The mitochondrial genomes of Crispatotrochus rubescens and Crispatotrochus rugosus (Hexacorallia; Scleractinia): new insights on the phylogeny of the family Caryophylliidae

Caryophylliidae is one of the most diverse scleractinian families, however it was recovered as polyphyletic in multiple molecular studies. Recently, the mitochondrial gene order was proposed as a character for a taxonomic revision of the family. Here we describe the first mitogenome of the caryophylliid genus Crispatotrochus, whose phylogenetic position remains uncertain. The complete mitochondrial genomes of Crispatotrochus rubescens and Crispatotrochus rugosus were sequenced, assembled, and annotated. The two mitogenomes are identical and circular, have a length of 16,536 bp, a GC content of 35.9%, and contain 13 protein-coding genes, 2 ribosomal RNAs and 2 transfer RNAs. Both species have a transposition of a three gene block - cob, nad2, and nad6 - similarly to a group of caryophylliid genera that were recovered as monophyletic, including the type genus (Caryophyllia) of the family. The phylogenetic analyses recovered Crispatotrochus within the clade that presents the gene rearrangement and specifically as sister taxa of the genus Caryophyllia, a result consistent with previous studies and the similar gross morphology of the two genera. We determined the mitochondrial genomes of the genus Crispatotrochus to investigate their relations within Scleractinia. Results from this study provide insights on the phylogenetic position of the genus and corroborate that the mitochondrial gene order could be used as taxonomic character for the family Caryophylliidae.


Introduction
Corals belonging to the order Scleractinia are distributed worldwide and are the engineers of complex shallow and deep-water reef ecosystems. Despite their importance, a well-resolved phylogeny of the order has not yet been achieved (e.g., [1]), hampering the study of longstanding evolutionary questions. Since the advent of the first molecular studies, scleractinian corals have been divided into two (e.g., [2][3][4]) or three (e.g., [5,6]) main clades at the suborder level, and several families and genera have been recovered as para-or polyphyletic (see [1]).
Within the Vacatina/Robust clade, the family Caryophylliidae Dana, 1846 is currently one of the most diverse, comprising solitary and colonial species, of which the latter ones include important components of deep-water reefs [7]. However, recent phylogenetic reconstructions recovered the family as polyphyletic with its members divided into several clades spread across the scleractinian phylogeny (e.g., [1,5]). Moreover, many taxa belonging to this family still lack molecular information. Therefore, the evolutionary history of this family is still obscure, and its taxonomic revision is in progress. In a recent study, Seiblitz and colleagues [6] showed that all components recovered in a monophyletic clade that harbors the caryophylliid genera Caryophyllia (type taxon of the family), Desmophyllum, Premocyathus, and Solenosmilia have a transposition of a three gene block ( Fig. 1) from the canonical scleractinian mitochondrial gene (mtgene) order (rearrangement previously known only for the genera Desmophyllum and Solenosmilia [8][9][10][11]). However, this rearrangement was not observed in any other analyzed mitogenome from genera formally belonging to the family but not to the molecular clade including Caryophyllia, such as Heterocyathus, Polycyathus, and Trochocyathus. Therefore, such a mtgene order was proposed as a taxonomic character/synapomorphy of the "true" Caryophylliidae, leading to the hypothesis that this family is much smaller than previously thought -but data from several genera are still needed in order to come to a definitive conclusion.
The caryophylliid genus Crispatotrochus Tenison-Woods, 1878 is distributed worldwide and comprises 14 extant azooxanthellate and solitary species. Despite being recovered as closely related to Caryophyllia in three studies [1,12,13], only Crispatotrochus rugosus has nucleotide sequences available to date (one nuclear marker -i.e., 28S rDNA, and two mitochondrial markers -i.e., 12S rDNA, 16S rDNA). Hence the phylogenetic position of the genus is still under debate. In this study, we obtained the complete mitochondrial genomes of Cr. rubescens and Cr. rugosus, compared them with other caryophylliid mitogenomes, and investigated the phylogenetic position of the genus. The data herein represent a step forward for untangling relationships between azooxanthellate species and will be fundamental for a future revision of the family. Muséum national d'Histoire naturelle, Paris, France) was extracted using the DNeasy Blood and Tissue kit (Qiagen) following the manufacturer's animal tissue protocol. DNA quality and integrity were assessed on a microvolume spectrophotometer (Nanodrop, Thermo Fisher Scientific) and in a 1% agarose gel electrophoresis, respectively. DNA concentration before and after library preparation was quantified with Qubit fluorometer (Thermo Fisher Scientific). Libraries were prepared using the TruSeq DNA Nano Library Preparation kit (Illumina) with modifications in index adapter concentration and the number of PCR cycles (see [6]). Libraries were then sequenced on an Illumina NovaSeq 6000 (150 bp PE reads, two lanes combined with 71 samples from other studies) at the Human Genome and Stem Cell Research Center (CEGH-CEL, USP).

Mitochondrial genome assembly and annotation
Quality control of sequencing data was performed with Trimmomatic [14] and trimmed sequences were assembled into contigs using SPAdes v 3.1 [15] (-careful parameter). For both species, the mitogenome was recovered as a single and circular contig. Genes were annotated using MITOS2 online tool [16] with the parameters genetic code 4 (mold) and RefSeq 89 Metazoa. Annotation was manually verified using Geneious Prime 2022.2.1 (Biomatters Ltd. Auckland, New Zealand) with four published Caryophylliidae mitogenomes used as reference sequences (Caryophyllia scobinosa, OL584334; Desmophyllum pertusum, KC875348; Desmophyllum dianthus, KX000893; Solenosmilia variabilis, KM609293). Boundaries of all genes were then confirmed using BLAST [17] against either the NCBI nucleotide database or non-redundant protein sequences database.

Phylogenetic analysis
Once mitogenomes were fully annotated, they were included in a phylogenetic reconstruction together with 25 published mitogenomes of species belonging to the Vacatina/Robust Protein-coding, tRNA, and rRNA genes were abbreviated as in the text. Blank regions between genes represent intergenic spacers. The NAD5 intron is indicated by the inner gray line. Transposed genes are marked in bold and an asterisk (*) indicates the canonical position of this gene block for Scleractinia clade and one outgroup (Porites lobata belonging to the Refertina/Complex clade) for a total of 28 mitogenomes. Sequence alignments of protein coding, transfer RNA, and ribosomal RNA genes were performed with MUS-CLE 3.8.425 [18]. Alignments were visually inspected for ambiguous sites and successively concatenated resulting in a final alignment of 14,741 bp. For the phylogenetic reconstruction a Maximum Likelihood analysis was performed with a gene partition set on RAxML v8.2.12 [19] using the GTR + GAMMA model, 1000 rapid bootstrap replicates and 20 random starting trees.

Mitochondrial genome features
The average assembly coverages for Cr. rubescens and Cr. rugosus were 232.2 and 245.1 X, respectively. The two determined mitochondrial genomes (Genbank accession numbers: OP594308; OP594309) are identical and circular, with a total length of 16,536 bp and a GC content of 35.9% ( Fig. 1 [6,20]). Hence, Crispatotrochus mitogenomes characteristics reflect those from vacatinian species. They represent the longest known mitogenomes within "true" caryophylliids (clade comprised by all species with the mitochondrial transposition; Fig. 2) and have a GC content most similar to Ca. scobinosa (Table 1).
Similar to other scleractinians, the studied mitogenomes contain 13 protein coding, 2 transfer RNA, and 2 ribosomal RNA genes (Fig. 1). A feature found in some scleractinian mitogenomes is the presence of an intron in the gene cox1 [21]. Nevertheless, the intron is absent in the mitogenomes from both Crispatotrochus, as in all caryophylliids sequenced to date [6]. Both Crispatotrochus mitogenomes have the transposition of three genes cob, nad2, and nad6, between the nad5 5′ and the trn-Trp (Fig. 1), similarly to the other "true" caryophylliid species [6]. An interesting feature of the Crispatotrochus mitogenomes is the presence of a long intergenic region (IGR) [879 bp] between atp8 and cox1 (Fig. 1). Such a long IGR, not observed in any other caryophylliid mitogenome, results in the longest mitogenomes known for all representatives of the clade of the "true" caryophylliids. Furthermore, Emblem and colleagues [8] reported the presence of repeated regions at the end of nad1 and the beginning of cob in the mitogenome of D. pertusum and proposed they played a role in the mechanism of the genes transposition. Later, Seiblitz and colleagues [6] found the same repeated regions in Caryophyllia and Solenosmilia, but not in Premocyathus, possibly due to a secondary loss of this characteristic. The aforementioned repeated regions were also found in the Crispatotrochus mitogenomes but when their repeated regions are aligned, they have more base pair differences than Caryophyllia scobinosa, for example. This result supports the hypothesis that these repeated regions might have been independently lost more than once in the family Caryophylliidae.
Interestingly, although both species have marked morphological autapomorphies (see [22]), their mitogenomes are identical. This is somehow unusual since base pair variations have been recovered even within specimens belonging to the same species (e.g., Desmophyllum pertusum -see [6,9]; and Solenosmilia variabilis -see [10]). Nevertheless, it is renowned that scleractinian mitochondrial genomes show  [6] slow rates of evolution [23], and their identical mitogenome sequences might indicate recent speciation.

Phylogenetic position
In the phylogenetic reconstruction (Fig. 2), the genus Crispatotrochus was recovered inside the clade composed by the caryophylliid genera that present the mtgene rearrangement, which is consistent with the mtgene order recovered for this genus. Specifically, Crispatotrochus was recovered as sister taxon of the genus Caryophyllia, a result that mirrors the similar gross morphological features of the two genera [24]. The phylogenetic position of Crispatotrochus is also consistent with the results from Romano and Cairns [12] and Barbeitos and colleagues [13] that recovered the genera Caryophyllia and Crispatotrochus as sister taxa using a combination of mitochondrial and nuclear markers (16S/28S and 12S/28S, respectively). Reliable phylogenetic reconstructions are the basis for better understanding diversification processes in different groups of animals, and evolutionary studies of the order Scleractinia have been long suffering an undersampling of azooxanthellate and deep-sea species [25]. In this scenario, results from this study expand our knowledge about molecular features of the azooxanthellate caryophylliid genus Crispatotrochus, provide the first molecular data available for the species Cr. rubescens and will be fundamental for a future revision of the taxonomic challenging family Caryophylliidae. Moreover, they add evidence to the hypothesis that the gene transposition could be a diagnostic feature and a synapomorphy of "true" caryophylliids and that, consequently, the family is in fact smaller than previously thought.
Acknowledgements MVK thanks the support of FAPESP (Processes #2014/01332-0 and #2021/06866-6) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (301436/2018-5). The first author is supported by a PhD scholarship from Conselho Nacional de Desenvolvimento Científico e Tecnológico (142149/2018-7). We are also grateful to the four anonymous reviewers, whose helpful and positive feedback has been incorporated into a revised text. This paper is a contribution of NP-BioMar, USP.