Macroalgal-Associated Dinoflagellates Belonging to the Genus Symbiodinium in Caribbean Reefs

Coral-algal symbiosis has been a subject of great attention during the last two decades in response to global coral reef decline. However, the occurrence and dispersion of free-living dinoflagellates belonging to the genus Symbiodinium are less documented. Here ecological and molecular evidence is presented demonstrating the existence of demersal free-living Symbiodinium populations in Caribbean reefs and the possible role of the stoplight parrotfish (Sparisoma viride) as Symbiodinium spp. dispersers. Communities of free-living Symbiodinium were found within macroalgal beds consisting of Halimeda spp., Lobophora variegata, Amphiroa spp., Caulerpa spp. and Dictyota spp. Viable Symbiodinium spp. cells were isolated and cultured from macroalgal beds and S. viride feces. Further identification of Symbiodinium spp. type was determined by length variation in the Internal Transcribed Spacer 2 (ITS2, nuclear rDNA) and length variation in domain V of the chloroplast large subunit ribosomal DNA (cp23S-rDNA). Determination of free-living Symbiodinium and mechanisms of dispersal is important in understanding the life cycle of Symbiodinium spp.


Introduction
Coral reefs maintain an incredible level of productivity considering the oligotrophic waters in which they are found. The high productivity and rapid growth exhibited by reef-building corals can be attributed to their association with symbiotic dinoflagellates referred to as zooxanthellae [1]. Other symbiotic reef dwellers include platyelminths, mollusks, sponges and protists [2,3]. There is an increasing understanding of Symbiodinium outside their hosts, documented from temperate waters of New Zealand [4], benthic sands in Hawaii [5], water column in China [6], and in coral reef sediments [7].
The complete life stages and sexual reproduction of most dinoflagellates, including Symbiodinium spp., are not well understood [8]. Symbiodinium spp. within the host exhibits asexual reproduction through mitosis. However, genetic evidence including no linkage disequilibrium and high allelic diversity suggest the presence of recombination and sexual reproduction [9,10]. Furthermore, Gonyaulacales and Gymnodiniales, Suessiales (Symbiodinium) sister groups, exhibit nuclear cyclosis and meiosis according to nuclear and mitochondrial DNA [11][12][13].
Characterization of free-living Symbiodinium spp. is important for understanding coral reefs and the coral-symbiont dynamics [14]. There are two modes of zooxanthellae acquisition, vertical (from the parental colony to the eggs) and horizontal (uptake from the environment by aposimbiotic eggs or larvae) [15]. It is important to know if there exists an environmental Symbiodinium spp. reservoir from where corals with horizontal symbionts acquisition mode will obtain Symbiodinium spp.
The second part of this study addresses dispersion exhibited by Symbiodinium spp. carried on by Sparisoma viride. This is especially interesting considering their limited motility capacities [16]. Dispersion mechanisms should include oceanic currents, nearshore currents, and/or some living disperser, such as reef fishes. There is compelling evidence showing corallivorous fish have the ability to disperse viable zooxanthellae capable of establishing symbioses with anemones [17]. If free-living Symbiodinium spp. live in other habitats such as macroalgal beds and turfs, herbivorous fishes should serve as a dispersing mechanism, disposing partially digested and viable free-living Symbiodinium in various habitats.
Sparisoma viride, known as the stoplight parrotfish, belongs to a functional group of herbivorous that has a significant effect in Caribbean reefs due to their bioerosion, sediment removal and foraging behaviors [18,19]. In addition, this fish is considered to have a big effect on corals, inflicting numerous injuries directly to coral colonies [20]. Bruggemann [21] found that S. viride feeds exclusively on algae; 95% of the bites registered were done on algae associated with dead coral, while only a 3.6% was done to live coral (''spot biting''), presumably a territorial behavior. It was noted that occasionally the fish spite out the food, particularly when they ingested live corals [21][22][23].
Coral reef scientists have started to appreciate the important role of parrotfishes as ecosystem engineers and keystone species, leading to cascade effects and positive feedbacks directly to coral reef health and resilience. Parrotfishes have record rates of bioerosion in some coral reef areas and their coral grazing effect promotes reef-building coral diversity [18]. Marine protected areas effectiveness, nowadays under the threats of global warming and coral bleaching, depends greatly on ecosystem resilience, where the grazing effect of parrotfishes, as well as other herbivores, is the key role to maintain the functional dominance of reef builders such as corals and coralline algae [24,25].
The aims of this paper were (i) to present both ecological and molecular evidence demonstrating the existence of macroalgalassociated demersal free-living Symbiodinium spp. populations in coral reefs and (ii) to examine the potential role of reef fishes (Sparisoma viride) as free-living Symbiodinium spp. dispersers.

Field Collections
Samples were collected by means of SCUBA diving from Cartagena, Santa Marta, and Tobago (Trinidad and Tobago) reefs, between 2005 and 2006. Sites (reefs), depths, and number of samples are summarized in the Table 1. Sampling was conducted by applying gentle pressure on several algal beds (mainly of Halimeda spp., Dictyota spp., Lobophora variegata, Caulerpa spp., and Amphiroa spp.) that where next to coral colonies, and withdrawing particles that were in suspension with a 60 ml syringe ( Figure 1). Water above the algal beds were sampled prior to agitation, these samples served as experimental controls. One set of water samples with the suspended particles was transferred to eppendorf tubes containing DMSO or 96% ethanol for molecular analysis. The other set of samples was cultured quickly after collection. The collection of Symbiodinium spp. from the feces of Sparisoma viride was done by means of SCUBA diving. Using 100 ml sterile tubes, feces from the water immediately after being released from the fish were collected and kept in sterile bags. Experimental controls from the water column were sampled. The collected samples were then cultured at the laboratory.

Cultures
F/2 medium [26], shown to be a suitable medium for culturing Symbiodinium spp. [27,28], was prepared using seawater passed through 0.45 micrometers filter paper. In addition, the antibiotic chloranphenicol was used (1 mg ml 21 ) to avoid bacteria contamination. The cultures were started in sterilized test tubes, sealed with cotton, and exposed to a photoperiod 12 hrs light: 12 darkness at 27-30uC. Direct observations under microscope of the cultures were carried out to verify zooxanthellae presence. When zooxanthellae growth was noticed, a sample was taken out and DNA extracted.

DNA Extraction and Molecular Analyses
DNA from cultured cells and environmental samples were extracted using the CTAB phenol-chloroform-isoamyl alcohol protocol [29]. Two gene regions, ITS2 rDNA and chloroplast 23S ribosomal DNA (rDNA) domain V (cp23S) were used to detect Symbiodinium spp. using polymerase chain reaction (PCR). The ITS2 region was amplified using the forward primer ''ITSintfor2'' (59GAATTGCAGA ACTCCGTG-39) and a modified reverse primer 'ITS2CLAMP (59CGCCCGCCGC GCCCCGCGCC CGTCCCGCCG CCCCCGCCC GGGATCCATA TGCTT-AAGTT CAGCGGGT-39) [30], with a GC clamp (underlined). PCR reactions were carried out using LaJeunesse (personal communication) ''Touchdown'' protocol, in a MyCycler thermocycler (BioRad) under the following conditions: an initial denature period at 92u for 3 min, followed by 35 cycles of 30 sec at 92uC, 40 sec at 48uC and 30 sec at 72uC and a final extension period of 10 min at 72uC. Primers used to amplify cp23S were 23SHY-PERUP (59-TCAGTACAAATAATATGCTG-39) and 23SHY-PERDNM13 (59-GATAACAATTTCACACAGGTTATCGCCC-CAATTAAACAGT-39) [28]. The PCR conditions were: an initial denature period of 3 min at 94uC, followed by 40 cycles of 1 min at 94uC, 30 s at 54.2uC, 30 s at 72uC and a final extension period of 5 min at 72uC [31]. The ITS2 products were loaded in a Denature Gradient Gel Electrphoresis (DGGE) containing a gradient of 3.15 M urea/18% deionized formamide to 5.6 M urea/37% deionized formamide; bands were excised from the gels, reamplified and sequenced. The cp23S rDNA were electrophoresed in polyacrylamyde gels. As in DGGE, the bands were excised from the gels, reamplified and sequenced.  To determine the identity of the ITS2 and cp23S sequences, phylogenetic trees were generated using sequences from clades A, B and C from the Caribbean. These sequences were taken from the Santos lab (Auburn University) webpage database and from Symbiodinium spp. isolated from corals of the same reefs that the water and fecal samples were collected (unpublished) ( Figure 1B). Phylogenetic trees were obtained using MrBayes [32] and maximum likelihood-ML (PAUP*) using the best-fit model and parameters according to Modeltest [33] and the Akaike Information Criterion.

Results
Bed-forming macroalgae, such as Halimeda spp., Lobophora variegata, Amphiroa spp., and Caulerpa prolifera ( figure 1B-D), were the most frequent non-coral habitats for free-living Symbiodinium spp. in coastal Colombia. Dictyota spp. was the most prevalent in Tobago (Table 2). Near bottom water column controls were negative for all Cartagena samples but were positive in the Tobago (Booby Island) experimental controls (30%). We also found occasional positive amplifications in sediments around corals (3%), like previous works conducted by Coffroth [7] show.
Positive molecular identifications of free-living Symbiodinium spp. associated with macroalgal beds were found both in environmental samples (table 2) and in cultures. Symbiodinium spp.-like cells were observed directly under the microscope further verifying this association ( Figure 1A). Overall, PCR amplifications of DNA isolated from samples collected at different locations and several habitat/depth replicates, yielded Symbiodinium spp. products from the planktonic demersal genomic DNA (Tables 2 and 3). Of the 9 sampled reefs, we detected Symbiodinium spp. in 7 reefs in several replicates, having 53% of positive PCR products from the noncultured samples (table 3). Sequencing (ITS2 accession numbers EU139607, EU139608 and cp23S accession numbers EU139605, EU139606) and comparative analyses with known sequences from cp23S and ITS2 corroborated the identity of the most frequent PCR amplification products belonging to Symbiodinium spp. clades A (Symbiodinium [ = Gymniodinium] linuchae, A4: ITS2), B (B184; cp 23S) and C (ITS2) (Figure 2A-B; Table 2). Those Symbiodinium spp. types are also present in symbiosis with cnidarian host species, such as Porites astreoides (A4) and Pseudopterogorgia acerosa (B184). The identity of isolated type of clade C is not precise, due to the large polytomy present in this group.
Free-living Symbiodinium spp. were successfully cultured ( Table 3). The cultures of Symbiodinium spp. cells from the feces of S. viride were also successful although a few experimental controls from the water column presented a positive band in the molecular analysis. Positive DNA extractions from commercial cultures (Carolina, USA) of other dinoflagellates such as Amphidinium carterae and Gymnodinium spp., as well as a diatom consortium, were always negative for the targeted cp23S DNA regions, reassuring on the zooxanthellae specificity of the molecular method used.

Discussion
This study found Symbiodinium spp. cells in demersal plankton habitats [34,35] in the coral reef areas of coastal Colombia (Cartagena and Santa Marta), Trinidad and Tobago (Charlotteville, Tobago). Water samples collected directly from macroalgal beds and parrotfish feces were both found to contain free-living Symbiodinium spp. Positive identification was achieved by molecular analysis and viability by culturing of Symbiodium spp. (Tables 2-3).
Macroalgae, with their intricate branching networks, have high surface to volume ratio, providing substrate, light attenuation, and refuge ( Figure 1B-D) and may be a major habitat for free-living Symbiodinium spp. on Caribbean coral reefs. It is important to mention that there is still uncertainty whether Symbiodinium spp. live as epiphytic dinoflagellates such as Gambierdiscus toxicus [36] or move freely inside the interstitial spaces within algae. Nonetheless, these findings are in accordance with the low motility and Table 2. Summary of the free-living Symbiodinium spp. clades found in the sampled reefs, depths, and substrates.  buoyancy of Symbiodinium spp. [35,37] as demersal organisms and with the occurrence of Symbiodinium spp. in gut contents of macroalgal-feeders such as the Queen conch Strombus gigas [30] and coralliivorous fish [17]. The occurrence of Symbiodinium spp. in other habitats helps to explain the large diversity that this genus has shown. It has been proposed that niche diversification of Symbiodinium spp. outside their cnidarian host could have maintained symbiont diversity through ecological shifts [30]. Zooxanthellae types A4 and B184 are two common symbionts of scleractinian and soft corals in the study area (unpublished), as well as other sampled Caribbean locations [38]. Despite the lack of precise type identification, the clade C symbiont is related to other types found in hard corals of the genus Agaricia spp. and Montastraea spp., both prevalent in the Caribbean reefs. The types found here are a small proportion of the high Symbiodinium spp. diversity that can be associated to reef dwellers [30,39].
In this study, the sampled macroalgal beds were directly adjacent to symbiotic corals. Despite the limited motility capacities of Symbiodinium spp., short distance dispersal of free-living Symbiodinium spp. from the macroalgal areas to the corals could be achieved by their diurnal swimming behavior or localized hydrodynamic conditions. Given that S. viride feces carry viable Symbiodinium spp. cells, dispersion on a larger scale could be carried out by coralivores fish. Fecal analysis of other coral predators like Arothron meleagris, Chaetodon auriga, Chaetodon unimaculatus and one nudibranch Berghia major from Hawaii have also been shown to contain viable Symbiodinum spp. cells. [17].
There is evidence of zooxanthellae uptake from the environment in octocorals (Briareum asbestinum) as adults and recruits [7,14], as well as, in the medusae (Cassiopea xamachana) [15] and in scleractinian (Acropora longicyathus) [40]. However, there is evidence that the changes of Symbiodinium types inside the host are due to shifts in the densities of preexisting types and not from the acquisition of environmental ''novel'' types [41,42]. Lastly, there is evidence of a stable symbiotic association over time [43,44], where there is no environmental Symbiodinium spp. uptake.
The term holobiont has been used to describe the tight symbiotic relationship between coral and algae. The reported advantages of the symbiosis are mainly for the coral [45][46][47] and the consequence of the disruption of this nearly obligate mutualism (e.g., coral bleaching) is catastrophic for corals. The advantages for Symbiodinium spp. are obtaining inorganic nutrients from host metabolism, including ammonium and phosphate [48], as well as, protection from predation and irradiation (UV). According to the findings in this work, the symbiosis may be facultative for some Symbiodinium spp. because they can find some of the benefits that they obtain from the coral in other habitats such as macroalgal beds.
Characterization of free-living Symbiodinium spp. populations is also important for understanding the life history of Symbiodinium spp. As mentioned above sexual reproduction in Symbiodinium spp. is not well understood [10]. However, molecular evidence suggests the presence of recombination and sexual reproduction [11]. It has been further suggested that this mode of reproduction might occur in the free-living stage [2], where the cells may encounter different genotypes.
From this characterization study, we can conclude that there are macroalgal-associated demersal free-living Symbiodinium spp. populations in coral reefs and that Sparisoma viride feces carry viable Symbiodinium cells. This information builds upon our limited understanding of the distribution and dispersion of demersal freeliving Symbiodinium spp. More research on Symbiodinium life history and ecology is urgently needed.