Reconstructing the dynamics of past coral endosymbiotic algae communities using coral ancient DNA (coraDNA)

18 19 Most scleractinian corals are drastically threatened due to global changes but some 20 colonies are intriguingly resistant to heat stress. Coral thermal tolerance partly relies on 21 genomic determinism among the cnidarian compartment but also on the physiology of their 22 associated symbiotic algae (Symbiodiniaceae). In fact, some corals can shift and/or shuffle their 23 associated Symbiodiniaceae communities to temporally cope with heat stress. So far coral 24 adjustments of their endosymbiotic algae were mainly observed at short-term evolutionary time 25 scales and we lack a general vision of coral holobiont evolution at broader timescales. We here 26 combined the use of ancient DNA from a coral core and a metabarcoding approach, to retrace 27 past Symbiodiniaceae communities associated with a living colony of Porites lobata from New 28 Caledonia over the last century. We were able to extract ancient DNA along the coral core at 29 19 time points dating back to the 1870’s. Overall, we detected 13 OTUs, nine of which were 30 affiliated to the Symbiodiniaceae Cladocopium clade, one to Azadinium spinosum 31 (Dinophycae); one to the host P. lobata, the two other OTUs remained unidentified. One OTU 32 was largely predominant and was ubiquitous over all samples. The number of OTUs was 33 marginally correlated to the total number of sequences per sample but not to the age of the 34 coraDNA sample. We found a generally stable core microbiota associated with P. lobata , 35 although drastic change in community composition was observed in coraDNA samples 36 corresponding to an extreme hot winter temperature event. More generally, this study paves the 37 way for further investigations on the evolutionary dynamics of coral holobionts at the colony 38 level over large temporal scales. 39 40

41 Coral; Symbiodiniaceae; Ancient DNA; metabarcoding, ITS2 42 43 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 3

44 45 Scleractinian corals and the associated reef ecosystems are among the most threatened 46 systems worldwide due to global changes. Almost half of all living corals have already been 47 destroyed in the last 150 years 1. Such massive loss has accelerated over the last three decades 48 under the influence of extreme climatic events, in particular heat waves, the frequency of which 49 is steadily increasing 1. Despite this alarming situation, coral colonies particularly resistant to 50 heat stress have been identified 2–4. These observations provide some hope for the maintenance 51 and/ or the restoration of corals and ecosystems they support. They also call for an urging need 52 to unravel the molecular mechanisms by which some coral colonies survive through time 53 despite recurrent events of environmental stresses and help adjust conservation plans. This task 54 is all but trivial in particular because corals are complex holobionts composed of cnidarians 55 associated with microbial communities including – among others – endosymbiotic algae 56 (Symbiodiniaceae), protista and bacteria, each of these partners potentially influencing the 57 thermotolerance of the colonies 5,6. 58 Variation in coral thermal tolerance across latitude, at least partly relies on genomic 59 variation among the cnidarian compartment of coral colonies 7. In this respect, it has been 60 suggested that resistant genotypes could emerge through intrinsic rapid genomic changes such 61 as somatic mutations 8, the activation of transposable elements 9 and/ or some modifications in 62 the proportions of genotypes coexisting within the same colony 10,11. Additionally to genomic 63 modifications, changes in DNA methylation patterns in cnidarians may also induce – at least 64 temporarily – adaptive phenotypic adjustments in coral colonies exposed to recurrent heat stress 65 12,13. Besides changes within the cnidarian compartments, coral thermal tolerance also depends 66 on the physiology of their associated symbiotic algae (Symbiodiniaceae), with which they form 67 a phototrophic mutualistic symbiosis 14,15. Symbiodiniaceae constitute a highly diversified 68 taxonomic group among which the identified species and even strains among species display 69 huge variation in thermal tolerance 16,17. Recently, a laboratory experiment has shown that 70 experimentally adapted strains of the Symbiodiniaceae Cladocopium goreaui to high 71 temperature, provide a better protection to heat stress in Acropora tenuis colonies after 72 reimplantation 18. This suggests that thermal tolerance may be acquired rapidly by natural coral 73 colonies via the acquisition of acclimatized or adapted Symbiodiniaceae from the surrounding 74 environment in response to heat stress. In fact, symbiont switching – the acquisition of new 75 (thermally resistant) species/strain from the environment – and symbiont shuffling – the 76 modification of the relative abundance of the inner Symbiodiniaceae strains within host – 77 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 4 constitute two alternative rapid responses for corals to temporarily cope with heat waves 19,20. 78 Finally, while other partners involved within the coral holobiont such as bacteria, fungi, viruses 79 and protists, certainly display higher adaptive capacity than corals in particular due to their short 80 generation time, their role in fostering adaptive response to heat stress at the holobiont scale 81 still remain to be demonstrated 5,13,21. 82 Most of our knowledge on the mechanisms underlying the adaptive response of coral 83 colonies to global changes relies on empirical studies based on controlled experiments 84 conducted either in laboratory or in the field, over extremely short-term evolutionary time 85 scales. These approaches are crucial to dissect the relative importance of specific molecular 86 mechanisms in action to foster coral responses. They have dramatically changed our vision on 87 the short-term adaptive capacity of corals 12,13. However, and complementary to these 88 approaches, we need a more general vision of coral holobiont evolution at broader timescales 89 to assess their evolutionary dynamics in response to the recent fast evolving environmental 90 changes. In this respect, ancient DNA (aDNA) based approaches are promising 22. These 91 approaches generally rely on DNA extracted from archaeological or paleontological remains or 92 from museum samples. For instance, museum samples of 8 octocoral species originally 93 collected from successive epochs revealed that mutualism between several coral species and 94 their associated Symbiodiniaceae remained stable overtime with no major changes in the last 95 two centuries despite major anthropogenic global change 23. In a more general context of 96 biodiversity, aDNA was also applied to reconstruct the community assemblages of reef 97 ecosystems from sediment cores 24,25. Surprisingly however, no attempts were made to extract 98 aDNA directly from cores excavated from a living massive coral colony that chronicles decades 99 and up to centuries, of its lifetime. Such an approach is particularly promising because it could 100 allow accessing the DNA material from individual holobiontic colonies over time and thus 101 reconstructing their intrinsic eco-evolutionary history. 102 We here provide the first proof of concept of the use of aDNA from a coral core, 103 hereafter called coraDNA (referring to the recent sedaDNA approach developed to study DNA 104 from sedimentary cores), to reconstruct past Symbiodiniaceae communities associated with a 105 living colony of Porites lobata over the last century. We discuss the benefit of such approach 106 to unravel the eco-evolutionary dynamics of coral holobiont and the current technical 107

that will need to be bypassed to have access to DNA from the cnidarian compartment 108 and hence reconstruct the full eco-evolutionary trajectories of coral holobionts over time. 109 110 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 5

111 112 Core sampling 113 Porites lobata is a massive reef-building scleractinian coral ubiquitous over the Tropical Pacific 114 Ocean. Colonies can be extremely long-lived (up to several centuries) and their growth strongly 115 depend on the environmental conditions in which they grow, especially seawater temperature 116 26. These characteristics make this species particularly suitable in the context of this study since 117 it makes it possible to obtain long cores and the marked seasonal successions of growth allow 118 the collection of core matrix at the year scale from which DNA can be extracted. A coral-matrix 119 core was collected from a 2.5-m high colony of Porites lobata at 10-m depth within the 120 Southern Lagoon of New Caledonia (22°17,146 et 166°11,004) the 17 th of December 2018 121 (Figure 1.A.). The excavated 8-cm diameter and ~ 80-cm long core was immediately stored in 122 dry ice and transferred under freezing conditions to the Institute de Recherche pour le 123 Developpement (IRD) in Nouméa, prior being sent in dry ice to the IHPE laboratory in 124 Perpignan where it was stored at -80°C until subsequent analyses. Thus, the core has been kept 125 at -80°C from its initial excavation until the sampling of coraDNA. 126 127 Figure 1: Details on the coral core of the studied Porites lobata colony. Pictures of its excavation in 128 December 2018 from the living colony in New Caledonia are provided in (A). In (B), the coraDNA 129 samples that were extracted along the core are represented from the older sample (S01) to the 130 contemporaneous living coral (control). In (C) Pictures of the interventional sampling using a CT scan 131 of the four coraDNA samples corresponding to successions before (E01) and after (E02) the 1997-1998 132 ENSO event; and before (E03) and after (E04) the anormal hot winter in 2010. Yellow arrows in the 133 last panel in (C) point toward the hole left by the drilling bit after the sampling of the E04 coraDNA 134 sample. 135 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 6 Dating successions and sampling ancient DNA from the coral core 136 The aging of the coral core was determined based on the counting of density bands as described 137 in Wu et al. 27. Annual skeletal density bands were observed using a CT scan at the imagery 138 service of the public Hospital of Perpignan (France). The observed density bands result from 139 variation in colony growth rate during the winter (slow) and during the summer (rapid) that lead 140 to a more or less dense aragonite skeleton along the year 26. According to this pattern, one year 141 was reconstructed by summing a clear (summer) and dark (winter) density bands 27. Based on 142 the reconstructed chronology of the sampled coral colony, a total of 23 samples were collected 143 along the coral core (Figure 1.B and C). Briefly, each sampling location along the core was first 144 rapidly washed using a 10% NaClO solution and ~ 3 mm of the external layer was removed 145 from the core surface using a sterile disposable scalpel to avoid possible DNA contamination 146 with contemporaneous genomic material. Once the surface of the core was decontaminated, an 147 electric driller with an individualized sterile 2-mm diameter drill bit was used to extract core 148 powder along a ~ 1.5 – 2 cm depth drilling hole perpendicular to the surface of the core. The 149 bits containing the powder was immediately transferred and rinsed in a 2 ml Eppendorf 150 (DNA/RNA and DNase/RNase free tube) containing 1 ml of Tissue Lysis buffer ATL from the 151 QIAamp DNA Micro Kit (Qiagen) and stored at 4°C until the DNA extraction process (i.e. at 152 most 24 h later). 153 Sampling along the core was achieved following two strategies. The first strategy 154 consisted in systematically sampling every 5 cm along the core without accurate dating. 155 According to the retraced chronology and to the thickness of the observed density bands, such 156 strategy roughly corresponded to a sample collected every 8 to 10 years back in time over the 157 colony history from December 2018 (collection date) until the 1870’s. A total of 17 coraDNA 158 samples were extracted following this design including one sample from the top of the coral 159 core (as positive control) which corresponds to the living colony at the time of the core sampling 160 (Sample S01 to S16 and control in Figure 1.B.). This strategy was used to specifically test our 161 ability to detect and obtain processable coraDNA samples over large timescales. Moreover, 162 two duplicates were sampled for two random samples (S04 and S08; Figure 1.B.). These 163 duplicates consisted in independent coraDNA samples obtained from the same section along 164 the core and using two independent sterile drill bits. 165 The second strategy aimed at studying the dynamics of the algae endosymbiotic 166 communities associated with coral colonies during well documented past extreme climatic 167 events. We targeted two well-known climatic anomalies that were easily observable along the 168 core in 2010 (abnormally hot winter) and in 1997-1998 which corresponded to a severe ENSO 169 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 7 event 27. Two samples were collected, one prior and the other after each of these two climatic 170 events (N = 4; Figure 1.C.). For this purpose, and to accurately sample the targeted coral 171 successions, the drilling procedure was achieved under interventional CT 3D scan at the 172 imagery services from the public Hospital of Perpignan. The coordinates of the targeted 173 successions were first retrieved, and the drilling sampling was achieved under interventional 174 scan to ensure that the same succession was sampled within the core during the sampling 175 process (Figure 1.C.). The precautions previously described were taken to avoid contamination 176 (treatment with bleached plus abrasion of the ~3 mm of the external layer of the core and a 177 unique sterile drill bit per sample). 178 Additionally to the overall 23 samples collected along the core, a negative control was 179 prepared at the time of the sampling collection to check for possible contaminations over the 180 whole process from DNA sampling to library preparation. This negative control was obtained 181 following the same protocol as described above except that the drill bit did not touch the core 182 prior to the rinsing step into the Tissue Lysis buffer ATL. 183 184 DNA extraction 185 The overall DNA extraction process and DNA amplification steps were performed at the 186 degraded DNA platform (Institut des Sciences de l’Evolution de Montpellier, France; 187 http://club.quomodo.com/plateforme-adn-degrade) offering facilities dedicated to the study of 188 ancient DNA. DNA extractions were performed using the QIAamp DNA Micro Kit (Qiagen) 189 following the “ Purification of genomic DNA from bones ” protocol. Briefly, 25 µl of Qiagen 190 Proteinase K were added to each sample. Samples were then incubated at 56°C with shaking at 191 1200 rpm overnight. The day after, 1 ml of AL buffer was added and samples were incubated 192 at 70°C during 10 minutes to inactivate enzymes. After a centrifugation step at full speed during 193 one minute, the supernatant was transferred to a QIAamp MinElute column. Once the DNA 194 attached on the column membrane (centrifugation step at 8000 rpm for 1 minute), the membrane 195 was successively washed using 600 µl of the AW1 and the AW2 buffer. After these washing 196 steps, the membrane was dried by centrifugation during 3 minutes at full speed. The DNA was 197 eluted using 30 µL of sterile water. The eluted DNA samples were then stored at -20°C until 198 subsequent molecular processes. The quality of DNA extracts wase assessed on a Bioanalyzer 199 High Sensitivity DNA kit (Agilent, USA) and DNA was quantified using a Qubit fluorometric 200 quantification with the ds DNA High Sensitivity assay kit (Thermo Fisher Scientific, USA). 201 202 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 8 Library preparation and sequencing 203 Each DNA sample was used as initial template for amplifying a portion of the ITS2 gene using 204 a set of 4 primers derived from the literature 28,29 with Illumina adapters and spacers (Table 1). 205 206 ITS2-MiSeqNN-F: TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGNNGTGAATTGCAGAACTCCGTG ITS2-MiSeqNN-R: GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGNNCCTCCGCTTACTTATATGCTT ITS2-MiSeqNNN-F: TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGNNNGTGAATTGCAGAACTCCGTG ITS2-MiSeqNNN-R: GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGNNNCCTCCGCTTACTTATATGCTT Table 1: Sequence of the 4 primers (including adapters and spacers) used to amplify a ~ 350 207 bp fragment of the ITS2 gene 208 209 Concomitantly, the extraction negative control and a PCR negative control (consisting in 210 distilled water) were used in the same PCR reaction to control for potential contamination 211 during the overall process. PCRs were performed in 50 µL containing 5µL of Buffer 10X PCR 212 Gold (1X), 5 µL of MgCL 2 (25 mM), 1 µL of each dNTPs (200 µM), 1.35 µL of primer mix 213 (0.25 µM), 0.25 µL of Taq Polymerase (1.25 units), 4 µL of DNA template and 35.4 µL of 214 distilled sterile water. Amplification was performed using the following PCR conditions: an 215 initial denaturation step at 95°C during 7 minutes followed by 35 cycles each of which 216 constituted of a denaturation step at 95°C during 15 sec, an annealing step at 57°C during 30 217 sec and an elongation step at 72°C during 30 sec. After these 35 cycles, a final elongation step 218 at 72°C during 7 minutes was applied. The amplification was checked by migrating 5 µL of 219 each PCR reaction on a 2% agarose gel stained with ethidium bromide and visualized under 220 UVs. Expected size of the amplicon is around 350 bp. 221 Indexed libraries were generated using the standard Illumina two-step PCR protocol using Q5 222 high fidelity DNA polymerase (New England Biolabs). Paired-end sequencing with a 2x250 bp 223 read length was performed at the Bio-Environment platform (University of Perpignan Via 224 Domitia Perpignan, France) on a MiSeq system (Illumina) using v2 chemistry according to the 225 manufacturer’s protocol. Sequencing data are available for download on SRA under the 226 bioproject number ### 227 228 Data processing 229 The sequence datasets were uploaded to the Galaxy web platform 30 and processed using the 230 Finding Rapidly OTUs with Galaxy Solution (FROGS) pipeline at the GenoToul platform 231 (Toulouse, France) 31. The first pre-processing step of this pipeline consisted in demultiplexing, 232 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 9 dereplicating and cleaning all reads. Given the theoretical expected amplicon size (i.e., ~ 350 233 bp) and after a quick overview of the overall amplicon lengths, we kept all sequences which 234 sizes ranged from 150 to 490 pb. These filtered sequences were then clustered using the 235 SWARM algorithm using an aggregation distance set at 1. This iterative clustering approach 236 uses amplicons’ homology, structure and abundance hence limiting potential biases generated 237 by other approaches such as de novo clustering methods (i.e., input order dependencies and 238 arbitrary clustering) (Mahe et al., 2014). Clusters were then cleaned to remove potential 239 chimeras, singletons and under-represented clusters (i.e. <50 sequences) using VSEARCH 240 (Rognes et al., 2016). 241 Each OTU was identified based on a nucleotide megablast using the online standard 242 database (nt/nr) available from NCBI. To more precisely identify OTUs that were affiliated to 243 the Symbiodiniaceae genus Cladocopium (clade C) based on the results from Blast, we next 244 computed pairwise genetic distance between each OTU seed sequence and a subset of 245 sequences from different Cladocopium strains available from LaJeunesse et al. 32. Pairwise 246 genetic distances were computed using the ‘K80’ evolutionary model as implemented in the 247 ape package V5.0 in R 33. 248 249

250 251 Dating back coral successions along the core 252 The obtained coral-skeleton core followed the colony’s growth axis on more than 2/3 of its 253 length (see Figure 1.C.; middle panel for illustration). We were hence capable to accurately date 254 back the growth successions over the last seven decades, i.e., back to 1950. Density bands along 255 the oldest part of the coral core partly deviated from the central vertical axis preventing our 256 capability to date back accurately the oldest successions. Annual growth bands were 257 approximately 5-6 mm thick and generally homogenous along the core. According to this 258 measure, we estimated that the oldest coraDNA sample collected from the core (S01; Figure 259 1.B.) roughly correspond to coral holobiont that lived in the 1870’s. 260 261 Amplicon sequencing from coraDNA samples, sequence affiliation and identification of 262 endosymbiotic algae communities 263 Because of low DNA concentrations, DNA quality and quantity could not be assessed from all 264 samples except the positive control. PCR amplicons of expected size for ITS2 Symbiodiniaceae 265 marker were however obtained from all samples. After sequencing and filtration steps, 266 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 10 sequences were validated from 19 out of the 23 initial coraDNA samples including the positive 267 control (Table 2). The final number of filtered sequences obtained from these 19 coraDNA 268 samples ranged from 1476 (sample S12) to 76403 (sample S13) (Table 2). Four coraDNA 269 samples displayed negligible numbers of sequences (from 9 to 33) and were then excluded for 270 further analyses. Similarly, extraction and PCR negative controls exhibited negligible numbers 271 of sequences (8 and 17, respectively) with an expected size of 300 bp (190 – 450 bp; Table 1), 272 which indicates negligible cross-contaminations during all the process from sampling, 273 extraction to sequencing. 274 275 Sample % kept Paired-end assembled (%) With 5' primer (%) With 3' primer (%) With expected length (%) Without Ns (%) Nseq OTU > 50 seq Samples kept in the analyses E01 97.29 77.44 77.38 76.15 75.34 75.34 73241 E02 97.12 57.91 57.87 56.68 56.24 56.24 54571 E03 96.37 73.18 73.13 71.18 70.52 70.52 68484 E04 91.60 48.53 48.52 47.35 44.46 44.46 43454 S01 87.58 75.61 75.58 74.20 66.21 66.21 64448 S04 75.10 59.92 59.90 59.09 45.00 45.00 44203 S04" 85.18 15.65 15.62 15.30 13.33 13.33 12993 S05 4.44 89.98 89.96 89.83 3.99 3.99 3928 S06 67.49 65.52 65.50 63.76 44.22 44.22 42583 S08 96.01 77.93 77.90 75.11 74.82 74.82 71366 S08" 97.56 46.68 46.63 45.66 45.54 45.54 44501 S09 63.04 70.01 69.98 68.80 44.13 44.13 43147 S10 8.32 48.20 48.19 48.09 4.012 4.012 3884 S11 85.80 54.13 54.11 53.04 46.44 46.44 45236 S12 6.98 21.95 21.94 21.90 1.53 1.53 1476 S13 96.83 80.93 80.90 78.79 78.36 78.36 76403 S14 93.24 74.74 74.69 73.05 69.69 69.69 68031 S15 56.67 88.20 88.18 86.22 49.98 49.98 48388 Control 28.60 8.73 8.73 8.66 2.50 2.50 2454 Samples removed from the analyses S02 0.02 152.242 152.221 152.075 33 33 0 S03 0.02 58.41 58.406 58.344 9 9 0 S07 0.02 76.544 76.531 76.462 15 15 0 S16 0.02 52.119 52.104 52.046 11 11 0 Negative controls Extraction 0.02 45.646 45.644 45.597 8 8 0 PCR 0.15 11.01 11.007 10.998 17 17 0 276 Table 2: Details on the obtained and conserved sequences from each coraDNA sample. 277 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 11 The sequences clustered into 13 OTUs. Nine of these OTUS were affiliated to 278 Symbiodiniaceae and all of which were assigned to the Cladocopium clade. Based on the 279 genetic distances computed between each of these 9 clusters and ITS2 sequences from 280 LaJeunesse et al. 32, all OTUs were more specifically affiliated to the ITS2-C15 type (computed 281 distances ranged from 0 to 0.044; Table S1). Among the 4 other non-specific OTUs, one was 282 affiliated to Azadinium spinosum (Dinophycae; OTU_7; Query cover = 100%, E-Value = 6e -283 156; Per. Ident. = 98.1%); one to P. lobata (the coral host species; OTU_11; Query cover = 284 100%, E-Value = 8e-175; Per. Ident. = 100%), and the two others were not successfully affiliated 285 with any identified sequence available from the NCBI nr database (OTU_12 and OTU_13). 286 Sequences of Azadinium spinosum (OTU_7) were amplified only in the S08 and its duplicate 287 S08” coraDNA samples. Amplified sequences of P. lobata (OTU_11) were found in the S01, 288 S08, E02 and E03 coraDNA samples although at low abundance (Figure 2; Table S2). Finally, 289 OTUs that were not affiliated with any organism documented in the NCBI database (OTU_12 290 and OTU_13) were present in samples S08 and/or S08” again with low read numbers (Figure 291 2). 292 293 294 Figure 2: Proportion of sequences attributed to each of the 13 clusters among the 19 positive coraDNA 295 samples A. obtained from the systematic sampling (i.e. every 5 cm along the coral core) and B. collected 296 under CT SCAN corresponding to successions before (E01) and after (E02) the 1997-1998 ENSO event; 297 and before (E03) and after (E04) the anormal hot winter in 2010. 298 299 Distribution of Symbiodinium OTUs over the coraDNA samples 300 The total number of OTUs obtained from each coraDNA sample varied from 1 to 10. Among 301 the 9 Symbiodiniaceae OTUs identified, 1 OTU (OTU_1, Figure 2) was clearly predominant 302 representing 92% of the generated sequences overall samples (from 80.88 % in S15 to 100 % 303 in S12). Only OTU_1 is ubiquitous among all coraDNA samples. OTU_9 and OTU_10 could 304 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 12 also be reasonably considered as ubiquitous since they have been detected in all except the S12 305 and the positive control coraDNA samples. However, these two samples were the one harboring 306 the lowest sequence coverage which may explain this absence (Table 1). 307 The number of OTUs per sample is marginally correlated to the total number of 308 sequences obtained per sample (Spearman correlation, rho = 0.53; p-value = 0.04) but is not 309 correlated to the age of the coraDNA sample (using the position order at which the coraDNA 310 was sampled along the core as a proxy; rho = 0.148; p-value = 0.60). 311 312

313 314 Unlocking the access to past coral communities 315 According to our results, ancient DNA from a coral core ( coraDNA) excavated from living 316 Porites lobata colonies can be obtained and used for metabarcoding approaches to reconstruct 317 the Symbiodiniaceae endosymbiotic communities that have been successively associated to the 318 colonies over a century. This paves the way for unprecedented time-series studies to document 319 the eco-evolutionary dynamics ofthe coral/Symbiodiniaceae mutualistic interaction to most 320 massive coral species. 321 Nine of the 13 OTUs detected in this study belong to the Symbiodiniaceae and all were 322 affiliated to the ITS2-C15 type. Since pairwise genetic distances computed between each of 323 these OTUs were relatively low, it is likely that the observed diversity among these strains result 324 from intragenomic variation at the ITS2 marker as previously highlighted 34,35. Importantly 325 however, the possible resulting overestimation of the diversity in the Symbiodiniaceae 326 community documented here cannot be related to possible post-mortem DNA modifications 327 commonly observed in ancient DNA samples 36. Indeed, all but one OTU were detected in the 328 living part of the coral core (positive control) which means that the nature and state of the 329 coraDNA samples does not seem to artificially inflate diversity of the Symbiodiniaceae 330 community. 331 Moreover, among the 13 identified OTUs, 1 was attributed to P. lobata. This P. lobata 332 OTU was detected, although at low abundance, in 4 coraDNA samples that do not have obvious 333 commonalities (e.g. spatial location along the core) but a high number of sequences. The 334 detection of the coral host species using the present ITS2 metabarcoding confirms that host 335 DNA may also be extracted and amplified to some extent. This result makes it possible to 336 consider the exploitation of coraDNA samples to trace the eco-evolutionary dynamics of the 337 genome and/or epigenome of the host coral species P. lobata at the colony scale. 338 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 13 Among the OTUs not affiliated to Symbiodiniaceae or to the cnidarian, we unexpectedly 339 detected A. spinosum among two coraDNA duplicates collected from the same location along 340 the core. These duplicates correspond to coral holobionts that lived in the 1940’s – 1950’s. 341 Azadinium spinosum is a photosynthetic dinoflagellate that was first described in the late 2000’s 342 and since then, has been reported on the coasts of Northern Europe 37, South and Central 343 America 38,39 and very recently in Asian pacific 40. While we found no direct evidence of its 344 occurrence in New Caledonia or the Southern Pacific exists in the literature, Azadinium spp. 345 was recently observed in French Polynesia under a scanning electron microscope (Mirielle 346 Chanin; pers. comm). Importantly A. spinosum is one of the primary producer of azaspiracid 347 toxins causing important health issues to several animals, mainly vertebrates including humans, 348 through toxin bioaccumulation 41. In this context, azaspiracid-2 (AZA2) toxins were recently 349 detected in New Caledonia based on a SPATT (Solid Phase Adsorption Toxin Tracking) 350 approach, hence indirectly indicating that Azadinium spp. can be present locally at least 351 temporary 42. So far, no interactions were described between P. lobata or any other coral species 352 and A. spinosum. We here hypothesize that A. spinosum could have been accidentally captured 353 by coral hosts during feeding. Accordingly, the detection of A. spinosum at only one time period 354 (although detected in two coraDNA duplicate samples from this time period) could coincide 355 with a A. spinosum bloom episode as previously described at some localities 38. Thus, 356 additionally to DNA from Symbiodiniaceae and their coral hosts, our results suggest that we 357 can also detect the punctual presence back in time of other organisms (most likely present in 358 high abundance) in the environment and that may have been captured by corals. However, 359 (short) specific molecular markers tailored to the targeted organisms would be necessary, as the 360 marker used in this study was not initially designed for such purposes. 361 From a technical point of view, it is important to note that the number of OTUs detected 362 in the coraDNA samples was associated to sequencing output but not to the age of the coraDNA 363 samples (estimated from the position of the coraDNA samples along the core). This suggests 364 that a low number of sequences but probably not the age of the coraDNA could lead to an under 365 represented vision of the studied past Symbiodiniaceae communities (see also rarefaction 366 curves in Supplementary Figure SF1). This is clearly illustrated by the S12 coraDNA fairly 367 recent sample, for which we obtained the smallest number of sequences and that harbors only 368 one OTU (OTU_1; Figure 2) which is also predominant in all coraDNA samples (representing 369 80.9 % to 100 % percent of the overall filtered sequences). Conversely, we detected up to 5 370 different clusters in the oldest coraDNA sample (S01) including 4 OTUs affiliated to 371 Cladocopium and one affiliated to P. lobata . We thus advise to sequence libraries from 372 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 14 coraDNA samples with a high coverage to avoid possible false negatives or to use a rarefaction 373 approach to characterize the optimal sequencing depth. 374 375 Dynamics of Symbiodiniaceae at large temporal scale 376 In addition to breaking down a technical barrier, this study also traces back the dynamics of the 377 Symbiodiniaceae community associated with P. lobata over time at the colony scale. More 378 particularly, the Symbiodiniaceae community associated with this P. lobata colony is composed 379 of 1 OTU (OTU1) assigned to the ITS2-C15-type. This occurrence is largely predominant since 380 the earliest part of the coral core; that was estimated to date back to the 1870’s approximately. 381 Two other Symbiodiniaceae OTUs were found to be (nearly) ubiquitous although at very low 382 proportion even in the contemporaneous living coral (OTU9 and OTU10). These three OTUs 383 are thus likely to be in tight association with the coral host and probably part of the core 384 Symbiodiniaceae community of the P. lobata colony. This is in line with previous studies that 385 have documented a similar pattern in the symbiotic community associated with natural colonies 386 of P. lobata where the Symbiodiniaceae C15 strain was highly predominant 43. No other clear 387 pattern was observed in terms of community changes over time except for the most abundant 388 non-ubiquitous OTUs such as OTU2, OTU4 and OTU5 which are generally absent in the oldest 389 coraDNA samples and appear sporadically only from the S06 – S08 samples (i.e. since the 390 1920’s) and upward along the core. Two hypotheses could explain this temporal fluctuation of 391 these OTUs. First, the abundance of these OTUs in the environment display important temporal 392 fluctuations and the studied P. lobata colony adjusted its associated Symbiodiniaceae 393 community according to the most prevalent OTU present in the environment. Alternatively, the 394 abundance of the different Symbiodiniaceae strains in the environment is relatively stable in 395 time and the P. lobata colony adjusted its Symbiodiniaceae community according to its 396 interaction’ preference, physiological state and/or to the prevailing environment 44. Importantly 397 however, and considering that each OTU is a singular genetic entity, we lack information 398 regarding the functional and physiological consequence of the association between these 399 different Cladocopium strains and P. lobata to support one of these hypotheses. The use of 400 more resolutive genetic markers could make it possible to better distinguish between different 401 strains from the C15 clade potentially associated with P. lobata colonies(ref). Importantly 402 however, and because ancient DNA is more prone to degradation, we could be constrained by 403 the amplicon size. We would thus be more in favor of combining multiple small barcodes rather 404 than using a larger marker. 405 406 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 15 Dynamics of Symbiodiniaceae associated with extreme climatic events 407 No major changes in the Symbiodiniaceae community were observed after the targeted 408 severe ENSO event that occurred in 1997-1998 and which led to a decrease in growth visible 409 on the core, compared to that observed in the coraDNA sampled before this event. Two 410 hypotheses might explain this result. First, the targeted ENSO event did not stress the colony 411 enough (e.g. no thermal bleaching occurred) to temporally modify the associated 412 Symbiodiniaceae community. In this regard, and although not comparable to a bleaching event, 413 stability of the Symbiodiniaceae community associated with P. lobata during extreme storm 414 events was previously reported on Kiritimati Island in the central equatorial Pacific Ocean 34. 415 More generally, the association between P. lobata and their symbionts are generally 416 acknowledged to be stable over time even when facing environmental stress 34,43. Alternatively, 417 and despite all precautions, we might have missed the succession corresponding to the past 418 coral colony recovering from the ENSO event during the coraDNA sampling. Clearly, more 419 coraDNA replicates could have allowed teasing apart these two hypotheses. However, because 420 of the difficulty of sampling only a single succession of coral growth band, and the limited 421 amount of material that can be obtained, replicating coraDNA samples is particularly tedious. 422 One solution could be to use a drill to excavate larger-diameter cores from natural colonies. 423 The most important pattern observed in this study however, concerns the apparent loss of the 424 Symbiodiniaceae community diversity observed post 2010 abnormal warm winter period 27. In 425 fact, the E04 coraDNA sample harbor a very different symbiotic community compared to the 426 3 others recent coraDNA samples and to the control. Only 3 Symbiodiniaceae OTUs were 427 detected in this coraDNA sample while 8 to 10 were found in the other 3 recent samples and in 428 the contemporaneous control sample. Such low diversity found in E04 cannot be explained by 429 a lower sequencing coverage (Supplementary Figure SF1). One hypothesis could be that this 430 pattern reflects a temporary and partial loss of Symbiodiniaceae diversity by the coral colony 431 during this abnormally warm and stable period over the year. This is in line with a pattern of 432 decreased in symbiotic algae diversity observed in less variable environments in several coral 433 species 45. Moreover, the remaining 3 OTUs detected in E04 do not correspond to the most 434 abundant ones, which suggest that this loss in Symbiodinium OTUs is likely to be non-random. 435 These 3 OTUs are those that are ubiquitous to all coraDNA samples. This result hence supports 436 the hypothesis that these 3 OTUs are part of the core Symbiodiniaceae community of the P. 437 lobata colony irrespective to their (sometimes low) relative abundance among the colony 15. 438 439 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 16

and perspectives 440 441 Together our results highlight the fact that ancient DNA can be extracted from cores excavated 442 from stony corals, the so-called coraDNA, dating back to at least one century. Moreover, among 443 the DNA material extracted from the core, we detected DNA from the coral host species, the 444 dinoflagellate symbionts and a free-living environmental dinoflagellate. This result paves the 445 way for studying the temporal evolutionary dynamics of coral holobionts at the colony scale. 446 Combined with some geochemical analyses on the same samples collected along the coral core, 447 this approach could provide new insights on the mechanisms underlying the adaptive responses 448 of corals to several past stress events including temperature changes, pH and/or chemicals such 449 as heavy metals. 450 451 452 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 17 Supplementary material 453 Supplementary table S1: Pairwise genetic distances computed between each pair of the 454 sequences from the OTUs obtained in this study or from previously identified Symbiodinium 455 strains available from LaJeunesse et al. (2003). 456 OTU__1 OTU__2 OTU__3 OTU__4 OTU__5 OTU__6 OTU__8 OTU__9 OTU__10 OTU__1 0 0,014 0 0,007 0,007 0 0,044 0,022 0,022 OTU__2 0,014 0 0,014 0,022 0,022 0,014 0,059 0,036 0,036 OTU__3 0 0,014 0 0,007 0,007 0 0,044 0,022 0,022 OTU__4 0,007 0,022 0,007 0 0,014 0,007 0,051 0,029 0,029 OTU__5 0,007 0,022 0,007 0,014 0 0,007 0,051 0,029 0,029 OTU__6 0 0,014 0 0,007 0,007 0 0,044 0,022 0,022 OTU__8 0,044 0,059 0,044 0,051 0,051 0,044 0 0,067 0,067 OTU__9 0,022 0,036 0,022 0,029 0,029 0,022 0,067 0 0,029 OTU__10 0,022 0,036 0,022 0,029 0,029 0,022 0,067 0,029 0 AY239363.1_Symb_C1b 0,029 0,044 0,029 0,037 0,037 0,029 0,075 0,051 0,051 AY239364.1_Symb_C1c 0,029 0,044 0,029 0,037 0,037 0,029 0,075 0,051 0,051 AY239365.1_Symb_C3h 0,022 0,036 0,022 0,029 0,029 0,022 0,067 0,044 0,044 AY239366.1_Symb_C3j 0,022 0,036 0,022 0,029 0,029 0,022 0,067 0,044 0,044 AY239367.1_Symb_C8 0,037 0,052 0,037 0,044 0,044 0,037 0,083 0,059 0,059 AY239368.1_Symb_C8a 0,044 0,059 0,044 0,052 0,052 0,044 0,091 0,067 0,067 AY239369.1_Symb_C15 0 0,014 0 0,007 0,007 0 0,044 0,022 0,022 AY239370.1_Symb_C17 0,022 0,036 0,022 0,029 0,029 0,022 0,067 0,044 0,044 AY239372.1_Symb_C21 0,022 0,036 0,022 0,029 0,029 0,022 0,067 0,044 0,044 AY239378.1_Symb_C26 0,022 0,036 0,022 0,029 0,029 0,022 0,067 0,044 0,044 AY258488.1_Symb_C1d 0,036 0,051 0,036 0,044 0,044 0,036 0,082 0,059 0,059 AY258496.1_Symb_C31 0,029 0,044 0,029 0,037 0,037 0,029 0,075 0,051 0,051 AY258501.1_Symb_C35a 0,022 0,036 0,022 0,029 0,029 0,022 0,067 0,044 0,044 AY765398.1_Symb_C1m 0,037 0,052 0,037 0,044 0,044 0,037 0,083 0,059 0,059 AY765400.2_Symb_C33 0,044 0,059 0,044 0,052 0,052 0,044 0,091 0,067 0,067 AY765401.1_Symb_C35 0,036 0,051 0,036 0,044 0,044 0,036 0,082 0,059 0,059 AY765402.1_Symb_C42 0,029 0,044 0,029 0,037 0,037 0,029 0,075 0,051 0,051 AY765413.1_Symb_C78 0,036 0,051 0,036 0,044 0,044 0,036 0,082 0,059 0,059 AY765414.2_Symb_C79 0,044 0,059 0,044 0,052 0,052 0,044 0,091 0,067 0,067 AF333518.1_Symb_C4 0,014 0,029 0,014 0,022 0,022 0,014 0,059 0,036 0,036 AY589730.1_Symb_C1g 0,029 0,044 0,029 0,037 0,037 0,029 0,075 0,051 0,051 AY589734.1_Symb_C3f 0,029 0,044 0,029 0,036 0,036 0,029 0,075 0,051 0,051 AY589746.1_Symb_C31a 0,029 0,044 0,029 0,037 0,037 0,029 0,075 0,051 0,051 AY589752.1_Symb_C47 0,029 0,044 0,029 0,037 0,037 0,029 0,075 0,051 0,051 AY589771.1_Symb_C66 0,029 0,044 0,029 0,037 0,037 0,029 0,075 0,051 0,051 AY258473.1_Symb_C1h 0,029 0,044 0,029 0,037 0,037 0,029 0,075 0,051 0,051 AY258475.1_Symb_C8b 0,037 0,052 0,037 0,044 0,044 0,037 0,083 0,059 0,059 AY258481.1_Symb_C37 0,052 0,067 0,052 0,06 0,06 0,052 0,099 0,075 0,075 AY258485.1_Symb_C40 0,036 0,051 0,036 0,044 0,044 0,036 0,082 0,059 0,059 AY258487.1_Symb_C42 0,036 0,051 0,036 0,044 0,044 0,036 0,082 0,059 0,059 AF499789.1_Symb_C3 0,022 0,036 0,022 0,029 0,029 0,022 0,067 0,044 0,044 AF499794.1_Symb_C4 0,029 0,044 0,029 0,037 0,037 0,029 0,075 0,051 0,051 AF499797.1_Symb_C7 0,022 0,036 0,022 0,029 0,029 0,022 0,067 0,044 0,044 AF499801.1_Symb_C12 0,022 0,036 0,022 0,029 0,029 0,022 0,067 0,044 0,044 JQ003816.1_Symb_D 0,827 0,82 0,827 0,856 0,827 0,827 0,856 0,82 0,861 457 458 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 18 Supplementary table S2: Raw data of the sequence abundance of each OTU identified in this 459 study over all coraDNA samples. 460 OTU_ id Nseq total S01 S04 S04" S05 S06 S08 S08" S09 S10 S11 S12 S13 S14 S15 E01 OTU_1 748108 55576 44178 12986 3924 39658 62753 43406 37033 3493 41942 1476 67729 68005 39137 67107 OTU_2 32296 0 1 0 0 2912 2786 340 6099 0 2030 0 4994 3 9241 1460 OTU_3 23547 8773 0 0 0 0 3858 172 0 0 694 0 0 0 0 4078 OTU_4 6503 0 0 0 0 0 1574 84 0 0 553 0 3653 0 0 237 OTU_5 1193 0 0 0 0 0 0 333 0 386 0 0 0 0 0 225 OTU_6 264 0 0 0 0 0 0 0 0 0 0 0 0 0 0 123 OTU_7 230 0 0 0 0 0 211 19 0 0 0 0 0 0 0 0 OTU_8 141 0 0 0 0 0 113 0 0 0 0 0 0 0 0 0 OTU_9 139 8 11 5 1 6 10 1 9 2 9 0 15 15 6 7 OTU_10 114 10 13 2 3 7 8 8 6 3 8 0 12 8 4 4 OTU_11 92 81 0 0 0 0 4 0 0 0 0 0 0 0 0 0 OTU_12 94 0 0 0 0 0 0 94 0 0 0 0 0 0 0 0 OTU_13 93 0 0 0 0 0 49 44 0 0 0 0 0 0 0 0 461 462 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 19 SF1: Rarefaction curves obtained for each of the 19 samples. 463 464 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 20 Declarations 465 Availability of data and materials 466 Raw sequences from the metabarcoding approach are available on GenBank under the SRA 467 project id. PRJNA1086717. 468 Competing interests 469 The authors declare that they have no competing interests. 470 Funding 471 This study was fully funded by an internal grant from the IHPE laboratory in the context of 472 the challenge ‘Holobiont Dynamics and Fitness’. The drilling of the coral was funded through 473 the South West Pacific ‘chantier’ of the UMR LOCEAN, while sea access, core drilling, 474 preservation and transport, was allowed through the personal and facilities of the IRD 475 Research Cenrtre of New Caledonia. 476 Authors' contributions 477 O.R and J.V.D equally contributed to the conception and design of the work. The coral core 478 was sampled and sent to the IHPE laboratory by D.D and J.B with the help of G.I for 479 administrative aspects. D.D. and T.G. dated the coral core successions. T.G, M.S, O.R, M.B.Z 480 and J.V.D acquired the ancient DNA extracts with the help of C.T at the degraded DNA 481 platform (Montpellier). O.R, J.F, M.S and J.V.D processed the DNA extracts. O.R., T.G and 482 E.T. analyzed the metabarcoding datasets. O.R., T.G. and J.V.D interpreted the results. O.R 483 have drafted the manuscript and J.V.D, D.D and E.T. substantively revised it. All authors 484 approved the submitted version. 485 Acknowledgments 486 O.R and J.V.D would like to thank Diane Merceron for her kind help in getting the coral core 487 travelling from New Caledonia to Perpignan. We thank the personal (divers, boat pilot, etc,..) 488 of the IRD centre of New Caledonia who allowed for coral core extraction and transportation. 489 We thank the Bio-Environment platform (University of Perpignan Via Domitia) for the 490 sequencing process, Dr. Cristian Chaparro and the IHPE bioinformatics service and the 491 GenoToul bioinformatics platform, Toulouse Occitanie (Bioinfo Genotoul, doi: 492 10.15454/1.5572369328961167E12) for providing computing resources (Galaxy instance; 493 http://sigenae-workbench.toulouse.inra.fr). Finally, we would like to warmly thanks all staff 494 from the imagery service at the public Hospital of Perpignan for their kind welcome and help 495 during CT scan including Stephane Belfio, Marine Maurel and Cécilia Colomer. This study is 496 set within the framework of the “Laboratoire d’Excellence (LabEx)” TULIP (ANR-10-LABX-497 41) and with the technical support of the LabEx CeMEB (ANR-10-LABX-04-01). 498 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 21

499 500 1. Díaz, S., Settele, J., Brondízio, E., Ngo, H.T., Guèze, M., Agard, J., Arneth, A., 501 Balvanera, P., Brauman, K., Watson, R.T., et al. (2019). Résumé à l’intention des décideurs 502 du rapport sur l’évaluation mondiale de la biodiversité et des services écosystémiques de la 503 Plateforme intergouvernementale scientifique et politique sur la biodiversité et les services 504 écosystémiques. 53. 505 2. DeCarlo, T.M., Harrison, H.B., Gajdzik, L., Alaguarda, D., Rodolfo-Metalpa, R., 506 D’Olivo, J., Liu, G., Patalwala, D., and McCulloch, M.T. (2019). Acclimatization of massive 507 reef-building corals to consecutive heatwaves. Proc. R. Soc. B 286, 20190235. 508 10.1098/rspb.2019.0235. 509 3. Maggioni, F., Pujo-Pay, M., Aucan, J., Cerrano, C., Calcinai, B., Payri, C., Benzoni, 510 F., Letourneur, Y., and Rodolfo-Metalpa, R. (2021). The Bouraké semi-enclosed lagoon (New 511 Caledonia) – a natural laboratory to study the lifelong adaptation of a coral reef ecosystem to 512 extreme environmental conditions. Biogeosciences 18, 5117–5140. 10.5194/bg-18-5117-513 2021. 514 4. Palumbi, S.R., Barshis, D.J., Traylor-Knowles, N., and Bay, R.A. (2014). Mechanisms 515 of reef coral resistance to future climate change. Science 344, 895–898. 516 10.1126/science.1251336. 517 5. Clerissi, C., Brunet, S., Vidal-Dupiol, J., Adjeroud, M., Lepage, P., Guillou, L., 518 Escoubas, J.-M., and Toulza, E. (2018). Protists Within Corals: The Hidden Diversity. Front. 519 Microbiol. 9, 2043. 10.3389/fmicb.2018.02043. 520 6. Webster, N.S., and Reusch, T.B.H. (2017). Microbial contributions to the persistence 521 of coral reefs. ISME J 11, 2167–2174. 10.1038/ismej.2017.66. 522 7. Dixon, G.B., Davies, S.W., Aglyamova, G.V., Meyer, E., Bay, L.K., and Matz, M.V. 523 (2015). Genomic determinants of coral heat tolerance across latitudes. Science 348, 1460–524 1462. 10.1126/science.1261224. 525 8. Van Oppen, M.J.H., Souter, P., Howells, E.J., Heyward, A., and Berkelmans, R. 526 (2011). Novel Genetic Diversity Through Somatic Mutations: Fuel for Adaptation of Reef 527 Corals? Diversity 3, 405–423. 10.3390/d3030405. 528 9. Rey, O., Danchin, E., Mirouze, M., Loot, C., and Blanchet, S. (2016). Adaptation to 529 Global Change: A Transposable Element–Epigenetics Perspective. Trends in Ecology & 530 Evolution 31, 514–526. 10.1016/j.tree.2016.03.013. 531 10. Rinkevich, B. (2019). Coral chimerism as an evolutionary rescue mechanism to 532 mitigate global climate change impacts. Glob Change Biol 25, 1198–1206. 533 10.1111/gcb.14576. 534 11. Vidal-Dupiol, J., Harscouet, E., Shefy, D., Toulza, E., Rey, O., Allienne, J.-F., Mitta, 535 G., and Rinkevich, B. (2022). Frontloading of stress response genes enhances robustness to 536 environmental change in chimeric corals. BMC Biol 20, 167. 10.1186/s12915-022-01371-7. 537 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 22 12. Liew, Y.J., Howells, E.J., Wang, X., Michell, C.T., Burt, J.A., Idaghdour, Y., and 538 Aranda, M. (2020). Intergenerational epigenetic inheritance in reef-building corals. Nat. Clim. 539 Chang. 10, 254–259. 10.1038/s41558-019-0687-2. 540 13. Torda, G., Donelson, J.M., Aranda, M., Barshis, D.J., Bay, L., Berumen, M.L., 541 Bourne, D.G., Cantin, N., Foret, S., Matz, M., et al. (2017). Rapid adaptive responses to 542 climate change in corals. Nature Climate Change 7, 627–636. 10.1038/nclimate3374. 543 14. Berkelmans, R., and van Oppen, M.J.H. (2006). The role of zooxanthellae in the 544 thermal tolerance of corals: a ‘nugget of hope’ for coral reefs in an era of climate change. 545 Proc. R. Soc. B 273, 2305–2312. 10.1098/rspb.2006.3567. 546 15. Brener-Raffalli, K., Clerissi, C., Vidal-Dupiol, J., Adjeroud, M., Bonhomme, F., 547 Pratlong, M., Aurelle, D., Mitta, G., and Toulza, E. (2018). Thermal regime and host clade, 548 rather than geography, drive Symbiodinium and bacterial assemblages in the scleractinian 549 coral Pocillopora damicornis sensu lato. Microbiome 6, 39. 10.1186/s40168-018-0423-6. 550 16. LaJeunesse, T.C., Parkinson, J.E., Gabrielson, P.W., Jeong, H.J., Reimer, J.D., 551 Voolstra, C.R., and Santos, S.R. (2018). Systematic Revision of Symbiodiniaceae Highlights 552 the Antiquity and Diversity of Coral Endosymbionts. Current Biology 28, 2570-2580.e6. 553 10.1016/j.cub.2018.07.008. 554 17. Swain, T.D., Chandler, J., Backman, V., and Marcelino, L. (2017). Consensus 555 thermotolerance ranking for 110 Symbiodinium phylotypes: an exemplar utilization of a novel 556 iterative partial‐rank aggregation tool with broad application potential. Funct Ecol 31, 172–557 183. 10.1111/1365-2435.12694. 558 18. Buerger, P., Alvarez-Roa, C., Coppin, C.W., Pearce, S.L., Chakravarti, L.J., 559 Oakeshott, J.G., Edwards, O.R., and van Oppen, M.J.H. (2020). Heat-evolved microalgal 560 symbionts increase coral bleaching tolerance. Sci. Adv. 6, eaba2498. 10.1126/sciadv.aba2498. 561 19. Cunning, R., Silverstein, R.N., and Baker, A.C. (2015). Investigating the causes and 562 consequences of symbiont shuffling in a multi-partner reef coral symbiosis under 563 environmental change. Proc. R. Soc. B. 282, 20141725. 10.1098/rspb.2014.1725. 564 20. Matthews, J.L., Oakley, C.A., Lutz, A., Hillyer, K.E., Roessner, U., Grossman, A.R., 565 Weis, V.M., and Davy, S.K. (2018). Partner switching and metabolic flux in a model 566 cnidarian–dinoflagellate symbiosis. Proc. R. Soc. B. 285, 20182336. 10.1098/rspb.2018.2336. 567 21. Voolstra, C.R., and Ziegler, M. (2020). Adapting with Microbial Help: Microbiome 568 Flexibility Facilitates Rapid Responses to Environmental Change. BioEssays 42, 2000004. 569 10.1002/bies.202000004. 570 22. Hofreiter, M., Paijmans, J.L.A., Goodchild, H., Speller, C.F., Barlow, A., Fortes, G.G., 571 Thomas, J.A., Ludwig, A., and Collins, M.J. (2015). The future of ancient DNA: Technical 572 advances and conceptual shifts: Prospects & Overviews. BioEssays 37, 284–293. 573 10.1002/bies.201400160. 574 23. Baker, D.M., Weigt, L., Fogel, M., and Knowlton, N. (2013). Ancient DNA from 575 Coral-Hosted Symbiodinium Reveal a Static Mutualism over the Last 172 Years. PLoS ONE 576 8, e55057. 10.1371/journal.pone.0055057. 577 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 23 24. DiBattista, J.D., Reimer, J.D., Stat, M., Masucci, G.D., Biondi, P., De Brauwer, M., 578 and Bunce, M. (2019). Digging for DNA at depth: rapid universal metabarcoding surveys 579 (RUMS) as a tool to detect coral reef biodiversity across a depth gradient. PeerJ 7, e6379. 580 10.7717/peerj.6379. 581 25. Gomez Cabrera, M., Young, J.M., Roff, G., Staples, T., Ortiz, J.C., Pandolfi, J.M., and 582 Cooper, A. (2019). Broadening the taxonomic scope of coral reef palaeoecological studies 583 using ancient DNA. Mol Ecol 28, 2636–2652. 10.1111/mec.15038. 584 26. Smith, L.W., Barshis, D., and Birkeland, C. (2007). Phenotypic plasticity for skeletal 585 growth, density and calcification of Porites lobata in response to habitat type. Coral Reefs 26, 586 559–567. 10.1007/s00338-007-0216-z. 587 27. Wu, H.C., Dissard, D., Douville, E., Blamart, D., Bordier, L., Tribollet, A., Le Cornec, 588 F., Pons-Branchu, E., Dapoigny, A., and Lazareth, C.E. (2018). Surface ocean pH variations 589 since 1689 CE and recent ocean acidification in the tropical South Pacific. Nat Commun 9, 590 2543. 10.1038/s41467-018-04922-1. 591 28. Quigley, K.M., Davies, S.W., Kenkel, C.D., Willis, B.L., Matz, M.V., and Bay, L.K. 592 (2014). Deep-Sequencing Method for Quantifying Background Abundances of Symbiodinium 593 Types: Exploring the Rare Symbiodinium Biosphere in Reef-Building Corals. PLoS ONE 9, 594 e94297. 10.1371/journal.pone.0094297. 595 29. Lajeunesse, T., and Trench, R. (2000). Biogeography of two species of Symbiodinium 596 (Freudenthal) inhabiting the intertidal sea anemone Anthopleura elegantissima (Brandt). The 597 Biological Bulletin 199, 126–134. 10.2307/1542872. 598 30. Afgan, E., Baker, D., Batut, B., Van Den Beek, M., Bouvier, D., Čech, M., Chilton, J., 599 Clements, D., Coraor, N., and Grüning, B.A. (2018). The Galaxy platform for accessible, 600 reproducible and collaborative biomedical analyses: 2018 update. Nucleic acids research 46, 601 W537–W544. 602 31. Escudié, F., Auer, L., Bernard, M., Mariadassou, M., Cauquil, L., Vidal, K., Maman, 603 S., Hernandez-Raquet, G., Combes, S., and Pascal, G. (2018). FROGS: Find, Rapidly, OTUs 604 with Galaxy Solution. Bioinformatics 34, 1287–1294. 10.1093/bioinformatics/btx791. 605 32. LaJeunesse, T.C., Loh, W.K.W., van Woesik, R., Hoegh-Guldberg, O., Schmidt, 606 G.W., and Fitt, W.K. (2003). Low symbiont diversity in southern Great Barrier Reef corals, 607 relative to those of the Caribbean. Limnol. Oceanogr. 48, 2046–2054. 608 10.4319/lo.2003.48.5.2046. 609 33. Paradis, E., and Schliep, K. (2019). ape 5.0: an environment for modern phylogenetics 610 and evolutionary analyses in R. Bioinformatics 35, 526–528. 611 34. Claar, D.C., Tietjen, K.L., Cox, K.D., Gates, R.D., and Baum, J.K. (2020). Chronic 612 disturbance modulates symbiont (Symbiodiniaceae) beta diversity on a coral reef. Sci Rep 10, 613 4492. 10.1038/s41598-020-60929-z. 614 35. Thornhill, D.J., Lajeunesse, T.C., and Santos, S.R. (2007). Measuring rDNA diversity 615 in eukaryotic microbial systems: how intragenomic variation, pseudogenes, and PCR artifacts 616 confound biodiversity estimates. Molecular Ecology 16, 5326–5340. 10.1111/j.1365-617 294X.2007.03576.x. 618 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint 24 36. Dabney, J., Meyer, M., and Pääbo, S. (2013). Ancient DNA damage. Cold Spring 619 Harbor perspectives in biology 5, a012567. 620 37. Salas, R., Tillmann, U., John, U., Kilcoyne, J., Burson, A., Cantwell, C., Hess, P., 621 Jauffrais, T., and Silke, J. (2011). The role of Azadinium spinosum (Dinophyceae) in the 622 production of azaspiracid shellfish poisoning in mussels. Harmful Algae 10, 774–783. 623 38. Akselman, R., and Negri, R.M. (2012). Blooms of Azadinium cf. spinosum Elbrächter 624 et Tillmann (Dinophyceae) in northern shelf waters of Argentina, Southwestern Atlantic. 625 Harmful Algae 19, 30–38. 10.1016/j.hal.2012.05.004. 626 39. Hernández-Becerril, D.U., Barón-Campis, S.A., and Escobar-Morales, S. (2012). A 627 new record of Azadinium spinosum (Dinoflagellata) from the tropical Mexican Pacific. 628 Revista de Biología Marina y Oceanografía 47, 553–557. 629 40. Fu, Z., Piumsomboon, A., Punnarak, P., Uttayarnmanee, P., Leaw, C.P., Lim, P.T., 630 Wang, A., and Gu, H. (2021). Diversity and distribution of harmful microalgae in the Gulf of 631 Thailand assessed by DNA metabarcoding. Harmful Algae 106, 102063. 632 41. Tillmann, U., Elbrächter, M., Krock, B., John, U., and Cembella, A. (2009). 633 Azadinium spinosum gen. et sp. nov.(Dinophyceae) identified as a primary producer of 634 azaspiracid toxins. European Journal of Phycology 44, 63–79. 635 42. Sibat, M., Mai, T., Tanniou, S., Biegala, I., Hess, P., and Jauffrais, T. (2023). Seasonal 636 Single-Site Sampling Reveals Large Diversity of Marine Algal Toxins in Coastal Waters and 637 Shellfish of New Caledonia (Southwestern Pacific). Toxins 15, 642. 10.3390/toxins15110642. 638 43. Stat, M., Pochon, X., Franklin, E.C., Bruno, J.F., Casey, K.S., Selig, E.R., and Gates, 639 R.D. (2013). The distribution of the thermally tolerant symbiont lineage ( Symbiodinium clade 640 D) in corals from Hawaii: correlations with host and the history of ocean thermal stress. Ecol 641 Evol 3, 1317–1329. 10.1002/ece3.556. 642 44. Baker, A.C. (2003). Flexibility and Specificity in Coral-Algal Symbiosis: Diversity, 643 Ecology, and Biogeography of Symbiodinium. Annu. Rev. Ecol. Evol. Syst. 34, 661–689. 644 10.1146/annurev.ecolsys.34.011802.132417. 645 45. Thornhill, D.J., LaJeunesse, T.C., Kemp, D.W., Fitt, W.K., and Schmidt, G.W. (2006). 646 Multi-year, seasonal genotypic surveys of coral-algal symbioses reveal prevalent stability or 647 post-bleaching reversion. Marine Biology 148, 711–722. 10.1007/s00227-005-0114-2. 648 649 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 22, 2024. ; https://doi.org/10.1101/2024.03.21.584420doi: bioRxiv preprint