Previous Article | Next Article ![]()
Journal of Bacteriology, June 2004, p. 3649-3652, Vol. 186, No. 11
0021-9193/04/$08.00+0 DOI: 10.1128/JB.186.11.3649-3652.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Département de Biochimie et de Microbiologie, Faculté des Sciences et de Génie, Groupe de Recherche en Ecologie Buccale (GREB), Faculté de Médecine Dentaire, Université Laval, Québec, Canada, G1K 7P4
Received 16 October 2003/ Accepted 18 February 2004
|
|
|---|
|
|
|---|
Two general types of phage mutants can be obtained through the selective pressure of Abi systems. The first class is obtained only with P335-like phages, and these mutants result from homologous recombination with phage-related sequences present in the host chromosome (5, 6, 11, 12, 21). To date, these recombination mutants have been isolated during the study of AbiA, AbiC, AbiK, and AbiT only. The second class is found in members of the three groups, and these phages carry only point mutations. Their characterization led to the identification of phage elements that are involved in the sensitivity to lactococcal Abi systems. Two AbiA-insensitive derivatives of phage
31 (species P335) bearing point mutations have been characterized, and the presence of the mutated intergenic region or orf245 of the bacteriophage reduced the efficacy of AbiA (11). On the other hand, the efficacy of AbiD1 was increased by the expression of the wild-type orf1 of phage bIL66 (species 936) (3). AbiK-insensitive class I phage mutants have been characterized previously (6). Here, 10 AbiK-insensitive class II phage derivatives were isolated and analyzed.
Isolation and characterization of AbiK-insensitive phage mutants.
The lactococcal virulent phages ul36 (species P335 [17, 20]), P335 (species P335 [8]),
31 (species P335 [1]), and p2 (species 936 [22]) were selected for challenge experiments because they are sensitive to AbiK (efficiency of plating [EOP], 105 to 107) and have been used in a number of studies on phage-host interactions. First, L. lactis SMQ-88 (15), a strain containing AbiK, was challenged (21) with a high titer of phage ul36 (1010 PFU/ml). Fifty-one ul36 AbiK-insensitive derivatives were isolated (at a frequency of 106), and their genomic restriction profiles were compared as described previously (21). Forty-four derivatives had a different restriction profile (class I phage mutants), while seven mutants had the wild-type ul36 restriction profile (class II mutants). A similar experiment was performed by using another phage-host system. L. lactis F7/2 (8) expressing AbiK in the vector pSRQ817 (15) was challenged with phage P335. Ten AbiK-insensitive phages were isolated. Only one phage was a class I mutant, whereas the nine others were class II mutants. Thus, the proportion of class I and II mutants was different for this phage-host system.
A third P335 phage-host system including phage
31 and L. lactis SMQ-481 (6) was studied. The six AbiK-insensitive
31 derivatives analyzed were class II mutants. Finally, L. lactis MG1363, harboring AbiK, was challenged with the virulent phage p2 (species 936); the nine AbiK-insensitive mutants had the wild-type restriction profile (class II). Ten class II mutants were selected for further study: five AbiK-insensitive derivatives of phage ul36, three of phage p2, one of phage P335, and one of phage
31.
Microbiological characterization of the AbiK-insensitive phage mutants. The EOP on the AbiK+ strains varied between 0.3 and 0.5 for the three p2 derivatives and between 0.7 and 1.0 for the seven class II mutants of the P335 species. These results indicate that the p2 mutants are still slightly inhibited by AbiK. Then, L. lactis SMQ86 (AbiK [15]) and L. lactis SMQ88 (AbiK+) were simultaneously infected with wild-type phage ul36 and one class II mutant to determine the efficiency with which they form a center of infection (ECOI) (19). The multiplicity of infection was six in infection assays with one phage (ul36 or its mutant) and three for each phage in coinfection experiments. For the wild-type phage ul36, the ECOI was 12% ± 5%. Thus, as expected, AbiK reduces drastically the percentage of cells that release at least one infectious particle (7, 15). In the case of class II mutants ul36.15 and ul36.16, almost all infected AbiK+ cells released phages (ECOI of 90% ± 4% and 107% ± 25%, respectively). When the AbiK+ host was coinfected with ul36 and a mutant, the formation of infectious centers was intermediate (51% ± 24% for ul36.15 and 38% ± 9% for ul36.16), indicating that the alleles are codominant. Similar results were obtained with phage p2 (7).
Identification of sak in phage ul36. Genome analysis of class I mutants of phage ul36 previously showed the replacement of a set of three genes (orf131, orf252, and ssb) that conferred insensitivity to AbiK (6). This genomic region was sequenced in five class II ul36 AbiK-insensitive derivatives. Each mutant had one point mutation in orf252, which was renamed sak (mnemonic for sensitivity to AbiK). Altogether, only three different amino acid substitutions were observed in the 252-amino-acid deduced protein (Table 1). The Sak protein possesses the conserved domain involved in DNA binding of the eukaryotic protein RAD52 (pfam04098.3). Recently, it was proposed to include Orf252 (Sak) of phage ul36 in the RAD52-like superfamily of single-strand annealing proteins (SSAPs) (16). Two other homologues of Sak were found in Salmonella enterica serovar Typhimurium and Escherichia coli bacteriophages, but the amino acid sequence similarity was confined to the N-terminal region (Fig. 1A and Table 1). Interestingly, the last 56 amino acids (aa) of both homologues share similarities with the C-terminal portion of Erf from Salmonella phage P22, which is also an SSAP (Fig. 1).
|
View this table: [in a new window] |
TABLE 1. Characteristics of the four phage genes involved in sensitivity to AbiK
|
![]() View larger version (28K): [in a new window] |
FIG. 1. Putative modules of recombination according to bioinformatic analyses, genome organization, and the classification of SSAPs into three superfamilies (16). (A) ORFs encoding homologues of Sak (RAD52 superfamily); (B) homologues of Sak2 or Erf; (C) homologues of Sak3; (D) homologues of Sak4; (E) members of the RecT/Beta superfamily. Neighboring genes represented by arrows of the same color have amino acid sequence similarity, except for ORFs represented in white, which have no significant similarity with other ORFs in the figure. Phages sensitive to AbiK are in blue, and those that are insensitive to AbiK are in red. The sensitivity to AbiK was not determined for phage names typed in black. The hosts of the bacteriophages are E. coli (E.c.), S. enterica serovar Typhimurium (Sa.t.), L. lactis (L.l.), Streptococcus pyogenes (S.p.), Lactobacillus johnsonii (L.j.), Streptococcus thermophilus (St.t.), and Bacillus subtilis (B.s.).
|
31 was also identified (Fig. 1D and Table 1). Thus, at least four nonsimilar proteins are involved in sensitivity to AbiK, and in two cases (sak of ul36 and sak2 of P335) their homologues are single-strand annealing proteins. Overexpression of sak3 on a plasmid. To confirm the role of sak in the AbiK mechanism, the gene was cloned on a plasmid downstream of a strong constitutive lactococcal promoter (31). When gene sak3 (phage p2) was expressed from the P59 promoter (31), the efficacy of AbiK was increased (EOP, 106 to 107). Moreover, the expression of the mutated Sak3 protein (S144Y) reduced the impact of the AbiK system on p2 (EOP, 106 to 105). These data confirm the involvement of sak3 in the sensitivity to the phage defense mechanism.
Growth of Sak mutants on recombination-deficient L. lactis strains. Two L. lactis MG1363 derivatives that have altered recombination activity (13, 14) were tested as hosts for phage p2 and its AbiK-insensitive mutants. In L. lactis VEL1222 and REX2, the RecA recombinase and the RexAB exonuclease (analog of RecBCD), respectively, have been inactivated. Phage p2 and the three class II mutants had an EOP of 1 on both strains. AbiK (pSRQ817) was also introduced in the recombination-deficient strains to examine its antiphage activity. The efficacy of AbiK was not modified by the inactivation of RecA or RexAB.
Single-strand annealing proteins. SSAPs are a group of recombinases that share similar quaternary structures and functions despite the lack of sequence similarity. SSAPs of bacteriophages are involved in genome circularization following DNA entry, in DNA repair of concatemers, and in some cases, in DNA replication (2, 4, 30). They bind single-stranded DNA and promote the annealing of complementary strands (16, 23). Assuming that sak genes encode proteins with similar functions or biochemical activities, the flanking genes of unknown function are therefore likely to play a role in homologous recombination or DNA repair (Fig. 1) (26, 27). The point mutations that confer insensitivity to AbiK were essentially located in the N-terminal region (first 100 aa) of the Sak proteins. According to the alignment of ul36 Sak with RAD52, the substitutions are located in nonconserved residues of the DNA binding and ring formation domain (16). The fact that an intermediary phage resistance phenotype was observed in coinfection experiments suggests that wild-type and mutated alleles of Sak do not interact with each other or form hetero-oligomers with a modified single-strand annealing activity. Moreover, the three phage p2 mutants replicated normally on recombination-deficient strains. It is likely that mutated Sak, Sak2, and Sak3 proteins allow the phage to circumvent AbiK while maintaining its recombination activities.
The first 209 aa of the human RAD52 protein, which correspond to the first 166 aa in ul36 Sak, promote single-strand annealing in vitro (29). This domain resides next to the RPA-binding domain in RAD52 (aa 220 to 280). RPA is a single-strand binding protein, and its interaction with RAD52 is essential for in vivo homologous recombination (25). Therefore, the last 86 aa of Sak could also interact with another protein involved in the recombination pathway. The latter protein could be an SSB protein, as with RAD52, or an exonuclease, as observed with the Beta protein of lambda (28). Several proteins of the Erf superfamily share a low level of sequence similarity over their C-terminal regions. For example, Erf of P22 and Erf of HK022 have 93% similarity over their first 152 aa but only 39% for the last 50 aa. Erf of P22 and Erf of P335 also display 61% similarity over the 133-aa N-terminal region but less than 30% for the last 67 aa. Moreover, members of the RAD52 or RecT/Bet superfamilies possess C-terminal regions (56 to 60 aa) that are homologous to Erf-like proteins (16). We propose that SSAPs of bacteriophages possess two domains. The first domain closely corresponds to the motifs described by Iyer et al. (16) and is sufficient for DNA binding and ring formation. The second domain is located at the C-terminal region and is necessary for homologous recombination in vivo (24, 25).
Nucleotide sequence accession numbers. The sequences were submitted to the GenBank database under accession numbers AY365422 (phage p2) and AY365423 (P335).
J.D.B is a recipient of graduate scholarships from the Fonds FCAR and the NSERC of Canada. This study was funded by grants (CRD and Strategic) from NSERC and Agropur to S.M.
|
|
|---|
red-dependent growth and recombination of phage P22. Virology 134:161-167.[CrossRef][Medline]
and P22. Genetics 134:1013-1021.[Abstract]
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»