Introduction

Dinoflagellates in the family Symbiodiniaceae (formerly the genus Symbiodinium) comprise a highly diverse group of unicellular eukaryotic algae, most of which form an endosymbiotic association with a wide range of marine invertebrates, including scleractinian corals (Trench 1997; LaJeunesse et al. 2018). Recent studies have estimated that the timing of the earliest diversification of extant lineages of Symbiodiniaceae occurred during the mid-Jurassic period. This timing coincides with a second and much larger adaptive radiation of modern shallow water corals (Stanley and Fautin 2001; Simpson et al. 2011; LaJeunesse et al. 2018). These endosymbioses have shaped coral holobiont biology by giving rise to specific adaptations, as evident, for example, the mode of intracellular reproduction and metabolism of the symbiont (Fitt and Trench 1983; Liu et al. 2018) and the host’s ability to use translocated photosynthates to meet most of its energy requirements and exhibit photosynthesis-enhanced calcification (Muscatine and Porter 1977; Falkowski et al. 1984). Some level of specificity between both entities exists (Trench 1993), which combined with the differential capacity to use light energy by the symbiont, creates niche diversification and largely explains vertical distribution in corals (Iglesias-Prieto et al. 2004).

Coral reefs are among the most diverse ecosystems providing important economic revenue through fisheries, tourism and pharmacology industries, and they also have a roll protecting seashores from erosion (Moberg and Folke 1999). However, coral reef ecosystems thrive under a relatively narrow temperature range (Buddemeier and Kinzie 1976; VeronQuery 1995), and when local temperatures surpass the normal average in conjunction to high light exposure, disruption of the symbioses by exocytosis and loss of light harvesting pigments occur (Weston et al. 2015), leading to coral bleaching (Glynn 1993, 1996; Goreau and Hayes 1994; Fitt et al. 2001). This process can be reversible or catastrophic depending on the geographic location, extent and duration of the perturbation (Boylan and Kleypas 2008; Donner 2011; Sully et al. 2019). It is predicted that more frequent and severe thermal anomalies will occur due to anthropogenic climate change (Spalding and Brown 2015). Therefore, it is imperative to better understand the potential plasticity and variability of the key biological traits of these symbioses under such thermal stress scenarios.

Based on early morphological studies in hospite, it was believed that Symbiodiniaceae algae isolated from different hosts were the same species, i.e., Symbiodinium (= Gymnodinium) microadriaticum Freudenthal (Taylor 1969, 1974). Other studies using a combination of biochemical and morphological characterisation of cultured Symbiodiniaceae found intrinsic differences of genetic origin (Schoenberg and Trench 1980a, b, c). The incongruence of in hospite and in vitro studies along with the relative slow-developing classical morphological descriptions and some difficulties to culture some symbionts led to the development of gene-based approaches using the nuclear Small Subunit ribosomal DNA (SSU rDNA) to characterise the diversity and evolution of Symbiodiniaceae (Rowan and Powers 1991a, b, 1992). These studies showed the existence of a vast genetic diversity, and the existence of divergent evolutionary lineages (Clades A–C, now expanded to up to twelve different genera and clades in the family Symbiodiniaceae (LaJeunesse et al. 2018)). This genetic diversity was potentially linked to different physiology, later demonstrated by the differential response to light intensity of different cultured symbionts (Iglesias-Prieto and Trench 1994, 1997). Since this initial SSU rDNA characterisation, there have been a number of different approaches to elucidate the Symbiodiniaceae genetic attributes by using sequence-based methods using other markers such as the internal transcribed spacer or ITS regions of the ribosomal cistron and chloroplast non-coding regions of the psbAncr gene, coding for de D1 protein involved in photosynthesis (Zurawski et al. 1982; LaJeunesse 2001; LaJeunesse and Thornhill 2011), and sequence-independent techniques such as amplified fragment length polymorphism or AFLP and allozyme electrophoretic patterns (Coffroth et al. 1992; Baillie et al. 1998). The ITS2 marker has been most frequently used and more recently in high throughput sequencing aimed at detecting cryptic symbiont populations as well as improving on the discriminating power for those intragenomic variants of the tandemly repeated ribosomal DNA (Arif et al. 2014; Quigley et al. 2014; Cunning et al. 2017). There are also population studies using polymorphic microsatellites (Thornhill et al. 2017). Along with this uncovered genetic diversity, there is an extensive body of literature showing that physiological diversity among symbiont types is also large, both among members of the same as well as different genera (reviewed in (Quigley et al. 2018)). In line with this, to have a point of comparison with other related protozoa, the expression of distinct phenotypes by the same genotype is a commonly recognised phenomenon among free-living eukaryotic dinoflagellates, with potential implications for growth (Doblin et al. 2000), but may have further implications within the coral holobiont.

In the coral reef community, the Symbiodiniaceae taxa harboured by a coral colony have commonly been interpreted as a measure of its environmental stress tolerance, with physiological variation within types being largely ignored (Rowan 2004; Berkelmans and van Oppen 2006). However, a recent study has demonstrated that large differences in thermal tolerance can exist between allopatric populations of the same symbiodiniacean species (related to Cladocopium goreaui, and referred here as Cladocopium C1acro), both in vitro and in hospite, a result of adaptation of these populations to their local environmental conditions (Howells et al. 2012; Levin et al. 2016). Here, we provide evidence for the existence of variation in the Symbiodiniaceae response to temperature at an even smaller spatial scale, among Cladocopium symbionts extracted from conspecific and sympatric hosts. We successfully adapted a non-sequence-based approach (TE-AFLP; Three Endonuclease Amplified Fragment Length Polymorphism) (van der Wurff et al. 2000) to distinguish individual isolates, with promising future applications for population-based studies of reef ecosystems. Our results highlight that a large portion of the physiological variation present within the Symbiodiniaceae is undocumented and mostly overlooked, and that more caution is required when interpreting species/strain identity based on DNA sequence data of one or a small number of loci in the context of physiological performance.

Methods

Establishment of clonal cultures

Cladocopium C1acro cells were isolated from four coral colonies growing at similar depths and photic habitats from the Central Great Barrier Reef (GBR): two Acropora tenuis (~ 3–4 m depth) and one Acropora millepora (~ 3 m depth) from Nelly Bay, Magnetic Island, and one A. tenuis (~ 3 m depth) from South Molle Island in the Whitsunday Islands (collected under the permit G10-33440.1). Host skeletons as well as living cells are kept at Symbiont Culture Facility, part of the Australian Institute of Marine Science, AIMS. Before the culturing of the symbionts, freshly isolated cells from the host were genotyped. These results were latter contrasted with the cultured dinoflagellates being both made of the same genetic makeup. Isolation of Cladocopium cells was carried out by air brushing the coral tissue off the skeleton, centrifugation (5 min at 1600 g) and resuspension of the pellet in 0.2 µm filtered seawater. Centrifugation and resuspension were repeated three times to remove organic debris. Symbiont cells were inoculated into sterile IMK growing media (Wako Chemicals USA, Richmond) with antibiotics (penicillin, neomycin, streptomycin, nystatin, final concentration [100 µg mL-1] each, amphotericin final [2.5 µg mL-1], plus GeO2 final [50 µM] (Beltran et al. 2012). Cultures were maintained in 24-well plates at 26 °C, 60 ± 10 µmol quanta m-2 s-1 of photosynthetically active radiation (PAR) (light coming from the bottom), 14:10 light: dark photoperiod inside environmental chambers (Steridium e500). After a month of growing in vitro, only those cultures showing no signs of contamination were pooled and re-inoculated in antibiotic-rich media or IMK+. Cells were allowed to grow for one more month before transferring them four more times into fresh IMK+ (total of five months) to minimize bacterial contamination. After this antibiotic treatment, cultures were expanded in volumes of 50 mL antibiotic-free media or IMK −, considering this the source population to establish clonal strains. For the generation of clonal cultures, two methods were used. Three of the four clonal cultures (Aten-MI-1, Aten-WSY, and Amil-MI) were obtained after plating the heterogeneous population onto IMK − 1% agar plus antibiotics and incubating them under the conditions as above and picking a single colony-forming unit. The fourth clonal culture (Aten-MI-2) was produced from a single cell isolated using a micro-manipulator (MN-153, Narishige) under an inverted microscope at 20X magnification (Leica DM IRB/E, Germany).

Genetic characterisation of Symbiodiniaceae cultures

For genotyping, DNA was extracted from exponentially growing cultures (10 mL culture centrifuged as above after which the culture medium was removed) using the Qiagen Plant DNA kit following manufacturer’s protocol. Single-strand conformation polymorphism (SSCP) analysis of the ITS1 region was carried out following the protocol described in van Oppen et al. (2001), and direct sequencing using the dideoxy chain terminator method (Sanger et al. 1977) of the ITS2 and psbAncr regions following Stat et al. (2013) and LaJeunesse and Thornhill (2011), respectively. Sequence data from the chloroplast marker were deposited at the genbank with the accession numbers MW691103-06 for Aten-MI-2, Aten-MI-1, Amil-MI and Aten-WSY, respectively. A phylogenetic analysis of the psbAncr data was conducted using PAUP employing the maximum likelihood algorithm. The cultures were also characterised by TE-AFLP, (van der Wurff et al. 2000) and their products were analysed using GeneMarker Ver 2.6.4 and R studio Ver. 1.2.1.3 for WARD and UPGMA grouping algorithms. We chose these approaches, the first ones to distinguish at the genus and species level, and the last one to be able to distinguish at the individual level (for the latter, assuming asexual reproduction after clonal isolation).

Growth kinetics

The ratchet approach to compare the strains was chosen because we were interested on examining the differences in response to selection. For each strain, 16 falcon flat bottom tubes (50 mL) with a gas permeable PTFE membrane cap (Bioreactor 50, TPP) were inoculated with 2 × 105 cell mL−1 in a total volume of 30 mL. The tubes were incubated upright, light coming from the bottom at 60 ± 10 µmol quanta m−2 s−1 of PAR (Sylvania FHO24W/T5/865 fluorescent tubes) and a 14:10 light: dark photoperiod and 26 °C. To estimate growth, cultures were subsampled (4 replicates of 50 µL per tube) at regular intervals and cell density assessed by measuring absorbance at 750 nm (Pomati et al. 2003) using a plate reader (PowerWave, Bio-Tek). Absorbance was transformed to cell density using a calibration equation per strain (Aten-MI-1 = 13389863.995*OD @ 750 nm, Aten-MI-2 = 8452519.023*OD @ 750 nm, Aten-WSY = 14325213.125*OD @ 750 nm, Amil-MI = 9016121.922*OD @ 750 nm). After approximately 30 days of incubation, the four replicates of each strain showing the highest growth rate were selected for further experimentation. Cells from each replicate were equally divided over 16 tubes and cell densities adjusted as above. These fast 26 °C growing cultures were grown/sampled for the same period of time at 28 °C, selecting the fastest replicates, splitting them in 16 tubes, density adjusted to 2 × 105 cell mL−1 then incubated for a month at 30 °C, repeating the procedure mentioned before incubating at 32 °C the best growing cultures and for those surviving this last temperature, the final incubation growth was 34 °C, with lighting and photoperiod conditions in all “ratchet” steps as above. It is worth mentioning that for the last incubation temperatures, the replicates number were reduced to 8–10 tubes due to the poor growth observed under these stressful temperatures. These experiments were repeated three independent times from May to November of 2011, beginning with a one month old 26 °C culture as starter for the temperature ratcheting. Growth rate was calculated by least-squares fit of a straight line to the data (cell density of four consecutive sampling days), logarithmically transformed, based on the formula: µ = Ln (N1/N0)/t1 − t0 where µ is the growth rate in cell divisions * day−1, and N0, N1 represent initial and final density of cells at t0 and t1, initial and final time, respectively (Guillard 1973). Because temperature could affect the growth kinetics, in our experiment we used two time windows to better characterise the timing of exponential growth. We called phase 1 and phase 2 from day 0 to day 11 (day 0, 3, 6, and 11) and from day 11 to day 18 (day 11, 13, 14, and 18) post inoculation, respectively. Growth rates were statistically compared using two-way ANOVA (temperature and phase), and the post-hoc Tukey test using the R-studio software. After Tukey test, the data generated were used to calculate the heterogeneity values by normalising with all six possible different pair combinations as a 100%.

Stability of thylakoid membranes

For the quantification of the fluidity of the thylakoid membranes, we followed the protocol of Diaz-Almeyda et al. (2011). Cultures growing exponentially at 26 °C and 31 °C (strait from 26 to 31 °C, with one month of acclimation) for one month under the conditions mentioned above were harvested by centrifugation at 1600 g for 5 min and resuspended in fresh sterile IMK − media to a density of 3 × 106 cell mL-1. In addition to the four Cladocopium C1acro cultures we included a heterogeneous culture of Durusdinium trenchii (formerly D1a) (isolated from Acropora muricata at Magnetic Island) to benchmark results with a phylogenetically distinct symbiont species. All cultures were dark adapted for 10 h at their growth temperature (either 26 °C or 31 °C) prior to the measurements. Six replicates of 100 µL each per culture per temperature were tested three independent times. Using a thermocycler (PTC-100, MJ Research), culture aliquots were exposed to 1 °C stepwise temperature increments (5 min each) from 32 to 43 °C under darkness. After the incubation period, maximum quantum yield, Fv/Fm of photosystem II, calculated according to Schreiber et al. (1994) and Warner et al. (1996), was measured (thermocycler heating block within the chamber of the fluorometer) using an imaging-PAM (Maxi-PAM, Walz Gmbh, Effeltrich, Germany). The Fv/Fm in response to temperature measures the uncoupling of energy transfer from the antennae complex to the reaction center as an indication of melting of the chloroplast membrane (Schreiber and Armond 1978). Differences in melting points among strains were statistically compared using a one-way ANOVA, followed by the post-hoc Tukey test using the NCSS software.

Results

Growth kinetics

Temperature-dependent growth performance for each strain shows a progressive decline, in response to temperature increments, as shown by the total biomass towards the end (23–30 days post inoculation) of the incubation period. Aten-MI-1 showed between 33 and 55% more biomass in comparison with the other three strains (Fig. 1). As growth rates were clearly affected by temperature, we decided to analyse in detail the window of time where exponential growth took place. Growth curves of the four clonal strains showed two distinct phases. The first was observed during the first 11 days following inoculation into fresh media, and the second was from day 11 to 18 (Fig. 1, dark and light shaded areas, respectively). By dividing in two windows of time, late and early growth rate in response to temperature can be seen, but also when there were no differences between these two phases we can conclude that for particular temperature the strain shows an extended maximum growth rate meaning that temperature does not have an effect on the timing and duration of growth (see for example Aten-MI-2 and Aten-WSY for the temperatures 26–30 °C, Fig. 2). Maximum growth rates were achieved in either the first or the second phase depending on strain and temperature. For instance, Aten-MI-1 and Amil-MI at 26 °C and 28 °C achieved their exponential growth in the first phase, while Aten-WSY and Aten-MI-2 at 30 °C reached their maximum growth rate in the second phase (Fig. 1). Temperature-dependent growth data were adjusted to different polynomial equations depending on the high coefficient value observed. These equations describe the growth behaviour (Fig. 1, coloured lines as best fit model) at different temperatures (Table 3).

Fig. 1
figure 1

Growth kinetics of four monoclonal Cladocopium C1acro strains at five different temperatures (mean ± SD). Temperature treatments were consecutive, based on a ratchet design (see methods). The shaded areas represent the two windows of time or phases (dark and light gray for phase 1 and 2 respectively) analysed in this study. Growing equations and correlation index per each curve are depicted in Table 3

Full size image
Fig. 2
figure 2

Growth rate comparison among Cladocopium C1acro strains at two different time points (mean ± SD). Dark and light bars represent phase 1 (0–11 days post inoculation) and phase 2 (11–18 days post inoculation), respectively. Significant differences (see Table 1) are depicted as asterisks and letters (red coloured letters against its counterpart in black coloration), representing statistically significant differences between two phases at one given temperature per strain, and differences among all growing temperatures for a particular strain (top of the bars) and differences per phase and temperature among strains (matrix below the chart), respectively

Full size image

Growth behaviour differed significantly among strains (Fig. 2, Table 1). In general, during phase 1 at 26 °C, we observe the highest heterogeneity (all six possible strain-growth combinations were different) in growth rate among the strains, with Aten-MI-1 being the strain with the highest growth rate (Fig. 2, comparative matrix under the main chart). High heterogeneity of 83% was also observed during the second phase at 32 °C followed by 67% observed during the 2nd and 1st phases at 26 °C and 28 °C, respectively. The less variable, or where most of the strains were “similar” to each other were the combination of temperature and time observed towards the end, during the second phase at 28 °C and the first phase at 30 °C (Fig. 2, matrix below the chart).

Table 1 Two-way ANOVA results. Temperature and time (phase) effects in the growth rate “µ” (day-1) of the different symbiont strains
Full size table

When comparing within strains, in none of the cases, the growth observed in both phases at 26 °C and 28 °C was different (Fig. 2, chart). This low variability extends further at 30 °C in Amil-MI. The last temperature of 32 °C was the most variable within strain for all strains.

At the temporal scale, at each temperature, we found that Aten-MI-2 and Aten-WSY showed no differences between phases at 26 °C and 28 °C and the same was observed for the pair Amil-MI and Aten-WSY at 30 °C. This means these strains should have an extended exponential growth phase compared with Aten-MI-1 and Amil-MI for the same temperatures (Fig. 2, asterisk above bars).

Finally, it is worth noticing that the strain Amil-MI was the only one capable to survive long enough at 32 °C to be able to be transferred to 34 °C where there were no dramatic increases in biomass, but the cells did not die during the 30 day incubation test (Fig. 1, dark carmine line).

Stability of the photosynthetic membranes

The stability of chloroplast thylakoid membranes was assessed by pulse amplitude modulation (PAM) fluorometry of cultures maintained at 26 °C and acclimated to 26 °C and 31 °C for one month. All strains showed evidence of an acclimation response, as the melting of chloroplast thylakoid membranes were significantly higher in cultures maintained at 31 °C compared to the 26 °C (see Table 2). Melting temperatures increased by 5.3% for the strains Aten-MI-1 and Aten-WSY and by 3.1% for the strains Aten-MI-2 and Amil-MI. Melting temperature change was the lowest in the D. trenchii strain with only 1.72% increase (Table 3). It is worth mentioning that two distinct patterns of acclimation were observed in symbionts isolated from the same reef and conspecific species (Fig. 3, see data for Aten-MI-1 vs. Aten-MI-2).

Table 2 One-way ANOVA results. Chloroplast melting temperature comparison between different symbiont strains
Full size table
Table 3 Temperature dependent growing equations of four strains of Cladocopium C1acro, where “y” represents the density in cells per millilitre at a given “x” time
Full size table
Fig. 3
figure 3

Melting temperatures (mean ± SD) of the photosynthetic membranes of five strains of Symbiodiniaceae cultures after having been grown at 26 °C (closed symbols) and at 31 °C (open symbols) for 30 days. Significant differences (see Table 2) are depicted beside the symbols with red coloured letters against its counterpart in black colouration

Full size image

Genetic characterisation

The four Cladocopuim C1acro strains showed identical ITS1 SSCP profiles (not shown) as well as an identical ITS2 sequence, which had 100% identity to Cladocopium goreaui (GenBank accession number AB778664.1). In line with this, there are at least 3 different lineages of Cladocopium from the GBR diagnosed by the sample ITS2 C1 sequence that are not C. goreaui, whose ribosomal array is also dominated by this same sequence variant. The four cultured strains belong to just one of these species lineages (T. LaJeunesse, personal communication, December, 2020). As for validating that all of our strains were actually one species, the phylogenetic analysis using the psbAncr marker clearly shows two distinctive clusters (Fig. 4a), one formed by sequences derived from C. goreaui isolated from Caribbean invertebrates, and the other formed with data from this communication along with other sequences from invertebrates sampled at the central GBR, proving we were dealing with a single undescribed Cladocopium species. This latter result justifies the validity of our DNA-fingerprinting approach using TE-AFLP in order to discriminate single genotypes. TE-AFLP analysis demonstrated that the four cultures represent distinct genotypes with an approximate 24% of shared bands (from a total 135 scored bands) below the 150 bp size mark. The analysis grouped the coral symbionts at between 77 and 84% of similarity (calculated using UPGMA. WARD algorithm also rendered same groups with similar distance among them) (Fig. 4b). The overall level of similarity of all strains was 65% (Fig. 4b, dendrogram). From a total of 135 events registered (bands of different sizes) between 60 to 460 bp, there was a total and average variation of 39.2 and 17.6%, respectively. The strains Aten-WSY and Aten-MI-2 showed more large sized (over 350 bp) fragments compared to the other two (Fig. 4b, see chart below the dendrogram).

Fig. 4
figure 4

Genetic identification of the symbionts. a Phylogenetic tree of the chloroplast psbAncr marker using maximum likelihood. The analysis includes some C. goreaui isolates along with Cladocopium C1cro. Accession numbers for sequences not from this research are depicted in brackets. Accession numbers for the strains used in this research can be found in the methods section. b TE-AFLP analysis of the monoclonal Cladocopium C1acro strains. Tree created using UPGMA showing percentage of similarity using the size window of 60–470 bp. The chart below represents the visual signature of the combination of the DNA band number as it appears in the electrophoresis and its molecular weight

Full size image

Discussion

Growth rate

The coral holobiont thrives under a narrow temperature range, which is why we are interested to understand the relationship of temperature and growth of the photosynthetic partner of this symbiosis. Recent Symbiodiniaceae genome data, including for the strain Aten-MI-1, points to positive selection for genes with functions related to photosynthesis, possibly selecting to increase thermal resilience (Liu et al. 2018), confirming an important fitness measure are growth rate in relation to temperature. Members of the Symbiodiniaceae display in hospite growth rates spanning one order of magnitude from 0.1 to 0.01 d−1. These rates are lower compared to growth rates of many planktonic dinoflagellates (Wilkerson et al. 1983). As for temperature effects in cultured symbionts, distantly related Symbiodinium microadriaticum (type A1) at 26 °C and Effrenium voratum (clade E) (previously known as S. californium (Jeong et al. 2014)) at 16 °C showed highly similar growth rates of 0.199 d−1 and 0.196 d−1, respectively (Banaszak and Trench 1995). Conversely, different growth rates at the same temperature of 25 °C were measured in members of the genus Brevolium (clade B) and Fugacium (clade F) (0.40 d−1 and 0.55 d−1, respectively (Fuchinoue et al. 2012)), and similar values of 0.25 d−1 and 0.20 d−1 were recorded for S. pilosum (type A2) and Fugacium kawagutti (type F1), respectively (Rodriguez-Roman and Iglesias-Prieto 2005). In regards to Cladocopium (clade C), a highly diverse genus (LaJeunesse et al. 2018), also known to associate with a number of invertebrates including scleractinian corals (Yuyama and Higuchi 2014), temperature-dependent growth observed in symbionts belonging to this genus showed a decline in kinetics towards temperatures above 30 °C (Karim et al. 2015; Grégoire et al. 2017) similar to our results (Fig. 2), however, with a different magnitude, potentially explained because they likely represent different genotypes and were isolated from molluscs and not corals. Intraspecific strain variation in growth rate have been observed in Breviolum psygmophilum strains, with an average growth of 0.31 d−1 under 12:12 light photoperiod, 25 °C and 60 µmol quanta m−2 s−1 of PAR (Parkinson and Baums 2014), and also responses under a variety of temperatures has recently been documented for S. microadriaticum (Diaz-Almeyda et al. 2017; Klueter et al. 2017). Similarly, our growth rate estimate for the strain Aten-MI-1 at 26 °C was 0.12 d−1 (Fig. 2), compared with 0.16 d−1 (Chakravarti et al. 2017), and 0.17 d−1 for the original heterogeneous C1 culture (Raina et al. 2017), both data taken at 27 °C. When exposed to a relatively high temperature above 30 °C, our results showed a less dramatic decline in growth, assuming an average between both phases to compare with densities from day 3 and 17 used in Chakravarti et al. (2017). These deviations can be explained in terms of timing of measurements and potential acclimation to different conditions in vitro. Our research is the closest to the isolation time from the coral host and potentially more likely to resemble the actual growth rate in hospite. However, Aten-MI-1 along with our other three strains are part of a broad symbiont collection isolated from a number of hosts from the GBR and the incubation conditions were standardised for all of them. The combination of changes in incubation temperature and potential acclimatory/adaptive effects due to the elapsed time growing in vitro may have affected the deviations observed in growth rates.

The two cultures that were isolated from different but conspecific coral colonies (A. tenuis) growing on the same reef (Nelly Bay, Magnetic Island) and in the similar photic habitat (~ 3–4 m depth), show remarkable differences in growth behaviour and thylakoid membrane fluidity (Figs. 1, 2, 3). We hypothesise that previous history along with micro-ambient and genetic differences could explain the observed behaviour. This level of variation was previously described for conspecific coral and its algal symbionts (heterogeneous cultures) isolated from different geographical locations with different temperature regimes (Howells et al. 2012). In our research, clonal strains derived from these two heterogeneous cultures were still showing differences in thylakoid membrane fluidity (Fig. 3, see Aten-MI-2 vs. Aten-WSY). Remarkably, the response of an apparently cryptic symbiont, Amil-MI matches one of the two well-observed behaviours. These symbionts were discovered by chance because the coral colony from which they were originally isolated was genotyped as type D1a (VHB, unpublished data), in agreement with a previous report for this coral species in this location (Little et al. 2004).

Photosynthetic response to elevated temperatures

In the same way, we characterised growth kinetics under different temperatures, we wanted to extend our understanding of the behaviour of the photosynthetic apparatus under thermal stress. Loss of photosynthetic activity due to thermal stress has been documented for a number of primary producers, including Symbiodiniaceae (Iglesias-Prieto et al. 1992; Sato et al. 1996; Inoue et al. 2001). Our Cladocopium C1acro clonal strains growing at 26°C showed a considerable level of variation (up to 1.24 °C) in the decrease of 50% of maximum quantum yield (Fig. 3, see pair Aten-MI-2 & Aten-WSY, for example). Similar responses were observed by Diaz-Almeyda et al. (2011) between Cladocopium goreaui and the distantly related Symbiodinium pilosum (type A2) strains, with melting values of 36.7 °C and 37.7 °C, respectively. The values observed by these authors for their C. goreaui strain were 1.7 °C lower than the average (38.4 °C) for the four clonal symbiont cultures in our study. Diaz-Almeyda et al. (2011) used a strain of C. goreaui and therefore representative of a different species from ours (T. LaJeunesse, personal communication, December 2020), also the fact that they used a lower growing temperature (24 °C instead of the 26 °C employed in our experiment) may have contributed to the difference in results. The difference between the melting temperature of the thylakoid membranes and the growing temperature was very similar in both studies (taking the average of the four strains in this study), with a value of 12.5 °C. When cultures were grown at 31 °C, the averaged melting value of our four clonal Cladocopium cultures was 39.9 °C and very close to that reported for C. goreaui by Diaz-Almeyda et al. (2011) of 40.2 °C. This increase in thermal tolerance after a period of acclimation has also been observed for other Symbiodiniaceae species in a similar experimental setting (Takahashi et al. 2013). Despite these similarities with other studies, we observed significant differences in the acclimation capacity as well as the absolute thylakoid membrane melting temperatures among the four Cladocopium strains (Fig. 3, Table 2), with the difference being of the same order of magnitude to those observed between different genera (Diaz-Almeyda et al. 2011). These observations highlight the physiological diversity among sympatric genotypes of the same Symbiodiniaceae species as observed based on their growth behaviour.

In our experiment, we used Durusdinium trenchii as a point of comparison for the Cladocopium C1acro clonal strains. In line with this, the genus Durusdinium has been typically reported as being thermotolerant (Glynn et al. 2001, Baker et al. 2004, Rowan 2004, Ulstrup et al. 2006), and in particular D. trenchii (LaJeunesse et al. 2009, Hoadley et al. 2019). This culture showed a lower melting point compared with the clonal Cladocopium strains (Fig. 3). D. trenchii strain, however, showed a faster recovery rate after thermal stress compared with heterogeneous Cladocopium C1acro culture (VHB, unpublished data). In line with this, it is been observed that Cladocopium C1acro confers the holobiont with higher thermotolerance than D. trenchii symbionts in Acropora tenuis stressing the importance of the host-symbiont interaction (Abrego et al. 2008). Finally, the observed differences between species could be explained by different regulatory mechanisms altering composition of fatty acids and antioxidant systems which seem to control the extent of acclimation and response to thermal stress (Tchernov et al. 2004, Diaz-Almeyda et al. 2011, Krueger et al. 2014).

Genetic identification of Symbiodiniaceae

The field of taxonomic characterisation of the members of the Symbiodiniaceae family has been in part impaired due to the intrinsic “coccoid” morphology in hospite, this along with some uncultivable isolates (Goulet and Coffroth 2003) witness the birth of biochemical and genetic characterisation (Schoenberg and Trench 1980a, b, c; Rowan and Powers 1991b; Baillie et al. 1998). The fact that our strains behaved differently (see Figs. 1 and 3) while having the same ITS1 SSCP profile and ITS-2 sequence led us to adopt a DNA fingerprinting similar to the AFLP, previously used in reef symbiotic invertebrates (Coffroth et al. 1992; Amar et al. 2008) but with the difference that the TE-AFLP uses three endonucleases instead of two in AFLP, with the combination of selective amplification that renders more discriminatory power with less number of bands (van der Wurff et al. 2000). Our results showed differences in their DNA fingerprint, forming two distinct clusters (Fig. 4b). Long or short (differences between growth phases) exponential growth periods (see Fig. 2) matched the clusters based on the TE-AFLP data; however, the chloroplast melting temperature data did not (i.e., Aten MI-2 vs. Aten-WSY in the TE-AFLP and Aten-WSY vs Aten MI-1 in the melting chloroplast experiment). This implies some level of caution when interpreting melting temperature data in the context of “relatedness” between genotypes and reassures the growing kinetics data as potentially more informative in terms of taxonomic similarities. The TE-AFLP approach has the advantage over ITS sequence analysis that this technology is the outcome of a large portion of the genome being analysed rather than a single locus, which has sometimes revealed an incongruence between genetic relatedness physiological response, for instance, the differential thermotolerance observed in closely related ITS2 types (Tchernov et al. 2004). We advocate other Symbiodininiaceae research to apply this method to characterise their cultures/isolates.

Implications of our findings

In the present communication, we demonstrated that different strains of Cladocopium C1acro differ physiologically in response to temperature treatments. These differences in growth and photobiology likely influence the overall physiological tolerance among coral colonies living in heterogeneous environments. The combined inter-individual variation among genotypes of host and symbiont has important implications with regard to processes of natural selection and resilience to changing environmental conditions. Further investigations are needed to examine the breadth of physiological diversity created by the interactions of different host and symbiont genotype combinations.