Flavobacterium flabelliforme sp. nov. and Flavobacterium geliluteum sp. nov., Two Multidrug-Resistant Psychrotrophic Species Isolated From Antarctica

Despite unfavorable Antarctic conditions, such as cold temperatures, freeze-thaw cycles, high ultraviolet radiation, dryness and lack of nutrients, microorganisms were able to adapt and surprisingly thrive in this environment. In this study, eight cold-adapted Flavobacterium strains isolated from a remote Antarctic island, James Ross Island, were studied using a polyphasic taxonomic approach to determine their taxonomic position. Phylogenetic analyses based on the 16S rRNA gene and 92 core genes clearly showed that these strains formed two distinct phylogenetic clusters comprising three and five strains, with average nucleotide identities significantly below 90% between both proposed species as well as between their closest phylogenetic relatives. Phenotyping revealed a unique pattern of biochemical and physiological characteristics enabling differentiation from the closest phylogenetically related Flavobacterium spp. Chemotaxonomic analyses showed that type strains P4023(T) and P7388(T) were characterized by the major polyamine sym-homospermidine and a quinone system containing predominantly menaquinone MK-6. In the polar lipid profile phosphatidylethanolamine, an ornithine lipid and two unidentified lipids lacking a functional group were detected as major lipids. These characteristics along with fatty acid profiles confirmed that these species belong to the genus Flavobacterium. Thorough genomic analysis revealed the presence of numerous cold-inducible or cold-adaptation associated genes, such as cold-shock proteins, proteorhodopsin, carotenoid biosynthetic genes or oxidative-stress response genes. Genomes of type strains surprisingly harbored multiple prophages, with many of them predicted to be active. Genome-mining identified biosynthetic gene clusters in type strain genomes with a majority not matching any known clusters which supports further exploratory research possibilities involving these psychrotrophic bacteria. Antibiotic susceptibility testing revealed a pattern of multidrug-resistant phenotypes that were correlated with in silico antibiotic resistance prediction. Interestingly, while typical resistance finder tools failed to detect genes responsible for antibiotic resistance, genomic prediction confirmed a multidrug-resistant profile and suggested even broader resistance than tested. Results of this study confirmed and thoroughly characterized two novel psychrotrophic Flavobacterium species, for which the names Flavobacterium flabelliforme sp. nov. and Flavobacterium geliluteum sp. nov. are proposed.