Whole transcriptome analysis reveals changes in expression of immune-related genes during and after bleaching in a reef-building coral

Climate change is negatively affecting the stability of natural ecosystems, especially coral reefs. The dissociation of the symbiosis between reef-building corals and their algal symbiont, or coral bleaching, has been linked to increased sea surface temperatures. Coral bleaching has significant impacts on corals, including an increase in disease outbreaks that can permanently change the entire reef ecosystem. Yet, little is known about the impacts of coral bleaching on the coral immune system. In this study, whole transcriptome analysis of the coral holobiont and each of the associate components (i.e. coral host, algal symbiont and other associated microorganisms) was used to determine changes in gene expression in corals affected by a natural bleaching event as well as during the recovery phase. The main findings include evidence that the coral holobiont and the coral host have different responses to bleaching, and the host immune system appears suppressed even a year after a bleaching event. These results support the hypothesis that coral bleaching changes the expression of innate immune genes of corals, and these effects can last even after recovery of symbiont populations. Research on the role of immunity on coral's resistance to stressors can help make informed predictions on the future of corals and coral reefs.


Summary
Climate change is negatively affecting the stability of natural ecosystems, especially coral reefs. The dissociation of the symbiosis between reef-building corals and their algal symbiont, or coral bleaching, has been linked to increased sea surface temperatures. Coral bleaching has significant impacts on corals, including an increase in disease outbreaks that can permanently change the entire reef ecosystem. Yet, little is known about the impacts of coral bleaching on the coral immune system. In this study, whole transcriptome analysis of the coral holobiont and each of the associate components (i.e. coral host, algal symbiont and other associated microorganisms) was used to determine changes in gene expression in corals affected by a natural bleaching event as well as during the recovery phase. The main findings include evidence that the coral holobiont and the coral host have different responses to bleaching, and the host immune system appears suppressed even a year after a bleaching event. These results support the hypothesis that coral bleaching changes 2015 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited.

Introduction
Environmental changes associated with climate change are affecting natural ecosystems [1][2][3]. Stressors, such as elevated sea surface temperature (i.e. thermal stress) and ocean acidification, are major causes of the decline of coral populations and deterioration of coral reefs [4][5][6][7][8]. Thermal stress has been associated with coral bleaching (i.e. disruption of the coral-algae symbiosis) [9], one of the most serious threats to coral health [10]. Coral bleaching affects the reproduction [11], growth, development [12] and health [13] of corals and weakens the structure and functionality of the reefs [14], ultimately affecting other reef inhabitants [6]. Coral bleaching events have become more frequent and devastating in the last several decades [15][16][17].
In a symbiotic relationship, the survival of both partners depends on their individual physiological capabilities and their combined resilience and resistance. In corals, the dissociation of the coral-algae symbiosis is associated with an increase in temperature of only a few degrees over a prolonged period of time [9]. During bleaching, the host tissue loses its symbiont cells, giving colonies a white appearance [18][19][20]. Depending on the intensity and duration of the stress, coral colonies can either recover normal symbiont densities and gain back typical coloration, or lose tissue, or die. Aside from colony mortality or tissue necrosis, other immediate effects of coral bleaching include: cessation of skeletal growth [21]; reduction in epithelial tissue thickness [22], larval survival [12,23] and protein synthesis [24]; appearance of diseases [25]; and increase in disease-related mortality [26]. Some of the effects can extend years after bleaching with bleached colonies showing a reduction in tissue biomass, as well as in protein and lipid content [27]. Bleached corals can also halt or delay the onset of oogenesis [11] and show an increase in disease prevalence compared with corals that did not bleach [28,29].
Evidence suggests that as coral bleaching become more common, so do disease outbreaks [28]. Diseases, such as white plague and yellow band disease in Orbicella faveolata, dark spots in Siderastraea siderea and black band in Colpophyllia natans, occur after bleaching events [28,29]. Additionally, coral bleaching can initiate the appearance of new infections (white plague in O. faveolata), or an increase in disease severity in diseases that were already present (yellow band and white plague in Orbicella spp. [28,29]). The relationship between coral bleaching and disease outbreaks suggests that the host's innate immune system is affected by bleaching and the changes persist long after the stressful conditions are over [30].
It has been suggested that within a species, corals living in naturally warm environments have an increased tolerance to temperature stress compared with colonies inhabiting cooler environments [31][32][33][34]. Evidence from transcriptomic analyses has shown that colonies exposed to high temperatures, for relatively low periods of time, have the capacity to increase expression of temperature-tolerant genes during thermal stress [31]. Additionally, under experimental conditions, two Caribbean corals (Acropora palmata and O. faveolata) appear to have similar responses in gene expression during thermal stress [35,36]. Responses of these species to increased temperatures include: increases in heat shock and antioxidant gene expression; decrease in expression of calcium homeostasis and ribosomal proteins; restructuring of the extracellular matrix; and rearrangement of the actin cytoskeleton [35,36]. However, the impacts of natural thermally induced bleaching on a coral's cellular and molecular machinery remain largely unknown.
Although genomic and transcriptomic resources have become common tools to study coral responses and tolerance to environmental changes, impacts of these events on the coral immune system remain largely understudied [30,37,38]. The purpose of this study was to assess the effects of a natural bleaching event on genes involved in the innate immune system of the Caribbean coral O. faveolata. In 2010, corals and coral reefs around the world experienced thermal stress that resulted in widespread coral bleaching [39][40][41]. In the Caribbean, reefs off the Puerto Rican coastline were no exception. In La Parguera (southwest Puerto Rico), reefs experienced elevated temperatures (up to 2 • C above average) from November 2009 through June/July 2010 (figure 1). Following this prolonged temperature anomaly, colonies from many coral species bleached   In La Parguera, the 2010 bleaching event (brown box) lasted from June-July to November-December and is linked to the continuous temperature anomaly (red box) observed between November 2009 and June-July 2010. During the temperature stress period, temperatures were 1 to 2 • C higher than the average for the region.
(November 2010), and at two time periods after the bleaching (March and October 2011). Results from the comparisons between bleached and unbleached colonies support the hypothesis that coral bleaching, due to thermal stress, affects the expression of innate immune genes of corals, and these effects can last at least 1 year after the event. Coral fragments (approx. 2 cm 2 ) were collected from all tagged colonies on three occasions: during bleaching (November 2010), after recovery of symbionts (four months-March 2011) and approximately a year after the first collection (11 months-October 2011). Samples were excised from the top of the colonies with a hammer and chisel, placed in sterile plastic bags and transported in seawater, at the collection site temperature, to the laboratory at the Department of Marine Sciences-University of Puerto Rico Mayagüez. Fragments were immediately flash frozen in liquid nitrogen, transported to Cornell University in dry ice and stored at −80 • C until further analysis.

RNA extractions, cDNA library preparation and sequencing
Each individual O. faveolata fragment was ground in liquid nitrogen using a mortar and pestle, and the resulting powder was placed in a 2.0 ml microcentrifuge tube. Total RNA was extracted using a modified Trizol/Qiagen RNeasy protocol as in Burge et al. [42]. After the ethanol precipitation step, following the manufacturer's instructions (Invitrogen, Life Technologies Corporation, Grand Island, NY, USA), RNA was cleaned from aqueous solution using an RNeasy column (Qiagen, Valencia, CA, USA). DNA was removed from the extracted solution using the Turbo DNA-free treatment according to the manufacturer's instructions (Ambion, Life Technologies Corporation). Removal of DNA was confirmed by using RNA (1 µl) as template in a quantitative PCR targeting 18S ribosomal DNA as previously described [43]. RNA

Transcriptome assembly
After removing adapters and low-quality reads (Trimmomatic [44]), the resulting reads from all sequenced libraries were combined and used to assemble a de novo metatranscriptome using the package Trinity [45]. This metatranscriptome included genes from the coral host, the algal symbiont (i.e. 'Symbiodinium spp.') and 'other-eukaryotes' (e.g. fungi, ciliates, endolithic algae, etc.) associated with O. faveolata. The poly-A selection in the library preparation reduces prokaryotic sequences, thus herein we primarily discuss the eukaryotic holobiont (i.e. coral host, 'Symbiodinium spp.' and 'other-eukaryotes').
To characterize the coral host transcriptome and elucidate the impacts of bleaching to the coral innate immune system, the metatranscriptome was filtered with genome and transcriptome data from different species or types of Symbiodinium and the O. faveolata genome. Symbiodinium data included the draft genome of Symbiodinium minutum (type B1, strain Mf1.05b-21 899 genes, 603 716 798 bp [46] [47]). Symbiodinium sequences were combined to create a single Symbiodinium reference data file (349 925 contigs-776 581 808 bp; electronic supplementary material, S1) that was aligned against the metatranscriptome using BLAT [48] with 90% identity and e-value < 0.000001 to filter spurious hits. Identities of the hits were filtered and duplicates removed, sequences were then retrieved from the metatranscriptome using the tool cdbfasta/cdbyank (http://sourceforge.net/projects/cdbfasta/), resulting in the 'Symbiodinium spp.' transcriptome.
metatranscriptome without the Symbiodinium-only genes was aligned (using BLAT as above) against the host genome (approx. 700 000 000 bp from non-symbiotic gametes) to acquire the 'O. faveolata' transcriptome and the 'other-eukaryotes' transcriptome. This genome is available on the O. faveolata Genome Consortium website: http://montastraea.psu.edu/.

Gene expression analysis and gene ontology
Reads from each of the samples (n = 12) were aligned against each of the transcriptomes to determine the expression levels within each of the components of the holobiont. Estimates of genes/contigs abundances for each sample and comparative gene expression analyses across samples and colony conditions through time were performed using Tophat, Cufflink and CummeRbund [49]. Changes in expression of genes involved in immunity or immune-related processes (e.g. immunity, signalling, response to stimulus) were further explored. We use the following designations to discern different conditions and time points in our sampling: bleached refers to corals that appeared white after losing their associated Symbiodinium cells in November 2010, during the height of the bleaching event. Even though these colonies regained their algal symbionts in March 2011, they are still referred to as bleached colonies or previously bleached colonies through the subsequent collection periods (March 2011 and October 2011). Corals that kept their pigmentation and algal cell populations in their tissues are referred to as unbleached.
Gene ontology annotations were initially determined using BLAST [50] for the metatranscriptome contigs/genes, and further explored with Protein Analysis Through Evolutionary Relationships [51] and Blast2GO [52] for genes showing significant gene expression differences (corrected p-values greater than 0.05). The metatranscriptome was blasted against the Swiss-Prot database. In Blast2Go, the annotations were obtained from the NCBI's nucleotide database, InterPro, GO, Enzyme Codes and KEGG. Enrichment tests among the differentially expressed genes were performed for the biological processes using the Fisher's exact test on Database for Annotation, Visualization and Integrated Discovery-DAVID v. 6.7 [53]. All biological processes except locomotion and biological regulation show p-values smaller than 0.05, suggesting that the processes are significantly enriched. Pathways involving genes with significant differences were obtained using PathVisio [54] from WikiPathways [55] and the Pathway Interaction Database [56].

Symbiodinium spp. type identity
The identity of the associated Symbiodinium types in each sample was determined with BLAST [50]. Reads from each of the samples were aligned against sequences of the internal transcribed spacer 2 (ITS2) of Symbiodinium types known to inhabit O. faveolata (S. fitti, D1a, S. minutum, C3, C3d, C3e, C7, C12). Alignments with 100% match were use as the correct identity. The symbiont identity of additional samples of the same colonies but collected in other months (September and December 2010 and August 2011) was determined using denaturing gradient gel electrophoresis of the ITS2 region [57][58][59][60][61][62].

Results
After trimming and quality filtering, a total of 387 512 512 pair-end reads ( (table 1 and electronic supplementary material, S3, S4 and S5, respectively). Some conserved genes may have been classified as being part of more than one transcriptome, resulting in overlap across the three transcriptomes. As expected, the coral was highly represented within samples with an average of 77.5% of the raw reads aligning with the 'O. faveolata' transcriptome (table 2). The other two components of the holobiont aligned with 7.6% ('Symbiodinium spp.' transcriptome) and 8.5% ('other-eukaryotes' transcriptome) of the raw reads.

Discussion
High-throughput sequencing is becoming a very common tool to answer important questions about the impacts of climate change on scleractinian corals [31,[63][64][65][66]. Our metatranscriptome analysis of the important reef-building coral O. faveolata is enhanced by the draft genome of the same host species. Although limited in sample numbers per time and condition, this study incorporates data from a natural coral bleaching event and subsequent recovery phase in the same coral colonies, resulting in a more comprehensive analysis of the processes involved in bleaching and recovery. The results from this study highlight the lasting effect coral bleaching has on key biological, physiological and immune pathways. maintained the same Symbiodinium types (S. fitti and D1a) and their profiles were similar to each other. On the other hand, bleached colonies showed shifts in Symbiodinium. During March 2011, bleached colonies acquired D1a in addition to S. fitti they already had and their gene expression profiles resembled that of the unbleached colonies that harboured the same types throughout. D1a has been proposed as a stress tolerant symbiont and may have helped these colonies recover [67,68]. The 'other-eukaryotes' portion of the holobiont may also have similar community shifts, but more resolution in the identity and function of these communities is needed.
The response of the holobiont reflected the expression levels of the less represented portions in our RNA libraries, leading to the masking of the expression of the coral host. Reads aligning to the 'Symbiodinium spp.' and 'other-eukaryotes' transcriptomes represented a low percentage of the total number of reads obtained during sequencing. This observation suggests that the overall condition of the colony is a result of the physiological tolerance of each of the elements of the holobiont, but can also be a reflection of the different genetic composition of each portion across time [69][70][71][72][73].

Bleaching affects several biological processes in the coral host, even a year after the event
Coral bleaching not only affects the coral-algae relationship but also acts on several aspects of the physiology and ecology of corals [12,[21][22][23][24][25][26]. Here, the regulation in the levels of gene expression in bleached colonies provides evidence of some of the affected processes. For example, reduction in epithelial tissue thickness [22] and protein synthesis [24] can be related to the downregulation of transcription, RNA processing and translation and protein synthesis and degradation during bleaching (November 2010; figure 6). Most of these processes are still affected (i.e. downregulated in bleached compared to unbleached colonies) a year after coral bleaching. However, a group of genes involved in protein synthesis and transport were upregulated in the bleached colonies 1 year after bleaching, perhaps in an effort to overcompensate the observed downregulation observed during the bleaching event.

Immune-related genes are affected by bleaching
Analyses of specific genes indicate that immune-related pathways such as apoptosis and the complement system are suppressed during bleaching and a year later in bleached colonies. Apoptosis plays a role in life-history processes, such as metamorphosis [74] and symbiosis [75], and it has been suggested to play a role in the defence of corals against pathogens [76,77]. Components of the tumour necrosis factor pathway and of capsase-8 were suppressed in O. faveolata bleached colonies. It is likely that the initial suppression of apoptosis in the bleached colonies (i.e. November 2010) is related to the mechanisms controlling bleaching. When apoptosis is blocked, bleaching can be reduced [78,79]. Although initially a mechanism to mitigate bleaching, continued downregulation of apoptosis 11 months after bleaching can have an immunosuppressive effect [76,77]. Bleached O. faveolata colonies are known to have higher disease prevalence than unbleached colonies during and after bleaching [11].
The complement system tags or selects foreign molecules for destruction [80,81]. A key component of this system is the mannan-binding lectin serine protease 1 (MASP1), which appears to be the exclusive activation factor of the pathway and produces large amounts (60%) of C2a, a compound responsible of C3 convertase formation [82]. MASP1 in bleached O. faveolata colonies was downregulated early during the recovery phase (March 2011), compared to during bleaching (November 2010) and 11 months after bleaching (October 2011). The pattern however was the opposite in unbleached colonies. Differences in expression of MASP1 between bleached and unbleached colonies suggest that the complement system is inactive, or less active, in bleached colonies. A less active component system indicates the lack, reduction or suppression of the immune system. In addition to apoptosis and the complement system, the cytoskeleton and translation are affected and genes such as RuvB [83,84] are downregulated in bleached colonies.
Environmental stressors have been linked to immunosuppression in other invertebrates [85]. The increase in new diseases and disease prevalence after bleaching [25,26] can be the result of the host immunosuppression during and after bleaching. The immune system is a well-regulated network of processes and pathways [86] that can interact with other cellular pathways. Upregulation of some genes involved in protein synthesis (e.g. peptidyl-prolyl cis-trans isomerase B, asparagine synthetase 3) and transport (e.g. ADP-ribosylation factor 2) and structural proteins (e.g. tubulin) a year after bleaching might be an attempt by the bleached colonies to compensate for their immunosuppression.

Concluding remarks
This study provides evidence that the coral holobiont and the coral host have different responses in terms of gene expression, during bleaching and through the recovery process. Here, we present evidence on previously unknown effects of bleaching; (i) Results of the metatranscriptome analysis indicate that each portion of the holobiont (i.e. 'O. faveolata', 'Symbiodinium spp.' and 'other-eukaryotes') has different responses to and recovery from bleaching; (ii) the coral host response appears to be masked by the responses of the associated organisms (i.e. 'Symbiodinium spp.' and 'other-eukaryotes'); (iii) bleached colonies may not successfully recover from bleaching, while unaffected colonies do not experience as intense changes in gene expression; and (iv) the effects of bleaching on the host immune system extend beyond recovery of the Symbiodinium population and appear to result in immune suppression. These results support the hypothesis that coral bleaching affects the expression of innate immune genes of corals, and these effects can last up to a year after the event.
Analyses on thermal resistance of corals suggest that some individuals might be able to overcome rising temperatures associated with climate change [31][32][33][34]. Bleaching impacts have proven erratic, and corals that in the past survived such events have been locally exterminated in the same locations after new bleaching events [87]. Results in this paper suggest that bleaching has long-term effects, but at the same time provide evidence that unbleached corals remain better prepared to fight pathogenic infections. The relation between coral bleaching and immunity in corals is complex and variable [30]. Studies emphasizing the role of coral immunity as an important aspect in coral's resistance to stressors can help improve predictions of the future of corals and coral reefs [88,89].
Data accessibility. All raw reads have been submitted to the National Center for Biotechnology Information Short Read Archive under the SRP022773 accession number.