Construction of H-19B derivatives which encode ampicillin resistance.

An antibiotic resistance marker was introduced into the genome of the Stx1-encoding phage H-19B to facilitate the study of the intestinal transmissibility of this lambdoid phage. Tn10d-bla, an 861-bp minitransposon which was constructed to allow identification of phage-encoded exported proteins (19), was used for this purpose. This small element generates translational fusions between the mature portion of β-lactamase and the amino-terminal portion of the target gene product (19). Plasmid pJR207 (19), which harbors Tn10d-bla, was introduced into E. coli C600 H-19B in seven independent transformations, and ampicillin-resistant (Ap^r) colonies were selected as described previously (19). Approximately 10,000 Ap^r colonies were then pooled, washed, and UV induced to produce seven independent H-19B phage pool lysates. The streptomycin-resistant (Sm^r) E. coli strain MC4100 (5) was then used as a recipient strain for transduction with these lysates. Ampicillin-resistant colonies were purified and tested for lysogeny by cross-streaking. Lysogens were then used to produce new phage stocks which were subsequently used to lysogenize MC4100. All lysogens were resistant to ampicillin, thereby confirming the linkage of the Ap^r gene to the H-19B phage genome.

Subsequently one representative H-19B Tn10d-bla insertion from each of the seven independent phage pool lysates was characterized by restriction enzyme and Southern blot analysis. All Tn10d-bla insertions were located within the same 8.1-kb EcoRI fragment known to contain the stxAB operon (9) (Fig. ​(Fig.1).1). Two of the Tn10d-bla fusions were further characterized by subcloning the fusions from the corresponding phage DNA and determining the precise location of the 5′ fusion junctions by using an oligonucleotide corresponding to the 5′ end of the bla gene as a sequencing primer. We found that the Tn10d-bla insertion in H-19B-Ap1 was fused to the coding sequence corresponding to amino acid 70 of the A subunit of Stx1, while the Tn10d-bla insertion in H-19B-Ap2 was fused to the sequence corresponding to amino acid 167 of the A subunit of Stx1 (Fig. ​(Fig.1).1). As anticipated, both of these Tn10d-bla insertions were fused to the correct reading frame of stxA. These results reconfirm the utility of Tn10d-bla insertion mutagenesis for the identification of phage-encoded exported proteins.

Open in a separate window

FIG. 1

Tn10d-bla insertions in the H-19B genome. The 3′ end of the stxAB operon is encoded on a 2.8-kb HindIII (H)-EcoRI (E) fragment (10). The orientation of the stxAB operon along with the locations of the Tn10d-bla insertions in H-19B-Ap1 and H-19B-Ap2 is shown.

The properties of MC4100 H-19B, MC4100 H-19B-Ap1, and MC4100 H-19B-Ap2 were compared. All three lysogens spontaneously produced approximately the same number of virions as measured by PFU per cell (Table ​(Table1).1). Also, virion production from all three lysogens was induced more than 4 orders of magnitude with mitomycin C (Table ​(Table1).1). The Tn10d-bla insertion in MC4100 H-19B-Ap1 and MC4100 H-19B-Ap2 resulted in the complete abrogation of detectable Vero cell cytotoxicity in cell-free supernatants of mitomycin C-induced cultures (data not shown). However, these insertions did not prevent production of the B subunit of Stx1 from MC4100 H-19B-Ap1 or MC4100 H-19B-Ap2 as detected by an enzyme-linked immunosorbent assay (1). Thus, MC4100 H-19B-Ap1 and MC4100 H-19B-Ap2 are Sm^r E. coli strains harboring Ap^r gene-marked detoxified H-19B prophages which, in vitro, produce numbers of infectious H-19B particles similar to those produced by MC4100 H-19B. We used these two Ap^r gene-marked H-19B lysogens to study the intraintestinal transmission of H-19B.

TABLE 1

Virion production by Ap^r gene-marked H-19Blysogens

Strain	Phage titer (PFU/104 CFU)a		Fold inductionc
	Before mitomycin treatment	After mitomycin treatmentb
MC4100 H-19B	4	25×104	6.3×104
MC4100 H-19B-Ap1	11	64×104	5.8×104
MC4100 H-19B-Ap2	7	18×104	2.6×104

Open in a separate window

^aSerial dilutions of chloroform-treated supernatants were plated on E. coli C600. 

^bCultures with an optical density at 600 nm of 0.5 were treated with mitomycin C at 250 ng/ml for 4 h. 

^cCalculated by dividing the number of PFU after mitomycin C treatment by the number of PFU prior to mitomycin C treatment. 

Intraintestinal transduction of phage H-19B.

Oral administration of streptomycin to adult mice has been shown to facilitate the colonization of the murine large intestine with laboratory strains of E. coli (4). We used this model of intestinal colonization in streptomycin-treated mice to investigate whether MC4100 H-19B-Ap1 and MC4100 H-19B-Ap2 produce infectious virions within the intestine. Recovery of intestinally derived transductants of a recipient strain was used as a means to detect intraintestinal virion production.

Streptomycin (0.6 mg/ml) was added to water given to CD-1 mice for 48 h prior to the intragastric administration of MC4100 H-19B-Ap1 (Sm^r Ap^r LacZ⁻), the phage donor strain. Stools collected from mice immediately prior to administration of the phage donor strain were found to have no detectable Sm^r LacZ⁻ CFU. Forty-five minutes after the administration of the phage donor strain, the mice were intragastrically inoculated with a differentially marked potential phage recipient strain, AK16 (MC4100 srl::Tn5), which is identifiable as Sm^r kanamycin-resistant (Kn^r) LacZ⁻ CFU. Approximately 1 × 10⁹ cells of the MC4100 H-19B-Ap1 lysogen and 2 × 10⁹ cells of AK16 were inoculated per mouse. Individual stool samples (pellets) were collected on each of the four days following inoculation, homogenized in 1 ml of Luria broth, and then plated on L agar containing antibiotics to allow the determination of the number of donor CFU (Sm^r Ap^r LacZ⁻), recipient CFU (Sm^r Kn^r LacZ⁻), and transductant CFU (Sm^r Kn^r Ap^r LacZ⁻) in each pellet. Transductants (Ap^r colonies of MC4100 srl::Tn5) were recovered in stool samples on each of the four days following inoculation. Twenty-four hours postinoculation, there was approximately one transductant per 10⁴ donor cells recovered in stool (Table ​(Table2).2). Each of 10 randomly picked Ap^r recipient colonies tested was found to produce StxB by an enzyme-linked immunosorbent assay (1). In addition, filtered supernatants from mitomycin C-induced cultures from each of these AK16 Ap^r transductant colonies was capable of transducing Ap^r to a new recipient strain. These results indicate that H-19B-Ap1 phage transduction accounted for the intestinal transmission of the Ap^r marker. There was a decrease in the frequency of transductants per recipients during the 4-day period when stools were tested (data not shown). At the end of this period the animals were sacrificed, and the frequency of transductants in cecal homogenates was found to be similar to the frequency of transductants in the stool homogenates.

TABLE 2

Detection of intestinally derived transductants at 24h

Donor straina	Mouse	No. of cells in stool (CFU/ml)b			Transductant ratiof
		Recipientc	Donord	Transductantse
MC4100 H-19B-Ap1	1	6.2×105	2.1×106	2.0×102	9.5×10−5
	2	8.8×106	1.1×107	4.2×103	3.8×10−4
	3	4.4×106	4.8×106	8.0×102	1.7×10−4
	4	2.2×106	9.1×105	3.8×101	4.2×10−5
MC4100 H-19B-Ap2	5	7.6×106	5.8×104	4.0×101	6.8×10−4
	6	6.8×106	8.9×104	2.1×102	2.3×10−3

Open in a separate window

^aApproximately 1 × 10⁹ MC4100 H-19B-Ap1 cells or 4 × 10⁶ MC4100 H-19B-Ap2 cells were inoculated intragastrically as donors. Approximately 10⁹ MC4100 srl::Tn5 (Kn^r) cells were intragastrically inoculated 45 min later as potential recipients. 

^bIndividual pellets were resuspended in 1 ml of Luria broth. 

^cThe recipient cell number was determined by counting the number of Sm^r Kn^r LacZ⁻ CFU. 

^dThe donor cell number was determined by counting the number of Sm^r Ap^r LacZ⁻ CFU. 

^eThe number of transductants was determined by counting the number of Sm^r Kn^r Ap^r LacZ⁻ CFU. The assay detection limit is 10 CFU/ml. 

^fCalculated by dividing the number of transductants by the number of donor cells. 

A similarly designed experiment was carried out with MC4100 H-19B-Ap2 as a donor strain in two mice. In this experiment, only 4 × 10⁶ donor cells (more than 2 log units fewer donor cells than used in the previous experiment) were inoculated per mouse. Twenty-four hours later stool samples from both mice contained transductants at approximately the same frequency per donor cell as was observed when MC4100 H-19B-Ap1 was used as a donor (Table ​(Table2).2). Again all tested intestinally derived transductants produced StxB. The detection of H-19B-Ap^r transductants in stool samples and cecal homogenates demonstrates that infectious virions are produced within the host intestine and furthermore that these virions are capable of transducing other E. coli strains within the gastrointestinal tract.

Host environments may be the common site where new pathogenic strains evolve. Our demonstration that an E. coli H-19B lysogen produces infectious phage within the murine gastrointestinal tract suggests that the site of the emergence of the many Stx-producing E. coli serotypes and the Stx-producing non-E. coli genera may have been within the mammalian gastrointestinal tract. Similarly, a recent report by members of our group that Vibrio cholerae CTX lysogens produce infectious virions within the gastrointestinal tract (12) may suggest that new toxinogenic V. cholerae serotypes, such as V. cholerae O139, may arise through phage-mediated transduction of nontoxinogenic strains within the host gastrointestinal tract. Similar phage conversion events may occur on other mucosal surfaces as well. Pappenheimer and Murphy reported a case of corynephage conversion of a nontoxinogenic Corynebacterium diphtheriae strain to toxinogenicity that probably occurred within a woman’s upper respiratory tract (17). It has been proposed that phage conversion within a nonimmune host provides an efficient mechanism for the rapid dissemination of phage-encoded virulence genes (3). It is possible that specific mammalian host signals induce bacterial gene transfer events, as has been described for the plant pathogen Agrobacterium tumefaciens (13).

In the present study, the antibiotic resistance gene-marked H-19B facilitated our ability to monitor intraintestinal phage production by lysogens. The bla-marked H-19B phage will also facilitate the study of the bacterial host range for this virus and the study of intestinal factors, such as diet, and therapeutic interventions, such as antibiotics, that may influence the rate of intestinal H-19B transduction. Our previous work indicates that SOS induction of Stx-converting bacteriophages exerts a regulatory effect on Stx production both by increasing the number of toxin gene copies and by the increased expression of a phage-encoded regulatory molecule (14). Future studies will address the question of whether intraintestinal Stx production and Stx-related disease require the in vivo induction of Stx-converting bacteriophages from lysogens.