Virulence Characterisation of Salmonella enterica Isolates of Differing Antimicrobial Resistance Recovered from UK Livestock and Imported Meat Samples

Salmonella enterica is a foodborne zoonotic pathogen of significant public health concern. We have characterized the virulence and antimicrobial resistance gene content of 95 Salmonella isolates from 11 serovars by DNA microarray recovered from UK livestock or imported meat. Genes encoding resistance to sulphonamides (sul1, sul2), tetracycline [tet(A), tet(B)], streptomycin (strA, strB), aminoglycoside (aadA1, aadA2), beta-lactam (blaTEM), and trimethoprim (dfrA17) were common. Virulence gene content differed between serovars; S. Typhimurium formed two subclades based on virulence plasmid presence. Thirteen isolates were selected by their virulence profile for pathotyping using the Galleria mellonella pathogenesis model. Infection with a chicken invasive S. Enteritidis or S. Gallinarum isolate, a multidrug resistant S. Kentucky, or a S. Typhimurium DT104 isolate resulted in high mortality of the larvae; notably presence of the virulence plasmid in S. Typhimurium was not associated with increased larvae mortality. Histopathological examination showed that infection caused severe damage to the Galleria gut structure. Enumeration of intracellular bacteria in the larvae 24 h post-infection showed increases of up to 7 log above the initial inoculum and transmission electron microscopy (TEM) showed bacterial replication in the haemolymph. TEM also revealed the presence of vacuoles containing bacteria in the haemocytes, similar to Salmonella containing vacuoles observed in mammalian macrophages; although there was no evidence from our work of bacterial replication within vacuoles. This work shows that microarrays can be used for rapid virulence genotyping of S. enterica and that the Galleria animal model replicates some aspects of Salmonella infection in mammals. These procedures can be used to help inform on the pathogenicity of isolates that may be antibiotic resistant and have scope to aid the assessment of their potential public and animal health risk.

Salmonella enterica is a foodborne zoonotic pathogen of significant public health concern. We have characterized the virulence and antimicrobial resistance gene content of 95 Salmonella isolates from 11 serovars by DNA microarray recovered from UK livestock or imported meat. Genes encoding resistance to sulphonamides (sul1, sul2), tetracycline [tet(A), tet(B)], streptomycin (strA, strB), aminoglycoside (aadA1, aadA2), beta-lactam (bla TEM ), and trimethoprim (dfrA17) were common. Virulence gene content differed between serovars; S. Typhimurium formed two subclades based on virulence plasmid presence. Thirteen isolates were selected by their virulence profile for pathotyping using the Galleria mellonella pathogenesis model. Infection with a chicken invasive S. Enteritidis or S. Gallinarum isolate, a multidrug resistant S. Kentucky, or a S. Typhimurium DT104 isolate resulted in high mortality of the larvae; notably presence of the virulence plasmid in S. Typhimurium was not associated with increased larvae mortality. Histopathological examination showed that infection caused severe damage to the Galleria gut structure. Enumeration of intracellular bacteria in the larvae 24 h postinfection showed increases of up to 7 log above the initial inoculum and transmission electron microscopy (TEM) showed bacterial replication in the haemolymph. TEM also revealed the presence of vacuoles containing bacteria in the haemocytes, similar to Salmonella containing vacuoles observed in mammalian macrophages; although there was no evidence from our work of bacterial replication within vacuoles. This work shows that microarrays can be used for rapid virulence genotyping of S. enterica and that the Galleria animal model replicates some aspects of Salmonella infection in mammals. These procedures can be used to help inform on the pathogenicity of isolates that may be antibiotic resistant and have scope to aid the assessment of their potential public and animal health risk.

INTRODUCTION
Salmonella enterica is a Gram-negative bacterium of significant public and animal health concern worldwide (Majowicz et al., 2010). The species is highly diverse, consisting of over 2,600 serovars (Grimont and Weill, 2007) which can be divided into typhoidal and non-typhoidal Salmonella (NTS) serovars. The typhoidal serovars comprise Salmonella serovar Typhi, Salmonella serovar Sendai, and Salmonella serovar Paratyphi A, B (excluding variant Java), or C that are highly adapted to the human host, in which they cause enteric fever (Gal-Mor et al., 2014). NTS comprise the majority of serovars, some of which are adapted to non-human hosts, e.g., Salmonella serovar Gallinarum in poultry and Salmonella serovar Dublin in cattle; while others have broad host ranges, e.g., Salmonella serovar Typhimurium and Salmonella serovar Enteritidis (Uzzau et al., 2000). Many NTS are zoonotic and human infections usually arise via a foodborne route (Majowicz et al., 2010), commonly resulting in a self-limiting gastroenteritis but in approximately 3-10% of cases the infection becomes invasive (Hohmann, 2001) resulting in a serious and potentially fatal condition that often requires antimicrobial treatment (Okeke et al., 2005).
The pathogenic capacity of different NTS serovars is dependent upon the virulence potential of the bacterium and the susceptibility of the host. S. enterica harbor a diverse assortment of virulence genes required for adhesion, invasion, intra-cellular survival, and replication. These genes are located on various elements of the genome including the chromosome, plasmids, integrated bacteriophage DNA, Salmonella pathogenicity islands (SPIs), and Salmonella genomic islands (SGIs). Some of these elements are conserved throughout the species, such as SPI-1 (encoding factors important for cell adhesion) and SPI-2 (encoding factors required for intracellular survival and replication); while other elements are serovar specific, e.g., SPI-7 in S. Typhi (Anjum et al., 2005). Salmonella virulence plasmids have also been identified in several serovars, including S. Typhimurium and S. Enteritidis, and, although they can vary in genetic content, all carry the spv operon which contributes to pathogenesis by promoting intra-macrophage survival (Gulig et al., 1998;Feng et al., 2012;Haneda et al., 2012). The presence of other plasmids harboring antimicrobial resistance (AMR) genes have been described in many serovars (Hopkins et al., 2007;Anjum et al., 2011). Many of these plasmids are readily transmissible between bacteria via conjugation and their presence limits the choice of antimicrobial treatments suitable for treatment. Therefore tools that can easily determine the resistance profile and inform on the potential for pathogenicity of Salmonella isolated from food animals or food products can inform on the possible risk posed to public or animal health.
The aim of this study was to screen 95 S. enterica isolates recovered from UK livestock or imported meat, and representing 11 NTS serovars, for the presence of virulence and AMR genes by DNA microarray Figueiredo et al., 2015) and to investigate the virulence potential of selected isolates in an insect model. Isolates included S. Typhimurium from different host animals, poultry-invasive isolates of S. Enteritidis and S. Gallinarum, and a multidrug resistant Salmonella serovar Kentucky of Sequence Type 198 which was highly ciprofloxacin resistant and has been epidemiologically associated with outbreaks in poultry and humans in Europe and Africa (Le Hello et al., 2013). The virulence pathotype of 13 isolates was investigated using the Galleria mellonella (greater wax moth) pathogenesis model, which we have described previously (Kirchner et al., 2013;Figueiredo et al., 2015). This model has been used to assess virulence of several bacterial species including S. enterica (Bender et al., 2013;Viegas et al., 2013;Figueiredo et al., 2015), Campylobacter jejuni (Champion et al., 2010;Senior et al., 2011), Escherichia coli (Kirchner et al., 2013), and Legionella pneumophila (Harding et al., 2012). A good correspondence with mammalian models has been established for G. mellonella and its innate immune response has many similarities to that of vertebrates, including the presence of haemocytes that play a similar role to that of macrophages in vertebrates and circulate in the insect haemolymph (Kavanagh and Reeves, 2004). The haemolymph fills the insect body cavity and functions analogously to blood in mammals. In this study we also used light and electron microscopy to examine the histopathology in the larvae following Salmonella infection.

Isolate culture, DNA Extraction, and Microarray
The isolates were recovered from storage on Dorset egg medium by culture on blood agar for ∼24 h at 37 • C. For DNA extraction, the isolates were cultured overnight at 37 • C on Luria-Bertani (LB) agar. A 10 µl loop of bacteria was lysed and DNA quantified using a nanodrop apparatus, as described previously (Card et al., 2013). For testing on the microarray the DNA was labeled in a linear multiplex reaction using the AMR and virulence gene primers described previously Figueiredo et al., 2015). For the microarray hybridization procedures, the HybPlus Kit (Alere Technologies, Jena, Germany) was used according to the manufacturer's instructions with an adapted protocol described previously . Microarray signals were detected with the ArrayMate device (Alere Technologies) using IconoClust software (Standard version; Alere Technologies). Mean signal intensities of two replicate spots per probe were used for analysis. For the AMR gene probes intensities ≥0.6 were considered present and for the virulence gene probes mean signal intensities ≥0.5 were considered present. The microarray data were converted to a binary format whereby 1 indicated gene presence and 0 indicated gene absence. Strains were clustered based on their virulence determinants using the Jaccard similarity coefficient and UPGMA cluster analysis method with Bionumerics software (Version 6.6, Applied Maths, Sint-Martens-Latem, Belgium) to generate a dendrogram.

Galleria mellonella Pathogenesis Model
To determine an effective infectious dose, strain LT2 and the S. Typhimurium field isolate S03659-10 were used. Bacterial strains were cultured for approximately 18 h in LB broth with shaking at 37 • C. Cells were harvested to give a final concentration of 10 8 CFU/ml in 0.1 M phosphate-buffered-saline (PBS) (pH 7.2) and were further diluted in PBS in 10-fold dilution series to 10 4 CFU/ml. For each strain, ten G. mellonella larvae were injected with 10 µl of bacterial suspension, from 10 2 to 10 6 CFU, using a previously described micro-injection technique (Champion et al., 2010). Larvae were incubated at 37 • C and assessed for survival at 24 h post-infection (hpi). Larvae were assessed as 'dead' if they displayed no movement in response to touch. The inoculating dose was verified by plating on LB agar. At 24 hpi S. enterica was enumerated in three larvae for each isolate as described previously (Kirchner et al., 2013). A further eleven S. enterica strains (see Supplementary  Table S1 and Results) were tested in the model using a single dose of 10 2 CFU per larvae, with subsequent verification of the inoculating dose. Each isolate was tested with 10 larvae in three separate experiments and every experiment included, as controls, uninfected and PBS-injected larvae. Survival data were plotted and standard deviations calculated using GraphPad Prism (version 6.04) (GraphPad Software Inc, San Diego, CA, USA). Differences in survival following infection was assessed for significance (P < 0.05) using the Welch-corrected 2-tailed t-test.

Time course, Histopathogy, and Transmission Electron Microscopy Following Infection of Galleria mellonella
A time-course study was performed to examine the progress of Salmonella infection in G. mellonella and to investigate the ensuing histopathology. Four isolates (see Supplementary  Table S1 and Results) were selected for this based on results obtained from the microarray and the G. mellonella model. The time course ran for 24 h with samples collected at 0, 5, 20, and 24 hpi. At each time point five larvae were taken for haemolymph extraction and three for formalin fixation. Larvae were chilled at 4 • C for at least 2 h prior to fixation in formalin for 1 week at room temperature. Each larva was cut longitudinally and both parts were embedded in paraffin. Four micron thick histological sections were stained with haematoxylin and eosin or Gram-stain, as previously described (Senior et al., 2011), and examined under a light microscope (Nikon 80i) and photographed using a Nikon Ds-Ri1 digital camera attached to a PC equipped with Nikon NIS-Br imaging software. For each Salmonella isolate tested the extent of degeneration of the gut structure in the larvae was recorded as: no change/normal (scored as zero), moderate (scored as 1), or severe (scored as 2) and plotted as the aggregate score obtained from the three formalin fixed larvae at each timepoint.
Haemolymph was extracted from five larvae at each time point, pooled and 500 µl 3% glutaraldehyde added, mixed, and transferred to a 1 ml sterile tube. The sample was centrifuged at 1180 g for 3 min and the pellet re-suspended in 200 µl of 3% glutaraldehyde. After fixation for at least 24 h the pellets were washed in 0.1 M PBS, post-fixed in 1% osmium tetroxide, dehydrated through a gradual series of alcohol up to 100% alcohol and placed in propylene oxide prior to embedding in araldite resin. The resin was polymerised at 60 • C for 48 h. One-micron sections, stained with toludine blue were prepared for light microscopy examination. Areas observed containing bacteria and cells were selected for ultrastructural examination. Ultrathin sections at 70-90 nm thickness were then prepared onto copper grids using a diamond knife, contrasted with uranyl acetate and lead citrate prior to examination using a Tecnai 12 BioTwin (FEI) transmission electron microscope (TEM).
A total of 105 Salmonella virulence genes were represented on the microarray (Figueiredo et al., 2015) and 40 genes were present in all isolates, two genes were not detected in any isolate while the remaining 63 were variably present (Supplementary  Table S1). Strains were clustered based on the presence/absence of these virulence determinants and they grouped largely according to their serogroup and serovar, forming separate branches in the dendogram (Figure 1). The S. Typhimurium isolates clustered into two subclades within their branch, according to presence or absence of plasmid virulence genes (Figure 1). Thus three broad virulence genotype groupings of S. Typhimurium were identified: isolates without the virulence plasmid (n = 36), isolates with the virulence plasmid (n = 8), and an isolate with both the plasmid and SGI-1 (n = 1). Virulence plasmid genes were also detected in group D S. enterica isolates.
Chromosomal and islet genes were highly conserved amongst isolates (Figure 1; Supplementary Table S1). Some genes were present in a specific serovar, e.g., lpfD (encoding a fimbrial protein) was present in S. Infantis, S. Typhimurium, and serogroup O:9 (D1) isolates. As expected, there was a high degree of conservation of genes located on SPI-1, SPI-2, SPI-3, SPI-4, and SPI-5, and most genes were present in the majority of isolates. The vexE gene from S. typhi specific SPI-7 was absent in all isolates. The SPI-10 genes sefA and sefR were only detected in group D Salmonella isolates. The cdtB gene, encoding a cytolethal distending toxin, was present in the S. Montevideo isolates only.

Mortality of Galleria mellonella Following Infection with S. enterica: Dose and Isolate Dependence
To assess the potential differences in the virulence pathotype of isolates a G. mellonella pathogenisis model was used. Dose-dependent mortality of larvae was observed and at 10 2 CFU/larva there was a significant difference (P < 0.018) in the survival of larvae infected with S03659-10 (mean survival 93.3%; range 80-100%) compared to those infected with LT2 (mean survival 56.7%; range 50-70%; Figure 2A). Infection with higher CFU/larva resulted in higher, up to 100%, larval mortality rates (Figure 2A). For both strains, enumeration of the intracellular bacteria in larvae infected with 10 2 CFU showed large increases, with up to a seven log increase in bacterial numbers over the inoculum (Supplementary Table S2). An additional 11 isolates were subsequently tested in the G. mellonella pathogenesis model at an infectious dose of 10 2 CFU/larva and scored for larval survival ( Figure 2B). Seven of these isolates represented the three S. Typhimurium virulotypes identified by microarray: three with the virulence plasmid (S05303-11, S06298-12, S02385-12); three without the plasmid (S05981-10, S01762-12, S04904-11); and one with the virulence plasmid and SGI-1 (S03099-12). Also included were four non-S. Typhimurium isolates: one S. Montevideo (L00776-10); a multidrug resistant S. Kentucky (S02107-12); and two poultry invasive isolates (S. Enteritidis, S01079-10; S. Gallinarum, S05886-06).
The mean G. mellonella survival rate differed between the 13 isolates and a significant difference (P < 0.05) was obtained for six strains compared to the uninfected control: the four non-S. Typhimurium isolates and two S. Typhimurium isolates (LT2 and S03099-12). A mean survival rate of 30% was obtained following infection with S. Typhimurium isolate S03099-12% (range 10-40%) and with the S. Enteritidis isolate (range 20-40%). The S. Montevideo isolate, carrying the cdtB gene, resulted in the lowest survival rate for Galleria (mean 26.7%; range 20-30%). The mean survival rates of the remaining seven S. Typhimurium isolates did not differ significantly from the uninfected control, and these included three with virulence plasmid genes and four without. There were 11 virulence genes variably present in the nine S. Typhimurium isolates tested in the G. mellonella model. The effector sopE1 was only present in 3/7 S. Typhimurium isolates with a low pathogenicity phenotype but there were no other virulence genes uniquely present in this group. LT2 was the only isolate harboring sodCIII, whereas only S03099-12 carried hldD_DT104 (a marker for DT104) and STM14_1441 (encoding a putative transcriptional regulator).

Histopathology of G. mellonella Infected with S. enterica: Gut Degradation and the Formation of Salmonella-Containing Vacuoles
The progress of S. enterica infection in G. mellonella and the ensuing histopathological changes were examined for four isolates over a 24 h period: S. Montevideo L00776-10 (with the lowest survival rate), S. Typhimurium S03659-10 (without virulence plasmid), LT2 (with virulence plasmid), and S. Typhimurium S03099-12 (with virulence plasmid and SGI-1). Bacteria were observed in the haemolymph by light microscopy at 5, 20, and 24 hpi and in close proximity to haemocytes in larvae infected with each isolate (Figure 3). Following infection there was moderate deterioration in the structure of the larval gut at 5 h for LT2, L00776-10 and S03099-12 (Supplementary Figure S1), characterized by degenerating epithelial cells which were sloughed from the gut wall and filled the gut lumen (Figure 3). Moderate degeneration was also noted in two larvae at 0 h post-infection with L00776-10 (Supplementary Figure S1). At 20 h severe gut degradation, characterized by complete destruction of the intestinal epithelium with necrotic cells and cellular debris present in the lumen (Figure 3), was noted in one larva infected with LT2, two larvae infected with L00776-10, two larvae infected with S03659-10 and all three larvae infected with S03099-12. At 24 h severe gut degradation was observed in all Frontiers in Microbiology | www.frontiersin.org FIGURE 1 | Virulence determinant microarray data for 95 Salmonella enterica field isolates and reference strain LT2. Genes are grouped by genomic location along the top and the hybridization result of each isolate is shown by row. The order of strains represents their relatedness according to UPGMA cluster analysis using Jaccard similarity coefficient performed in BioNumerics 6.6. A white box indicates the absence and a black box indicates the presence of the target sequence in the strain.
Frontiers in Microbiology | www.frontiersin.org FIGURE 2 | (A) Dose-dependent killing of G. mellonella after S. Typhimurium infection. Strain LT2 (with virulence plasmid; white bar) and isolate S03659-10 (without virulence plasmid; black bar) were inoculated with decreasing bacterial loads ranging from 10 6 to 10 2 CFU and larvae were scored for survival at 24 h post-infection. PBS injected and non-infected controls are also shown. Data as shown are mean results of three independent experiments and error bars represent standard deviation. (B) G. mellonella survival following infection with different S. enterica isolates. Isolates tested comprised four S. Typhimurium without the virulence plasmid (black bars); four S. Typhimurium with the virulence plasmid (white bars); one S. Typhimurium with the virulence plasmid and SGI-1 (horizontal stripes); and four non-S. Typhimurium isolates as indicated (hatched bars). Isolates with a significant difference (P < 0.05) in larval survival compared to the uninfected control are indicated with an asterisk. Data as shown are mean results of three independent experiments and error bars represent standard deviation.
TEM showed the presence of haemocytes and dividing bacterial cells in the haemolymph at 20 and 24 hpi ( Figure 4A). Furthermore, haemocytes containing one or more small discreet vacuoles in which one (or rarely two) bacteria resided were observed by TEM at 20 or 24 hpi (Figures 4B,C) in larvae infected with S03659-10, S03099-12, or L00776-10, but not with LT2. These intracellular bodies may be Salmonella containing vacuoles (SCV), but it is possible that they were phagolysosomes located within granular haemocytes, as evidence of intra-cellular bacterial cell division or replication was not observed. However, granular haemocytes observed by TEM had an appearance that was distinct from that of haemocytes containing SCV ( Figure 4C).

DISCUSSION
In this study we have characterized virulence and AMR genes present in 95 UK S. enterica field isolates from 11 serovars.
The diversity of AMR genes detected was similar to that reported previously for S. enterica isolates from UK livestock (Liebana et al., 2004;Randall et al., 2004;Anjum et al., 2011) and one isolate harbored SGI-1 which is commonly found in S. Typhimurium DT104 (Briggs and Fratamico, 1999). Carriage of large numbers of resistance genes (<9) was limited to S. Typhimurium mainly from pigs in which multidrug resistance is common (Anjum et al., 2011), and to the S. Kentucky isolate from turkey. Carriage of conserved sets of AMR genes associated with mobile genetic elements such as plasmids and integrons was noted in many isolates. For example, strB, bla TEM , sul2, and tet(B), which generally reside together on a chromosomal island (Hopkins et al., 2010) but have also been described on conjugable elements (Anjum et al., 2011), were detected in 25 S. Typhimurium isolates. The genes aadA1, cmlA1, dfrA12, and sul3 often reside together in integrons located on plasmids (Garcia et al., 2014) and were present in 6 S. Typhimurium isolates and the S. Kentucky isolate. Notably, isolate S. Typhimurium S01762-12 carried both of these sets of AMR genes. The two S. Paratyphi B var. Java isolates, recovered from imported chicken meat, carried aadA1, dfrA1, and the class 2 integron, which can be present following integration into the chromosome of poultry isolates (Miko et al., 2003). This carriage of genes encoding resistance to antibiotics used in clinical or veterinary medicine predicts that isolates would be recalcitrant to certain antibiotic therapies following infection. However, it is the presence of virulence determinants that help predict the pathogenic capacity of the isolates and we examined this capacity by virulence genotyping and pathotyping.
There was considerable conservation of the virulence gene content amongst isolates with SPI -1 to -5 and many chromosomal genes conserved in all isolates, as previously reported (Anjum et al., 2005;Litrup et al., 2010;Beutlich et al., 2011;Zou et al., 2011;Suez et al., 2013;Figueiredo et al., 2015). However, differences between serovars were noted, e.g., SPI-10 genes were only detected in group D Salmonella isolates, in concordance with other studies (Suez et al., 2013;Figueiredo et al., 2015). Intra-serovar differences in gene carriage were most notable for S. Typhimurium, chiefly arising from the presence/absence of virulence plasmid genes, the variable presence of which is well documented (Litrup et al., 2010;Suez et al., 2013;Figueiredo et al., 2015). Virulence genes residing on prophages also showed some variation between isolates, as noted by others (Litrup et al., 2010;Beutlich et al., 2011;Suez et al., 2013;Figueiredo et al., 2015), indicating that these elements make a noteworthy contribution to the plasticity of the S. enterica genome.
To explore functional differences in virulence gene content we employed the G. mellonella pathogenesis model and investigated the virulence pathotype of 13 isolates, selected to represent the three virulence genotypes of S. Typhimurium identified by microarray, and four serovars with animal or public health significance. We demonstrated a dose-dependent effect of S. enterica infection on the mortality of the larvae, consistent with other studies (Bender et al., 2013;Viegas et al., 2013). Infection with each of the four non-S. Typhimurium isolates resulted in high mortality of the larvae. These isolates included an egg-invasive S. Enteritidis (Arnold et al., 2014) and a S. Gallinarum isolate, a serovar that is highly invasive in chickens and can also infect parasitic red mites (Dermanyssus gallinae), a vector for blood-borne Salmonella transmission in poultry houses (Zeman et al., 1982). The S. Montevideo isolate harbored cdtB, which encodes a cytolethal distending toxin that functions in a tripartite complex to cause cell cycle arrest (Mezal et al., 2014). cdtB was first described in S. Typhi but subsequently identified in a number of NTS serovars including S. Montevideo (Suez et al., 2013;Mezal et al., 2014;Figueiredo et al., 2015). The presence of this toxin may have contributed to S. Montevideo virulence in G. mellonella, as was previously shown for an S. Typhimurium isolate harboring cdtB (Figueiredo et al., 2015). The significant pathogenicity of the multidrug resistant S. Kentucky isolate is notable as it was phenotypically and genotypically similar to epidemic isolates that have been reported in humans and poultry in Europe and Africa (Le Hello et al., 2013).
Infection of G. mellonella with a S. Typhimurium isolate harboring both the virulence plasmid and SGI-1 resulted in low survival rates which were significantly different to the uninfected control. SGI-1 has been reported to confer increased Salmonella virulence and invasiveness in vertebrate hosts (Carlson et al., 2007) and may make a similar contribution during infection of G. mellonella. S. Typhimurium LT2, a well characterized strain harboring several virulence determinants including the virulence plasmid and the Fels-1 prophage (McClelland et al., 2001), also produced significant mortality. This contrasts with the virulence phenotype of LT2 in experimentally infected mice, in which it is attenuated for systemic infection, e.g., in the liver and spleen (Swords et al., 1997;Wilmes-Riesenberg et al., 1997). Importantly, however, LT2 demonstrates several other virulence phenotypes, including the ability to adhere to and invade human epithelial (Wilmes-Riesenberg et al., 1997;Jolivet-Gougeon et al., 2003) and colon cells (Roberts et al., 2013;Figueiredo et al., 2015) in vitro, as well as survive and replicate within macrophages (Wilmes-Riesenberg et al., 1997). Additionally, Salmonella strains SL1344 and ATCC 14028, which are virulent in mice, also demonstrate virulence in Galleria (Bender et al., 2013;Viegas et al., 2013). Interestingly, infection with the remaining seven S. Typhimurium isolates, including three possessing the virulence plasmid, did not significantly increase larval mortality. Thus, although the virulence plasmid encodes functions that enable survival in macrophages and facilitate systemic infection (Gulig et al., 1998;Haneda et al., 2012), it may not function in insect haemocytes in the same manner and/or other virulence determinants (e.g., SGI-1, Fels-1, or others) may be required for extensive S. Typhimurium pathogenesis in this model.
Histopathological examination of infected larvae demonstrated that S. enterica replicates successfully in the haemolymph, correlating with the enumeration data that showed large increases in bacterial numbers 24 hpi. In vertebrates the formation of SCV in macrophages is an important stage for the onset of systemic infection. We have demonstrated the presence of SCV-like structures in G. mellonella following infection with S. enterica. The presence of SCVs indicates that the Salmonella virulence factors which enable adhesion and invasion, such as those encoded on SPI-1, function in G. mellonella; a hypothesis supported by previous work in which a S. Infantis isolate lacking 38 SPI-1 genes gave reduced mortality in this model (Figueiredo et al., 2015). We did not observe SCV following infection with LT2, but given the significant pathogenicity this strain demonstrated in the larvae it may have been due to insufficient sampling for TEM with this sample. The absence of evidence for S. enterica replication within the haemocytes may also have arisen from low sampling numbers. Alternatively, the Salmonella may have been unable to replicate as they were trapped in phagolysosomes located within granular haemocytes, although this is less probable as the haemocytes containing SCV-like structures were distinctly different to the granular haemocytes observed by TEM. It is also possible that virulence factors that promote intracellular replication and survival, such as those located on the virulence plasmid and SPI-2, do not function in haemocytes as they do in macrophages and warrants further investigation. However, haemocytes are able to support the replication in vacuoles of other bacteria, such as L. pneumophila (Harding et al., 2012).

CONCLUSION
We have used virulence and AMR genotyping to provide an assessment of the pathogenic potential of a variety of S. enterica serovars and isolates and their susceptibility to antibiotic therapy. Although there was no correlation between AMR carriage and virulence determinants present, it is pertinent to determine AMR gene carriage as it will impact therapeutic options. Thus the microarray continues to provide a rapid and easy-to-use epidemiological surveillance tool that could be used to help inform intervention measures and can be more accessible for many diagnostic laboratories than whole genome sequencing. We have shown that Salmonella isolates of clinical importance in humans and poultry can replicate and produce significant pathology in G. mellonella, indicating that this virulence model may replicate some aspects of Salmonella infection in mammals. However, as noted by Kavanagh and Reeves (2004) this model may not be suitable for the study of disease processes or microbial dissemination pathways that are specific to mammals. For example, the insect haemolymph fills the body cavity and may not provide the same barrier to systemic infection as the closed circulatory system of mammals. Furthermore confirmatory work, including a direct comparison of LT2 and a virulent strain such as SL1344 at the same infectious dose, is required before the Galleria model can be used to help predict the virulence potential of Salmonella for public or animal health. To our knowledge this study reports the first demonstration of SCV-like structures in G. mellonella and supports the further investigation of the potential of this model for preliminary investigation of the pathogenicity of S. enterica, due to the low cost and simplicity of the procedure.

FUNDING
This work was funded by the Department for Environment, Food & Rural Affairs. The funders had no role in study design, data collection, and interpretation.