The Role of the Exo-Xis Region in Oxidative Stress-Mediated Induction of Shiga Toxin-Converting Prophages

Previous studies indicated that these genetic elements could be involved in the regulation of lysogenization and prophage induction processes. The effects were dramatic in Shiga toxin-converting phage Φ24B after treatment with oxidative stress-inducing agent, hydrogen peroxide, while they were less pronounced in bacteriophage λ and in both phages irradiated with UV. The hydrogen peroxide-caused prophage induction was found to be RecA-dependent. Importantly, in hydrogen peroxide-treated E. coli cells lysogenic for either λ or Φ24B, deletion of the exo-xis region resulted in a significant decrease in the levels of expression of the S.O.S. regulon genes. Moreover, under these conditions, a dramatic decrease in the levels of expression of phage genes crucial for lytic development (particularly xis, exo, N, cro, O, Q, and R) could be observed in Φ24B-, but not in λ-bearing cells. We conclude that genes located in the exo-xis region are necessary for efficient expression of both host S.O.S regulon in lysogenic bacteria and regulatory genes of Shiga toxin-converting bacteriophage Φ24B.


Introduction
Infection of humans by enterohemorrhagic Escherichia coli (EHEC) strains causes hemorrhagic colitis, and in some patients it may result in various complications, including, the most severe of them, the hemolytic-uremic syndrome and neurological dysfunctions [1][2][3]. The main causes of EHECmediated complications are Shiga toxins, produced by the infecting bacteria [4]. The severity of EHEC infections and significance of the medical problem related to them are exemplified by local outbreaks, occurring in various geographical regions around the world. One of the most famous of them took place in 2011 in Germany, where over 4,000 patients developed severe symptoms, and 54 died [5][6][7][8][9][10].
In EHEC strains, Shiga toxins are encoded by genes (called stx genes) located in genomes of prophages [11,12]. The phages bearing stx genes are referred to as Shiga toxinconverting bacteriophages, and all of them belong to the family of lambdoid phages (with phage serving as a paradigm) [12]. stx genes are present between antiterminator gene and the genes coding for proteins causing cell lysis; thus, in the lysogenic state, these genes are not transcribed [13][14][15] and Shiga toxins are not produced. Their expression is possible only after prophage induction [11,12] which usually requires activation of the bacterial S.O.S. response, mediated by RecA protein, though RecA-independent induction of Shiga toxinconverting prophages by EDTA has also been reported [16].
During infection of human intestine by EHEC, the oxidative stress appears to be the most likely condition causing the bacterial S.O.S. response and subsequent induction of Shiga toxin-converting prophages [17]. In fact, it was demonstrated that hydrogen peroxide (which is produced by neutrophils as a response to infection) enhanced production of Shiga toxins by EHEC [18] due to oxidative stress-mediated induction of Shiga toxin-converting prophages [19,20].
Since many antibiotics not only kill bacteria and inhibit their growth but also induce prophage lytic development, their use is not recommended when EHEC infection is confirmed or even suspected (reviewed in [12]). Therefore, there are serious problems with treatment of patients, indicating that searching for new targets of potential therapies against Shiga toxin-producing bacteria is important. One might consider that such therapies should be focused on inhibition of Shiga toxin-converting prophage induction which would impair production of the EHEC virulence factor. All lambdoid phages, including Shiga toxin-converting bacteriophages, contain the region in their genomes which is dispensable for the development under standard laboratory conditions [14,15]. Inside this part of the phage genome, there is an evolutionarily conserved fragment, located between exo and xis genes and transcribed from promoter, called the exo-xis region ( Figure 1). This region encompasses several genes and open reading frames whose functions in phage development are largely unknown, and only a few articles are available in the literature that focused on them. Nevertheless, some interesting observations have been reported. Namely, induction of expression of genes from the exo-xis region resulted in synchronization of the host cell cycle [21] and inhibition of host DNA replication [22]. Moreover, overexpression of these genes impaired lysogenization of E. coli by bacteriophage [23] and enhanced induction of prophages and Φ24 B (one of Shiga toxin-converting phages) [24]. Ea8.5 protein, encoded by a gene located in the exo-xis region, contains a fused homeodomain/zinc finger fold [25] which suggests a regulatory role for this protein. Interestingly, prophage induction with mitomycin C or hydrogen peroxide caused different expression patterns of genes from the exoxis region; such differences were observed in both phages, and Φ24 B [26]. In this work, we used the deletion mutants to investigate the role of the exo-xis region in induction of and Φ24 B prophages under oxidative stress conditions.

Bacteria and
Bacteriophages. E. coli MG1655 strain [27] and its derivatives, used in this work, are listed in Table 1. Bacteria were routinely cultured in the Luria-Bertani (LB) medium at 30 ∘ C (most experiments) or 37 ∘ C (lysogenization and recombination procedures during construction of strains and SOS ChromoTest, according to the instructions of kits' manufacturers), under aerobic conditions. Where appropriate, the following antibiotics were added: chloramphenicol up to 20 g/mL, kanamycin up to 50 g/mL, and/or tetracycline up to 12.5 g/mL.
The deletion mutants were constructed as described previously [29], by using the Quick and Easy E. coli Gene Deletion Kit (from Gene Bridges). The deletion of the indicated region was performed according to the manufacturer's protocol using primers listed in Table 2. In the first step, the targeted sequence has been replaced with the FRT-flanked kanamycin resistance cassette, and the selection marker was subsequently removed in the FLP-recombinase step, leaving only 87 nucleotides of the cassette in the place of the original sequence. Each deletion was confirmed by DNA sequencing.
Lysogenic strains were constructed according to the procedure described previously [24], with slight modifications. Briefly, host bacteria were cultured to A 600 of 0.5 in LB medium supplemented with MgSO 4 and CaCl 2 (to final concentrations of 10 mM each) at 37 ∘ C with shaking. At this point, one milliliter of the culture was withdrawn and centrifuged (10 min, 2000 ×g). Pellet was washed twice with TCM buffer (10 mM Tris-HCl, pH 7.2, 10 mM MgSO 4 , 10 mM CaCl 2 ) and then suspended in 1 mL of the same buffer. Next, bacteria were incubated for 30 min at 30 ∘ C and mixed with phage suspensions at multiplicity of infection (m.o.i.) = 5. Mixtures of bacteria and phages were incubated in TMC buffer for 30 min at 30 ∘ C; then serial dilutions were prepared in TM buffer (10 mM Tris-HCl, 10 mM MgSO 4 ; pH 7.2), and the mixture was plated onto LB agar. Plates were incubated at 37 ∘ C overnight. Lysogens were verified by sensitivity to UV irradiation and confirmed by PCR with primers designed to amplify phage sequence (Table 3).     [31,32].
In the case of UV light irradiation, the production ofgalactosidase was evaluated immediately after the exposure, without 2 h incubation at 37 ∘ C (recommended by the manufacturer), as, without this modification, the visual detection of the blue color was not possible due to rapid S.O.S. response after UV irradiation. Before use, the SOS-ChromoTest bacterial strain (E. coli PQ37, provided with the kit) was lysogenized by following phages: , Δexo-xis, Φ24 B , or Φ24 B Δexo-xis, according to procedure described above.

Preparation of RNA and cDNA from Bacteria.
For the isolation of total RNA, the previously described [26] procedure was employed. Briefly, the prophage induction was performed with 1 mM H 2 O 2 or UV irradiation (at the dose of 50 J/m 2 ). Following induction, the samples were withdrawn at indicated times and the growth of bacteria was inhibited by the addition of NaN 3 (Sigma-Aldrich) to a final concentration of 10 mM. Total RNA was isolated from 10 9 bacterial cells by using the High Pure RNA Isolation Kit (Roche Applied Science). Bacterial genomic DNA carryover was removed by incubation with TURBO DNase from TURBO DNA-free Kit (Life Technologies) for 60 min at 37 ∘ C, according to the manufacturer's guidelines. To evaluate the quality and quantity of the isolated RNA, a NanoDrop spectrophotometer was employed, considering the absorbance ratio (which should be 1.8 ≤ A260/A280 ≤ 2.0). Moreover, band patterns of total RNA were visualized by electrophoresis. The absence of DNA from RNA samples was controlled by PCR amplification and by real-time PCR amplification (all analyzed genes were tested). RNA preparations were stored at −80 ∘ C. cDNA was obtained with Transcriptor Reverse Transcriptase and random hexamer primers (Roche Applied Science), using total RNA samples (1.25 g) as templates. cDNA reaction mixtures were diluted 10-fold for use in real-time PCR.

Real-Time PCR Assay and Data
Analysis. The patterns of genes' expression were determined by quantitative real-time reverse transcription-PCR (qRT-PCR), using the LightCycler 480 Real-Time PCR System (Roche Applied Science) and cDNA samples from lysogenic bacteria. Transcripts of tested phage and bacterial genes were compared in parallel to 16S rRNA housekeeping gene (according to a procedure described previously [34]), whose expression was found to be constant. Primers were developed by Primer3web version 4.0.0 and produced by Sigma-Aldrich or GENOMED. The transcriptional analysis of phage and bacterial genes from lysogenic strains was performed with primers presented in Table 3. Real-time PCR amplifications were carried out for 55 cycles in 20 L reaction volume, using LightCycler 480 SYBR Green I Master (Roche Applied Science) as a fluorescent detection dye. Reactions were performed in Roche 96-well plates containing 10 L 2x SYBR Green I Master Mix, 6.25 ng/ L cDNA, and 200 nM of each genespecific primer (Table 3). Relative quantification assays were performed with cDNA of 16S rRNA and phage/bacterial genes multiplex assay. All templates were amplified using the following program: incubation at 95 ∘ C for 5 min, followed by 55 cycles of 95 ∘ C for 10 s, 60 ∘ C for 15 s, and 72 ∘ C for 15 s. No template control was included with each run. The specificity of amplified products was examined by melting curve analysis immediately after the final PCR cycle and confirmed by gel electrophoresis. Each experiment was conducted in triplicate. The relative changes in gene expression revealed by quantitative real-time PCR experiments were analyzed using the calibrator, normalizing relative quantification method with efficiency correction (called the E-Method). This method has been used and described in detail previously [26,29,35]. The values obtained at time 0 (representing the conditions of spontaneous prophage induction) were used as calibrators. Thus, the following equation was employed to calculate the final results: where is target and is reference.
2.6. In Silico Analyses. The multiple sequence alignment was performed using the ClustalW algorithm available at the website http://www.genome.jp/tools/clustalw/. The Pfam protein families database [36], available at the website http://pfam .xfam.org/, was used to identify protein domains.

Deletion of the Exo-Xis Region Impairs Φ24 B but Not Prophage Induction after Treatment with Hydrogen Peroxide.
Until now, all in vivo studies on effects of the exo-xis region on host or phage development were performed with the use of strains overexpressing genes from this region [21][22][23][24]26]. In this work, we have constructed a series of bacteriophage and Φ24 B mutants with deletions of either the whole exo-xis region or individual genes or open reading frames (Table 1). When wild-type and Φ24 B prophages were induced by UV irradiation (employed in this work as positive control conditions causing effective prophage induction) or hydrogen peroxide treatment of the lysogenic cells, efficiencies of induction and further phage lytic development were comparable in both phages, though some differences were observed in the duration of the lag phase of the phage development ( Figure 2). Induction of Δexo-xis mutant with UV irradiation was similar to that observed for the wild-type , and treatment with hydrogen peroxide caused only a slight delay in the mutant phage development. The decrease in the phage titer at later times of the experiments is characteristic for and most probably arises from adsorption of the progeny virions on fragments of disrupted cell envelopes [15,24]. However, induction of Φ24 B Δexo-xis prophage by UV irradiation was less efficient than that of the wild-type Φ24 B , and induction Oxidative Medicine and Cellular Longevity   of the mutant by hydrogen peroxide was severely impaired (Figure 2). More detailed analyses, based on the full-plate phage titration method, allowing detection of 10 pfu/mL (see Section 2 for details), indicated that the number of pfu per mL of Φ24 B Δexo-xis phage after induction with hydrogen peroxide was at the same range (10 3 /mL) as that measured without specific induction (i.e., representing efficiency of spontaneous prophage induction). Nevertheless, the titer of Φ24 B Δexo-xis measured at 240 and 360 min after induction was 9.0 ± 0.2 × 10 3 and 7.9 ± 0.9 × 10 3 , respectively, that Alkaline phosphatase activity (identified by yellow spots) evaluates viability of tested bacteria. Quantification of -galactosidase activity was performed by densitometry, using the ImageJ software (available at http://imagej.nih.gov/ij/index.html). The results (in arbitrary units reflecting value = 1 ascribed to samples with no inductor), presented as numbers inside the corresponding spots, are mean values from three measurements (with SD < 10% in each case). All these values were significantly ( < 0.001 in t-test) higher than that in the control experiments with no inductor. When Δexo-xis mutants were compared to wild-type phages, the only significant difference ( < 0.05 in t-test) occurred between Φ24 B and Φ24 B Δexo-xis lysogens induced with hydrogen peroxide. is, still 3-4 times higher than that without induction, which was 2.1 ± 0.3 × 10 3 and 2.6 ± 0.4 × 10 3 , respectively (note that the titer of the wild-type Φ24 B after prophage induction was several orders of magnitude higher than that without induction, Figure 2).
Deletions of individual genes and open reading frames from the exo-xis region in Φ24 B did not affect significantly the phage titer. However, such deletions resulted in delays in prophage induction by hydrogen peroxide (Table 4). Interestingly, when prophage induction was stimulated by UV irradiation, such effect was not observed, and in some cases even more rapid induction of the mutant prophages occurred. In bacteriophage , only slight effects of deletions of individual genes and open reading frames were detected (Table 4).
We conclude that the genes and open reading frames from the exo-xis region play important roles in the regulation of lambdoid prophage induction, as deletions of the whole region or single loci caused significant changes in efficiency and timing of this process. The effects of mutations are more pronounced in Shiga toxin-converting phage Φ24 B than in and in lysogenic E. coli cells treated with hydrogen peroxide than in UV-irradiated ones. Thus, the exo-xis region seems to be particularly important for Φ24 B phage under conditions of the oxidative stress, the most likely conditions causing Shiga toxin-converting prophage induction during infection with EHEC.

Hydrogen Peroxide-Mediated Prophage Induction Is a
RecA-Dependent Process. Efficient induction of lambdoid prophages is a RecA-dependent process due to a molecular mimicry between the phage cI repressor and the hostencoded LexA repressor which is self-cleaved after stimulation by the activated form of RecA protein under the S.O.S. response conditions [12][13][14][15]. Such a mimicry is well known for bacteriophage cI protein and LexA [12,13], and we found that both domain structure and amino acid residues crucial for the self-cleavage are also conserved in cI repressor of phage 24 B (Figure 3) (note that cI sequence of 24 B is identical to that of another Shiga toxin-converting bacteriophage, 933 W [37]). Nevertheless, since RecA-independent induction of Shiga toxin-converting prophages has also been reported [16], we asked whether hydrogen peroxide-caused prophage induction depends on the activation of the S.O.S. response.
When testing H 2 O 2 -or UV-dependent induction of prophages , Δexo-xis, Φ24 B , and Φ24 B Δexo-xis in recA13 Oxidative Medicine and Cellular Longevity  mutant host, in the assays analogous to those presented in Figure 2, pfu/mL values were at the levels of those estimated for the uninduced controls (1.8 ± 0.1 × 10 2 , 4.1 ± 1.4 × 10 2 , 1.3 ± 0.1 × 10 3 , and 2.8 ± 0.3 × 10 3 pfu/mL for , Δexoxis, Φ24 B , and Φ24 B Δexo-xis, resp.). Therefore, we conclude that induction of the investigated prophages under conditions of the oxidative stress (treatment with hydrogen peroxide) strongly depends on RecA function. Indeed, in cells lysogenic for or Φ24 B and treated with UV light or hydrogen peroxide, efficient induction of the S.O.S. response was evident, as estimated with the SOS ChromoTest ( Figure 4). Intriguingly, while induction of the S.O.S. response by hydrogen peroxide in Δexo-xis lysogen was comparable to that in lysogen, the signal in the SOS ChromoTest in Φ24 B Δexo-xis lysogen was considerably weaker than in the analogous experiment with Φ24 B lysogen (Figure 4). No such difference could be observed in UV-irradiated bacteria (Figure 4).  Figure 2 and Table 4, particularly delayed induction of Φ24 B prophage devoid of certain genes and open reading frames, and less pronounced effects of their lack in than in Φ24 B . Indications that overexpression of some genes from the exo-xis region of can influence host cell cycle and DNA replication have been reported previously [21,22]. Suggestions that some genes of Φ24 B prophage may affect host growth were also published [38]. However, the results described in this subsection demonstrate for the first time that the exo-xis region can significantly modulate one of global cellular responses, the S.O.S. response, after treatment with hydrogen peroxide.  regulatory processes and lytic development, has been tested under the same conditions as described in the preceding subsection. The specific conditions and time after addition of hydrogen peroxide into the cell culture at which samples were withdrawn were chosen on the basis of similar experiments reported previously [26]. Interestingly, different effects of the deletion of the exo-xis region were observed for phages and 24 B . In , deletion of genes and open reading frames located between exo and xis genes did not cause considerable effects on mRNA levels for xis, exo, and Q, whereas expression of int, cIII, N, cI, cro, cII, O, and R was enhanced upon treatment with hydrogen peroxide (Figure 6). Completely different results were obtained when Φ24 B and Φ24 B Δexo-xis lysogens were studied; namely, expression of all tested genes was drastically impaired in hydrogen peroxide-treated bacteria in the absence of the exo-xis region on the prophage ( Figure 6). While negative regulation of transcription from cIIdependent promoters by overexpression of the exo-xis region has been reported previously in phage [23], this study demonstrated for the first time significant effects of this region on expression of a battery of phage genes under conditions of the oxidative stress. The results presented in Figure 6 for phage are compatible with those published previously (though obtained with different methods) [23], as overexpression of the exo-xis region had opposite effects to those observed in its absence. On the other hand, severely impaired expression of all tested phage genes in Φ24 B Δexoxis was unexpected. However, these results ( Figure 6) can explain a strong defect in the induction of Φ24 B Δexoxis prophage (and perhaps further lytic development) by hydrogen peroxide (Figure 2). Similarly, drastic differences between effects of Δexo-xis mutations on hydrogen peroxidemediated prophage induction between and 24 B (Figure 2) can be ascribed to opposite regulation of expression of phage genes in the absence of the exo-xis region.

Effects of the Exo-Xis Region on Expression of Host and
Phage Gene in UV-Irradiated Lysogenic Cells. Experiments analogous to those described in two preceding subsections were performed with lysogenic cells irradiated with UV. Interestingly, in both and 24 B phages, deletion of the exoxis region caused only moderate effects on expression of most genes from the S.O.S. regulon (Figure 7), contrary to hydrogen peroxide-treated bacteria where the differences were significantly higher (compare Figures 5 and 7). The exceptions in UV-irradiated cells were rpoD, rpoH, and rpoS genes in and rpoH and rpoS in 24 B , whose expressions were at considerably lower level in the absence of the exo-xis region ( Figure 7). One should also note that the induction of the S.O.S. regulon with UV irradiation was quicker than that with hydrogen peroxide. These results indicate that the influence of the exo-xis region on the S.O.S. response is particularly well pronounced under conditions of the oxidative stress. Unlike the S.O.S. regulon expression, levels of mRNAs of bacteriophage genes in UV-irradiated cells were affected similarly to those in hydrogen peroxide-treated lysogenic bacteria by the absence of the exo-xis region ( Figure 8). Again, although some differences were observed between and Δexo-xis, the differences between Φ24 B and Φ24 B Δexo-xis were dramatic. This indicates that the influence of the exo-xis region on expression of phage genes after prophage induction does not depend on the induction agent.

Conclusions
The exo-xis region is necessary for effective, RecA-dependent induction of Shiga toxin-converting bacteriophage Φ24 B under conditions of the oxidative stress. In hydrogen peroxide-treated E. coli, this region positively influences expression of the S.O.S. regulon in both Φ24 B and lysogens and expression of phage genes crucial for lytic development (particularly xis, exo, N, cro, O, Q, and R) in Φ24 B , but not in . Since the oxidative stress appears to be the major cause of induction of Shiga toxin-converting prophages during infections of human intestine by enterohemorrhagic E. coli (EHEC), the exo-xis region and/or products of its expression might be considered as potential targets for anti-EHEC drugs.