Symbiodinium Photosynthesis in Caribbean Octocorals

Symbioses with the dinoflagellate Symbiodinium form the foundation of tropical coral reef communities. Symbiodinium photosynthesis fuels the growth of an array of marine invertebrates, including cnidarians such as scleractinian corals and octocorals (e.g., gorgonian and soft corals). Studies examining the symbioses between Caribbean gorgonian corals and Symbiodinium are sparse, even though gorgonian corals blanket the landscape of Caribbean coral reefs. The objective of this study was to compare photosynthetic characteristics of Symbiodinium in four common Caribbean gorgonian species: Pterogorgia anceps, Eunicea tourneforti, Pseudoplexaura porosa, and Pseudoplexaura wagenaari. Symbiodinium associated with these four species exhibited differences in Symbiodinium density, chlorophyll a per cell, light absorption by chlorophyll a, and rates of photosynthetic oxygen production. The two Pseudoplexaura species had higher Symbiodinium densities and chlorophyll a per Symbiodinium cell but lower chlorophyll a specific absorption compared to P. anceps and E. tourneforti. Consequently, P. porosa and P. wagenaari had the highest average photosynthetic rates per cm2 but the lowest average photosynthetic rates per Symbiodinium cell or chlorophyll a. With the exception of Symbiodinium from E. tourneforti, isolated Symbiodinium did not photosynthesize at the same rate as Symbiodinium in hospite. Differences in Symbiodinium photosynthetic performance could not be attributed to Symbiodinium type. All P. anceps (n = 9) and P. wagenaari (n = 6) colonies, in addition to one E. tourneforti and three P. porosa colonies, associated with Symbiodinium type B1. The B1 Symbiodinium from these four gorgonian species did not cluster with lineages of B1 Symbiodinium from scleractinian corals. The remaining eight E. tourneforti colonies harbored Symbiodinium type B1L, while six P. porosa colonies harbored type B1i. Understanding the symbioses between gorgonian corals and Symbiodinium will aid in deciphering why gorgonian corals dominate many Caribbean reefs.

Despite being prominent members of Caribbean reef communities, studies on gorgonian coral physiology are sparse, predominantly focusing on the gorgonian coral hosts without addressing their symbiosis with the unicellular dinoflagellates in the genus Symbiodinium.Studies tracked digested material [15,16], described sclerite (microscopic skeletal elements) formation [17,18], or measured growth [19,20], and feeding rates [21,22].In addition, gorgonian secondary metabolites have been extensively studied due to their medical and economic importance [23].Two early studies measured the photosynthetic rates of Symbiodinium in a few Caribbean gorgonian species [24,25], while key photosynthetic characteristics, including light absorption efficiencies, have not been measured.Furthermore, the handful of studies that investigated the physiology of Caribbean gorgonian corals and their Symbiodinium [24][25][26][27][28][29] did not identify the Symbiodinium present.
Collecting baseline physiological data on coral-algal symbioses [39], and not just data when the symbioses are stressed, are critical to evaluating the effects of stressors on symbioses [40].The objective of this study was to characterize the photosynthesis of Symbiodinium, in hospite and in isolation, in four common Caribbean gorgonian species: Pterogorgia anceps, Eunicea tourneforti, Pseudoplexaura porosa, and Pseudoplexaura wagenaari.Studying the physiology of the symbiosis between gorgonian corals and Symbiodinium may shed light on why gorgonian corals dominate Caribbean reefs while scleractinian coral abundance is declining.

Sample collection and acclimation
In June 2010, gorgonian branches were collected at 3 m depth from a patch reef near Puerto Morelos, Quintana Roo, Me ´xico (20u 52' N, 86u 51' W, permit number DGOPA.11519,121109.3949,Secretaria de Medio Ambiente y Recursos Naturales, Me ´xico).One branch was removed from each sampled colony of the four gorgonian species: Pterogorgia anceps (n = 9), Eunicea tourneforti (n = 9), Pseudoplexaura porosa (n = 9), and Pseudoplexaura wagenaari (n = 6).In order to maintain their natural orientation, the branches were attached to vertical PVC stands and were held in outdoor aquaria with flowing seawater for 11 days for acclimation.The temperature in the aquaria was determined using Hobo pendant data loggers (Onset Computer Corporation, MA, USA), and it was maintained at 29.560.5uC,similar to the ambient temperature on the reef, by using an aquarium chiller (0.5hp Delta Star, Aqua Logic, USA) and heaters (1000 W and 1800 W EasyPlug heater, Process Technologies, USA).Light levels in the aquaria were maintained at levels similar to those at the collection site (,900 mmol quanta m 22 s 21 at solar noon) by shading the aquaria with window screening.

Photochemical efficiency of photosystem II in Symbiodinium
Throughout the acclimation period, the photochemical efficiency of photosystem II of the Symbiodinium in each branch was measured using a diving pulse amplitude modulated (PAM) fluorometer (WALZ, Effeltrich, Germany) at solar noon (DF/F m '; effective yield) and after sunset (F v /F m ; maximum yield).Photochemical efficiency was measured approximately 2 cm below the branch tip, on the side of the branch, at a constant distance from the surface of the branch.Using the effective and maximum yield values obtained for each species, we calculated the maximum excitation pressure over photosystem II, Q m , whereby Q m = 1-((DF/ F m ')/(F v /F m )) [41].

Oxygen flux of Symbiodinium in gorgonian branches
After 11 days of acclimation, the oxygen fluxes of the branches were measured using Clark-type oxygen electrodes in twin 0.5 l acrylic chambers.The electrodes were connected to a 782 Oxygen Meter (Strathkelvin Instruments Ltd., North Lanarkshire, Scotland).The chambers were filled with 0.45 mm filtered seawater with 4 mM sodium bicarbonate [42].Water was circulated inside each chamber using a small pump.Electrodes were calibrated by bubbling oxygen and nitrogen gas to define 100% and 0% oxygen concentrations, respectively.A water jacket, connected to a water recirculator, maintained the temperature of 29.0-29.5uCinside the chambers.
To measure the oxygen flux of a gorgonian branch, a branch was sealed in a chamber.Following a 20 min acclimation, respiration was measured for 10 min in darkness.Then, the branch was illuminated from one direction with three 6 W LED light bulbs, while white acrylic on the sides and rear of the chamber reflected the light within the chamber.The branch was exposed to 11 irradiance levels (0-2200 mmol quanta m 22 s 21 ) by progressively removing sheets of window screening.The irradiance levels in each chamber were measured using a 4p quantum sensor (WALZ, Effeltrich, Germany).A gorgonian branch was exposed to each irradiance level for up to 15 min, or until a linear change in oxygen concentration was recorded in the chamber.Following the last light level, respiration was measured again in darkness.
Oxygen flux measurements required 2.5 h per sample and were collected at approximately 09:00 or 12:00 local time.Due to the lengthy data collection time, it took 10 days to measure the oxygen fluxes of all the gorgonian branches.To account for changes in photosynthetic rate at different times of day, the photosynthesis of Symbiodinium in each gorgonian species was measured at alternate times on consecutive days.We then used the oxygen flux data to produce photosynthesis versus irradiance (P-E) graphs (see below).
After measuring the oxygen flux, the surface areas of the gorgonian branches were calculated.For gorgonian species with cylindrical branches (E.tourneforti, P. porosa, and P. wagenaari), we measured the length and diameter of each branch and calculated the surface area of a cylinder.For P. anceps, whose branches are blade like, we measured the diameter and length of the blade and calculated the surface area by combining the surface area of all rectangular sides.

Isolation and oxygen flux of isolated Symbiodinium
Branches were returned to shaded aquaria for 1 h before Symbiodinium isolation.Symbiodinium were isolated by homogenizing, in a mortar and pestle, a 2 cm section of the gorgonian branch, obtained 3 cm below the tip of the branch.The homogenate was diluted in 20 ml of 0.45 mm filtered seawater and centrifuged for 1 min at 500 rpm.The liquid fraction was filtered through 150, 74, and 20 mm nitex meshes.The filtered fraction was then centrifuged for 1 min at 3500 rpm and the Symbiodinium pellet was washed two times with 10 ml of filtered seawater.Following the last wash, the Symbiodinium pellet was resuspended in 5 ml of filtered seawater and divided into aliquots for oxygen flux measurements (0.5 ml), chlorophyll content (1.5 ml), cell density (0.25 ml), and genetic identification (2 ml).
To measure isolated Symbiodinium oxygen flux, 0.5 ml of the Symbiodinium slurry was mixed with 2.5 ml filtered seawater containing 4 mM NaHCO 3 and loaded into a water-jacketed glass cell respirometry chamber (StrathKelvin Instruments Ltd).The Symbiodinium samples were stirred with magnetic stir bars and maintained at 29uC with a water recirculator.Oxygen flux measurements were recorded in two chambers simultaneously using Clark-type oxygen electrodes.Respiration was measured in darkness for 10 min before and after a series of irradiance levels.The isolated Symbiodinium were exposed to 13 increasing irradiance levels (0-2200 mmol quanta m 22 s 21 ) for 10 minutes per level or until a linear change in oxygen concentration was observed in the chamber.For each sample, the irradiance levels were measured using a 4p quantum sensor (WALZ, Effeltrich, Germany).Oxygen fluxes of isolated Symbiodinium were mea-sured after those of the intact branches, at approximately 15:00 or 16:00 local time.

Calculation of photosynthetic rates
Net and gross rates of oxygen flux were plotted against irradiance to generate photosynthesis versus irradiance (P-E) graphs for both in hospite and isolated Symbiodinium.Coefficients from P-E curves (P s , the maximum photosynthetic rate in the absence of photoinhibition; a, the initial slope of the curve; and b the photoinhibition coefficient) were determined for each sample by fitting the equation of Platt et al. [43] using the nlsList function in the nlme package of the R statistical software.Photosynthetic rates were standardized to the surface area of the gorgonian branch, the total number of Symbiodinium, and the total amount of chlorophyll a. a was standardized to the amount of chlorophyll a.To calculate the diurnal balance between gross photosynthesis and respiration, we integrated the P-E curve over the irradiance for a 24 h period.

Symbiodinium density
Symbiodinium density in a gorgonian branch was estimated from averaging four replicate hemocytometer counts (0.4 mm 3 each) of the Symbiodinium cell density aliquot.Oxygen fluxes of gorgonian branches were standardized to the total number of Symbiodinium cells in a branch, which was estimated using the density of cells in the homogenized piece (cells cm 22 ) multiplied by the surface area of the entire branch.For isolated Symbiodinium, the number of cells in the respirometry chamber was estimated using the cell density determined from the Symbiodinium cell density aliquot.

Chlorophyll content in Symbiodinium
For chlorophyll quantification, the Symbiodinium cells in the 1.5 ml aliquot of the Symbiodinium slurry were pelleted, and the supernatant was decanted.Chlorophylls were extracted from the Symbiodinium cells by adding 950 ml of cold 100% acetone and 50 ml DMSO.After 24 h of extraction at 220uC in the dark, the absorbance of the extract was measured at 630, 660, and 750 nm using an ELYPTICA model ELy-2000 spectrophotometer.Absorbance at 750 nm was subtracted from the absorbance at 630 nm and 660 nm for each sample and chlorophylls a and c 2 were estimated using the equations of Jeffrey and Humphrey [44].Chlorophyll concentrations were standardized to surface area (mg chl cm 22 ) and to cell density (pg chl cell 21 ).Oxygen flux of gorgonian branches was standardized to the total amount of chlorophyll a in the gorgonian branches, which was calculated using the concentration of chlorophyll per surface area of the homogenized piece multiplied by the surface area of the entire branch.For isolated Symbiodinium, the amount of chlorophyll a in the respirometry chamber was estimated using the chlorophyll a concentration obtained from the 1.5 ml chlorophyll content aliquot.

Estimated absorbance and chlorophyll a specific absorption
Following the oxygen flux measurements, the reflectance spectrum (400-750 nm) of each branch was measured using an Ocean Optics USB 4000 fiber optic spectrophotometer.The fiberoptic cable was held at a 45u angle above a branch, which was illuminated on all sides to produce a homogeneous light field (designed by T. Scheufen, UNAM).A dried gorgonian branch, painted with white acrylic paint, was used to correct for light  scattered by the surface of the gorgonian branch.The painted branch reflected ,90% of PAR compared to a similarly shaped object wrapped in Teflon.Surface-corrected reflectance spectra were standardized to the reflectance value at 750 nm.Estimated absorbance spectra (D e ) were calculated as the negative log of the corrected reflectance.Chlorophyll a specific absorption (a* chl a ), was calculated using the equation a* chl a = (D e 675 /r) x ln (10), where r is mg m 22 of chlorophyll a [45].

Genetic identification of Symbiodinium
For each gorgonian branch, 2 ml of the isolated Symbiodinium slurry was centrifuged at 10,000 rpm for 1 min to pellet the Symbiodinium cells.The supernatant was removed and replaced with 2 ml 100% EtOH to preserve the Symbiodinium.DNA was extracted from an aliquot of this solution using the Qiagen DNeasy Plant Mini kit.The internal transcribed spacer 2 (ITS2) region of ribosomal DNA was amplified using the primers ITSintfor2, ITS-Reverse, and ITS2CLAMP following the protocol of LaJeunesse [34].The PCR products were separated using denaturing gradient gel electrophoresis (DGGE) with a 45-80%  denaturing gradient and run at 120 V for 13 h.Dominant ITS2 sequence variants in unique profiles were excised, re-amplified using ITSintfor2 and ITS-Reverse, and then sequenced using an Applied Biosystems 3730XL automated sequencer at the DNA Laboratory at Arizona State University.Since we found that at least one colony of each of the four gorgonian species harbored Symbiodinium type B1, we determined whether colonies of the four different species harbored the same B1 lineage [36].Microsatellite flanking regions from B1 Symbiodinium were PCR amplified with the primers B7Sym15 [46] and CA4.86 [35] and directly sequenced using an Applied Biosystems 3730XL automated sequencer.The flanking sequences were concatenated and aligned with 16 published Symbiodinium type B1 sequences from Supplementary tables S1 in [36,47].Six samples, five from our study (1 P. anceps, 2 P. porosa, 2 P. wagenaari) and one published (B1 lineage 1.4) [36], were only represented by B7Sym15 sequences.For each flanking region, the Jukes Cantor model of sequence evolution was chosen using AIC scores from jModelTest (2.1.4)[48,49] and substitution rates were assumed to follow a gamma distribution with four categories.In order to determine the phylogenetic relationships among B1 Symbiodinium, a Bayesian phylogenetic tree was generated using the MrBayes (3.2.2) [50].Two sets of four independent chains were run for 1,000,000 generations, but the model converged after 916,000 generations.Trees were sampled every 100 generations (916 trees per run) and the first 25% of trees were discarded as burn-in.The average standard deviation of split frequencies was less than 0.01.To confirm the topology generated by MrBayes, 100 bootstrap replicates of a maximum-likelihood phylogenetic tree were generated in Garli 2.0 using the Jukes Cantor model of sequence evolution [51].Symbiodinium type B19, from supplementary table S1 in [36] was used as an outgroup for both analyses.

Statistical methods
In a given host species, different host-symbiont combinations can differ in their physiology [52].Therefore, we excluded from statistical analyses gorgonian-Symbiodinium ITS2 type combinations represented by three or fewer colonies.The sample sizes given in the results reflect the number of colonies used in the statistical analyses.Each parameter was analyzed using a one-way ANOVA with gorgonian species as a factor.Residuals for most parameters were not normally distributed and/or had unequal variance among species.Therefore, most data required transformation using a log, square root, or reciprocal function.When significant differences were found among gorgonian species, Tukey HSD post hoc tests were used to identify significant differences among all pairwise species combinations using Bonferonnicorrected p-values.a* chl a data could not be transformed to meet the assumptions of ANOVA and was tested by determining the frequency of obtaining an F statistic greater than or equal to the observed F statistic in 10,000 permutations of the data.

Photochemical efficiency of photosystem II in Symbiodinium
Measuring the effective (DF/F m ') and maximum (F v /F m ) quantum yield of PSII throughout the acclimation period enabled us to calculate maximum excitation pressure over photosystem II, Q m.Q m ranged from 0.25 in P. wagenaari to 0.32 in P. anceps and did not significantly differ between the four species.The Q m values demonstrated that there was no detrimental tank effect on the gorgonians.

Oxygen flux in gorgonian branches
P-E curves for gorgonian branches did not reach saturation despite being exposed to more than 1800 mmol quanta m 22 s 21 .Therefore, the maximum photosynthetic rate could not be determined as in Platt et al. [43].Since 1800 mmol quanta m 22 s 21 was the highest irradiance shared between the two respirometric chambers, the fitted values at 1800 mmol quanta m 22 s 21 were used as a proxy for the maximum photosynthetic rates.P. porosa and P. wagenaari had the highest average photosynthetic rates per cm 2 at 1800 mmol quanta m 22 s 21 (Figure 1A), with both species having similar gross photosynthetic and respiration rates per cm 2 , although P. porosa had significantly higher net photosynthetic rates per cm 2 than P. wagenaari (Table 1).P. anceps had significantly higher respiration rates per cm 2 than E. tourneforti, but the two species did not differ significantly in their maximum photosynthetic rates per cm 2 (Figure 1A, Table 1).When integrated on a diurnal cycle, the total oxygen produced via photosynthesis (10.3 to 25.9 mmol O 2 ) was less than the oxygen consumed via respiration (13.4 to 29.7 mmol O 2 ) for each gorgonian species and therefore the 24 h gross P/R were less than 1 (Table 1).
The gross and net photosynthetic rates per Symbiodinium cell exhibited a different pattern than the photosynthetic rates per cm 2 (Table 1).P. anceps had significantly higher gross and net photosynthetic rates per cell than all other species (Figure 1B, Table 1).P. porosa had the second highest average photosynthetic rates per cell, which were significantly higher than those of P. wagenaari, but not E. tourneforti.Photosynthetic rates per cell of E. tourneforti were similar to both P. porosa and P. wagenaari.
Photosynthetic rates per chlorophyll a produced a similar pattern as photosynthetic rates per cell (Table 1).P. anceps had the highest average photosynthetic rates per chlorophyll a (gross and net), significantly higher than the two Pseudoplexaura species, but similar to E. tourneforti.P. wagenaari had significantly lower rates than P. porosa (Table 1).Differences in the initial slope of the P-E curve, a, mirrored differences in photosynthetic rates per chlorophyll a, as a was significantly greater for P. anceps than all other species (Table 1).a of E. tourneforti was significantly greater than in both Pseudoplexaura species, which had statistically similar a.

Oxygen flux of isolated Symbiodinium
Photosynthetic rates of isolated Symbiodinium were lower than the in hospite photosynthetic rates except for Symbiodinium from E. tourneforti (Table 2).With the exception of Symbiodinium from E. tourneforti, isolated Symbiodinium had gross P max /R less than or equal to one (Table 2).The initial slopes of the P-E curves (a) were greater for isolated Symbiodinium (Table 2) than for Symbiodinium within gorgonian branches (Table 1).

Symbiodinium density
P. wagenaari had the highest Symbiodinium densities and these differed significantly from the Symbiodinium densities in P. anceps and E. tourneforti (Figure 2A, Table 3).P. anceps had the lowest Symbiodinium densities.

Chlorophyll content
P. porosa and P. wagenaari had significantly more chlorophyll a and c 2 per cm 2 than P. anceps and E. tourneforti (Table 3).On the other hand, P. porosa, P. wagenaari, and P. anceps had statistically similar chlorophyll a per Symbiodinium cell, with Symbiodinium in E. tourneforti having significantly lower chlorophyll a per cell than all other Symbiodinium (Figure 2B, Table 3).While chlorophyll c 2 per Symbiodinium cell exhibited a similar pattern to chlorophyll a per cell, the statistical results were slightly different: Symbiodinium in P. porosa had significantly greater chlorophyll c 2 per cell than those in P. anceps (Figure 2B, Table 3).Symbiodinium in E. tourneforti had lower chlorophyll c 2 per cell than Symbiodinium in both Pseudoplexuara species, but similar chlorophyll c 2 per cell as those in P. anceps (Figure 2B, Table 3).Symbiodinium in P. porosa and P. wagenaari had statistically similar ratio of chlorophyll a to chlorophyll c 2 to those in P. anceps and E. tourneforti, but Symbiodinium in E. tourneforti had a significantly lower ratio of chlorophyll a to chlorophyll c 2 than those in P. anceps (Table 3).
Chlorophyll a specific absorption P. porosa had the highest estimated absorbance at 675 nm (D c 675 ), which was significantly greater than that in P. anceps and E. tourneforti, but similar to P. wagenaari (Figure 3A, Table 3).Consistent with chlorophyll density data, chlorophyll a specific absorption (a* chl a ), was significantly higher in P. anceps and E. tourneforti than in the two Pseudoplexaura species (Figure 3B, Table 3).P. wagenaari had significantly lower a* chl a than all other species.
Since at least one colony from all the four studied gorgonian species hosted Symbiodinium type B1, we compared these B1 Symbiodinium amongst the four gorgonian species in our study and to published B1 sequences from other cnidarians.Analysis of microsatellite flanking region sequences revealed that Symbiodinium B1 from 16 of 19 gorgonian colonies, representing all four gorgonian species, formed a phylogenetic group with high posterior probability (Figure 4).The B1 Symbiodinium in this group included Symbiodinium from eight of nine P. anceps colonies, one colony each of E. tourneforti and P. porosa and all six P. wagenaari colonies.B1 Symbiodinium from three gorgonian colonies (from P. anceps and P. porosa) placed outside this group, but did not cluster with the examined Symbiodinium B1 lineages from scleractinian corals.
Oxygen flux data demonstrated differences in the Symbiodinium in the four studied gorgonians.For example, the two Pseudoplexaura species had the highest average photosynthetic rates per cm 2 , probably due to the higher Symbiodinium and chlorophyll densities compared to P. anceps and E. tourneforti.On the other hand, the two Pseudoplexaura species had lower initial slopes of the P-E curves, low photosynthetic rates per Symbiodinium cell and per chlorophyll a, and lower chlorophyll a specific absorption.Taken together, these data suggest that Symbiodinium in P. porosa and P. wagenaari are less efficient in light absorbtion and utilization than Symbiodinium in P. anceps and E. tourneforti.The light levels available for Symbiodinium could differ due to Symbiodinium self-shading or to host tissue characteristics such as tissue thickness [53,54] or pigmentation.
The possibility that Symbiodinium in P. porosa and P. wagenaari are less efficient in light absorbtion and utilization is corroborated by looking at changes in chlorophyll a specific absorption coefficient as a function of chlorophyll a density.In the two Pseudoplexaura species, the a* chl a values as a function of chlorophyll a density were very low, comparable to those values reported for phytoplankton and freshly isolated Symbiodinium from Porites banneri [45,55].On the other hand, E. tourneforti a* chl a values fell within those previously reported for scleractinian corals.Lastly, the a* chl a values of Symbiodinium in P. anceps demonstrated a very high pigment light absorption efficiency, comparable to that of Symbiodinium in the scleractinian coral Porites banneri [45].
Of the four gorgonian species, Symbiodinium in P. anceps exhibited twice the photosynthetic rate per Symbiodinium cell than that of the next gorgonian species (P.porosa) and the highest average photosynthetic rates per chlorophyll a.The relatively high photosynthesis per Symbiodinium in P. anceps may be related to the low density of Symbiodinium and chlorophyll a per cm 2 .In addition, the thin, angular branches and low polyp density may aid Symbiodinium photosynthesis in P. anceps by maximizing gas exchange and/or reducing self-shading of Symbiodinium.
The four Caribbean gorgonian species produced comparable photosynthetic and respiration rates per cm 2 to the average rates of eight shallow scleractinian coral species (P = 2.0, R = 0.64) [39].Conversely, the Mediterranean gorgonian Eunicella singularis at 15 m depth had lower average Symbiodinium photosynthetic (,1) and respiration (,0.55) rates per cm 2 [56].The differences could be due to E. singularis in deeper waters being exposed to lower irradiance levels than those in the current study [56,57].In our study, the four gorgonian species did not exhibit photoinhibition, similar to what has been observed in other symbioses between Symbiodinium and cnidarians [58].The lack of photoinhibition may be due to branch tissue thickness, as was seen in the octocoral Capnella gaboensis [53,59].The gross P/R ratios in the four gorgonian symbioses, ranging from 2 to 4, were comparable to ratios for other octocorals [60][61][62][63], anemones [52], and shallow water scleractinian corals [39].On the other hand, the ratios of diurnal integrated gross photosynthesis to respiration were below 1.It remains to be determined the extent of the contribution of the Symbiodinium autotrophic production to the energy budget of these Caribbean gorgonians.
Symbiodinium photosynthesis within a host may also be affected by host characteristics [54,64].For example, the scleractinian coral skeleton enhances light absorption by Symbiodinium [45,65], and Symbiodinium chlorophyll a specific absorption is higher in symbiosis than in isolation [45].Chlorophyll a specific absorption of Symbiodinium in the studied gorgonian corals was comparable to that in scleractinian corals [45,54], even though gorgonian skeletal structure (sclerites and an axial rod) substantially differs from the calyx structure in scleractinian corals.The calcite sclerites  [47], from scleractinian and hydrozoan corals [36], as well as Symbiodinium minutum from Aiptasia, a sea anemone [47].Branch tips are labeled with host species and sample sizes when n.1.Gorgonian and scleractinian coral species are shown in black and red, respectively, and the other cnidarians are shown in blue.B1 lineages described by Finney et al. [36] are listed besides the host taxa.Numbers above the branches are the posterior probability above the maximum likelihood consensus support for each group.B1 Symbiodinium from 16 of 19 gorgonian colonies sampled clustered in a phylogenetic group with high posterior probability (top gray box).Three gorgonian colonies were placed outside of this clade (bottom gray box) and were most closely related to Symbiodinium isolated from Pseudoplexaura porosa from Florida (indicated with (+1)) and cultured Symbiodinium from Gorgonia ventalina [47] indicated with *. (#) indicates a group recovered in the maximum likelihood tree, but not the Bayesian phylogenetic tree.doi:10.1371/journal.pone.0106419.g004within gorgonian tissues may produce the same effect as the light scattering of the scleractinian coral skeleton, perhaps similar to the influence of the siliceous spicules of sponges on light transmission [66,67].
Furthermore, in symbioses between cnidarians and Symbiodinium, photosynthesis is dependent upon the genetic identities of both the host and symbiont [52].Therefore, it is imperative to identify the Symbiodinium.In the Caribbean, scleractinian corals host Symbiodinium belonging to clades A, B, C, or D [34,36].On the other hand, the majority of Caribbean gorgonian species host only clade B Symbiodinium [33,68].The four gorgonian species in this study were no exception; harboring clade B Symbiodinium belonging to types B1, B1i, and B1L.Types B1i and B1L have only been reported in Caribbean gorgonian species [36,69].Sequencing of the B1 Symbiodinium from the four gorgonian species revealed that they harbor B1 Symbiodinium with distinct microsatellite flanking region sequences from those in scleractinian coral species.The B1 Symbiodinium obtained from most gorgonian colonies in our study formed a distinct group amongst the previously identified B1 lineages from scleractinian corals (Figure 4).
Although both P. anceps and P. wagenaari hosted type B1 Symbiodinium, the photosynthetic characteristics differed between these two symbioses.Photosynthetic variability within Symbiodinium type B1 has also been observed in cultures [70], and may be associated with distinct genetic lineages within type B1 [35,36].In our study, eight of the nine P. anceps colonies, and all P. wagenaari colonies, harbored symbionts from the same, highly supported, phylogenetic group within B1 Symbiodinium.Therefore, the observed photosynthetic variability between P. anceps and P. wagenaari was not due to different B1 lineages but probably due to the physiology of different host-symbiont combinations [52].
In E. tourneforti, Symbiodinium had comparable photosynthetic rates per cell in the intact symbiosis and in isolation.On the other hand, maximum photosynthetic rates per cell in Symbiodinium isolated from P. anceps, P. porosa, and P. wagenaari were lower than those measured in hospite, although the average a was higher in isolation.Diminished photosynthetic rates in isolated Symbiodinium cells may occur in the absence of host carbon concentrating mechanisms [71] or differences in carbonic anhydrase activity among Symbiodinium types [72].Exposure to the ionic environment of seawater [73,74] or bacteria [75] may also reduce photosynthesis in isolated Symbiodinium.In the sea anemone Aiptasia pallida, at ambient temperatures, the photosynthetic rates of isolated Symbiodinium were also lower compared to the intact association [52], although not to the degree measured here.Conversely, secondary metabolites released from homogenized gorgonian corals may impair isolated Symbiodinium, as reported for the homogenate of the soft coral C. gaboensis that lysed Symbiodinium cells [76].Therefore, secondary metabolites produced by many gorgonians may limit the utility of investigating freshly isolated Symbiodinium.
In conclusion, our results contribute to consequential data on Symbiodinium physiological performance in their mutualism with four Caribbean gorgonian species at ambient temperature.Given that gorgonian corals are either maintaining or increasing their abundance on Caribbean coral reefs, understanding aspects of their symbiosis is imperative to understanding the future of Caribbean coral reefs.This study demonstrates differences between Symbiodinium photosynthetic characteristics in the four gorgonian species, collected from the same site, maintained under identical conditions, and with two of the gorgonian species containing the same Symbiodinium type.The differences observed between the gorgonian symbioses emphasize the influence of the host physiology and architecture on Symbiodinium photosynthesis.

Supporting Information
Figure S1 Denaturing gradient gel electrophoresis gel of ITS2 DNA from Symbiodinium associated with Pseudoplexaura porosa, Pseudoplexaura wagenaari, and Pterogorgia anceps.Type B1 Symbiodinium was recovered from all colonies of P. wagenaari and P. anceps, but from only 3 of 9 P. porosa colonies.B1 Symbiodinium exhibited distinct DGGE profiles.($) denotes a band that is faint in these samples, but is typically visible in B1 Symbiodinium from P. anceps.(*) denotes the band of the type B1i ITS2 sequence variant.(PDF)

Figure 2 .
Figure 2. Symbiodinium parameters in four Caribbean gorgonian species.(A) Cell densities and, (B) Concentration of chlorophylls a (circles) and c 2 (squares) per Symbiodinium cell.Points represent sample means 6 standard error.Gorgonian species that do not share a letter are significantly different from each other in either density or chlorophyll a per cell (a = 0.05, see Figure 1 for full species names and sample sizes).See Table 3 for significant differences in chlorophyll c 2 per cell.doi:10.1371/journal.pone.0106419.g002 sizes for each gorgonian species (see Figure 1 legend for full genus names) are given in parentheses next to the species name.Parameters include the abundant Symbiodinium (Sym) internal transcribed spacer region two (Sym ITS2) type, Sym cell density, chlorophyll content (chl), estimated absorbance at 675 nm (D e 675 ), as well as the pressure over photosystem II (Q m ).Table cells contain the sample mean 6 standard deviation.The mean square (MS), F statistics (F), and significance value (p) from a one-way ANOVA, using gorgonian species as a factor, are shown for each variable.All variables were transformed prior to conducting the ANOVA.Means with different superscript letters were statistically different (a = 0.05) and this was diagnosed by Tukey's HSD post hoc tests.Power (1-b) is shown for non-significant results.doi:10.1371/journal.pone.0106419.t003

Figure 4 .
Figure 4.A maximum likelihood phylogenetic tree based on microsatellite flanking regions of B1 Symbiodinium.The phylogeny includes B1 Symbiodinium from the four gorgonian species in this study (highlighted in gray), from other gorgonian corals[47], from scleractinian and hydrozoan corals[36], as well as Symbiodinium minutum from Aiptasia, a sea anemone[47].Branch tips are labeled with host species and sample sizes when n.1.Gorgonian and scleractinian coral species are shown in black and red, respectively, and the other cnidarians are shown in blue.B1 lineages described by Finney et al.[36] are listed besides the host taxa.Numbers above the branches are the posterior probability above the maximum likelihood consensus support for each group.B1 Symbiodinium from 16 of 19 gorgonian colonies sampled clustered in a phylogenetic group with high posterior probability (top gray box).Three gorgonian colonies were placed outside of this clade (bottom gray box) and were most closely related to Symbiodinium isolated from Pseudoplexaura porosa from Florida (indicated with (+1)) and cultured Symbiodinium from Gorgonia ventalina[47] indicated with *. (#) indicates a group recovered in the maximum likelihood tree, but not the Bayesian phylogenetic tree.doi:10.1371/journal.pone.0106419.g004

Table 1 .
Photosynthesis-irradiance (P-E) curve parameters for gorgonian species.Sample sizes for each gorgonian species are given in parentheses next to the species name (see Figure1legend for full genus names).P and R represent photosynthetic and respiration in represents the initial slope of the P-E curve.Table cells contain the sample mean 6 standard deviation.The mean square (MS), F statistics (F), and significance value (p) are from a one-way ANOVA using gorgonian species as a factor.All variables were transformed prior to conducting ANOVA.Means with different superscript letters are statistically different (a = 0.05).Power (1-b)

Table 2 .
Photosynthesis-irradiance (P-E) curve parameters for isolated Symbiodinium.Sample sizes for Symbiodinium isolates from each gorgonian species (see Figure1legend for full genus names) are given in parentheses next to the species' name.P . a is the initial slope of the P-E curve.Table cells contain the sample mean 6 standard deviation.The mean square (MS), F statistics (F), and significance value (p) from a one-way ANOVA using gorgonian species as a factor are shown for each variable.All variables were transformed prior to conducting the ANOVA.Means with different superscript letters are statistically different (a = 0.05).doi:10.1371/journal.pone.0106419.t002