Comparative genomic analyses identify the Vibrio harveyi genome sequenced strains BAA-1116 and HY01 as Vibrio campbellii

Abstract

Three notable members of the Harveyi clade, Vibrio harveyi, Vibrio alginolyticus and Vibrio parahaemolyticus, are best known as marine pathogens of commercial and medical import. In spite of this fact, the discrimination of Harveyi clade members remains difficult due to genetic and phenotypic similarities, and this has led to misidentifications and inaccurate estimations of a species' involvement in certain environments. To begin to understand the underlying genetics that complicate species level discrimination, we compared the genomes of Harveyi clade members isolated from different environments (seawater, shrimp, corals, oysters, finfish, humans) using microarray-based comparative genomic hybridization (CGH) and multilocus sequence analyses (MLSA). Surprisingly, we found that the only two V. harveyi strains that have had their genomes sequenced (strains BAA-1116 and HY01) have themselves been misidentified. Instead of belonging to the species harveyi, they are actually members of the species campbellii. In total, 28% of the strains tested were found to be misidentified and 42% of these appear to comprise a novel species. Taken together, our findings correct a number of species misidentifications while validating the ability of both CGH and MLSA to distinguish closely related members of the Harveyi clade.

Fig. 1

Hierarchically clustered heat maps based on CGH profiles demonstrating the presence and absence of genes within Harveyi clade members with respect to V. harveyi BAA-1116. A. Chromosome I, 2999 CDS. B. Chromosome II, 1765 CDS. The CDS in each heat map are ordered according to the genome structure of strain BAA-1116. Each CDS is depicted by one of five possible hybridization states (scale bar): (i) positive hybridization (CDS present call) = bright red bars, (ii) between positive and intermediate hybridization (uncertain ‘intermediate high’ call) = dark red bars, (iii) intermediate hybridization (uncertain ‘intermediate’ call) = grey bars, (iv) between intermediate and no hybridization (uncertain ‘intermediate low’ call) = dark green bars, and (v) no hybridization (CDS absent call) = bright green bars. Presence/absence designations generated from the hybridization profiles were calculated using the avgdiff method and clustered and visualized using MultiExperiment Viewer (MeV v4.4) software. ‘Para./algin.’ = parahaemolyticus and alginolyticus subclade.

Fig. 2

Multilocus sequence analysis (MLSA) of Harveyi clade members. A. Phylogenetic tree based on the Neighbour-Joining method using concatenated sequences from the ftsZ, mreB and topA genes (1528 nt) and MEGA software v4.0. Original species designations are in brackets. Strains lacking species designations were originally identified as V. harveyi. This analysis includes all of the strains used in the CGH analyses (with the exception of strain 50A) and four additional strains that are denoted with an asterisk ‘*’. The primary sequence information has been submitted to the GenBank database and the relevant accession numbers can be found in Table S1. B. Phylogenetic tree based on the Neighbour-Joining method using concatenated sequences from the rpoD, rctB and toxR genes (1848 nt) and MEGA software v4.0. Strain identifiers ending in ‘**’ denote type strains. With the exception of BAA-1116, HY01 and AND4 (bold type), all sequences used in this MLSA were downloaded from the ‘Taxonomy of the Vibrios’ database (http://www.taxvibrio.lncc.br/). Alignments for both analyses were generated using the clustalw program and bootstrap percentages > 50% from 1000 simulations are shown to the left of each branch point. The scale bar represents the number of substitutions per site.