) were measured in sediments from
Checker Reef samples was 58–66% of DAPI-stained cells Hawaiian reefs at Kilo Nalu and Moku o Loe (data not
(Fig. 3) and appeared to be independent of the redox state shown).The porewater [DIN]:[DIP] ratio may indicate the
of the porewater. For comparison, 13–84% of microbial source and state of degradation of the OM respired in the
cellswereFISH-detectableinthefewpermeablesediments sediment, but is also influenced by denitrification and P
analyzed previously, all of them terrigenous silicate sands bindingontomineralphases,especiallyCaCO (Ogrincand
3
(Llobet-Brossa et al. 1998; Rusch et al. 2003; Ishii et al. Faganeli 2006). The values for this ratio obtained in the
2004; Bu¨hring et al. 2005). Aside from methodical limi- present study, ranging between 8 and 52, envelope a pre-
tations, these values may indicate that varying parts of the vious estimate of 37 for Checker Reef porewater at 1 m
microbialpopulationweretemporarilyinactiveinresponse depth(Tribbleet al.1990).BasedonthedatafromChecker
to the quickly changing porewater geochemistry of per- Reef (Haberstroh and Sansone 1999; Falter and Sansone
meable sediments. 2000b), it has been argued that [DIN]:[DIP] in oxic reef
DNA-based molecular analyses manifest community porewaters is one-third to one-half of that in deeper pore-
structure by cell abundance and operon number, whereas waters(AtkinsonandFalter2003).Measurementsfromthe
RNA-based data reflect gene expression and growth rate. present study cannot supportthisclaim.
The census of sedimentary DNA pools (e.g., Sørensen Chemoorganotrophs capable of hydrolyzing biopoly-
et al. 2007) includes inactive and dead cells as well as mers, of fermenting or respiring organic compounds with
extracellular DNA attached to the sand grains (Naviaux NO - or other electron acceptors are well represented in
3
et al. 2005), whereas RNA-targeted methods probe the Planctomycetes (Fuerst 1995; Schlesner et al. 2004),
active fraction of the population. Between DNA- and Pseudomonadales, Oceanospirillales, b-Proteobacteria
RNA-based clone libraries established from the same (excl.Nitrosomonadales),Bacteroidetes,andFlavobacteria
sample of polluted soil (Nogales et al. 2001) or marine (Shewan and McKeekin 1983; Cottrell and Kirchman
bacterioplankton (Moeseneder et al. 2005), fewer than 30 2000; Weller et al. 2000; Kirchman 2002). These promi-
% of phylotypes were found in both libraries. The differ- nent groups (Table 2, Fig. 3) could be involved in various
enceinoutcomebetweenDNA-andRNA-basedsurveysof steps and pathways of OM degradation in Checker Reef
coral reef bacterial benthos (Table 2) may be particularly and Moku o Loe Reef sediments.
pronounced due to rapid redox fluctuations that affect ThebenthicmicrobialcommunitiesofCheckerReefalso
microbial activity and hence the RNA but not the DNA holdthepotentialforchemoautotrophicmetabolism,suchas
pool. In addition, there are reef-specific differences aerobic and anaerobic NH ? oxidation. Nitrosomonadales,
4
123
24 CoralReefs(2009)28:15–26
with all cultured members characterized as chemolithotro- BiddleJF,LippJS,LeverMA,LloydKG,SørensenKB,AndersonR,
phic or mixotrophic NH ? oxidizers (Garrity et al. 2005), Fredricks HF, Elvert M, Kelly TJ, Schrag DP, Sogin ML,
4
Brenchley JE, Teske A, House CH, Hinrichs K-U (2006)
were the dominant fraction of b-Proteobacteria inChecker
Heterotrophic Archaea dominate sedimentary subsurface eco-
Reef.MarineCrenarchaeota,someofwhichcarrygenesfor systemsoffPeru.ProcNatlAcadSciUSA103:3846–3851
NH ? oxidizing enzymes (Francis et al. 2007), constituted Breitbart M, Bhagooli R, Griffin S, Johnston I, Rohwer F (2005)
4
themajorpartofthearchaealbenthos(Fig. 3).Thesignifi- Microbial communities associated with skeletal tumors on
Poritescompressa.FEMSMicrobiolLett243:431–436
cantly enhanced occurrence of both groups in interfacial
Bu¨hring SI, Elvert M, Witte U (2005) The microbial community
sediment suggests that aerobic NH4? oxidation may be an structure of different permeable sandy sediments characterized
importantbacterialandarchaealmetabolisminhabitatsthat by the investigation of bacterial fatty acids and fluorescence
experiencethespatialortemporalproximityofbothoxicand insituhybridization.EnvironMicrobiol7:281–293
CaponeDG,DunhamSE,HorriganSG,DuguayLE(1992)Microbial
anoxic conditions. Bacteria with the putative capacity to
nitrogentransformationsinunconsolidatedcoralreefsediments.
oxidizeNH4?anaerobicallywithNO2-werepresentinthe MarEcolProgSer80:75–88
rRNA pool of interfacial reef sediment (Fig. 3d). More ChaoA(1987)Estimatingpopulationsizeforcapture-recapturedata
detailed investigations on these reef-dwelling anammox withunequalcatchability.Biometrics43:783–791
ChaoA(2005)Speciesrichnessestimation.In:BalakrishnanN,Read
bacteriaareunderwayintheauthors’laboratory.
CB,VidakovicB(eds)Encyclopediaofstatisticalsciences.John
This ribotype survey and FISH quantification of physi- Wiley&Sons,NewYork,pp7909–7916
ologically active prokaryotes in permeable reef sediments Cottrell MT, Kirchman DL (2000) Natural assemblages of marine
suggest phylogenetically diverse benthic communities. proteobacteria and members of the Cytophaga-Flavobacter
cluster consuming low- and high-molecular weight dissolved
Members of all detected groups have been found previ-
organicmatter.ApplEnvironMicrobiol66:1692–1697
ously in temperate marine environments, but all Checker DaimsH,Bru¨hlA,RudolfA,SchleiferK-H,WagnerM(1999)The
Reef ribotypes were novel. It remains an open question domain-specific probe EUB338 is insufficient for the detection
whether these reef microbes are endemic to reefs (i.e., ofallBacteria:developmentandevaluation ofamorecompre-
hensiveprobeset.SystApplMicrobiol22:434–444
specializedatthespecieslevel)orarecosmopolitanspecies
DaimsH,NielsenJL,NielsenPH,SchleiferK-H,WagnerM(2001)
forming assemblages that are reef-specialized at the com- Insitu characterization of Nitrospira-like nitrite oxidizing bac-
munity level. The microbial community structure in oxic, teria active in wastewater treatment plants. Appl Environ
interfacial, and anoxic benthic habitats of Checker Reef Microbiol67:5273–5284
D’Elia CF, Wiebe WJ (1990) Biogeochemical nutrient cycles in
largely reflected the O -related ecophysiology of most
2 coral-reef ecosystems. In: Dubinsky Z (ed) Ecosystems of the
detected groups. Members of several abundant groups are world.25.Coralreefs.Elsevier,Amsterdam,pp49–74
likelycontributorstothebenthicdecompositionofOMand Falter JL, Sansone FJ (2000a) Hydraulic control of pore water
regeneration of nutrients that are crucial for sustaining the geochemistry within the oxic-suboxic zone of a permeable
sediment.LimnolOceanogr45:550–557
high productivity of coral reefs in oligotrophic tropical
Falter JL, Sansone FJ (2000b) Shallow pore water sampling in reef
marine environments. sediments.CoralReefs19:93–97
Francis CA, Beman JM, Kuypers MMM (2007) New processes and
Acknowledgments The authors gratefully acknowledge logistical playersinthenitrogencycle:themicrobialecologyofanaerobic
support by the Friends of Heeia (Kaneohe), who made kayaks andarchaealammoniaoxidation.TheISMEJournal1:19–27
availableforoursamplingexcursions.SeanOgataassistedwithTOM Fuerst JA (1995) The planctomycetes - emerging models for
analysesandcellcounts.WethankBrianPoppforsupplyingapure microbial ecology, evolution and cell biology. Microbiology
culture of Pseudomonas chlororaphis. The German Research 141:1493–1506
Foundation (DFG) supported A.R. with a research scholarship, and GarrityGM, BellJA,LilburnT(2005)OrderV:Nitrosomonadales.
A.K.H.wasfundedbyNSFgrantsOCE-0327332andGEO-0311912. In:GarrityGM,BrennerDJ,KriegNR,StaleyJT(eds)Bergey’s
manualofsystematicbacteriology,vol.2partC.Springer,New
York,pp863–886
Grenz C, Denis L, Boucher G, Chauvaud L, Clavier J, Fichez R,
References Pringault O (2003) Spatial variability in sediment oxygen
consumption under winter conditions in a lagoonal system in
AlongiDM, Tirendi F,Goldrick A(1996)Organic matter oxidation NewCaledonia(SouthPacific).JExpMarBiolEcol285:33–47
andsedimentchemistryinmixedterrigenous-carbonatesandsof HaberstrohPR,SansoneFJ(1999)Reefframeworkdiagenesisacross
NingalooReef,WesternAustralia.MarChem54:203–219 wave-flushed oxic-suboxic-anoxic transitionzones. CoralReefs
AltschulSF,MaddenTL,Scha¨fferAA,ZhangJ,ZhangZ,MillerW, 18:229–240
Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a new HannidesAK(2008)Organicmattercyclingandnutrientdynamicsin
generation of protein database search programs. Nucleic Acids marinesediments.Ph.D.thesis,UniversityofHawaiiatManoa,
Res25:3389–3402 p439
Atkinson MJ, Falter JL (2003) Biogeochemistry of coral reefs. In: Hansen JA, Alongi DM, Moriarty DJW, Pollard PC (1987) The
Black KP, Shimmield GB (eds) Biogeochemistry of marine dynamics of benthic microbial communities at Davies Reef,
systems.CRCPress,BocaRaton,pp40–64 centralGreatBarrierReef.CoralReefs6:63–70
Bendschneider K, Robinson RJ (1952) A new spectrophotometric Hewson I, Fuhrman JA (2006) Spatial and vertical biogeography of
methodforthedeterminationofnitriteinseawater.JMarRes coral reef sediment bacterial and diazotroph communities. Mar
11:87–96 EcolProgSer306:79–86
123