Salmonella enterica is a Gram-negative pathogen that infects humans and animals. S. enterica is divided into six sub-species on the basis of genetic content (Wain and O’Grady, 2017), and contains more than 2,500 serovars. S. enterica subspecies enterica is a major cause of enteric disease in humans and animals. The majority of illnesses caused by S. enterica are foodborne. Globally, S. enterica causes 93 million gastroenteritis cases and 150,000 deaths annually (Majowicz et al., 2010). In Canada, salmonellosis accounts for 87,510 human cases, and 17 deaths each year (Thomas et al., 2013, 2015). However, while there are more than 2,500 serovars of S. enterica subspecies I, the majority (75%) of all salmonellosis cases in Canada are caused by only 10 serovars. And three serovars, S. Enteritidis (30%), S. Heidelberg (15%), and S. Typhimurium (12%), account for 57% of all salmonellosis cases in Canada (Public Health Agency of Canada [PHAC], 2012). In the United States, where Salmonella accounts for 1.2 million illnesses and 450 deaths annually (Scallan et al., 2011), the top 10 serovars cause approximately 57% of illnesses. Characterizing Salmonella serovars into monophyletic and polyphyletic lineages is essential for linking outbreaks (Timme et al., 2013). While the implementation of Hazard Analysis and Critical Control Point (HACCP) programs in the food industry has reduced contamination of foods of animal origin, there has been increased recognition of Salmonella contamination associated with fresh produce, which accounts for approximately half of all fresh produce outbreaks due to bacteria in the United States and European Union (Callejón et al., 2015).

These statistics have led to much work aimed at identifying virulence and fitness markers in S. enterica, as well as questions regarding genome diversity. S. enterica genomes are highly diversified due to insertions and deletions (indels) (Zhou et al., 2013). Survival in different habitats, as evidenced by the large number of host species colonized by S. enterica, and the ability to successfully transmit through food and water, or directly from host to host has driven this diversity. High levels of intra-serovar diversity has also been recognized, as demonstrated by the acquisition of indels (51 prophages, 10 plasmids, and 6 integrative conjugational elements) by S. enterica Agona (Zhou et al., 2013).

Prophages are bacteriophages which have integrated into bacterial chromosomes, by means of an integrase gene, and they have been found to contribute to interstrain genetic variability (Brüssow et al., 2004). Bacteriophages are the most abundant organisms on earth, and it is estimated that there are 10³¹ phage particles in the biosphere (McNair et al., 2012). Phages are ubiquitous and can be found in any environment where their bacterial hosts are present. It has been estimated that there are 100 million phage species (Rohwer, 2003). As such, phages likely play a major role in defining the dynamics of microbial community structure and function. In fact, much of the diversity observed in closely related bacterial strains is a result of the incorporation of diverse prophages into the core bacterial genome (Brüssow et al., 2004). Prophages enhance bacterial fitness by encoding many proteins important in virulence and antibiotic resistance.

Many studies have demonstrated the presence of numerous prophages within S. enterica (Figueroa-Bossi and Bossi, 1999; Kropinski et al., 2007). In one such study, Thomson et al. (2004) characterized prophages within the S. enterica serovar Typhi CT18 chromosome. In silico analyses were used to compare prophage regions in S. Typhi CT18, to prophages within 40 other Salmonella isolates using DNA microarray technology. The results indicated that the S. Typhi CT18 prophages had similarity to the lambda, Mu, P2 and P4 phage families. Other S. Typhi isolates also had similar prophages, supporting a clonal origin of this serovar. In contrast, distinct prophage variation was detected within a broad range of Salmonella serovars, suggesting that these phages may confer a level of specialization on their host. The authors concluded that prophages therefore play a crucial role in the generation of genetic diversity within S. enterica. This statement, and the lack of an universal phylogenetic marker for phages provides the rationale for this study in which we evaluated the use of prophage integrase genes as reliable indicators of genome diversity in S. enterica.

We compared two serovars (Enteritidis and Heidelberg) whose genomes are reported to be clonal, to a set of S. enterica isolates from diverse food sources (Table 1). These food associated isolates are considered to be “rare” because they belong to serovars that are not within the top 70 serovars that cause illness in Canada. We hypothesized that the integrase gene is associated with prophage diversity within Salmonella. Hence, screening integrase genes (via a PCR assay) would detect few and similar patterns of integrase genes in the non-diverse and clonal isolates of S. Enteritidis and S. Heidelberg but a larger number and diversity of integrase genes in the food associated isolates that are rare and diverse.

The study of phage diversity trails far behind similar studies of the bacterial and eukaryotic kingdoms, and only a small fraction of phages have so far been characterized (Adriaenssens and Cowan, 2014). This is largely due to the absence of any universal phylogenetic marker (or signature gene) for phages. Unlike bacteria, in which the 16S rRNA gene is a universal gene that can be used for taxonomy and phylogeny, phages have no universally present gene that can be used for taxonomic analysis (Rohwer and Edwards, 2002). Various signature genes have been used to investigate phage diversity, including genes encoding structural proteins (portal proteins, major capsid proteins, and tail sheath proteins), auxiliary metabolism genes (psbA, psbB, and phoH), and several polymerase genes (Adriaenssens and Cowan, 2014). The majority of work conducted on signature gene analysis to assess phage diversity has focused on virulent phages. With respect to diversity conferred on bacterial hosts by temperate phages, Verghese et al. (2011) investigated the use of comK prophage junction fragments as markers for Listeria monocytogenes genotypes that persisted in individual meat and poultry processing plants (Verghese et al., 2011). In this work, the authors demonstrated that sequences in comK prophage junction fragments could be used to differentiate strains of epidemic clones (ECs), which, when identified, were shown to be specific to individual meat and poultry processing plants. The authors concluded that comK prophage junction fragment sequences may permit accurate tracking of persistent strains within individual food processing operations and thus allow the design of more effective intervention strategies to reduce contamination and enhance food safety (Verghese et al., 2011).

The absence of a universal marker for phages and the fact that that signature genes are only specific for certain phages or phage species, limits the use of such approaches to related phages, meaning that these approaches are not useful when assessing phages that are genetically diverse. Hence, in this work, we used the prophage integrase gene as a signature gene to measure prophage diversity, and therefore the diversity of host genomes that carry the prophages. Many factors such as pathogenicity islands, virulence factors, and antimicrobial and heavy metal resistance genes are encoded by prophages and provide genetic diversity among bacterial species (Brüssow et al., 2004). We therefore hypothesized that prophages may contribute to the diversity in the Salmonella isolates from rare serovars, and a lack of diversity in the clonal (clinical) serovars. The tyrosine integrase gene can be used as an indicator for phage diversity because it is carried by over 300 prophages (Balding et al., 2005). Although prophages can carry serine integrase genes, only approximately 30 members have been identified and they are not typically carried by foodborne bacteria (Fogg et al., 2014). The C-terminal of the tyrosine integrase contains two highly conserved regions designated “box 1” and “box II.” The tyrosine residue is located at residue 342 in “box II” and is conserved in every family member encoding a tyrosine integrase and is responsible for the DNA cleavage on the bacterial attachment site (attB) and the phage attachment site (attP), which facilitate the integration of the prophage into its host (Balding et al., 2005). In addition, two arginine residues R212 and R311 that assist in the facilitation of prophage integration are also conserved in all family members. Bobay et al. (2013) investigated the number and type of integration loci of prophages within E. coli and S. enterica genomes. They observed that the most prevalent (46%) integration site flanking prophages in S. enterica was a protein coding sequence, lepA, followed by tRNA and tmRNA genes (37%) and sRNA genes (17%). Of 24 prophage integration sites observed within S. enterica, 19 were also present in E. coli, suggesting that prophage integration sites are restricted to a few bacterial sites. The authors also observed that 423 of 500 prophages (83%) contained an integrase, of which all were tyrosine integrases, and a phylogenetic analysis demonstrated that closely related integrases integrate at the same integration sites. Since phages restrict their integration to conserved sites on the host genome, this further supports our findings that the integrase gene can serve as an indicator of prophage diversity (Bobay et al., 2013).

Core genome SNP analysis of the Salmonella isolates demonstrated a high degree of clonality in the S. Enteritidis and S. Heidelberg isolates (Figure 1), in agreement with previous studies, while also showing that the food associated Salmonella were highly diverse. Ogunremi et al. (2014) observed the clonality of 11 S. Enteritidis isolates, and demonstrated that they were all characterized by a comparably sized genome, an estimated 23-905 SNPs among genome pairs and five prophages or prophage remnants. Hoffmann et al. (2014) observed the clonality of 44 S. Heidelberg isolates, in which nearly 30 were indistinguishable by pulsed-field gel electrophoresis (PFGE), had significantly fewer core genome SNPs then S. Newport and, S. Typhimurium and contained similar prophages.

In this study, the results of the core genome SNP analysis agreed with a prophage integrase gene bioinformatic approach, in which prophages and integrase genes were identified within WGS using PHASTER, and used to construct unrooted maximum likelihood trees. The integrase gene was a good predictor of the diversity of the entire prophage sequence, as prophages that clustered together (i.e., Gifsy-2, HP2, and RE-2010) (Figure 2), contained integrase genes that also clustered together (Figure 4). Gifsy-2 is a highly studied phage known to integrate into various strains of S. Typhimurium. HP2 is a P2-like phage known to infect Salmonella spp., while RE-2010 has high nucleotide homology to Fels-2 prophages that have been observed in various Salmonella genomes (Switt et al., 2015). Also, both the entire prophage sequences and the integrase gene sequences were good predictors of Salmonella core genome diversity, as only four intact prophages and their integrase genes were identified in the clonal Salmonella (Enteritidis and Heidelberg) serovars, while 29 different types of prophages and 23 different integrase genes, were identified in the food associated Salmonella isolates. These results agreed with the core genome SNP analysis of the Salmonella isolates, in which similar results regarding the genome diversity of the clinical and food associated isolates were observed.

As with the bioinformatics approach, the PCR assay detected more integrase genes in the food associated isolates than the clinical isolates. Additionally, when the PCR assay was used to evaluate the Salmonella isolates for the presence of integrase genes, the results showed that the PCR assay detected 147 integrases in the 49 Salmonella isolates, while 74 integrases were detected by PHASTER in the intact prophages encoded by the Salmonella isolates (Figure 5). The differential ability of the two methods to detect prophage integrase genes is due to the fact that the primer sets used in the PCR assay detected the integrase genes of intact and cryptic phages, while our bioinformatic analysis focused only on the intact phages detected by PHASTER. The prophages encoded by the Salmonella isolates were identified as intact, questionable or incomplete by PHASTER. Incomplete phages suggest that they may represent cryptic phages, which may no longer encode an integrase gene. Questionable phages do not contain sufficient prophage genes to be considered complete functional phages. Thus, only intact phages were selected and blasted for confirmation to ensure an accurate multiple sequence alignment.

Other groups have demonstrated the use of prophages as markers of diversity in their bacterial hosts. For example, (Shan et al., 2012) conducted phylogenetic analysis of the holin sequences of Clostridium difficile prophages, and identified three groups of C. difficile phages, two within the Myoviridae and a divergent group within the Siphoviridae. The marker also produced homogenous groups within temperate phages that infect other taxa, including Clostridium perfringens, Clostridium botulinum, and Bacillus spp., indicating the potential use of the holin gene to study prophage carriage in other bacteria. This study demonstrated the high incidence of prophage carriage and diversity in clinically relevant strains of C. difficile.

Basu et al. (2000) conducted a molecular analysis of CTX prophages in clinical, classical biotype strains of Vibrio cholerae O1 that were isolated between 1970 and 1979 (Basu et al., 2000). Restriction fragment length polymorphism (RFLP) of rRNA genes and PFGE showed clonal diversity among the strains. The authors observed that one strain (GP13) had three CTX prophages while another (GP147) had four CTX prophages, indicating heterogeneity in the arrangement of the CTX prophages among classical strains of V. cholerae O1.

Collectively, our data shows that the prophage integrase PCR assay may be a good indicator of genome diversity in S. enterica, and that the PCR assay is a rapid and cost-effective rapid screening tool that may be used as a high throughput screen to evaluate large numbers of Salmonella isolates as a way to reduce the numbers of isolates that are submitted for whole genome sequencing to evaluate genomic diversity.

SB and ST supplied bacterial isolates and AC, YD, ET, LG, J-GE-R, JJ, LF, IK-I, and RL performed the analyses and drafted the manuscript. BB provided support for sequencing and analysis. All authors revised the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The handling Editor declared a shared affiliation, though no other collaboration, with one of the authors ST, and the handling Editor states that the process met the standards of a fair and objective review.