Highly contiguous genome assemblies ofPhotobacteriumstrains isolated from fish light organs using nanopore sequencing technology

AbstractSeveral species of luminous bacteria in the genusPhotobacteriumare the light organ symbionts of teleost fishes.Photobacterium leiognathiand its subspecies,P. mandapamensis, in particular, commonly form bioluminescent symbioses with fish hosts in the Leiognathidae and Acropomatidae families as well as with cardinalfish in the genusSiphamia(Apogonidae). These two closely related lineages ofPhotobacteriumare right at the cutoff average nucleotide identity used to delimit bacterial species (95-96%) and show overlapping ecological niches, including their host fish range. However, there are only a few whole genome assemblies available for these bacterial species, particularly for symbiotic strains isolated from fish light organs, that can be used to explore genome evolution of these two lineages. Here we used Oxford Nanopore Technologies sequencing to produce long reads for assembling highly contiguous genomes ofPhotobacteriumstrains isolated from fish light organs, including severalP. kishitaniistrains isolated from deep water fishes. We were able to assemble 31 high-quality genomes with near complete BUSCO scores, many at the chromosome-level, and compare their gene contents, including plasmid genes. In doing so, we discovered a new candidate species ofPhotobacterium, CandidatusPhotobacterium acropomis, which originated from the light organ of the acropomid fish,Acropoma japonicum. We also describe a lack of congruency between the presence of theluxFgene, which is involved in light production, and the phylogenetic relationships between closely relatedP. leiognathiandP. mandapamensisstrains. In contrast, there was strong congruency betweenluxFand the host fish family of origin, suggesting this gene might be essential to initiate bioluminescent symbioses with certain hosts, includingSiphamiaandAcropomaspecies. Our study shows the benefit of using long reads in the assembly of bacterial genomes and outlines an assembly pipeline that results in highly contiguous genomes, even from low-coverage ONT reads.