Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Identification and Characterization of Novel Salmonella Mobile Elements Involved in the Dissemination of Genes Linked to Virulence and Transmission

  • Andrea I. Moreno Switt,

    Affiliation Department of Food Science, Cornell University, Ithaca, New York, United States of America

  • Henk C. den Bakker,

    Affiliation Department of Food Science, Cornell University, Ithaca, New York, United States of America

  • Craig A. Cummings,

    Affiliation Life Technologies Corporation, Foster City, California, United States of America

  • Lorraine D. Rodriguez-Rivera,

    Affiliation Department of Food Science, Cornell University, Ithaca, New York, United States of America

  • Gregory Govoni,

    Current address: AvidBiotics Corporation, South San Francisco, California, United States of America

    Affiliation Life Technologies Corporation, Foster City, California, United States of America

  • Matthew L. Raneiri,

    Affiliation Department of Food Science, Cornell University, Ithaca, New York, United States of America

  • Lovorka Degoricija,

    Affiliation Life Technologies Corporation, Foster City, California, United States of America

  • Stephanie Brown,

    Affiliation Department of Food Science, Cornell University, Ithaca, New York, United States of America

  • Karin Hoelzer,

    Affiliation Department of Food Science, Cornell University, Ithaca, New York, United States of America

  • Joseph E. Peters,

    Affiliation Department of Microbiology, Cornell University, Ithaca, New York, United States of America

  • Elena Bolchacova,

    Affiliation Life Technologies Corporation, Foster City, California, United States of America

  • Manohar R. Furtado,

    Affiliation Life Technologies Corporation, Foster City, California, United States of America

  • Martin Wiedmann

    Affiliation Department of Food Science, Cornell University, Ithaca, New York, United States of America


The genetic diversity represented by >2,500 different Salmonella serovars provides a yet largely uncharacterized reservoir of mobile elements that can contribute to the frequent emergence of new pathogenic strains of this important zoonotic pathogen. Currently, our understanding of Salmonella mobile elements is skewed by the fact that most studies have focused on highly virulent or common serovars. To gain a more global picture of mobile elements in Salmonella, we used prediction algorithms to screen for mobile elements in 16 sequenced Salmonella genomes representing serovars for which no prior genome scale mobile element data were available. From these results, selected mobile elements underwent further analyses in the form of validation studies, comparative analyses, and PCR-based population screens. Through this analysis we identified a novel plasmid that has two cointegrated replicons (IncI1-IncFIB); this plasmid type was found in four genomes representing different Salmonella serovars and contained a virulence gene array that had not been previously identified. A Salmonella Montevideo isolate contained an IncHI and an IncN2 plasmid, which both encoded antimicrobial resistance genes. We also identified two novel genomic islands (SGI2 and SGI3), and 42 prophages with mosaic architecture, seven of them harboring known virulence genes. Finally, we identified a novel integrative conjugative element (ICE) encoding a type IVb pilus operon in three non-typhoidal Salmonella serovars. Our analyses not only identified a considerable number of mobile elements that have not been previously reported in Salmonella, but also found evidence that these elements facilitate transfer of genes that were previously thought to be limited in their distribution among Salmonella serovars. The abundance of mobile elements encoding pathogenic properties may facilitate the emergence of strains with novel combinations of pathogenic traits.


Salmonella is a widely distributed foodborne pathogen and one of the most common causes of bacterial foodborne illnesses and deaths globally [1]. In the United States, Salmonella causes approximately 11% of foodborne illnesses, and is the principal cause of hospitalizations and deaths due to foodborne diseases [1], [2]. The genus Salmonella includes two species (S. enterica and S. bongori) and more than 2,500 different serovars [3]; within S. enterica, a total of six subspecies have been reported to date. In addition to species, subspecies and serovar classification, Salmonella can also be classified based on their ability to cause disease in different hosts. For example, Salmonella has been classified into host-restricted, host adapted, and unrestricted serovars according to the degree of host specificity, or into typhoidal and non-typhoidal serovars, according to clinical presentation of systemic disease in humans [4][6]. Typhoidal Salmonella serovars (i.e., Typhi, Paratyphi A, B, and C) have been characterized by a unique complement of virulence determinants, including the type IVb pilus operon [7].

Comparisons of the Salmonella pan-genome have shown that the sequenced strains share a conserved genomic backbone, and that the vast majority of the genomic variation can be assigned to specific genomic regions in the accessory genome [4]. Mobile elements are part of the accessory genome and have been associated with the emergence of strains with novel pathogenicity phenotypes in a number of foodborne pathogens, including Salmonella [8], [9] and E. coli [10]. The importance of mobile elements has recently been illustrated by the European E. coli O104:H4 outbreak in 2011; this strain appears to have become highly virulent by acquisition of a stx2-harboring prophage, as well as a plasmid encoding virulence and resistance determinants [10]. The importance of mobile elements in the evolutionary history of Salmonella can be illustrated by the acquisition and rearrangement of different Salmonella pathogenicity islands (SPIs) in S. enterica and S. bongori [11], [12]. In addition to SPIs, which encode genes with experimentally validated virulence functions, genomic islands have been associated with the emergence of endemic strains, for instance, the multidrug resistant phage type S. Typhimurium DT104 [13]. Genome analysis of a strain within this phage type has allowed identification of Salmonella genomic island 1 (SGI1), which has subsequently been recognized as a globally distributed, integrative mobilizable element containing an array of antimicrobial resistance genes and present in multiple Salmonella serovars [8], [13][16]. Plasmids have also been shown to play an important role in dispersal and acquisition of virulence and antimicrobial resistance genes in Salmonella [17][19]. The distribution of plasmids is restricted by their replicons; this forms the basis of classifying plasmids into multiple incompatibility types (Inc types) [20], [21]. Previously reported plasmids in Salmonella include IncFIB serovar-specific virulence plasmids and IncA/C conjugative plasmids conferring resistance to multiple antimicrobials [17], [18], [22], [23]. Prophages are common in Salmonella and also play important roles in the evolution of this pathogen. For example, S. Typhimurium and S. Typhi prophages can encode genetic traits that increase pathogenicity (e.g., genes encoding SopE, SodC-1, SspH1, and SseI) or fitness in certain hosts (e.g., genes that lead to O-antigen conversion) [24][27].

While different mobile elements (e.g., plasmids, phages, transposons, and mobilizable islands) are clearly important for the evolution of Salmonella, including the emergence of strains with novel antimicrobial resistance and pathogenicity-associated phenotypes, our current understanding of mobile element distribution and diversity is still limited. Most descriptions of Salmonella mobile elements to date have focused on common or highly virulent serovars (e.g., serovars Typhimurium, Typhi) or strains with multidrug resistance phenotypes. To improve our understanding of mobile element diversity in Salmonella, we searched for mobile elements in the genome sequences of 16 different Salmonella serovars [5] for which no in-depth mobile element analysis has been performed to date. This study identified new mobile elements in Salmonella, including a novel IncI1-IncFIB cointegrated virulence plasmid, two novel genomic islands (SGI2 and SGI3), mosaic prophages carrying virulence genes, as well as an integrative conjugative element encoding the serovar Typhi type IVb pilus operon in three non-typhoidal serovars.


Characterization of mobile elements in 16 Salmonella serovars

Comprehensive mobile element analysis of the draft genome assemblies of 16 Salmonella enterica subsp. enterica serovars identified a large number of mobile elements including (i) two antimicrobial resistance plasmids (both found in the same serovar Montevideo strain) and four highly similar virulence plasmids found in four different serovars, (ii) two novel genomic islands (designated SGI2 and SGI3), (iii) three integrative conjugative elements (ICE), and (iv) 35 transposons. In addition, we identified 42 prophages that are each contained in a single sequence contig and thus could be described and characterized with a high level of confidence (Table 1).

Table 1. Summary of mobile elements identified in this study.

Identification of antimicrobial resistance gene-carrying plasmids and chromosomally integrated mobile elements that have not been previously reported in Salmonella

Plasmids encoding antimicrobial resistance were only identified in the serovar Montevideo strain FSL S5-403, the only antibiotic-resistant isolate in this set of strains (Table 1). The two plasmids found in this strain–designated pS5-403-1 and pS5-403-2–were predicted in silico and subsequently validated by PFGE, PCR, and sequencing. The antimicrobial resistance genes identified on these two plasmids (Fig. 1A and Table 2) were consistent with the resistance phenotype (i.e., resistance to aminoglycosides, tetracycline and sulfonamides). Plasmid pS5-403-1, a 53 kb IncN2 type plasmid, carries the resistance genes sul1, aacC, aadA, strAB, and qacEΔ1 in a class 1 integron (Fig. 1B). This integron shows similarity with a class 1 integron, which was previously reported as a component of pSN254, a plasmid found in multidrug resistant Salmonella Newport. Plasmid pS5-403-1 has the same backbone as the Escherichia coli plasmid p271A [28] . This backbone has typically been associated with IncN1 plasmids [28]; p271A and pS5-403-1 are the only described IncN2 plasmids with this backbone and, to our knowledge, pS5-403-1 is the first IncN2 plasmid reported for Salmonella. Whereas pS5-403-1 has a class 1 integron inserted in the accessory region, p271A has a transposon encoding the New Delhi metallo-β-lactamase, blaNDM-1, inserted in this region (Fig. 1A).

Figure 1. Circular representation of plasmid pS5-403-1 identified in Salmonella Montevideo.

(A) Circular representation of pS5-403-1. Circles from inside to outside: the inner circle represents the GC%, the second circle is pS5403-1, a 53 kb plasmid that encodes antimicrobial resistance genes, the outer circle represents the shared backbone with p271A plasmid in E. coli. The curly bracket represents a cluster of genes in a transposon insertion in p271A that includes the NDM-β-lactamase. (B) Cluster of genes representing a comparison of a class 1 integron in pS5-403-1 homologous to an integron inserted in pSN254 (plasmid in S. Newport). Genes were color coded according to function as follows: resistance (pink), plasmid transfer (blue), transposition/IS (orange), replication (green), plasmid stability (turquoise), metabolism (brown), and hypothetical proteins (grey).

The second plasmid, pS5-403-2 is a large (299 kb) conjugative plasmid of the IncHI replicon type. This plasmid encodes resistance to tetracycline, and predicted resistance to acriflavin, copper, silver, and cadmium. This plasmid shares the backbone region (including regions encoding Tra1 and Tra2 transfer functions and resistance to tellurite, silver, and copper) with previously described IncHI plasmids [29]. Using a guide tree created from an alignment made using the Mauve algorithm (used to cluster plasmids based on their overall identity, see methods) of pS5-403-2 and previously sequenced IncHI plasmids (Table S1), we found that pS5-403-2 is more similar to IncHI plasmids in E. coli (pAPCE-O1-R), Serratia marcescens (pR478), and Enterobacter cloacae (pEC-IMPQ), than to IncHI1 plasmids from Salmonella Typhi, Paratyphi A, and Choleraesuis (Fig. S1), indicating that Salmonella may have acquired IncHI plasmids from multiple sources.

Interestingly, a Tn7-like transposon inserted in pS5-403-2 carries three heavy metal resistance genes (i.e., cusR, silE and pcoE) among several other genes (Table 2). A highly similar Tn7-like transposon, which also carries cusR, silE and pcoE, was identified in Salmonella Senftenberg; in this strain this element is chromosomally integrated and inserted at the 3′ end of a gene that encodes a NAD-utilizing dehydrogenase. The transposons found in these two serovars have a similar open reading frame (ORF) content (24 out of 29 ORFs are found in both transposons) and highly similar tnsABCD genes (99% nucleotide sequence identity). Two other IncH1 plasmids (i.e., pAPCE-O1-R and pR478) also carry a Tn7-like transposon. Cluster analysis shows that these plasmids are most similar to pS5-403-2 (Fig. S1). Closely related Tn7-like transposons (based on TnsABCD homology) that are chromosomally inserted have been found in S. Tennessee str. CDC07-0191, Escherichia albertii TW07627, and Enterobacter cloacae ATCC 13047 (GenBank accessions NZ_ACBF01000002, ABKX01000002, NC_014121, respectively). Interestingly, in all these organisms Tn7-like transposons are not inserted at the previously described attTn7 site (i.e., downstream of the glmS gene) [30], but are inserted at the 3′ end of the gene that encodes a NAD-utilizing dehydrogenase, similar to what we observed for the integration site in the S. Senftenberg chromosome. Given that these Tn7 elements are so similar, but found in the exact same location in different genera, it suggests the possibility that they reside in a new attachment site that is recognized by one of the two TnsD proteins encoded in this element.

IncI1-IncFIB cointegrated plasmids that carry virulence genes in the accessory region are found in serovars that are rarely isolated from animal hosts in the US

In addition to the two resistance plasmids identified in Salmonella Montevideo, we identified plasmids with cointegrated IncI1-IncFIB replicons (i.e., “IncI1-IncFIB cointegrated plasmids”) in isolates representing serovars Inverness (plasmid pR8-3668), Mississippi (pA4-633), Rubislaw (pA4-653), and Urbana (pR8-2977). These plasmids were initially predicted by de novo assembly as scaffolds without homology to the reference genomes [5] (Table 2 and Fig. 2). We confirmed the presence of these plasmids through in silico and experimental approaches, including identification of essential plasmid genes (e.g., genes encoding replication and conjugation functions) through inspection of the annotation of the relevant contigs, PFGE-based estimates of plasmids sizes, and PCR-based validation of a circular molecule (see Materials and Methods for details). These four plasmids have the same backbone previously described for other Salmonella and E. coli IncI1 plasmids [31]; this backbone includes a type IVa pilus operon as well as genes encoding a type IV secretion system, a shufflon recombinase, an antirestriction protein, plasmid stability proteins, and DNA repair systems (error-prone repair proteins UmuC and UmuD).

Figure 2. Circular representation of IncI1-IncFIB cointegrated plasmids.

Circular representations of IncI1-IncFIB plasmids identified in Salmonella serovars Urbana pR8-2977 (inner circle), Inverness pR8-3668 (second circle) and Mississippi pA4-633 (third circle). Plasmids were aligned against the Urbana plasmid. Gene cluster (1) represents insertions in the accessory region in plasmids in serovars Mississippi and Inverness relative to the S. Urbana plasmid. Gene cluster (2) represents insertions in plasmids in serovars Mississippi and Inverness relative to the S. Urbana plasmid. For plasmid in serovar Rubislaw pA4-653, there is a 90 kb insertion (3) that differentiates this plasmid from other plasmids (outer cluster of genes), and encodes several putative virulence genes, arrows indicate the insertion location for this region in plasmid pA4-653. Genes were color coded according to function as follows: virulence (red), plasmid transfer (blue), transposition/IS (yellow), replication (green), plasmid stability (turquoise), metabolism (purple), phage origin proteins (brown), antibiotic production (black), hypothetical proteins (grey).

While IncI-IncFIIA cointegrated plasmids have previously been described in enterotoxinogenic E. coli (ETEC) [20], [31], this is the first description of Salmonella plasmids that encode IncI1 as well as the IncFIB replicons. In addition to this unique replicon arrangement, the four IncI1-IncFIB cointegrated plasmids described here are also unique with regard to the virulence gene arrays found in the accessory region of these plasmid (i.e., the region between the replicon and a site-specific recombinase [encoded by LTSEURB_6814 in serovar Urbana]). In the serovars Inverness, Mississippi and Urbana the IncI1-IncFIB cointegrated plasmids contain a region of 20 to 30 kb encoding similar putative virulence proteins such as adhesins, fimbrial proteins, and iron uptake proteins (Table 2). The same region on the IncI1-IncFIB cointegrated plasmid of serovar Rubislaw contains a different set of virulence genes, including sopE, as well as genes encoding a serine/threonine phosphatase, a hemolysin, and an adhesin. Moreover, the cointegrated plasmid of serovar Rubislaw also encodes a gene involved in the biosynthesis of antibiotics (i.e., 3,3′-neotrehalosadiamine, ntdA). A guide tree created using the Mauve algorithm (used to cluster plasmids based on their overall identity), clustered the four plasmids identified here on a branch that is clearly separated from the branch that contains the previously reported Salmonella IncI1 plasmids [20], [31] (Fig. S2). In addition, this comparison showed that four IncI-IncFIB cointegrated plasmids identified here are more similar to the ETEC cointegrated plasmid p557 than to plasmids that carry only the IncI1 replicon.

PCR assays targeting the IncI1 and IncFIB replicons were used to further screen for the distribution of this IncI1-IncFIB cointegrated virulence plasmid among a broader Salmonella subsp. enterica population, using a set of 107 isolates representing 84 serovars (Table S2). Nine isolates representing serovars Enteritidis (2 isolates) as well as Rubislaw, Berta, Hindmarsh, Holcomb, Paratyphi C, Wandsworth, and Typhimurium (1 isolate each) were positive for IncFIB replicon, but not for IncI1. Both IncFIB and IncI1 replicons were detected in three serovar Inverness isolates and in one serovar Manhattan isolate (Table 3). The detection of these two replicons could indicate the presence of two plasmids, an IncI1 and an IncFIB, or the presence of an IncI1-IncFIB cointegrated plasmid. For these four isolates, the amplified incI1 fragment (976 bp) was sequenced. The newly sequenced IncI1 sequences, along with the IncI1 sequences for the four IncI1-IncFIB cointegrated plasmids identified by whole genome sequencing and previously reported IncI1sequences (Table S1), were used to construct a IncI1 nucleotide sequence maximum likelihood phylogeny. This phylogeny placed the IncI1 sequences into two well supported clades (98 and 100% bootstrap support) (Fig. 3); (i) a clade (Clade I) containing IncI1 sequences for 8 plasmids (4 Inc1-IncFIB cointegrated plasmids for which the full genome sequence was described here and four plasmids for which only IncI1 was sequenced), and (ii) a clade (Clade II) containing 15 IncI1 sequences, representing 5 Salmonella and 10 E. coli plasmids which were fully sequenced previously (see Table S1). The eight plasmids of Clade I were obtained from strains classified as serovars Inverness (4 plasmids), Mississippi, Rubislaw, Urbana, and Manhattan. Interestingly, these five serovars are rarely isolated from animal and human hosts in the United States. Based on a CDC report for 2009 [2], [32] these serovars represented only 0.4% of the nonhuman and 1.6% of human isolates obtained over 10 years in the US. The four plasmids in Clade I for which full genome sequences were available contained predominantly putative virulence genes in the accessory region (see Table 2), including hlyD (secretion of hemolytic toxins) and eal (cellular adhesion) (both found in three of the four plasmids). Two plasmids of the clade II plasmids represented co-integrated plasmids found in E. coli. The 5 Salmonella serovars represented in this clade are Typhimurium, Thompson, Kentucky, and Heidelberg. Based on a CDC report for 2009 [32] these serovars represented 31% of the 13,006 nonhuman and 22% of human isolates obtained over 10 years in the US. Interestingly, the 15 clade II plasmids did not include any putative virulence genes in their accessory region, but often carried antibiotic resistance genes, including genes conferring resistance to beta lactam antibiotics (e.g., blaCTX, blaCMY, blaTEM [3 plasmids]); aminoglycosides (e.g., aadA1, aadA2, strA, strB [3 plasmids]; tetracycline (e.g., tetA [3 plasmids]); chloramphenicol (e.g., cmlA [2 plasmids]); sulfonamide (e.g., sul2 [1 plasmid]); as well as heavy metals and disinfectants (e.g., sugE, arsR, cusR, silE, qacH [5 plasmids]). Unfortunately further confirmation of plasmid classification into clade I and II with pMLST, which has been used previously to characterize IncI1 plasmids [33] was not possible as plasmids described here have diverged considerably from IncI1 plasmids. For example, among the five pMLST loci, two loci are absent (i.e., ardA, and sogS) in these IncI1 plasmids and the other three loci have diverged considerably (e.g., 60–70% for repI1).

Figure 3. Phylogeny inferred with maximum likelihood of the IncI1 replicon.

Maximum likelihood phylogeny conducted with IncI1 sequences for plasmids found in this study and currently available sequences. Analysis was conducted with RAxML. Two clades were observed (clade I and clade II), with clade I containing the Salmonella plasmids analyzed here.

Table 3. Isolates positive for type IVb pili, and/or IncI1-IncFIB replicons.

Identification of novel genomic islands and transposons encoding genes that may contribute to host specificity and Salmonella transmission

We used SIGI-HMM [34] and comparative genomic analyses with the Mauve algorithm and RAST [35][37] to predict and identify putative genomic islands among the 16 genomes analyzed here. As expected, a number of putative Salmonella genomic islands were identified using this approach. We defined genomic islands as chromosomal regions more than 15 kb in length, that represent variable gene content across Salmonella strains and differ in codon usage compared to the rest of the genome, which is a definition similar to Juhas et al. [38]. To focus on novel genomic islands that may be relevant for emergence of strains with unique pathogenicity characteristics, we did not further analyze regions representing previously reported Salmonella pathogenicity islands (SPIs). The remaining nine putative genomic islands were designated as Salmonella genomic islands (SGI) 2 to 10; SGI1 was previously assigned to a genomic island that encodes antibiotic resistance genes found in S. Typhimurium DT104. One putative genomic island (SGI4) was found in most Salmonella genomes reported so far, while six putative genomic islands (SGI6 to 10) did not contain apparent features relevant to host specificity or transmission (see Table S3 for more details on these islands). However, we identified two novel genomic islands in these 16 genomes (SGI2 and SGI3) that contained genes that may be linked to unique host specificity and transmission characteristics and in particular an ability to interact with plant hosts. These two regions are discussed in more detail below.

SGI2 was predicted in nine genomes representing Salmonella serovars Montevideo, Johannesburg, Urbana, Baildon, Uganda, Minnesota, Mississippi, Give, and Hvittingfoss. The insertion site for SGI2 was identified as tRNA-leu (CAA), except in serovar Hvittingfoss (Table 4). SGI2 always encodes one or two toxin-antitoxin modules and different repertoires of restriction-modification systems, suggesting similar functionality despite a difference in gene repertoire. SGI2 shows high sequence similarity only in three genomes (serovars Montevideo, Urbana, and Johannesburg). However, distinct variants with conserved regions of SGI2 were found in the other six genomes, and were designated as SGI2.1 to SGI2.6 (Fig. S3). Four of these variants encode genes that may facilitate propagation in animal and plant hosts, including SGI2.3 (Give), SGI2.4 (Baildon), SGI2.5 (Uganda), and SGI2.6 (Hvittingfoss) (Table 4) [39][41]. SGI2.3 and SGI2.5 encode different guanine-binding proteins; SGI2.4 carries a gene encoding a pectin lyase fold protein, which may facilitate interaction with plant hosts, and SGI2.6 includes an insertion of a toll-interleukin receptor (see discussion for specific interpretations).

SGI3 is a novel 31 kb genomic island that was only identified in the serovar Mississippi genome. SGI3 is inserted at the 3′ end of the GDP-mannose pyrophosphorylase gene, and resembles a region in the chromosome of Yersinia intermedia ATCC 29909 (GenBank accession NZ_AALF00000000). SGI3 encodes an integrase, transposases, OpgC (succinyl modification of osmoregulated periplasmic glucan), a cellulose synthesis protein, beta-galactosidase, and several proteins involved in stress response and regulation (i.e., sensory box histidine kinase/response regulator, an anti-sigma factor antagonist, and a serine phosphatase regulator) (see discussion for specific interpretations).

In addition to the genomic islands described above, we also identified 35 putative transposons (<1 kb to >15 kb), which encode a range of proteins that may facilitate transmission by favoring their maintenance, including restriction modification systems, efflux pumps, O-antigen conversion, and disinfectant resistance (Table S4). One of these transposons, a 10 kb Tn-31-like transposon, was present in all 16 genomes. This transposon encodes a potassium efflux system and also contains an operon encoding acriflavin resistance. In 11 genomes (serovars Baildon, Gaminara, Give, Inverness, Johannesburg, Mississippi, Montevideo, Senftenberg, Uganda, Urbana, and Wandsworth), we identified a transposon insertion downstream of SPI-1. This transposon encodes genes involved in transposition and a serine/threonine specific phosphatase-1.

Identification of prophages with highly mosaic genome architecture in selected Salmonella serovars

Lysogenic phages and/or remnants of phages were initially detected by Prophinder [42] in all the 16 Salmonella genomes, and subsequently confirmed in silico by manual annotation and comparative analysis. Only phages that were found in a single contig and with a size of at least 20 kb were further analyzed. A total of 42 putative phages were identified in the 16 genomes; these phages were classified into 12 groups; 11 groups represent phages that are similar to previously reported phages (e.g., PSP3-like, P22-like, Gifsy-1-like), while 1 group comprises 6 phages that do not show similarity with any previously reported Salmonella phages (Table S5).

Analysis of “morons” (i.e., genes not required for the phage infective cycle) [43], [44] among the 42 phages, identified genes encoding DNA-methylases in 15/42 phages (including PSP3-like, HP2-like, Fels2-like, PhiCTX-like, and HK97-like phages) as well as genes with potential functions in fitness and virulence. For instance, O-antigen conversion genes were identified in five phages (see Table S5). While all four Gifsy-1-like phages identified here (i.e., PhBail-1 in S. Baildon, PhWands-3 in S. Wandsworth, PhHvi-1 in S. Hvittingfoss, and PhInv-2 in S. Inverness) carried putative virulence genes, Gifsy-1 does not appear to be fully conserved in these four genomes. The two virulence genes previously reported in Gifsy-1-gtgA and gipA, which facilitate Salmonella growth or survival in Peyer's patches and enterophatogenesis [45][47]-, are both present in PhWand3 and PhHvi1, while only gtgA and only gipA are present in PhBai-1 and PhInv-2, respectively (Fig. 4).

Figure 4. Representation of prophages carrying virulence genes.

Blast comparison of Gifsy-1-like prophages detected in the 16 genomes. Prophage order from top to bottom is PhInv-2 (S. Inverness), Gifsy-1 (S. Typhimurium LT2), PhWands-3 (S. Wandsworth), PhHvi-1 (S. Hvittingfoss), and PhBai-1 (S. Baildon). Arrows in orange represent coding regions, and grey shaded regions represent regions with homology. The two virulence genes (gipA and gtgA) are labeled in yellow and grey, respectively.

In addition, we identified putative virulence genes in three different phages in the genomes of serovars Inverness, Uganda and Gaminara. In S. Inverness, we identified a 48 kb phage (PhInv-1b) that encodes a secreted effector protein and the pertussis-like toxin ArtAB. In S. Uganda, we identified a 29 kb phage resembling HK97 (PhUga-3) that encodes two copies of virulence protein msgA and an O-antigen conversion protein. Finally, we identified a 52 kb phage in S. Gaminara (PhGam-1), which resembles a Stx-2 phage. While this prophage does not encode Stx-2, it does encode a secreted effector protein and an attachment invasion protein (Fig. 5).

Figure 5. Linear representations of three phage genomes identified to have virulence genes.

At the top is phage PhInv-1b in S. Inverness, the middle is phage PhUga-3 in S. Uganda and in the bottom is PhGam-1 in S. Gaminara. Genes were color coded according to function as follows: virulence (red), replication (yellow), phage structural genes (green), metabolism (purple), and hypothetical protein (grey).

Comparison of phages belonging to the same group indicated a highly mosaic composition of the phage genomes. For instance, a comparison of six phages that resemble Enterobacteria phage PSP3 shows that phages in the genomes of serovars Uganda (PhUga-5), Johannesburg (PhJoh-3), Adelaide (PhAde-1), and Gaminara (PhGam-2) are conserved and more similar to phage PSP3, while related phages in Urbana (PhUrb-1) and Montevideo (PhMont-3) are more distinct from these phages and PSP3 (Fig. 6A). Both PhMont-3 and PhUrb-1 have several gene duplications (i.e., eight in PhMont-3 and six in PhUrb-1), they have nine genes in common that are absent in the rest of phages in this group, as well as 8 and 4 new unique genes, respectively. Similar to the PSP3-like phages, a mosaic architecture was also identified in the six phages that resemble Salmonella phage P22 (Fig. 6B).

Figure 6. Comparison, made using the Blast algorithm, of PSP3-like and P22-like phages.

(A) Comparison of six phages that resemble PSP3; phages were identified in serovars Uganda, Johannesburg, Adelaide, Gaminara, Montevideo, and Urbana. Coding regions are represented as blue arrows and regions of homology are shaded in grey. (B) Comparison of six phages that resemble P22, identified in serovars Wandsworth, Mississippi, Rubislaw, Uganda, Johannesburg, and Montevideo. Coding regions are represented as orange arrows and regions of homology are shaded in grey.

ICES1, a novel Salmonella integrative conjugative element, encodes a Salmonella Typhi virulence factor, the type IVb pilus operon

We identified a novel mobile element in the genomes of the three serovar Inverness, Urbana and Rubislaw strains (Fig. 7). This novel element is a putative integrative conjugative element (which we designated ICES1) inserted adjacent to the tRNA-CAA gene in these three genomes. ICES1 is flanked on one side by an integrase, and encodes phage genes (e.g., activator of prophage gene expression IbrB) as well as genes involved in transfer of the ICE (e.g., genes encoding an ICE relaxase and a conjugation system). ICES1 carries a type IVb pilus operon that resembles the Salmonella Typhi SPI-7, island which also has the SopEΦ prophage and the capsular operon [48]. The full type IVb pilus operon is present in the Salmonella Inverness, Urbana and Rubislaw ICES1, however the prophage is incomplete, and the capsular operon (found in S. Typhi) appears to be absent. The presence and location of the full type IVb pilus operon in these three strains was also experimentally confirmed using a long-range PCR mapping strategy. Recently, an ICE with similar characteristics was reported in Salmonella bongori as well as S. enterica serovars Senftenberg, Hadar, and S. enterica subspecies VII [49] (Table S1); all three of these ICEs also encode the type IVb pilus operon, which is less conserved in S. Senftenberg. Phylogenetic analysis based on the sequences of the ICES1-encoded pilQ found in this study and publicly available pilQ sequences revealed three pilQ clades with low sequence divergence within clades, including (i) a clade consisting of serovar Typhi, Paratyphi C, Dublin, and S. bongori (ii) a divergent clade composed of serovars Inverness, Urbana, Rubislaw, and subspecies VII, and (iii) a separate clade representing serovar Senftenberg (Fig. S4). This phylogenetic pattern suggests multiple horizontal gene transfer events of the type IVb pilus operon; in particular the serovar Senftenberg type IVb pilus operon appears to have a distinct origin, while the type IVb pilus operons in serovars Typhi, Paratyphi C, Dublin, and S. bongori may represent another origin.

Figure 7. Comparison, made using the Blast algorithm, of SPI-7 in S. Typhi, ICES1 identified in this study, and previously sequenced ICE.

At the top is SPI-7 of S. Typhi with its three main regions (i.e., type IVb pilus operon, SopEΦ phage and capsular operon). Black arrows indicate the position of the primers used to validate the presence of the type IVb pilus operon in ICES1 in S. Rubislaw, S. Inverness and S. Urbana. Blue arrows are the coding regions and grey shaded are regions that present homology.

To further probe the distribution of the type IVb pilus operon among a broader S. enterica population, a set of 107 isolates representing 84 serovars was screened, using PCR, for the presence of the most conserved genes in the type IVb pilus operon (i.e., pilQ, pilV, and rci). Six of the 107 isolates were positive for all three genes including 3/3 serovar Inverness and 2/5 Manhattan isolates as well 1/2 Urbana isolates tested. One Paratyphi C isolate tested positive for pilQ, but negative for pilV and rci (Table 3), which are less conserved than pilQ.

Based on these data, we investigated whether isolates encoding the type IVb pili may show a particular ability to invade human intestinal epithelial Caco-2 cells. Caco-2 invasion assays were performed with (i) one Manhattan isolate, one Urbana isolate, and one Rubislaw isolate, that all contained the type IVb pilus operon, and (ii) one Manhattan isolate, one Urbana isolate, and one Rubislaw isolate that did not contain the type IVb pilus operon (as supported by the absence of pilQ, pilV, and rci; Table 3). No difference (p = 0.347) in invasion efficiency was observed between isolates with or without the type IVb pilus operon.

Remarkably, all four isolates of Salmonella Inverness tested in this study that were positive for the type IVb pilus operon (as supported by the presence of pilQ, pilV, and rci) were also positive for the IncI1 and IncFIB replicons, suggesting presence of two plasmids (one IncI1 and one IncFIB) or of a IncI1-IncFIB cointegrated plasmid (see above), further validation is needed to confirm if these replicons are found on two or one plasmid. While this could indicate that these four serovar Inverness isolates represent the same clonal strain, XbaI PFGE analysis revealed four different PFGE patterns for these isolates (Fig. S5), suggesting that they may not be closely related. The presence of both the type IVb pilus operon and IncI1-IncFIB replicons may therefore be conserved in the genome of serovar Inverness.


Salmonella strains are subject to frequent integration of new mobile elements. These integration events may give rise to strains with novel pathogenic phenotypes and are often associated with the emergence of new endemic or epidemic strains. Genome scale prediction of mobile elements conducted in this study using 16 different Salmonella serovars [5] for which no in-depth mobile element analyses have been available to date, considerably expands our understanding of the diversity of mobile elements found in this important foodborne and animal pathogen. Key findings from this study include (i) a new type of co-integrated Salmonella plasmid, which is found among multiple serovars, that contains accessory genes with putative virulence functions, (ii) previously unidentified genomic islands and prophages that encode functions that may facilitate Salmonella propagation among animal and plant hosts, and (iii) an ICE encoding the type IVb pilus operon, which could facilitate transfer of this Salmonella Typhi virulence factor to non-typhoidal Salmonella serovars.

A new type of co-integrated Salmonella plasmid, which is found among multiple serovars, contains accessory genes with putative virulence functions

In this study we identified a type of IncI1-IncFIB cointegrated plasmid that has not previously been reported in Salmonella. A key unique feature of this plasmid is that it contains both the IncI1 and the IncFIB replicon, but contains the IncI1 backbone; this plasmid thus represents a IncI1-IncFIB cointegrated plasmid. IncI1 and IncFIB plasmids have previously been described in Salmonella [17], [50]. Other cointegrated plasmids previously reported in Salmonella include IncFIIA-IncFIB in Salmonella Kentucky (pCVM29188_146), as well as the Salmonella virulence plasmid in serovars Typhimurium, Enteritidis and Choleraesuis [17], [50]. The presence of multiple replicon types was recently reported for F plasmids in E. coli [51], indicating that the presence of multiple replicons could be a characteristic of F plasmids. The IncI1-IncFIB cointegrated plasmid described here has a novel array of virulence genes in the accessory region. Prior to our description of this plasmid, the accessory region of IncI1-plasmids, which have the same backbone as the IncI1-IncFIB cointegrated plasmid described here, was found mostly to contain antimicrobial resistance genes and specifically genes that encode resistance to cephalosporins [20].

Interestingly, we also found initial evidence for differential accessory gene “content” on related plasmids with the IncI1 replicon. IncI1-IncFIB plasmids that were obtained from Salmonella serovars that are rarely isolated from food-producing animals in the US (i.e., serovars Inverness, Rubislaw, Mississippi, and Urbana) [2], [32] all harbored virulence genes in the accessory region of the plasmid. For example, three of these plasmids harbor hlyD, encoding a protein essential for secretion of hemolytic toxins in E. coli [52], and eal, which encodes a protein involved in cellular adhesion [53]. On the other hand, plasmids with an IncI1 replicon that grouped into a separate Clade (clade II, Fig. 3) typically have antimicrobial resistance genes integrated in the accessory region of these IncI1 plasmids [31], [50], [54], [55]. The Salmonella isolates that carried these antimicrobial resistance plasmids represent serovars commonly isolated from food-producing animals, including serovars Typhimurium, Kentucky, Heidelberg and Thompson [2]. These observations suggest that additional studies on accessory genes of plasmids found in serovars that are predominant in different host types will provide further insights in the evolution of antibiotic resistance and virulence associated characteristics in Salmonella.

In addition to IncI1-IncFIB cointegrated plasmids, we also identified two additional plasmids in S. Montevideo FSL S5-403. This is the only strain that showed phenotypic antibiotic resistance in this study. This Salmonella Montevideo strain contained an IncHI plasmid that is more similar to E. coli and E. cloacae plasmids, than to IncHI plasmids previously identified in Salmonella serovars Typhi and Paratyphi A [56][58]. This Salmonella Montevideo strain also contained an IncN2 plasmid (pS5-403-1). In addition to this description, IncN plasmids carrying antimicrobial resistance (quinolone) were previously reported in four Salmonella serovars (i.e., Bredeney, Typhimurium, Saintpaul, and Kentucky) [59]. Importantly, the IncN2 plasmid identified here shares its backbone with a plasmid that was previously identified in E. coli as carrying the blaNDM-1 extended spectrum beta-lactamase, which confers resistance to multiple antimicrobials [28]. blaNDM-1 was also recently identified in a single Salmonella isolate from the United States using PCR [28], [60], but the genomic location of this gene was not identified. Identification of an IncN2 plasmid with an identical backbone to the plasmid that was previously found to carry blaNDM-1, suggests that Salmonella has the potential to acquire the E. coli IncN2 plasmid carrying the blaNDM-1.

Previously unidentified genomic islands and prophages may encode functions that facilitate Salmonella propagation among animal and plant hosts

In this study we identified two novel genomic islands; one of which (SGI2) has evolved six different variants. SGI2 is conserved in only three of the genomes. These three genomes represent closely related serovars (i.e., Montevideo, Johannesburg, and Urbana) classified into the clade B group proposed by den Bakker et al. [5]. Interestingly, different SGI2 variants harbor genes encoding proteins with similar functions, including restriction modification systems [61], [62], toxin-antitoxin systems [61], [63][65], and genes partially involved in host specificity and virulence, even though the specific genes encoding a given function can differ considerably between SGI2 variants. For example, different SGI2 variants encode different restriction modification (RM) systems, including two different type I RMs, two different type II RMs, and one type III RM, consistent with previous reports that RM system substitution is common in bacterial genomes and with a report of a cassette-like variations of restriction enzyme genes in the same genome region in several E. coli strains [61], [62], [66]. SGI2 is inserted at the tRNA-Leu (CAA) in eight of the nine genomes, suggesting that this region is an insertion “hot spot” for genomic islands that differ in gene content, but that are very similar in the functions they encode. Remarkably, tRNA-Leu (CAA) has been previously defined as a highly variable hot spot for gene acquisition in Salmonella enterica, including restriction modification systems [67].

Interestingly, SGI2 also appears to encode distinct propagation associated functions in different genomes. For instance, SGI2 variants in serovars Give and Uganda encode a guanine-binding protein, while the serovar Hvittinfgoss SGI2 encodes a toll-interleukin receptor. These proteins have been shown to affect cellular functions involved in host intracellular signaling and in the mammalian immune response [40], [68], [69], suggesting a possible role in infection of mammalian hosts. The serovar Baildon SGI2 encodes a pectin lyase, an enzyme that degrades the plant cell wall and thus may facilitate bacterial invasion of plant cells [39]. Interestingly, the serovar Mississippi SGI3 also appears to encode functions that may facilitate interaction with plant hosts, including a cellulose synthesis protein, which was shown to be required for adherence and aggregation to plant roots in Agrobacterium [70] and OpgC (osmoregulated periplasmic glucan), which is also associated with host-pathogen interaction in plant pathogens [71], [72]. While our findings suggest that SGI2 and SGI3 encode functions that may facilitate interactions with plant and animal hosts, phenotypic characterization and infection studies in different hosts are needed to test this hypothesis and to better understand the role of mobile elements in transmission of Salmonella among animal and plant hosts.

Previous reports of Salmonella phage-borne virulence genes have focused on Salmonella Typhimurium, Typhi and Paratyphi [24], [27], [73] and have revealed important roles of prophages in the pathogenicity of these serovars. For example, Gifsy-2 prophage has two major virulence determinants (i.e., sodC1 and sseI/gtgE/srfH), that increase the capability of S. Typhimurium to produce systemic infection in mice [46], [74]. In this study, we found a number of phage morons that encode virulence factors, which have previously been thought to be limited to a few serovars. For example, we identified, in the serovar Inverness genome, a 48 kb phage that encodes a pertussis-like toxin, ArtAB; a phage with this putative virulence gene has previously only been reported in S. Typhimurium DT104 [73]. Our comparative analysis of the chromosomally inserted phage genomes reported here along with previously reported Salmonella prophages also found Gifsy and λ-like phages with putative virulence genes (e.g., sopE, sodC1, gtgA) in a number of serovars. We also identified six phages of the P22 family, including two that encode for O-antigen conversion genes, consistent with previous studies that found O-antigen conversion genes in P22 phage genomes [75][77].

An ICE encoding the type IVb pilus operon could facilitates transfer of this Salmonella Typhi virulence factor to non-typhoidal Salmonella serovars

In this study we described a unique mobile element in Salmonella that harbors S. Typhi virulence genes. This mobile element, which we designated ICES1, encodes the type IVb pilus operon, an important virulence factor that facilitates S. Typhi invasion of human intestinal cells [78], [79]. While we identified this mobile element in the genomes of isolates representing serovars Inverness, Rubislaw, and Urbana, another recent study [49] also identified a very similar mobile element, which also encodes the type IVb pilus operon, in S. bongori as well as in S. enterica serovars Senftenberg and Hadar and a subspecies VII isolate. While a unique pathogenicity island repertoire (i.e., presence of SPI-18, SPI-7 and the cdtB islet) has typically been considered to be responsible for the unique clinical symptoms associated with S. Typhi infection [4], [5], [80][82], our study, along with other recent studies [5], [49], provides increasing support for the notion that S. Typhi virulence factors are more widely distributed among non-typhoidal Salmonella serovars than previously assumed and are often present on mobile elements, which facilitate dispersal.

While two types of type IV pili (i.e., IVa, and IVb) have been described in different Gram-negative and Gram-positive bacteria (e.g., Neisseria gonorrhoeae, Pseudomonas aeruginosa, Mycobacterium bovis, and Vibrio cholerae) [83], type IVb pili have only been reported in human pathogens [83]. Type IVa pili, on the other hand, have been detected in a wider range of bacteria. Numerous virulence-related functions are associated with type IV pili, including surface motility, bacterial aggregation, biofilm production, adhesion, invasion, and immune evasion [83]. In S. Typhi, the type IVb pilus operon is located in SPI-7, an island containing several independent mobile regions (i.e., the pilus operon, the SopEΦ phage, and the capsular operon) [4], [84]. Our finding of ICES1, along with the recent description of a very similar element in additional Salmonella serovars [49], indicates putative horizontal gene transfer of the type IVb pilus operon across diverse Salmonella strains. Dispersal of the type IVb pilus operon was also supported by phylogenetic analysis of its nucleotide sequences, which showed lower sequence divergence in the type IVb pilus operon among serovars Inverness, Rubislaw, and Urbana than expected given their phylogenetic relationships previously described [5]. This indicates likely recent transfer of the type IVb operon between these strains [49]. Horizontal transfer of this mobile element containing the IVb pilus operon is also supported by the recent report showing presence of this element in additional Salmonella serovars, and a demonstration of transfer by conjugation from the host strain into S. Typhimurium [49]. If one key trait making typhoidal strains pathogenic is also found in non-typhoidal strains, this suggests that a combination of traits makes typhoidal strains so dangerous [11]. Given that many of these traits are on mobile elements, it makes the emergence of new pathogenic strains likely.

While a diversity of virulence associated genes have been found in different Salmonella serovars [4], [5], [11], [12], our understanding of the association of virulence and host specificity associated phenotypes with different gene repertories in Salmonella is still limited. In this study we found the presence of IncI1-IncFIB cointegrated plasmids to be associated with the genomic presence of the type IVb pilus operon and ICES1. Remarkably, the IncI1-IncFIB cointegrated plasmid encodes a type IVa pilus operon, indicating that at least some strains classified into serovars Inverness, Rubislaw, and Urbana encode both type IVa and IVb pili; for serovar Inverness it appears that both of these pili are found in a several strains within this serovar, as supported by the PCR-based population screen performed here. To our knowledge this is the first report of presence of both of these pili in Salmonella strains. Future experiments focusing on the expression and the potential roles in virulence and transmission of these two pili in Salmonella strains, that encode both pili, are warranted to probe the functional importance of these genetic elements.

Materials and Methods


We analyzed 16 Salmonella serovars that were previously sequenced with the SOLiD™ next generation sequencing technology by our group [5]. These isolates represent the serovars Montevideo, Inverness, Rubislaw, Give, Mississippi, Urbana, Uganda, Senftenberg, Gaminara, Baildon, Minnesota, Hvittingfoss, Adelaide, Alachua, Wandsworth, and Johannesburg. For details about sequencing workflow, isolate sources, and accession numbers see den Bakker et al. [5].

Antimicrobial susceptibility was determined according to the National Antimicrobial Resistance Monitoring System (NARMS) protocol, at the New York State Animal Health Diagnostic Center (NYSAHDC). Minimal inhibitory concentrations (MIC) were determined using the Sensititre system (TREK Diagnostic Systems, Cleveland, OH), for the following 15 antimicrobials: amikacin, ampicillin, amoxicillin clavulanate, cefoxitin, ceftiofur, ceftriaxone, chloramphenicol, ciprofloxacin, gentamicin, kanamycin, nalidixic acid, streptomycin, sulfisoxazole, tetracycline and trimethoprim-sulfonamide. MIC values were interpreted according to the Clinical and Laboratory Standards Institute (CLSI).

Previously described transposable element nomenclature was used to classify chromosomally inserted mobile elements [38], [85], [86]. Briefly, transposons were classified as either composite or unit transposons [86], integrative conjugative elements (ICE) were defined as chromosomally inserted transposons that carry genes for insertion, excision and conjugative transfer [85], [86], and genomic islands were defined as relatively large regions of DNA (>15 kb), that presented variable gene content across Salmonella strains, and that were recognized by having different nucleotide codon usage than the rest of the genome [38].

Prediction of mobile elements

To predict mobile elements, we prepared a pseudogenome (contigs after scaffolding were connected with a pseudomarker “NNNCACACACTTAATTAATTAAGTGTGTGNNN” that was added to identify the different scaffolds). These pseudogenomes were used for mobile element prediction. We used SIGI-HMM, a program that identifies regions that differ in codon usage compared to the rest of the genome [34], and Prophinder, a web server that compares query sequences against a phage database [42]. This methodology was combined with a search for mobile element-related key words from the RAST annotations (e.g., integrase, transposase and phage). Most plasmids were found as large scaffolds that could not be aligned to the reference genome. These large scaffolds were considered putative plasmids and were further analyzed by comparative analysis. Briefly, RAST annotations were manually examined for essential plasmid backbone genes (e.g., genes involved in plasmid partitioning and replication). All the predicted mobile elements were analyzed by comparative analysis using a Mauve algorithm [35], [36], and with the comparative analysis tools of RAST [37]. Sequence data from selected plasmids, prophages, genomic islands and transposons were extracted from pseudogenomes. Because these are draft genomes, there is not an easy option to deposit these predicted elements in GenBank; however, all contigs are available in GenBank (see den Bakker et al.) [5]. For the predicted elements (plasmids, genomic islands, and prophages), sequences were extracted from the pseudogenome, and genbank files with the sequences and annotations of selected mobile elements are available at

Comparative and phylogenetic analysis of predicted plasmids

IncI1-IncFIB cointegrated plasmids were compared with previously sequenced plasmids of the IncI1replicon in Salmonella and E. coli (Table S1). Plasmids were aligned with the Progressive Mauve algorithm, and the guide tree was used to as an indicator of overall sequence similarity of these plasmids. Briefly, Mauve calculates a guide tree, which is a neighbor joining tree computed based on an estimate of the shared sequences among each pair of input genomes. In addition, a maximum likelihood phylogeny was inferred based on the replicon nucleotide sequences, using RAxML 7.03 [87]. The nucleotide substitution model was a general time reversible model with a gamma parameter for rate heterogeneity. For the two plasmids predicted in Salmonella Montevideo, we compared each plasmid individually with similar publicly available plasmids using Mauve and the comparative tools in RAST.

Prophage comparative analysis

Prophages of at least 20 kb predicted by prophinder were extracted from the original contigs, but only if they were found in a single contig. Sequences were annotated in RAST and comparative analyses were performed as described above. Figures were prepared with Easyfig [88].

PCR confirmation of predicted mobile elements

Predicted plasmids were validated by PCR. Regions of approximately 500 bp from the contig's end and start on predicted plasmids were used as template for primer design. A PCR with these newly developed primers was used to validate the circular conformation of plasmids. Primer sets and PCR conditions are provided in Table S6. To confirm that the amplicon obtained correspond to plasmids, PCR products were purified (Exonuclease I and shrimp alkaline phosphatase [USB, Cleveland, OH] and sequenced (Applied Biosystems Automated 3730 DNA Analyzer at the Cornell University Life Sciences Core Laboratories Center). Sequences were assembled with SeqMan (DNAStar Inc., Madison, WI), and overlap with the end and start of the plasmid contig was examined. Two plasmids were not validated with this approach (i.e., pA4-653 in serovar Rubislaw, and pS5-403-2 in serovar Montevideo). For these two plasmids we used long-range PCR, and estimated the size of the plasmid with PFGE as described below. Takara LA Taq (Takara BIO Inc., Shiga, Japan) was used for long-range PCR, according to manufacturer recommendations.

Pulsed field gel electrophoresis estimation of plasmid sizes

Plasmids on serovars Montevideo (pS5-403-2) and Rubislaw (pA4-653) were not validated by the traditional PCR approach. To obtain an approximation on the size of these plasmids, we conducted PFGE. Briefly, plasmids were extracted using Plasmid Midi Kit (Qiagen Inc., Valencia, CA). After isolation, two plugs were prepared per isolate by mixing 80 ng/µl and 160 ng/µl of plasmid DNA with melted 1.0% of SeaKem Gold (SKG) agarose (Lonza Walkersville Inc., Walkersville, MD) in TE (pH 8.0). In silico restriction of the predicted plasmids was performed with NEBcutter V2.0, and enzymes that cut one or two sites in the predicted plasmids were used. Specifically, Not1 and Fse1 were used to digest pS5-403-2, and AvrII and FseI were used to digest pA4-653. All enzymes were acquired from New England Biolabs, Ipswich, MA. Restriction digestion was conducted on 2 mm plugs slices in 200 µl restriction mixture with 10–50 units of enzyme, at 37°C for 1 hour. Electrophoresis was conducted with the following conditions: 1.0% SKG agarose, 14°C, initial switch of 0.55 s, final switch of 5 s, 6 V/cm, for 22 h.

Type IVB pili and plasmid screening

pilQ was chosen as the initial target for a population wide screening for the type IVb pilus operon because it is one of the most conserved genes found in this operon. Using PCR we screened for the presence of type IVb pili in a collection of 107 human and animal clinical isolates, representing 84 different Salmonella serovars (Table S2). These isolates were selected to represent multiple isolates of the serovars identified to carry the type IVb pilus operon and to represent serovar diversity. If the PCR results were positive for the presence of pilQ, additional PCRs targeting the rci and pilV genes were performed. For IncI1-IncFIB plasmid screening, the same 107 isolates were used. Primers were designed based on the IncI1 and IncFIB replicons, primers and PCR conditions are available in Table S6.

Supporting Information

Figure S1.

Tree generated using the Mauve algorithm with IncH1 plasmids. Alignment was generated of plasmids found in Salmonella serovars (i.e., Montevideo, Paratyphi A, Typhi, and Choleraesuis), S. marcescens, E. coli and E. cloacae.


Figure S2.

Tree generated using the Mauve algorithm of IncI1 plasmids in E. coli and Salmonella, and IncI1-IncFIB plasmids found in this study.


Figure S3.

Blast comparison of SGI2 and its six variants. Green arrows indicate coding regions, and regions with >50% homology are linked by grey shaded areas.


Figure S4.

Phylogenetic tree inferred with Maximum Likelihood showing evolutionary relationships between pilQ sequences found in this study and previously reported pilQ sequences.


Figure S5.

PFGE dendogram for the four S. Inverness isolates positive for the type IVb pilus operon and the IncI1-IncFIB cointegrated plasmid. Four different XbaI PFGE patterns were identified for these four S. Inverness isolates harboring the type IVb pilus operon and the IncI1-IncFIB replicons.


Table S1.

Table containing plasmids, replicon types, host and accession number used for comparative analysis.


Table S2.

Table containing list of isolates used for PCR screening, and results of PCR screening for pilQ, pilV, rci, IncI1, and IncFIB.


Table S3.

List of putative genomic islands detected among the 16 genomes analyzed in this study.


Table S4.

List of composite and unit transposons detected among the 16 genomes analyzed in this study.


Table S5.

List of the 42 phages identified in this study that represent a size of at least 20 kb and that were identified on only one contig. Phages were classified based on homology to previously described phages in 12 groups.


Table S6.

Table containing primers & PCR conditions used for plasmids and ICES1 validation and for population PCR-based screen.



We would like to thank Barbara Bowen for her help with plasmid extraction and long range PCR experiments.

Author Contributions

Conceived and designed the experiments: AIMS HCdB LDR-R MW. Performed the experiments: AIMS HCdB LDR-R CAC MLR LD SB. Analyzed the data: AIMS HCdB CAC JEP. Contributed reagents/materials/analysis tools: CAC GG LD EB MRF MW. Wrote the paper: AIMS HCdB CAC KH MW.


  1. 1. Scallan E, Hoekstra RM, Angulo FJ, Tauxe RV, Widdowson MA, et al. (2011) Foodborne illness acquired in the United States–major pathogens. Emerg Infect Dis 17: 7–15.
  2. 2. [CDC] Center for Disease Control and Prevention. National Enteric Disease Surveillance: Salmonella Surveillance Overview (2011) Available.
  3. 3. Grimont P, Weill F (2007) Antigenic Formulae of the Salmonella Serovars. 9th. Paris: World Health Organization Collaborating Centre for Reference and Research on Salmonella, Pasteur Institute.
  4. 4. Jacobsen A, Hendriksen RS, Aaresturp FM, Ussery DW, Friis C (2011) The Salmonella enterica Pan-genome. Microb Ecol 62: 487–504.
  5. 5. den Bakker HC, Moreno Switt AI, Govoni G, Cummings CA, Ranieri ML, et al. (2011) Genome sequencing reveals diversification of virulence factor content and possible host adaptation in distinct subpopulations of Salmonella enterica. BMC Genomics 12: 425.
  6. 6. Didelot X, Bowden R, Street T, Golubchik T, Spencer C, et al. (2011) Recombination and population structure in Salmonella enterica. PLoS Genet 7: e1002191.
  7. 7. van Asten AJ, van Dijk JE (2005) Distribution of “classic” virulence factors among Salmonella spp. FEMS Immunol Med Microbiol 44: 251–259.
  8. 8. Doublet B, Boyd D, Mulvey MR, Cloeckaert A (2005) The Salmonella genomic island 1 is an integrative mobilizable element. Mol Microbiol 55: 1911–1924.
  9. 9. Chu C, Doublet B, Lee Y-L, Cloeckaert A, Chiou C-S, et al. (2011) Salmonella genomic island 1-J variants associated with change in the antibiotic resistance gene cluster in multidrug-resistant Salmonella enterica serovar Virchow isolated from humans, Taiwan, 2004-2006. Clin Microbiol Infect 18: 47–53.
  10. 10. Mellmann A, Harmsen D, Cummings CA, Zentz EB, Leopold SR, et al. (2011) Prospective genomic characterization of the German Enterohemorrhagic Escherichia coli O104:H4 outbreak by rapid next generation sequencing technology. PLoS One 6: e22751.
  11. 11. Sabbagh SC, Forest CG, Lepage C, Leclerc JM, Daigle F (2010) So similar, yet so different: uncovering distinctive features in the genomes of Salmonella enterica serovars Typhimurium and Typhi. FEMS Microbiol Lett 305: 1–13.
  12. 12. Fookes M, Schroeder GN, Langridge GC, Blondel CJ, Mammina C, et al. (2011) Salmonella bongori provides insights into the evolution of the Salmonellae. PLoS Pathog 7: e1002191.
  13. 13. Doublet B, Weill FX, Fabre L, Chaslus-Dancla E, Cloeckaert A (2004) Variant Salmonella genomic island 1 antibiotic resistance gene cluster containing a novel 3′-N-aminoglycoside acetyltransferase gene cassette, aac(3)-Id, in Salmonella enterica serovar Newport. Antimicrob Agents Chemother 48: 3806–3812.
  14. 14. Hall RM (2010) Salmonella genomic islands and antibiotic resistance in Salmonella enterica. Future Microbiol 5: 1525–1538.
  15. 15. Douard G, Praud K, Cloeckaert A, Doublet B (2010) The Salmonella genomic island 1 is specifically mobilized in trans by the IncA/C multidrug resistance plasmid family. PloS One 5: e15302.
  16. 16. Doublet B, Praud K, Weill FX, Cloeckaert A (2009) Association of IS26-composite transposons and complex In4-type integrons generates novel multidrug resistance loci in Salmonella genomic island 1. J Antimicrob Chemother 63: 282–289.
  17. 17. Rychlik I, Gregorova D, Hradecka H (2006) Distribution and function of plasmids in Salmonella enterica. Vet Microbiol 112: 1–10.
  18. 18. Chu C, Chiu CH (2006) Evolution of the virulence plasmids of non-typhoid Salmonella and its association with antimicrobial resistance. Microbes Infect 8: 1931–1936.
  19. 19. Fernández-Alarcón C, Singer RS, Johnson TJ (2011) Comparative genomics of multidrug resistance-encoding IncA/C plasmids from commensal and pathogenic Escherichia coli from multiple animal sources. PloS One 6: e23415.
  20. 20. Johnson TJ, Nolan LK (2009) Pathogenomics of the virulence plasmids of Escherichia coli. Microbiol Mol Biol Rev 73: 750–774.
  21. 21. Welch TJ, Fricke WF, McDermott PF, White DG, Rosso ML, et al. (2007) Multiple antimicrobial resistance in plague: an emerging public health risk. PloS One 2: e309.
  22. 22. Herrero A, Mendoza MC, Threlfall EJ, Rodicio MR (2009) Detection of Salmonella enterica serovar Typhimurium with pUO-StVR2-like virulence-resistance hybrid plasmids in the United Kingdom. Eur J Clin Microbiol Infect Dis 28: 1087–1093.
  23. 23. Folster JP, Pecic G, McCullough A, Rickert R, Whichard JM (2011) Characterization of bla(CMY)-encoding plasmids among Salmonella isolated in the United States in 2007. Foodborne Pathog Dis 8: 1289–94.
  24. 24. Zou QH, Li QH, Zhu HY, Feng Y, Li YG, et al. (2010) SPC-P1: a pathogenicity-associated prophage of Salmonella Paratyphi C. BMC Genomics 11: 729.
  25. 25. Ehrbar K, Friebel A, Miller SI, Hardt WD (2003) Role of the Salmonella pathogenicity island 1 (SPI-1) protein InvB in type III secretion of SopE and SopE2, two Salmonella effector proteins encoded outside of SPI-1. J Bacteriol 185: 6950–6967.
  26. 26. Mirold S, Rabsch W, Tschäpe H, Hardt WD (2001) Transfer of the Salmonella type III effector sopE between unrelated phage families. J Mol Biol 312: 7–16.
  27. 27. Ehrbar K, Hardt WD (2005) Bacteriophage-encoded type III effectors in Salmonella enterica subspecies 1 serovar Typhimurium. Infect Genet Evol 5: 1–9.
  28. 28. Poirel L, Bonnin RA, Nordmann P (2011) Analysis of the resistome of a multidrug-resistant NDM-1-producing Escherichia coli by high-throughput genome sequencing. Antimicrob Agents Chemother 55: 4224–4229.
  29. 29. Johnson TJ, Wannemeuhler YM, Scaccianoce JA, Johnson SJ, Nolan LK (2006) Complete DNA sequence, comparative genomics, and prevalence of an IncHI2 plasmid occurring among extraintestinal pathogenic Escherichia coli isolates. Antimicrob Agents Chemother 50: 3929–3933.
  30. 30. Parks AR, Peters JE (2009) Tn7 elements: engendering diversity from chromosomes to episomes. Plasmid 61: 1–14.
  31. 31. Johnson TJ, Shepard SM, Rivet B, Danzeisen JL, Carattoli A (2011) Comparative genomics and phylogeny of the IncI1 plasmids: A common plasmid type among porcine enterotoxigenic Escherichia coli. Plasmid 66: 144–51.
  32. 32. [CDC] Center for Disease Control and Prevention. Salmonella: annual summary (2009) Atlanta, GA: CDC, Division of Bacterial and Mycotic Diseases; 2003. Available:
  33. 33. García-Fernández A, Chiaretto G, Bertini A, Villa L, Fortini D, et al. (2008) Multilocus sequence typing of IncI1 plasmids carrying extended-spectrum beta-lactamases in Escherichia coli and Salmonella of human and animal origin. J Antimicrob Chemother 61: 1229–1233.
  34. 34. Waack S, Keller O, Asper R, Brodag T, Damm C, et al. (2006) Score-based prediction of genomic islands in prokaryotic genomes using hidden Markov models. BMC Bioinformatics 7: 142.
  35. 35. Darling ACE, Mau B, Blattner FR, Perna NT (2004) Mauve: multiple alignment of conserved genomic sequence with rearrangements. Genome Res 14: 1394–1403.
  36. 36. Darling AE, Mau B, Perna NT (2010) progressive Mauve: multiple genome alignment with gene gain, loss and rearrangement. PloS One 5: e11147.
  37. 37. Aziz RK, Bartels D, Best AA, DeJongh M, Disz T, et al. (2008) The RAST Server: rapid annotations using subsystems technology. BMC Genomics 9: 75.
  38. 38. Juhas M, van der MeerJR, Gaillard M, Harding RM, Hood DW, et al. (2009) Genomic islands: tools of bacterial horizontal gene transfer and evolution. FEMS Microbiol Rev 33: 376–393.
  39. 39. Fries M, Ihrig J, Brocklehurst K, Shevchik VE, Pickersgill RW (2007) Molecular basis of the activity of the phytopathogen pectin methylesterase. EMBO J 26: 3879–3887.
  40. 40. Cirl C, Wieser A, Yadav M, Duerr S, Schubert S, et al. (2008) Subversion of Toll-like receptor signaling by a unique family of bacterial Toll/interleukin-1 receptor domain-containing proteins. Nat Med 14: 399–406.
  41. 41. Boquet P (2000) Small GTP binding proteins and bacterial virulence. Microbes Infect 2: 837–843.
  42. 42. Lima-Mendez G, Van Helden J, Toussaint A, Leplae R (2008) Prophinder: a computational tool for prophage prediction in prokaryotic genomes. Bioinformatics 24: 863–865.
  43. 43. Hendrix RW, Lawrence JG, Hatfull GF, Casjens S (2000) The origins and ongoing evolution of viruses. Trends Microbiol 8: 504–508.
  44. 44. Brüssow H, Canchaya C, Hardt WD (2004) Phages and the evolution of bacterial pathogens: from genomic rearrangements to lysogenic conversion. Microbiol Mol Biol Rev 68: 560–602.
  45. 45. Figueroa-Bossi N, Bossi L (1999) Inducible prophages contribute to Salmonella virulence in mice. Mol Microbiol 33: 167–176.
  46. 46. Ho TD, Figueroa-Bossi N, Wang M, Uzzau S, Bossi L, et al. (2002) Identification of GtgE, a novel virulence factor encoded on the Gifsy-2 bacteriophage of Salmonella enterica serovar Typhimurium. J Bacteriol 184: 5234–5239.
  47. 47. Stanley TL, Ellermeier CD, Slauch JM (2000) Tissue-specific gene expression identifies a gene in the lysogenic phage Gifsy-1 that affects Salmonella enterica serovar Typhimurium survival in Peyer's patches. J Bacteriol 182: 4406–4413.
  48. 48. Seth-Smith HMB (2008) SPI-7: Salmonella's Vi-encoding Pathogenicity Island. J Infect Dev Ctries 2: 267–271.
  49. 49. Seth-Smith HMB, Fookes MC, Okoro CK, Baker S, Harris SR, et al. (2012) Structure, diversity and mobility of the SPI-7 family of integrative and conjugative elements within Enterobacteriaceae. J Bacteriol 194: 1494–504.
  50. 50. Fricke WF, McDermott PF, Mammel MK, Zhao S, Johnson TJ, et al. (2009) Antimicrobial resistance-conferring plasmids with similarity to virulence plasmids fromavian pathogenic Escherichia coli strains in Salmonella enterica serovar Kentucky isolates from poultry. Appl Environ Microbiol 75: 5963–5971.
  51. 51. Doumith M, Dhanji H, Ellington MJ, Hawkey P, Woodford N (2012) Characterization of plasmids encoding extended-spectrum β-lactamases and their addiction systems circulating among Escherichia coli clinical isolates in the UK. J Antimicrob Chemother 67: 878–885.
  52. 52. Pimenta AL, Racher K, Jamieson L, Blight MA, Holland IB (2005) Mutations in HlyD, part of the type 1 translocator for hemolysin secretion, affect the folding of the secreted toxin. J Bacteriol 187: 7471–7480.
  53. 53. Jenal U (2004) Cyclic di-guanosine-monophosphate comes of age: a novel secondary messenger involved in modulating cell surface structures in bacteria? Curr Opin Microbiol 7: 185–191.
  54. 54. Komano T, Kim SR, Yoshida T, Nisioka T (1994) DNA rearrangement of the shufflon determines recipient specificity in liquid mating of IncI1 plasmid R64. J Mol Biol 243: 6–9.
  55. 55. Takahashi H, Shao M, Furuya N, Komano T (2011) The genome sequence of the incompatibility group Iγ plasmid R621a: Evolution of IncI plasmids. Plasmid 66: 112–121.
  56. 56. Tracz DM, Tabor H, Jerome M, Ng L-K, Gilmour MW (2006) Genetic determinants and polymorphisms specific for human-adapted serovars of Salmonella enterica that cause enteric fever. J Clin Microbiol 44: 2007–2018.
  57. 57. Holt KE, Phan MD, Baker S, Duy PT, Nga TV, et al. (2011) Emergence of a globally dominant IncHI1 plasmid type associated with multiple drug resistant typhoid. PLoS Negl Trop Dis 5: e1245.
  58. 58. Holt KE, Thomson NR, Wain J, Phan MD, Nair S, et al. (2007) Multidrug-resistant Salmonella enterica serovar Paratyphi A harbors IncHI1 plasmids similar to those found in serovar Typhi. J Bacteriol 189: 4257–4264.
  59. 59. García-Fernández A, Fortini D, Veldman K, Mevius D, Carattoli A (2009) Characterization of plasmids harbouring qnrS1, qnrB2 and qnrB19 genes in Salmonella. J Antimicrob Chemother 63: 274–281.
  60. 60. Savard P, Gopinath R, Zhu W, Kitchel B, Rasheed JK, et al. (2011) The First NDM-positive Salmonella spp. identified in the United States. Antimicrob Agents Chemother 55: 5957–5958.
  61. 61. Kobayashi I (2001) Behavior of restriction-modification systems as selfish mobile elements and their impact on genome evolution. Nucleic Acids Res 29: 3742–3756.
  62. 62. Furuta Y, Abe K, Kobayashi I (2010) Genome comparison and context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements. Nucleic Acids Res 38: 2428–2443.
  63. 63. Bukowski M, Rojowska A, Wladyka B (2011) Prokaryotic toxin-antitoxin systems–the role in bacterial physiology and application in molecular biology. Acta Biochim Pol 58: 1–9.
  64. 64. Wang X, Wood TK (2011) Toxin/Antitoxin systems influence biofilm and persister cell formation and the general stress response. Appl Environ Microbiol 77: 5577–5583.
  65. 65. Brown JM, Shaw KJ (2003) A novel family of Escherichia coli toxin-antitoxin gene pairs. J Bacteriol 185: 6600–6608.
  66. 66. Sibley MH, Raleigh EA (2004) Cassette-like variation of restriction enzyme genes in Escherichia coli C and relatives. Nucleic Acids Res 32: 522–534.
  67. 67. Bishop AL, Baker S, Jenks S, Fookes M, Gaora PO, et al. (2005) Analysis of the hypervariable region of the Salmonella enterica genome associated with tRNA(leuX). J Bacteriol 187: 2469–2482.
  68. 68. Aktories K (2011) Bacterial protein toxins that modify host regulatory GTPases. Nat Rev Microbiol 9: 487–498.
  69. 69. Gough NR (2008) Bacterial TIR Domains. Sci Signal 1: ec129–ec129.
  70. 70. Römling U (2002) Molecular biology of cellulose production in bacteria. Res Microbiol 153: 205–212.
  71. 71. Cogez V, Gak E, Puskas A, Kaplan S, Bohin J-P (2002) The opgGIH and opgC genes of Rhodobacter sphaeroides form an operon that controls backbone synthesis and succinylation of osmoregulated periplasmic glucans. Euro J Biochem 269: 2473–2484.
  72. 72. Lee S, Cho E, Jung S (2009) Periplasmic glucans isolated from Proteobacteria. BMB Rep 42: 769–775.
  73. 73. Uchida I, Ishihara R, Tanaka K, Hata E, Makino S, et al. (2009) Salmonella enterica serotype Typhimurium DT104 ArtA-dependent modification of pertussis toxin-sensitive G proteins in the presence of [32P]NAD. Microbiology 155: 3710–3718.
  74. 74. McLaughlin LM, Govoni GR, Gerke C, Gopinath S, Peng K, et al. (2009) The Salmonella SPI2 effector SseI mediates long-term systemic infection by modulating host cell migration. PLoS Pathog 5: e1000671.
  75. 75. Casjens S (2003) Prophages and bacterial genomics: what have we learned so far? Mol Microbiol 49: 277–300.
  76. 76. Casjens SR, Gilcrease EB, Winn-stapley DA, Schicklmaier P, Schmieger H, et al. (2005) The generalized transducing Salmonella bacteriophage ES18: complete genome sequence and DNA packaging strategy. J Bacteriol 187: 1091–1104.
  77. 77. Kropinski AM, Sulakvelidze A, Konczy P, Poppe C (2007) Salmonella phages and prophages–genomics and practical aspects. Methods Mol Biol 394: 133–175.
  78. 78. Morris C, Tam CKP, Wallis TS, Jones PW, Hackett J (2003) Salmonella enterica serovar Dublin strains which are Vi antigen-positive use type IVB pili for bacterial self-association and human intestinal cell entry. Microb Pathog 35: 279–284.
  79. 79. Zhang XL, Tsui IS, Yip CM, Fung aW, Wong DK, et al. (2000) Salmonella enterica serovar Typhi uses type IVB pili to enter human intestinal epithelial cells. Infect Immun 68: 3067–3073.
  80. 80. Parkhill J, Dougan G, James KD, Thomson NR, Pickard D, et al. (2001) Complete genome sequence of a multiple drug resistant Salmonella enterica serovar Typhi CT18. Nature 413: 848–852.
  81. 81. Deng W, Liou S-R, Plunkett G, Mayhew GF, Rose DJ, et al. (2003) Comparative genomics of Salmonella enterica serovar Typhi strains Ty2 and CT18. J Bacteriol 185: 2330–2337.
  82. 82. Townsend SM, Kramer NE, Edwards R, Baker S, Hamlin N, et al. (2001) Salmonella enterica serovar Typhi possesses a unique repertoire of fimbrial gene sequences. Infect Immun 69: 2894–2901.
  83. 83. Craig L, Pique ME, Tainer JA (2004) Type IV pilus structure and bacterial pathogenicity. Nat Rev Microbiol 2: 363–378.
  84. 84. Pickard D, Wain J, Baker S, Line A, Chohan S, et al. (2003) Composition , acquisition , and distribution of the Vi exopolysaccharide-encoding Salmonella enterica pathogenicity island SPI-7. J Bacteriol 185: 5055–5065.
  85. 85. Wozniak RAF, Waldor MK (2010) Integrative and conjugative elements: mosaic mobile genetic elements enabling dynamic lateral gene flow. Nat Rev Microbiol 8: 552–563.
  86. 86. Roberts AP, Chandler M, Courvalin P, Guédon G, Mullany P, et al. (2008) Revised nomenclature for transposable genetic elements. Plasmid 60: 167–173.
  87. 87. Stamatakis A (2006) RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22: 2688–2690.
  88. 88. Sullivan MJ, Petty NK, Beatson SA (2011) Easyfig: a genome comparison visualizer. Bioinformatics 27: 1009–1010.