Assessing the Genotypic Differences between Strains of Corynebacterium pseudotuberculosis biovar equi through Comparative Genomics

Abstract

Seven genomes of Corynebacterium pseudotuberculosis biovar equi were sequenced on the Ion Torrent PGM platform, generating high-quality scaffolds over 2.35 Mbp. This bacterium is the causative agent of disease known as "pigeon fever" which commonly affects horses worldwide. The pangenome of biovar equi was calculated and two phylogenomic approaches were used to identify clustering patterns within Corynebacterium genus. Furthermore, other comparative analyses were performed including the prediction of genomic islands and prophages, and SNP-based phylogeny. In the phylogenomic tree, C. pseudotuberculosis was divided into two distinct clades, one formed by nitrate non-reducing species (biovar ovis) and another formed by nitrate-reducing species (biovar equi). In the latter group, the strains isolated from California were more related to each other, while the strains CIP 52.97 and 1/06-A formed the outermost clade of the biovar equi. A total of 1,355 core genes were identified, corresponding to 42.5% of the pangenome. This pangenome has one of the smallest core genomes described in the literature, suggesting a high genetic variability of biovar equi of C. pseudotuberculosis. The analysis of the similarity between the resistance islands identified a higher proximity between the strains that caused more severe infectious conditions (infection in the internal organs). Pathogenicity islands were largely conserved between strains. Several genes that modulate the pathogenicity of C. pseudotuberculosis were described including peptidases, recombination enzymes, micoside synthesis enzymes, bacteriocins with antimicrobial activity and several others. Finally, no genotypic differences were observed between the strains that caused the three different types of infection (external abscess formation, infection with abscess formation in the internal organs, and ulcerative lymphangitis). Instead, it was noted that there is a higher phenetic correlation between strains isolated at California compared to the other strains. Additionally, high variability of resistance islands suggests gene acquisition through several events of horizontal gene transfer.

Fig 1. Map of the circular genomes of C. pseudotuberculosis, C. ulcerans and C. diphtheriae and the analysis of GEIs.

The genomes were compared by blastn, and the percentage of identity between them was determined by the intensity of the color in the circular map. The genomes of C. pseudotuberculosis were identified only by the name of the lineage. (a) The innermost ring to the outermost is presented in this figure, as follows: the GC skew of the MB14 strain; the GC contents of the MB14 strain, the genomes of the C. pseudotuberculosis strains MB14, MB20, MB30, MB44, MB66, MB122, MB154, MB278, MB295, MB302, and MB336, and the genomes of C. ulcerans BR-AD22 and C. diphtheriae NCTC 13129. The two outermost rings comprise CDSs identified in the genome of C. pseudotuberculosis MB14. The graph displayed by GC skew is common to chromosomes that have bidirectional replication. (b) A circular genome map of the MB20 strain was constructed by comparing the position of 18 PAIs in the MB122 strain and 21 PAIs the in MB336 strain through tblastx. The PAIs shown in black were detected in both strains by GIPSy, while the PAIs in green were detected only in MB122 and the PAIs in red were detected only in MB336. A later comparison using blastn showed that the conservation of these PAIs was higher than that observed by GIPSy. (c) A dendrogram was calculated with the Neighbor-Joining model from the comparison of the nucleotide sequences of the RIs of 12 Californian isolated genomes.

Fig 2. Synteny analysis of C. pseudotuberculosis pili gene clusters.

From top to the bottom of the figure the genome of strains 1002, MB11, MB14, MB30 and CIP 52.97 are shown. Genes are represented by arrows of different color. Genes without standardized names are identified as gene_1, gene_2, and so on. Conserved genes are connected by lines. (a) spaA cluster. (b) spaD cluster.

Fig 3. Circular map of C. pseudotuberculosis prophages identified with the online tool PHAST.

Each circular map represents an identified prophage. The outermost ring describes the genes of each prophage that were interspersed (colored in blue and red). The acronym "unk" identifies uncharacterized proteins. The innermost ring to the outermost represents the results of blastn for the genomes of C. pseudotuberculosis MB11, MB14, MB20, MB30, MB44, MB66, MB122, MB154, MB278, MB295, MB302, and MB336. The level of identity, nearly 100% for all genes, indicated high conservation of these prophages in the 12 analyzed genomes.

Fig 4. Phylogenomic analysis of the Corynebacterium genus.

(a) The entire Neighbor-Joining tree containing the Corynebacterium RefSeq genomes, demonstrating the closeness between the species C. diphtheriae, C. ulcerans and C. pseudotuberculosis. This latter species comprises two distinct clades composed of strains from biovar ovis (marked in orange) and biovar equi (marked in purple). (b) An expanded view of the biovar equi tree. Strains isolated in California (in color) are closely related to each other. The strains CIP 52.97 and 1/06-A formed the outermost clade of biovar equi. The colors of the names from strains isolated in California distinguish the types of infection caused by the pathogen. (c) Dendrogram calculated with the Neighbor-Joining model from heatmap values generated in the Gegenees program. The values shown in the heatmap indicate the percentage of similarity between the analyzed genomes. The name of the species isolated in California is colored according to the type of infection caused by the pathogen.

Fig 5. Prediction of the core genome and pangenome of eighteen genomes of C. pseudotuberculosis biovar equi.

The graphic bars represent the number of core genes (red bars) and pangenome (blue bars) for n combinations of the analyzed genomes. The dotted lines indicate the standard deviation of these combinations and white circles outside the dotted lines indicate discordant values of the analysis. The calculation of the pangenome and core genome size showed very similar values using either the mean (yellow line) or median (green line) of the possible combinations.

Fig 6. Graphical representations of the pangenome characteristics of biovar equi.

(a) Flower graph representing the number of core and accessory genes for each strain of biovar equi. The strains are listed by their names along with each set of singletons. Those strains causing internal or external abscesses are indicated by black lines. (b) Bar graph of singletons classified in COG categories. The numbers in the bars indicate the amount of genes not classified in COGs.