The bacterial strains and plasmids used in this study are listed in Table 1. All bacteria were grown in LB (Luria-Bertani) medium (Nacalai tesque, Kyoto, Japan). For aerobic culture, EHEC strains were grown in loosely capped glass tube with shaking at 160 rpm. For anaerobic culture, we grew EHEC in a sealed container with carbon dioxide gas generators, AnaeroPack-Anaero (Mitsubishi Gas Chemical Co., Inc., Tokyo, Japan) for MIC assays. When we need to monitor the cell growth by measuring absorbance at 600 nm for RNA extraction and cyclic AMP assays, we used Anaerobic Hungate culture tubes equipped with a rubber stopper and a screw cap (Chemglass Life Sciences, Vineland, NJ). For marker selection and maintenance of plasmids, antibiotics were added to growth media at the following concentrations; 150 μg/ml ampicillin, 25 μg/ml kanamycin and 30 μg/ml chloramphenicol.

We obtained antibiotics and reagents used in this study except ampicillin, kanamycin, chloramphenicol and isopropyl β-D-1-thiogalactopyranoside (IPTG) from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Ampicillin, kanamycin, chloramphenicol and IPTG were purchased from Nacalai tesque (Kyoto, Japan).

In-frame deletions of crp, cyaA, and uhpA were constructed by sequence overlap extension PCR according to a strategy described previously (Link et al., 1997), with primer pairs, delta1/delta2 and delta3/delta4 primers for each gene as described in Table 2. The upstream flanking DNA included 450 bp and the first three amino acid codons. The downstream flanking DNA included the last two amino acid codons, the stop codon, and 450 bp of DNA. These deletion constructs were ligated into BamHI and SalI-digested temperature sensitive vector pKO3 (Link et al., 1997) and introduced into HH-H7-008, the parent strain (Hirakawa et al., 2009). We selected sucrose-resistant/chloramphenicol-sensitive colonies at 30°C and confirmed the resulting mutant strains using PCR analysis and DNA sequencing.

To construct His₆-CRP expression plasmid pQE80crp, the crp gene was amplified with the primer pair shown in Table 2. The product was digested with BamHI and HindIII and ligated into similarly digested pQE80 plasmid.

To construct pTrc99Afnr, pTrc99AcyaA, pTrc99Acrp, or pTrc99AuhpA for complementation tests, we amplified the fnr, cyaA, crp or uhpA gene, and digested with NcoI and BamHI for pTrc99Afnr, pTrc99Acrp, and pTrc99AuhpA or EcoRI and BamHI for pTrc99AcyaA, then ligated into similarly digested pTrc99A plasmid. All constructs were confirmed by DNA sequencing.

To test susceptibility of EHEC to fosfomycin, MIC assays were performed by a serial agar dilution method with the standard method of the Clinical and Laboratory Standards Institute (CLSI) (Clinical Laboratory Standards Institute, 2011). To conduct MIC assays under anaerobic conditions, we grew EHEC strains without shaking in a sealed container with AnaeroPack-Anaero for 20 h. After taking the cultures out the anaerobic container, five thousand cells were immediately inoculated on Mueller-Hinton agar (EIKEN Chemical Co., Ltd., Tochigi, Japan) containing fosfomycin, and the agar plates were incubated in the sealed container with fresh AnaeroPack-Anaero gas generator pack at 37°C. The MICs were determined as the lowest concentration at which growth was inhibited.

EHEC strains carrying pNN-glpT-P or pNN-uhpT-P, the LacZ reporter plasmid, were aerobically or anaerobically grown at 37°C in LB medium. To measure lacZ expression from pNN-uhpT-P, we added 25 μg/mL of glucose-6-phosphate into the medium because the promoter activity of uhpT gene was too low to be detected when glucose-6-phosphate is absent. β-Galactosidase activities from lacZ expression in cell lysates were determined as Miller's method (Miller, 1992).

Bacteria were grown to mid-logarithmic growth phase (OD₆₀₀ ~0.4) in LB medium. Total RNA extraction and cDNA synthesis were performed using SV Total-RNA Isolation System and GoScript™ Reverse Transcription System as described by the manufacturer (Promega Corp., Madison, WI). Real-time PCR included 2.5 ng cDNA and 200 nM primers in SYBR Select Master Mix (Applied Biosystems, Foster City, CA) and were run on an ABI Prism 7900HT Fast Real Time PCR System. Constitutively expressed rrsA and rpoD genes that respectively encode 16S ribosomal RNA and RNA polymerase sigma 70 factor were used as an internal control (Kurabayashi et al., 2014). Primers are listed in Table 2. Amplification plot and melting curve data are available upon request.

Intracellular cAMP levels of EHEC grown under aerobic or anaerobic conditions were determined by using Cyclic AMP EIA Kit (Cayman Chemical, Ann Arbor, MI). Bacteria were grown in 20 ml of LB medium to late logarithmic phase and harvested. The cell pellets were once washed in phosphate buffered saline (PBS) and suspended in 0.5 ml of EIA buffer supplied in the kit, then lysed by sonication. The lysate was centrifuged and the amount of cAMP in the resulting supernatant was quantified using cAMP AChE Tracer, cAMP EIA Antiserum and Ellman's reagent according to the manufacturer protocol. cAMP levels in fraction of the cell lysate were represented as pmol per cells producing 1 μg of protein.

N-Terminal histidine tagged D154AFNR and CRP, His₆-D154AFNR, and His₆-CRP were expressed in and purified from E. coli Rosetta™(DE3) (Novagen/EMD Bioscience, Philadelphia, PA), respectively. We supplied 100 μM cAMP into the buffer for purification of His₆-CRP. Bacteria containing recombinant plasmid were grown at 37°C to an OD₆₀₀ of 0.4 in LB, 0.5 mM IPTG was then added, and culture growth was continued for 3 h. Cells were harvested and stored at −80°C overnight. The cell pellet was suspended in lysis buffer (20 mM Tris [pH 7.9], 500 mM NaCl and 10% glycerol) and lysed by sonication. The lysate was centrifuged and the resulting supernatant was mixed with Ni-NTA agarose (Qiagen, Valencia, CA) for 1 h. The agarose was washed with 50 mM imidazol twice and then protein was eluted with 200 mM imidazol. The protein was >95% pure as estimated by SDS-PAGE electrophoresis and Coomassie brilliant blue staining. Protein concentration was determined using the Bio-Rad protein assay (Bio-Rad, Hercules, CA). The protein was diluted into gel-shift/footprinting assay buffer (20 mM Tris [pH 7.5], 50 mM KCl, 1 mM dithiothreitol, 10% glycerol and 200 μM cAMP) to a concentration of 4 pmol/μl, and stored at 4°C.

Gel-shift assays were performed as previously described (Kurabayashi et al., 2015a). We used DNA probes containing the 300 bp region upstream of the glpT and uhpT gene, respectively. We also used a DNA fragment from Pseudomonas aeruginosa rhlR gene as a non-specific control probe. The probe DNA fragments (0.30 pmol) were mixed with purified CRP and/or D154AFNR in a 10 μl reaction mixture containing gel-shift/footprinting assay buffer described above. After incubation for 20 min at room temperature, samples were separated by electrophoresis on a 5% polyacrylamide Tris-glycine/EDTA (10 mM Tris [pH 8.0], 380 mM glycine and 1 mM EDTA) gel in Tris-glycine/EDTA buffer at 4°C. DNA bands in the gel were stained by 10,000-fold-diluted SYBR Green I nucleic acid stain (Lonza, Walkersville, MD), and visualized under UV light at 300 nm.

DNase I footprinting was performed using a previously described non-radiochemical capillary electrophoresis method on an ABI PRISM genetic analyzer equipped with an ABI PRISM GeneScan (Hirakawa et al., 2005). The 6-carboxyfluorescein (6-FAM)-labeled (5′), 300-bp DNA fragments (starting at 300 bp upstream of the glpT start codon and ending 1 bp upstream of the glpT start codon for glpT probe and starting at 300 bp upstream of the uhpT start codon and ending 1 bp upstream of the uhpT start codon for uhpT probe, respectively) were generated by PCR amplification with primer pairs a 6-FAM labeled forward primer (glpT-footprintF-6FAM) and an unlabeled reverse primer (glpT-footprintR) for glpT probe and a 6-FAM labeled forward primer (uhpT-footprintF-6FAM) and an unlabeled reverse primer (uhpT-footprintR) for uhpT probe. Each DNA fragment (0.45 pmol) was mixed with purified protein (20 pmol) in a 50 μl reaction mixture containing the same buffer as above. After incubation for 20 min at room temperature, DNase I (0.3 units, Promega Corp., Madison, WI) was added. After incubation for 60 s at room temperature, samples were purified for electrophoresis on the ABI PRISM genetic analyzer apparatus.

P-value in each assay was determined by unpaired t-test and two-way ANOVA with the GraphPad Prism version6.00.

cAMP-bound CRP (CRP-cAMP) activates expression of both glpT and uhpT genes while UhpA activates expression of uhpT but not glpT. The role of these activators on expression of glpT and uhpT has been demonstrated in aerobically grown E. coli laboratory strains (Schumacher and Bussmann, 1978; Merkel et al., 1995). To investigate the effect of these regulators on expression of glpT and uhpT in EHEC grown under anaerobic conditions, we constructed crp, which encodes the regulatory protein, cyaA, which encodes cAMP synthase and uhpA deletion mutants in EHEC O157:H7 background, and measured promoter activities of glpT and uhpT via lacZ expression from the reporter plasmids pNN-glpT-P and pNN-uhpT-P in these mutant strains when aerobically or anaerobically grown. Previously, we found that expression of glpT and uhpT is activated by FNR during anaerobic culture (Kurabayashi et al., 2015a). Initially, we showed that levels of lacZ expression from both pNN-glpT-P and pNN-uhpT-P in the wild-type parent when grown under anaerobic conditions were indeed higher than those when grown under aerobic conditions, and expression levels from these two genes under anaerobic conditions were decreased by deletion of fnr (Figures 1A,B). We note that contribution of fnr to uhpT expression under anaerobic conditions is partial because significant level of lacZ expression from pNN-uhpT-P in the fnr mutant was still observed. This observation implies that there is yet another unknown factor that elevates expression of uhpT during EHEC anaerobiosis.

Levels of lacZ expression from pNN-glpT-P in crp and cyaA mutants were under detectable limit under both aerobic and anaerobic conditions while those from pNN-uhpT-P in these mutants was still detectable under anaerobic conditions, however approximately 5-fold lower than the wild-type parent (Figures 1A,B). Since CRP requires cAMP for its activity, it is convicting that crp and cyaA mutants showed same phenotype for glpT and uhpT expression. On the other hand, deletion of uhpA completely abolished lacZ expression from pNN-uhpT-P, but did not affect that from pNN-glpT-P (Figures 1A,B). These results suggest that CRP-cAMP activates expression of both glpT and uhpT during anaerobic cultures as aerobic cultures, and UhpA only activates that of uhpT.

To confirm that lowered expression of glpT and uhpT in deletion of crp, cyaA, or uhpA leads to decrease of susceptibility to fosfomycin, we determined the MICs of fosfomycin for these mutants with the wild-type parent, Δfnr, ΔglpT, ΔuhpT, and ΔglpTΔuhpT strains. As observed in our previous study, the wild-type parent grown under anaerobic conditions exhibited a lower MIC than that grown under aerobic conditions, and the fnr mutant was 4-fold lower susceptible than the wild-type parent under anaerobic conditions, but no difference in the MIC between these strains was seen under aerobic conditions (Table 3). In this study, we found that deletions of crp and cyaA decreased the susceptibility to fosfomycin up to the level of ΔglpTΔuhpT strain under aerobic and anaerobic conditions whereas the susceptibility in uhpA mutant was same as that in uhpT mutant (Table 3). To confirm the contribution of these regulatory genes to glpT and uhpT-dependent susceptibility of EHEC to fosfomycin, we introduced pTrc99Afnr, pTrc99AcyaA, pTrc99Acrp, or pTrc99AuhpA plasmid into each mutant, and determined the MICs of fosfomycin. We observed that MIC phenotypes on fnr, cyaA, crp, and uhpA mutants were actually complemented by introducing exogenous gene expression plasmids, respectively (Table 3). These observations support the results observed in the promoter assay that deletions of crp and cyaA decreased expression of both glpT and uhpT, and deletion of uhpA only decreased expression of uhpT.

Activity of CRP-cAMP to regulate expression of target genes is affected by levels of both cAMP and CRP in cells (Hogema et al., 1997). These levels might be elevated during anaerobic culture, then lead to increase of glpT and uhpT expression. To test this hypothesis, we measured levels of intracellular cAMP and cyaA and crp transcripts in aerobically and anaerobically grown EHEC. We found that intracellular concentration of cAMP in EHEC when anaerobically grown was 2-fold higher than that when aerobically grown (Figure 2A). The data of qPCR analyses showed higher transcripts of cyaA and crp in EHEC grown under anaerobic conditions than aerobic conditions (Figure 2B). To confirm that increase of CRP-cAMP activity elevates expression of glpT and uhpT in anaerobic cultures, we measured levels of these genes expression in the wild-type parent grown with or without glucose under anaerobic conditions because addition of glucose as a PTS sugar reduces cAMP level associating with CRP-cAMP activity (Botsford and Harman, 1992). As expected, glucose addition reduced both expression of glpT and uhpT (Figure 2C). When the EHEC was grown with glycerol instead of glucose, we did not observe the reduction of glpT and uhpT expression levels because glycerol does not affect cAMP level (Botsford and Harman, 1992; Figure 2C). These results suggest that activity of CRP-cAMP in EHEC when grown under anaerobic conditions is higher than aerobic conditions, and leads to elevated expressions of glpT and uhpT.

FNR is a transcriptional regulator that is able to induce expression of particular genes during anaerobic growth (Green et al., 1996). To investigate whether FNR induces expression of crp and cyaA genes during anaerobic growth or not, we compared levels of these genes expression between the wild-type parent and the fnr mutant when aerobically or anaerobically grown by qPCR analyses. Levels of crp and cyaA transcript in the fnr mutant when anaerobically grown were still higher than those when aerobically grown as the wild-type parent exhibited (Figure 3). Thus, FNR is not responsible for the increased activity of CRP-cAMP under anaerobic conditions.

We previously found that FNR binds to the region upstream of glpT and uhpT genes (Kurabayashi et al., 2015a). To determine whether CRP-cAMP activates expression of glpT and uhpT genes in cooperation with FNR or not, we investigated an ability of CRP-cAMP binds to the region upstream of glpT and uhpT genes in the presence of FNR by gel-shift assays using recombinant proteins. To maintain the CRP protein in active form that cAMP binds (designated CRP-cAMP), we supplied cAMP into the buffer through the process of protein purification and gel-shift assays. For the FNR protein, we used the mutant protein designated D154AFNR that aspartate at amino acid residue 154 in FNR is replaced alanine because the mutant protein is still able to form an active conformation even in the presence of oxygen and then to bind to the target DNA in the same fashion as that of wild-type FNR (Ziegelhoffer and Kiley, 1995). Consistent with results of our previous study, DNA probe bands from glpT and uhpT, but not rhlR were shifted in the presence of the D154AFNR protein (Figure 4). We also observed a retarded motility of the glpT and uhpT probes when the CRP protein is present, and the motility of these probes was further retarded in the presence of both CRP and D154AFNR (Figure 4). These results suggest that CRP-cAMP and FNR together bind to the region upstream of glpT and uhpT genes, then cooperatively activate expression of glpT and uhpT genes.

CRP-cAMP controls target genes in two different manners depending on locations that it binds within region upstream of target genes (Zhou et al., 1994). When CRP-cAMP binds to a site which is typically distant more than 90 bp from the transcription start point of the target gene, CRP-cAMP requires a regulon specific activator protein for transcriptional activation. It has been known that the uhpT gene is in this case. CRP-cAMP binds to the inverted repeat centered at −103 position which site is located at 30 bp region upstream of the DNA site that UhpA binds, then UhpA is essential for uhpT gene activation (Figure 5D; Merkel et al., 1995; Olekhnovich et al., 1999). The data of our promoter analyses supports it because lacZ expression from pNNuhpT in the uhpA mutant was under detectable limit (Figure 1B). Thus, CRP-cAMP is an enhancer for uhpT gene activation.

As the other manner for CRP-cAMP, it alone activates target genes without any other transcriptional activators. In this case, CRP-cAMP binds to the DNA at near position −61.5, −72.5. −82.5, or −92.5 where CRP-cAMP and RNA polymerase bind on the same face of the DNA helix, or CRP-cAMP partly overlays RNA binding site, then it enables to form direct interaction with RNA polymerase (Straney et al., 1989; Gaston et al., 1990; Ushida and Aiba, 1990). We have shown that levels of glpT expression in crp and cyaA mutants were under detectable limit (Figure 1A). Therefore, we hypothesized that CRP-cAMP may act as an essential element for glpT gene expression interacting with RNA polymerase. To test whether this hypothesis is correct or not, we determined the DNA site which CRP-cAMP binds on the region upstream of glpT gene by DNase I footprinting analyses. We found that a 25-bp region including sequence (AATGTGTGCGGCAATTCACATT) which is similar to CRP-binding consensus motif (AA-TGTGA-[N]-TCACA-TT) was protected from DNase I digestion by CRP-cAMP (Figure 5A), suggesting that the CRP-cAMP binds to this site (Gunasekera et al., 1992). The site is located at −30 to −52 position overlaying −35 box (TCACAT) that RNA polymerase binds (Figure 5C). Thus, CRP-cAMP interacts RNA polymerase, and it acts as an essential element for glpT gene activation. In addition to CRP-cAMP binding site, we determined the DNA site which FNR binds. The D154AFNR protein protected a 70-bp region which locates at 32 bp region upstream of CRP-cAMP binding site (Figure 5B). Within this protected region, there was a sequence (TTGAT-TTCG-CGCAT) located at −123 to −136 position that is relatively similar to the proposed FNR-binding motif (TTGAT-[N]₄-ATCAA) (Ziegelhoffer and Kiley, 1995; Figure 5C).

We also performed DNase I footprinting experiment to determine FNR binding site on uhpT region upstream. However, contrary to the result of gel-shift assays, we did not observe specific binding of the D154AFNR protein on the DNA probe from uhpT region upstream in footprinting experiment (data not shown).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.