Diversity and quorum-sensing signal production ofProteobacteria associated with marine sponges
Naglaa M. Mohamed,1† Elisha M. Cicirelli,2†
Jinjun Kan,1 Feng Chen,1 Clay Fuqua2 andRussell T. Hill1*1Center of Marine Biotechnology, University of MarylandBiotechnology Institute, Baltimore, MD 21202, USA.2Department of Biology, 1001 E. 3rd St., Jordan Hall142, Indiana University, Bloomington, IN 47405, USA.
Summary
Marine sponges are hosts to diverse and dense bac-terial communities and thus provide a potential envi-ronment for quorum sensing. Quorum sensing, a keyfactor in cell–cell communication and bacterial colo-nization of higher animals, might be involved in thesymbiotic interactions between bacteria and theirsponge hosts. Given that marine Proteobacteria areknown to produce N-acyl homoserine lactone (AHL)signal molecules, we tested the production of AHLsby Alpha- and Gammaproteobacteria isolated frommarine sponges Mycale laxissima and Ircinia stro-bilina and the surrounding water column. We usedthree different AHL biodetection systems in diffusionassays: Chromobacterium violaceum, Agrobacteriumtumefaciens and Sinorhizobium meliloti with optimalsensitivity to short-chain (C4–C6), moderate-chain(C8–C12) and long-chain (� C14) AHLs respectively.Thirteen of 23 isolates from M. laxissima and five of25 isolates from I. strobilina were found to produceAHLs. Signals were detected from two of eight pro-teobacterial strains from the water column. Thin-layerchromatographic assays based on the A. tumefaciensreporter system were utilized to determine the AHLprofiles of the positive isolates. The types andamounts of AHLs synthesized varied considerablyamong the strains. Small ribosomal rRNA genesequencing revealed that the AHL-producingalphaproteobacterial isolates were mainly from theSilicibacter–Ruegeria subgroup of the Roseobacterclade. Two-dimensional gel electrophoresis (2DGE)-based proteomic analyses were congruent with phy-
logenetic relationships but provided higher resolutionto differentiate these closely related AHL-producingstrains.
Introduction
As one of the oldest multicellular animals, marinesponges are sessile filter feeders that consume bacteria,phytoplankton, algae and other particulate marine matteras food. Bacteria filtered from the surrounding water pri-marily serve as a food source for the sponge, and undi-gested bacteria are either returned to the water column orretained within the sponge. Despite the fact that spongesacquire nutrients by phagocytosis of bacteria (van Soest,1996), large numbers of bacteria inhabit the mesohylmatrix of sponges (Webster et al., 2001; Montalvo et al.,2005). These endosymbionts can constitute up to 60%of the sponge biomass (Vacelet, 1975; Vacelet andDonadey, 1977; Wilkinson, 1978; Hentschel et al., 2006).Marine sponges act as microbial fermenters (Hentschelet al., 2006) and are potential models to study symbiosis.The modes of acquisition of symbionts include filtrationfrom the surrounding seawater and vertical transmissionof symbionts from parent sponges through larvae or germcells. Vertical transmission has been shown conclusivelyin several cases (Enticknap et al., 2005; Maldonado et al.,2005; Oren et al., 2005; Usher et al., 2005; Hentschelet al., 2006; Sharp et al., 2007). Symbionts are hypoth-esized to contribute to the health and nutrition of spongesin a variety of ways including production of protectiveantibiotics and the acquisition of limiting nutrients (Wilkin-son, 1983; Corredor et al., 1988; Diaz and Ward, 1997;Hoffmann et al., 2005). Marine sponges provide a pro-tected and nutrient-rich niche in which extensive interac-tion among the dense and diverse microbial populations isfostered and perhaps unavoidable. One mechanism bywhich microbes can productively interact is throughchemical signalling.
Quorum sensing allows bacteria to communicate andregulate gene expression in a population density-dependent manner through the accumulation of signals,which are often diffusible (Fuqua and Greenberg, 2002;Waters and Bassler, 2005). Common quorum-sensingsignals among the Gram-negative Proteobacteria are theN-acyl homoserine lactones (AHLs). N-acyl homoserinelactones were discovered from studies of the symbiotic
Received 6 April, 2007; accepted 2 August, 2007. *Forcorrespondence. E-mail [email protected]; Tel. (+1) 410 234883;Fax (+1) 410 2348896. †These authors contributed equally. Thisarticle is contribution No. 06-154 from the Center of MarineBiotechnology.
association between the bioluminescent marine het-erotrophic Gammaproteobacterium, Vibrio fischeri andcertain marine fishes and squids (Nealson et al., 1970;Eberhard et al., 1981). N-acyl homoserine lactone-basedquorum sensing is broadly distributed among Proteobac-teria, but the V. fischeri system has remained an importantmodel, particularly through studies of its symbiosis withthe Hawaiian bobtail squid, Euprymna scolopes (Visicket al. 2000). Despite the intense study of AHL signalling inV. fischeri and related marine vibrios, there is only limitedinformation on AHL-dependent regulation in other marinebacteria. Recently, a mixture of long- and short-chainAHLs were chemically characterized for a collection ofmarine alphaproteobacterial isolates (Wagner-Dobleret al., 2005). Gram and colleagues (2002) found thatthree Roseobacter isolates from marine snow producedcompounds capable of activating an AHL-responsive bio-logical reporter. In a separate study the cultivation effi-ciency of heterotrophic bacteria from the Baltic Sea wasenhanced on addition of cyclic AMP and AHLs, suggestinga broad role in adaptation of marine bacteria to conditionsin laboratory culture (Bruns et al., 2002). N-acyl homo-serine lactones can also affect the behaviour of marineeukaryotes. For example, bacterial biofilms that produceAHLs stimulate the settlement of motile zoospores of themarine alga Ulva intestinalis, and the inhibition of AHLsignalling can block this settlement (Joint et al., 2002; Taitet al., 2005).
The cell–cell communication afforded by AHLs oftencomes into play in the bacterial colonization of metazoanorganisms (Passador et al., 1993; Zhang et al., 1993;Ruby, 1996; Fuqua et al., 2001). Many pathogens utilizeAHLs to modulate specific aspects of infection. Likewise,symbiotic microbes often utilize AHL quorum sensing.N-acyl homoserine lactones have an important role in thedevelopment of symbiotic root nodules on leguminousplants by rhizobia (Oldroyd et al., 2005). Based on theseand many similar examples, it is appealing to hypothesizethat AHLs produced by sponge-associated bacteria mightbe involved in regulating symbiotic interactions with theirhosts. Quorum sensing might regulate colonization of thesponge by the bacteria or some of the important interac-tions among sponge microbiota, such as the production ofbioactive compounds. In a small-scale screen of several
marine invertebrates, Taylor and colleagues (2004)reported production of AHLs by sponge-associatedbacteria. One AHL-producing bacterial isolate from thesponge Cymbastela concentrica collected off of the coastof south-eastern Australia was found to be a member ofthe Roseobacter–Ruegeria marine bacterial subgroup.The first objective of this study was to survey the produc-tion of AHLs by Proteobacteria from the shallow watermarine sponges, Mycale laxissima and Ircinia strobilina.The second objective was to determine the phylogenyof AHL-producing bacteria utilizing 16S rRNA genesequencing. Finally, a proteomics approach was used toinvestigate fine-scale diversity among more closelyrelated AHL-producing proteobacterial isolates.
Results
Identification of AHL activities in bacteria isolated fromM. laxissima and I. strobilina
Cultivation of bacteria from sponge tissue of M. laxissimaand I. strobilina yielded a collection of bacterial isolatesfrom the groups Alpha- and Gammaproteobacteria andBacterioidetes (N.M. Mohamed, V. Rao, M.T. Hamann,M. Kelly and R.T. Hill, unpublished; N.M. Mohamed, J.J.Enticknap, J.E. Lohr, S.M. McIntosh and R.T. Hill,unpublished). Initial screening of the isolates showed thata portion of the proteobacterial isolates might produceAHLs. Based on these preliminary findings, we focusedon alpha- and gammaproteobacterial isolates for thedetection of AHL-signalling molecules.
Cultivable Proteobacteria from M. laxissima and I. stro-bilina were analysed by bioassays using three differentAHL reporter bacteria (Table 1). We used Chromobacte-rium violaceum strain 026 (Cv-AHL), which cannot synthe-size AHLs, but responds to the presence of short-chainAHLs (C4-C6) by producing violacein, a purple pigment(McClean et al., 1997). We also utilized a sensitive andbroader-spectrum AHL-responsive reporter derived fromAgrobacterium tumefaciens and described as an ultrasen-sitive AHL reporter (At-AHL) (Zhu et al., 2003). The At-AHLreporter responds efficiently to AHLs with acyl chainlengths of C6–C14, with decreasing response to AHLsshorter and longer than its cognate ligand 3-oxo-octanoyl-
a. N-acyl homoserine lactones produced by cognate AHL synthase to which reporter has the greatest sensitivity.b. Range of AHL side-chain length to which the reporter will detectably respond (specificity for R-group at b position).
L-homoserine lactone (3OC8-HSL). Finally, a Sinorhizo-bium meliloti AHL reporter (Sm-AHL), which responds tolong-chain AHLs, was also used (Llamas et al., 2004)(Tables 1 and 2, and Fig. 1). We utilized these AHL-responsive biodetection systems to score the sponge iso-lates for production of AHL-type inducers. In marine agaroverlay assays with the At-AHL reporter in large Petridishes (150 ¥ 15 mm), we defined three distinct ‘activa-tion’ phenotypes (Fig. 1): very strong activation (an activa-tion zone radius greater than 2.5 cm of activity), moderateactivation (a radius of 0.5–2.5 cm) and minimal activationof the reporter (a radius less than or equal to 0.5 cm). Dueto the overall weak coloration imparted by the Sm-AHLbiosensor, its activity was recorded as relative intensity of
positive strains (++ indicating strongest intensity, + indicat-ing roughly 50% of the intensity of strongest producingstrains). Twenty-three proteobacterial isolates from M. lax-issima were assayed for AHL production and 13 of the 23strains (57%) activated the At-AHL reporter, varying fromless than 0.5 cm to greater than 2.5 cm for the radius of theactivation zone. One of the 13 strains, N04ML5, anAlphaproteobacterium from M. laxissima, displayed thestrongest AHL activity, activating all three reporter strainsincluding the Cv-AHL reporter (Fig. 1). As further confir-mation of AHL production by the sponge isolates, theS. meliloti (Sm-AHL) reporter detected long-chain AHLproduction in 11 of the 13 AHL-positive strains (Fig. 1Cand Table 2). Four of the 25 proteobacterial strains iso-
Table 2. N-acyl homoserine lactones (AHL) production by 20 bacterial strains isolated from: seawater collected in 2001 (SWKLH7 and SWKLH8);M. laxissima collected in 2001 (KLH11), 2004 (N04ML) and 2005 (N05ML); and I. strobilina collected in 2004 (N04IS) and 2005 (N05IS).
a. Percent identity to the closest well-described taxa. GenBank accession numbers are listed before each reference sequence.b. Overlay bioassays using sensor strains A. tumefaciens (At-AHL) and S. meliloti (Sm-AHL).c. N-acyl homoserine lactones production activity is presented as: +++, radius > 2.5 cm of activity, ++, radius of 0.5–2.5 cm of activity, and +, radiusof < 0.5 cm of activity. Activation of the S. meliloti reporter strain is presented as relative intensity of positive strains. ++ = the strongest intensity,+, roughly 50% of the intensity of the strongest producing strains.d. AHL profile as assessed by thin-layer chromatography bioassay. Numbers are lane numbers referring to AHL profiles shown in Fig. 2.ND, not determined.
Fig. 1. Example overlay marine agarbioassays for AHL activity. Isolates areindicated around each plate; 8 – N04ML8(AHL negative), 5 – N04ML5 (AHL+), 32 –N05ML13 (weak AHL+) and 11 – KLH11(AHL+). Reporter strains (A) Cv-AHL, (B)At-AHL and (C) Sm-AHL.
BA C
8 11 8 11 8 11
5 32 5 32 5 32
Diversity and AHLs in sponge-associated Proteobacteria 77
lated from I. strobilina activated the At-AHL reporter inoverlay assays (Table 2). Two of the strains, which acti-vated the At-AHL reporter, were also able to activate theSm-AHL reporter (N04IS3 and N05IS9), indicating long-chain AHL production. One strain, N05IS8, a Gammapro-teobacterium from I. strobilina, activated only the Sm-AHLreporter, indicating primarily long-chain AHL production,outside the detection range of the other AHL biosensors.Of the eight proteobacterial isolates from the surroundingwater column assayed for AHL production, two Alphapro-teobacteria, SWKLH7 and SWKLH8, activated the At-AHLreporter.
Profiling of AHL production by thin-layerchromatography and bioassays
Reverse phase (RP) thin-layer chromatography (TLC)was utilized to determine the signal molecule profiles ofAHL+ isolates that tested positive in the plate-basedoverlay assays. A comparison of the AHL profiles revealedseveral common patterns (Table 2). A representative ofeach observed unique AHL profile from Alpha- and Gam-maproteobacteria is included in Fig. 2. Isolate KLH11(Fig. 2A, lane 3) from M. laxissima produced greater thansix different spots of activity, ranging in migration fromthose that do not move from the site of application (longchain, non-polar), to those that migrated the furthest(short, highly polar molecules), with significantly higher Rfvalues than the N-butanoyl homoserine lactone (C4-HSL)standard. KLH11 shared a nearly identical AHL profilewith five other alphaproteobacterial strains (N04ML2,
N04ML4, N04ML6, N04ML9 and N04IS3). N04ML5 andN05IS9, Alphaproteobacteria from M. laxissima andI. strobilina, respectively, had very distinct profiles. Incomparison with KLH11, N05IS9 produced two prominentactivities that had similar Rf values to 3O-C6 and 3O-C8AHL standards (Fig. 2A, lane 5). The AHL profiles ofstrains N04ML11 (lane 4) and N04ML5 (lane 6) wereclearly distinct from each other, but shared an apparentlack of AHL activities that migrated with high Rf valueson RP-TLC. The AHL+ Gammaproteobacteria producedmoderate-chain-length AHLs, lacking the non-polar activi-ties with low Rf values observed for the sponge-associated Alphaproteobacteria (Fig. 2B). Analysis of theAHL+ isolates from the water column resulted in AHLprofiles different from those of any of the sponge isolates.
Phylogenetic analysis and diversity of AHL-producingisolates
A phylogenetic tree of partial 16S rRNA gene sequenceswas constructed for AHL-producing proteobacterialstrains (Fig. 3). Eleven AHL+ alphaproteobacterial isolatesfell within the Silicibacter–Ruegeria (SR) subgroup of theRoseobacter clade and three isolates were in genera(Pseudovibrio and Erythrobacter) that are distinct fromthis subgroup (Table 2, Fig. 3). N-acyl homoserinelactone-producing gammaproteobacterial isolates fellwithin the genera Vibrio, Thalassomonas and Spongio-bacter. 16S rRNA sequence analysis indicated thatSWKLH7 and SWKLH8, the AHL+ seawater isolates,shared the greatest similarity to an Alphaproteobacterium
1 2 3 4 5 6 7 8 9 10 11 12
O12
O8
O6
4
6
8
10O12
O8
O6
4
6
8
10
A B
Fig. 2. Reverse phase (RP) TLC profiling of AHL+ isolates. Thin-layer chromatography plates were overlaid with At-AHL reporter strain.Mixtures of synthetic 3-oxo-AHL and fully reduced AHLs standards were run on each plate (labelled on plate), lanes 1 and 7 and 2 and 8respectively.A. Alphaproteobacterial strains; lanes 3–6 are KLH11, N04ML11, N05IS9 and N04ML5 respectively. Lane 6 is a 2000-fold dilution of N04ML5extract.B. Gammaproteobacterial strains; lanes 9 and 10 are N05ML4 and N05IS15 respectively. Lane 11 is a marine broth control and lane 12 isAHL+ water column isolate (SWKLH8). N-acyl homoserine lactones standard concentrations are: fully reduced, C4, 1 mM; C6, 500 mM; C8,50 nM; C10, 125 mM, and 3-oxo derivatives, 3-oxo-C6, 50 nM; 3-oxo-C8, 42 nM; 3-oxo-C12, 68 mM.
species in the genus Erythrobacter. Nine Roseobacterisolates were chosen to represent the AHL-producing SRsubgroup for further phylogenetic analysis including full16S rRNA phylogenetic analysis and proteomics. A phy-logenetic tree of the full 16S rRNA gene sequences of the
nine SR isolates is shown in Fig. 4. KLH11, N04ML9,N04ML6 and N04ML4 shared 99–100% identity andformed a cluster (cluster 1), which was closely related tothe roseobacters from I. strobilina, N04IS3 and N05IS9.N04ML2 and N04ML11 (cluster 2) were 99.2% identical
Fig. 3. Rooted neighbour-joining tree of partial 16S rRNA gene sequences of AHL-producing Alphaproteobacteria and Gammaproteobacteriafrom M. laxissima, collected in 2001 (KLH11), 2004 (N04ML) and 2005 (N05ML), I. strobilina, collected in 2004 (N04IS) and 2005 (N05IS) andseawater (SWKLH8). Bootstrap confidence values > 50% are shown at the nodes. Approximately 500 bp was used in the phylogeneticanalysis and branches indicated by f and p were found using Fitch–Margoliash and maximum parsimony methods respectively. The outgroupused in this analysis was Escherichia coli BL21 (NCBI Accession No. AJ605115). Scale bar indicates 0.10 substitutions per nucleotideposition. GenBank numbers are not given for partial 16S rRNA sequences from nine Roseobacter isolates; these GenBank numbers foralmost complete 16S rRNA sequences are provided in Fig. 4.
Diversity and AHLs in sponge-associated Proteobacteria 79
and clustered with Silicibacter sp. JC1077 (NCBI Acces-sion No. AF201086, 99.8–99.9% identity). N04ML5 wasnotably distinct with 98.9% identity to cluster 1 and 98.6%identity to cluster 2. This strain was closely related to aSR-type bacterium isolated from coral mucus (NCBIAccession No. AY654746, 99.3% identity). It is importantto note that 100% of the sponge-associated SR-typeRoseobacteria were AHL+, while two roseobacters iso-lated from the water column did not produce detectableAHLs (data not shown).
Biochemical characterization of Roseobacter strains
API 20NE strips were used for biochemical characteriza-tion of six AHL+ SR strains from M. laxissima. API 20NEprofiles were identical for strains, N04ML2, N04ML4,N04ML6, N04ML9 and N04ML11. The strains werepositive for b-galactosidase, negative for denitrification,indole production, glucose acidification, arginine dihydro-lase, urease, esculin hydrolysis, gelatin hydrolysis andoxidase, and negative for assimilation of the follow-ing sugars: glucose, arabinose, mannose, mannitol,N-acetylglucosamine, maltose, potassium gluconate,capric acid, malic acid, trisodium citrate and phenylaceticacid. N04ML5 was different from the other strains inbeing positive for nitrate reduction, urease and esculinhydrolysis and weakly positive for arginine dihydrolase.
Proteome pattern analyses
Proteomic analysis was used to investigate fine-scalediversity among the more closely related AHL-producingalphaproteobacterial isolates KLH11, N04ML9, N04ML4,N04IS3, N05IS9, N04ML5, N04ML2, N04ML11 and Silici-bacter pomeroyi DSS-3. As a first step, protein expression
profiles of KLH11 at three different growth stages (earlyand mid-exponential, and stationary phases) were com-pared (Fig. 5), to understand how functional proteins varyat the different growth stages. Although some modulationof protein levels was apparent at different growth periods,the two-dimensional gel electrophoresis (2DGE) imagesat the three different stages shared higher than 90%similarity. We decided to acquire samples at mid-logphase for other bacterial strains because functional pro-teins at this phase were more uniformly expressed andproduced a greater number of spots compared with theother time points. We tested the reproducibility of our2DGE system, and found that triplicates yielded highlyreproducible results (> 95% similarity, data not shown).The protein profiles of nine AHL-producing SR roseobac-ters were different from each other (Fig. S1), but could be
Fig. 4. Rooted neighbour-joining tree ofalmost complete 16S rRNA gene sequencesof AHL-producing SR roseobacters fromM. laxissima and I. strobilina. Bootstrapconfidence values > 50% are shown at thenodes. Approximately 900 bp was used in thephylogenetic analysis and branches indicatedby f and p were found using Fitch–Margoliashand maximum parsimony methodsrespectively. The outgroup used in thisanalysis was Agrobacterium sp. PNS-1 (NCBIAccession No. AY762361). Scale barindicates 0.10 substitutions per nucleotideposition.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0 10 20 30 40 50 60 70
Incubation time
T1
T2
T3
Fig. 5. Proteomes of KLH11 at three different growth stages, T1,T2 and T3, which correspond to early exponential, mid-exponentialand stationary phases respectively.
compared using the ImageMaster software. Multidimen-sional scaling (MDS) analysis showed that some bacterialstrains (i.e. N04ML4 and N04ML9; N04ML2 andN04ML11) share more similar 2DGE patterns, whileothers are more distinct to each other (Fig. 6). The relat-edness of these nine isolates based on the MDS analysissupported the branch points and general topology of the16S rRNA gene phylogeny shown in Fig. 4. For example,N04IS3 and N05IS9 were close to KLH11, N04ML9 andN04ML4, while N04ML5 was more similar to N04ML2 andN04ML11 (Fig. 6).
Discussion
The dense bacterial community harboured within thesponge mesohyl lends itself to interbacterial communica-tion mechanisms. The prevalence of sponge-associatedproteobacteria in this environment suggested that AHL-based signalling systems might be important. It is likelythat some sponge-specific bacteria are initially acquiredfrom the oligotrophic water column where they exist in lowabundance. Sponge tissues provide increased nutrientavailability relative to bulk seawater, and the colonizingbacteria can grow to achieve high population densitiesthat allow for a bacterial quorum. Opportunities for com-petition and collaborative metabolism are also providedwithin the sponge. We cultivated 48 individual proteobac-terial isolates from sponge tissues, 38% (18 total) of whichwere AHL+. Among the bacteria tested, AHL productionwas more frequently observed for the Proteobacteriaassociated with M. laxissima than those with I. strobilina,although the sample sizes were not sufficient to establishthis as a definitive trend. Of the AHL+ bacteria, 67%were Alphaproteobacteria and 33% were Gammaproteo-bacteria. The AHL+ gammaproteobacterial isolatesincluded taxa closely related to species of Vibrio, Thalas-somonas and Spongiobacter. It is not surprising that
AHL+ Vibrio spp. were identified, but to our knowledge thisis the first report suggesting that bacteria related toThalassomonas and Spongiobacter produce AHL-typeactivities. Alphaproteobacterial isolates were predomi-nantly from the SR subclade of Roseobacter lineage (11strains). One Alphaproteobacterium from M. laxissimathat resulted in a weak response in the AHL reporterbacteria was from the genus Pseudovibrio (N05ML11).Most of the AHL+ bacteria were detected with the At-AHLreporter, the most sensitive and least specific AHL biode-tection system available, while a subset of these werepositive for long-chain AHL synthesis in the Sm-AHLassay. A single Alphaproteobacterium (N04ML5) wasfound to produce AHLs detected by the Cv-AHL system(Fig. 1A), specific for short-chain AHLs, and the least sen-sitive biodetection strain we used. Proteobacterial isolatesfrom the water column adjacent to the site of spongecollection were also characterized and eight isolates weretested. Two isolates identified from the genus Erythro-bacter (SWKLH7 and SWKLH8) proved to be AHL+. Twowater column strains SWKLH6 and SWKLH14 fell withinthe Roseobacter clade, based on 16S rRNA sequence,but both of these strains were negative for AHLproduction. Although far larger data sets are required totest the idea, our observations bear facile similarity toreports for plant-associated pseudomonads in which host-associated bacteria are disproportionately AHL+ relative tothose in bulk soil (Elasri et al., 2001).
Although all of the AHL+ sponge-associated isolateswithin the SR subclade were similar at the 16S rRNAsequence level, a surprising range of AHL profiles wasobserved following comparison of the strains with RP-TLC(Fig. 2A). Not only did the set of AHLs synthesized varyamong the strains, but also some of the isolates producedvastly different amounts of the signals. The AHL profilesobtained for the alphaproteobacterial isolates were highlycomplex, often exhibiting four to six or more species sepa-rated on the TLC plate. On the other hand, the profilesobtained for the signal-positive gammaproteobacterialisolates were similar and fairly simple with one to twoprimary species, generally in the medium chain lengthrange (Fig. 2B). It is important to recognize that this analy-sis will only detect AHLs synthesized under laboratoryconditions. There is ample evidence that AHL synthesiscan be tightly regulated by environmental conditionsincluding host-released cues and nutritional status. Ourfindings may therefore be an under-representation of thetrue prevalence of AHL synthesis.
In a small-scale survey of several marine sponges, anAlphaproteobacterium that produced AHL signal mol-ecules was isolated from the marine sponge C. concen-trica (Taylor et al., 2004). Strikingly, this single AHL+ isolatefrom an Australian sponge falls within the same SR groupas the dominant AHL+ microbes we have isolated from the
-1.5
-1
-0.5
0
0.5
1
1.5
-1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
Dimension 1
Dim
ensi
on 2
N04IS3N05IS9
KLH11
N04ML9N04ML4
N04ML5
S. pomeroyi
N04ML2
N04ML11
Fig. 6. Two-dimensional MDS plots showing the relativerelationships of the nine Roseobacter strains based on their proteinexpression patterns. Bacterial isolates are indicated next toeach point. Stress = 0.039.
Diversity and AHLs in sponge-associated Proteobacteria 81
Florida sponges M. laxissima and I. strobilina. Synthesis ofAHLs is apparently a prevalent characteristic of thebroader group of Alphaproteobacteria within the Roseo-bacter clade (Wagner-Dobler et al., 2005). As with therhizobia, terrestrial plant-associated Alphaproteobacteria,the production of long-chain AHLs is often observed in theSR group. We hypothesize that AHL-based signalling inthose SR bacteria associated with marine sponges con-trols functions that have significance to the sponge–microbe interaction and also the structure of thepolymicrobial community within the sponge.
Differentiation of closely related microbial species is notstraightforward. Although identical 16S rRNA gene sequ-ences were found among actinobacterial strains isolatedfrom different thermal niches in temperate, subtropical andtropical freshwater habitats, the isolates had sequencedifferences (Hahn and Pockl, 2005). Differences wereobserved at zero to five positions in a 2310-nucleotidefragment of the ribosomal operon, including part of theintergenic spacer upstream of the 16S rRNA gene, thecomplete 16S rRNA gene, the complete 16S-23S internaltranscribed spacer and a short part of the 23S rRNA gene.Thus, bacterial species characterization cannot solely relyon sequence analysis of a single gene marker. The ecologyof closely related bacterial species should be considerednot only at phylogenetic and genetic levels, but also atmetabolic and functional levels. Our results suggest thatspecies or strain-specific protein expression profilesprovide high resolution for closely related bacterial strains.The nine AHL-producing SR bacteria examined sharedgreater than 98% identity in terms of their 16S rRNA genesequences, and they would be considered as the samespecies (> 97%; Rossello-Mora and Amann, 2001). Somestrains (i.e. KLH11, N04ML9 and N04ML4) could not beresolved based on the 16S rRNA phylogeny, but metabolicor functional differences were observed based on the AHLprofiles or proteome patterns. Our results are congruentwith previous studies in which protein profiles could beused to resolve and identify closely related Ferroplasmaisolates (Dopson et al., 2004), and even for phylogeneticstudies (Navas and Albar, 2004). In our study, all thebacterial strains were cultivated and harvested under thesame growth conditions with identical culture media and,therefore, the protein profiles should reflect real differ-ences in protein expression between strains.
Conclusions
Isolates from the SR subclade within the Alphaproteobac-teria are the dominant cultivatable producers of AHLs inmarine sponges M. laxissima and I. strobilina. This is con-sistent with previous studies showing the production ofAHLs by members of the Roseobacter clade from themarine environment (Wagner-Dobler et al., 2005). All the
Roseobacter isolates included in the proteomic analysesare AHL producers. The 2DGE reference maps for AHLproducers can be generated as a reference standard forAHL-producing species recognition. Characteristic proteinspots shared by all the AHL producers might be used as‘AHL indicators’ to differentiate these strains from non-AHLproducers. More importantly, characterization of uniquesignature proteins for AHL-positive strains or species mayhold a key to understanding the mechanisms of the AHLproduction pathway. Quorum sensing-regulatory genesincluding AHL synthases have been isolated and directedmutant strains of AHL-producing SR bacteria have beengenerated (E.M. Cicirelli, J. Herman, N.M. Mohamed,M.E.A. Churchill, R.T. Hill and C. Fuqua, unpublished). Theproteomes of these mutants are under comparison withwild-type SR bacteria. Sponge colonization assays willalso be carried out using wild-type and mutant strains.These future studies will begin to provide an understandingof the functional roles of different proteins in AHL produc-tion and regulation in the sponge environment, phenotypictargets of this regulation, and symbiotic interactionsbetween sponges and their bacterial communities.
Experimental procedures
Sponge collection and taxonomic identification
Three individuals of each of M. laxissima and I. strobilina werecollected by SCUBA at Conch Reef, Key Largo, Florida in July2001, June 2004 and August 2005 at a depth of 40–50 ft.Water samples were collected from the vicinity of the collectedsponges, at a depth of 60 ft. Water salinity was 36 ppt andtemperature was 27°C.
Sponge processing for isolation of culturable bacteria
Immediately after collection of sponges, each sample wasrinsed with sterile artificial seawater (ASW) to remove anytransiently associated bacteria. Sponge tissue (1 cm3) wasground in ASW and serial dilutions of sponge homogenatewere plated on Difco Marine Agar 2216 (BD Biosciences,Franklin Lakes, NJ, USA). Plates were incubated at 30°C for aweek. Water samples were processed similarly for bacterialisolation.
Identification of bacterial isolates by 16S rRNA genesequence analysis
Single pure colonies of each isolate were transferred to 20 mlof Marine Broth 2216 (BD Biosciences) and incubated over-night at 30°C in a shaking incubator. DNA was extracted fromisolates using the Ultra-Clean microbial kit (MoBio Laborato-ries, Carlsbad, CA, USA). Isolates were stored at -80°C inmarine broth supplemented with 30% glycerol. 16S rRNAgene was PCR-amplified using universal primers 27F and1492R (Lane, 1991) as described by Montalvo and colleagues(2005). Amplicons were extracted from 1% (w/v) agarose gelswith a QIAquick gel extraction kit (Qiagen, Valencia, CA, USA)and sequenced with 27F and 1492R primers.
Sequences were visualized and annotated usingSequencherTM version 4.2.2 (Gene Codes, Ann Arbor, MI,USA). Forward and reverse sequences were assembledusing AssemblyLIGN sequence assembly software (EastmanKodak, Rochester, NY, USA) into a single consensussequence. 16S rRNA gene sequences from isolates wereanalysed using the BLASTN tool at the National Center ofBiotechnology Information and then imported to ARB soft-ware package (Ludwig et al., 2004) to align homologousregions of 16S rRNA gene sequences from different isolates.The sequences were aligned using the PT-server with a dataset containing the nearest relative matches. Multiple align-ments were checked manually and improved by the ARBeditor tool. Trees were constructed using the neighbour-joining (Jukes–Cantor correction) (Saitou and Nei, 1987),Fitch–Margoliash (Fitch and Margoliash, 1967) andmaximum parsimony (Kluge and Farris, 1969) algorithmsimplemented in ARB. The robustness of the inferred treestopologies was evaluated after 1000 bootstrap replicates ofthe neighbour-joining data. Bootstrap values were generatedusing Phylip (Felsenstein, 2004). Identification of isolateswas based on the closest cultured taxon in the top BLAST hits.
Bioassays for AHL synthesis
Isolates were tested for the production of AHLs with a seriesof bioreporter systems, each with different optimal sensitivi-ties to short-chain, moderate-chain and long-chain AHLs(Table 1). The At-AHL reporter expresses a lacZ fusion moststrongly in response to medium-chain-length AHLs (C6–C12,although weakly to C4), with limited distinction of AHLs car-rying a hydrogen, a hydroxyl or a carbonyl as the R-group atthe b carbon. The Cv-AHL derivative produces the purplepigment violacein in response to fully reduced short-chainAHLs (C4–C6 side-chains) with a hydrogen as the R-group atthe b carbon. The Sm-AHL reporter also expresses a lacZfusion in response to long-chain AHLs (C14–C18) with orwithout unsaturated bonds in the side-chains, and with hydro-gens and carbonyls as the R-group (hydroxyls have not beentested) (McClean et al., 1997; Zhu et al., 2003; Llamas et al.,2004). Preliminary screening protocols involved cultivation ofsponge isolates on Marine Agar and incubation at 28°C forup to 7 days until a healthy colony size or patch of growthwas obtained. In most cases standard 10 cm Petri disheswere used although the assay was occasionally scaled up to15 cm plates. An overlay of soft agar was either inoculateddirectly with the AHL reporter strain (At-AHL and Sm-AHL),or the strain (Cv-AHL) was gently spread on the overlaysurface. Cleavage of X-Gal (5-bromo-4-chloro-3-indolyl-galactopyranoside) (Promega, Madison, WI, USA) and bluestaining (At-AHL and Sm-AHL) or purple pigmentation(Cv-AHL) within the agar overlay in a diffusible ring around acolony was measured and scored as AHL production.
To prepare cells for the bioassays, the At-AHL reporterstrains were grown to mid-exponential phase [optical densityat 600 nm (OD600) = 0.6]. Cells were harvested by centrifuga-tion and the pellet was thoroughly washed and re-suspendedwith sterile deionized distilled water, followed by a final cen-trifugation and re-suspension in 30% glycerol to a final OD600
of 12. For the overlay assays, 100 ml of AT minimal medium
(Tempé et al., 1977) soft agar (0.6%) with glucose as acarbon source (0.5%) and ammonium sulfate (15 mM) as anitrogen source, and 40 mg ml-1 X-Gal, was inoculated with1 ml of the concentrated At-AHL reporters. Marine Agar 2216plates with colonies were overlaid with 20–25 ml of the ATGNagar and incubated at 28°C overnight.
Sm-AHL cells were grown in LB (supplemented with2.5 mM CaCl2 and 2.5 mM MgSO4), and the cells were pre-pared in a similar manner to the At-AHL strains as describedpreviously (Llamas et al., 2004). For overlays, 0.6% TY agarwas supplemented with 2.5 mM CaCl2, 2.5 mM MgSO4,40 mg ml-1 X-gal and 1 ml of concentrated Sm-AHL cells.Marine agar plates were overlaid with 25 ml of the inoculatedTY soft agar and incubated at 28°C overnight.
Due to salt intolerance the Cv-AHL assays were performedslightly differently. Marine Agar 2216 plates with fully growncolonies were overlaid with 25 ml of 0.6% TY (Tryptone YeastExtract) agar and this was allowed to solidify. Following thisstep, overnight cultures of CV026 grown in Luria–Bertani (LB)broth were centrifuged and concentrated sixfold in fresh LBfollowed by spread plating of approximately 100 ml of thesuspension per plate. Plates were incubated at 28°Covernight.
TLC profiling of AHLs
Five millilitres of MB 2216 cultures were grown to an OD600 of1.5–2.0, followed by extraction with an equal volume ofdichloromethane. Culture pH was monitored and was withinthe range of 7.6 � 0.2 at the time of harvesting (sterile MarineBroth 2216 is pH 7.6 � 0.2). Following centrifugation, theorganic phase was removed and allowed to evaporate in afume hood. Extracts were concentrated 1000-fold and nor-malized to an OD600 of 1.5 and re-suspended in a final volumeof approximately 5 ml of acidified (0.01%) ethyl acetate andloaded onto a C18 RP-TLC plate (Mallinckrodt Baker, Phil-lipsburg, NJ, USA). Thin-layer chromatography plates weredeveloped in a 60% methanol water mobile phase, dried, andoverlaid with 100 ml of 0.6% ATGN media supplemented with40 mg ml-1 X-gal and 1 ml of an OD600 = 12.0 suspension ofthe highly sensitive At-AHL reporter. Thin-layer chromatogra-phy plate overlays were placed in a sealed container andincubated at 28°C for 16–24 h.
Protein extraction
Nine isolates belonging to the SR subgroup were selectedfor proteomic analysis. They included KLH11, N04ML2,N04ML4, N04ML5, N04ML6, N04ML9 and N04ML11 fromM. laxissima and N04IS3 and N05IS9 from I. strobilina. Thegrowth curve of each isolate was monitored by absorbance(OD600) and samples (1.8 ml) were harvested at mid-exponential phase for proteomic analysis. Cells were centri-fuged at 10 600 g for 3 min and rinsed with washing buffer(Tris-Cl 10 mM, Sucrose 250 mM, pH 7.6). Pellets werestored at -80°C until further processing. To extract proteins,frozen cells were thawed on ice and re-suspended with0.5 ml of extraction buffer, consisting of 0.01 M Tris-HCl,pH 7.4, 1 mM EDTA, 7 M urea and 2 M thiourea, 10%(v/v) glycerol, 2% CHAPS, 0.2% amphylotes, 0.002 M
Diversity and AHLs in sponge-associated Proteobacteria 83
Tributyl phosphine (TBP), DNase (0.1 mg ml-1), RNase(0.025 mg ml-1) and proteinase inhibitor cocktail (CalBio-chem, San Diego, CA, USA). TBP, DNase, RNase and pro-teinase inhibitor cocktail were freshly prepared and added tothe extraction buffer prior to the experiment. Samples wereincubated on ice and vortexed for 15 s every 5 min. Cellulardebris was removed by centrifugation (10 600 g, 4°C for5 min). Protein concentration was estimated by measuringthe absorbance at 280 nm using a spectrophotometer(Bio-Rad, Hercules, CA, USA).
Isoelectric Focusing and SDS-PAGE
The first-dimension separation of proteins (approximately200 mg each sample) was conducted in the immobilized pHgradient strips (11 cm, pH 4–7) on a Bio-Rad Protean Iso-electric Focusing (IEF) Cell system (Bio-Rad, Hercules, CA,USA). The IEF program was run at 250 V for 20 min followedwith a linear ramp to 8000 V for 2.5 h, and 8000 V for a total40 000 V h-1 with a rapid ramp. After the first dimension, theIEF strips were equilibrated in freshly made Buffer 1 (6 Murea, 2% SDS, 0.05 M Tris/HCl pH 8.8, 50% Glycerol) andBuffer 2 (6 M urea, 2% SDS, 0.375 M Tris/HCl pH 8.8, 20%Glycerol and 0.5 g of iodoacetamide) (Bio-Rad) respectively.The second dimension of 2D-PAGE were performed using8–16% gradient pre-cast polyacrylamide gels (Bio-Rad) fol-lowing the manufacturer’s instructions. Gels were stainedwith SYPRO Ruby (Bio-Rad) after electrophoresis andscanned using a Typhoon 9410 gel imager (GE HealthcareBio-Sciences, Piscataway, NJ, USA) with 488 nm excitationand emission filter 610 BP30.
Image analyses
Two-dimensional gel electrophoresis images were analysedusing the ImageMaster software (GE Healthcare Bio-Sciences) following the manufacturer’s instructions. The gelimages were imported into the workspace designated foranalysis. As the gel images acquired with the Typhoon 9410were already calibrated, no intensity calibration wasnecessary. Spot detection parameters including smooth,saliency and minimum area were adjusted and optimized inorder to detect all the real spots and filter out the noise. Gelimages were compared using the total density in gel methodfor spot quantification. A single landmark was annotated andused for image automatic matching procedure. Matches wereedited manually to eliminate poor quality matches. Pairwisesimilarities were obtained based on the percentage of spotsthat were matched, which was calculated as follows: per centmatches = 2m/(ng + nm), where m was the total number ofmatches between the gel and the Master image (referencegel), and ng is the number of spots in the gel and nm is thenumber of spots in the Master. The pairwise similarities werefurther used to construct the distance matrix for MDSanalysis.
Non-metric MDS analysis
The MDS analysis was performed using the SAS System(SAS Institute, Cary, NC, USA). The relationship between any
two isolates is reflected by the relative distance betweenthem in two-dimension MDS plots. Bacterial isolates withmore similar proteome patterns are plotted closer, while iso-lates with less similar proteome patterns are further apart.The stress value (measure of goodness-of-fit) was recorded.In general, stress value less than 0.1 indicated a good ordi-nation with little risk of misinterpretation of data (Clarke,1993).
Characterization of AHL-producing roseobacters
Isolates N04ML2, N04ML4, N04ML5, N04ML6, N04ML9and N04ML11 were characterized using API 20NE strips(bioMérieux, Paris, France).
Nucleotide sequence accession numbers
16S rRNA gene sequence fragments from AHL-producingisolates were submitted in the GenBank and accessionnumbers are given in Figs 3 and 4.
Acknowledgements
We thank Matthew Anderson, Julie Enticknap, Jayme Lohr,Naomi Montalvo and Olivier Peraud for their help in spongecollection and processing. We acknowledge Xuesong He forthe initial AHL screening and Karen Tait for guidance inseveral aspects of this work. Silicibacter pomeroyi DSS-3was a gift from Mary Ann Moran. Funding for this researchwas provided by the Microbial Observatories Program,National Science Foundation (MCB-0238515).
References
Bruns, A., Cypionka, H., and Overmann, J. (2002) CyclicAMP and acyl homoserine lactones increase the cultivationefficiency of heterotrophic bacteria from the central BalticSea. Appl Environ Microbiol 68: 3978–3987.
Clarke, K.R. (1993) Nonparametric multivariate analysesof changes in community structure. Aust J Ecol 18: 117–143.
Corredor, J.E., Wilkinson, C.R., Vicente, V.P., Morell, J.M.,and Otero, E. (1988) Nitrate release by Caribbean reefsponges. Limnol Oceanogr 33: 114–120.
Diaz, M.C., and Ward, B.B. (1997) Sponge-mediated nitrifi-caton in tropical benthic communities. Mar Ecol Prog Ser156: 97–109.
Dopson, M., Baker-Austin, C., Hind, A., Bowman, J.P., andBond, P.L. (2004) Characterization of Ferroplasma isolatesand Ferroplasma acidarmanus sp. nov., extreme acido-philes from acid mine drainage and industrial bioleachingenvironments. Appl Environ Microbiol 70: 2079–2088.
Eberhard, A., Burlingame, A.L., Eberhard, C., Kenyon, G.L.,Nealson, K.H., and Oppenheimer, N.J. (1981) Structuralidentification of autoinducer of Photobacterium fischeriluciferase. Biochemistry 20: 2444–2453.
Elasri, M., Delorme, S., Lemanceau, P., Stewart, G., Laue,B., Glickmann, E., et al. (2001) Acyl-homoserine lactoneproduction is more common among plant-associated
Pseudomonas spp. than among soilborne Pseudomonasspp. Appl Environ Microbiol 67: 1198–1209.
Enticknap, J.J., Kelly Shanks, M., Peraud, O., and Hill, R.T.(2005) Characterization of a culturable Alphaproteobacte-rial symbiont common to many marine sponges and evi-dence for vertical transmission via sponge larvae. ApplEnviron Microbiol 72: 4105–4119.
Felsenstein, J. (2004) PHYLIP (Phylogenetic InferencePackage), Version 3.6. Seattle, WA, USA: Department ofGenetics, University of Washington.
Fitch, W.M., and Margoliash, E. (1967) Construction of phy-logenetic trees: a method based on mutation distances asestimated from cytochrome c sequences is of generalapplicability. Science 155: 279–284.
Fuqua, C., and Greenberg, E.P. (2002) Listening in on bac-teria: acyl-homoserine lactone signaling. Nat Rev Mol CellBiol 3: 685–695.
Fuqua, C., Parsek, M.R., and Greenberg, E.P. (2001) Regu-lation of gene expression by cell-to-cell communication:acyl- homoserine lactone quorum sensing. Annu RevGenet 35: 439–468.
Gram, L., Grossart, H.P., Schlingloff, A., and Kiørboe, T.(2002) Possible quorum sensing in marine snow bacteria:production of acylated homoserine lactones by Roseo-bacter strains isolated from marine snow. Appl EnvironMicrobiol 68: 4111–4116.
Hahn, M.W., and Pockl, M. (2005) Ecotypes of planktonicActinobacteria with identical 16S rRNA genes adapted tothermal niches in temperate, subtropical, and tropicalfreshwater habitats. Appl Environ Microbiol 71: 766–773.
Hentschel, U., Usher, K.M., and Taylor, M.W. (2006) Marinesponges as microbial fermenters. FEMS Microbiol Ecol 55:167–177.
Hoffmann, F., Larsen, O., Thiel, V., Rapp, H.T., and Reitner,J. (2005) An anaerobic world in sponges. Geomicrobiol J22: 1–10.
Joint, I., Tait, K., Callow, M.E., Callow, J.A., Milton, D.L.,Williams, P., and Camara, M. (2002) Cell-to-cell communi-cation across the prokaryote–eukaryote boundary. Science298: 1207–1207.
Kluge, A.G., and Farris, F.S. (1969) Quantitative phyleticsand the evolution of annurans. Syst Zool 18: 1–32.
Lane, D.J. (1991) 16S/23S rRNA sequencing. In Nucleic AcidTechniques in Bacterial Systematics. Stackebrandt, E., andGoodfellow, M. (eds). Chichester, UK: John Wiley & Sons,pp. 115–175.
Llamas, I., Keshavan, N., and Gonzalez, J.E. (2004) Use ofSinorhizobium meliloti as an indicator for specific detectionof long-chain N-acyl homoserine lactones. Appl EnvironMicrobiol 70: 3715–3723.
Ludwig, W., Strunk, O., Westram, R., Richter, L., Meier, H.,Yadhukumar, et al. (2004) ARB: a software environment forsequence data. Nucleic Acids Res 32: 1363–1371.
McClean, K.H., Winson, M.K., Fish, L., Taylor, A., Chhabra,S.R., Camara, M., et al. (1997) Quorum sensing and Chro-mobacterium violaceum: exploitation of violacein produc-tion and inhibition for the detection of N-acylhomoserinelactones. Microbiology 143: 3703–3711.
Maldonado, M., Cortadellas, N., Trillas, M.I., and Rutzler, K.(2005) Endosymbiotic yeast maternally transmitted in amarine sponge. Biol Bull 209: 94–106.
Montalvo, N.F., Mohamed, N.M., Enticknap, J.J., and Hill,R.T. (2005) Novel actinobacteria from marine sponges.Antonie Van Leeuwenhoek 87: 29–36.
Navas, A., and Albar, J.P. (2004) Application of proteomics inphylogenetic and evolutionary studies. Proteomics 4: 299–302.
Nealson, K.H., Platt, T., and Hastings, J.W. (1970) Cellularcontrol of the synthesis and activity of the bacterial lumi-nescent system. J Bacteriol 104: 313–335.
Oldroyd, G.E.D., Harrison, M.J., and Udvardi, M. (2005)Peace talks and trade deals. Keys to long-term harmony inlegume-microbe symbioses. Plant Physiol 137: 1205–1210.
Oren, M., Steindler, L., and Ilan, M. (2005) Transmission,plasticity and the molecular identification of cyanobacterialsymbionts in the Red Sea sponge Diacarnus erythraenus.Mar Biol 148: 35–41.
Rossello-Mora, R., and Amann, R. (2001) The speciesconcept for prokaryotes. FEMS Microbiol Rev 25: 39–67.
Ruby, E.G. (1996) Lessons from a cooperative, bacterial–animal association: The Vibrio fischeri-Euprymna scolopeslight organ symbiosis. Annu Rev Microbiol 50: 591–624.
Saitou, N., and Nei, M. (1987) The neighbor-joining method:a new method for reconstructing phylogenetic trees.Mol Biol Evol 4: 406–425.
Sharp, K.H., Eam, B., Faulkner, D.J., and Haygood, M.G.(2007) Vertical transmission of diverse microbes in thetropical sponge Corticium sp. Appl Environ Microbiol 73:622–629.
Tait, K., Joint, I., Daykin, M., Milton, D.L., Williams, P., andCamara, M. (2005) Disruption of quorum sensing inseawater abolishes attraction of zoospores of the greenalga Ulva to bacterial biofilms. Environ Microbiol 7: 229–240.
Taylor, M.W., Schupp, P.J., Baillie, H.J., Charlton, T.S., deNys, R., Kjelleberg, S., and Steinberg, P.D. (2004) Evi-dence for acyl homoserine lactone signal production inbacteria associated with marine sponges. Appl EnvironMicrobiol 70: 4387–4389.
Tempé, J., Petit, A., Holsters, M., Montagu, M.V., and Schell,J. (1977) Thermosensitive step associated with transfer ofthe Ti plasmid during conjugation: possible relation totransformation in crown gall. Proc Natl Acad Sci U S A 74:2848–2849.
Usher, K.M., Sutton, D.C., Toze, S., Kuo, J., and Fromont, J.(2005) Inter-generational transmission of microbial sym-bionts in the marine sponge Chondrilla australiensis(Demospongiae). Mar Freshwater Res 56: 125–131.
Vacelet, J. (1975) Etude en microscopie electronique del’association entre bacteries et spongiaires du genreVerongia (Dictyoceratida). J Microscopie de BiologieCellulaire 23: 271–288.
Vacelet, J., and Donadey, C. (1977) Electron microscopestudy of the association between some sponges andbacteria. J Exp Mar Biol Ecol 30: 301–314.
Van Soest, R. (1996) Porifera, Schwämme. In SpezielleZoologie, Teil 1: Einzeller and Wirbellose Tiere. Westheide,
Diversity and AHLs in sponge-associated Proteobacteria 85
W., and Rieger, R. (eds). Stuttgart, Germany: GustavFischer Verlag, pp. 98–119.
Visick, K.L., Foster, J., Doino, J., McFall-Ngai, M., and Ruby,E.G. (2000) Vibrio fischeri lux genes play an important rolein colonization and development of the host light organ.J Bacteriol 182: 4578–4586.
Wagner-Dobler, I., Thiel, V., Eberl, L., Allgaier, M., Bodor, A.,Meyer, S., et al. (2005) Discovery of complex mixtures ofnovel long-chain quorum sensing signals in free-living andhost-associated marine alphaproteobacteria. Chembio-chem 6: 2195–2206.
Waters, C.M., and Bassler, B.L. (2005) Quorum sensing:cell-to-cell communication in bacteria. Annu Rev Cell DevBiol 21: 319–346.
Webster, N.S., Wilson, K.J., Blackall, L.L., and Hill, R.T.(2001) Phylogenetic diversity of bacteria associated withthe marine sponge Rhopaloeides odorabile. Appl EnvironMicrobiol 67: 434–444.
Wilkinson, C.R. (1978) Microbial associations in sponges. II.Numerical analysis of sponge and water bacterialpopulations. Mar Biol 49: 169–176.
Wilkinson, C.R. (1983) Net primary productivity in coral reefsponges. Science 219: 410–412.
Zhang, L., Murphy, P.J., Kerr, A., and Tate, M.E. (1993)Agrobacterium conjugation and gene regulation by N-acyl-L-homoserine lactones. Nature 362: 446–448.
Zhu, J., Chai, Y., Zhong, Z., Li, S., and Winans, S.C. (2003)Agrobacterium bioassay strain for ultrasensitive detectionof N-acylhomoserine lactone-type quorum-sensingmolecules: detection of autoinducers in Mesorhizobiumhuakuii. Appl Environ Microbiol 69: 6949–6953.
Supplementary material
The following supplementary material is available for thisarticle online:Fig. S1. Two-dimensional gel electrophoresis protein pat-terns of soluble proteins from AHL-producing SR roseo-bacters isolated from M. laxissima and I. strobilina. (A)KLH11, (B) N04ML9, (C) N04ML4, (D) N04IS3, (E) N05IS9,(F) N04ML5, (G) N04ML2, (H) N04ML11 and (I) S. pomeroyiDSS-3.
This material is available as part of the online article fromhttp://www.blackwell-synergy.com
