Ultrastructure, Molecular Phylogenetics, and Chlorophyll a Content of Novel Cyanobacterial Symbionts in Temperate Sponges

Marine sponges often harbor photosynthetic symbionts that may enhance host metabolism and ecological success, yet little is known about the factors that structure the diversity, specificity, and nature of these relationships. Here, we characterized the cyanobacterial symbionts in two congeneric and sympatric host sponges that exhibit distinct habitat preferences correlated with irradiance: Ircinia fasciculata (higher irradiance) and Ircinia variabilis (lower irradiance). Symbiont composition was similar among hosts and dominated by the sponge-specific cyanobacterium Synechococcus spongiarum. Phylogenetic analyses of 16S–23S rRNA internal transcribed spacer (ITS) gene sequences revealed that Mediterranean Ircinia spp. host a specific, novel symbiont clade (“M”) within the S. spongiarum species complex. A second, rare cyanobacterium related to the ascidian symbiont Synechocystis trididemni was observed in low abundance in I. fasciculata and likewise corresponded to a new symbiont clade. Symbiont communities in I. fasciculata exhibited nearly twice the chlorophyll a concentrations of I. variabilis. Further, S. spongiarum clade M symbionts in I. fasciculata exhibited dense intracellular aggregations of glycogen granules, a storage product of photosynthetic carbon assimilation rarely observed in I. variabilis symbionts. In both host sponges, S. spongiarum cells were observed interacting with host archeocytes, although the lower photosynthetic activity of Cyanobacteria in I. variabilis suggests less symbiont-derived nutritional benefit. The observed differences in clade M symbionts among sponge hosts suggest that ambient irradiance conditions dictate symbiont photosynthetic activity and consequently may mediate the nature of host–symbiont relationships. In addition, the plasticity exhibited by clade M symbionts may be an adaptive attribute that allows for flexibility in host–symbiont interactions across the seasonal fluctuations in light and temperature characteristic of temperate environments.


Introduction
Invertebrate-photosymbiont associations are common in shallow water marine environments, typically involving sponge or cnidarian hosts and cyanobacterial or algal symbionts [61,68]. The photosynthetic capacity of these symbionts and translocation of fixed carbon to host organisms can boost invertebrate metabolism and increase overall holobiont fitness. Similar to photosymbionts in scleractinian corals [3] and ascidians [23], the relationship between host sponges and their associated Cyanobacteria are often mutually beneficial. Indeed, some host sponges acquire supplemental nutrition from the by-products of symbiont photosynthesis [16,69] while cyanobacterial symbionts receive a sheltered habitat within sponge tissue (e.g., reduced grazing pressure and UV exposure) and possibly benefit from the nitrogenous end products of host (animal) metabolism. In addition to nutrient translocation, symbiotic Cyanobacteria may also provide a source of defensive secondary Electronic supplementary material The online version of this article (doi:10.1007/s00248-012-0047-5) contains supplementary material, which is available to authorized users. metabolites [15,62]. Accordingly, cyanobacterial symbionts appear to contribute to the competitive ability and ecological success of host sponges and represent a key functional component of the complex sponge microbiota.
Photosymbionts are prevalent in sponge communities of coastal ecosystems worldwide [63], accounting for one third to three fourths of coral reef sponges in the tropical regions [11,53,70] and over half of sponges from temperate ecosystems [27,41]. In general, Cyanobacteria are the dominant photosynthetic symbiont group in sponge hosts [11,63], although zooxanthellae and filamentous algae are also found in association with marine sponges [7,22]. The genetic diversity of sponge-associated Cyanobacteria spans multiple phylogenetic lineages and forms ten monophyletic and sponge-specific sequence clusters related to the genera Synechococcus, Synechocystis, Oscillatoria, Lyngbya, and Cyanobacterium [52,54,59].
The most commonly reported and widespread cyanobacterial symbiont is "Candidatus Synechococcus spongiarum" [66], a single-celled cyanobacterium that occurs in peripheral (ectosomal) regions of the sponge body in diverse hosts from tropical and temperate marine environments across the globe [20,52,54,59]. S. spongiarum symbionts account for up to 85 % of sponge-photosymbiont associations in Caribbean reefs [11] and exhibit variable functional significance to host sponges [2,12,16,30,71]. Molecular evidence from 16S to 23S ribosomal RNA (rRNA) internal transcribed spacer (ITS) sequences recently revealed cryptic diversity among populations of S. spongiarum, with 12 distinct symbiont clades structured by both geography and host phylogeny [13]. Additional studies targeting clade-level diversity in the S. spongiarum species complex may shed new light on the variability of host-symbiont interactions described for this widespread cyanobacterium.
Cyanobacterial symbionts related to the genera Synechocystis and Prochloron have also been described from marine sponges, primarily based on microscopic observations and ultrastructural morphology [9,44]. To date, molecular characterization of Synechocystis symbionts in marine sponges has been conducted for hosts in the genus Lendenfeldia from the Indo-Pacific [40] and Western Indian Ocean [54], Spongia sp. and Mycale sp. from Western Australian [27], and Ectyoplasia ferox from the Caribbean [50], while a single Prochloron-affiliated sequence has been reported in the Japanese sponge Halichondria okadai (GenBank acc. no. HM100971). In fact, the best studied Synechocystis and Prochloron symbionts, Synechocystis trididemni and Prochloron didemni, are associated with didemnid ascidian hosts [26,28,29,32]. Among sponge hosts, the specificity and ecological importance of Synechocystis and Prochloron symbionts are currently unknown and further studies are needed to understand the biodiversity of these Cyanobacteria and their interactions with host sponges.
In this study, we examined the diversity and activity of cyanobacterial symbionts in Mediterranean Ircinia spp. using electron microscopy, molecular characterization, and chlorophyll a quantification. The host sponges Ircinia fasciculata and Ircinia variabilis were chosen due to previous reports of cyanobacterial symbionts in these species [8,46,64,65], their close phylogenetic relationship [14], and their distinct zonation patterns within the littoral benthos of the NW Mediterranean Sea [14]. Typical of a phototrophic sponge species, I. fasciculata occurs preferentially in exposed and high irradiance zones, while I. variabilis is more common in semi-sciophilous ("shade-loving") communities of vertical walls and shaded crevices. However, distribution patterns associated with light availability can also occur in non-phototrophic sponge species [4], necessitating a detailed study of putative photosymbiont communities to confirm their presence and activity in host sponges. The objective of our study was to compare the genetic diversity, ultrastructural morphology, and chlorophyll a content of cyanobacterial symbionts in two conspecific temperate sponges. By targeting both partial 16S rRNA and entire 16S-23S ITS gene sequences, our study allowed for both comparative phylogenetic analysis and fine-scale resolution of closely related cyanobacterial symbionts.

Chlorophyll a Quantification
Chlorophyll a (chl a) concentrations were determined for I. fasciculata (n03) and I. variabilis (n03), following Erwin and Thacker [12]. Briefly, 0.25 g of freshly collected ectosomal tissue (blotted wet weight) from each individual was separately extracted in 5 ml of 90 % acetone, held overnight at 4°C. Absorbance values of supernatant aliquots were determined at 750, 664, 647, and 630 nm and chl a concentrations were calculated using the equations of Parsons et al. [35], standardized by sponge mass extracted. Chl a concentrations were compared between host sponge species with a Student's t test using the software SigmaPlot (version 11).

Transmission Electron Microscopy
To visualize the diversity and ultrastructure of cyanobacterial symbionts in I. fasciculata and I. variabilis, transmission electron microscopy (TEM) observations were conducted on small ectosome tissue pieces (ca. 4 mm 3 ), following the methods of Erwin et al. [14]. Briefly, tissue pieces were fixed and incubated (overnight at 4°C) in a solution of 2.5 % glutaraldehyde and 2 % paraformaldehyde (buffered with filtered seawater) then rinsed and stored in filtered seawater. Following dehydration in a graded ethanol series, samples were embedded in Spurr resin (room temperature), sliced into ultrathin sections (ca. 60 nm), and contrasted with uranyl acetate and lead citrate for ultrastructural observation [39]. For Synechocystis sp. symbionts, a total of seven cells were recovered and measured in I. fasciculata. Two measurements were recorded for each cell: the maximum cell diameter (hereafter, "length") and the cell diameter perpendicular to the maximum (hereafter, "width"). Cell dimensions were compared between host sponge species with a Student's t test using the software SigmaPlot (version 11).

DNA Extraction and PCR Amplification
Metagenomic DNA extracts were prepared from samples of sponge tissue (ectosome and choanosome) from I. fasciculata (n03) and I. variabilis (n03) using the DNeasy® Blood & Tissue Kit (Qiagen®), following the manufacturer's animal tissue protocol. Diluted DNA extracts (1:10) were used as templates in PCR amplification with the universal cyanobacterial forward primer 359F [34] and reverse primer 23S1R [24] to amplify a cyanobacterial rRNA gene fragment corresponding to the 3′ end of the 16S region (1,140 to 1,142 bp), the entire 16S-23S ITS region (258 to 443 bp), and the 5′ end of the 23S region (25 bp). Total PCR reaction volume was 50 μl, including 10 pmol of each primer, 10 nmol of each dNTP, 1× reaction buffer (Ecogen), and 5 units of BIOTAQ™ polymerase (Ecogen). Thermocycler reaction conditions were an initial denaturing time of 2 min at 94°C, followed by 30 cycles of 1 min at 94°C, 0.5 min at 50°C, 1.5 min at 72°C, and a final extension time of 2 min at 72°C. To minimize PCR amplification biases, a low annealing temperature and low cycle number were used and three separate reactions were conducted for each sample. PCR amplification products were gel-purified and cleaned using the QIAquick Gel Extraction Kit (Qiagen®), then triplicate PCR products were combined and quantified using a Qubit™ fluorometer and Quant-iT™ dsDNA Assay Kit (Invitrogen™).

Clone Library Construction and Sequencing Analysis
Purified PCR products (ca. 75 ng) were ligated into plasmids using the pGEM®-T Vector System (Promega). Individual clones were PCR-screened using vector primers and clones with ca. 1,650 bp inserts were purified and sequenced at Macrogen, Inc. Bidirectional sequencing with vector primers provided two overlapping sequence reads per clone and allowed the retrieval of the entire cloned amplicons. Raw sequence data were processed in Geneious [10] by aligning high-quality forward and reverse reads to yield a final consensus sequence for each clone. Quality-checked sequences are archived in GenBank under accession nos. JQ410235 to JQ410319. Consensus sequences were subject to nucleotide-nucleotide BLAST searches [1] to recover closely related sequences in the GenBank database and pairwise genetic distance (uncorrected p-distance) among sequences and rarefaction analysis by individual sponge host were conducted using the software package mothur [49].

Phylogenetic Analyses
Phylogenetic reconstructions were based on different regions of the recovered rRNA gene fragments for Synechococcus and Synechocystis, due to the resolution required for each symbiont phylogeny and the availability of reference sequences in the GenBank database. For S. spongiarum clones, phylogenies were constructed using 16S-23S rRNA ITS sequences, since 16S rRNA genes sequences do not exhibit sufficient variability for clade-level resolution of S. spongiarum and ITS sequences from S. spongiarum are available for comparative analyses [12,13]. For Synechocystis symbionts, phylogenies were constructed using 16S rRNA gene sequences. 16S rRNA gene sequences were sufficient to resolve a novel clade of Synechocystis symbionts in I. fasciculata, and few ITS sequences from Synechocystis spp. are available for comparative analyses.
Consensus 16S-23S rRNA ITS sequences from S. spongiarum clones recovered herein were compared with 12 previously described S. spongiarum clades (A to L) [13] to determine the subspecific clade affiliations of S. spongiarum symbionts in I. fasciculata and I. variabilis. A single consensus sequence was used for identical clones (100 % identity) from the same individual. Unique sequences from I. fasciculata (n020) and I. variabilis (n026), representative sequences from S. spongiarum clades A to L (n039), and congeneric outgroup sequences from cultures (n04) and environmental sources (n03) were aligned using MAFFT [25]. Maximum likelihood phylogenies were constructed in PHYML [19] with the Hasegawa-Kishino-Yano model of nucleotide substitution and a gamma distribution of variable substitution rates among sites (HKY+G), as suggested by FINDMODEL; data were resampled using 100 bootstrap replicates. Bayesian inference was used to calculate posterior probabilities of branch nodes in MrBayes [42] implemented with the HKY+G model. Markov Chain Monte Carlo Markov (MCMC) analysis was performed with four chains (temp00.2) and run for 2,000,000 generations, with a sampling frequency of 400 generations (burn-in value0 1,250). After 1,957,000 generations, the average standard deviation of split frequencies among chains reached less than 0.01.
Consensus partial 16S rRNA gene sequences from Ircinia-derived clones related to the genera Synechocystis and Prochloron were compared with previously published sequences from top GenBank matches (n064) and outgroup sequences from related cyanobacterial genera (n017) to determine the phylogenetic affiliation of Synechocystis-like symbionts in I. fasciculata and I. variabilis. Sequence alignment and phylogenetic analyses were conducted as described above for S. spongiarum clones. During MCMC analysis, the average standard deviation of split frequencies among chains reaches less than 0.01 after 1,278,000 generation cycles. Additional sponge-derived sequences related to Synechocystis have been reported from Spongia sp. (GenBank acc. nos. EU383035 and EU383036), Mycale sp. (EU383038), and E. ferox (EF159744); however, these partial sequences were excluded because their short length (<600 bp) precluded accurate phylogenetic placement, destabilized phylogenetic reconstructions, and obscured relationships among the remaining sequences in the dataset.

Morphology, Abundance, and Activity of Cyanobacterial Symbionts
Dense populations of cyanobacterial cells were observed in the ectosome of I. fasciculata and I. variabilis (Figs. 1 and 2) and corresponded to two distinct symbiont cell morphologies. The dominant symbiont cells represent the spongespecific symbiont, "Candidatus S. spongiarum" [66], diagnosed by the characteristic spiral thylakoids that occur in the cell perimeter surrounding a finely granulated cell center (Fig. 1). Few central cytoplasmic inclusions were identifiable in these cells. In both host sponges, S. spongiarum cells occurred in intercellular mesohyl areas and were actively reproducing via cell elongation, central constriction (yielding a figure 8 shape), and separation into two daughter cells. S. spongiarum cells also appeared to interact with host cells, often seen surrounding and interfacing with sponge archeocytes (Fig. 1). Symbiont cells were occasionally engulfed by host archeocytes, although no clear evidence of symbiont consumption (phagocytosis) was observed (Fig. S1).
Comparing S. spongiarum symbiont populations between the two host sponge species revealed two differentiating factors: cell size and glycogen abundance. Symbiont cells in I. variabilis were significantly larger than resident cells in I. fasciculata (Table 1) 14 μm in I. variabilis) and were within previously reported cell size ranges from different host species (Table 1). A more consistent difference among symbiont populations occurred in the abundance of glycogen granules (fine black dots, 20 to 35 nm in diameter) between the lamellae of the thylakoids in S. spongiarum cells [45]. In I. fasciculata, symbionts exhibited a high abundance of glycogen granules and similar granules were observed in host cells interfacing with symbiont cells (Fig. 1f). In I. variabilis, glycogen granules were also present in S. spongiarum cells and neighboring sponge cells, though in much lower abundance (Fig. 1e).
A second morphotype of symbiotic Cyanobacteria was observed in I. fasciculata and occurred rarely (n07) within the host tissue (Fig. 2). Cell shape was spherical and cell size was over three times larger than S. spongiarum cells (Fig. 2a), averaging 7.11±1.36 μm in length and 7.11± 1.43 μm in width. Parallel thylakoids occurred around the cell periphery and multiple cytoplasmic inclusions were observed in the cell center, including carboxysomes and polyphosphate bodies (Fig. 2b). These characteristics are diagnostic of Cyanobacteria in the genus Synechocystis and matched previous descriptions of sponge symbionts [46,65]. Unlike S. spongiarum, Synechocystis symbionts were not observed in reproductive processes and no close contact with sponge cells occurred, due to separation of symbionts from the sponge mesohyl by a lacunar space (0.5 to 2 μm) surrounding each Synechocystis cell (Fig. 2c).

Genetic Diversity of Cyanobacterial Symbionts
Consistent with electron microscopy observations, clone libraries revealed the presence of two distinct cyanobacterial symbionts in I. fasciculata and I. variabilis hosts. Rarefaction analysis revealed sufficient sampling to reach saturation in all host sponge individuals examined (Fig. S2). Analysis of the 16S rRNA gene regions (1,140 to 1,142 bp) revealed that the majority of clones from I. fasciculata (n034, 85 %) and all clones from I. variabilis (n045, 100 %) corresponded to the sponge-specific cyanobacterium "Candidatus S. spongiarum" (99 % sequence identity) [66]. The remaining clones (n06 from a single I. fasciculata individual) corresponded to the genus Synechocystis, matching most closely (>97 %) to uncultured Synechocystis symbionts from marine sponges and ascidians, including the cyanobacterium S. trididemni. In addition, Synechocystis clones matched nearly identically (>99 %) to symbiont clones derived from another dictyoceratid sponge, Spongia sp. (GenBank acc. nos. EU383035 and EU 383036); however, these partial 16S rRNA gene sequences were short (<430 bp), precluding their inclusion in subsequent phylogenetic analyses.
Phylogenetic analysis of 16S-23S rRNA ITS sequences recovered from I. fasciculata and I. variabilis revealed a novel clade of S. spongiarum distinct from all previously described symbiont clades (A to L; Fig. 3). This new symbiont clade, here labeled clade "M", exhibited reciprocal monophyly and greater than 3 % sequence divergence (average09.6 %, range06.2-21.5 %) from sister clades, thus satisfying the precedent criteria for defining a new clade of S. spongiarum [13]. Within clade M, sequence divergence values were low (average00.41 %, range00-1.3 %) and no consistent genetic differentiation by host species was observed, as clones from I. fasciculata and I. variabilis formed a mixed cluster (Fig. 3).
Phylogenetic analysis of partial 16S rRNA gene sequences recovered from I. fasciculata revealed a novel symbiont clade within the Synechocystis evolutionary lineage (Fig. 4). Four robust and distinct Synechocystis symbiont clades were resolved: a sponge symbiont clade specific to the host species I. fasciculata, two closely related sponge symbionts clades specific to the host genus Lendenfeldia, and an ascidian symbiont clade corresponding to S. trididemni (Fig. 4). Additional sequences derived from sponges (n02) and ascidians (n02) were positioned within the Synechocystis lineage, although their relationships with other Synechocystis clades were unresolved. Synechocystis sequences clustered as a sister lineage to Prochloron sequences and together formed a well-supported monophyletic clade comprised solely of symbiont-derived sequences (Fig. 4). Prochloron sequences were closely related to ascidian symbiont, P. didemni, and recovered almost exclusively from ascidian hosts, with the exception of one sponge-derived (H. okadai) and one coral-derived (Muricea elongata) sequence (Fig. 4).

Discussion
Symbiotic Cyanobacteria in the temperate sponge hosts, I. fasciculata and I. variabilis, were shown to exhibit similar species composition yet different levels of photosynthetic pigments (chl a) and storage products (glycogen granules) that correlated with the irradiance conditions of preferred host habitats. In both hosts, symbiont communities were dominated by a novel clade (M) of the unicellular cyanobacterium, S. spongiarum. A second single-celled cyanobacterium, Synechocystis sp., was also observed in I. fasciculata, though rarely (seven total cells) and sporadically (one of the three host individuals). Symbiont communities associated with the photophilic host I. fasciculata exhibited nearly twice the chl a concentrations of I. variabilis and abundant accumulation of glycogen granules, a polysaccharide storage product of photosynthetic carbon assimilation. Notably, similar (putatively glycogen) granules were also observed in host cells interfacing with active symbionts in I. fasciculata, indicating the potential transfer of surplus carbon stores to the host sponge. These results suggest that ambient irradiance conditions play a role in dictating the photosynthetic activity of sponge-associated Cyanobacteria and possibly mediate the nature of hostsymbiont interactions among different host sponge species.
The phylogenetic signature of cyanobacterial symbionts in temperate Ircinia spp. was quite different from congeneric species in the Caribbean, which lack Synechocystis symbionts, and host distinct S. spongiarum clades [13]. The fine-scale phylogenetic resolution afforded by 16S-23S rRNA ITS sequence data has important implications in host specificity, as the interpretation of symbiont specificity varies with molecular marker resolution. For example, based on 16S rRNA gene sequences, clade M symbionts recovered from Mediterranean Ircinia spp. herein matched nearly  identically (99.1-99.5 % sequence identity) to 16S rRNA gene sequences from clade J symbionts described in Caribbean Ircinia hosts. In contrast, 16S-23S rRNA ITS gene sequences showed that clade M symbionts were clearly differentiated from clade J symbionts based on sequence similarity (90.2-91.1 % identity) and phylogenetic analysis (distinct monophyletic clades), revealing cryptic biogeographic trends in symbiont structure among Ircinia hosts. The biogeographic distribution of S. spongiarum suggests that unique clades inhabit hosts from different regions. In addition to the Mediterranean clade M symbionts described herein, distinct S. spongiarum clades have also been Figure 3 Phylogeny of cyanobacterial 16S-23S rRNA ITS gene sequences from the sponge-specific symbiont S. spongiarum highlighting a new subspecific clade ("M") from I. fasciculata and I. variabilis. Terminal node labels denote the host species of each sequence, followed by the number of clones (in parenthesis) and the sponge individual for sequences from this study (bold) or the GenBank accession nos. for representative clones from the 12 previously described clades. Tree topology was constructed using maximum likelihood (ML) inference. Bootstrap support values for ML analyses (upper) and posterior probabilities (PP) for Bayesian inference (lower) are shown on internal nodes, with double asterisks indicating bootstrap values <50 % or PP <0. 50. Bold values indicate support for distinct symbiont clades (gray boxes), with clade labels shown on the right (dark bars). Outgroup sequences include three environmental and four cultured Synechococcus sequences. Scale bar represents 0.04 substitutions per site reported in the Indo-Pacific (Palau, clade F) and eastern Atlantic (Canary Islands, clade E) [13]. In fact, the majority of clades described to date are specific to a single geographic region (Fig. 5). However, the clade diversity within a region is strongly correlated with the number of host species surveyed (Fig. 5), indicating that additional sampling is required to fully elucidate the diversity and distribution of S. spongiarum clades. Indeed, even in the well-studied coral-zooxanthellae symbioses, the sampling of new hosts and environments continues to reveal novel subclades and expand the distribution of known clades of Symbiodinium symbionts [6].
Clade-level differentiation of S. spongiarum symbionts is a recently described phenomenon [13] and whether this cryptic genetic diversity relates to differences in symbiont functioning and host benefit is currently unknown. The  presence of a single, shared S. spongiarum clade in Ircinia spp. provides new insight into clade-level symbiont physiology through the comparative analyses of cellular ultrastructure and photosynthetic pigment concentrations in clade M symbionts from hosts in high (I. fasciculata) and low (I. variabilis) irradiance habitats. Clade M symbionts were smaller in I. fasciculata compared to I. variabilis, suggesting morphological plasticity in cell size in response to the ambient irradiance levels. I. fasciculata exhibited greater chl a concentrations compared to I. variabilis and dense aggregations of glycogen granules in symbiont cells, indicators of higher photosynthetic activity. Glycogen accumulation is consistent with high photosynthetic output, representing a key storage polysaccharide for fixed carbon in Cyanobacteria [33,57]. Together, these data suggest flexibility among populations of clade M symbionts and acclimation to environmental irradiance gradients, rather than expulsion and compositional shifts as reported for coralzooxanthellae symbioses [43].
Previous investigations of sponges hosting S. spongiarum have reported high variability in the functional role of symbiotic Cyanobacteria and dependence of host sponges on photosymbiont communities. Among some host sponges, symbiont loss has little effect on host growth rates [12,60,71], secondary metabolite production [18], stress response [30], and host mortality [31]. In contrast, other host species exhibit decreased growth rates [12], metabolic collapse [2], and mass mortality of local host populations [17] in response to the reduction or loss of S. spongiarum symbionts. Similar data and experiments are unavailable for Ircinia hosts; however, symbiont loss due to temperature extremes has been suggested to contribute to mass mortality events of I. fasciculata in the Mediterranean [8]. Further, the sponge Petrosia ficiformis hosts a related, facultative cyanobacterium, Synechococcus feldmanni [66], that occurs in hosts from light-exposed habitats yet is absent in conspecific sponges from dark caves [47]. The plasticity of this unique sponge-cyanobacteria symbiosis allows for comparative analyses of symbiotic and aposymbiotic P. ficiformis individuals and has yielded insight into the genetic regulation and metabolic implications of host-symbiont interactions [47,55].
Notably, our results also indicate high photosynthetic capacity in some temperate sponge-cyanobacteria symbioses, as chl a levels in I. fasciculata were consistent with values reported for tropical Ircinia spp. [11] and similar glycogen accumulation has been observed in symbionts of the tropical sponge C. nucula [45]. In contrast, low symbiont activity in I. variabilis hosts may indicate less dependence on symbiont photosynthetic output. Alternatively, symbiont populations may be actively regulated by I. variabilis to avoid specialist predators attracted to cyanobacteriarich sponges [5] and reduce the oxidative stress of reactive oxygen species produced by symbiont photosynthesis [37,38]. Additional studies are required to resolve specific hostsymbiont interactions, with emphasis on fine-scale symbiont characterization, as well as, more refined metrics of hostsymbiont metabolic interactions [16,58].
Synechocystis symbionts formed a monophyletic clade specific to the Mediterranean host I. fasciculata and distinct from related sponge and ascidian-associated symbionts. The morphology of these symbionts matched previous descriptions of symbionts characterized as Aphanocapsa raspaigellae in I. variabilis [46,65]. Reclassification of this cyanobacterium to the genus Synechocystis was suggested by previous authors [26,63] and is supported by the phylogenetic analyses herein. Synechocystis symbionts formed a monophyletic cyanobacterial lineage and exhibited a close phylogenetic relationship with Prochloron symbionts, consistent with previous molecular phylogenies [51] and the ultrastructural similarity of these symbiont genera [9,21].
The ecological significance of Synechocystis symbionts in marine sponges remains unclear, as these photosymbionts exhibit variable incidence and abundance among different populations of Ircinia host sponges. In the current study, Synechocystis symbionts in I. fasciculata occurred rarely and sporadically among the sponge individuals studied and were observed to have limited interactions with host cells. Previous investigations of Ircinia spp. from different regions of the Mediterranean have reported more abundant Synechocystis populations in host sponges from Bari, Italy [46] and Marseille, France [65], suggested biogeographic variability in these photosymbiont communities. Consistent among these previous reports and the present study is the physical separation of Synechocystis from host sponge tissue by a lacunar space surrounding each symbiont cell. The lack of direct contact with sponge tissue may limit host-symbiont interactions, although cellular secretion from intact Synechocystis symbionts and Figure 5 Linear regression of the clade-level diversity in the spongeassociated cyanobacterium, S. spongiarum, and the number of host species studied for five geographic regions leakage from disintegrating cells has been observed [46]. Further, the occurrence of Synechocystis symbionts as secondary to dominant cyanobacterial populations in Ircinia spp. [46, 65; this study] and Lendenfeldia spp. hosts [40] suggests that these symbionts may be opportunistic and exploiting host habitats receptive to photosymbionts (e.g., distributed in high irradiance zones, tolerant of oxidative stress) while providing minimal ecological benefit. Additional study is clearly required to test such hypotheses as well as assess potential contributions to host ecology beyond symbiont-derived photosynthates (e.g., secondary metabolite production).
Environmental irradiance gradients represent an important factor in structuring invertebrate-photosymbiont associations and may dictate the nature of host-symbiont interactions. In the symbiosis between cnidarian hosts and zooxanthellae, irradiance exposure and intensity have been shown to influence the density, physiology, and composition of photosymbiont communities [3,48]. Similarly, the cyanobacterial symbionts studied herein exhibited physiological and morphological differences in related host sponges from different light environments, including photosynthetic pigment content (chl a concentrations), fixed carbon accumulation (glycogen granules), and symbiont cell size. In contrast to cnidarian symbionts, whose cladal composition can shift across smallscale irradiance gradients [43], sponge photosymbiont communities were dominated by clade M symbionts in both host sponges regardless of irradiance conditions, suggesting that less dynamic and more versatile host-symbiont interactions occur in Ircinia-Synechococcus associations. Thus, the temperate clade M of S. spongiarum described herein appears to represent a flexible symbiont able to survive in sponge hosts under different environmental conditions, a potential hallmark of symbiotic Cyanobacteria occurring in temperate ecosystems that must tolerate large seasonal fluctuations in light and temperature. Future research on temperate sponge-cyanobacteria interactions targeting the fine-scale characterization of symbiont communities and temporal monitoring of hostsymbiont interactions are required to test such hypotheses and determine the contributions of these symbiotic systems to host ecology and microbial biodiversity.