to dark-induced coral bleaching in the reef coralStylophora pistillata
Alice Rouan1, Mélanie Pousse1, Eric Tambutté2, Nadir Djerbi1, William Zozaya1, Laura Capasso2,3, Didier Zoccola2, Sylvie Tambutté2*, Eric Gilson1,4*
1 Université Côte d’Azur-CNRS-Inserm, IRCAN, Nice, France.
2 Department of Marine Biology, Centre Scientifique de Monaco, Monte Carlo, Principality of Monaco.
3 Sorbonne Université, Collège Doctoral, F-75005 Paris, France
4 Department of Medical Genetics, CHU, Nice, France.
* Corresponding authors : Eric.GILSON@univ-cotedazur.fr, stambutte@centrescientifique.mc

Abstract

Telomere DNA length is a complex trait controlled both by multiple loci and environmental factors. Even though the use of telomere DNA length measurement, as a method of assessing stress accumulation and predicting how this will influence survival, is currently being studied in numerous human cohort studies, the importance of telomere length for stress response in ecological studies remains at its infancy. Here, we investigated the telomere changes occurring in the symbiotic coralStylophora pistillata that has experienced a continuous darkness over six months. This stress condition led to the loss of its symbionts, as what is also observed when bleaching occurs in the field at a large-scale due to climate changes and anthropogenic activities, threatening the worldwide reef ecosystem. We found that the continuous darkness condition was associated with telomere DNA length shortening and a downregulation of the expression of the telomere-associated protein Pot2. These results pave the way for future studies on the role of telomere in coral stress response and the importance of telomere dysregulation in endangered coral species.

Introduction

Scleractinian corals are metazoans that successfully built complex ecosystems, the coral reefs, which are among the most important biodiversity hotspots of the planet. However, all over the world, coral reefs are under threat due to global climate change and anthropogenic activities (Hughes et al. 2018b). These stressors can cause imbalances in the community of endosymbiotic partners hosted by corals such as viruses, bacteria, fungi and most notably, members of the algal dinoflagellate family Symbiodiniaceae. The breakdown of this intricate symbiosis leads to the loss of the intracellular symbionts and/or their pigments, a process that is called bleaching (Downs et al. 2013; Sully et al. 2019; Downs et al. 2009; Danovaro et al. 2008; Kushmaro et al. 1996; Rosenberg et al. 2009). Corals can either recover from bleaching by acquiring new symbionts or die in case of prolonged or repeated stresses. Of note, bleaching can be triggered in the laboratory under different experimental set-up (increases in temperature or light intensity, continuous darkness, hyposalinity, chemical pollutants and pathogen infection) but if food is provided, the corals can survive for long time periods even though different physiological processes can be affected (body mass loss, stop of calcification, growth arrest and polyp mortality) (Lyndby et al. 2020; Gardner et al. 2017).
In order to mitigate coral loss, there is currently an urgent need to understand the mechanisms involved in the response of corals to stress and among them, the pathways maintaining genome stability are essential. An accurate response to DNA damages caused by environmental stressors (radiation, heat, pollutants, nutrient fluctuations…) as well as normal metabolic processes (such as replication, respiration and, in the case of symbiotic corals, photosynthetic activity leading to daily oxidative switches), requires efficient DNA damage checkpoints, DNA repair and survival pathways to maintain genome stability and health (López-Otín and Kroemer, 2021). Therefore, the efficiency of DNA repair mechanisms is one of the main regulators of stress resistance and longevity in many organisms (Ma et al.,2016; Tian et al., 2019). However, aside from antioxidant defenses (Furla et al., 2005) and UV DNA repair (Reef et al., 2009) and recent report on increased levels of phosphorylated H2AX (a marker of DNA damage response) (Rodriguez‐Casariego et al., 2018) and Rad51 (a DNA double strand break repair protein) (Maor-Landaw and Levy, 2016) during heat stress, our knowledge on the role of genome maintenance pathways in environmental tolerance in corals remains limited.
Among the genome stability mechanisms, the chromosome ends, or telomeres, play a key role in stress response. Telomeres are composed of DNA repeats, (TTAGGG)n in most metazoans (Traut et al. 2007), bound by a protective protein complex, the shelterin, which protects the chromosome against unwanted DNA damage response (de Lange 2005). In humans, the shelterin complex is made of six proteins (TRF1, TRF2, TIN2, TPP1, POT1 and RAP1) (Giraud-Panis et al. 2013). TRF1 and TRF2 (Telomeric-repeat binding factor 1 and 2) directly bind the duplex telomeric DNA repeats, POT1 (Protection of Telomere 1) binds the single stranded telomeric DNA forming the 3’ overhangs of chromosomal DNA, while TIN2 and TPP1 bridge the double and single stranded parts of telomeric DNA. In addition, RAP1 is associated with TRF2. The telomerase holoenzyme, comprising the reverse transcriptase subunit TERT and its RNA template, is able to maintain telomere length during DNA replication (Gilson and Geli 2007). Functional telomeres protect chromosomes against instability and senescence and are also involved in a wide range of important processes such as inflammation, immunity, energetic metabolism and mitochondrial integrity, stemness as well as cellular differentiation (Ye et al., 2014). Changes in telomere structure occur during development, stress response and aging. In general, exposure to unfavorable environments leads to accelerated telomere shortening (Young et al., 2017). In birds, telomere DNA length can predict key life-history traits like growth, reproduction and lifespan (Monaghan, 2014). Thus, it was proposed that change in telomere DNA length is an adaptive strategy based on life-history regulation and environmental adaptation (Young 2018) and that telomere DNA length is a useful biomarker for past stress (Bateson, 2016).
Here we investigated the impact on telomere structure of continuous darkness stress (six months) in the model coral Stylophora pistillata (Esper, 1797). Notably, we observed that this stress condition leads to telomere DNA length shortening as well as a downregulation of the expression of a coral ortholog of a shelterin subunit. Since this stress is associated with the bleaching of the coral colony (Supplementary Figure 2), we discuss the possibility that the telomere response to continuous darkness is a consequence of the bleaching state.

Material and methods

Sampling
Colonies of the tropical coral S. pistillata were exposed to long-term darkness in an experimental aquarium setup for a long term exposure of six months. Briefly, coral fragments were kept in aquaria supplied with Mediterranean seawater (exchange rate of 70% per hour) at a salinity of 38 g liter−1, temperature of 25°C, pHT 7.94 ± 0.02 and for the control condition at an irradiance of 230 μmol photons m−2s−1 on a 12:12 photoperiod. Both dark and control conditions corals were fed daily with frozen rotifers and twice a week with live artemia nauplii. The first experiment D1 was conducted from July 2018 to January 2019 were a part of the S1 colony was separated and put in dark condition, 6 branches of the bleached colony were sampled and 5 of the control one, the second experiment D2 was conducted from June 2019 to December 2019 were a different S. pistillata colony was divided in 2 part, each one kept in either of the two experimental condition, 4 branches from each colony part were sampled.
DNA extraction
DNA was extracted on the first 3-4 cm of a colony branch. All steps until cell lysis were done on ice. The fresh coral tissue was removed from the skeleton using the “air_brush” technique in 50mL of extraction buffer (NaCl 550mM, EDTA 0,2M (Tambutté et al., 2007)) in a Whirl-Pak bag (Nasco) under the hood, then centrifuged for 10 minutes at 5 000g (Beckman Coulter, Avanti J-E centrifuge, rotor JA-18) at 4°C. The tissue pellet was kept on ice to perform high molecular weight DNA extraction following the Midi Kit Cell and Blood culture (QIAGEN). Briefly, cell lysis was carried out for 2 hours at 50°C after adding 19 mL of G2 buffer (800 mM guanidine HCl ; 30 mM Tris·Cl, pH 8.0 ; 30 mM EDTA, pH 8.0 ; 5% Tween-20 ; 0,5% Triton X-100, QIAGEN) with 190 mL RNase A (ThermoFisher, ref : 12091039 , 20mg/mL) and 400 mL of protease (QIAGEN, 1 UA/mL) to the tissue pellets. Supernatant was poured in the anionic column after a 5 min centrifuge at 5 000g and 4°C, columns were washed following the kit instructions, DNA was eluted with 5mL of warm QF buffer (1.25 M NaCl ; 50 mM Tris·Cl, pH 8.5; 15% isopropanol, QIAGEN). Precipitation was performed adding 3.5mL of 2-propanol (Sigma), after mixing by inverting, tubes were stored overnight at 4°C. Two washing steps were performed adding 1mL of 70% ethanol (sigma) to DNA pellets after 30 min centrifuge at 15 000g and 4°C before adding 200 mL of TE buffer (Tris-Cl pH8 10mM, EDTA 1mM) for long-term storage. Pellets were resuspended at 50°C for 1 hour. To test DNA quality and quantity both Nanodrop (ThermoFisher) and agarose gel were used, respectively 1 mL of DNA and 1,5 mL of loading buffer (New England Biolabs, 6X) in a final volume of 10 mL were loaded in a 1.2% agarose gel run for 30 minutes at 100V in Tris-Borate-EDTA (TBE) 1X. Agarose gel was incubated for 10 minutes in 4 mg/mL ethidium bromide and DNA was revealed under UV (GelDoc Transilluminator, BioRad), the average of three measurements was used. A diluted aliquot was prepared for enzymatic digestion before storing the stock DNA at -20°C.
Telomere Restriction Fragment assay:
Dilution and DNA digestion
To digest non telomeric DNA sequences, we added 60U HinfI (R0155M, New England Biolabs) and 60U RsaI (R4374, Promega) to of 2,5 mg of DNA diluted in mqH2O and “CutSmart Buffer” 1X (New England Biolabs) in a final volume of 40mL, left overnight at 37°C.
Southern Blot
We measured telomere length using the Telomeric Restriction Fragment assay (TRF) using the Southern blotting procedure (Herbert et al., 2003) digested DNA was loaded on a 1,2% agarose gel for a 3h30 run in TBE 0.5X at 9 volts/cm in a CHEF-DR-II (BioRad). The run timing makes it suitable for 2 runs a day. After a 5 min wash in mQ H2O agarose gel was stained in Ethidium bromide bath (4 mg/mL) under agitation for 30min and imaged the gel (Typhoon, GE Healthcare, Fluorescence, Method=Alexa Fluor 555, Laser=532nm, PMT=ch.1,700V, Resolution = 50 µm). Denaturation was done in (1 M NaCl ; 0.5 M NaOH) for two incubations of 20 min followed by a neutralization step with two 20 min incubation in 1M ammonium acetate. The DNA was transferred to a Hybond N+membrane (GE Healthcare) overnight in SSC (20X). DNA was cross-linked on the membrane using the Ultraviolet Crosslinker (UVP) at 1200 x 100µ Joules. Membrane was gently washed in mQH2O before storage at room temperature for (12h up to 48h) before 1h pre-hybridization in Denhardt’s buffer (10X) at 40°C. We prepared stocks of Denhardt’s 100X buffer (2% Ficoll 400, 2% Bovine Serum Albumine, 2% Polyvinylpyrrolidone) filtered (0.2µm) and stored at -20°C.
Probe synthesis
We prepared probe template 40uM stocks annealing two primers per probe. Probe targeting (TTAGGG) sequence will be called T2 and the one targeting (TTTAGGG) sequence will be referred to as T3. We mixed 25 of Primer F and 25 of Primer R (Primer T2AG3_F : GGGTTAGGGTTAGGGTTAGGGAAA and T2AG3_R : TTTCCCTAACCCTAA, PrimerT3AG3_F : GGGTTTAGGGTTTAGGGTTTAGGGAAA and T3AG3_R : TTTCCCTAAACCCTAAA) with STE 5X buffer (Tris 50mM pH8, NaCl 250mM, EDTA 5mM) and heat it to 95°C to therefore gradually let it cool down to room temperature before being stored at -20°C. We mixed on ice 2.5 of dATP (10 mM) and 2.5 of dTTP (10mM, New England Biolabs, R0192) with 3of NEB 2 buffer (10X), 15of mQH2O, 5 of gamma-dCTP32 and 1 Klenow (3’–>5’ exo-) (M0212S, NEB) incubated in a thermocycler (ThermoFisher Scientific) for a (30min 25°C, 5min 98°C, 5min 25°C) program. Then we purified the probes following the ProbeQuant G50-micro columns (GE, Healthcare) instructions. Probes were denatured at 95°C for 5 min and used or stored at -20°C.
Hybridization
Membranes were pr hybridized for 1h at 40°C in hybridization Denhardt’s buffer (10X). We added 5µL of denatured T3 probes to 15mL of hybridization buffer for an overnight hybridization at 40°C. Washing steps included a 2x20min wash in 2X SSC, a 30min step in (2X SSC, SDS 0,2%) and a final H2O washing step. Membrane was carefully wrapped in plastic sarran film and applied to a phosphor-screen (brand) for 60h. We imaged the phospho-screen on the Typhoon (GE Healthcare, Phosphorimaging, Method=[Phosphor], Laser=635nm, PMT=ch.1,1000V, Resolution = 50 µm), the screen was flashed (brand of device). Membrane was stripped with NaOH 0,4M for 30min at 42°C, then for 30min in (SDS 1%, SSC 0.1X, Tris-HCl pH7.4 1M) to remove the T3 probe. To control the stripping, membrane we exposed the membrane to a flashed phosphor-screen for 12h. The screen was imaged as described above, in the absence of signal the membrane was pre hybridized for 1h and hybridized overnight at 45°C with the T2 probe, otherwise the membrane was stripped again until no signal was imaged after 12h of exposure. Washing steps and imaging were carried out as described above with a 24h exposure of the phosphor-screen.
Image analysis
Telomere Restriction Fragment (TRF) images signal was extracted using ImageQuant (GE Healthcare) 1D gel analysis mode and manual lane creation. Ladders signal intensity were extracted from Ethidium bromide gel images, setting the lane upper limit at the gel wells bottom. Pixel position of ladder peaks were manually reported. Host and symbiont telomere signals were extracted from the phospho screen, exposed to radioactive labelled membrane, images setting sample lane upper limit at the top of the membrane. Efficiency of stripping step between the two probes hybridization was assessed imaging phospho screens overnight exposed to stripped membranes.
Telomere Length measurements
Single lane intensity files were fused in RStudio. Ladder peaks exact position was extracted in R, searching for the maximum intensity in a 10-pixel perimeter around the manually reported peaks. Fitted linear model coefficients (a,b) of log2 ladder size (kb) against peak pixel position were calculated for high molecular weight (48.5-15kb) and low molecular weight (10-1kb) using the lm function of “stats” R package. Coefficients were used to transform samples intensity scale from pixel to base pair (bp) () in Excel using the high molecular weight coefficient for the upper part and the low molecular coefficient for the lower one. Depending on their position on the membrane samples were divided in left, right and middle to be scaled to the closest ladder. Intensity signals were imported in R using the read_excel from “readxl” package, background correction was automatically computed to level the signal by subtracting to each position the minimal intensity. Intensity was normalized by the size to avoid probe number hybridization bias, intensity above 2 kb were discarded to avoid genomic noise, interstitial telomeric sequence noise and normalization bias (<1kb). Ponderate telomere mean was calculated as in (Li and de Lange 2003), median, quartiles, intensity signal, smoothed signal and telomere measurements were plotted for each sample using the ggplotfunction from “ggplot2” package and measurements were saved ascsv files.
Transcriptome
RNA extraction
Coral pieces of approximately 2 cm were sampled just below the coral piece dedicated to DNA extraction, at least 4 cm from the apex of the branch. Samples were flash frozen in liquid nitrogen in WhirlPack bags, and then crushed in small pieces with a press and fragments were put in 600uL of lysis buffer RLT in PowerBeads Glass 0.1mm (Qiagen), samples were lysed in Bead Beating instrument for 4 minutes with 30 cycles/s. Samples were spin at max speed for 3 minutes and the 600uL of supernatant was transferred to a fresh tube. Add 600uL (1V) of cold EtOH 70% to the column (Qiagen, RNeasy Mini Kit), gently mixed by pipetting, immediately apply 600uL on the column and centrifuge 15s at 8 000g (10 000 rpm), discard the flow-through, repeat this step. Add 350 uL of RWI buffer to the column and centrifuge 15s at 8 000g (10 000 rpm), discard the flow-through. Take a DNAse, Rnase free, (Qiagen, 79254) stored at -20°C, add 70ul of buffer RDD 4°C and apply to the column. Leave the column at RT for 15min before adding 350 uL of RWI buffer and centrifuge 15s at 8 000g (10 000 rpm), discard the flow-through and add 500uL of RPE buffer, repeat the centrifugation, add again 500uL of RPE buffer and make a 2min at 8 000g (10 000 rpm) centrifuge before discarding the flow-through and putting the column in a clean 1.5ml tube. Add 30ul of RNase free water to the column and centrifuge for 1min at 8 000g (10 000 rpm), keep the tube on ice for further experiment and store at -80°C. RNA quantity was measured on a microplate spectrophotometer Epoch (Bioteck) and the quality was assessed by running a 1% agarose gel labelled with Ethidium Bromide and imaged under UV-exposure in a Fusion Fx7.
RNA sequencing
Paired-end sequencing (read length: 2x150bp) was performed by NovoGene company using HiSeq sequencer, raw data are available with accession number GSE171268 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE171268). Three biological replicates per condition were performed with roughly 45 million reads per sample. Reads were trimmed using fastp software (Chen et al., 2018). Reads were mapped on the S.pistillatatranscriptome from NCBI (GCF_002571385.1) using Salmon software (v0.11.4) (Patro et al., 2017). The differential genes expression analysis was performed using DESeq2 R package (Love et al., 2014). A cutoff of 0.05 was applied on adjusted p-values, obtained using Benjamini and Hochberg correction.
Ontology and pathway analysis
Gene Ontology (GO) analysis was performed using the “topGO” R package (Alexa et al., 2006). For the cnidarian analysis, all GO terms identified in S.pistillata were used as background. Ingenuity Pathway Analysis (Krämer et al. 2014) (QIAGEN,https://www.qiagenbioinformatics.com/products/ingenuitypathway-analysis) was performed on human ortholog genes. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was performed using “gage” and “KEGGREST” R packages (Kanehisa, 2000). The human ortholog genes associated with the S.pistillata log2FoldChange values were used to conduct the analysis. Human orthologs were identified using the protein sequences. Blastp (Altschul et al. 1990) on UniProtKB/Swiss-Prot database restricted to Homo sapiens species was performed with a 1e-5 e-value threshold. For each protein, the best hit was selected, based on e-value and bit-score.
Phylogeny
Ortholog sequences of S.pistillata shelterin proteins were curated using PSI-BLAST (Altschul et al, 1997) on the non-redundant (Nr) protein database from the NCBI. The search was restricted to Metazoa group. The retrieved sequences were aligned using MAFFT v7.310 with L-INS-I algorithm (Katoh and Standley, 2013). The multiple alignment was trimmed with TrimAl v1.4 with the gt 0.6 option (Capella-Gutierrez et al., 2009). Then the best substitution matrix was assessed using ModelTest (Darriba et al., 2020). Finally, the trees were built using raxmlGUI 2.0, a graphical interface for the implementation of RAxML Next Generation and the transfer bootstrap expectation branch support (Edler et al., 2020). Tree visualization was done with iToL v6.1 (Letunic and Bork, 2019). Multiple alignment for POT1 and POT2 were wrapped using Alignment Annotator (Gille et al., 2014).
Quantitative Polymerase Chain Reaction (PCR)
Primers were designed using PrimerBlast (Ye et al., 2012) and sequences are available in (Supplementary Table 3). Reverse transcription was carried for 1ug of RNA samples using the Applied Biosystems reverse transcription kit for 1h at 37°C followed by 5min of denaturation at 95°C. qPCRs were made on cDNAs obtained using ROCHE’s Fast universal Sybr Greenmaster (ROX) on a StepOne plus thermocycler (Applied Biosystems). Each sample was triplicated on the 96 wells PCR plate containing the interest gene primers and the two control genes. In addition to the samples from the RNAseq, samples from the same experiment were added to the qPCR, two additional control samples (S1n2 and S1n4) as well as two treated samples (S1Bn2 and S1Bn6).
Statistics
Telomere length measurements were averaged from two independent Telomeric Restriction Fragment assay, homoscedasticity and normal distribution were tested on R using (fligner.test andshapiro.test functions) and significant differences was calculated using the t_test from the “rstatix” packages with a Bonferroni pvalue adjustement. Significant differences are reported as stars ( (ns) non significant, (*) P<0.05, (**) P<0.01, (***) P<0.001.

Results