Background: Population outbreaks of the crown-of-thorns starfish (Acanthaster planci sensu lato; COTS), a primary predator of reef-building corals in the Indo-Pacific Ocean, are a major threat to coral reefs. While biological and ecological knowledge of COTS has been accumulating since the 1960s, little is known about its associated bacteria. The aim of this study was to provide fundamental information on the dominant COTS-associated bacteria through a multifaceted molecular approach. Methods: A total of 205 COTS individuals from 17 locations throughout the Indo-Pacific Ocean were examined for the presence of COTS-associated bacteria. We conducted 16S rRNA metabarcoding of COTS to determine the bacterial profiles of different parts of the body and generated a full-length 16S rRNA gene sequence from a single dominant bacterium, which we designated COTS27. We performed phylogenetic analysis to determine the taxonomy, screening of COTS27 across the Indo-Pacific, FISH to visualize it within the COTS tissues, and reconstruction of the bacterial genome from the hologenome sequence data. Results: We discovered that a single bacterium exists at high densities in the subcuticular space in COTS forming a biofilm-like structure between the cuticle and the epidermis. COTS27 belongs to a clade that presumably represents a distinct order (so-called marine spirochetes) in the phylum Spirochaetes and is universally present in COTS throughout the Indo-Pacific Ocean. The reconstructed genome of COTS27 includes some genetic traits that are probably linked to adaptation to marine environments and evolution as an extracellular endosymbiont in subcuticular spaces.

The crown-of-thorns starfish (COTS) is a coral predator that is widely distributed in Indo-Pacific Oceans. A previous phylogenetic study using partial mitochondrial sequences suggested that COTS had diverged into four distinct species, but a nuclear genome-based analysis to confirm this was not conducted. To address this, COTS species nuclear genome sequences were analyzed here, sequencing Indian Ocean (NIO) and Red Sea (RS) species genomes for the first time, followed by a comparative analysis with the Pacific Ocean (PO) species. Phylogenetic analysis and ADMIXTURE analysis revealed clear divergences between the three COTS species. Furthermore, within the PO species, the phylogenetic position of the Hawaiian sample was further away from the other Pacific-derived samples than expected based on the mitochondrial data, suggesting that it may be a PO subspecies. The pairwise sequentially Markovian coalescent model showed that the trajectories of the population size diverged by region during the Mid-Pleistocene transition when the sea-level was dramatically decreased, strongly suggesting that the three COTS species experienced allopatric speciation. Analysis of the orthologs indicated that there were remarkable genes with species-specific positive selection in the genomes of the PO and RS species, which suggested that there may be local adaptations in the COTS species.

The estimation of larval dispersal on an ecological timescale is significant for conservation of marine species. In 2018, a semi-population outbreak of crown-of-thorns sea star, Acanthaster cf. solaris, was observed on a relatively isolated oceanic island, Ogasawara. The aim of this study was to assess whether this population outbreak was caused by large-scale larval recruitment (termed secondary outbreak) from the Kuroshio region. We estimated larval dispersal of the coral predator A. cf. solaris between the Kuroshio and Ogasawara regions using both population genomic analysis and simulation of oceanographic dispersal. Population genomic analysis revealed overall genetically homogenized patterns among Ogasawara and other Japanese populations, suggesting that the origin of the populations in the two regions is the same. In contrast, a simulation of 26-year oceanographic dispersal indicated that larvae are mostly self-seeded in Ogasawara populations and have difficulty reaching Ogasawara from the Kuroshio region within one generation. However, a connectivity matrix produced by the larval dispersal simulation assuming a Markov chain indicated gradual larval dispersal migration from the Kuroshio region to Ogasawara in a stepping-stone manner over multiple years. These results suggest that the 2018 outbreak was likely the result of self-seeding, including possible inbreeding (as evidenced by clonemate analysis), as large-scale larval dispersal from the Kurishio population to the Ogasawara population within one generation is unlikely. Instead, the population in Ogasawara is basically sustained by self-seedings, and the outbreak in 2018 was also most likely caused by successful self-seedings including possible inbreeding, as evidenced by clonemate analysis. This study also highlighted the importance of using both genomic and oceanographic methods to estimate larval dispersal, which provides significant insight into larval dispersal that occurs on ecological and evolutionary timescales.

Estimating the spatial extent of gamete and larval dispersal of deep-sea coral species, is challenging yet important for their conservation. Spatial autocorrelation analysis is useful for estimating the spatial range of dispersal of corals; however, it has not been performed for deep-sea coral species using genome-wide single nucleotide polymorphisms (SNPs). In this study, we examined the spatial genetic structure of a deep-sea coral species—the Japanese red coral, Corallium japonicum, sampled off the coast of Kochi, which lies to the southwest of the Shikoku Island in Japan; the Kochi region suffers from over-harvesting because of its high commercial value. We also examined the power of detecting significant spatial genetic structure by changing the number of loci and the proportion of missing data using both de novo analysis and mapping analysis. Similar results were obtained for both de novo and mapping analysis, although a higher number of loci were obtained by the mapping method. In addition, “many SNPs with a lot of missing data” was generally more useful than “a small number of SNPs with a small amount of missing data” to detect significant fine-scale spatial genetic structure. Our data suggested that more than 700 neutral SNPs were needed to detect significant fine-scale spatial genetic structure. The maximum first distance class that can detect significant spatial genetic structure within Kochi for the C. japonicum population was less than 11 km, suggesting that the over-harvesting of C. japonicum within a diameter of approximately 11 km in the Kochi area should be avoided, because this can cause the local extinction of this species.

No abstract