Temporal variation and photochemical efficiency of species in Symbiodinaceae associated with coral Leptoria phrygia (Scleractinia; Merulinidae) exposed to contrasting temperature regimes

The Symbiodinaceae are paradoxical in that they play a fundamental role in the success of scleractinian corals, but also in their dismissal when under stress. In the past decades, the discovery of the endosymbiont’s genetic and functional diversity has led people to hope that some coral species can survive bleaching events by associating with a stress-resistant symbiont that can become dominant when seawater temperatures increase. The variety of individual responses encouraged us to scrutinize each species individually to gauge its resilience to future changes. Here, we analyse the temporal variation in the Symbiodinaceae community associated with Leptoria phrygia, a common scleractinian coral from the Indo-Pacific. Coral colonies were sampled from two distant reef sites located in southern Taiwan that differ in temperature regimes, exemplifying a ‘variable site’ (VS) and a ‘steady site’ (SS). We investigated changes in the relative abundance of the dominant symbiont and its physiology every 3–4 months from 2016–2017. At VS, 11 of the 12 colonies were dominated by the stress-resistant Durusdinium spp. (>90% dominance) and only one colony exhibited co-dominance between Durusdinium spp. and Cladocopium spp. Every colony displayed high photochemical efficiency across all sampling periods, while showing temporal differences in symbiont density and chlorophyll a concentration. At SS, seven colonies out of 13 were dominated by Cladocopium spp., five presented co-dominance between Durusdinium spp./Cladocopium spp. and only one was dominated by Durusdinium spp. Colonies showed temporal differences in photochemical efficiency and chlorophyll a concentration during the study period. Our results suggest that VS colonies responded physiologically better to high temperature variability by associating with Durusdinium spp., while in SS there is still inter-colonial variability, a feature that might be advantageous for coping with different environmental changes.


Introduction
The success of coral reefs in tropical oligotrophic waters is often attributed to the symbiotic relationship between scleractinian corals and dinoflagellates algae. The Symbiodiniaceae are intracellular photosynthetic organisms that supply up to 95% of the coral host's energy requirements in optimum conditions [1]. However, this subtle relationship is commonly disrupted under stressful environmental conditions, such as abnormally high seawater temperatures, in what is known as coral bleaching. Due to climate change and anthropogenic increases in seawater surface temperatures, bleaching events are becoming more frequent and severe [2]. Many coral reefs worldwide are now recurrently affected by mass bleaching events and mortality and, consequently, may become scarce within the next 20-30 years [3]. One wellknown asset that some coral species have to survive bleaching events is their ability to associate with a functionally diverse community of symbionts, and to adjust their relative abundances to favour those better fitted to endure temperature variations [4][5][6].
The family Symbiodiniaceae is highly diverse and corals typically associate with members from the genera Symbiodinium (formerly Clade A), Breviolum (formerly Clade B), Cladocopium (formerly Clade C), and Durusdinium (formerly Clade D) [7]. Durusdinium spp. are extremophile endosymbionts that have been found in stressed habitats, such as those with high temperature [8][9][10][11], high turbidity [8] or located in high latitudinal marginal reefs [12][13][14]. In order to survive, some corals are able to increase the relative abundance of Durusdinium trenchii during and after bleaching events [15]. Some examples of this mechanism are found in species such as Acropora millepora in Australia [16]; Orbicella annularis, Siderastrea siderea, Agaricia spp., and Montrastraea cavernosa in Barbados [17]; and Isopora palifera in Taiwan [10]. It is a species-and location-specific mechanism that changes with environmental conditions [15,18,19]. Some examples include Pocillopora damicornis, Seriatopora hystrix, Stylophora pistillata, Favites abdita, Goniastrea favulus, A. millepora, I. palifera in Australia [18] and O. faveolata and O. annularis in the Florida Keys and the Bahamas [19], which did not change their symbiotic composition during and after thermal stress. Because the systematics of the Symbiodinaceae family is still a work in progress, there are many dinoflagellates that have not yet been identified to the species level. Therefore, the former sub-clade or type is used after the genera. Within the Cladocopium genus there are some species that have been recognized to be thermally resistant, such as Cladocopium C15. In Australia, Porites lutea colonies revealed that, when experimentally heated, corals hosting Cladocopium C15 maintained higher maximum photochemical efficiency (Fv/Fm) than those hosting Cladocopium C3 [20]. Similar results were found in Porites lobata from Hawaii; colonies that were experimentally bleached could maintain gross photosynthetic rates similar to control colonies when hosting Cladocopium C15. Moreover, this thermally tolerant symbiont helped the coral recover faster from bleaching by contributing 96% of the host's daily metabolic demand, even when chlorophyll a levels were significantly lower than in the control [21].
The capacity to associate with multiple Symbiodinaceae genera is considered a widespread phenomenon [22,23]. Some coral species can maintain a stable association with their dominant symbiont across their lifetime [24], including during and after stress events [18,19]. Alternatively, some species are capable of shifting the relative abundance of their dominant partner to background symbionts when exposed to stress [5,6,25]. Those background symbionts are usually considered to initially represent <10% of the overall Symbiodinaceae community within one colony [26,27]. A third scenario, documented for a very few number of species is to have a different symbiont dominating distinct areas within the coral colony (Table 1). In this study, we refer to multi-symbiont dominance on the colony scale, but each microhabitat within the colony can have its own dominant symbiont. For example, the dominant symbiont may be different between the top part of the colony and the lowest part. In the Caribbean, this has been documented in the Orbicella spp. complex (including O. annularis, O. faveolata and O. franski) since the 90's (Table 1) [28][29][30][31]. These massive corals associate with multiple dominant symbionts in response to different light gradients, creating different microhabitats within the same coral colony. In O. faveolata-the top part of the colony, with high-irradiance-was dominated by Symbiodinium sp. and Breviolum sp., while the side or shaded parts of the colony were dominated by Cladocopium sp. [29,31]. The same pattern was observed between colonies living at different depths: those living in shallow waters with high light intensity were dominated by Symbiodinium sp. and Breviolum sp., whereas Cladocopium sp. dominated those colonies living in deep waters [29]. In the Pacific, Isopora palifera colonies from southern Taiwan presented multi-symbiont dominance between Cladocopium C3 and D. trenchii (Table 1) [10,32]. The relative abundance of both symbionts varied across the study period and D. trenchii became highly abundant in some colonies after the  [32]. The relative abundance of D. trenchii decreased afterwards and Cladocopium C3 became dominant almost 10 years after the bleaching event [10]. These differences in the dominant symbiont between Symbiodiniaceae genera explain why certain colonies, or parts of the colony, bleach and some do not during natural bleaching events [29]; they also explain how Symbiodiniaceae can cope with and recover from a bleaching event by shuffling their relative abundance [10]. These species are able to cope with environmental fluctuations, presenting a long-term ecological and evolutionary coral-Symbiodinaceae specialization strategy [29,31].
In the present study, we describe the temporal dynamics of Cladocopium spp. and Durusdinium spp. associated with the sub-massive brain coral Leptoria phrygia (Ellis and Solander, 1786) in Kenting National Park, southern Taiwan. We monitored symbiont community changes and performance in coral colonies from two shallow reef flats exposed to contrasting seawater temperature regimes. The average shallow water temperatures at the 'Outlet' reef flat can be up to 2-3˚C higher than other coral reef sites during the summer [39,40]. Furthermore, a tidal upwelling [41,42] causes daily temperature fluctuation that can reach 6-8˚C; we considered this our "variable site" (VS). Temperatures at the 'Wanlitong' reef flat are steadier (daily seawater temperature fluctuations <3˚C); we considered this our "stable site" (SS). Our specific objectives were to: (1) investigate whether the relative abundance of endosymbionts changes are in line with temporal changes in local environments from both sites, and (2) characterise the dynamics of the endosymbionts' physiology (photochemical efficiency, symbiont density, and chlorophyll a concentration) caused by these temporal changes.

Study sites and temperature data
This study was conducted with proper permissions and permits (Field permit number: 1040008112) issued by the Kenting National Park authority.
Both Variable Site (VS) and Stable Site (SS) are located within Kenting National Park (KNP) in southern Taiwan (Fig 1). They are less than 20 km apart but have very different seawater temperature regimes. VS is located next to a nuclear power plant outlet in Nanwan Bay (21˚55' 53.7" N-120˚44' 42.7" E), where it is influenced not only by the hot-water effluent from the power plant throughout the year, but also by a spring tide upwelling which reduces the daily seawater temperature [40,42]. SS is located outside Nanwan Bay (21˚59' 43.9" N-120˚42' 23.2" E) and is protected from the effects of the upwelling and the thermal pollution from the nuclear plant [11]. Leptoria phrygia is a common species at both sites, found in shallow waters and associated with Durusdinium trenchii in VS and Cladocopium C1 in SS [11]. Similarly, other species such as Isopora palifera and Platygyra verweyi living in VS were dominated by D. trenchii and associated with a mixture of D. trenchii and Cladocopium spp. or only with Cladocopium spp. at other sites in KNP [10,11,43,44].
Temperature loggers (Onset HOBO 64K Pendant data loggers, accuracy ±0.5˚C, resolution 0.14˚C, USA) were deployed at a depth of 3 m at each site and temperatures were recorded at 1 hr intervals from June 2016 to June 2017. The mean monthly temperature, mean monthly maximum and minimum temperatures and mean monthly temperature variability (defined as the mean monthly maximum temperature minus the mean monthly minimum temperature) were calculated at each site for comparison.

Sample collection, photochemical efficiency and preservation
Twelve and 13 large colonies were randomly selected at VS and SS, respectively, and tagged with a minimum distance of 7-10 m among them to avoid picking up clones. Coral colonies were sampled at an interval of 3-4 months (in August, December 2016 and March 2017) to characterize seasonal differences typifying the tropical monsoon climate of Kenting. Five nubbins were collected from each colony using a pneumatic drill connected to a regulator on a scuba diving tank and equipped with a core bit (; 2.8 cm). Nubbins were sampled at distances of 10-20 cm apart longitudinally from the top to the bottom of the colony (top part of the colony, top middle part, middle part, bottom middle part and bottom part). All collected nubbins were transported in coolers filled with seawater to the wet laboratory of the nearby National Museum of Marine Biology and Aquarium (NMMBA) and placed in tanks under flowing seawater to emulate field conditions. Dark-adapted maximum photochemical efficiency (Fv/Fm) of PSII was quantified using a diving-PAM (Heinz Walz GmbH, Germany; settings: Saturating intensity = 8, Saturating width = 0.8, Gain = 4, Damping = 2), keeping a fixed distance between the fibre and the sample by attaching a 1 cm probe extension. Dark-adapted (Fv/Fm) measurements were done approximately two hours after sunset, due to logistical constraints. A test was performed to see if (Fv/Fm) values after two hours were different from those after only 30 minutes for L. phrygia; results showed similar values. Nubbins were then immediately snap-frozen in liquid nitrogen and stored at -80˚C for transportation and further processing.

Laboratory analyses
Each coral sample was defrosted, inspected for contaminants, and a small subsample of coral tissue (~6 mm) was cut and stored in an Eppendorf tube with 95% Ethanol for molecular analysis. Remaining coral tissue was then air-brushed [45] from the skeleton using approximately 15 ml of filtered seawater (0.2 μm). Resulting slurry (Vi) was homogenized using a Homogenizer Stirrer (WiseStir HS-30E, Germany).

Surface area
The surface area of the remaining coral skeleton was measured using a 3D scanner (HP David SLS-2 3D structured light scanner Pro S3, resolution/precision up to 0.1% of scan size down to 0.06 mm, max mesh density: 1,200,000 vertices per scan; USA) equipped with a 360˚turntable and area calculations were done using the open source software Meshlab [46]; data were presented in cm 2 .

Symbiont density
Symbiont density was determined from a 10% aliquot of Vi by counting cells with 6-8 replicates per sample using a Neubauer hemocytometer (Assistant, Germany) under a light microscope (BX40 Olympus, Japan). This cell count was then normalized to the surface area of the nubbin, and presented as number of symbiont cells per cm 2 .

Chlorophyll a concentration
Another 10% aliquot of Vi was centrifuged at high speed (5 min, 14000 xg). The supernatant was removed; 1 ml of 90% acetone was added to the pellet and incubated in total darkness at -20˚C overnight for photosynthetic pigment extraction [47]. After centrifugation for 1 min at 10000 xg, supernatant absorbance was read at 630, 647 and 664 nm wavelengths from triplicate aliquots of 200 μl each, using a spectrometer (SpectroStar nano absorbance reader BMG-Lab-Tech, Germany) with three wells containing 200 μl of 90% acetone as a blank. The chlorophyll a concentration was then calculated using the equations from Jeffrey and Humphrey [48]: Chla = 11.85 A 664 −1.54 A 647 −0.08 A 630 . Chlorophyll a concentration was then normalized per cm 2 of the nubbin's surface area and the symbiont density and presented as pg of chlorophyll a concentration per symbiont cell per cm 2 .

Molecular analysis
DNA extraction. Genomic DNA was extracted using a modified high salt method [49]. Briefly, for tissue incubation, 30 mg of coral tissue (~3-4 polyps) was cut and incubated overnight at 55-60˚C using 200 μl lysis buffer (1M Tris-Boric 25 ml, 0.5M EDTA pH8 10 ml, 20% SDS 10 ml, 5M NaCl 2 ml, ddH 2 O 53 ml) and 10 μl proteinase E (10 mg/ml). For DNA precipitation, 210 μl 7M-NaCl was added, vortexed for 30 seconds at maximum speed and centrifuged for 30 minutes at 10000 xg. The supernatant was transferred to a new tube and 420 μl of 100% isopropanol were added and gently mixed for five minutes. Samples were incubated at -20˚C for at least two hours. For DNA purification, samples were centrifuged (30 minutes at 16000 xg) and rinsed with 150 μl of 70% ethanol stored at -20˚C and centrifuged again (5 minutes at 16000 xg); this rinsing step was repeated three times. Ethanol was removed from the tubes, which were dried under the hood for 1-2 hours. The DNA was eluted with 150 μl of preheated (65˚C) 1X TE buffer. The total concentration of genomic DNA was determined using a NanoDrop 2000 (Thermal Scientific, USA).

Quantitative PCR (qPCR).
Relative abundances of Cladocopium sp. and Durusdinium sp. were measured by amplifying the ITS1 region using the LightCycler 480 Instrument II (Roche, Switzerland) with a modified protocol from Mieog et al. [26]. The following primers were used: nuclear ITS1 universal forward primer (UF, 5'-AAGGAGAAGTCGTAACAAGGTTTCC-3'), nuclear ITS1 C-specific reverse primer (CR, 5'-AAGCATCCCTCACAGCCAAA-3'), and ITS1 D-specific reverse primer (DR, 5'-CACCGTAGTGGTTCACGTGTAATAG-3') [34]. Each 10 μl qPCR reaction consisted of 5 μl of 1x SYBR Fast Master Mix, 0.5 μl of UF primer (2 nM/μl), 0.5 μl of CR or DR primer (2 nM/μl), 1.5 μl of ddH 2 O and 2.5 μl of DNA templates (equal to 1 ng of genomic DNA). The two-step qPCR reactions were set at 95˚C for 15 seconds and 60˚C for one minute. 40 cycles were performed in total. Melting curves were generated to start at 60˚C with an increase of 0.11˚C/s until it reached 95˚C, then followed by a cooling step to 40˚C for 30 seconds. The ratio of Durusdinium (D) to Cladocopium (C) was calculated using the formula: ratio = D/(C+D) from triplicate measures per sample. A correction was done to account for the differences in copy numbers between Cladocopium and Durusdinium genera following Mieog et al. [26].
Symbiont identification-(DGGE). Denaturing gradient gel electrophoresis (DGGE) was used for symbiont identification. The ribosomal internal transcribed spacer 2 (ITS2) region was amplified with the primers ITSintfor2: 5'-GAATTGCAGAACTCCGTG-3', and ITS2clamp: 5'-CGCCCGCCGCGCCCCGCGCCCGTCCCGCGGGATCCAT-ATGCTTAAGTTCA GCGGGT-3' and touch-down PCR protocol [50]. Each PCR product was loaded onto an acrylamide denaturing gradient gel (45-80%) and then electrophoresed at a voltage of 115 for 15 hours (CBS Scientific system, USA). Gels were stained with SYBR gold (Invitrogen, USA) for 30 minutes and photographed for further analysis. The most significant band of each PCR product was cut from the gel and eluted in 100 μl of distilled water for a few hours. Re-amplification was performed with 1 μl of the sample following a protocol similar to touch-down PCR-DGGE, but using a reverse ITS2 primer without the clamp [50], and the product was sent for sequencing. The resulting sequences were trimmed, cleaned and blasted against the database of the National Center for Biotechnology Information (NCBI) to determine Symbiodiniaceae genus.

Statistical analysis
Mean monthly temperature, mean maximum/minimum temperature and mean monthly temperature variability (defined as the mean monthly maximum minus the mean monthly minimum temperature) were calculated for each site and the differences between VS and SS during the study period were tested using the non-parametric Mann-Whitney-Wilcoxon test after confirming that data from most months were not normally distributed. Data from all measured physiological parameters were checked for normality and homogeneity of variances (Q-Q plot and Fligner test) and all were normally distributed and presented homoscedasticity. A linear mixed model was used to test the effect of sampling times as the fixed effect (with three levels) at each site, using colony as the random effect. Similarly, to compare between sites at each sampling time, a linear mixed model was used with site as the fixed effect (with two levels) and using colony as the random effect. P-values were obtained by likelihood ratio test of the full model against the model without the effect in question. This was performed for each of the physiological parameters measured. All samples at both sites were combined and the nonparametric Spearman correlation was used to test for monotonic correlations between each temperature measurement independently (mean monthly temperature, mean monthly maximum temperature, mean monthly minimum temperature and mean monthly temperature variability) and Durusdinium spp. association. All statistical analyses were performed in R (version 3.4.1) [51], using the 'lme4' package for the linear mixed model analysis [52]. All data were presented as mean ± standard deviation (S.D.).

Seawater temperature differences between sites
Both sites displayed different seawater temperature regimes (Fig 2) from June 2016 to June 2017. Mean monthly temperature in VS ranged from 25.4 ± 0.9˚C to 31.0 ± 1.1˚C while in SS it ranged from 23.7 ± 0.7˚C to 29.7 ± 0.6˚C, making it significantly higher in VS than in SS (Mann Whitney test, p<0.05). June 2017 was the only month in which there was no significant difference in mean monthly temperatures between sites (Mann Whitney test, p = 0.58). During the summer of 2016, July exhibited the maximum mean temperature of the year in VS (32.3 ± 0.9˚C), which was significantly different to that in SS (30.4 ± 0.6˚C). In the winter of 2017, January showed the minimum mean temperature of the year in VS (24.6 ± 0.9˚C), which was significantly different from that in SS (23.1 ± 0.7˚C; Table in S1 Table). The mean monthly temperature variability (mean monthly maximum temperature minus the mean monthly minimum temperature) was also significantly different between sites. Mean monthly temperature variability was significantly higher in VS (1.5 ± 0.5˚C to 3.2 ± 1.1˚C) than in SS (0.7 ± 0.4˚C to 1.6 ± 0.8˚C) during the study time (Mann Whitney test, p<0.05) (Fig 2, S1  Table).

Symbiodinaceae association
Among the 12 colonies sampled in VS, Durusdinium spp. dominated in 11 (>90%), and none showed temporal variation across the entire study period. One colony exhibited  Fig 3). Of the 13 colonies at SS, seven were dominated by Cladocopium spp. (>90%), five presented co-dominance between genera in Cladocopium and Durusdinium, and only one was Durusdinium spp.-dominant (>90%; Fig 3). Similar to what was observed in VS, there was no variation in those colonies that were either Cladocopium spp.-or Durusdinium spp.-dominant during the entire study period. Only those five Cladocopium spp./Durusdinium spp. co-dominated colonies presented temporal variability (Fig 3). (For more detailed information see S2 Table).
DGGE and subsequent sequencing of the major bands confirmed that D. glynnii (GenBank accession number MK127922) was the most dominant Durusdinium endosymbiont found at both sites. There was also inter-colonial variation in those Cladocopium spp.-dominated colonies in SS. Cladocopium C3w (GenBank accession number MK127920) and Cladocopium C21a (GenBank accession number MK127921) were the two most dominant endosymbionts found in different colonies (Table 2, Fig 4). This difference between Cladocopium spp. was identified in the DGGE band profiling (Fig 4) and later confirmed by sequencing. Cladocopium C3w showed only one dominant band in the DGGE profile (Fig 4) and was dominant in five colonies (COL4, COL5, COL7, COL9, and COL12; Table 2). Cladocopium C21a displayed two bands (heteroduplex) in the DGGE profile (Fig 4) and was dominant in three colonies (COL6, COL10 and COL11; Table 2).

Correlation between Durusdinium spp. and temperature
We found significant positive monotonic correlation (r s = 0.37, p<0.05) between the mean maximum temperature and the percentage of Durusdinium spp (Fig 5A), as well as between the temperature variability and the percentage of Durusdinium spp. (r s = 0.65, p<0.05) during the study period (Fig 5B).

Discussion
In this study we characterised Symbiodiniaceae communities of Leptoria phrygia from two sites with contrasting temperature regimes. This species has been reported to associate with Cladocopium and Durusdinium in American Samoa [9], and, in KNP, southern Taiwan, L. phrygia is also known to host both genera within the same colony [11]. Our study demonstrates that the stress-resistant Durusdinium spp. dominated almost all colonies in VS. In contrast, coral colonies in SS presented three different types of associations: 1) dominated by Cladocopium (54%), 2) co-dominated by Cladocopium and Durusdinium (39%) and 3) dominated by Durusdinium (7%). Nonetheless, in both VS and SS, those colonies dominated by a single genus-i.e., Durusdinium or Cladocopium only-continued to be dominated by that genus through time. Even the dominant species within those colonies remained the same, e.g., Cladocopium C3w or Cladocopium C21a in the Cladocopium-dominated colonies and D. glynnii in the Durusdinium-dominated colonies.
However, there was distinct temporal variation in colonies with multi-symbiont dominance (co-dominated by two genera), particularly those from SS. Durusdinium can tolerate 1.0 to 1.5˚C higher than Cladocopium in Acropora millepora [5]. Our results showed that there is a significant positive correlation between the mean maximum temperature and the percentage of Durusdinium spp. a phenomenon also documented under natural conditions in previous studies [9,53,54]. We also found significantly higher correlation between high temperature variability and the percentage of Durusdinium spp. That might be the main reason why almost all colonies in VS are associated with the stress-resistant Durusdinium spp. Other studies also found that coral communities living at this thermally variable site were dominated by D. trenchii [10,43,44]. Another survey done at the same site reported that species from the genera Acropora, Cyphastrea, Goniastrea, Isopora, Platygyra, Favites, Pocillopora, Acanthastrea, and Leptoria associated with D. trenchii at 3 m deep in VS, while at 7 m deep the same species associated with Cladocopium spp. (C1, C3, C21a and C15) [11]. This difference in symbiont communities between depths implies that the hot water from the power plant outlet only affects the shallow waters in VS [11].
Our results indicate that, in addition to high temperature, large temperature variability also plays an important role in determining symbionts associations; this concurs with a study of four common taxa sampled from four different sites with a gradient of temperature variability in Kenya over 10 years [55]. The study showed that Pavona and Pocillopora were the genera that best survived bleaching events, including the global bleaching event of 1998; similar to our study, most colonies in these two genera lived in the sites with the highest temperature variability and were associated with Durusdinium sp. [55].
Those colonies in SS presenting intra-colonial variation, plus a Durusdinium spp.-dominated colony (46% of all colonies sampled), have the stress-resistant symbiont that provides higher resilience to bleaching events than other taxa. Those Cladocopium/Durusdinium codominated colonies presented temporal variability in the dominant symbionts, which offers a ground for corals to shuffle the relative abundances of their dominant symbiont [56]. This suggests that L. phrygia's inter-colonial variation might be an ecological advantage, because coral Symbiodinaceae community of Leptoria phrygia species that host symbionts resistant to variable conditions will most likely survive climate warming events [6,23,24,57].
This theory that inter-colonial variation provides ecological advantage has been experimentally simulated with two species in the Caribbean with multi-symbiont associations, Montastraea cavernosa and O. faveolata [6,57]. Corals were stressed with high temperatures to induce bleaching. They recovered by increasing their abundance of the resistant symbiont D. trenchii and maintaining this association after recovery for three months at 29˚C [6,57]. Researchers exposed M. cavernosa corals to a subsequent bleaching heat stress and determined that those D. trenchii-dominated corals were more resistant to bleaching and lost fewer symbionts than before [6].
Durusdinium trenchii has been found to increase bleaching thresholds by 1.0 to 1.5˚C under high temperature stress [5,6,58] by maintaining a high photochemical efficiency compared to other symbionts from the Breviolum and Cladocopium genera [25,59]. Our study suggests that species living in VS are able to resist high summer temperatures and temperature variability by hosting Durusdinium spp. Colonies in VS did not display any signs of stress when their maximum photochemical efficiency was measured, even during the summer, when mean maximum temperatures reached 32.3 ± 0.9˚C and were significantly higher than in SS. All colonies in this highly variable thermal environment showed mean Fv/Fm higher than 0.66 ± 0.03 (see Fig 6A).
In the southern Great Barrier Reef, A. valida, a species dominated by multiple Cladocopium and Symbiodinium symbionts, did not present seasonal variation in photochemical efficiency [37]. But the lack of seasonal variation was explained by changes in the relative abundance of its symbionts, because most colonies increased their percentage of Symbiodinium sp. after a bleaching event in the summer. Therefore, they maintained high photochemical efficiency during the winter [37]. In contrast, SS colonies varied their photochemical efficiency through sampling times by increasing Fv/Fm during the winter. Similar results were found in the Caribbean in Orbicella spp. [60] and in the South China Sea in three Acropora spp., Pavona decussata and Porites lutea [61]. Durusdinium spp. not only are advantageous to corals under high temperature, they can also resist cold temperature stress [62].
M. cavernosa corals dominated by either D. trenchii or Cladocopium C3 were experimentally exposed to temperatures down to 15˚C and, when measuring the symbiont photochemical efficiency, both coral colonies showed photodamage. Those corals dominated by D. trenchii neither bleached nor died, while Cladocopium C3-dominated corals bleached, losing 94% of their symbionts [62]. Similarly, Pocillopora spp. corals associated with D. glynnii better survived a cold bleaching event in 2008 in the Gulf of California [63]. All colonies associated with Cladocopium C1b-c bleached during the cold event and 56% suffered partial or total mortality, whereas those colonies associated with D. glynnii neither bleached nor showed any mortality [63].
During our study period, there was no sign of bleaching from stress at either site, and corals in VS exhibited an increase in symbiont density during December 2016 following natural temporal fluctuations, which caused a significant difference between sites. In SS, there was high variability between colonies in 2016, therefore mean values of symbiont density in December were comparable to those in August.
Similar results as in VS were found for A. muricata in a long-term monitoring study (6 years) in Mauritius: symbiont densities were three times higher during the autumn and winter than the spring and summer, and the authors found that symbiont density positively correlated with nitrate concentrations [64]. In the Caribbean, a long-term monitoring study (4 years) found similar results: A. cervicornis, A. palmata, O. annularis and O. faveolata presented higher values during the winter [47]. A study in the Red Sea described this same seasonal variation in P. verrucosa, with higher densities during the winter [65].
This difference in symbiont density between seasons has been explained by dynamic photoinhibition, where carbon fixation decreased due to down-regulation of photosynthesis when there is high light, such as during the summer [47]. Additionally, chlorophyll a concentration values showed temporal fluctuations due to this dynamic photoinhibition related to high light and temperature during the summer [60,66,67]. Both sites in our study presented lower values during the summer and increased values in the winter.
In summary, understanding the physiological plasticity of a species such as Leptoria phrygia living in different environments allows us to establish different mechanisms that species could use to withstand future climate change. Here, we examined the multi-symbiont association in L. phrygia in a stable environment and compared it to the same species living at a site with high temperature variability. Our results suggest that those corals living in variable environments provide important information on coral resilience during environmental perturbations. Coral colonies in VS dominated by Durusdinium spp. had better physiological responses and were able to cope better with the high variability in seawater temperature. These results raise the question of whether the symbiont community of coral species such as L. phrygia may have been selected by the environmental conditions they live in. In addition, only those Durusdinium/Cladocopium-dominated colonies presented temporal variability. No matter the environmental conditions they lived in, those Durusidinum spp.-and Cladocopium spp.-dominated colonies maintained the same dominant symbiont at all times. These results also question the role of the coral host in deciding which Symbiodinaceae community to associate with. In order to answer these questions and to have a better understanding of how highly variable environments will respond to future climate change, we recommend further experimental work between both sites, such as transplantation experiments; we also recommend measuring the physiological parameters of all partners in the holobiont. It will also be interesting to investigate if species already dominated by Durusdinium spp. will be able to survive more recurrent bleaching events [2] or prolonged thermal stresses [44] in the future. would like to thank Andy Zhang and Weber Gao for their help during fieldwork, all members in the Coral Reef Evolutionary Ecology and Genetics Laboratory that contributed to this study, Dr. Pei-Jie Meng at the National Museum of Marine Biology and Aquarium for providing space in his wet laboratory to keep the corals and the Kenting National Park for issuing collection permits.