Keywords:Salmonella, host adaptation, pseudogene, metabolism

ClassicalS. Enteritidis Is Not the Ancestor of theS. Gallinarum Biovars.

To establish an accurate phylogeny, we selected isolates ofS. Enteritidis,S. Gallinarum,S. Pullorum, andS. Dublin representing different MLSTs, phage types (PTs), and geographical and temporal spread to ensure that we sampled a broad diversity in this lineage. In total, genomic DNA from 59 isolates was used to generate whole genomes by Illumina multiplex sequencing (Table S1). Single-nucleotide polymorphisms (SNPs) were detected by mapping sequences against a pan-chromosome and virulence plasmid pseudomolecule (SI Methods). The maximum-likelihood phylogenetic tree shown inFig. 1 was constructed based upon the core regions of the referenceS. Enteritidis strain P125109 genome.

The phylogeny of these serotypes is largely consistent with previous studies showing thatS. Dublin represents an independent clade (9,11). Our data show that there are ∼23,000 SNPs differentiatingS. Dublin from the most recent common ancestor ofS. Enteritidis and the avian-adapted salmonellae,S. Gallinarum andS. Pullorum. However, rather than being a single clade itself,S. Enteritidis appears to be more structurally complex: isolate 01-00493-2 occupies a position on the tree basal to all otherS. Enteritidis andS. Gallinarum/S. Pullorum complex isolates, and a second clade ofS. Enteritidis appears within theS. Gallinarum/S. Pullorum complex (Fig. 1).

Statistical clustering supports the two clades ofS. Enteritidis depicted inFig. 1; a Bayesian analysis of population structure (BAPS) was used to identify significant differences in polymorphism sharing across strains (12). This initially indicated that members of clade 2 clustered separately from clade 1. A subsequent level of clustering then separated theS. Enteritidis 01-00493-2 isolate labeled “ancestral” inFig. 1 from the rest of clade 2 (Dataset S1). No further substructure inS. Enteritidis was found to be significant according to the Bayesian statistical model.

We refer to members of the first clade as “classic”S. Enteritidis because this clade includes the most commonly isolated MLST, ST11, and PTs affecting humans and animals. It is of note that both PT11 isolates (99-02302 and 00-03508) occupy a deep branch within the classic clade. PT11s are unusual in that they are rare in human infections and have been reported to be associated with hedgehogs (Erinaceus europaeus) (13).

The second clade (Fig. 1, second clade) includes isolates SARB18, 118-2012K-186, 105-2011K-1654, andD24359, which share the same O and H antigen designations as classicS. Enteritidis and are closely related at the sequence level. It is this second clade ofS. Enteritidis from which the ancestors of theS. Gallinarum/S. Pullorum complex emerged. This phylogeny suggests thatS. Gallinarum andS. Pullorum did not directly evolve from within the most commonly isolated (classic) clade ofS. Enteritidis, as was previously postulated (8), and that there is much more substructure in the phylogeny of these important human and animal pathogens than was previously known.

Chromosomal Macroevolution.

Evolution of pathogenicity bySalmonella is strongly associated with the acquisition of mobile genetic elements calledSalmonella pathogenicity islands (SPIs). Many of these SPIs were acquired very early in the evolution ofS.enterica (14), and so, perhaps unsurprisingly, their complement was found to be conserved across this entire lineage, with the exception of SPI-6 and SPI-19. Among other functions, these SPIs both encode type VI secretion systems (T6SSs), which are known to promote colonization in the avian gut (15). As previously reported forS. Enteritidis PT4 P125109, the majority of the classicS. Enteritidis isolates sequenced here harbor degenerate versions of both these SPIs. In these isolates, little more than thesaf fimbrial operon is maintained on SPI-6, whereas on SPI-19 only 13 out of 33 genes, mainly encoding exported proteins, remain intact. The borders of these deletions are the same for all isolates. We also observed that two isolates of the classicS. Enteritidis clade, both PT11, carry intact versions of both SPIs. This indicates that the loss of SPI-6 and SPI-19 is restricted and suggests that it occurred following the divergence of the main classicS. Enteritidis group from these strains. That the common ancestor of the entire lineage contained both loci is also supported by their (intact) presence inS. Dublin.

SPI-19 is intact in all other isolates sequenced in this study (Fig. 1), and forS. Gallinarum andS. Pullorum, the retention of a functional SPI-19 is consistent with the T6SS it encodes being important forS. Gallinarum survival in chicken macrophages (16). However, the degradation of SPI-6 and loss of the T6SS, seen within the second clade ofS. Enteritidis and theS. Gallinarum/S. Pullorum complex, may have only occurred once. Examining the recombination events across the entire lineage revealed that the region surrounding SPI-6 appears to have been independently donated from theS. Gallinarum lineage toS. Pullorum and isolates in the second clade ofS. Enteritidis (Fig. 1, SPI-6 version ii;Dataset S1).

Within the chromosome, other clade-specific whole gene losses were related to fimbriae. InSalmonella, the loss or inactivation of fimbriae is associated with host adaptation (8,17).S. Enteritidis harbors 13 fimbrial operons, and the variable presence of 6 of these occurred in patterns that matched the phylogeny of isolates in this lineage (Fig. 1). The majority of these differences result from partial- or whole-gene losses in host-restrictedS. Gallinarum andS. Pullorum, but partial losses of thessf andsaf operons were also observed in the second clade ofS. Enteritidis.

Extrachromosomal Macroevolution.

Only four isolates in this study do not carry a version of theSalmonella virulence plasmid (pVIR) encoding the virulence-associatedspv operon (Fig. 1). Using the alignment produced by mapping to a pseudomolecule representing the pan-chromosome combined with the pVIR, we extracted the subset of sequences aligning to the pVIR ofS. Gallinarum 287-91, which is the largest and most complete virulence plasmid in this lineage. The SNPs identified in this alignment were used to construct a pVIR phylogenetic tree (Fig. S1). The plasmid phylogeny mirrors that of the chromosome for all isolates except for the classic Enteritidis clade, which appears to harbor a pVIR more distantly related to the plasmids carried within this lineage (∼1,000 SNPs) than the outgroup, pSLT (∼300 SNPs).

The smaller size of the classicS. Enteritidis pVIR (∼60 kb) relative to that ofS. Gallinarum (∼90 kb) can be largely attributed to major deletions in thetra region responsible for conjugal transfer (18). Other differences that contribute to the phylogenetic distance between pVIRs include the presence of genes for a partial K88-like fimbria (19), which was found in all of theS. Dublin,S. Gallinarum, andS. Pullorum pVIRs. ApefA-like remnant upstream of SG_p0053 was also conserved across these plasmids, which we believe to be a scar of thepef fimbrial operon. This remnant has not been annotated in any published K88-like–containing pVIRs to date but has been randomly mutated inS. Gallinarum by transposon mutagenesis. In contrast to the parentalS. Gallinarum, no morbidity or mortality was observed in chickens inoculated with the resultant mutants (19). This suggests that thepefA remnant retains some functionality, although it is not clear what interaction this may have with the K88-like replacement. Thepef fimbrial operon itself is present only in pVIR-carrying strains ofS. Enteritidis from the classic clade. Alongside the pVIR phylogeny, this suggests that the ancestral pVIR plasmid of the entire lineage was replaced by a different plasmid before the expansion of the classicS. Enteritidis clade. In further support of this, we observe that the second clade and ancestralS. Enteritidis all have a pVIR more closely related to those carried byS. Gallinarum andS. Pullorum than to those found in the classicS. Enteritidis clade.

Pseudogene Formation Largely Occurred After Serovar Diversification.

Manual, whole-genome comparisons were used to define the extent to which gene degradation has occurred along the branches of the chromosomal phylogenetic tree. In this context, we defined a pseudogene as a gene harboring a mutation (i.e., premature stop codon, frame shift, truncation, or syntenic deletion) relative to an intact version of that gene.Table 1 summarizes the number of pseudogenes present in each of the respective serovars, and how many shared pseudogenes are due to identical or nonidentical mutations. Full pseudogene sets are listed inDataset S2. In total, 98 ancestral pseudogenes are common to the entire set of serovars. These are typically fragments of nonfunctional genes and are likely due to stochastic loss. However, pseudogenes that are found in particular regions of the tree are more likely to have resulted from a primary event interrupting metabolic or other pathways, potentially leading to further degradation of related genes.

S. Enteritidis is generally regarded as a promiscuous serovar, and our analysis found only one nonancestral pseudogene shared by all classic isolates (Fig. 1). When considered alongside the maintenance of intact fimbriae, replacement of the ancestral plasmid, and loss of both T6SSs, it is apparent that these are markers of adaptation to host generalism in classicS. Enteritidis.Fig. S2 depicts the differing routes toward host generalism and adaptation seen across the entire lineage, from a hypothetical shared ancestor.

Previous studies have indicated that the degree of host specificity displayed by particularSalmonella serovars loosely correlates with the level of gene degradation (8,17,20,21). In keeping with this observation, host-restrictedS. Gallinarum andS. Pullorum harbor 231 and 212 pseudogenes, respectively, whereasS. Dublin, which is associated with but not restricted to cattle, harbors 95 (Table 1).

Significantly, given their shared host restriction, analysis of pseudogene positioning across the tree indicated that over 60% of the genome degradation found inS. Gallinarum (151 of 231) andS. Pullorum (132 of 212) occurred after the two diverged. Our recombination analysis revealed that there has been one large recombination event between them, representing a region of around 180 kb (Dataset S1), but this accounts for only six shared pseudogenes, indicating that recombination has not been a significant cause of shared gene loss. However, this region encompasses 165 genes, including four fimbrial operons (saf,sti, stf, andstb) and SPI-6, suggesting that recombination has played some role in these two pathogens converging on the same host niche.

Loss of Metabolic Capacity.

Functionally, the 98 identified ancestral pseudogenes consist mainly of phage, insertion sequences, and genes of unknown function. This is in stark contrast to the functional categories represented by pseudogenes found toward the tips of the tree (Figs. 1 and2A), where membrane/surface structure and central/intermediary metabolism genes are more commonly inactivated. This latter pattern is broadly consistent forS. Gallinarum andS. Pullorum, both individually and in their shared 78 pseudogenes, and also in theS. Dublin pseudogenes, suggesting that loss of related functions is associated with adaptation to a host organism.

We established that ∼15–20% of pseudogenes from each serovar had functional locations according to a database of metabolic pathways predicted to be present inS. Enteritidis (Dataset S3). This allowed us to establish whether any metabolic functions were commonly affected (i.e., by the same or different pseudogenes in the same pathway) between serovars. The overlap of the affected pathways is shown inFig. 2B and indicates that many specific pathways or transport reactions are affected in more than one of the host-adapted serovars in our set, and indeed in other host-adapted salmonellae (see below).

One sodium/galactoside transporter has been lost in all of the host-adapted serovars within our set, but degradation of the carbon sourced-glucarate is the only full pathway that has been affected (by different mutations) in them all. Interestingly, the human-restricted serovarsS. Typhi andS. Paratyphi A also harbor pseudogenes related to the transport ofd-glucarate into the cell, suggesting that the utilization of this carbon source may not be advantageous in causing invasive infection. Two isoenzymes catalyze a key reaction in this pathway, one of which is mutated inS. Dublin, whereas both are inactivated inS. Gallinarum.

The differing impact of these pseudogenes was supported by a phenotyping screen (Dataset S4), becauseS. Dublin has only suffered a loss of redundancy and therefore remains capable of usingd-glucarate, whereasS. Gallinarum has lost the function entirely.

There are 13 metabolic pathways and two transport reactions commonly affected inS. Gallinarum andS. Pullorum (Dataset S3). The majority of these (12 of 15) are due to identical pseudogenes that occurred before their divergence. One of these is in the biosynthesis of the siderophore enterobactin, used to scavenge iron from the environment (22). Iron acquisition in the host is key to many bacterial infections (23), and other iron uptake and transport systems remain intact. This is in contrast to previous findings that human-restrictedS. Typhi is dependent upon iron uptake through enterobactin via thefep genes (24) and therefore suggests that enterobactin is not required in the avian host. However, another of the pathways, putrescine biosynthesis, is also affected inS. Typhi, and the same enzyme, ornithine decarboxylase, is mutated in all three. Because the alternative putrescine pathway remains intact, loss of redundancy may have an important role alongside loss of function in host adaptation. This is further supported by the phenotyping screen performed inS. Dublin. In accordance with the lower absolute number of pseudogenes,S. Dublin has only three pathways affected, and four transport reactions (Dataset S3). Of the four relevant substrates present in the phenotyping screen,S. Dublin cannot use one, but remains capable of using three others, indicating that reduction in redundancy is more common than loss of function in this host-adapted serovar.

Three metabolic pathways are affected inS. Gallinarum andS. Pullorum due to different pseudogenes that occurred after the two diverged. Two of these functions, allantoin degradation and adenosylcobalamin (vitamin B12) biosynthesis, are also mutated inS. Typhi andS. Paratyphi A. Given the pseudogene accumulation in these pathways in host-restricted serovars, they emerge as strong contenders for markers of a switch to invasive rather than enteric disease.

In birds, allantoin can be found in the serum and is used as a carbon source duringS. Enteritidis infection of chickens (25). Inactivation of the genes encoding the regulatorallS and the degradative enzymeallD inS. Gallinarum andS. Pullorum, respectively, means that neither can use allantoin. Although the relevance of allantoin in mammalian hosts is unknown, pseudogenes relating to allantoin utilization have also been identified in a strain ofS. Typhimurium belonging to ST313, which causes invasive nontyphoidal salmonellosis in humans (26).

Vitamin B12 is required for the anaerobic degradation of 1,2-propanediol, which uses tetrathionate as an electron acceptor (27). Tetrathionate plays an important role in enteric infection: its production is triggered by the inflammatory response toS. Typhimurium in the gut and provides a niche for respiration, in competition with other gut microbes (28). A recent report by Nuccio and Bäumler (21) identified loss of function in central anaerobic metabolism as a key indicator of invasive versus gastrointestinal salmonellae. From plotting pseudogenes onto the chromosomal phylogeny, we know thatS. Gallinarum andS. Pullorum lost the ability to reduce tetrathionate before their divergence, and that their subsequent loss of vitamin B12 synthetic ability occurred independently. The knock-on effect of those initial pseudogenes in tetrathionate reduction is illustrated inFig. 2C, as further pseudogenes have also arisen in other related functions, consistent with Nuccio and Bäumler’s findings.

Second Clade and AncestralS. Enteritidis Display Intermediate Characteristics.

Of the fiveS. Enteritidis isolates outside the classic clade, one (ancestral) was basal to allS. Enteritidis andS. Gallinarum, whereas the remainder (second clade) were basal to bothS. Gallinarum andS. Pullorum (Fig. 1). To assess whether their position on the phylogenetic tree had any phenotypic consequences, we initially looked at colony morphology and motility of the ancestral isolate (01-00493-2) and a representative isolate from the second clade (SARB18). These indicated that both have intermediate characters betweenS. Enteritidis andS. Gallinarum (Fig. S3). One explanation for this could be recombination; multiple separate events have recombined ∼175 kb fromS. Gallinarum orS. Pullorum into the second clade isolates, but not into any from the classicS. Enteritidis clade (Dataset S1). Approximately 160 kb originating fromS. Gallinarum recombined into all four of the second clade isolates, around the SPI-6 locus, resulting in the presence of a different degenerate version of this pathogenicity island compared with that carried by the classic clade isolates. A separate event was responsible for the recombination of ∼85 kb into the ancestral isolate 01-00493-2 fromS. Gallinarum, also including SPI-6. Unlike classicS. Enteritidis, SPI-19 remains intact in the ancestral and second clade isolates ofS. Enteritidis, as it does inS. Gallinarum andS. Pullorum (Fig. 1).

Our pseudogene analysis was dependent upon a single reference strain per serovar, and therefore new pseudogenes could not be confirmed in theS. Enteritidis isolates outside the classic clade. However, metabolic phenotyping evidence suggests numerous differences are present between the SARB18 second clade and 01-004-93-2 ancestral isolate, with each able to use six or eight different carbon sources, respectively, that could not be used by the other isolate (Dataset S4).

Given the genotypic and phenotypic differences observed, we tested the pathogenesis of these two isolates in comparison with the reference strains ofS. Enteritidis andS. Gallinarum in the natural avian host. TheS. Gallinarum-infected group showed signs of systemic salmonellosis and reached the humane endpoint of this experimental protocol at 5 d postinfection. The remaining groups showed no signs of ill health and the experiment continued for the full 7 d. At postmortem, theS. Gallinarum-infected birds displayed considerable hepatosplenomegaly, white spot lesions on the spleen and discoloration of the liver accompanied by “bronzing” on exposure to air, consistent with fowl typhoid. In contrast, the groups infected withS. Enteritidis strains showed mild hepatosplenomegaly consistent with infection by thisSalmonella serovar. Also as expected,S. Gallinarum was the most invasive, indicated by significantly higher colony counts in both the spleen and liver on day 5 (P < 0.01) (Fig. 3A andB). However, the classicS. Enteritidis strain P125109 showed significantly higher levels of colonization (bacteria present in the caeca) compared with both the nonclassicS. Enteritidis strains andS. Gallinarum (P < 0.05) (Fig. 3C). This was also borne out by histopathological scoring of the tissue (Table S2). This reduced colonization phenotype from the ancestral and second cladeS. Enteritidis isolates further suggests that they represent intermediate stages in the evolution of host adaptation.

In conclusion, whole-genome comparisons across an entireSalmonella lineage have enabled us to establish the progression of pseudogene formation that has resulted in differently host-adapted pathogens. The strongest signal of metabolic loss was seen inS. Gallinarum andS. Pullorum, the fully host-restricted members of the lineage, consistent with patterns seen in other host-restricted salmonellae. It is therefore plausible that pseudogene formation progresses in a predictable fashion, given a particular host. We also observed the presence ofS. Enteritidis isolates that fall outside the classic host generalist clade, instead occupying positions basal to the avian-restricted strains. Their metabolic, genomic, and infection phenotypes all suggest that these represent intermediate, but extant, stages in the process of pathogen–host adaptation, demonstrating an unexpected diversity in this serovar.

Sequencing and Phylogenetic Analysis.

All isolates were sequenced on the Illumina platform, with additional 454 sequencing where required. A nonredundant pseudomolecule was used a reference for mapping and to produce genome alignments. The genetic structure of the population was estimated with the software BAPS, version 6.0 (12,29), and recombination detected using the BratNextGen method (30). Pseudogene identification was performed by manual genome comparison, and metabolic pathway reconstruction was based upon the annotated genome ofS. Enteritidis, using Pathway Tools software (SRI International). High-throughput phenotype screening was performed using the Biolog system (Biolog). Details are described inSI Methods.

Chick Infection Models.

All experiments were conducted in accordance with United Kingdom legislation governing experimental animals under project licenses PPL 40/3063 and PPL 40/3652 and were approved by the University of Liverpool ethical review process before the award of the license. Four or five chicks were inoculated per experimental group and were killed at 5 or 7 d postinfection. Colony-forming units were quantified from the liver, spleen, and cecal contents, and pathology scoring on hematoxylin-and-eosin–stained sections was performed. Full details are described inSI Methods.

• 1.Parkhill J, et al. Comparative analysis of the genome sequences of Bordetella pertussis, Bordetella parapertussis and Bordetella bronchiseptica. Nat Genet. 2003;35(1):32–40. doi: 10.1038/ng1227. [DOI] [PubMed] [Google Scholar]
• 2.Rohmer L, et al. Comparison of Francisella tularensis genomes reveals evolutionary events associated with the emergence of human pathogenic strains. Genome Biol. 2007;8(6):R102. doi: 10.1186/gb-2007-8-6-r102. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 3.Cole ST, et al. Massive gene decay in the leprosy bacillus. Nature. 2001;409(6823):1007–1011. doi: 10.1038/35059006. [DOI] [PubMed] [Google Scholar]
• 4.Moran NA, Plague GR. Genomic changes following host restriction in bacteria. Curr Opin Genet Dev. 2004;14(6):627–633. doi: 10.1016/j.gde.2004.09.003. [DOI] [PubMed] [Google Scholar]
• 5.Majowicz SE, et al. International Collaboration on Enteric Disease “Burden of Illness” Studies The global burden of nontyphoidal Salmonella gastroenteritis. Clin Infect Dis. 2010;50(6):882–889. doi: 10.1086/650733. [DOI] [PubMed] [Google Scholar]
• 6.Boyd EF, et al. Salmonella reference collection B (SARB): Strains of 37 serovars of subspecies I. J Gen Microbiol. 1993;139(Pt 6):1125–1132. doi: 10.1099/00221287-139-6-1125. [DOI] [PubMed] [Google Scholar]
• 7.Porwollik S, et al. Differences in gene content between Salmonella enterica serovar Enteritidis isolates and comparison to closely related serovars Gallinarum and Dublin. J Bacteriol. 2005;187(18):6545–6555. doi: 10.1128/JB.187.18.6545-6555.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 8.Thomson NR, et al. Comparative genome analysis of Salmonella Enteritidis PT4 and Salmonella Gallinarum 287/91 provides insights into evolutionary and host adaptation pathways. Genome Res. 2008;18(10):1624–1637. doi: 10.1101/gr.077404.108. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 9.Achtman M, et al. S. Enterica MLST Study Group Multilocus sequence typing as a replacement for serotyping in Salmonella enterica. PLoS Pathog. 2012;8(6):e1002776. doi: 10.1371/journal.ppat.1002776. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 10.Selander RK, et al. Molecular evolutionary genetics of the cattle-adapted serovar Salmonella dublin. J Bacteriol. 1992;174(11):3587–3592. doi: 10.1128/jb.174.11.3587-3592.1992. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 11.Olsen JE, Skov M. Genomic lineage of Salmonella enterica serovar Dublin. Vet Microbiol. 1994;40(3-4):271–282. doi: 10.1016/0378-1135(94)90116-3. [DOI] [PubMed] [Google Scholar]
• 12.Corander J, Marttinen P, Sirén J, Tang J. Enhanced Bayesian modelling in BAPS software for learning genetic structures of populations. BMC Bioinformatics. 2008;9(1):539. doi: 10.1186/1471-2105-9-539. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 13.Nauerby B, Pedersen K, Dietz HH, Madsen M. Comparison of Danish isolates of Salmonella enterica serovar enteritidis PT9a and PT11 from hedgehogs (Erinaceus europaeus) and humans by plasmid profiling and pulsed-field gel electrophoresis. J Clin Microbiol. 2000;38(10):3631–3635. doi: 10.1128/jcm.38.10.3631-3635.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 14.Fookes M, et al. Salmonella bongori provides insights into the evolution of the Salmonellae. PLoS Pathog. 2011;7(8):e1002191. doi: 10.1371/journal.ppat.1002191. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 15.Blondel CJ, et al. Contribution of the type VI secretion system encoded in SPI-19 to chicken colonization by Salmonella enterica serotypes Gallinarum and Enteritidis. PLoS One. 2010;5(7):e11724. doi: 10.1371/journal.pone.0011724. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 16.Blondel CJ, et al. The type VI secretion system encoded in Salmonella pathogenicity island 19 is required for Salmonella enterica serotype Gallinarum survival within infected macrophages. Infect Immun. 2013;81(4):1207–1220. doi: 10.1128/IAI.01165-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 17.Parkhill J, et al. Complete genome sequence of a multiple drug resistant Salmonella enterica serovar Typhi CT18. Nature. 2001;413(6858):848–852. doi: 10.1038/35101607. [DOI] [PubMed] [Google Scholar]
• 18.Rodríguez-Peña JM, Buisan M, Ibáñez M, Rotger R. Genetic map of the virulence plasmid of Salmonella enteritidis and nucleotide sequence of its replicons. Gene. 1997;188(1):53–61. doi: 10.1016/s0378-1119(96)00776-7. [DOI] [PubMed] [Google Scholar]
• 19.Rychlik I, Lovell MA, Barrow PA. The presence of genes homologous to the K88 genes faeH and faeI on the virulence plasmid of Salmonella gallinarum. FEMS Microbiol Lett. 1998;159(2):255–260. doi: 10.1111/j.1574-6968.1998.tb12869.x. [DOI] [PubMed] [Google Scholar]
• 20.Chiu C-H, et al. The genome sequence of Salmonella enterica serovar Choleraesuis, a highly invasive and resistant zoonotic pathogen. Nucleic Acids Res. 2005;33(5):1690–1698. doi: 10.1093/nar/gki297. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 21.Nuccio S-P, Bäumler AJ. Comparative analysis of Salmonella genomes identifies a metabolic network for escalating growth in the inflamed gut. MBio. 2014;5(2):e00929-14. doi: 10.1128/mBio.00929-14. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 22.Pollack JR, Ames BN, Neilands JB. Iron transport in Salmonella typhimurium: Mutants blocked in the biosynthesis of enterobactin. J Bacteriol. 1970;104(2):635–639. doi: 10.1128/jb.104.2.635-639.1970. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 23.Skaar EP. The battle for iron between bacterial pathogens and their vertebrate hosts. PLoS Pathog. 2010;6(8):e1000949. doi: 10.1371/journal.ppat.1000949. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 24.Barquist L, et al. A comparison of dense transposon insertion libraries in the Salmonella serovars Typhi and Typhimurium. Nucleic Acids Res. 2013;41(8):4549–4564. doi: 10.1093/nar/gkt148. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 25.Dhawi AA, et al. Adaptation to the chicken intestine in Salmonella Enteritidis PT4 studied by transcriptional analysis. Vet Microbiol. 2011;153(1-2):198–204. doi: 10.1016/j.vetmic.2011.07.013. [DOI] [PubMed] [Google Scholar]
• 26.Kingsley RA, et al. Epidemic multiple drug resistant Salmonella Typhimurium causing invasive disease in sub-Saharan Africa have a distinct genotype. Genome Res. 2009;19(12):2279–2287. doi: 10.1101/gr.091017.109. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 27.Price-Carter M, Tingey J, Bobik TA, Roth JR. The alternative electron acceptor tetrathionate supports B12-dependent anaerobic growth of Salmonella enterica serovar typhimurium on ethanolamine or 1,2-propanediol. J Bacteriol. 2001;183(8):2463–2475. doi: 10.1128/JB.183.8.2463-2475.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 28.Winter SE, et al. Gut inflammation provides a respiratory electron acceptor for Salmonella. Nature. 2010;467(7314):426–429. doi: 10.1038/nature09415. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 29.Tang J, Hanage WP, Fraser C, Corander J. Identifying currents in the gene pool for bacterial populations using an integrative approach. PLOS Comput Biol. 2009;5(8):e1000455. doi: 10.1371/journal.pcbi.1000455. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 30.Marttinen P, et al. Detection of recombination events in bacterial genomes from large population samples. Nucleic Acids Res. 2012;40(1):e6. doi: 10.1093/nar/gkr928. [DOI] [PMC free article] [PubMed] [Google Scholar]

Supplementary Materials