Widespread Distribution of Poribacteria in Demospongiae

Poribacteria were found in nine sponge species belonging to six orders of Porifera from three oceans. Phylogenetic analysis revealed four distinct poribacterial clades, which contained organisms obtained from several different geographic regions, indicating that the distribution of poribacteria is cosmopolitan. Members of divergent poribacterial clades were also found in the same sponge species in three different sponge genera.Recently, a novel bacterial phylum, termed “Poribacteria,” was discovered, and members of this phylum have been found exclusively in sponges (2). Phylogenetic analyses of 16S rRNA genes indicated that poribacteria are evolutionarily deeply branching organisms and related to a superphylum composed of Planctomycetes, Verrucomicrobia, and Chlamydia (11). Poribacterial 16S rRNA genes contain 13 of 15 planctomycete signature nucleotides, but a level of sequence divergence of more than 25% compared to any other bacterial phylum, including the Planctomycetes, justifies the status of this taxon as an independent phylum. A consistent treeing pattern is difficult to resolve in comparative phylogenetic sequence analyses, making the poribacteria an unusual line of phylogenetic descent. In addition to their divergent status as a separate phylum on the basis of the 16S rRNA sequence, poribacteria are also divergent because they may have a compartmentalized cell structure, a cell plan they share only with members of the phyla Planctomycetes and Verrucomicrobia (2). They are also of interest for understanding the potential contribution of obligate sponge-associated bacteria to the sponges harboring them and as an example of a yet-to-be-cultured group of bacteria associated with invertebrate tissue apparently exclusively but for unknown reasons. This study aimed to further explore the presence and diversity of poribacteria in different marine demosponge genera using samples from around the world.The Mediterranean sponges were collected by scuba divers offshore at Banyuls sur Mer, France (42°29′N, 03°08′E). The Caribbean sponges were collected offshore at Little San Salvador Island, Bahamas (24°32′N, 75°55′W). The eastern Pacific sponge Aplysina fistularis was collected offshore at San Diego, CA (32°51′N, 117°15′W). The western Pacific sponge Theonella swinhoei was collected offshore at Palau (07°23′N, 134°38′E). All non-Great Barrier Reef (non-GBR) sponges were collected between May and July 2000, and once individual sponge specimens were brought to the surface, they were frozen in liquid nitrogen on board ship and stored at −80°C until microbiological processing (9). The GBR marine sponges were collected off Heron Island Research Station (23°27′S, 151°5′E) in April 2002 (5). Pseudoceratina clavata was collected by scuba divers at a depth of 14 m, and Rhabdastrella globostellata was collected at a depth of ca. 0.5 m after a reef walk consisting of a few hundred meters. The samples were immediately placed in plastic bags and brought to Heron Island Research Station, where they were stored at −80°C until processing. Sponge DNA was extracted as described previously (2, 5).Total sponge-derived genomic DNA was screened by PCR for the presence of poribacteria using a 16S rRNA gene primer set. Poribacterial 16S rRNA genes were amplified by employing a pair of Poribacteria-specific primers, POR389f (5′-ACG ATG CGA CGC CGC GTG-3′) and POR1130r (5′-GGC TCG TCA CCA GCG GTC-3′) (2). The poribacterial PCR products that were ca. 740 bp long derived from one sponge individual were cloned into the pGEM-T Easy vector (Promega, Madison, WI). Clone inserts were digested with restriction endonucleases MspI and HaeIII (New England Biolabs, Inc., United States), characterized to obtain restriction profiles and unique profiles, and sequenced. The compiled partial 16S rRNA gene sequences were then analyzed using BLASTN to select the most closely related poribacterial reference sequences.The sequences exhibiting levels of similarity of less than 97% were used for further analysis. Poribacterial 16S rRNA gene sequences were aligned using the ARB software package (7). The resulting alignment was imported into PAUP (10) and analyzed by using distance, maximum parsimony, and maximum likelihood algorithms together with bootstrap resamplings (3,000, 3,000, and 200 resamplings, respectively), and the resulting bootstrap values were applied to nodes on the ARB neighbor-joining tree. Signature sequences were detected using the ARB software package. A signature sequence is defined here as a short sequence that is present in a group of poribacterial sequences in a phylogenetic clade but is not found in any other clade in the poribacterial tree.Analysis of the 16S rRNA gene clone library sequences generated from sponge tissues revealed the presence of poribacteria in sponge individuals belonging to the orders Verongida, Astrophorida, Dictyoceratida, Haplosclerida, Lithistida, and Homosclerophorida, while poribacteria could not be detected in sponges belonging to the orders Hadromerida and Agelasida. In the order Halichondrida, poribacteria were detected in Xestospongia muta but not in Haliclona sp. Altogether, nine sponge species were added to the list of Poribacteria-containing sponges (Table ​(Table1).1). Three distinct clades were observed that were clearly supported by bootstrap values greater than 75 with every tree-building algorithm applied (Fig. ​(Fig.1),1), and one clade (clade I) was supported by bootstrap values of 64, 98, and 71 in distance, maximum parsimony, and maximum likelihood trees, respectively. Similarity calculations using approximately 740-bp amplified poribacterial 16S rRNA gene fragments and other poribacterial sequences from the NCBI database showed that the dissimilarity between clades was consistent with their separation in phylogenetic trees. For example, the levels of dissimilarity between members of clade I and clade II were 3 to 8%, while the levels of dissimilarity between members of clades I and III and between members of clades I and IV were 10 to 14% and 11 to 15%, respectively.Open in a separate windowFIG. 1.Neighbor-joining phylogenetic tree for poribacterial clones based on Poribacteria-specific PCR products (740 bp) of the 16S rRNA gene, showing relationships of poribacterial clones from different global regions. The poribacterial clones on the right are additional clones belonging to the same clades as strains in the tree at the same level. Bootstrap confidence values of >75% for distance, maximum parsimony, and maximum likelihood algorithm analyses are indicated by filled circles at nodes, and open circles indicate unsupported nodes. Prefixes for clones: A, Aplysina aerophoba; C, Aplysina cavernicola; F, Aplysina fistularis; L, Aplysina lacunosa; S, Ircinia sp.; P, Plakortis sp.; PC, Pseudoceratina clavata; RG, Rhabdastrella globostellata; T, Theonella swinhoei; X, Xestospongia muta. Scale bar = 0.1 nucleotide substitution per site.

TABLE 1.

Distribution of poribacteria in different demosponge orders

Sponge species or seawater	Order	Geographic locationa	Presence of poribacteriab	Reference
Aplysina aerophoba	Verongida	MED	+	2
Aplysina lacunosa	Verongida	BAH	+	2
Aplysina fistularis	Verongida	EPAC or BAH	+	2
Aplysina insularis	Verongida	BAH	+	2
Verongula gigantea	Verongida	BAH	+	2
Smenospongia aurea	Dictyoceratida	BAH	+	2
Aplysina cauliformis	Verongida	BAH	+	This study
Aplysina archeri	Verongida	BAH	+	This study
Aplysina cavernicola	Verongida	MED	+	This study
Pseudoceratina clavata	Verongida	WPAC	+	This study
Rhabdastrella globostellata	Astrophorida	WPAC	+	This study
Ircinia sp.	Dictyoceratida	BAH	+	This study
Xestospongia muta	Haplosclerida	BAH	+	This study
Theonella swinhoei	Lithistida	EPAC	+	This study
Plakortis sp.	Homosclerophorida	BAH	+	This study
Chondrilla nucula	Hadromerida	BAH	−	2
Agelas wiedenmayeri	Agelasida	BAH	−	2
Agelas cerebrum	Agelasida	BAH	−	This study
Axinella polypoides	Halichondrida	MED	−	This study
Ptilocaulis sp.	Halichondrida	BAH	−	2
Dysidea avara	Dictyoceratida	MED	−	This study
Haliclona sp.	Haplosclerida	MED	−	This study
Ectyoplasia ferox	Poecilosclerida	BAH	−	2
Seawater	NAc	MED	−	This study

Open in a separate window^aMED, Mediterranean Sea; BAH, Bahamas; WPAC, western Pacific Ocean; EPAC, eastern Pacific Ocean.^bThe presence of poribacteria was evaluated by sequencing and phylogenetic analysis of amplified PCR products. +, present; −, absent.^cNA, not applicable.Within each clade in the phylum Poribacteria, there were higher similarity values, including 94 to 100% among members of clade I, 94 to 99% among members of clade II, 96 to 99% among members of clade III, and 96 to 99% among members of clade IV. When members of the the phylum Poribacteria were compared to members of the Planctomycetes (Fig. ​(Fig.1),1), the 16S rRNA genes exhibited levels of sequence dissimilarity of up to 38%, consistent with the conclusion of Fieseler et al. concerning the separate phylum level status of poribacteria based on a similarity value of <75%. A phylogenetic correlation between sponge phylogeny and poribacterial phylogeny is not evident, since, for example, clones from A. fistularis and Aplysina aerophoba occurred in both clade I and clade II and one clone from A. aerophoba also occurred in clade III, while clones from P. clavata and R. globostellata occurred in clades I, II, and III but not in clade IV. Clades I and II included poribacterial clones derived from all sponge species occurring in all of the widely separated geographic regions examined in this study (Fig. ​(Fig.2).2). Clade III represented poribacterial clones derived from sponge species obtained in the eastern Pacific region, GBR, and the Bahamas but not in the Mediterranean region. The majority of poribacterial clones in clade IV were derived from sponge species obtained in the Bahamas, and one clade IV clone was obtained from a sponge species collected in the Mediterranean region.Open in a separate windowFIG. 2.Neighbor-joining phylogenetic tree for poribacterial clones based on Poribacteria-specific PCR products (740 bp) of the 16S rRNA gene, showing the internal relationships of and occurrence of clade I members in distinct sponge species representing cosmopolitan geographic regions. For an explanation of the colors, see Fig. ​Fig.1.1. Bootstrap confidence values of >75% for distance, maximum parsimony, and maximum likelihood algorithm analyses are indicated by filled circles at nodes, and open circles indicate unsupported nodes. Prefixes for clones: A, Aplysina aerophoba; C, Aplysina cavernicola; F, Aplysina fistularis; L, Aplysina lacunosa; S, Ircinia sp.; P, Plakortis sp.; PC, Pseudoceratina clavata; RG, Rhabdastrella globostellata; T, Theonella swinhoei; X, Xestospongia muta. Scale bar = 0.1 nucleotide substitution per site. Clones PC15, L8, T6, C2, P3, S2, and X1 were removed from this analysis to allow better branch resolution.Poribacterial clones from different sponges from widely separated marine habitats belonged to at least four major clades with similarities ranging from 94 and 96%. For clade III (Fig. ​(Fig.1),1), we detected a signature sequence characteristic of poribacterial clones from the GBR sponges R. globostellata and P. clavata. This signature sequence (CCA GTT AGC TTG ACG GTA) (Table ​(Table2)2) at E. coli positions 469 to 487 targeted 10 sequences, 5 of which were from GBR marine sponges generated in this study (clones RG68, RG112, RG105, PC96, and PC8). Another five poribacterial sequences were detected in an unpublished study investigating the microbial diversity in GBR sponges. This signature sequence indicates a specific geographic presence of poribacteria belonging to clade IV in the GBR region. In addition, a sequence (GAG TGT GAA ATG GCT TGG at E. coli positions 599 to 617) characteristic of clade IV was found in 11 sequences derived from sponges from the Bahamas and one sequence (A7) from a Mediterranean sponge.

TABLE 2.

Poribacterial signature sequences for clades III and IV, including a GBR-specific signature sequence (pori_SSIII) and a signature sequence specific to 11 of 12 sequences from the Bahamas (pori_SSIV)

Signature sequence	Name	Full namea	E. coli position	Sequenceb
pori_SSIII	Pla101P	Pla101P*	469	GGUGAUAAG-==================-CCAUAGUA
	Pla131P	Pla131P*	469	GGUGAUAAG-==================-CCAUAGUA
	Pla134P	Pla134P*	469	GGUGAUAAG-==================-CCAUAGUA
	Pla50P	Pla50P*	469	GGUGAUAAG-==================-CCAUAGUA
	Pla82P	Pla82P*	469	GGUGAUAAG-==================-CCAUAGUA
	PO68	Pori clone RGPo68	469	GGUGAUAAG-==================-GAGAAAAG
	PO112	Pori clone RGPo112	469	GGUGAUAAU-==================-CCAUAGUA
	PO105	Pori clone RGPo105	469	GGUGAUAAG-==================-CCAUAGUA
	PO96	Uncultured Pori clone	469	GGUGAUAAG-==================-CCAUAGUA
	PCPO8	Pori clone PCPo8	469	GGUGAUAAG-==================-CCAUAGUA
pori_SSIV	AY485286	Uncultured Pori clone	599	ACAUUAGUC-==================-CUCAACCA
	AY485285	Uncultured Pori clone	599	ACAUNAGUC-==================-CUCAACNA
	AY485284	Uncultured Pori clone	599	ACAUUAGUC-==================-CUCAACCA
	AY485281	Uncultured Pori bacterium	599	ACAUUAGUC-==================-CUCAACCA
	A7	A7	599	AUAUUAGUC-==================-CUCAACCA
	F2	F2	599	ACAUAAGUC-==================-CUCAACCA
	L16	L16	599	ACAUUAGUC-==================-CUCAACCA
	P20	P20	599	ACAUAAGUC-==================-CUCAACCA
	P38	P38	599	ACAUUAGUC-==================-CUCAACCA
	S6	S6	599	AUAUUAGUC-==================-CUCAACCA
	S10	S10	599	AUAUUAGUC-==================-CUCAACCA
	X18	X18	599	ACAUUAGUC-==================-CUCAACCA

Open in a separate window^aAsterisks indicate poribacterial clones derived from the GBR sponge R. globostellata in a separate study.^bThe internal sequence (indicated by equals signs) of each pori_SSIII clone is CCAGUUAGCUUGACGGUA, and that of each pori_SSIV clone is GAGUGUGAAAUGGCUUGG.Based on the data presented here, Poribacteria appears to be a bacterial phylum that is specifically found in several demosponge genera of the phylum Porifera (Table ​(Table1).1). To our knowledge, this is the only case of a bacterial phylum specifically associated with a marine invertebrate phylum. Certain phylum members appear to be widely distributed among sponges belonging to different species and in different geographic regions, forming sponge-specific lineages (3), but these are individual species level or at most genus level clades in a subdivision of a phylum rather than in a whole phylum.PCR analyses of seawater samples collected in this study (Table ​(Table1)1) and searches using nucleotide sequence databases of seawater metagenomes were negative for poribacteria. This is consistent with the concept that Poribacteria is a sponge-specific phylum. Within the sponges poribacteria are distributed among members of distinct demosponge orders that occur in various geographic locations, indicating that there is wide distribution of poribacteria among marine demosponges. Very similar 16S rRNA clone sequences that cluster in clade I were found in sponges from all geographic regions sampled in this study, including locations in the Northern and Southern hemispheres (Fig. ​(Fig.2).2). Similarly, clade II contains poribacterial clones from the Mediterranean Aplysina species and from GBR Pseudoceratina and Rhabdastrella species. This appears to contrast, albeit at a lower level of resolution, with results suggesting that bacterial populations are endemic in different geographic regions, e.g., with the findings that marine bacterioplankton communities include few cosmopolitan operational taxonomic units (8), that fluorescent Pseudomonas genotypes from soil are endemic at different geographic sites (1), and that hyperthermophilic Sulfolobus archaea from different geothermal areas are genetically divergent (12). Judging the endemicity of populations in different geographic regions may depend on the taxonomic scale used to distinguish populations (1). In this study we provide evidence that at least some clades may be relatively characteristic of particular regions, e.g., GBR clade III (Table ​(Table2).2). It is remarkable that in the case of the sponge species R. globostellata and P. clavata from a single geographic region (GBR), the microbial communities include representatives of distantly related poribacterial clades II and III, whose sequences exhibit levels of dissimilarity ranging from 10 to 13%. In another case poribacteria belonging to clades I, II, and IV were found in a single host, A. aerophoba, from the Mediterranean. Thus, members of widely divergent poribacterial clades occur in the same specimen in sponges in widely separated geographic regions in the world''s oceans. Three different sponge species belonging to three different genera exhibit this phenomenon.The morphology and life strategy of sponges have remained unchanged for the past 580 million years, as judged by the dramatic similarity of the morphologies of Precambrian fossils to the morphologies of recent sponges (6). Adaptation of the poribacteria to this niche might have taken place early in evolution before the various sponge orders separated from each other. It seems likely that poribacteria diverged from other bacterial phyla long before evolution of the metazoans as part of the fan-like radiation by which all bacterial phyla appear to have arisen (4). This bacterial radiation may have resulted in the divergence of the clades that we have observed for the poribacteria, but there is no indication of cospeciation between host sponges and the poribacteria.In summary, poribacteria exhibit considerable diversity and are classified into four phylogenetic clades. Poribacteria seem to be widely distributed among many different marine demosponge genera, and further studies are needed to explain the nature of the poribacterium-sponge interaction.