Azooxanthellate Palythoa (Cnidaria: Anthozoa) Genomes Reveal Toxin-related Gene Clusters and Loss of Neuronal Genes in Hexacorals

Abstract Zoantharia is an order among the Hexacorallia (Anthozoa: Cnidaria), and includes at least 300 species. Previously reported genomes from scleractinian corals and actiniarian sea anemones have illuminated part of the hexacorallian diversification. However, little is known about zoantharian genomes and the early evolution of hexacorals. To explore genome evolution in this group of hexacorals, here, we report de novo genome assemblies of the zoantharians Palythoa mizigama (Pmiz) and Palythoa umbrosa (Pumb), both of which are members of the family Sphenopidae, and uniquely live in comparatively dark coral reef caves without symbiotic Symbiodiniaceae dinoflagellates. Draft genomes generated from ultra-low input PacBio sequencing totaled 373 and 319 Mbp for Pmiz and Pumb, respectively. Protein-coding genes were predicted in each genome, totaling 30,394 in Pmiz and 24,800 in Pumb, with each set having ∼93% BUSCO completeness. Comparative genomic analyses identified 3,036 conserved gene families, which were found in all analyzed hexacoral genomes. Some of the genes related to toxins, chitin degradation, and prostaglandin biosynthesis were expanded in these two Palythoa genomes and many of which aligned tandemly. Extensive gene family loss was not detected in the Palythoa lineage and five of ten putatively lost gene families likely had neuronal function, suggesting biased gene loss in Palythoa. In conclusion, our comparative analyses demonstrate evolutionary conservation of gene families in the Palythoa lineage from the common ancestor of hexacorals. Restricted loss of gene families may imply that lost neuronal functions were effective for environmental adaptation in these two Palythoa species.


Introduction
Hexacorallia, a cnidarian class of the subphylum Anthozoa, is a major and diversified group that includes the orders Scleractinia (stony corals), Corallimorpharia (mushroom anemones), Antipatharia (black corals), Actiniaria (true anemones), Zoantharia (colonial anemones), and Ceriantharia (tube anemones).Among these orders, Zoantharia is a sister group to a clade consisting of four orders: Actiniaria, Antipatharia, Corallimorpharia, and Scleractinia (McFadden et al. 2021).Scleractinia and Actiniaria have been comparatively well studied at the molecular level for a relatively long period of time (Shinzato and Yoshioka 2024).For example, Forêt et al. (2010) examined the genomes of scleractinians, discussing selective gene loss in Cnidaria, and proposed the conservation of ancestral genes in Anthozoa and lineagespecific gene families as the genomic basis for the diversification of stony corals (Yoshioka et al. 2022).Additionally, robust molecular studies on the origins of anthozoans have recently suggested that Hexacorallia had a common ancestor in the Cryogenian (711Ma), far older than had previously been estimated (McFadden et al. 2021).Among anemones, Nematostella and Exaiptasia have become model organisms for many genomic studies, and there are many data available for both species (Putnam et al. 2007;Baumgarten et al. 2015).Whole genomes of hexacorals have been reported from sea anemones and scleractinian corals (Putnam et al. 2007;Shinzato et al. 2011).However, genomic studies on zoantharians remain sparse, and this lack of information inhibits our ability to properly understand genomic evolution within the hexacorals.
Although genomes of zoantharians have been assembled with short reads (Santos et al. 2023), comparative analyses have been restricted to transposable elements (Fourreau et al. 2023), and thus far, gene models from genome assemblies remain unavailable.As well, whole transcriptome shotgun assemblies (TSA) of zoantharians have been analyzed, focusing on genes related to venoms and toxins (Huang et al. 2016;Liao et al. 2018Liao et al. , 2019)).Although novel venomrelated transcripts, novel functional toxins, and six groups of expressed peptide toxins were found (Liao et al. 2018), their genomic bases and regulatory mechanisms remain unclear.
Zoantharians can broadly be separated into two main suborders, although this classification remains controversial (McFadden et al. 2021;Fourreau et al. 2023).Members in the suborder Macrocnemina are found from shallow waters to the deep sea, and many species (but not all) are epibiotic, and as a group, they are known to be in symbioses with a wide variety of different marine phyla (Kise et al. 2023).On the other hand, the suborder Brachycnemina includes mostly shallow water tropical and subtropical species, with the large majority of species being zooxanthellate, in symbioses with Symbiodiniaceae (Davies et al. 2023).Among the Brachycnemina, the genera Zoanthus and Palythoa are the most speciose and well-known (Reimer et al. 2023), and are often common species in coral reef ecosystems (Reimer et al. 2023).Palythoa spp.have received research attention for their ecological role on coral reefs (Irei et al. 2015;Reimer et al. 2023), as well as their ability to produce palytoxin (PTX) (Deeds et al. 2011), one of the most potent toxins known from nature.
Recently, two closely related species of zoantharians with exceptional features, Palythoa mizigama and Palythoa umbrosa, were described from the Ryukyu Archipelago, Japan (Irei et al. 2015).These two species inhabit low-light environments such as coral reef caves and have no associations with photosymbiotic Symbiodiniaceae algae, unlike their congeners.As until now, no studies have reported on whether these species harbor PTX, and there have only been limited phylogenetic studies on them (Irei et al. 2015).As azooxanthellate and congeneric species, the loss of photoendosymbionts and their evolution to live in caves make them unique among Palythoa.Combined with the overall lack of zoantharian genomic information and these two species make good targets for investigating genomic evolution under such conditions.
Accordingly, here we report the whole-genome assembly of the two zoantharians, P. mizigama and P. umbrosa (Fig. 1a and b).By comparative genomic analyses, we provide insight into genome evolution of hexacorals by focusing on expanded gene families and putative gene loss.These genomes will also serve as the future basis for comparisons with zooxanthellate congeneric Palythoa spp.

Genome Assembly and Gene Models
We obtained 56 Gb of PacBio HiFi reads for P. mizigama and 53 Gb for P. umbrosa (supplementary table S1,

Significance
Anthozoan hexacorals are an important animal group in many marine environments, and include at least ∼3,500 extant species, including reef-building scleractinian corals.We generated two Palythoa genomes from the order Zoantharia within Hexacorallia, providing novel insights into early hexacorallian evolution by comparing with genomes of diversified scleractinian corals and actiniarian sea anemones.These first available gene-sets from zoantharians demonstrated genome conservation with restricted neuronal gene loss, and the suggested expansion of enzyme genes in Palythoa may be related to the production of unique chemicals and toxins such as palytoxin.Overall, our analyses imply that lineagespecific tandem duplication of enzyme genes may have occurred in the genome evolution of Zoantharia.Supplementary Material online).We successfully assembled complete mitochondrial genomes with a length of 21,122 bp and 21,145 bp, encoding 13 protein-coding genes for P. mizigama and P. umbrosa, respectively (supplementary fig.S1, Supplementary Material online).
After removing mitochondrial and contaminated HiFi reads, we performed nuclear genome assembly of the Palythoa species, resulting in draft genome assemblies of 373 and 319 Mb for P. mizigama and P. umbrosa, with mean depth of 102 to 107x, respectively (Table 1, supplementary fig.S2, Supplementary Material online).K-mer profiles suggested that estimated genome sizes for P. mizigama and P. umbrosa were ∼330 Mb and heterozygosity rate of ∼3.79% for P. mizigama and ∼3.39% for P. umbrosa (supplementary fig.S3, Supplementary Material online).The estimated genome sizes around 300 Mb were supported based on other k-mer profiles (Fourreau et al. 2023), suggesting that assembly sizes obtained in this study did not deviate from the expected sizes.When we compared assembly statistics with other zoantharian genomes, the numbers of contigs were significantly reduced and the indices of continuity (N50 and mean contig size) were significantly improved (supplementary table S2, Supplementary Material online).While single copy category in BUSCO completeness (Huang and Li 2023) in reported zoantharian genomes (Santos et al. 2023) were 17% to 49%, they were larger than 90% in our assemblies (supplementary table S2, Supplementary Material online).QV scores calculated with Inspector were over 50 (Table 1).These results indicate that the assemblies presented in this study are the first cases that achieved high continuity in zoantharian genomes.We predicted 30,394 protein-coding genes for P. mizigama and 24,800 protein-coding genes for P. umbrosa (Table 1) based on protein-based gene prediction (see supplementary materials, Supplementary Material online).BUSCO completeness scores were 93.8% (of which 2.4% were duplicated) for P. mizigama and 91.8% (of which 0.9% were duplicated) for P. umbrosa (Table 1).This high BUSCO completeness score of gene models is comparable with those of gene models in the other cnidarians (supplementary table S3, Supplementary Material online), supporting acquisitions of high-quality gene models, and enabling more accurate comparative genomics to infer molecular bases of zoantharians.

Toxin-Related Genes in Palythoa Genomes
Prior studies using transcriptome assemblies have focused on genes encoding toxin-like polypeptides (Huang et al. 2016;Liao et al. 2018Liao et al. , 2019)).Putative toxin-related genes have been categorized into six main groups, with neurotoxin, hemostatic and hemorrhagic toxins, protease inhibitors, membrane-active peptides, mixed function enzymes, and peptides related to allergens and innate immunity components (Liao et al. 2018).We searched for their homologs in cnidarians based on orthogroup classification.Genes from each of these six groups were conserved in the other available cnidarian genomes (supplementary table S4, Supplementary Material online).Interestingly, some genes in these groups were tandemly arranged in both P. mizigama and P. umbrosa genomes (Fig. 2a; supplementary table S4, Supplementary Material online), suggesting that some of these tandem duplications occurred in the common ancestor of the genus Palythoa.Huang et al. (2016) suggested that putative toxins in Palythoa species are highly likely to be employed as an antipredatory armamentarium.Several species of fishes and turtles have been confirmed as predators of zoantharians (Stampar et al. 2007;Francini-Filho and Moura 2010).These duplications may enable rapid transcription of the genes when they face predators (Mathers et al. 2017).
The well-known toxin, PTX, has been detected in many different Palythoa specimens.It has been implied that polyketide synthases (PKSs) are related to the biosynthesis of PTX-like compounds (Verma et al. 2019).Our preliminary surveys of PKS genes found two PKS genes that encode multiple domain proteins (data not shown), suggesting no expansions of PKSs in these two Palythoa genomes.As the symbiosis-related function of a unique chemical that was biosynthesized with animal PKSs has been reported (Torres et al. 2020), zooxanthellate Palythoa spp.harboring symbiotic Symbiodiniaceae might have more than two PKS genes for the biosynthesis of unique chemicals (Deeds et al. 2011).

Gene Family Expansion in Palythoa Lineage
In order to reveal the molecular basis underlying evolution of zoantharians, we inferred evolutionary relationships of genes among anthozoans (using genomes of two zoantharians [this study], seven scleractinians, three actiniarians, and two octocorallians as outgroup).Here, we included anthozoan genome assemblies available in RefSeq (supplementary table S3, Supplementary Material online).We used the orthogroups produced by OrthoFinder as putative gene families in this study.To gain a more comprehensive overview of gene families in Palythoa (zoantharians), we also included three transcriptome assemblies from P. variabilis, P. caribaeorum, and Zoanthus sp. in the analyses, resulting in a total of 66,871 gene families (supplementary data, Supplementary Material online).Of these, 3,036 gene families were conserved in the hexacorallian genomes, i.e. all hexacorallians used in this study possessed at least one gene per gene family, of which 149 were single copy, which were also conserved in the two octocorallian genomes as single copies (supplementary table S5, Supplementary Material online).Using the 149 single copy gene families, we performed molecular phylogenetic analyses.The tree topology was identical to the reported phylogeny of class Anthozoa (McFadden et al. 2021), and that five zoantharians formed a cluster, with two clear groups, Palythoa (n = 4) separate from the single Zoanthus (Fig. 1c).
Gene expansion (gene duplication) contributes to the evolution of organisms (Conant and Wolfe 2008).Using 3,036 conserved gene families, we identified gene families whose size (the number of paralogs) in the genus Palythoa were twotimes different compared with that of the hexacorallian average.As transcriptome assemblies do not cover whole genes and there are difficulties in reducing redundancy (high duplicates BUSCO completeness, Table 1), we used P. mizigama and P. umbrosa as the representatives for the zoantharian lineage in the analyses.The sizes of 111 and 138 gene families in the two Palythoa genomes were two times larger or smaller than the average size in the other 10 hexacorallian genomes, respectively (Fig. 1d; supplementary tables S6 and S7, Supplementary Material online).Five functional categories, including immunity and cell adhesion, were identified by enrichment analysis as being significantly (FDR < 0.05) enriched in gene families with average sizes smaller than those in other hexacorallians (Table 2).In case of scleractinians, gene expansions for complex immune systems have previously been discussed to possibly be due to endosymbiosis with Symbiodiniaceae dinoflagellates (Shinzato et al. 2011).As our dataset included six scleractinians, increases of average gene family sizes of immune-related genes, including NOD-like receptors (OG00066), which have been shown to expand in a scleractinian coral lineage (Hamada et al. 2013), were observed (Fig. 1d, supplementary table S7, Supplementary Material online), confirming the reproducibility of the previous report.On the other hand, 10 functional

GBE
categories, including peptide transport, chitin degradation, and prostaglandin biosynthesis, were identified by enrichment analysis as being significantly (FDR < 0.05) enriched in gene families with average sizes larger than those in other hexacorallians (Table 2).Diverse toxin-like peptides have been reported from zoantharians (Liao et al. 2019), and expansions of peptide transport may be related to the diversification of toxin-like peptides by tandem gene duplications in zoantharians (Fig. 2a).Chitin is the second-most abundant polysaccharide in nature (Tharanathan and Kittur 2003) and serves as a structural element of the exoskeleton of crustacean, and in cell walls in fungi and algae (Gooday 1990).Zoantharians are known to incorporate sand and/or detritus into their tissues to help strengthen their structure (Haywick and Mueller 1997), possibly increasing their encounters with fungal pathogens, as various components, including algae, accumulate more in sediment than in the water column (Littman et al. 2008;Amend et al. 2019).Chitinase can hydrolyze chitin into chitin oligosaccharides and/or monosaccharides and is widely distributed in marine organisms including scleractinian corals (Yoshioka et al. 2017), octocorals (Douglas et al. 2007), and Palythoa caribaeorum (Souza et al. 2008).The tentacle feeding response of Palythoa species, as well as of scleractinian corals, in the presence of zooplankton has been reported (Goreau et al. 1971), suggesting that they utilize chitinases to consume various zooplankton, such as copepods, with chitinous exoskeletons.As P. mizigama and P. umbrosa have no algal symbionts, utilization of duplicated chitinases might be beneficial to heterotrophically obtain their energy budget from plankton prey.In addition, possible functions of chitinase as protection against fungal pathogens in cnidarians have been hypothesized (Yoshioka et al. 2017;van de Water et al. 2018), suggesting that these chitin degradation-related genes (supplementary table S8, Supplementary Material online) may also act in defense systems in zoantharians, and that gene expansion of these genes may be related to the adaptive evolution of zoantharians in order to rapidly degrade invasive organisms with chitin via gene expression.It has been hypothesized that prostaglandins in octocorals could function as chemical defense against predators but this remains unclear (Di Costanzo et al. 2019).Some of these enzyme genes may be related to the production of unique chemicals and toxins in addition to prostaglandin biosynthesis.
suggesting restrictive gene losses from the ancestral gene repertoire in the early zoantharian lineage.This result also suggests that genes involved in the biosynthesis of essential amino acids are conserved in the genomes of P. mizigama and P. umbrosa (Shinzato et al. 2011).
Our detailed genomic analyses suggest the genomes of two Palythoa from dark coral reef environments have lost eight common gene families (Fig. 2).These events include the loss of MEGF11, of which orthologs play a critical role in the formation of retinal interneuron in humans (Kay et al. 2012).
The loss of genes involved in light sensing and neuronal function is a possible case of environmental variability in these Palythoa species that lack algal Symbiodiniaceae symbionts.Our results indicate that light recognition system appears to have retrogressed at least in P. mizigama and P. umbrosa.It should be noted that as some genes were also not detected in the three zoantharian transcriptome assemblies and as these three species are zooxanthellate, genomic information from other zooxanthellate zoantharians are needed for further discussion.

Conclusions
In this study, we successfully obtained high quality gene models for two dark environment-adapted zoantharians, P. mizigama and P. umbrosa.Comparative genomic analyses among cnidarian anthozoans revealed putative gene losses, including a photoreceptor-related gene in both P. mizigama and P. umbrosa.Our genome assemblies identified expansions of enzyme genes related to prostaglandins and possibly toxins, although unique chemicals such as PTX still have not been identified in these two Palythoa species.For understanding zoantharians ecology and evolutionary success, further genomic information from sister-group species living in light environments will be needed and may illuminate new insight into genome evolution of hexacorals.

Materials and Methods
The method details are shown in Supplementary Material online at "Genome Biology and Evolution" online (supplementary Material online).

Fig. 1 .
Fig. 1.P. mizigama and P. umbrosa (Anthozoa: Hexacorallia: Zoantharia) in situ.a) A colony of P. mizigama in a marine cave.The specimens were used for genomic analyses in this study.The inset shows one illuminated polyp (∼5 mm in height).Image taken by C. J. L. Fourreau on October 9, 2023, at Mizugama.b) A colony of P. umbrosa used for the treatment of cell dissociations as the sample for genomic DNA extraction.The polyp is covered with grains of sand.Dissociated cells are indicated in the inset (Scale bar, 50 µm).Symbiodiniaceae cells were not found.c) Molecular phylogenetic tree of anthozoans constructed with 149 single copy orthologs.Asterisk indicates that genes are from transcriptome assemblies.Bootstrap value for each node was 100%, except one node with 99%.The bar indicates expected substitution per site in aligned regions.d) Gene families that were expanded or reduced in the Zoantharia lineage.Dots indicate 3,036 conserved gene families among 12 hexacoral genomes.Red and blue indicate their gene family sizes are two times larger or smaller than that of average in the other Hexacorallia.

Fig. 2 .
Fig. 2. Tandem gene arrangement of toxin-related genes and putative gene loss in Palythoa lineage.a) Toxin-like genes belonging to orthogroup ID 000039 (Pumb contig0274 and Pmiz contig0202) and orthogroup ID 000067 (Pumb contig0333 and Pmiz contig0678).Contig ID is shown in the left and gene ID is shown in each arrowhead.Arrowhead indicates transcriptional direction.Red and blue indicate P. umbrosa and P. mizigama genes, respectively.Black indicates other genes (i.e.not related to toxin-like peptides).b) Possible gene name and orthogroup ID are shown in the upper row (NA indicates unknown gene

Table 1
Statistics for genome assembly and gene prediction of P. mizigama and P. umbrosa with reported transcriptome shotgun assembly of zoantharians

Table 2
Differential biological process between zoantharians and other hexacorals predicted by gene family enrichment analysis