Differential gene expression analysis of symbiotic and aposymbiotic Exaiptasia anemones under immune challenge with Vibrio coralliilyticus

Abstract Anthozoans are a class of Cnidarians that includes scleractinian corals, anemones, and their relatives. Despite a global rise in disease epizootics impacting scleractinian corals, little is known about the immune response of this key group of invertebrates. To better characterize the anthozoan immune response, we used the model anemone Exaiptasia pallida to explore the genetic links between the anthozoan–algal symbioses and immunity in a two‐factor RNA‐Seq experiment using both symbiotic and aposymbiotic (menthol‐bleached) Exaiptasia pallida exposed to the bacterial pathogen Vibrio coralliilyticus. Multivariate and univariate analyses of Exaiptasia gene expression demonstrated that exposure to live Vibrio coralliilyticus had strong and significant impacts on transcriptome‐wide gene expression for both symbiotic and aposymbiotic anemones, but we did not observe strong interactions between symbiotic state and Vibrio exposure. There were 4,164 significantly differentially expressed (DE) genes for Vibrio exposure, 1,114 DE genes for aposymbiosis, and 472 DE genes for the additive combinations of Vibrio and aposymbiosis. KEGG enrichment analyses identified 11 pathways—involved in immunity (5), transport and catabolism (4), and cell growth and death (2)—that were enriched due to both Vibrio and/or aposymbiosis. Immune pathways showing strongest differential expression included complement, coagulation, nucleotide‐binding, and oligomerization domain (NOD), and Toll for Vibrio exposure and coagulation and apoptosis for aposymbiosis.

have since used RNA-Seq data to produce some of the first profiles of anthozoan innate immunity (Anderson, Walz, Weil, Tonellato, & Smith, 2016;Burg, Prentis, Surm, & Pavasovic, 2016;Fuess, Mann, Jinks, Brinkhuis, & Mydlarz, 2018;Fuess, Pinzón, Weil, Grinshpon, & Mydlarz, 2017;Libro, Kaluziak, & Vollmer, 2013;Libro & Vollmer, 2016;Pinzón et al., 2015;Poole & Weis, 2014;Vidal-Dupiol et al., 2011;Weiss et al., 2013). To date, at least nine studies have profiled the immune response of corals and their anthozoan relatives, and the data suggest that the immune response varies across anthozoans and/or immune exposures. For example, Weiss et al. (2013) studied the response of the reef coral Acropora millepora to the bacterial cell wall derivative muramyl dipeptide (MDP) and observed the up-regulation of GTPases of immunity-associated proteins (GiMAPs), which are primarily associated with immunity in vertebrates (Wang & Li, 2009) and plants (Liu, Wang, Zhang, & Li, 2008). Vidal-Dupiol et al. (2011) compared the transcriptomic responses of the reef coral Pocillopora damicornis to thermal stress and Vibrio coralliilyticus infection and observed that immune pathways-including Toll/TLR, complement, prophenoloxidase, and the leukotriene cascade pathways-were up-regulated due to Vibrio exposure. Libro et al. (2013) compared the immune response of healthy and White Band Disease (WBD) infected Acropora cervicornis coral using RNA-Seq and found that C-type lectins, ROS production, arachidonic acid metabolism, and allene oxide production were strongly up-regulated in diseased corals (Libro et al., 2013). Up-regulation of C-type lectins and ROS production are hallmarks of phagocytosis, and the metabolism of arachidonic acid via the allene oxide pathway has been linked to eicosanoid synthesis in wounded corals (Lõhelaid, Teder, Tõldsepp, Ekins, & Samel, 2014). Interestingly, Libro et al. (2013) did not identify strong up-regulation of genes associated with the classic innate immune pathways such as Toll-like receptor pathway or prophenoloxidase pathway.
Reef-building corals and other anthozoans like the symbiotic anemone Exaiptasia are also well known for their symbiotic relationship with the dinoflagellate Symbiodinium (also called zooxanthellae).
This symbiosis presents a challenge with regard to the immune system because both pathogens and symbionts can elicit an allorecognition response, with the difference being that pathogens are typically eliminated, while symbionts are allowed to coexist within vacuoles in the endodermis of host cells (Kazandjian et al., 2008;Wakefield, Farmer, & Kempf, 2000) providing the anthozoan host up to 95% of its energy as translocated polysaccharides (Falkowski, Dubinsky, Muscatine, & Porter, 1984). Symbiosis requires clear communication between the host and symbiont. During the establishment of symbiosis, the anthozoan host must be able to recognize symbionts, engulf them in phagosomes, and shield these phagosomes from destruction (Davy, Allemand, & Weis, 2012). This suggests a clear link between symbioses and immunity wherein symbionts evade the immune response. Arrest of phagosomal maturation by Rab GTPases (Davy et al., 2012) and suppression of immune responses by transforming growth factor beta (TGFβ) (Detournay, Schnitzler, Poole, & Weis, 2012) have been identified as potential mechanisms by which symbionts are shielded from destruction by the immune system. Once symbiosis is established, the host must regulate the growth of the symbionts and remove dead or dying symbionts (Davy et al., 2012).
Regulation of nutrients has been identified as one mechanism by which the host can prevent overgrowth of the dinoflagellates (Davy et al., 2012).
Bleaching occurs when the symbionts are degraded or expelled by the coral host due to factors like thermal stress (Fitt, Brown, Warner, & Dunne, 2001), UV exposure (Gleason & Wellington, 1993), and disease (Libro et al., 2013). In addition to these naturally occurring stressors, chemical agents have been identified to deliberately induce bleaching in the laboratory for manipulative studies. These include menthol (Wang, Chen, Tew, Meng, & Chen, 2012) and photosynthesis inhibitors (Jones, 2004) that result in bleaching. Several mechanisms have been identified that result in the degradation and expulsion of the symbionts, including apoptosis, necrosis, and symbiont digestion via autophagy (symbiophagy; Dani et al., 2016), and the mechanisms vary depending on the type of stress. Apoptosis and necrosis predominate in heat-stress bleaching, while symbiophagy predominates in menthol bleaching (Dani et al., 2016;Wang et al., 2012). Arrest of phagosomal maturation is required for the establishment of symbiosis, and Dani et al. (2016) suggest that a re-engagement of phagosomal maturation is involved in the breakdown.
A number of transcriptomic studies of anthozoan bleaching have shown varied immune responses. Mansfield et al., 2017 found that NF-κβ protein levels increase after bleaching and decrease after recolonization in Exaiptasia. Pinzón et al. (2015) found that 1 year after a bleaching event in Orbicella faveolata colonies, 17 immune genes within tumor necrosis factor pathway, apoptosis, cytoskeleton, transcription, signaling, and cell adhesion and recognition were downregulated. Seneca and Palumbi (2015) found that the transcriptome response of Acropora hyacinthus exposed to heat varied widely between the initial heat exposure and the bleaching response 15 hr later, and the later response included up-regulation of immune and apoptosis pathways including Toll-like receptor and C-type lectins. In this study, we explore whether breakdown of symbiosis triggered by exposure to menthol alters the subsequent immune response to the coral pathogen Vibrio coralliilyticus.
The symbiotic anemone Exaiptasia has become a powerful model for studying symbiosis and immunity in symbiotic anthozoans because (a) it is a hardy animal that can be made aposymbiotic experimentally by exposing it to cold and heat stress (Lehnert et al., 2014), as well as by treating it with compounds like menthol (Matthews, Sproles, & Oakley, 2016), (b) it can be propagated clonally (Sunagawa et al., 2009), and (c) a well-annotated genome for Exaiptasia now exists (Baumgarten et al., 2015). Limited gene expression data also exist for Exaiptasia comparing aposymbiotic and symbiotic anemones (Lehnert et al., 2014), Exaiptasia exposed to pathogens (Poole, Kitchen, & Weis, 2016), and Exaiptasia colonized by heterologous symbionts . Lehnert et al. (2014) used RNA-Seq to compare symbiotic and aposymbiotic anemones and identified 900 differentially expressed genes involved in metabolite transport, lipid metabolism, and amino acid metabolism. Poole et al. (2016) used qPCR to compare complement activity in response to colonization with Symbiodinium and the response to pathogen exposure (Serratia marcescens). Within the complement pathway, B-factor 1 and MASP were up-regulated and B-factor 2b down-regulated in response to both pathogen exposure and symbiont colonization. Matthews et al. (2017) used RNA-Seq to profile immune and nutrient exchange activity in response to colonization with Symbiodinium trenchii versus its normal symbiont, Symbiodinium minutum. The expression pattern after colonization with the heterologous S. trenchii was intermediate between the aposymbiotic state and the normal (S. minutum) symbiotic state, with up-regulation of innate immune pathways in response to heterologous colonization.
In this study, we explore the genetic links between the anthozoan-algal symbioses and immunity in a two-factor RNA-Seq experiment using both symbiotic and aposymbiotic Exaiptasia exposed to the bacterial pathogen Vibrio coralliilyticus. Menthol bleaching was used to compare symbiotic (untreated) versus aposymbiotic (menthol-treated) anemones where the hypothesized mechanism of menthol bleaching is thought to be the activation of autophagic digestion of Symbiodinium cells (symbiophagy) as part of host innate immunity (Dani et al., 2016). The bacterial pathogen Vibrio coralliilyticus was used to initiate the immune response of Exaiptasia 72 hr after exposure to menthol. The two-factor design comparing Vibrio and aposymbiosis as factors allowed us to identify gene expression patterns that were due to Vibrio and/or symbiotic state as well as any interactions between pathogen exposure and symbiotic state.

| ME THODS
Wild Exaiptasia pallida were obtained from Carolina Biological Supply. These anemones collected off the coast of North Carolina are the source population from which the widely used cc7 clonal population was developed (Sunagawa et al., 2009). Anemones were maintained in 6-well culture plates and held under 24-watt t5 fluorescent lights for a 12-hr light cycle. To avoid any bias based on lighting intensity or other positional effects, the wells of the plates were randomly assigned to six groups ( Figure 1). Thirty-six anemones were divided into aposymbiotic and symbiotic. The aposymbiotic and symbiotic groups were then further subdivided with six symbiotic and six aposymbiotic anemones being sacrificed to estimate Symbiodinium densities due to menthol exposure leaving six symbiotic and six aposymbiotic anemones for Vibrio treatment and six symbiotic and six aposymbiotic controls. At the end of the Vibrio exposure, we produced RNA-Seq data for six replicate anemones for each of the four groups: Vibrio/symbiotic, control/symbiotic, Vibrio/ aposymbiotic, and control/aposymbiotic. After plating and group assignment, anemones were maintained in the wells for a 1-week acclimation phase. They were exposed to a 12-hr day/night cycle with a light intensity of 70 μmol quanta m −2 s −1 . To avoid contamination of the RNA with any partially digested food, the anemones were not fed during the acclimation phase. Aposymbiotic anemones were produced by exposure to 0.58 mM menthol/ASW using a modified version of the protocol outlined by Wang et al. (2012). The menthol exposure was on a 72-hr cycle, with a 24-hr menthol exposure followed by a 48-hr resting period in 0.2 μm filtered natural seawater (FNSW). During the menthol treatment cycle, 18 anemones had water replaced with 0.58 mM menthol/FNSW, and 18 had water replaced with fresh FNSW. The degree of menthol bleaching was measured by homogenizing anemones with BioMasher mortar and pestle sets in 1 ml After the menthol treatment and resting period were complete, 12 anemones were then exposed to live Vibrio at a concentration of 10 8 CFU/ml in 0.2 μm filtered natural seawater (FNSW) using Vibrio coralliilyticus strain BAA-450™ from ATCC ® (Ben-Haim et al., 2003), and 12 controls were exposed to FNSW. The Vibrio inoculate was produced by centrifuging marine broth cultures for 2 min at 5,000 rcf, drawing off the broth, and re-suspending the pellet in FNSW. The anemones were exposed to either Vibrio or FNSW for F I G U R E 1 Outline of experimental design -Thirty-six anemones were divided into aposymbiotic and symbiotic. The aposymbiotic and symbiotic groups were then further subdivided with six symbiotic and six aposymbiotic anemones being sacrificed to estimate Symbiodinium densities due to menthol exposure leaving six symbiotic and six aposymbiotic anemones for Vibrio treatment and six symbiotic and six aposymbiotic controls 24 hr and then immediately homogenized in Tri-Reagent for downstream extraction of total RNA.
After the menthol treatment and Vibrio exposures were complete, the anemones were homogenized using BioMasher mortar and pestle sets for total RNA extraction. Each anemone was first homogenized in 900 μl Tri-Reagent, and then, the 900 μl was divided into three separate tubes to which an additional 600 μl Tri-Reagent was added to ensure sufficient Tri-Reagent volume to lyse the cells completely. Total RNA was isolated using the Tri-Reagent manufacturer's protocol. Total RNA was quantified on an Agilent BioAnalyzer to obtain concentrations and RNA integrity number (RIN) scores. For each of the 24 samples, the RNA isolate with the highest RIN score (mean score 6.88) was selected to  (Patro, Duggal, Love, Irizarry, & Kingsford, 2017), and the transcript counts were imported into DESeq2 (Love, Huber, & Anders, 2014) using tximport (Soneson, Love, & Robinson, 2015). We performed two-way ANOVA on the counts of trimmed and aligned reads and found no significant differences based on Vibrio (F = 1.022, p = 0.324) or symbiotic state To facilitate KEGG (Kanehisa, Sato, Kawashima, Furumichi, & Tanabe, 2016) pathway analysis, the transcripts were mapped from Exaiptasia predicted protein IDs to KEGG ortholog IDs. Predicted coding sequences were extracted from the Exaiptasia genome annotation file using the gffread utility from the Cufflinks (Trapnell et al., 2012) package yielding 26,042 sequences. The FASTA file produced by gffread was aligned to Swiss-Prot (The Uniprot Consortium, 2017) using blastx (Camacho et al., 2009). Swiss-Prot hits were filtered using an e-value cutoff of 1e −10 and a minimum query coverage of 50%. Where blastx hits mapped to multiple KEGG orthologs, the ortholog with the query coverage closest to 100% was selected.
For multiple hits with identical coverage, the lowest e-value was chosen. This same process was then applied in the reverse direction to eliminate duplicate Exaiptasia to KEGG ortholog mappings.
The filtering and mapping were accomplished with custom Perl and R scripts, yielding 4,807 one-to-one Exaiptasia to KEGG ortholog mappings. This mapping table is available in the Dryad repository aipAnnot.csv.
PERMANOVA and MDS analyses were used to identify transcriptome-wide differences in gene expression due to symbiotic state or Vibrio exposure. Hellinger-transformed DESeq2-normalized counts were analyzed using PERMANOVA to identify transcriptome-wide differences in expression patterns using the adonis function within the R package Vegan (Oksanen et al., 2018) with 999 permutations and formula Vibrio * Symbiotic State. MDS was used to visualize transcriptome-wide differences between groups (Oksanen et al., 2018). A two-factor negative binomial GLM implemented in DESeq2 (Love et al., 2014) was used to identify differentially expressed (DE) genes that differed due to Vibrio and aposymbiosis as well as the interaction. The R package ESGEA (Alhamdoosh et al., 2017) was used to identify KEGG pathways that showed significant enrichment due to Vibrio or aposymbiosis, and the R package GOseq (Young, Wakefield, Smyth, & Oshlack, 2010) was used to identify pathways overrepresented in DE genes.

| RE SULTS
RNA-Seq data were produced for 24 anemones with six replicates each for the four treatment groups: (a) Vibrio/symbiotic, (b) con-  TA B L E 1 Multivariate PERMANOVAs show significant differences in transcriptome-wide gene expression patterns due to Vibrio exposure (p = 0.001) and aposymbiosis (p = 0.007), but not the Vibrio-aposymbiosis interaction

| Univariate analyses
Univariate, negative binomial GLMs in DESeq2 (Love et al., 2014) were used to identify genes that were differentially expressed due   For the transport and catabolism pathways (Table 4) For cell growth and death (

| Immune system response
Among the immune pathways, there was strong evidence that the complement and coagulation cascade was responding to both Vibrio and aposymbiosis, whereas expression of NOD/TLR pathway, chemokine, and antigen processing was initiated primarily by Vibrio exposure. The stimulation of complement and coagulation cascade pathway and NOD/TLR pathway indicates that bacterial immune challenge by Vibrio involves two of the three primary innate immune pathways in invertebrates; we did not find significant evidence for stimulation of the prophenoloxidase (PPO) activating system (i.e., melanization). The absence of a transcriptomic PPO response was surprising, because enzymatic assays of Vibrio-infected Exaiptasia (10 6 cfu/ml) showed a tenfold increase in PPO enzymatic activity relative to controls (Zaragoza et al., 2014), and PPO has been shown to be up-regulated in some hard and soft coral immune responses (Palmer & Traylor-Knowles, 2012). Differences in immune responses would be expected between Exaiptasia and other symbiotic anthozoans, but the conserved immune features identified in the Exaiptasia genome (Baumgarten et al., 2015) support Exaiptasia as a model for anthozoan immune responses.

| Complement and coagulation cascade
Patterns of gene expression in the complement and coagulation cascade indicate that coagulation is initiated by Vibrio and aposymbiosis, whereas the complement alternative pathway is initiated primarily by Vibrio exposure. Out of the four highly expressed DE coagulation genes that differed due to Vibrio exposure and aposymbiosis, three

Von Willebrand factor (VWF) is involved in cell adhesion and
collagen binding (Ruggeri, 2007) and has been associated with allogeneic rejection, pathogen exposure, and symbiotic state in cni-  Alpha-2-macroglobulin (A2MG) binds peptides including a wide range of proteinases (Borth, 1992). Its ability to bind the serine protease thrombin gives A2MG anticoagulant properties (Mitchell, Piovella, Ofosu, & Andrew, 1991), while its ability to inhibit proteins C and S gives it procoagulant properties (Cvirn et al., 2002). Many pathogen virulence factors act as proteases, and thus, A2MG's ability to inhibit proteases protects the host from these virulence factors (Armstrong & Quigley, 1999). A2MG has been associated with pathogen exposure in corals (Libro et al., 2013) and wound healing in anemones (Stewart et al., 2017), but had not been linked to aposymbioses.

and in WBD-infected
Acropora cervicornis (Libro & Vollmer, 2016  identified two variants of factor B in Exaiptasia, one of which was up-regulated in response to both onset of symbiosis and treatment with Serratia marcescens. Up-regulation of coagulation factor XII B chain (F13B), is associated with blood coagulation and hemostasis in vertebrates (Ivanov et al., 2017) and has also been observed in diseased Acropora cervicornis (Libro, 2014).
In the complement pathway, the classical and lectin pathways require specific recognition molecules for initiation, but in the alternative pathway, C3b is deposited on all cells (host as well as pathogenic) exposed to activated complement (Ferreira, Pangburn, & Cortés, 2010). Three remaining highly expressed DE genes within complement (membrane cofactor protein, complement decay-accelerating factor, factor H) that were up-regulated by Vibrio exposure are all involved in protecting host tissues from attack by the alternative complement pathway (Elvington, Liszewski, & Atkinson, 2016;Ferreira et al., 2010). Up-regulation of membrane cofactor protein (MCP/cd46), complement decay-accelerating factor (DAF/cd55), and component factor H in Exaiptasia would limit C3b deposition on healthy Exaiptasia cells (Elvington et al., 2016;Ferreira et al., 2010). Neither DAF, MCP, or CFAH have previously been associated with anthozoan immunity.
Out of the six DE genes in the NOD/Toll-like receptor pathway, only MyD88 has previously been observed to be DE in cnidarians due to immune exposure. Libro et al. (2013) observed up-regulation of MyD88 in WBD-infected Acropora cervicornis. In humans (Wang, Dziarski, Kirschning, Muzio, & Gupta, 2001), mouse (Deguine & Barton, 2014), and fly (Horng & Medzhitov, 2001), stimulation of Toll-like receptors (TLRs) causes MyD88 to associate with the intracellular domain of the TLR leading to downstream signaling of NF-kappa-β via IRAK and TRAF and production of pro-inflammatory cytokines (Akira, Uematsu, & Takeuchi, 2006 Franzenburg et al. (2012) in the hydrozoan Hydra vulgaris, which resulted in the down-regulation of TRAF3 but not IRAK.
The expression patterns of the remaining 3 NOD/Toll-like receptor pathway genes (Bcl-2, RIPK2, and CASR) suggest that they are acting to prevent apoptosis in Exaiptasia exposed to Vibrio. The B-cell lymphoma 2 (Bcl-2) family of apoptosis-regulating proteins includes both pro-and antiapoptotic members. Ainsworth et al. (2015) identified up-regulation of the pro-apoptotic Bcl-2 family member Bak in Acropora hyacinthus tissues exhibiting white syndrome, and Pernice et al. (2011) proposed that up-regulation of Bcl-2 is a protective response to heat-stress-induced apoptotic activity in Acropora millepora. Down-regulation of RIPK2 and CASR also suggests an antiapoptotic role in Exaiptasia exposed to Vibrio.
RIPK2 activates NF-kappa-β and induces cell death (McCarthy, Ni, & Dixit, 1998). Up-regulation of CASR leads to apoptosis in rat myocytes exposed to LPS (Wang et al., 2013). To our knowledge, we are the first to report differential expression of RIPK2 and CASR in anthozoans.
Even though Exaiptasia shows strong evidence for TLR pathway activation, no Exaiptasia TLRs met our annotation criteria (best hit, e-value < 1e −10 , coverage > 50%). Two Exaiptasia genes (KXJ18603.1, KXJ08560.1) annotated as a relaxin receptor 2 (RXFP2) and outer membrane protein OprM (OPRM) were up-regulated for Vibrio and had had blast hits for a TLR with an e-value < 1e −10 and coverage greater than 50%, but the TLR was not the best hit. While it is possible that these two genes represent TLRs, more data would be needed to confirm their putative functions. Toll-like receptors (TLRs) are transmembrane proteins consisting of an extracellular leucinerich repeat region (LRR) involved in pathogen recognition, and an intracellular Toll-interleukin receptor (TIR) which initiates downstream activation of NF-kappa-β via MyD88 (Brennan et al., 2017).
A single TLR has been identified in the model anemone Nematostella vectensis, and its activation and downstream signaling via NF-kappa-β have been demonstrated in response to Vibrio coralliilyticus (Brennan et al., 2017). We performed a Pfam domain search on the

| Chemokine and antigen processing
Chemokine and antigen processing pathways were also activated by Vibrio exposure. Chemokine pathway had three highly DE genessignal transducer and activator of transcription 1-alpha/beta (STAT1) was up-regulated while C-X-C chemokine receptor type 4 (CXCR4), and guanine nucleotide-binding protein subunit beta-4 (GBB4) were Sinkovics (2015) proposed a Cnidarian origin of STAT based on genomic studies on Nematostella vectensis, but its role in immunity had not been confirmed by expression analysis.
Two antigen processing genes were highly DE; proteasome activator subunit 2 (PSME2/PA28 beta) was up-regulated for Vibrio, and regulator factor X-associated ankyrin-containing protein (RFXK) was down-regulated for aposymbiosis. Proteasomes are involved in antigen processing (Michalek, Grant, Gramm, Goldberg, & Rock, 1993) and degradation of other intracellular proteins (Tanaka, 2009), including cytotoxic damaged proteins resulting from the oxidative stress of an immune response (Kammerl & Meiners, 2016). Traylor-Knowles, Rose, Sheets, and Palumbi (2017) observed up-regulation of proteasome components in Acropora hyacinthus exposed to heat stress. To our knowledge, we are the first to report up-regulation of proteasomal proteins in response to bacterial immune challenge in cnidarians.

| Transport and catabolism
Within transport and catabolism (peroxisome, endocytosis, lysosome), there were more genes highly down-regulated than up-regu-

| Endocytosis
Endocytosis pathway had six highly expressed DE genes; three genes were up-regulated for Vibrio (HRS, HSE1, CLH), two genes were down-regulated for Vibrio (VPS4, PLD2), and one gene was down-regulated for aposymbiosis (JUNO). Following recognition by Toll-like receptors, pathogens are engulfed by clathrin-mediated endocytosis (Husebye et al., 2006). In Drosophila, endocytosis is required for activation of the Toll pathway, and endosomal proteins Mop and Hrs colocalize with the Toll receptor in endosomes (Huang, Chen, Kunes, Chang, & Maniatis, 2010). Although we observed more down-regulation than up-regulation of genes within the endocytosis pathway, those which were up-regulated in response to Vibrio (hepatocyte growth factor-regulated tyrosine kinase substrate HRS, clathrin heavy-chain CLH, signal transducing adaptor molecule HSE1) are consistent with recognition by TLR pathway followed by clathrin-mediated endocytosis.
Two DE apoptosis genes for aposymbiosis (P53, BIRC5) suggest apoptosis is a mechanism for menthol bleaching. Tumor protein P53 regulates a number of cell-cycle functions including apoptosis, regulation of autophagy, cell-cycle arrest, and senescence (Zilfou & Lowe, 2009). Lesser and Farrell (2004) observed up-regulation of P53 in corals exposed to increased solar radiation, and Weis (2008) proposed activation of P53 by the reactive nitrogen species nitric oxide (NO) in thermally stressed corals as a mechanism of bleaching. The up-regulation of P53 in aposymbiotic anemones may indicate that the mechanism of menthol-induced bleaching is similar to the mechanisms of bleaching in thermal and solar radiation-stressed corals. The second DE apoptosis gene for menthol BIRC5, also known as survivin, is an antiapoptotic caspase inhibitor (Li et al., 1998). The down-regulation of BIRC5 for aposymbiotic anemones lends further support to apoptosis as a mechanism of menthol bleaching.

| CON CLUS ION
Exposure to live Vibrio coralliilyticus for both symbiotic and aposymbiotic anemones had strong and significant impacts on gene expression, but their effects were independent or additive, not interactive. to those involved in thermal stress-induced bleaching. While we did not see the interaction that we expected between symbiotic state and response to a pathogen, this study provides additional data points to better understand both bleaching and pathogen response in anthozoans.

ACK N OWLED G M ENTS
The authors would like to thank Stefan T. Kaluziak for assistance with Bioinformatics server support and assistance with Bioinformatics analysis. Research was funded by NSF award OCE-1458158 to SVV. Anemones were maintained and the experiment conducted using the seawater system and other equipment provided by NSF facilities grant OCE-0963010. This is contribution #393 from the Marine Science Center at Northeastern University.

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
CLR and SVV conceived and designed the experiment. CLR generated and analyzed the data and wrote the manuscript. CLR and SVV edited the manuscript.

DATA AVA I L A B I L I T Y
The Illumina RNA-Seq read data are available on NCBI SRA https ://www.ncbi.nlm.nih.gov/sra under BioProject accession number PRJNA547971. Normalized read count data, DESeq2, and an-