Three STEC O26:H11 strains (2976-1, F46-223, and 09QMA-277.2) (Bonanno et al., 2016), isolated from dairy products and containing stx1, stx2 and both stx1 and stx2 genes, respectively, were used in this study. E. coli strains were cultivated in Lysogeny broth (LB) at 37°C.

An overnight culture of STEC O26:H11 was inoculated at 2% in a fresh LB medium with 5 mM of CaCl₂ and incubated at 37°C. At the exponential growth phase (OD₆₀₀ of 0.3), cultures were further incubated at 37°C for 24 h with shaking at 240 rpm, in the presence of stress factors (listed below). All cultures were centrifuged at 7,200 × g for 10 min, and the supernatants were filtered through low-protein-binding 0.22 μm-pore-size membrane filters (Millex-GP PES; Millipore, St-Quentin-en-Yvelines, France) for phage purification.

The factors linked to cheese manufacturing that were studied for their impact on Stx phage induction were lactic acid at 0.05, 0.5, 1.5, and 3%, hydrogen peroxide (H₂O₂) at 0.25 and 3 mM, salt (NaCl) at a final concentration of 3% (taking into account NaCl from the LB broth). The effect of temperature, i.e., 9 and 42°C, was also tested.

The other studied factors were related to the analytical STEC detection procedure and included acriflavin (12 mg/l), novobiocin (20 mg/l) and bile salts (1.5 g/l). Finally, cefixime-tellurite (C-T; 0.05 and 2.5 mg/l), used as a supplement in the Rhamnose MacConkey agar (RMAC) adapted for the isolation of STEC O26 (Hiramatsu et al., 2002) was also evaluated for its ability to induce Stx phage.

All these experiments were performed in duplicate. Moreover, each strain was also cultured without inducing agent at 37°C as a control representing the spontaneous induction of Stx phages. Twelve to 14 replicates (n_R) per strain were performed. Bacteriophage spontaneous induction variability, expressed in log₁₀ stx gene copies per milliliter (GCs/ml), was characterized for each strain with normal distribution adjusted to n_R values. The effect of a factor was assessed by calculating the probability of observing log₁₀ (GCs/ml) if the factor has no effect on induction (Prob_i/s), i.e., if the production of Stx phage is only due to spontaneous induction. A factor was considered to significantly induce more Stx phages compared to spontaneous induction when both Prob_i/s values (associated to replicate values) were above 0.975. A factor was considered to significantly induce less Stx phages compared to spontaneous induction when both Prob_i/s values were below 0.025.

Filtered supernatants obtained after Stx phage induction were treated with DNase using the Turbo DNA-free^TM kit (Ambion^®, life technologies, Illkirch, France) and phage DNA was released by heat treatment for 10 min at 100°C (Bonanno et al., 2016). Quantitative PCR (qPCR) assays targeting stx1 and stx2 genes, were then used with the LightCycler^® 480 instrument (Roche Diagnostics, Meylan, France) as described previously (Derzelle et al., 2011). However, the fluorescent reporter Flc of the stx1 probe was replaced by Rox to eliminate any interference arising from the acriflavin solution.

DNA from strain EDL933 which carries only one copy of the stx1 and stx2 genes was used for preparation of the standard curve. Briefly, bacterial cells were collected by low speed centrifugation, and supernatants containing any spontaneously induced Stx phages were discarded. Genome DNA was then purified from the pelleted bacterial cells and its concentration determined using a Nanodrop spectrophotometer (Thermo Scientific, Illkirch, France). Genomic copy numbers were calculated from the known size of EDL933 genome (i.e., 5.44 Mb). The standard curve was prepared from serially diluted genomic DNA and used for the quantification of Stx phages expressed as log₁₀ stx gene copies per milliliter (GCs/ml). All the samples were run together with the standard-positive and negative controls. Moreover, the absence of inhibition was verified by the absence of Ct variation of the standard in the presence of DNA extracts prepared from the studied stress conditions.

Four factors related to the manufacturing of cheeses were selected in this study and tested for their ability to induce Stx phages. Lactic acid (at various concentrations) was used to mimic acid stress provoked by lactic acid bacteria during the coagulation step. Salt (NaCl) at 3% was tested due to its role in the conservation and flavor enrichment step. Hydrogen peroxide (H₂O₂, at various concentrations) was used to mimic oxidative stress caused by other bacteria present in the cheese matrix. Finally, the initial milk heating step and final ripening (cooling) step were tested by performing bacterial growth at 42 and 9°C, respectively.

Figure 1 presents the amount of Stx phages produced by each strain in the presence of the factors tested compared to the control representing the spontaneous induction of Stx phages (i.e., without inducing agent).

When STEC were grown in the presence of H₂O₂ to 3 mM, an increase in the concentrations of Stx phages was significantly observed for strains 2976-1 (Figure 1A) and 09QMA277.2 (Figures 1C,D). For strain F46-223, only one of the two replicate showed a significant induction effect (Figure 1B). Therefore, we considered that H₂O₂ at 3 mM induced Stx phages in vitro. This result was in agreement with previous reports which described the inducing effect on Stx phages by H₂O₂ 3 mM (Los et al., 2009, 2010).

Interestingly, 3% NaCl induced Stx1 phage only from strain 2976-1 (Figure 1A), but this was not the case for the Stx2 phage neither from strain F46-223, nor for both Stx1 and Stx2 phages from strain 09QMA277.2 (Figures 1B–D). Harris et al. (2012) described that 3% NaCl inhibited Stx phage induction presumably because of the inhibitory effect of 3% salt on vital physiological processes. On the other hand, they also demonstrated that the presence of a salt concentration (2%) equivalent to that found for meat processing induced Stx phages (Harris et al., 2012). Moreover, Wagner et al. (2002) demonstrated that Stx1 phages induction could be regulated by RecA-independent pathway and therefore not by the SOS response. In our study, Stx phage induction did not occur in the presence of 3% NaCl except for an Stx1 phage from one strain. Whether the alternative RecA-independent induction system was used in this strain remains to be determined. This phenomenon was not observed for the Stx1 phage from strain 09QMA277.2 which possesses two Stx phages (Stx1 and Stx2). However, it was shown previously that when two Stx prophages are integrated into the bacterial chromosome, the production of Shiga-toxin and the activation rate of the lytic cycle of each phage are both highly reduced (Muniesa et al., 2003; Serra-Moreno et al., 2008).

The presence of lactic acid at 0.05% resulted in Stx phage production similar to the spontaneous induction (Figure 1) and a slight decrease in Stx phage production was even observed with 0.5% lactic acid (Figure 1). This is in agreement with a previous study which showed that for pH lower than 5.5, inhibition of the induction of Stx phages occurred, even in the presence of mitomycin C (Imamovic and Muniesa, 2012). In addition, stx gene transcription was shown to be very low or non-existent at pH 5.5 (Olesen and Jespersen, 2010), consistent with Stx phage repression at low pH. This phenomenon could be linked to the RpoS system involved in the survival of bacteria triggered by acid stress (Cheville et al., 1996). No Stx phages could be detected when higher concentrations of lactic acid such as 1.5 and 3% were used but in both cases this was due to the inhibition of PCR.

Incubation of STEC at a low temperature (9°C) or at 42°C did not reveal significantly difference with the spontaneous induction (Figure 1) A previous study showed that the temperature had an effect on the induction of Stx phage, which was lower at 30°C than at 37°C whereas at 43°C, it was considerably increased (Los et al., 2009). However, this observation was made in the presence of inducing agents such as mitomycin C, UV irradiation and H₂O₂, i.e., in experimental conditions that differ from those tested here.

Finally, Stx phage induction was also tested in real cheese-making conditions where cheeses were produced in an experimental plant using milk inoculated with an stx1- or stx2-positive STEC O26:H11 strain. A total of 48 samples were collected at various time points (i.e., 6 h, 24 h, 8 and 28 days) during cheese production. Stx1 and Stx2 phages were detected from 3 and 7 samples, respectively. No correlation between the induction of Stx phage and the different sampling steps during the process could be observed.

The analytical STEC detection procedure based on the technical specification XP CEN ISO/TS 13136 relies on the use of several selective agents. Acriflavin (at 12 mg/l) and novobiocin (at 20 mg/l) are used for STEC isolation from dairy products and other food categories, respectively. They are added to the enrichment medium, triptic soy broth (TSB) modified with bile salts at 1.5 g/l (mTSB). Moreover, the supplement cefixime-tellurite (C-T, at 0.05 and 2.5 mg/l, respectively) is used as a selective agent in agar medium for the isolation of STEC O26 onto Rhamnose MacConkey agar (CT-RMAC) (Hiramatsu et al., 2002). These specific culture media allow the enrichment and isolation of STEC to the detriment of the background microflora.

The presence of C-T had no effect on Stx phage induction (Figure 1). It was also the case for novobiocin and acriflavin (Figure 1), with two exceptions. First, Stx1 phage production from 09QMA277.2 was considerably reduced in the presence of novobiocin compared to the control condition (Figure 1C). Then, Stx1 phage production from 2976-1 strain was also reduced but this occurred in the presence of acriflavin (Figure 1A). An additional test was performed therefore with two other stx1-positive STEC O26:H11 strains (i.e., 10d and 09QMA245.2). A decrease of Stx1 phage production was also observed in the presence of acriflavin but only for one of the two strains (i.e., 10d) (data not shown). Distinct genetic backgrounds of the various strains tested could explain the differences observed in Stx phage production in response to acriflavin. This selective agent was shown to induce cell wall changes in Staphylococcus aureus (Kawai and Yamagishi, 2009). Moreover, it has the capacity to bind on the cell wall of E. coli and the acriflavin-binding capacity is controlled by the acrA gene. A mutation of acrA leads to sensitivity not only to acriflavin but also to mitomycin C (Nakamura and Shinya, 1985). This observation may therefore suggest differences in the acrA gene and acriflavin sensitivity for 2976-1 and 10d strains compared to 09QMA277.2 and 09QMA245.2 strains. Moreover, such differences would also explain the higher production of Stx1 phage observed elsewhere in the presence of mitomycin C for strains 2976-1 and 10d compared to strains 09QMA277.2 and 09QMA245.2 (Bonanno et al., 2016).

Finally, induction of Stx phage by bile salts at 1.5 g/l could not be quantified by qPCR due to the inhibition of PCR by bile salts. However, a slight decrease of OD₆₀₀ (of -0.13 to -0.57 units) was observed with bile salts at 1.5 g/l for the three strains tested (data not shown) which might reflect induction of Stx phages.