Transcriptional Response of Salmonella enterica to Bacteriophage Treatments with Differential Multiplicities of Infection

: Salmonella enterica ( S. enterica ) is a causative agent of numerous foodborne outbreaks, as current industrial measures may be <90% effective. Therefore, bacteriophages have been suggested as an antimicrobial treatment against S. enterica , but it is currently unclear if there is an optimal bacteriophage multiplicity of infection (MOI) against S. enterica . Two bacteriophage cocktails at MOIs 1, 10, 100, 1000 and 10,000 were co-inoculated against four S. enterica strains ( S. Enteritidis, S. Newport, S. Muenchen and S. Typhimurium), and populations were estimated on days 0–3. The transcriptional profiles of 20 genes previously indicated to be differentially expressed after bacteriophage treatment were studied by extracting RNA from all four S. enterica strains after bacteriophage SE14, SF5 and SF6 treatment on days 0, 1 and 3, and RT-qPCR was conducted to determine the expression of the 20 selected genes. The results showed that an MOI of 1000 was the most optimal in reducing S. Enteritidis populations to undetectable levels from day 0 to 3. The cas1 (SOS response) and mod (DNA modification and recombination) genes were highly upregulated between 2.5-and 5-fold on day 0 for S. Enteritidis S5-483 and S. Typhimurium S5-536 at MOIs of 1000 and 10,000. On day 3, hsdS (DNA modification and recombination) was upregulated by ~1-fold for S. enteritidis S5-483 after an MOI of 1000. Understanding an optimal bacteriophage MOI can be beneficial to implementing effective and optimal bacteriophage treatments in the industry. Knowledge of S. enterica ’s transcriptional response after bacteriophage treatment provides further insight into how S. enterica can survive bacteriophage infection.


Introduction
From 2018 to 2023, Salmonella was the causative pathogen in 41 separate foodborne outbreaks in a multitude of food commodities in North America, including fresh produce; ready-to-eat meat; seafood; ground beef; chicken and chicken products; and more [1].Current antimicrobial methods to reduce Salmonella on food can be chemical, biological or non-thermal [2,3].Chemical antimicrobials can include sulphites, nitrites, organic acids, lysozymes, isothiocyanate and lactoferrin, and each can have metabolic targets in microorganisms such as genetic determinants or enzymes, the cell membrane or the cell wall [2].Biological antimicrobials can include bacteriocins, bacteriophages and lactic acid bacteria [2,3].Many chemical and biological antimicrobial methods can be used on farms and at the postharvest level, and there are also non-thermal methods that have been studied to reduce Salmonella in food: cold plasma, high-pressure processing, irradiation, ozone treatment and UV light treatment [3].
A steadily growing food trend among consumers in the past decade has been the consumption of organic foods and the use of non-chemical antimicrobial methods [4,5].The consumer trend regarding a desire for more natural foods has made the use of chemical antimicrobial methods undesirable, leaving only biological and postharvest antimicrobial methods.However, some biological antimicrobial methods, such as lactic acid bacteria, produce an acidic taste that may not be desirable to food products that are not acidic [6].
Biological methods such as bacteriophages are flavourless and odourless when added to food [7,8].Bacteriophages are cost-effective as they are abundant and easily located in the ecosystem [9].Bacteriophages are naturally occurring, can self-replicate, are generally non-toxic to humans and are highly specific to their bacterial targets [7,10].Since bacteriophages are highly specific, the understanding is that they will not alter the native food microbiota, will not inactivate desirable microorganisms and will not negatively affect food spoilage [10].
Bacteriophages have been shown in previous research to reduce Salmonella populations in a wide variety of food products.A cocktail of five bacteriophages reduced Salmonella populations in chicken meat by 1.5 and 1.2 log CFU/g after 48 h and 72 h at 10 • C, respectively [11].On postharvest lettuce and cantaloupe, a five-bacteriophage cocktail at a multiplicity of infection (MOI) of 1000 reduced Salmonella enterica (S. enterica) populations by 1-4 log CFU/cm 2 depending on the S. enterica strain.The authors of [12] used a single bacteriophage, SSP6, against S. Oranienburg on alfalfa seeds at an MOI of 70 and reduced Salmonella by 1 log CFU/g after 3 h.However, the inhibitory effect did not last, and there was a Salmonella population resurgence after 3 h [13].Similarly, Fong et al. [14] also used a single bacteriophage, SI1, against S. Enteritidis on sprouting alfalfa seeds at an MOI of 110 and reduced populations by ~2.5 log CFU/g immediately after bacteriophage treatment on day 1.However, there was also a population resurgence by day 6, where the bacteriophage-treated alfalfa seeds were <1 log CFU/g compared with the control [14].Given the Salmonella population resurgence after initial bacteriophage treatments shown in previous research, the authors of [15] implemented repeated daily bacteriophage applications at an MOI of 1000 against S. enterica on alfalfa sprouts during germination and found repeated applications to further reduce S. enterica populations, but the issue of population resurgence still arose [15].
Currently, it is unclear if there is an optimal bacteriophage MOI against S. enterica that can reduce or prevent population resurgence.In previous research, different studies have employed a range of different MOIs against Salmonella [12][13][14][15].However, to the best of our knowledge, there has not been a study that compares the efficacy of different bacteriophage MOIs.Therefore, the objective of this study is to (i) systemically determine if there is an optimal bacteriophage MOI against S. enterica for reducing or preventing population resurgence and (ii) to profile certain S. enterica genes expressed that may mechanistically provide an insight into the fate of S. enterica after bacteriophage treatment.

Bacteriophage Propagation and Titer Measurement
The bacteriophages used in this study were previously isolated in British Columbia, Canada, by [16].Four Salmonella bacteriophage isolates (SE14, SE20, SF5 and SF6) were selected based on their ability to lyse the S. enterica strains discussed in Section 2.2.The same S. enterica strains were also used for bacteriophage propagation.The bacteriophages were purified according to methods described by [14].S. enterica cultures were individually prepared with overnight incubation for 18 h at 37 • C in 10 mL of Tryptic Soy Broth (TSB, Difco, Becton Dickinson, Franklin Lakes, NJ, USA) under agitation at 175 rpm.From the overnight S. enterica cultures, 100 µL was inoculated into 10 mL of TSB and incubated at 37 • C under agitation at 175 rpm for 1.5 h until OD 600 was measured with a UV-1800 UV/Vis Spectrometer (Shimadzu, Columbia, MD, USA) to be between 0.2-0.4.In total, 50 µL of bacteriophage (SE14, SE20, SF5 or SF6) was pipetted into the inoculum at OD 600 0.2-0.4 and incubated at 37 • C for 6 h.After 6 h, the inoculum was poured into a 50 mL centrifuge tube (VWR International, Radnor, PA, USA) and sedimented at 4000× g for 10 min at 21 • C. The lysate was then poured into a separate centrifuge tube and filtered through a 0.4 µm filter (VWR International, Radnor, PA, USA) into a 15 mL centrifuge tube (VWR International, Radnor, PA, USA).A plaque assay was then used to measure the titers of the bacteriophages.The propagated bacteriophages were decimally diluted in 450 µL TSB (VWR International, Radnor, PA, USA).In total, 50 µL of a separate S. enterica overnight culture was prepared according to the methods provided above; added to each decimal dilution; and incubated for 1 h at 37 • C under agitation at 175 rpm.After 1 h, the contents were spread onto the surface of a Tryptic Soy Agar plate (TSA, Difco, Becton Dickinson, Franklin Lakes, NJ, USA).The plates were then incubated at 37 • C for 24 h prior to counting the plaques.

S. enterica Strain Storage Conditions
The four S. enterica strains used in this study were S. Enteritidis S5-483, S. Newport S5-639, S. Muenchen S5-504 and S. Typhimurium S5-536.The serovars of these strains are all serovars that are commonly implicated in S. enterica foodborne outbreaks [17].For longterm storage, the S. enterica strains were maintained at −80 • C in TSB, supplemented with 20% glycerol (VWR International, Radnor, PA, USA).The working stocks were maintained on TSA and stored for a maximum of 4 weeks at 4 • C.

S. enterica Growth Assay
For inoculation, each S. enterica strain was prepared with overnight incubation at 37 • C in 10 mL TSB under agitation at 175 rpm for 18 h.The overnight cultures were spun at 4000× g for 10 min, and the supernatant was decanted.The resulting pellets were washed twice with 10 mL of TSB and decimally diluted in TSB for a final concentration of 10 5 CFU/mL.In total, 1 mL of each S. enterica strain was then pipetted to each well in a 96-well dilution block (VWR International, PA, USA).In total, 1 mL of each of the two bacteriophage cocktails, (1) SE14, SE20 and SF6 and (2) SE14, SF5 and SF6, was separately added to each well at MOIs of 1, 10, 100, 1000 and 10,000.For the control, instead of a bacteriophage sample, 1 mL of TSB was pipetted into the wells with an S. enterica strain.The 96-well dilution blocks were incubated at 22 ± 1 • C from day 0 to 3. A temperature of 22 ± 1 • C was chosen to determine the efficacy of both bacteriophage cocktails at MOIs of 1, 10, 100, 1000 and 10,000 against S. enterica at a temperature that was within the range of ambient temperatures that food seeds and plants are commonly germinated and grown [18,19].On each day 0, 1, 2 and 3 samples from each well were pipetted into separate test tubes, serially diluted in TSB, plated on xylose lysine deoxycholate (XLD) agar (Oxoid, Nepean, ON, Canada) and incubated at 37 ± 1 • C for 24 h before population estimation.

RNA Extraction, Reverse-Transcription and Real-Time Quantitative PCR (RT-qPCR)
The preparation of the four S. enterica strains and inoculation with bacteriophage cocktails (1) SE14, SE20 and SF6 and (2) SE14, SF5 and SF6, were conducted as discussed in Section 2.3 with the exception that each S. enterica strain was decimally diluted in TSB for a final concentration of 10 9 CFU/mL.On days 0, 1 and 3, samples from each MOI and control were pelleted at 6000× g for 10 min and then preserved with RNAprotect Bacteria Reagent (Qiagen, Germantown, MD, USA).Total RNA was then isolated from S. enterica cells with the RNeasy Miniprep Kit (Qiagen, Germantown, MD, USA) by following the manufacturer's instructions.On-column Dnase digestion was performed with Rnase-Free Dnase I (Qiagen, Germantown, MD, USA).The quality and quantity of each extracted RNA sample was assessed with a Nanodrop 2000 Spectrophotometer (Thermo Fisher Scientific, Asheville, NC, USA) and an Agilent 2100 Bioanalyzer (Agilent, Santa Clara, CA, USA).Only RNA samples with an RNA Integrity Number (RIN) of >6.0 were reverse-transcribed to ensure the quantification of gene expression for RT-qPCR was reliable [20][21][22].The RNA samples were reverse-transcribed to cDNA with the LunaScript RT SuperMix Kit (New England Biolabs, Whitby, ON, Canada) in a Veriti 96-Well Thermal Cycler (Applied Biosystems, Waltham, MA, USA).The resulting cDNA was used for RT-qPCR analyses with the QuantiNova SYBR Green PCR Kit (Qiagen, Germantown, MD, USA) in a C1000 Touch Touch Thermal Cycler (Bio-Rad, Hercules, CA, USA).Three genes were used as reference genes, 16s, gapA and rsrnC, and the forward and reverse primers for each are shown in Table 1.Twenty genes were assessed for each MOI and control, and the primers for each target gene are listed in Table 2.The genes were chosen based on their function and have been shown to be differentially regulated in response to bacteriophage treatment from previous research listed in Table 2.The relative expression values of the genes were calculated using 2 −∆∆Ct .Melting curve analyses were performed to ensure that only a single gene was amplified for each target gene.

Statistical Analysis
Each treatment was conducted with six biological replicates.S. enterica populations in Section 2.3 were analyzed with log-transformed data using a one-way analysis of variance (one-way ANOVA), and Turkey's honestly significant difference (Turkey's HSD) was conducted for means separation with Rstudio, version 1.1463 (Rstudio, Inc., Boston, MA, USA).p-values of <0.05 were considered statistically significant.For visualization and comparison of gene expression patterns, heatmaps of the expression of 20 genes in Section 2.4 were created in Rstudio.The quantization of the expression of the 20 genes was conducted with K-means clustering in Rstudio.

MOI of 1000 Was the Optimal Bacteriophage Treatment against S. enterica
Two separate bacteriophage cocktails, (1) SE14, SE20 and SF6 and (2) SE14, SF5 and SF6, were co-inoculated at MOIs of 1, 10, 100, 1000 and 10,000 with four S. enterica strains, S. Enteritidis S5-483, S. Newport S5-639, S. Muenchen S5-504 and S. Typhimurium S5-536, in tryptic soy broth.For bacteriophage cocktail (1), SE14, SE20 and SF6, MOIs 100, 1000 and 10,000 were the most effective at significantly reducing S. Enteritidis S5-483 populations by ~2.9-3 log CFU/mL by day 1 compared with the control and MOI 1 (p < 0.05, Figure 1A).For S. Muenchen S5-504 and S. Newport S5-639, MOI 10,000 was the most effective MOI on day 0 with a significant population reduction of ~3 log CFU/mL and ~1.5 log CFU/mL compared with the control and MOI 1000, respectively (p < 0.05, Figure 1B,C).However, by day 1 for S. Newport S5-639, it was MOI 1000 that had the significantly highest population reduction of ~6 log CFU/mL compared with the control; ~3 log CFU/mL compared with MOI 10,000; and ~1 log CFU/mL compared with MOIs of 1, 10 and 100 (p < 0.05, Figure 1C).MOIs of 10, 100 and 1000 for bacteriophage cocktail SE14, SE20 and SF6 were the most effective against S. Muenchen S5-504 on day 1, where populations were significantly reduced to ~2.7 log CFU/mL compared with ~8.7 log CFU/mL for the control and ~5 log CFU/mL for MOIs of 1 and 10,000 (p < 0.05, Figure 1B).Both the S. Muenchen S5-504 and S. Newport S5-639 populations after the MOI 10,000 treatment had the fastest population resurgence compared with other bacteriophage MOI treatments by day 1 (Figure 1B,C).For S. Muenchen S5-504, the populations after the MOI 1 and 10,000 treatments on day 1 were not significantly different from each other (p > 0.05, Figure 1B).However, the population resurgence was slower for MOI 1 compared with MOI 10,000 for S. Muenchen S5-504 on day 1 because MOI 1 populations increased from ~4.5 log CFU/mL to ~5.1 log CFU/mL, but populations increased from ~1.8 CFU/mL to ~5.2 log CFU/mL for MOI 10,000 (Figure 1B).By day 3, the S. Muenchen S5-504 population after the MOI 1 and MOI 10,000 treatments were not significantly different from the control (p > 0.05, Figure 1B).This was unexpected because, at the highest bacteriophage MOI of 10,000, it was expected that surviving and new bacteriophage progenies would continue to reduce S. enterica populations on later days after inoculation, which unfortunately was not the case based on the population counts (Figure 1B).However, there is a theory called lysis from without, where there can be an overload of bacteriophages above the threshold (high MOI) attacking the bacteria simultaneously, which can lead to the bacteria lysing before new bacteriophage progeny can be created [44,45].With lysis from without, if new bacteriophage progenies are not created, then there is no bursting of the bacteria, and this could potentially decrease the efficacy of the high MOIs overloading the bacterial cell compared with lower MOIs [46].This is because new bacteriophage progeny from a burst can infect an average of five previously uninfected bacterial cells [46].Without the continued infection of bacteriophage progenies from a burst, this could be an explanation for why there can be faster S. enterica population resurgence with MOI 10,000 compared with lower MOIs, as shown with S. Muenchen S5-504 and S. Newport S5-639 with bacteriophage cocktails SE14, SE20 and SF6 and S. Enteritidis S5-483 and S. Typhimurium S5-536 with bacteriophage cocktails SE14, SF5 and SF6 (Figure 1B,C and Figure 2A,D).
The idea that the highest bacteriophage MOI may not always be the most effective is also evident with bacteriophage cocktail (2), SE14, SF5 and SF6.MOIs 10, 100, 1000 and 10,000 significantly reduced S. Enteritidis S5-483 populations to below the detection limit on day 0 (p < 0.05, Figure 2A).By day 1, MOI 10 had the fastest population resurgence followed by MOI 10,000 (Figure 2A).MOI 100 had a population resurgence by day 2, and it was only MOI 1000 that continued to have S. Enteritidis S5-483 populations below the detection limit up to day 3 (Figure 2A).For S. Muenchen S5-504 and S. Newport S5-639, MOI 10,000 further reduced populations significantly by ~1 log CFU/mL compared with MOIs 1, 10, 100 and 1000 and ~1.3-1.5 log CFU/mL compared with the control (p < 0.05, Figure 2B,C).However, by day 3, the S. Muenchen S5-504 and S. Newport S5-639 populations from all the MOI treatments were not significantly different from each other or the control (p > 0.05, Figure 2B,C).For S. Typhimurium S5-536, both MOI 1000 and 10,000 reduced populations to below the detection limit on day 0, but populations at an MOI of 1000 had a significantly slower population resurgence compared with MOI 10,000 and other MOIs on days 1 and 3 (p < 0.05, Figure 2D).
1B).This was unexpected because, at the highest bacteriophage MOI of 10,000, it was expected that surviving and new bacteriophage progenies would continue to reduce S. enterica populations on later days after inoculation, which unfortunately was not the case based on the population counts (Figure 1B).However, there is a theory called lysis from without, where there can be an overload of bacteriophages above the threshold (high MOI) attacking the bacteria simultaneously, which can lead to the bacteria lysing before new bacteriophage progeny can be created [44,45].With lysis from without, if new bacteriophage progenies are not created, then there is no bursting of the bacteria, and this could potentially decrease the efficacy of the high MOIs overloading the bacterial cell compared with lower MOIs [46].This is because new bacteriophage progeny from a burst can infect an average of five previously uninfected bacterial cells [46].Without the continued infection of bacteriophage progenies from a burst, this could be an explanation for why there can be faster S. enterica population resurgence with MOI 10,000 compared with lower MOIs, as shown with S. Muenchen S5-504 and S. Newport S5-639 with bacteriophage cocktails SE14, SE20 and SF6 and S. Enteritidis S5-483 and S. Typhimurium S5-536 with bacteriophage cocktails SE14, SF5 and SF6 (Figures 1B,C and 2A,D).Between the two different bacteriophage cocktails, (1) SE14, SE20 and SF6 and (2) SE14, SF5 and SF6, it has been seen that the highest MOI, 10,000, is not the most effective MOI, and a slight edge can be given to MOI 1000 as the MOI that was more effective at reducing S. enterica populations compared with other MOIs (Figures 1 and 2).For commercialized bacteriophages, SalmoFresh™ (a cocktail of six bacteriophages) against Salmonella gained recognition from the U.S. Food and Drug Administration in 2013 when it received GRAS (Generally Regarded as Safe) status for certain poultry products, fish, shellfish and processed fruits and vegetables at 10 7 PFU/g [47].In 2015, the U.S. Food and Drug Administration allowed SalmoFresh TM to be used on raw poultry at 10 7 PFU/g, and this allowance was expanded in 2019 for SalmoFresh TM to be used on ready-to-eat and raw red meat carcasses, subprimals and trimmings at 10 7 PFU/g [47].There was a further amendment to the GRAS status in 2021, where the bacteriophage titer was increased from 10 7 PFU/g to 10 8 PFU/g [47].Given the results of this study, it is important that the maximum approved bacteriophage titer of 10 8 PFU/g is only used when it is the most optimal MOI against S. enterica and not because of the notion that applying the highest titer of bacteriophages is the most effective against S. enterica.For food companies that can conduct microbiological testing for samples of their food products and environmental testing, there might be a preconception of the levels of pathogenic contamination, and bacteriophage titers can be adjusted to more optimal MOIs and possibly reduce antimicrobial costs.The idea that the highest bacteriophage MOI may not always be the most effective is also evident with bacteriophage cocktail (2), SE14, SF5 and SF6.MOIs 10, 100, 1000 and 10,000 significantly reduced S. Enteritidis S5-483 populations to below the detection limit on day 0 (p < 0.05, Figure 2A).By day 1, MOI 10 had the fastest population resurgence followed by MOI 10,000 (Figure 2A).MOI 100 had a population resurgence by day 2, and it was only MOI 1000 that continued to have S. Enteritidis S5-483 populations below the detection limit up to day 3 (Figure 2A).For S. Muenchen S5-504 and S. Newport S5-639, MOI 10,000 further reduced populations significantly by ~1 log CFU/mL compared with MOIs 1, 10, 100 and 1000 and ~1.3-1.5 log CFU/mL compared with the control (p < 0.05, Figure 2B,C).However, by day 3, the S. Muenchen S5-504 and S. Newport S5-639 populations from all the MOI treatments were not significantly different from each other or the control (p > 0.05, Figure 2B,C).For S. Typhimurium S5-536, both MOI 1000 and 10,000 reduced populations to below the detection limit on day 0, but populations at an MOI of 1000 had a significantly slower population resurgence compared with MOI 10,000 and other MOIs on days 1 and 3 (p < 0.05, Figure 2D).
Between the two different bacteriophage cocktails, (1) SE14, SE20 and SF6 and (2) SE14, SF5 and SF6, it has been seen that the highest MOI, 10,000, is not the most effective MOI, and a slight edge can be given to MOI 1000 as the MOI that was more effective at Besides the theory of lysis from without, to the best of our knowledge, there has not been any previous research conducted to compare the efficacy of different bacteriophage MOIs.Between the two different bacteriophage cocktails used, an MOI of 1000 for cocktail SE14, SF5 and SF6 was able to reduce and keep S. Enteritidis S5-483 populations below the detection limit throughout the 3 days (Figure 2A).On day 0 for S. Typhimurium S5-536, similar results were observed where both MOIs 1000 and 10,000 also reduced populations to below the detection limit (Figure 2D).In contrast, MOI 10,000 only significantly further reduced S. Muenchen S5-504 and S. Newport S5-639 populations by 1 log CFU/mL on day 0 (p < 0.05, Figure 2B,C).Given the contrasts between the S. enterica strains after bacteriophage cocktail (2), SE14, SF5 and SF6, treatment, gene expression was measured with this cocktail.
The high upregulation of the cas1 gene on day 0 for S. Enteritidis S5-483 and S. Typhimurium S5-536 after MOI 1000 and MOI 10,000 is not surprising.The cas1 gene codes for the CRISPR/Cas system-associated protein Cas1 and is an SOS gene [23].SOS genes are responsible for cell survival when faced with extensive DNA damage [23].When there is a viral infection, such as bacteriophage infection, S. enterica will upregulate CRISPRassociated genes such as cas1 to adapt its immune system [48].The immune system is adapted by acquiring a new sequence integrated into the CRISPR locus, which signals the assembly of Cas proteins into surveillance complexes that detect and degrade the viral target [48].The high upregulation of cas1 on day 0 for both S. Enteritidis S5-483 and S. Typhimurium S5-536 after MOI 1000 and MOI 10,000 is a defence mechanism meant to detect and degrade bacteriophages.This defence mechanism may have contributed to the population resurgence of S. Typhimurium S5-536 by day 1, as cas1 was highly upregulated in both MOI 1000 and MOI 10,000, and S. Typhimurium S5-536 populations on day 0 may have been below the detection limit, but both populations resurged by day 1 (Figure 2D).However, the populations of S. Enteritidis S5-483 after MOI 1000 did not resurge by day 1, but the S. Enteritidis S5-483 populations for MOI 10,000 did resurge (Figure 2A).Therefore, the effectiveness of the differential gene expression is very much dependent on the strain, treatment and/or time.
Another gene that was highly upregulated for both S. Enteritidis S5-483 and S. Typhimurium S5-536 on day 0 for MOI 1000 and MOI 10,000 was the mod gene (Figure 3).The mod gene, along with the res gene, codes for type III restriction and modification enzymes, Mod and Res, respectively [49].Restriction-modification genes are involved in protecting the modification of host DNA at specific recognition sites and interfering with bacteriophage DNA entering the cell [50].Both Mod and Res hemimethylate and cleave DNA, but the difference between the two is that Mod is required for both restriction and modification, and Res is only required for restriction [49,51].The res gene was highly upregulated in S. Typhimurium S5-536 for both MOI 1000 and MOI 10,000 and at an MOI of 10,000 for S. Enteritidis S5-483 but moderately downregulated for S. Enteritidis S5-483 on day 0 for MOI 1000 (Figure 3).The authors of [49] found that, when bacterial mutants were transcribing res but not transcribing mod, it was lethal to the mutant.This was similarly shown in the gene expression results from all four S. enterica strains across all MOI treatments and days where, when res was highly upregulated, mod was also highly upregulated (Figures 3-5).However, when mod was highly upregulated, res was not always highly upregulated (Figures 3-5).This supports the idea that mod and res are interlinked for the formation of type III restriction and modification enzymes after bacteriophage treatment, but because mod is required for both restriction and modification, it can be more highly expressed in situations when res is not.On day 0, all 20 genes were upregulated for S. Enteritidis S5-483 after MOI 1, whereas a higher number of genes were downregulated for MOIs 100, 1000 and 10,000 (Figure 3A).In contrast, all genes were highly upregulated for S. Enteritidis S5-483 after MOI 10,000 on day 1, whereas all genes were downregulated for the other MOIs on day 1 (Figure 4A).When comparing populations between the four strains on day 0, MOI 1 for S. Enteritidis S5-483 was the most effective MOI 1 for bacteriophage cocktail (2), SE14, SF5 and SF6, and was able to reduce populations by ~1 log CFU/mL (Figure 2).At a low MOI of 1, ~1 log CFU/mL reduction can be considered quite effective, and potentially, the upregulation of all 20 genes was a response intended for S. Enteritidis S5-483 to overcome the bacteriophage treatment and have a population resurgence by day 1 (Figure 3A).Both MOIs 10 and 10,000 for S. Enteritidis S5-483 had a population resurgence by day 1 after populations were below the detection limit on day 0 (Figures 3A and 4A).The high upregulation of all 20 genes for S. Enteritidis S5-483 after MOI 10,000 could be attributed to the theory of lysis from without, where there is an overload of bacteriophages, and the high upregulation of all genes may have been to overcome the high concentration of bacteriophages for population resurgence on day 1 (Figure 4A).Aside from cas1, which was moderately upregulated for S. Newport S5-639 at an MOI of 1000, all genes were moderately to highly downregulated across the MOIs for S. Newport S5-639 on day 0 (Figure 3C).This was unexpected because S. Newport S5-639 populations were reduced by <0.5 log CFU/mL for MOIs 1-1000 and ~1.5 log CFU/mL for MOI 10,000; therefore, it was expected that there would be more differential gene expression between the MOIs on day 0 (Figure 2C).However, it is important to note that there were only 20 genes studied, and there are many more genes that were not studied that may have been differentially expressed.The high upregulation of the cas1 gene on day 0 for S. Enteritidis S5-483 and S. Ty murium S5-536 after MOI 1000 and MOI 10,000 is not surprising.The cas1 gene code the CRISPR/Cas system-associated protein Cas1 and is an SOS gene [23].SOS genes responsible for cell survival when faced with extensive DNA damage [23].When the a viral infection, such as bacteriophage infection, S. enterica will upregulate CRISPR-a ciated genes such as cas1 to adapt its immune system [48].The immune system is adap On day 3, hsdS was the only gene highly upregulated for S. Enteritidis S5-483 after MOI 1000 and is another DNA restriction and modification gene (Figure 5).A type I restriction and modification system comprises of three gene products, hsdR, hsdM and hsdS [51].The first gene, hsdR is used for restriction, hsdM for methylation and hsdS for sequence specificity [51].Of the three gene products, only hsdM and hsdS are required for both restriction and methylation [51].By day 3, S. Enteritidis S5-483 populations were still below the detection limit, and the high upregulation of DNA modification and restriction genes throughout the three days may indicate that any surviving S. Enteritidis S5-483 cells were still highly expressing these genes as a defence mechanism against invading bacteriophages [52].Potentially, if a later timepoint, such as day 10 or 15, was used, there might be an S. Enteritidis S5-483 population resurgence, and populations might be above the detection limit; however, this will require further study.
Previous research indicates an upregulation in genes involved in SOS response (cas1) and restriction-modification (hsdS) in Salmonella Typhimurium with impaired sensitivity to the target bacteriophage, fmb-p1 [23].Even though Salmonella Typhimurium had an impaired sensitivity to fmb-p1, it was not completely resistant to it; the SOS response and restriction and modification genes were upregulated, but genes involved in virulence were downregulated [23].For the virulence genes (invA, ipfA, spvC, sopB) used in this study, all were moderately or highly downregulated for S. Enteritidis S5-483 and S. Typhimurium S5-536 after MOI 1000 on day 0, except for invA in S. Typhimurium S5-536, which was highly upregulated (Figure 3).All four virulence genes continued to be moderately or highly downregulated for S. Enteritidis S5-483 on day 1 for MOI 1000 (Figure 4).The differing expression of invA in S. Typhimurium S5-536 is unclear, but it is important to note that the MOI used by [23] was an MOI of 0.01, and the MOI used in this study was an MOI of 1000.The difference in MOIs could potentially cause differences in the gene expression of virulence genes in S. enterica.

Conclusions
When bacteriophage treatment is implemented in the food industry, it is important to determine the most optimal MOI for the bacteriophage cocktail used.For this study, MOI 1000 was the most optimal, but for some S. enterica strains, such as S. Typhimurium S5-536, after the bacteriophage cocktail SE14, SF6 and SF5 treatment on day 0, the most effective were both MOIs 1000 and 10,000.For this timepoint, MOI 10,000 may have similar effectiveness compared to MOI 1000, but it would not be the most optimal because a higher number of bacteriophages were implemented than necessary.An insight into the optimal bacteriophage MOI could play a role in reducing antimicrobial costs.For the gene expression of S. enterica after bacteriophage treatment, the high upregulation of SOS and DNA modification and recombination genes indicate that S. enterica activates defence mechanisms against bacteriophage infection for survival.However, it is important to note that the 20 genes studied are a small subset of all genes regulating S. enterica.There may be many other S. enterica genes that are differentially expressed after bacteriophage treatment that were not studied here.Therefore, future studies should include RNA sequencing and transcriptomics to gain a fuller picture of all the S. enterica genes differentially expressed after bacteriophage treatment.

Table 1 .
Primers for housekeeping genes used for RT-qPCR.

Table 2 .
Primers for the twenty target genes used in this study.