Hotspot mutations and ColE1 plasmids contribute to the fitness of Salmonella Heidelberg in poultry litter

Abstract

Salmonella enterica subsp. enterica serovar Heidelberg (S. Heidelberg) is a clinically-important serovar linked to food-borne illness, and commonly isolated from poultry. Investigations of a large, multistate outbreak in the USA in 2013 identified poultry litter (PL) as an important extra-intestinal environment that may have selected for specific S. Heidelberg strains. Poultry litter is a mixture of bedding materials and chicken excreta that contains chicken gastrointestinal (GI) bacteria, undigested feed, feathers, and other materials of chicken origin. In this study, we performed a series of controlled laboratory experiments which assessed the microevolution of two S. Heidelberg strains (SH-2813 and SH-116) in PL previously used to raise 3 flocks of broiler chickens. The strains are closely related at the chromosome level, differing from the reference genome by 109 and 89 single nucleotide polymorphisms/InDels, respectively. Whole genome sequencing was performed on 86 isolates recovered after 0, 1, 7 and 14 days of microevolution in PL. Only strains carrying an IncX1 (37kb), 2 ColE1 (4 and 6kb) and 1 ColpVC (2kb) plasmids survived more than 7 days in PL. Competition experiments showed that carriage of these plasmids was associated with increased fitness. This increased fitness was associated with an increased copy number of IncX1 and ColE1 plasmids. Further, all Col plasmid-bearing strains had hotspot mutations in 37 loci on the chromosome and in 3 loci on the IncX1 plasmid. Additionally, we observed a decrease in susceptibility to tobramycin, kanamycin, gentamicin, neomycin and fosfomycin for Col plasmid-bearing strains. Our study demonstrates how positive selection from poultry litter can change the evolutionary path of S. Heidelberg.

Fig 1. Survival curve and chromosomal mutation based phylogeny of evolved S. Heidelberg strains.

(a) S. Heidelberg concentration from poultry litter microcosms reported in colony forming units (CFU) and normalized by litter dry weight. (b) Venn diagram showing the total number of chromosomal SNPs/InDels shared between isolates with hotspot mutations. (c) Left. Maximum-likelihood tree of SH-2813 and SH-116 isolates based on 213 SNPs/InDels. Clades and taxa highlighted in green denote isolates without hotspot mutations or Col plasmids, while clades highlighted with cyan boxes carry hotspot mutations and 3 Col plasmids. Taxa in parenthesis represent isolates with identical SNPs/InDels and not used in the reconstruction of the tree shown. Isolates with sample ID’s 1–24 are evolved isolates from strain SH-2813 and ID’s 25–48 are evolved isolates from strain SH-116. The numbers shown next to the branches represent the percentage of replicate trees where associated taxa cluster together based on ~200 bootstrap replicates. Numbers (1–18) denote arbitrarily assigned clade numbers. Right: Corresponding venn diagram comparing the total number of SNPs/InDels shared across time points. Tree was rooted with ancestral SH-2813.

Fig 2. Deletion of 22-bp iteron of IncX1 is associated with a reduction in rep expression.

(a) Annotated IncX1 plasmid map of SH-2813_anc and SH-116_anc. Regions that acquired mutations in evolved strains carrying Col plasmids are highlighted in rectangular dashed boxes (b) Multiple alignment of the IncX1 iteron-based direct repeat (DR) origin of selected strains from this study, and 22 S. Heidelberg genomes available on NCBI. (c) Differences in expression of IncX1 genes in evolved SH-2813 and SH-116 relative to SH-2813_anc and SH-116_anc (log₂ scale, N = 3, *p.value = 0.086 (Wilcoxon signed-rank test). The error bars represent the standard error of the mean.

Fig 3. Correlation between IncX1 and ColE1-6kb plasmid copies.

Plasmid copy number (PCN) was determined from whole genome sequencing de-novo assembly coverage (see S6 Table).

Fig 4

Linear maps of Col plasmids and phylogeny based on conserved mobilization genes. (a) Annotated maps of Col plasmids carried by ancestral and evolved isolates of S. Heidelberg used for competition experiments. (b and c) Maximum-likelihood based tree of aligned conserved protein sequences of (b) relaxase (MbeA) and (c) MbeC genes of Col plasmid-carrying isolates. Plasmids with identical MbeA (n = 106) and MbeC (n = 112) sequences were not included in the reconstructed tree. The numbers shown next to the branches represent the percentage of replicate trees where associated taxa cluster together based on 1000 bootstrap replicates. Taxa in parenthesis represent selected isolates used for competition experiments. Trees were rooted with MbeA and MbeC sequences of E. coli (NCBI accession number: NC_001371).

Fig 5. Fitness and conjugation rates of evolved S. Heidelberg isolates.

(a) Fitness of Col plasmid-carrying strains of SH-2813 and SH-116 relative to a Col plasmid-free ancestor of SH-2813_nal. Bars represent means of four technical replicates for each population. The horizontal dashed red line represents the fitness of SH-2813_nal. (b) Conjugation rates of Col plasmids from Col plasmid-carrying strains to SH-2813_nal ancestor. Conjugation rates were estimated using data from fitness experiment. Bars represent means of four technical replicates for each population. The transfer rate for ColE1-4kb plasmid from SH-116_evol to SH-2813_nal was not determined. Symbols: Abortive phage infection protein of ColE1-6kb (Abi-ColE1-6kb), macrophage stimulating factor protein of ColE1-4kb (MSF-ColE1-4kb) and replication protein of ColpVC (rep-ColpVC).

Fig 6. Acquisition of ColE1 plasmids was associated with decreased susceptibility to aminoglycosides and fosfomycin.

(a) Multiple alignment of conserved N-terminal residues for aminoglycoside acetyltransferase (AAC-6’) genes encoded in the S. Heidelberg (SH-2813) chromosome and ColE1-6kb plasmid from this study, and AAC-6’ genes with known resistance phenotype. (b) Corresponding neighbor-joining tree of aligned AAC-6’ protein sequences. (c) Inhibition zone diameter (mm) distributions for evolved strains and selected antimicrobial drugs. SH-2813 (n = 27) and SH-116 (n = 18) isolates tested were recovered after 0 and 14 days of microevolution in PL. A decrease in inhibition zone corresponds to a decrease in susceptibility to selected antibiotic.

Fig 7. Col plasmids present in poultry litter.

(a) Concentration of plasmid-specific genes and 16S rDNA [48] in poultry litter of different origins. PL was collected as grab samples from multiple locations of various poultry houses/farms in USA and composited in 1-gallon Whirl-Pak bags. qPCR was performed on DNA extracted from 250 mg of PL in triplicates and pooled. Error bars represent the standard error of the means. Symbols: Conventional farming (Con), Backyard farming (Back), replication protein of IncX1 (IncX1-rep), replication protein of ColpVC (ColpVC-rep), abortive phage infection protein of ColE1-6kb (ColE1-6kb-Abi), macrophage stimulating factor protein of ColE1-4kb (ColE1-4kb-MSF) and 16S ribosomal DNA of Enterobacteriaceae (16S rDNA) [48]. (b) Concentration of plasmid-specific genes in PL used for microevolution experiment. DNA was extracted from 48 microcosms and qPCR was performed using primers targeting the incRNAI region of ColE1-6kb and ColE1-4kb, the replication protein of ColpVC and the 16S ribosomal DNA of Enterobacteriaceae [48]. Note. One uninoculated microcosm was included at day 0. ColE1 was not detected in SH-2813-BHI and SH-116-BHI microcosms. *p.value = 0.086 (Wilcoxon signed-rank test). The error bars represent the standard error of the mean.