Evidence of microevolution of Salmonella Typhimurium during a series of egg-associated outbreaks linked to a single chicken farm

Abstract

Background: The bacterium Salmonella enterica serovar Typhimurium (S. Typhimurium) is one of the most frequent causes of foodborne outbreaks of gastroenteritis. Between 2005-2008 a series of S. Typhimurium outbreaks occurred in Tasmania, Australia, that were all traced to eggs originating from a single chicken farm. We sequenced the genomes of 12 isolates linked to these outbreaks, in order to investigate the microevolution of a pathogenic S. Typhimurium clone in a natural, spatiotemporally restricted population.

Results: The isolates, which shared a phage type similar to DT135 known locally as 135@ or 135a, formed a clade within the S. Typhimurium population with close similarity to the reference genome SL1334 (160 single nucleotide polymorphisms, or SNPs). Ten of the isolates belonged to a single clone (<23 SNPs between isolate pairs) which likely represents the population of S. Typhimurium circulating at the chicken farm; the other two were from sporadic cases and were genetically distinct from this clone. Divergence dating indicated that all 12 isolates diverged from a common ancestor in the mid 1990 s, and the clone began to diversify in 2003-2004. This clone spilled out into the human population several times between 2005-2008, during which time it continued to accumulate SNPs at a constant rate of 3-5 SNPs per year or 1x10-6 substitutions site-1 year-1, faster than the longer-term (~50 year) rates estimated previously for S. Typhimurium. Our data suggest that roughly half of non-synonymous substitutions are rapidly removed from the S. Typhimurium population, after which purifying selection is no longer important and the remaining substitutions become fixed in the population. The S. Typhimurium 135@ isolates were nearly identical to SL1344 in terms of gene content and virulence plasmids. Their phage contents were close to SL1344, except that they carried a different variant of Gifsy-1, lacked the P2 remnant found in SL1344 and carried a novel P2 phage, P2-Hawk, in place SL1344's P2 phage SopEϕ. DT135 lacks P2 prophage. Two additional plasmids were identified in the S. Typhimurium 135@ isolates, pSTM2 and pSTM7. Both plasmids were IncI1, but phylogenetic analysis of the plasmids and their bacterial hosts shows these plasmids are genetically distinct and result from independent plasmid acquisition events.

Conclusions: This study provides a high-resolution insight into short-term microevolution of the important human pathogen S. Typhimurium. It indicates that purifying selection occurs rapidly in this population (≤ 6 years) and then declines, and provides an estimate for the short-term substitution rate. The latter is likely to be more relevant for foodborne outbreak investigation than previous estimates based on longer time scales.

Figure 1

Phylogenetic tree of S. Typhimurium genomes. Maximum likelihood tree for S. Typhimurium based on SNPs identified by mapping to the reference chromosome for S. Typhimurium SL1344 (isolate details in Tables 1 and 2), excluding those SNPs identified in phage or repeat regions. Inset shows a neighbour-joining split network of the same SNP data. Shaded region indicates S. Typhimurium 135@ isolates sequenced in this study, for which a high resolution tree is given in Figure 2; red branches indicate the farm clone. Bootstrap support is shown for those bipartitions with <100% bootstrap support; tree branches are labelled with the number of SNPs contributing to the branch.

Figure 2

Phylogeny and timeline of outbreak-related S. Typhimurium 135@ isolates. Maximum likelihood tree of the S. Typhimurium 135@ isolates, based on SNPs identified by mapping to the reference chromosome for S. Typhimurium SL1344. Bootstrap values for each bipartition were 100%. Isolates are labelled by their STm number and an icon indicating the specimen type (according to the legend provided); black numbers next to branches indicate the number SNPs contributing to that branch. Red bar graph indicates the timing and size of the seven outbreaks. Red branches in the tree indicate the farm clone; dashed lines link each isolate to its time of isolation and outbreak membership.

Figure 3

Phage content variation among S. Typhimurium 135@, SL1344, LT2, DT135 and DT12. BRIG diagram showing the STm1 chromosome as a reference; coloured rings indicate the coverage of STm1 sequences among contigs from other S. Typhimurium 135@ and the two reference chromosomes. The location of prophage sequences and phage remnants in the S. Typhimurium 135@ genome is indicated around the outside.

Figure 4

Variable P2-P4 phage region in S. Typhimurium 135@ compared to SL1344, DT135 and DT12. Orange arrows indicate genes encoded in the P2-Hawk prophage sequence from S. Typhimurium 135@; the two cargo genes are darkened. Purple arrows indicate genes encoded in the P2 SopEϕ phage from SL1344; the sopE effector gene is darkened. The P4 phage, conserved in SL1344, DT135 and S. Typhimurium 135@ is shown in blue. The integrase genes of each prophage are labelled (int).

Figure 5

Phylogenetic analysis of IncI1 plasmids identified in S. Typhimurium 135@. (A) Maximum likelihood phylogeny of 23 publicly available IncI1 plasmids (listed in Table 4) together with pSTM2 and pSTM7 identified in this study. Key plasmids from these 23 are labeled. The shaded region indicates the clade containing pSTM2 and pSTM7, shown in detail in (B). Bootstrap support for this clade is 88%. (B) Bootstrap values for each branch are 100%, except one node as indicated (99%).