Effect of a bacteriophage T5virus on growth of Shiga toxigenic Escherichia coli and Salmonella strains in individual and mixed cultures

A previously isolated a bacteriophage, vB_EcoS_AKFV33 of T5virus, demonstrated great potential in biocontrol of Shiga toxigenic Escherichia coli (STEC) O157. This study further evaluated its potential as a biocontrol agent in broth culture against other important non-O157 serogroups of STEC and Salmonella. AKFV33 was capable of lysing isolates of STEC serogroups O26 (n = 1), O145 (n = 1) and Salmonella enterica serovars (n = 6). In a broth culture microplate system, efficacy of AKFV33 for killing STEC O26:H11, O145:NM and Salmonella was improved (P < 0.05) at a lower multiplicity of infection and sampling time (6–10 h), when STEC O157:H7 was also included in the culture. This phage was able to simultaneously reduce numbers of STEC and Salmonella in mixtures with enhanced activity (P < 0.05) against O157:H7 and O26:H11, offering great promise for control of multiple zoonotic pathogens at both pre and post-harvest.


Background
Shiga toxin-producing Escherichia coli (STEC) and Salmonella are often carried by food-producing animals and remain leading causes of foodborne illness worldwide [1]. However, few effective on-farm interventions have been established. Moreover, with emergence of STEC and Salmonella that are resistant to conventional interventions (e.g. heat, acid and chemical sanitizers [1];), novel approaches are needed to control these pathogens in both primary and secondary food production. Bacteriophages (phages) are viruses that naturally use bacteria as hosts, and when virulent, induce lysis of the infected bacteria. Commercial phage-based products have been used in the biocontrol of important foodborne bacteria including STEC and Salmonella [2]. However, several challenges remain before phages could be widely used in the food industry. One major challenge is that the host range of phages is often limited to certain species and even strains within species. Although such specificity is often desirable, phage treatment to decontaminate foods adulterated with multiple pathogenic species would often require phage cocktails, a preparation including multiple phages with each targeting a specific pathogen. However, limitations in the formulation of phage cocktails such as interference among phages and high manufacturing costs [2] make the identification of polyvalent phages that kill multiple bacterial host species particularly desirable.
Previously, we identified and systematically characterized a phage vB_EcoS_AKFV33 (AKFV33), a T5virus that possesses many of the desired features of a biocontrol agent [3]. Moreover, we found AKFV33 to be superior to phages T4virus, T1virus and rV5virus used individually or as phage cocktails for inactivating O157 STEC on refrigerated beef [4].

Phage microplate virulence assay
Host range and lytic activities of phage AKFV33 were assessed using a microplate phage virulence assay [11]. High titer phage stocks (> 10 9 plaque forming units (PFU)/ml) were propagated and filter-purified as previously described [3]. To estimate multiplicity of infection (MOI), the filter-purified phage stocks were serially diluted and incubated at 37°C without shaking for 5 h with 10-fold diluted overnight cultures of bacteria in a 96-well microplate. After incubation, wells were examined visually for turbidity and the highest dilution that resulted in complete lysis (no discernable turbidity) of bacteria was recorded. The MOI for each phage-host assay was calculated by dividing the initial number of phages in the greatestdilution wells by the initial number of bacteria added, as determined from plate counts of serially diluted bacterial cultures. Sensitivity to phages was categorized as follows: extremely susceptible: (10 − 6 ≤ MOI < 10 − 2 ); highly susceptible: (0.01 ≤ MOI < 1); moderately susceptible: (1 ≤ MOI < 10); and minimally susceptible: (10 ≤ MOI < 100).

Phage lysis kinetics
To further assess dynamics of AKFV33 infection, a bacterial growth inhibition curve was conducted.
Two independent experiments were performed in duplicate.

Statistical analysis
Results from phage lysis kinetics and enumeration of bacteria from larger scale broth cultures were compiled from two independent experiments, respectively. The OD values at 600 nm were squareroot transformed and colony forming units were log-transformed. Influence of MOIs and time on phage efficacy were analyzed using the MIXED model with repeated measure. Least-squares were used to differentiate means (P < 0.05). The analyses were conducted with SAS (version 9.4, SAS Institute, Cary, NC).

Results
Of 36   Letters which differ after the mean values indicate differences (P < 0.05) among MOIs within each bacterial culture Asterisks * , ** and *** indicate a statistical difference between phage-treated individual and mixed culture within same MOI at P < 0.05, P < 0.01 and P < 0.001, respectively AKFV33 with MOIs ranging from 0.5 to 4. Although complete lysis of other strains was not observed after 5 h of phage treatment, phage-treated cultures (n = 3, 1 and 1, respectively), from STEC O26, S. Ago and S. Soerenga showed complete lysis at MOI = 2-10 at 2 h (data not shown). The subsequent re-growth after 2 h may indicate rapid emergence of phageresistant mutant strains in these cultures, which complies with previous studies of other T5virus strains [6,7]. The ability of AKFV33 to lyse some non-O157 STEC and Salmonella strains is consistent with other reports that T5viruses may have broad host ranges across multiple bacterial species [5][6][7][8][9][10].
For individual bacterial cultures, AKFV33 caused an overall reduction of 7.5 ± 0.4 log 10 CFU/ml in O26:H11, greater (P < 0.001) than those in STEC O157:H7 (2.5 ± 2.7 log 10 CFU/ml) or S. Typhimurium (2.2 ± 1.2 log 10 CFU/ml, Fig. 2). The greatest efficacy of the phage (P < 0.001) was at 4 and/or 7 h, but was reduced (P < 0.001) thereafter. Notably, after 24 h of incubation, phage treatment had no effect (P > 0.1) on the numbers of O157:H7 or S. Typhimurium. When exposed to a mixture of O157:H7, O26:H11 and S. Typhimurium, AKFV33 was able to simultaneously reduce (P < 0.01) numbers of each bacteria in the mixtures by 2-8 log 10 CFU/ml (Fig. 2). Moreover, both O157:H7 and O26:H11 in the mixture were undetectable (< 300 CFU/ml) at each sampling time, even after 24 h. This indicates that AKFV33 was more active and/or the targeted STEC were more vulnerable to the phages (P < 0.05) in mixed cultures. In contrast, S. Typhimurium was equally sensitive to the phages either alone or in a mixture with O157:H7.

Discussion
To our knowledge, this is the first study to evaluate effectiveness of a polyvalent phages T5virus in control of STEC and Salmonella in a mixed culture. In our previous studies, AKVF33 was shown to be highly virulent to various phage types of STEC O157 strains [3], but its virulence for other foodborne pathogens was unknown. Here we have found that AKVF33 is virulent for a broad host range that includes some non-O157 STEC and Salmonella serovars, and that in mixed cultures, AKVF33 not only simultaneously reduces numbers of STEC and Salmonella, but in some instances also has greater efficacy. Further study is required to understand mechanism(s) underlying this improved efficacy. Potentially, replication of AKVF33 in a preferred host (O157:H7) and enhanced concentrations of phage led to improved control of nonpreferred hosts (Salmonella and non-O157 E. coli). In addition, this finding was consistent with earlier reports that phage av-08 (unknown taxonomy) was able to decontaminate S. Montevideo and STEC O157:H7 on chicken skin [14]. Costa et al. [15] also found that single phage ELY-1 or phSE-5 (unknown taxonomy) reduced number of non-O157 E. coli and S. Typhimurium ATCC13311 in a mixture, although this reduction was less than produced by a cocktail of both of these phages in broth culture. The relative contribution of polyvalent phages vs phage cocktails to bacterial biocontrol remains unclear. However, Zhao et al. [16] reported that a polyvalent phage of the Siphoviridae was effective in decreasing population of E. coli K12 and Pseudomonas aeruginosa in a soilcarrot system. Although less effective than a cocktail of phages against these organisms, polyvalent phages were more capable than the phage cocktail of sustaining the diversity of the commensal bacterial community in the system. In another study, a polyvalent phage of the Podoviridae in combination with biochar treatment effectively eliminated E. coli K12 and P. aeruginosa in a soil-lettuce system, while synergistically enhancing indigenous bacterial communities [17]. This suggests that polyvalent phages such as AKFV33 may be used for simultaneous inhibition of various zoonotic bacterial pathogens without harming beneficial microbes resident in gastro-intestinal tracts of food animals or in food products.