Table 1.
Strains used in this study and general sequence information of different Edwardsiella strains.
Figure 1.
Pathogenic characteristics of Edwardsiella strains.
Cumulative mortality in zebra fish i.m. challenged with the indicated Edwardsiella strains. No mortality was observed after 7 days of observation (data not shown). Error bars showed the standard deviations calculated from three individual experiments.
Figure 2.
Schematic comparison of the Edwardsiella genomes.
(A) The outside circle represents E. tarda EIB202 GIs (pink). The next 7 circles ranging from outside to inside show the coordinated mapping of 2 complete genomes (E. tarda FL6-60 and E. ictaluri 93–146) and 5 contig sets of genomes against E. tarda EIB202 reference genome sequence. (B and C) Dot plot comparison of MUMmer nucmer output between E. tarda EIB202 (x-axis) and E. ictaluri 93–146 (y-axis) (B), or between E. tarda EIB202 (x-axis) and E. tarda FL6-60 (y-axis) (C). Red and blue plot means forward and reverse matches, respectively.
Figure 3.
Phylogenetic tree of Edwardsiella species.
(A) Maximum likelihood phylogeny based on all filtered SNPs across 8 Edwardsiella genomes. Branches are colored according to the main phylogeographic lineages of Edwardsiella bacteria. (B) NJ tree of 48 Edwardsiella strains inferred from concatenated alignments of partial coding sequences of glyA, mdh, pgi, fusA, aspA and tpi genes with 100 bootstrap iterations. Strains investigated in this study are indicated in bold font.
Figure 4.
The Venn diagram illustrating the number of genes unique or shared between two Edwardsiella lineages.
The associated pie charts showed the functional groups assigned for CDSs in relevant sections of the Venn diagram. The strains used for comparison were E. tarda EIB202, FL6-60, and 080813 in EdwGI lineage, DT, ATCC15947, and ATCC23685 in EdwGII lineage, and E. ictaluri ATCC33202 and 93–146.
Figure 5.
LPS related genes of Edwardsiella strains.
Mauve progressive alignment of the concatenated coding sequences of 8 sequenced Edwardsiella strains using E. tarda EIB202 as reference. Arrows indicate the gene coding orientation in EIB202 genome. The dendrogram is derived from NJ analysis of concatenated amino acid sequences of LPS biosynthesis genes with 1,000 bootstrap iterations. Gradient bar indicated the sequence similarity of LPS coding sequences of Edwardsiella strains to those of EIB202.
Figure 6.
Genome-wide nucleotide variations among the orthologs of sequenced Edwardsiella strains.
(A) Nucleotide diversity (π) for 8 Edwardsiella strains. The black line represents the average π value of all orthologs. Green line indicates π values above 1.5σ (standard deviation) from the average π values of all orthologs, respectively. The percent of genes with π values large than 1.5σ from the average π value in each function category are shown under x axis. (B) Analysis of the ratio of nonsynonymous (NonSyn) to synonymous (Syn) SNP rates according to the RAST-annotated categories. The set of genes which contain significant high ratios (p<0.05) of nonsynonymous (NonSyn) SNPs than synonymous (Syn) SNPs between E. tarda EdwGI and E. tarda EdwGII strains were as indicated.
Table 2.
Representative genes with high diversity or under positive selection.
Figure 7.
Contributive roles of representative diversified or positive selection genes in the virulence and colonization in zebra fish.
(A) LD50 values of the wild-type EIB202 (WT) and the null mutants of the indicated genes. LD50 is calculated by the method described elsewhere [36]. (B) Virulence comparison of parental E. tarda EIB202 with the mutants with gene disruption in ETAE_1081, fur, and pvuA, respectively. Graphs show survival curves of zebra fish following injected with varying dosages of E. tarda strains. All mutant strains are significantly attenuated compared to parental EIB202 strain (p<0.01, Mantel-Haenszel Chi-squared test). (C) Competitive indexes of the indicated strains against WT in zebra fish at 24 h after inoculation. WT was differentiated from the mutant strain based upon GFP label or Km resistance on DHL agar plates as detailed in the Materials and Methods [37]. The strain ΔesrB with significantly attenuated virulence while transitorily enhanced CI was included in the experiments as a control [36], [37]. The p value of the decreased growth competition of the mutants against WT are shown under x axis (p<0.01, one sample t-test).
Figure 8.
Proposed hypothetical evolutionary pathway of Edwardsiella species.
Probable insertions, deletions of GIs and gene clusters found in 8 Edwardsiella strains are indicated by yellow and red arrows, respectively. Host change events of different strains are indicated by green arrows. Hypothetical ancestral strains are indicated by open circles.