Salmonella enterica Phylogeny Based on Whole-Genome Sequencing Reveals Two New Clades and Novel Patterns of Horizontally Acquired Genetic Elements

Abstract

Using whole-genome sequence (WGS) data from the GenomeTrakr network, a globally distributed network of laboratories sequencing foodborne pathogens, we present a new phylogeny of Salmonella enterica comprising 445 isolates from 266 distinct serovars and originating from 52 countries. This phylogeny includes two previously unidentified S. enterica subsp. enterica clades. Serovar Typhi is shown to be nested within clade A. Our findings are supported by both phylogenetic support, based on a core genome alignment, and Bayesian approaches, based on single-nucleotide polymorphisms. Serovar assignments were refined by in silico analysis using SeqSero. More than 10% of serovars were either polyphyletic or paraphyletic. We found variable genetic content in these isolates relating to gene mobilization and virulence factors which have different distributions within clades. Gifsy-1- and Gifsy-2-like phages appear more prevalent in clade A; other viruses are more evenly distributed. Our analyses reveal IncFII is the predominant plasmid replicon in S. enterica Few core or clade-defining virulence genes are observed, and their distributions appear probabilistic in nature. Together, these patterns demonstrate that genetic exchange within S. enterica is more extensive and frequent than previously realized, which significantly alters how we view the genetic structure of the bacterial species.IMPORTANCE Rapid improvements in nucleotide sequencing access and affordability have led to a drastic increase in availability of genetic information. This information will improve the accuracy of molecular descriptions, including serovars, within S. enterica Although the concept of serovars continues to be useful, it may have more significant limitations than previously understood. Furthermore, the discrete absence or presence of specific genes can be an unstable indicator of phylogenetic identity. Whole-genome sequencing provides more rigorous tools for assessing the distributions of these genes. Our phylogenetic and genetic content analyses reveal how active genetic elements are dynamically distributed within a species, allowing us to better understand genetic reservoirs and underlying bacterial evolution.

Keywords: GenomeTrakr; Salmonella; phylogeny; plasmids; virulence; whole-genome sequencing.

FIG 1

phylogenetic analysis of S. enterica. Shown is a maximum-likelihood phylogeny of S. enterica calculated using a core genome alignment with RAxML. Bootstrap support is included for the major lineages. Named clades within Salmonella enterica subsp. enterica are indicated by color. The top 20 serovars found in the United States in 2015, plus serovar Typhi, are noted with asterisks. The remaining labeled serovars are nonmonophyletic. (a) Bayesian clustering for 6 assumed populations calculated using core SNPs found using kSNP with STRUCTURE. (b) Continent from which each isolate originated, indicated by color: Africa, red; Asia, green; Europe, blue; North America, purple; South America, yellow; Oceania, cyan; unknown, black.

FIG 2

Distribution of prophages. Distribution of prophages for which more than 10 related examples were found scored “intact” by PHASTER, with a cutoff of 50% of CDSs assigned to the indicated phage type used for phage assignment. Color indicates the percentage of CDSs within the phage region that matches the indicated phage on a linear scale from red at 50% to green at 100%. The phylogeny at the top uses the same colors as Fig. 1 for the different clades.

FIG 3

Plasmid replicon abundance in Salmonella enterica. Shown is the number of times the indicated replicon was identified within the isolates using the PlasmidFinder database and an 80% nucleotide identity cutoff threshold.

FIG 4

Distribution of genes encoding T3SS-delivered proteinaceous effectors and typhoid toxins. Shown is the presence or absence of genes coding for proteinaceous effectors delivered by the T3SS encoded in SPI-1, SPI-2, or delivered by both (“1/2”), as well as typhoid toxin subunit-encoding genes. The phylogeny at the top uses the same colors as Fig. 1 for the different clades. Color indicates the hit percentage of nucleotide identity on a linear scale from yellow at 80% to green at 100%. sspH1 and sspH2 were only shown above the 90% threshold to account for the high similarity of the alleles.

FIG 5

Distribution of fimbrial operons. Shown is the presence or absence of the fimbria-encoding operons. The phylogeny at the top uses the same colors as Fig. 1 for the different subclades. Positive hits had at least 80% nucleotide identity for greater than 75% of the genes in the operon, using the VFDB as a reference. A positive hit is indicated in green with no respect to sequence identity.