The Association of Waminoa with Reef Corals in Singapore and Its Impact on Putative Immune- and Stress-Response Genes

: Waminoa spp. are acoel ﬂatworms mainly found as ectosymbionts on scleractinian corals. Although Waminoa could potentially represent a threat to their hosts, not enough information is available yet regarding their ecology and effect on the coral. Here, the Waminoa sp.–coral association was analyzed in Singapore reefs to determine the prevalence, host range, and preference, as well as the ﬂatworm abundance on the coral surface. Moreover, the impact of Waminoa sp. on the expression of putative immune- and stress-response genes ( C-type lectin , C3 , Hsp70 and Actin ) was examined in the coral Lobophyllia radians . The association prevalence was high (10.4%), especially in sites with lower sedimentation and turbidity. Waminoa sp. showed a wide host range, being found on 17 coral genera, many of which are new association records. However, only few coral genera, mostly characterized by massive or laminar morphologies appeared to be preferred hosts. Waminoa sp. individuals displayed variable patterns of coral surface coverage and an unequal distribution among different host taxa, possibly related to the different coral growth forms. A down-regulation of the expression of all the analyzed genes was recorded in L. radians portions colonized by Waminoa individuals compared to those without. This indicated that Waminoa sp. could affect components of the immune system and the cellular homeostasis of the coral, also inhibiting its growth. Therefore, Waminoa sp. could represent a potential further threat for coral communities already subjected to multiple stressors.

In particular, we determined the prevalence of the association, the host range, the host preference of Waminoa, and the flatworm abundance on the coral surface. In addition, we tested a possible effect of Waminoa on components of the coral immune system and cellular stress response. For this purpose, the coral Lobophyllia radians was selected being a species highly colonized by the Waminoa in Singapore, and the gene expression of C-type lectin, C3, Hsp70 and Actin were examined and used as biomarkers to provide new insights on the nature of the association.  85 85 E), were selected as study sites ( Figure 1). Both reefs are characterized by a shore-adjacent reef flat leading seaward to the reef crest and down the reef slope to~8-10 m maximum depth because of high levels of suspended sediments that cause extreme light attenuation [55,59].

Materials and Methods
formed to assess significant differences in the prevalence of the Waminoa sp.-coral associ-ation among the different host genera. Coral genera showing a prevalence < 5% were not included in the analysis.
The host preference of Waminoa sp. in terms of coral genus was tested through the Van der Ploeg and Scavia Selectivity coefficient (Ei), following [64]. This coefficient is defined for a group i as: where Wi represents the value of Chesson's α and n represents the number of habitat types [63], here represented by the coral genera found in the study area. Chesson's α value (Wi) is defined as: where ri represents the frequency of a habitat category (coral genus) in the environment, and Pi represents the frequency of the same habitat category in which the organism of interest (Waminoa sp.) is found [65]. Values of selectivity coefficient range between -1 and 1, with -1 meaning complete avoidance of a host coral genus, and 1 meaning exclusive preference for a specific coral genus [66]. A one-way ANOVA followed by a Tukey's HSD post hoc test was performed to assess differences in the coverage percentages of the flatworms on the surface of the host corals, between the four coverage categories.  In both sites, extensive surveys were conducted by SCUBA diving between August and October 2019 to detect the occurrence of Waminoa individuals on different coral genera ( Figure 2). Images of Waminoa specimens were taken by a digital camera and analyzed. We distinguished a single morphotype of Waminoa in the surveyed area (as described in [8,15], Figure 2) and we treated it as Waminoa sp.

Coral Collection
To assess the effect of Waminoa sp. on coral's gene expression, five colonies of Lobophyllia radians showing patches of Waminoa sp. individuals in a fixed position on their surface were selected, monitored for several days and sampled on Kusu Island. All five colonies were moderately (11-25%) covered by the worms. For each colony, fragments of approximately 2 cm 2 were collected using a hollow-point stainless steel spike [67]. Two coral fragments were sampled from each colony: one fragment located just underneath a patch of Waminoa sp. individuals (marked as "W"), and the other from a colony portion without worms, at least 5 cm away from the Waminoa individuals (marked as "W/0"). In addition, three healthy colonies of L. radians not colonized by the flatworms were randomly sampled as controls to test primer efficiencies.
All coral samples were taken at the same hour (around 9:00 a.m.), at the same shallow depth, and during high tide, to minimize any possible effects of abiotic variables, such as water temperature and light intensity, on the gene expression. During the sampling period, the sea surface temperature was continuously logged and its mean was 30.36 ± 0.39 °C, with very slight oscillation between 29.66 and 30.42 °C. In "W" fragments, Waminoa Six 50 × 2 m belt transects (100 m 2 each), spaced 10 to 20 m apart and placed parallel to the coast at a constant depth between 5 and 8 m (depending on the tides) along the reef slope, were randomly laid at each site. Within belt transects all the coral colonies were identified to genus level, according to Huang et al. [60] and Wong et al. [61], and the number of colonies of each genus found colonized by Waminoa sp. was recorded. Moreover, in coral colonies hosting Waminoa individuals, their abundance was estimated by determining the percentage of coral surface covered by flatworms and was indicated with four coverage categories, as suggested in [62]: low (coral surface covered 1-10% by worms), moderate (11-25%), severe (26-50%), and extreme (>50%).
In each belt transect, the point intercept transect (PIT) method was also performed (by recording data every 10 cm [63]) to determine the composition and structure of the benthic community, as well as the cover percentage of each coral genus. Data were collected using the following benthic categories: algae, dead coral, coral, coral rubble, rock, sand, and other (sponges, soft corals, tunicates, zoantharians, and unknown). Furthermore, for the macro-category "coral", the genus was also recorded. For each transect, the prevalence of the association was calculated as the ratio between the number of corals colonized by Waminoa sp. and the total number of colonies. By averaging the corresponding prevalence values measured on the six random belt transects for each site, both an overall and a series of taxon (coral genus)-specific prevalence values were determined. Data normality was verified using the Shapiro-Wilk test. A one-way ANOVA was used to test significant differences in the overall association prevalence between the two sites analyzed. The same analysis, followed by Tukey's honestly significant difference (HSD) post hoc tests for multiple pairwise comparisons of means, was performed to assess significant differences in the prevalence of the Waminoa sp.-coral association among the different host genera. Coral genera showing a prevalence < 5% were not included in the analysis.
The host preference of Waminoa sp. in terms of coral genus was tested through the Van der Ploeg and Scavia Selectivity coefficient (Ei), following [64]. This coefficient is defined for a group i as: where Wi represents the value of Chesson's α and n represents the number of habitat types [63], here represented by the coral genera found in the study area. Chesson's α value (Wi) is defined as: where ri represents the frequency of a habitat category (coral genus) in the environment, and Pi represents the frequency of the same habitat category in which the organism of interest (Waminoa sp.) is found [65]. Values of selectivity coefficient range between −1 and 1, with −1 meaning complete avoidance of a host coral genus, and 1 meaning exclusive preference for a specific coral genus [66]. A one-way ANOVA followed by a Tukey's HSD post hoc test was performed to assess differences in the coverage percentages of the flatworms on the surface of the host corals, between the four coverage categories.

Coral Collection
To assess the effect of Waminoa sp. on coral's gene expression, five colonies of Lobophyllia radians showing patches of Waminoa sp. individuals in a fixed position on their surface were selected, monitored for several days and sampled on Kusu Island. All five colonies were moderately (11-25%) covered by the worms. For each colony, fragments of approximately 2 cm 2 were collected using a hollow-point stainless steel spike [67]. Two coral fragments were sampled from each colony: one fragment located just underneath a patch of Waminoa sp. individuals (marked as "W"), and the other from a colony portion without worms, at least 5 cm away from the Waminoa individuals (marked as "W/0"). In addition, three healthy colonies of L. radians not colonized by the flatworms were randomly sampled as controls to test primer efficiencies.
All coral samples were taken at the same hour (around 9:00 a.m.), at the same shallow depth, and during high tide, to minimize any possible effects of abiotic variables, such as water temperature and light intensity, on the gene expression. During the sampling period, the sea surface temperature was continuously logged and its mean was 30.36 ± 0.39 • C, with very slight oscillation between 29.66 and 30.42 • C. In "W" fragments, Waminoa individuals were removed from their hosts by using the pipettes as previously described [8] and the morphological condition of the coral tissues, as well as the presence of any physical damages or lesions, were evaluated. All the coral portions were immediately placed in pre-labeled tubes and, at the end of the underwater sampling, the seawater was decanted and replaced with RNAlater (Thermo Fisher Scientific, Singapore) with a tissue:RNAlater ratio of 1:10. The maximum time between collection and placement in RNAlater was about 30 min. The sample tubes were inverted to mix for 30 s and kept at 4 • C overnight to allow complete penetration of RNAlater into the coral tissues. The tubes were subsequently stored at −80 • C until RNA extraction.

RNA Extraction and REVERSE Transcription (RT)
Total RNA was extracted from all the coral samples without homogenization to reduce RNA fragmentation using TRIzol Reagent (Life Technologies, Sigma-Aldrich, Singapore) following the manufacturer's protocol. RNA quality was checked by examining with gel electrophoresis for presence of clear bands of ribosomal RNAs and RNA concentration was estimated using Qubit (RNA Broad Range Assay Kit, Thermo Fisher Scientific). Complementary DNA (cDNA) was immediately prepared from 1 µg of total RNA for each sample, using a one-tube format of iScript RT Supermix (Bio-Rad, Hercules, CA, USA) for a reverse transcription quantitative polymerase chain reaction (RT-qPCR). Reaction setup was composed of iScript RT Supermix (4 µL), RNA template (varied depending on RNA sample concentration; 14.6-200 ng/µL), and nuclease-free water (variable), with a final volume of 20 µL as previously described [68]. Incubation of the reaction mix was performed in a Labcycler (Sensoquest, Göttingen, Germany) following the manufacturer's protocol: priming for 5 min at 25 • C, RT for 20 min at 45 • C, and RT inactivation for 1 min at 95 • C.

Primer Design and Validation
A local search in the L. radians transcriptome previously sequenced (raw sequencing data available at NCBI Sequence Read Archive under accession number PRJNA512601) and assembled by [69] using BLASTn against the GenBank database was performed. Coral transcripts matching genes from the genomes of Acropora digitifera, Orbicella faveolata, or Stylophora pistillata were identified and orthologous genes on the L. radians transcriptome selected. The accuracy of the sequence of each gene of interest (GOI) was checked using the NCBI Nucleotide BLAST tool (https://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 15 November 2019). A megablast search was performed to ensure that the sequences producing significant alignments with the query sequence were corresponding to the GOIs. A BLASTn search was performed in an open-access coral genomic database at http://reefgenomics.org/blast/ [70,71] with Goniastrea aspera genome as subject and the sequence of each GOI as query. A suitable region within exons was selected and the query corresponding to the longer hit with the highest identities value was selected to design primers. Primers for each gene were designed using the online tool by NCBI that incorporated Primer3 and BLAST (https://www.ncbi.nlm.nih.gov/tools/primer-blast). The set of primers given as output by the NCBI tool were analyzed and used in http:// reefgenomics.org/blast/to perform a BLASTn search against Symbiodiniaceae nucleotide databases (Supplementary materials Table S1) to verify primer specificity for coral DNA. The selected set of primer was then used for a megablast search using the NCBI Nucleotide BLAST tool to verify the absence of significant alignment with marine species found in the same environment as L. radians. The designed primers are shown in Table 1. The specificity of the selected primers for L. radians was tested by performing PCR using the GoTaq Green Master Mix in a LifeECO Thermal Cycler (Bioer Technology, Hangzhou, China). PCR reaction steps were (1) denaturation: 95 • C for 45 s, (2) annealing: 50-55 • C for 45 s, and (3) extension: 72 • C for 3 min, repeated for 35 cycles. Table 1. List of genes of interest (GOIs) with primer designs. The melting temperature (Tm), the % of guanine and cytosine (GC), and the length of PCR product (bp) are also reported for each gene. Additional information of these sequences can be found in Supplementary Materials.

Gene
Sequence To test primer efficiencies, a series of twofold dilutions of L. radians cDNA starting from 5 ng/µL were performed for the samples collected from the control colonies [38]. Each dilution was used in triplicate for each primer to assess the primer efficiency through an RT-qPCR using the CFX96 Real-Time PCR System (Bio-Rad). Calculations of efficiencies (E, the amplification factor per PCR cycle) needed to correct for amplification efficiencies per primer were undertaken using an MCMC.qpcr package in R developed by Matz et al. [72]. The function, PrimEff( ), calculates E and plots the regression slopes and E based on dilution series. GOIs with E values outside the 1.85-2.16 range had primers redesigned and re-validated. GOIs that failed amplification were excluded from downstream analyses [68].

Gene Expression Quantification
RT-qPCRs were performed in a CFX96TM Real-Time PCR System (Bio-Rad) using SsoAdvanced inhibitor-tolerant SYBR Green Supermix following the manufacturer's protocol (polymerase activation and DNA denaturation: 3 min at 98 • C, denaturation: 15 s at 95 • C, and annealing/extension: 30 s at 60 • C) repeated for 40 cycles. The reaction mix was prepared as in Table S2. Each sample was tested in duplicates for each of the four genes. To control for variations in expressions of genes caused by differences in RNA concentration of each sample, the amount of cDNA template was standardized to~10 ng of cDNA for every reaction mix.

Data Analysis
Data obtained from RT-qPCRs expressed as "cycle of quantification values" (i.e., Cq values) were collated and sorted for subsequent analysis. RT-qPCR data were analyzed using generalized linear mixed models based on lognormal-Poisson error distribution, fitted using the MCMC.qpcr package Version 1.2.3 in R Studio Version 1.1.463 as previously reported [72]. Molecule count data with corrections for primer efficiencies were derived with amplification efficiencies (E) per gene and Cq for a single target molecule using the formula: Count = E(Cq1 − Cq). The Cq-to-counts conversion is the key transformation in this method, in which higher variation at the low gene expression values is properly accounted for by the relative quantification model. The transformation makes it possible to fit the resulting data to generalized linear mixed models to account for Poisson-distributed fluctuations when the number of the molecule count is low. Similar Bayesian approaches for analyzing qPCR data have been used in several other reports [68]. Results of the mcmc.qpcr() function were then plotted with HPDsummary() to visualize fold changes in gene expression in response to the presence of Waminoa worms. HPDsummary() also calculates all the pairwise differences between treatments and their statistical significance for each gene. Each gene profile was examined to determine the differential expression level between coral samples underneath the surface of the coral colonized by Waminoa sp. and samples of the same colonies at least 5 cm apart from the flatworms. A multivariate analysis was performed using the statistical package PRIMER-E v.7 with the PERMANOVA+ add on [73,74] to investigate together the modulation of all biomarkers in response to Waminoa sp. colonization. In particular, data related to the levels of all the biomarkers were square root transformed to calculate a matrix based on the Bray-Curtis similarity. To test for differences in biomarker levels between corals with and without worms, a non-parametric permutational multivariate analysis of variance (PERMANOVA) was performed using 999 permutations with partial sum of squares and unrestricted permutation of raw data. Values were considered statistically significant at p < 0.05.
Overall, 1044 coral colonies were observed, and the overall prevalence of the coral-Waminoa sp. association was 10.4 ± 2.3%, with the site on Kusu Island showing a significantly higher prevalence compared to Pulau Hantu (ANOVA, F(1,34) = 3.1, p = 0.012; Figure 3). In total, 17 out of the 39 scleractinian genera recorded in the study area were found in association with Waminoa sp. and significant differences in the association prevalence were detected among the host coral genera (ANOVA, F(12,143) = 2.39, p = 0.008; Figure 4A). In particular, Lobophyllia clearly displayed the highest prevalence, followed by Goniastrea and Favites, and later by Mycedium, Platygyra, Oxypora, and Pachyseris, while for the five other host genera the prevalence recorded was lower than 10% but higher than 5% (Pectinia, Echinopora, Fungia, Ctenactis, and Podabacia, Figure 4A). However, significant differences in the association prevalence were observed only between Lobophyllia and Echinopora, and Fungia, Ctenactis and Podabacia ( Figure 4A). In addition, the five other scleractinian genera, namely Merulina, Porites, Dipsastraea, Hydnophora, and Montipora, were found associated with Waminoa sp. but with prevalence < 5% (4.2, 3.7, 3.5, 2.5, and 1.9%, respectively). The prevalence patterns recorded in both sites were not uniform and did not fully reflect those recorded in the whole study, with Goniastrea, Pachyseris and Lobophyllia showing the higher prevalence in Pulau Hantu, and Lobophyllia, Favites, Mycedium, and Oxypora in Kusu Island ( Figure S2). The selectivity coefficient Ei allowed the comparison of the relative abundance of coral genera colonized by Waminoa sp. with the relative abundance of the same coral genera recorded in the whole study area ( Figure 4B). The analysis was performed only for coral genera showing an association prevalence > 5%. It revealed that Mycedium was the preferred host for Waminoa sp., followed by Lobophyllia, Oxypora, and, surprisingly, Ctenactis, which was among the coral genera that showed the lowest prevalence of the association ( Figure 4B). Moreover, Waminoa sp. showed a marked avoidance for high/medium-prevalence genera, such as Platygyra, Pachyseris and Pectinia ( Figure 4B).    Most of the coral colonies were moderately (from 11 to 25% of the coral surface) or severely (26-50%) covered by Waminoa sp., while a few colonies showed an extreme colonization of worms (>50%) on their surface ( Figure 5A). However, no significant differences in the abundance percentages were recorded among the Waminoa sp. coverage categories (ANOVA, F(3, 44) = 1.83, p = 0.158, Figure 5A). Therefore, the distribution of the flatworms on the coral colonies was heterogeneous, although Waminoa sp. mostly occupied less than 50% of the coral's surface ( Figure 5A). This pattern was also found in the coral genera showing the highest prevalence of the association, such as Lobophyllia, Goniastrea, and Favites, as well as in the preferred genus Mycedium ( Figure 5B). In particular, in Lobophyllia corals the flatworms mostly colonized from 26 to 50% of the colony surface, while in both Goniastrea and Favites about 40% of the colonies had less than 10% of their area occupied by Waminoa sp. individuals. Almost all the colonies of Mycedium hosting the worms had less than 10% of their surface covered, while the Waminoa sp. infestation on the coral surface was mostly severe in Pachyseris and extreme in Fungia ( Figure 5B). -corals associations in Pulau Hantu and Kusu Island. Numbers above each bar indicate the total number of coral colonies (both with and without Waminoa) analyzed per site. Data are expressed as the mean ± SEM. One-way ANOVA was performed between sites. Letters on the bars denote significant difference among sites. Different letters indicate significant difference (p < 0.05), while same letter indicates no significant difference (p ≥ 0.05).  Data are expressed as mean ± SEM. Letters denote Tukey's significant differences among the different groups (p < 0.05); the same letter indicates no significant difference (p ≥ 0.05). (B) Host preferences of Waminoa sp. according to the Van der Ploeg and Scavia selectivity coefficient Ei (−1 = complete avoidance; 0 = random choice; +1 = exclusive preference) for each coral genus. In both graphs, only the coral genera with an association prevalence > 5% are reported. Most of the coral colonies were moderately (from 11 to 25% of the coral surface) or severely (26-50%) covered by Waminoa sp., while a few colonies showed an extreme colonization of worms (>50%) on their surface ( Figure 5A). However, no significant differences in the abundance percentages were recorded among the Waminoa sp. coverage categories (ANOVA, F(3, 44) = 1.83, p = 0.158, Figure 5A). Therefore, the distribution of the flatworms on the coral colonies was heterogeneous, although Waminoa sp. mostly occupied less than 50% of the coral's surface ( Figure 5A). This pattern was also found in the coral genera showing the highest prevalence of the association, such as Lobophyllia, Goniastrea, and Favites, as well as in the preferred genus Mycedium ( Figure 5B). In particular, in Lobophyllia corals the flatworms mostly colonized from 26 to 50% of the colony surface, while in both Goniastrea and Favites about 40% of the colonies had less than 10% of their area occupied by Waminoa sp. individuals. Almost all the colonies of Mycedium hosting the worms had less than 10% of their surface covered, while the Waminoa sp. infestation on the coral surface was mostly severe in Pachyseris and extreme in Fungia ( Figure 5B).

Molecular Analysis
All the sampled coral portions occupied by worms showed no visible surface damage or lesions. All the candidate genes showed reliable amplification, since the efficiency of amplification was within the range of acceptable values of 1.49-2.2 (Table S4, [38]).
Significant differences in biomarker levels among portions of coral tissue colonized

Molecular Analysis
All the sampled coral portions occupied by worms showed no visible surface damage or lesions. All the candidate genes showed reliable amplification, since the efficiency of amplification was within the range of acceptable values of 1.49-2.2 (Table S4, [38]).
Significant differences in biomarker levels among portions of coral tissue colonized or not by Waminoa sp. were recorded (PERMANOVA: df = 1, F = 3.372, p = 0.007). All the genes showed a lower expression level in samples collected underneath Waminoa sp. compared to samples of the same coral colony taken at least 5 cm apart from the flatworms ( Figure 6A). Therefore, the presence of Waminoa sp. on coral caused a down-regulation of the expression of all the investigated genes in the portion of coral tissue directly in contact with the flatworms. The effect of the presence of Waminoa sp. on the gene expression was significant for C3 (pMCMC = 0.03), Hsp70 (pMCMC = 0.01), and Actin (pMCMC = 0.005), but not for C-type lectin (pMCMC = 0.13), ( Figure 6A). The highest fold change in expression levels was observed for Actin, followed by Hsp70, while the lowest change was observed for C-type lectin, which showed a non-significant down-regulation ( Figure 6B). Fold changes were calculated with respect to levels detected in "W" fragments and were log2-transformed. In both graphs, data are expressed as means (n = 5).

Ecology of the Waminoa-Coral Association in Singapore Reef
Our data contributed to extend the geographic distribution of the Waminoa-coral association. In fact, to date it has been recorded only in the reefs of the Red Sea [9,11,12,17], Indonesia [13,20,75], Micronesia [76], Australia [77,78], Japan [8,14], and Taiwan [79]. In addition, our results indicate that the association appeared to be abundant in the study area, with a prevalence greater than 10%. In areas close to Singapore, such as Taiwan, less than 1% of the corals analyzed were found colonized by Waminoa sp. [79], while in Wakatobi (Sulawesi, Indonesia) a total of 4.8% of all observed hard corals were associated with the acoel worm in 2006 and 2.6% of hard and soft corals in 2007 [13]. However, a comparison between the prevalence obtained in these studies may not be completely reliable, since the investigated geographic areas were characterized by diverse habitat structure and ecological traits, and the survey methods and approaches used were not entirely the same.
Although the presence of the flatworms was reported in both the investigated sites, a significantly higher prevalence of the association was observed on Kusu Island compared to Pulau Hantu. This difference could be explained by the greater diversity of coral genera, rugosity, and reef complexity on Kusu Island [59], which may have contributed to available niches for Waminoa sp., as well as by the different environmental and physical characteristics of the sites. Pulau Hantu is sheltered by adjacent and heavily developed islands in an area of intense industrialization and ship traffic, while Kusu Island experiences comparatively lower anthropogenic impacts and higher exposure to wave action [80,81]. This generates a significantly higher average turbidity, sedimentation, and light Fold changes were calculated with respect to levels detected in "W" fragments and were log2-transformed. In both graphs, data are expressed as means (n = 5).

Ecology of the Waminoa-Coral Association in Singapore Reef
Our data contributed to extend the geographic distribution of the Waminoa-coral association. In fact, to date it has been recorded only in the reefs of the Red Sea [9,11,12,17], Indonesia [13,20,75], Micronesia [76], Australia [77,78], Japan [8,14], and Taiwan [79]. In addition, our results indicate that the association appeared to be abundant in the study area, with a prevalence greater than 10%. In areas close to Singapore, such as Taiwan, less than 1% of the corals analyzed were found colonized by Waminoa sp. [79], while in Wakatobi (Sulawesi, Indonesia) a total of 4.8% of all observed hard corals were associated with the acoel worm in 2006 and 2.6% of hard and soft corals in 2007 [13]. However, a comparison between the prevalence obtained in these studies may not be completely reliable, since the investigated geographic areas were characterized by diverse habitat structure and ecological traits, and the survey methods and approaches used were not entirely the same.
Although the presence of the flatworms was reported in both the investigated sites, a significantly higher prevalence of the association was observed on Kusu Island compared to Pulau Hantu. This difference could be explained by the greater diversity of coral genera, rugosity, and reef complexity on Kusu Island [59], which may have contributed to available niches for Waminoa sp., as well as by the different environmental and physical characteristics of the sites. Pulau Hantu is sheltered by adjacent and heavily developed islands in an area of intense industrialization and ship traffic, while Kusu Island experiences comparatively lower anthropogenic impacts and higher exposure to wave action [80,81]. This generates a significantly higher average turbidity, sedimentation, and light attenuation rate in Palau Hantu, resulting in an overall lower light intensity and shallower euphotic depth than Kusu Island [81]. These conditions potentially affect the presence of photosynthetic dinoflagellatehosting organisms such as Waminoa flatworms, which may also have more difficulty in colonizing sediment-covered surfaces and may themselves be vulnerable to environmental disturbances.
The Waminoa sp. host range was updated with additional scleractinian genera, many of which are new records. Indeed, in Singapore, Waminoa sp. was in association with 17 coral genera belonging to six families, namely Lobophylliidae, Merulinidae, Agariciidae, Poritidae, Fungiidae, and Acroporidae. Among them, the family Merulinidae was largely the most represented, as also recorded in Taiwan, despite only six coral genera in total being found infested by Waminoa sp. [79]. In Sulawesi (Indonesia), the association with Waminoa was confirmed for 21 coral taxa (Wakatobi [13]), but in Bangka Island it was recorded for only 4 coral genera (Gardineroseris, Platygyra, Porites, Turbinaria [75]). In the Red Sea, 13 coral genera were found infected [11]. In Japan, Waminoa individuals were found on 4 coral genera only, namely Cycloseris, Echinomorpha, Echinophyllia, and Pachyseris [14], and 13 scleractinian hosts all belonging to Lobophylliidae [8]. Therefore, Waminoa sp. in Singapore coral reefs showed a wide host range. However, only few coral genera such as Lobophyllia, Mycedium, Oxypora, and to a lesser extent Goniastrea and Ctenactis, appeared to be preferred hosts. These coral genera are characterized mainly by massive/submassive or laminar/encrusting colony morphologies, while corals without Waminoa sp. are typically branching or columnar. On the contrary, in previous studies Waminoa sp. was predominantly observed on branching Acropora and Stylophora corals, as well as in the columnar Tubastrea [11,13]. The question of why Waminoa sp. colonizes and/or prefers only specific coral taxa remains largely unanswered. We hypothesize that the coral skeleton morphology could represent a factor driving the choice of the flatworms, given that some structures could favor protection from predators, allowing the worm to hide. Since coral mucus represents a possible food source for Waminoa spp. [17], the different mucus production among different coral taxa could also represent an additional host selection factor. This is even more relevant in Singapore's turbid reefs, as some corals can increase or decrease mucus production as a defense mechanism in response to persistent sediment stress [82][83][84][85]. In this regard, it would be interesting to analyze the mucus production and composition of the different coral taxa to explore possible correlations with the Waminoa presence. Finally, since Waminoa also feed on zooplankton caught by corals [18], the ability of a coral species to capture zooplankton, which is determined by its morphology, coral polyp size, and the type of tentacles and nematocysts (reviewed in [86]), may play an important role in the Waminoa host selection. However, in addition to these hypotheses, we cannot exclude that the Waminoa individuals analyzed here, albeit being of a single morphotype, did not belong to a single species but represented a complex of cryptic species, each of which specialized in a different host.
Corals of Singapore showed variable patterns of flatworm density, ranging from colonies that were densely and extremely infested to others that were only moderately and sparsely populated, as previously observed [11]. However, as also occurred in Okinawa [14], Waminoa sp. individuals were not equally distributed among different host taxa. In particular, in Singapore we detected that different Waminoa infestation rates could be related to the coral growth form. Indeed, in corals with massive growth forms (such as Lobophyllia, Goniastrea, Favites, and Platygyra), Waminoa sp. showed a heterogeneous pattern of distribution (but in general < 50% of the coral surface was occupied). Corals with a foliose and/or encrusting growth form (Mycedium, Podabacia, Oxypora, and Echinopora) were sparsely or moderately covered by flatworms, while Fungidae corals (Ctenactis and Fungia) appeared extremely colonized by Waminoa sp., as also previously observed [20].

Effect of Waminoa sp. on Coral Putative Immune-and Stress-Response Genes
Waminoa spp. can cause physiological damage to corals by inhibiting photosynthesis, reducing the coral tolerance to environmental stress, and impairing coral respiration and feeding [11,17,20]. Our analysis on coral gene expression produced a detailed description of the early response to stress at cell/tissue level, since changes at the molecular level occur before morphological and physiological impairment appear evident [87,88].
Our results show that Waminoa sp. affected the analyzed host molecular pathways associated with the coral's stress tolerance and immunity response, causing a uniform down-regulation of the expression of all the investigated genes. Moreover, this modulation was only observed in the physically undamaged coral tissue portions colonized by Waminoa individuals and not in those free from flatworms. This might suggest that, as previously observed in corals infected by bacteria or protozoans [24,46,67], the stress response was confined in a restricted area just below the flatworm, even though polyps are linked together by common tissue in the coral colony.
The complement pathway of the immune system is triggered by lectins binding a pathogen-associated molecule and results in the activation of the complement component factor C3 and C4 [89]. Indeed, in corals, C-type lectin and C3 protein were usually up-regulated and activated in response to epizootic diseases [47,48,90,91]. However, their down-regulation may reflect suppression of host immunity, as previously observed for the association between corals and the microalga Chromera [92]. Mohamed et al. [92] also suggested that the down-regulation of some PRRs could reflect the host attempting to limit interactions with non-beneficial organisms, since both complement C3 and the C-type lectin have been implicated in symbiont recognition and in host-symbiont communication [34]. In addition, the down-regulation of C-type lectin and C3 has been observed in corals subjected to temperature/light stress, suggesting that these stresses might compromise the coral's immune defenses and therefore increase the coral's susceptibility to diseases [38,[93][94][95]. Likewise, the decreased expression of the C-type lectin and C3 here suggests that the presence of Waminoa sp. individuals on coral tissue might interfere with the ability of the whole host coral to respond to the attack of various pathogens and at the same time could make it more vulnerable to environmental stressors. However, while the C3 appeared to be significantly down-regulated by the flatworm presence, the decrease in expression of the lectin was not significant. Considering that C-type lectins have been shown to respond immediately following an immune challenge [34,90] but may not show any significant response at later times [47], we hypothesize that the observed modulation could be influenced by the sampling times.
The cytoplasmic chaperonin Hsp70 is involved in assembly of newly synthesized proteins and in the refolding of misfolded or aggregated proteins, contributing to the protein transfer to different cellular compartments or to the proteolytic machinery and acting as cellular defensive mechanism [96]. Up-regulation of the Hsp70 has been proposed as an activator of other components of the coral effector immune systems, such as the prophenoloxidase cascade, in corals infected by pathogens [47]. On the contrary, the down-regulation of Hsp70 in corals reflected the impairment of the cellular defense mechanisms that is due to severe and intolerable stress [97][98][99], and may indicate a reduced activity of the immune system because of diseases [24]. In addition, since Hsps are ATPdependent chaperones, the decrease of Hsp70 expression may be related to the high-energy expenditure necessary to reduce the deleterious effects of Waminoa sp. and restore cellular damage. However, it is important to underline that, since the roles of Hsp70 in organismal function are broad, changes in expression of this gene could also reflect changes in other physiological processes.
Actin was the most responsive gene, showing the greatest down-regulation. In addition to being fundamental for cell motility, contractibility, mitosis, and intracellular transport, Actin is also an important part of the nuclear complex, being required for the transcription of RNA polymerases and in the export of RNAs and proteins from the nucleus [100]. Down-regulation of Actin has previously been observed in corals subjected to thermal stress and acidification [38,68,101,102]. Since Actin is a major cytoskeletal component involved in growth, down-regulation of this gene could be indicative of growth inhibition caused by the presence of Waminoa sp. Moreover, the reduced expression of Actin may reflect a change in the regulation of gene transcription of proteins involved in cytoskeletal interactions and may imply changes in intracellular transport and cell shape/integrity, as previously suggested [102]. The overall down-regulation of all the analyzed genes may reflect a negative effect of the acoelomate ectosymbiont Waminoa sp. on the host coral L. radians. However, alternative scenarios should also be considered. For example, it could be possible that Waminoa did not cause detectable cellular stress to hosts, preferentially colonizing polyps with reduced defense responses, or interfering with polyp feeding and causing the observed gene down-regulation, as reduced resources would lead to reduced investment in defense.
In conclusion, our study demonstrated that Waminoa sp. showed a high prevalence and wide host range in Singapore coral reefs and its distribution patterns were specific to certain scleractinian host genera. Moreover, Waminoa sp. could impair both the cellular homeostasis and components of the immune system of the host, thus representing a potential further threat for coral communities living in an area already subjected to multiple stresses, such as sedimentation and light limitation. However, further studies analyzing more genes and biomarkers in different hosts are necessary to have a more complete picture of the association.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/d14040300/s1, Figure S1: Coverage percentage of the different benthic categories in the two sites analyzed and in the whole study area. Data are expressed as the mean ± SEM; Figure S2: Prevalence (%) of Waminoa sp.-corals association by coral genus in the two sites. Data are expressed as the mean ± SEM; Table S1: Symbiodiniaceae nucleotide databases; Table S2: RT-qPCR Mastermix used for determining gene expression and efficiency; Table S3: Coverage percentage of each coral genus in the two sites and in the whole study area; Table S4 Institutional Review Board Statement: All applicable permits and institutional guidelines required to perform the work were followed. Collections were made under permit NP/RP16-156.

Informed Consent Statement: Not applicable.
Data Availability Statement: The datasets generated during and/or analyzed during the current study are available from the corresponding authors on reasonable request.