Resuscitation of a Defective Prophage in Salmonella Cocultures

Abstract

Widely studied Salmonella enterica serovar Typhimurium strains ATCC 14028s and SL1344 harbor a cryptic ST64B prophage unable to produce infectious virions. We found that coculturing either strain with an isogenic sibling lacking the prophage leads to the appearance of active forms of the virus. Active phage originates from reversion of a +1 frameshift mutation at a monotonous G:C run in a presumptive tail assembly pseudogene.

Prophages account for most of the genomic diversity among closely related Salmonella strains. As an example, the two Salmonella enterica serovar Typhimurium isolates generally used as model systems for virulence studies worldwide, strains ATCC 14028s (3) and SL1344 (7), have partially distinct prophage repertoires. In addition to the Gifsy-1 and Gifsy-2 prophages found in both strains, strain ATCC 14028s contains prophage Gifsy-3, which is absent from SL1344 (5). This prophage encodes type III secreted effector protein SspH1 and the stress response regulator irsA (5; F. Heffron, personal communication). Conversely, strain SL1344 carries sopEΦ, which is not found in ATCC 14028s and encodes secreted G nucleotide exchange factor SopE (5, 6). All of the above prophages are fully functional and produce infectious particles upon induction by standard treatments or spontaneously at lower levels (4, 5).

Strains ATCC 14028s and SL1344 carry a defective ST64B prophage.

While characterizing the prophage complement of strain ATCC 14028s, a few years ago, we tested whether a derivative cured for Gifsy-1, Gifsy-2, and Gifsy-3 would still release phage. Aliquots from the supernatant of a culture treated with mitomycin C (MitC) were spotted on an array of tester strains including isolates from serovars other than Typhimurium. Only a single plaque could be identified (with a serovar Gallinarum host). Phage propagated from the plaque was subjected to limited DNA sequence analysis. The results did not reveal any significant similarity with sequences known at the time. Recently, however, new perusal of DNA databases showed our sequence segments to correspond to various portions of the genome from Salmonella phage ST64B (GenBank accession number AY055382). Mmolawa and coworkers identified ST64B in an induced culture from an epidemic lysotype DT64 strain (9). The authors of that work visualized the virion particles by electron microscopy and could purify phage DNA for sequence analysis; intriguingly, however, they were unable to propagate the virus in any strain tested, leading them to conclude that their ST64B isolate lacked infection capability, possibly due to a tail defect (9).

The phage ST64B sequence includes a serine tRNA gene (serU) segment near the 3′ end of the putative int gene, suggesting that the prophage is inserted at the serU locus, adjacent to the umuCD operon. We used oligonucleotides complementary to regions from either side of the predicted phage chromosomal boundaries as primers for PCR. Fragments corresponding to the predicted attL and attR sequences were amplified with strains ATCC 14028s and SL1344, whereas a fragment corresponding to the attB sequence, which is diagnostic for the absence of the prophage, was obtained with strain LT2 (data not shown). To assess the functional status of ST64B-like prophages in strains ATCC 14028s and SL1344, prophage-deleted derivatives expected to become sensitive to infection were constructed by Red-mediated exchange with a PCR-amplified fragment (2). This work was carried out with Gifsy-cured strains MA6052 (ATCC 14028s) and MA6247 (SL1344). The resulting strains, MA7549 and MA7551 (Table 1 and Fig. 1), were used as recipients for phage infection. Aliquots (0.1 ml) from the supernatants of either unchallenged or MitC-treated cultures of MA6052 and MA6247 were overlaid on lawns of the corresponding recipient strains. Despite repeated attempts (testing different plating procedures), plaques were never observed with the samples from the untreated cultures, while a total of two plaques were obtained from cultures exposed to MitC in separate experiments. Altogether, these data suggested that the ST64B prophage of both ATCC 14028s and SL1344 is defective in regard to some step needed for the formation of infectious particles, but it can occasionally revert to produce functional phage.

TABLE 1.

List of strains used in this study^a

Strain	Relevant genotypeb	Source or reference
ATCC 14028s	Wild type	3
MA6052	Gifsy-1[−] Gifsy-2[−] Gifsy-3[−]	5
MA7549	Gifsy-1[−] Gifsy-2[−] Gifsy-3[−] Δ[ST64B]::kan	This work
MA7566	Gifsy-1[−] Gifsy-2[−]Gifsy-3[−] ST64B1344[+]	This work
MA7747	Gifsy-1[−] Gifsy-2[−] Gifsy-3[−] ST64B14028[+]	This work
SL1344	Wild type	7
MA6247	Gifsy-1[−] Gifsy-2[−]	4
MA7551	Gifsy-1[−]Gifsy-2[−] Δ[ST64B]::kan	This work
MA7569	Gifsy-1[−]Gifsy-2[−] ST64B1344[+]	This work
SSM859 (DT104)	Wild type	S. Uzzau
MA7726	Δ(ST64B)::kan	This work
RKI4898/03 (DT104)	Wild type	W. Rabsch
MA7727	Δ(ST64B)::kan	This work
RKI6256/03 (DT104)	Wild type	W. Rabsch
MA7728	Δ(ST64B)::kan	This work

^aAll strains belong to Salmonella enterica serovar Typhimurium. Strain SSM859, a phage-type DT104 isolate from Zimbabwe, was a gift of S. Uzzau, Department of Biomedical Sciences, University of Sassari, Sassari, Italy. Strains RKI4898/03 and RKI6256/03, isolated in Germany, were kindly donated by W. Rabsch, Robert Koch Institute, Wernigerode, Germany. Transfer of the Δ(ST64B)::kan swap construct into these strains was achieved by phage P22-mediated transduction (selecting for kanamycin resistance) by using a donor lysate made with strain MA7549.

^bOnly the changes from the native prophage complement of the strain are indicated. Prophages with the same name don't have identical sequences in different strains (e.g., compare GenBank accession numbers AY552603, AY552604, and AY552605). A superscript number specifies a functional variant of an initially inactive ST64B prophage (the number refers to the strain the phage originates from).

FIG. 1.

Schematic representation of the ST64B prophage (left and right portions) and the Δ(ST64B)::kan swap construct. The ORF map combines information from the reports of McClelland et al. (8) and Mmolawa et al. (9), from the Sanger Institute, (http://www.sanger.ac.uk/; Salmonella spp. comparative sequencing), and from this work. The Δ(ST64B)::kan swap construct was made according to the method of Datsenko and Wanner (2). PCR Primers were 60 bases long, with the last 20 bases (indicated by italics) annealing to template plasmid pKD4 (2). Primer sequences used were ACTGTACTTCTGCTTGTCTTTTGCCGTTCCCTCATAGTCTCATATGAATATCCTCCTTAG (pp339) and TTAACTCCCTTCCGGTTAGCCGATAACAGAATCCAGTACATGTAGGCTGGAGCTGCTTCG (pp340). Red-mediated recombination resulted in the replacement of the segment between coordinates 293 and 39819 of the ST64B genome map with a kanamycin resistance gene (kan). The procedure was carried out separately in strains MA6052 and MA6247 (Table 1) and yielded strains MA7549 and MA7551, respectively. A filled arrow above the prophage map shows the ORF generated by a reversion of the +1 frameshift mutation (see the text).

Accumulation of ST64B phage revertants in cocultures.

We have shown that phage originating from spontaneous induction of a lysogenic strain can multiply if this strain is cocultured with an isogenic sibling that lacks the prophage (1). We figured that this feature might be exploited to select for amplification of rare ST64B revertants that may arise spontaneously in a culture. Cocultures were inoculated with a 1:5 mixture of bacteria from parental strain MA6052 or MA6247 and the respective ST64B prophage-deleted derivative MA7549 or MA7551. Bacteria (approximately 1,000 in the inoculum) were allowed to reproduce at 37°C without agitation until stationary phase. Cultures were then diluted 200-fold and incubated further under the same conditions for 6 to 12 h. The subculturing routine was repeated four times. At each dilution step, aliquots from supernatants were tested for the presence of plaque-forming particles. As predicted, plaques were observed as early as the first subculture, and their number rose sharply, reaching a peak between the second and third subculturing step. Phage obtained from resuspended plaques was confirmed to have ST64B immunity. This phage was used to isolate newly lysogenized derivatives. Interestingly, in most instances, the lysogenization event was accompanied by the loss of kanamycin resistance, suggesting that the incoming virus is able to dislodge the kan cassette occupying the att site in the recipient strains (Fig. 1). (This property is not unique to ST64B phage; work in our laboratory has shown that Salmonella Fels-2 and SopEΦ bacteriophages [11] can dislodge each other's prophage from the chromosome [our unpublished data]. Unlike the ATCC 14028s- and SL1344-derived parents, newly lysogenized strains release active ST64B phage spontaneously at significant frequencies (approximately 3 × 10⁴ PFU/ml in a full-density grown culture), and phage yield increases by three orders of magnitude following exposure to MitC. As initially postulated, the coculture conditions appeared to have led to the amplification of an ST64B phage revertant.

Reactivation of ST64B phage results from reversion of a +1 frameshift mutation in a tail operon gene.

To gather clues as to the possible nature of prophage alteration, the ST64B sequence (9) was compared to other phage genomes by using the BLASTX program. This analysis revealed that two putative genes lying adjacent to each other in different frames in the ST64B tail operon (coding for proteins SB21 and SB22) (Fig. 1) are similar to the two halves of an uninterrupted open reading frame (ORF) in other phages, including phages P27 and Mu (see below). This finding suggested the presence of a frameshift mutation in the ST64B ORF. Consistent with this hypothesis, a monotonous G:C run (a typical frameshift mutation hotspot [13]) could be seen immediately upstream of the stop codon terminating SB21 translation. To analyze the structure of this region in the ST64B prophage of ATCC 14028s- and SL1344-derived strains, DNA fragments spanning the G:C repeat were amplified by PCR and subjected to sequence analysis. Results showed that in the case of both parental strains (MA6052 and MA6247), the same 8-bp G:C run found in the published ST64B sequence is present (Fig. 2). In contrast, the G:C repeat is 1 bp shorter in the derivatives lysogenized by the active form of the virus (strains MA7566 and MA7569), resulting in the fusion of the SB21 and SB22 ORFs in a single reading frame (Fig. 2). These findings strongly suggest that reversion of the +1 frameshift mutation is directly responsible for the reactivation of the phage. The full-length version of the SB21 ORF is hereafter referred to as SB21*. Sequences were deposited in the National Center for Biotechnology Information database (GenBank accession numbers AY552603 and AY552604).

FIG. 2.

DNA sequence analysis of the region containing the end of the SB21 ORF of prophage ST64B in strains MA6052 (ATCC 14028s) and MA6247 (SL1344) and in strains lysogenized by active phage variants. The region of interest was PCR amplified from the chromosome of the indicated strains and subjected to DNA sequence analysis. The following oligonucleotides were used as primers for PCR amplification and sequencing: pp439 (TGCCGGTTATTGCTGATG) and pp440 (CGGCAAAATATGGTCACG).

Epidemic DT104 strains carry a functional ST64B-like prophage with an uninterrupted SB21* ORF.

The presence of the ST64B prophage in clinical, phage-type DT104 isolates was assessed by PCR amplification of the predicted attL and attR boundaries. This analysis showed that all three strains analyzed (Table 1) carried the prophage (data not shown). We then examined the status of the SB21-SB22 region. DNA sequence analysis of a PCR-amplified fragment showed that all three strains contain a unique ORF with a G:C repeat of the same length (7 bp) as that in the revertants described above (data not shown) (GenBank accession numbers AY552605, AY574196, and AY574197). To test whether these strains produce functional ST64B phage, derivatives in which the prophage was deleted were constructed by transduction (Table 1) and used as recipients to detect the presence of plaque-forming particles in cultures of the respective parent strains that were either unchallenged or exposed to MitC. Plaques were observed in all cases (less than 10³ PFU/ml in untreated cultures grown to stationary phase, increasing to approximately 10⁶ PFU in the MitC-treated samples). These values are lower than those observed with the strains carrying the prophage revertants; nonetheless, they clearly indicate that all three DT104 strains harbor a functional ST64B prophage. In each case, the resulting phage was found to have the same immunity as ST64B phage isolated from ATCC 14028s and SL1344 (ST64B¹⁴⁰²⁸ and ST64B¹³⁴⁴); like the latter, they showed the ability to displace a kanamycin resistance marker from the att site upon integration (data not shown).

The predicted product of the SB21* ORF is a 359-amino-acid protein which resembles proteins l52 and gp47 of phages P27 (12) and Mu (10), respectively (87 and 44% overall similarity). Although the function of these proteins is unknown, the location of their coding sequences within the tail operon of the respective phages, in the vicinity of a putative baseplate assembly gene, suggests their possible involvement in tail assembly. The apparent lack of visible tails in the ST64B phage preparation of Mmolawa et al. (9) is generally consistent with this idea. Also consistent with this idea is the recent finding that cultures from strain SL1344 dam methylase mutants accumulate ST64B capsid protein in their supernatants, but no infectious virions are produced (J. Casadesus and F. Garcia del Portillo, personal communication). Apparently, the defective ST64B prophage in strain SL1344 is induced in the dam mutant background, confirming that the SB21* alteration does not hamper the induction process. The data presented here strongly suggest that the +1 frameshift mutation in the SB21* gene is the sole defect of the ST64B prophage in strains SL1344 and ATCC 14028s. Coculturing either of these strains with strains that support ST64B phage growth rapidly selects for revertants of the mutation. Conceivably, this mechanism allows regeneration of the virus in an environment where mixed strain conditions may not be uncommon. In contrast, since the +1 frameshift in the SB21* gene does not relieve the genomic burden associated with prophage presence, or the capacity to undergo spontaneous induction (lethal to the host), the forces that selected for the mutation in strains SL1344 and ATCC 14028s, or in a common ancestor, remain elusive.

Nucleotide sequence accession numbers. The nucleotide sequences for the portions of the SB21 gene analyzed in this study have been deposited in GenBank under accession numbers AY552603, AY552604, AY552605, AY574196, and AY574197.