Induction of Prophages of Enterohemorrhagic Escherichia coli O157:H7 with Norfloxacin

Abstract

Norfloxacin (NFLX) caused induction of prophages VT1 and VT2 of enterohemorrhagic Escherichia coli O157 at subinhibitory concentrations. In time course experiments, we observed the following sequential events: upon induction, the phage genomes underwent multiplication; the amount of stx genes increased; and subsequently, large quantities of toxins VT1 and VT2 were produced. Further studies showed that the molecular mechanism of prophage induction is closely related to the RecA system since the prophage VT2 was not induced with NFLX in a recA mutant strain.

In the summer of 1996 in Japan, large outbreaks of enterohemorrhagic Escherichia coli (EHEC) O157:H7 infection took place, especially in Sakai City, Osaka Prefecture. Many school children, totalling thousands of cases, suffered from diarrhea and hemorrhagic colitis, and some developed hemolytic-uremic syndrome (HUS) characterized by acute renal failure, thrombocytopenia, and microangiopathic hemolytic anemia, resulting in 12 deaths. Mechanisms causing the above conditions have been investigated from various aspects. Vero toxin (VT) (4, 17) and other possible virulence factors of EHEC O157 (8, 11, 12) and directions for the use of drugs, especially antimicrobial agents, are also being re-evaluated. The 1997 guideline for medical treatment of EHEC O157 infection issued by the Ministry of Health and Welfare in Japan (13) has recommended the use of kanamycin, fosfomycin, and a new quinolone antimicrobial agent, norfloxacin (NFLX). However, Walterspiel et al. (22) reported that, instead, these various antimicrobial agents, including the new quinolone, induced the production of VT under certain conditions. Thus, to solve the apparent contradictions, we have started to investigate the induction of the prophage carrying the stx gene by NFLX. In this communication, we show that NFLX indeed induces the stx-converting phages and suggest the following underlying mechanism: NFLX is an inhibitor specific for the A subunit of DNA gyrase (7, 23), the drug inhibits DNA synthesis in E. coli and causes an accumulation of single-stranded DNA fragments capable of activating the RecA protein (2, 18), and the activated RecA protein promotes the cleavage of a phage repressor, allowing prophage induction (2, 19). Thus, as a result of this process, the phage in vegetative growth starts VT phage DNA replication and toxin production.

NFLX triggers prophage induction.

We have examined whether NFLX could cause prophage induction, resulting in multiplication of stx genes and mass production of VT. Experiments were performed with EHEC O157 strain RIMD0509894. This strain, isolated during the EHEC outbreak in Sakai City in 1996 and provided by the Research Center for Emerging Infectious Diseases of Osaka University, was doubly lysogenic for phages VT1 and VT2, designated VT1-Sa and VT2-Sa, respectively. Overnight cultures in Luria-Bertani broth were subcultured by shaking for 1 h to the early logarithmic growth phase. They were treated with various subinhibitory concentrations of NFLX (provided by Kyorin Pharmaceuticals, Inc.) for prophage induction for 30 min, centrifuged, resuspended in drug-free Luria-Bertani broth, and subsequently cultured for phage growth and toxin production for a further 2 h. Infective-center assays were carried out by plating cells within a few minutes after induction together with the indicator E. coli K-12 W3110, which is nonlysogenic for λ, and on the following day, the plaques that appeared were counted. In the experiment, the efficiency of prophage induction corresponded to the number of infective centers divided by the number of plated induced cells. The number of infective centers increased with the concentration of NFLX up to 1 μg/ml and then declined (Fig. 1a). In parallel with prophage induction, production of two toxins, VT1 and VT2, showed patterns similar to that of prophage induction, although the level of VT2 was always higher than that of VT1 (Fig. 1b). These results indicated that both the infective centers and the toxin amount produced increased 100- to ∼1,000-fold in the presence of subinhibitory concentrations (0.1 to 10 μg/ml) of NFLX compared to those found in the absence of the drug and that those obtained with NFLX (1 μg/ml, 30 min) were almost same as those obtained with an optimal dose of UV light irradiation (1.66 J/m², 30 s) (Fig. 1a).

FIG. 1.

Prophage induction in EHEC O157 strain RIMD0509894. (a) Efficiency of prophage induction (closed circles, NFLX; open circle, application of UV light for 30 s). The assay was performed as described in the text. Since this strain is a double lysogen for phages VT1 and VT2, infective centers contain both phages. (b) Toxin production. For assay of toxins, E. coli cultures, after incubation for 2 h, were centrifuged at 5,000 × g for 5 min. The amounts of VT1 and VT2 in the supernatants were then separately determined by using a reversed passive latex agglutination test kit (E. coli Verotoxin Detection Kit; Denka Seiken Co., Tokyo, Japan) as described in the manufacturer’s manual.

Time course analysis after induction.

Figure 2 illustrates results of time course experiments with EHEC strain RIMD0509894. After induction, toxin production and synthesis of stx-specific DNA were monitored. All of these values increased 10- to ∼100-fold with time and reached plateaus after 60 to 90 min. In this experiment, we employed the probe specific for VT1 or VT2, neither of which cross-hybridized with the other. When a probe of the other fragment derived from VT2 phage, which was irrelevant to the stx DNA was used, essentially the same results were obtained (data not shown). Thus, the increase in the stx gene indicated replication of the phage DNA. Concerning the two kinds of VT in this strain, VT1-Sa and VT2-Sa, a difference of 1 order of magnitude between the production of VT1 and that of VT2 was observed, production of VT2 being higher than that of VT1. This result was consistent with the observation that the synthesized amounts of the DNAs encoding these toxins also differed by roughly 1 order of magnitude. Since the copy numbers of prophages VT1 and VT2 per genome were identical to each other, the growth of phage VT2 should be much faster than that of VT1.

FIG. 2.

Time course experiments using EHEC O157 strain RIMD0509894. Panels: a, toxins VT1 and VT2; b, VT1- and VT2-specific DNAs. The VT1 and VT2 toxin DNA assay method used is as follows. After induction for 30 min with 1-μg/ml NFLX, 1-μl aliquot of a culture was spotted onto each of two nitrocellulose filters (Nitroplus 2000; Micron Separations Inc., Westborough, Mass.) every 30 min. One filter was hybridized with fluorescein-labeled fragments containing VT1 gene DNA (100 ng), and the other was hybridized with VT2 gene DNA in 10 ml of 6× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–5× Denhardt’s solution–0.5% sodium dodecyl sulfate–4 mM EDTA–100-μg/ml salmon sperm DNA. After being washed in 0.5× SSC at 60°C, the filter was incubated with an anti-fluorescein–alkaline phosphatase conjugate and visualized with the fluorescence substrate Atto Phos (Amersham Pharmacia Biotech, Ltd., Uppsala, Sweden). The fluorescence signals were detected with a FluorImager (Molecular Dynamics Inc., Sunnyvale, Calif.). The amount of VT1- and VT2-specific DNA was expressed in arbitrary fluorescence intensity units as measured at 540 to 560 nm (excitation at 488 nm).

Concerning toxin gene multiplication and toxin protein production after prophage induction, the following two points are worth mentioning. Firstly, we have determined the complete nucleotide sequence of phage VT2-Sa. This result (14) and recent reports (3, 16) have together demonstrated, through a search for sequence homology between VT2-Sa and phage λ, that the stx gene is located downstream of the Q gene homologue and upstream of the homologues of the S and R genes encoding lytic enzymes. This indicates that the stx gene is likely transcribed from the p_R′ promoter as a part of late gene expression under the influence of the Q gene product synthesized after inactivation of the phage repressor (6, 16, 20). We observed that from 90 to 120 min after induction with NFLX, phage maturation and the subsequent burst were always accompanied by complete cell lysis (data not shown). These results led to the following conclusion: NFLX causes prophage induction, resulting in an increase in the copy number and transcription of the stx gene and cell lysis-mediated massive toxin release.

Secondly, we further confirmed the toxin production by using a different strain. We have examined VT1-Sa and VT2-Sa from strain RIMD0509894, which was isolated during the outbreak in Sakai City in 1996. Another EHEC O157:H7 strain, V141, which is singly lysogenic for VT2 and was isolated in Tokyo by Tae Takeda in 1996, is known to produce phage VT2-141, which has immunity different from that of VT2-Sa. Analyses of these strains indicated that the two strains have similar patterns of prophage induction by NFLX, as well as similar time course profiles for stx gene multiplication and VT production. Thus, induction of the toxin-converting phage by NFLX appears to be a general phenomenon.

Mitomycin-induced increases in toxin production were reported in EHEC H18 (O128:H12) (15, 24) and Shigella species (21). However, in these cases, the stx gene was located on the host chromosome. Some of these cases may be accounted for by the heteroimmune induction by prophages other than the toxin-converting phages.

RecA protein plays a key role in prophage induction.

VT2-Sa single lysogens were prepared by spotting phage VT2-Sa on lawns of E. coli K-12 W3110 recA⁺ and isogenic recA mutant strains and verified by PCR with primers based on a part of the VT2 gene sequence. These lysogens were examined for prophage induction, and the results are shown in Fig. 3. An increase in VT2 toxin DNA was observed after induction in the recA⁺ strain but not in the recA mutant strain. These results indicated a crucial role of RecA protein in VT2 prophage induction.

FIG. 3.

Induction with NFLX of E. coli K-12 W3110 recA⁺ and recA mutant strains both singly lysogenic for VT2-Sa. Curves: a, VT2-specific DNA synthesized in the recA⁺ strain (VT2-Sa); b, VT2-specific DNA synthesized in the recA mutant (VT2-Sa); c, VT2-specific DNA synthesized in the presence of NFLX.

stx-specific DNA synthesis in the recA⁺ lysogen was also examined in the continued presence of NFLX (Fig. 3c). However, under these conditions, neither toxin gene multiplication nor toxin protein synthesis took place. It was concluded that NFLX not only triggers prophage induction but also inhibits DNA synthesis of the toxin genes in its continued presence. Hence, continued inhibition of DNA synthesis in the bacteria and phages leads to abrogation of toxin production. However, it should be pointed out that although NFLX is effective for treatment of O157 infection, discontinuation of medication halfway in the course of treatment may occasionally stimulate toxin production. This conclusion should be a useful guideline for the medical treatment of EHEC O157 infection.

As far as we have determined, of the antimicrobial agents available, only an inhibitor of peptidoglycan synthesis, fosfomycin, and an inhibitor of protein synthesis, kanamycin, have not caused any increase in toxin production. In contrast, DNA synthesis inhibitors such as the new quinolones trimethoprim (9) and mitomycin C (1, 24) have been reported to augment toxin production by various EHEC strains in vitro, as well as in vivo. A high incidence of cancer-associated HUS after mitomycin treatment (5, 10) could be related to such toxin production.

As described in this report, the relative amount of VT produced by noninduced cells was less than 10⁻³ compared with that produced by induced cells. Release of toxin proteins in large quantities outside the cell may be due to a burst of phage-mediated lytic-enzyme released by E. coli. However, every case of HUS is not necessarily brought about by antimicrobial agents; there must also be some inducers or triggers in the intestinal lumen that cause serious systemic toxin-related complications like HUS.