Type III restriction is alleviated by bacteriophage (RecE) homologous recombination function but enhanced by bacterial (RecBCD) function

Abstract

Previous works have demonstrated that DNA breaks generated by restriction enzymes stimulate, and are repaired by, homologous recombination with an intact, homologous DNA region through the function of lambdoid bacteriophages lambda and Rac. In the present work, we examined the effect of bacteriophage functions, expressed in bacterial cells, on restriction of an infecting tester phage in a simple plaque formation assay. The efficiency of plaque formation on an Escherichia coli host carrying EcoRI, a type II restriction system, is not increased by the presence of Rac prophage-presumably because, under the single-infection conditions of the plaque assay, a broken phage DNA cannot find a homologue with which to recombine. To our surprise, however, we found that the efficiency of plaque formation in the presence of a type III restriction system, EcoP1 or EcoP15, is increased by the bacteriophage-mediated homologous recombination functions recE and recT of Rac prophage. This type III restriction alleviation does not depend on lar on Rac, unlike type I restriction alleviation. On the other hand, bacterial RecBCD-homologous recombination function enhances type III restriction. These results led us to hypothesize that the action of type III restriction enzymes takes place on replicated or replicating DNA in vivo and leaves daughter DNAs with breaks at nonallelic sites, that bacteriophage-mediated homologous recombination reconstitutes an intact DNA from them, and that RecBCD exonuclease blocks this repair by degradation from the restriction breaks.

FIG. 1.

Gene organization of recET-lar region of Rac prophage. pRAC3 carries a deletion that connects the N terminus of racC with the Cterminus of the recE gene. This fused gene encodes a functional RecE (6, 7, 76). The kanamycin resistance gene is inserted into the lar gene in pNH271. The C terminus of lar is deleted in pNH263. The C-terminal region of pJC980 derivatives is deleted by ClaI digestion, and connected with the ClaI site of the vector. Hin: HindIII; Cla: ClaI; RI: EcoRI; Pst: PstI; Age: AgeI; Xho: XhoI; Hpa: HpaI.

FIG. 2.

Type II restriction activities in the presence and absence of Rac prophage. The bacterial strains carry no plasmid, an EcoRI plasmid (pIK172) or its restriction-negative version (pIK173). The titer of EcoRI-unmodified lambda on each strain relative to that on the strain without any plasmid is shown.

FIG. 3.

Type III restriction activities in bacterial strains with different recombination genotypes. The relative titer of EcoP1 and EcoP15-unmodified lambda on the strain is shown. The left graphs are for E. coli carrying the EcoP1 plasmid (pNR201) or its restriction-negative version (pNH224). The right graphs are for E. coli carrying the EcoP15 plasmid (pNR301) or its restriction-negative version (pNH225). pRAC3, which carries the recET-lar part of Rac (Fig. 1), is present in some of the strains. The values were normalized by the titer on AB1157 carrying pNH224. The recombination or alleviation function coded by the plasmid is indicated in Fig. 1.

FIG. 4.

Effects of lar on type I and type III restriction. A. Type I (EcoKI) restriction alleviation by Lar, encoded by Rac prophage. These values were normalized to the titer of DH5 in which EcoKI restriction is defective. All of the remaining strains are derived from AB1157. B. Type III restriction is not affected by Lar. Restriction activities were measured in strains harboring the EcoP1 plasmid (pNR201) or its restriction-negative version (pNH224) in the presence or absence of lar-expressing plasmid (pIK187). All of the strains are derived from AB1157. The ratio of the number of lambda plaques to that of the strain without any plasmid is shown.

FIG. 5.

Gene requirements for type III restriction alleviation. EcoP1 restriction activities were measured in strains that harboring EcoP1 plasmid (pNR201) or its restriction-negative derivative (pNH224) and a part of Rac prophage. The ratio of the number of lambda plaques on the strain to that of the strain without any plasmid is shown. A. rec⁺ background (AB1157). B. recA mutant background (BIK733). In the recE (exonuclease) columns a and b, Δ(racC-recE)188 and Δ(racC-recE)191 showed the Exo⁺ phenotype (Fig. 1) (6, 40, 76). In the recT column, 950, 951, and 959 are the allele names in Clark et al. (7) as shown in Fig. 1 (also see Materials and Methods).

FIG. 6.

Gene requirements for type III restriction alleviation. JC8679 and its recombination-negative derivatives were examined. Shown on the left is plaque formation in E. coli that carries the EcoP1 plasmid (pNR201) or its restriction-negative derivative (pNH224), while shown on the right is plaque formation in E. coli that carries the EcoP15 plasmid (pNR301) or its restriction-negative derivative (pNH225). Plaque formation efficiencies were normalized to that in the strain without a plasmid.

FIG. 7.

Type I restriction alleviation by recE and recT recombination function. Restriction by EcoKI, a type I system, was measured in strains that express recE, recT, or lar from a plasmid. These values were normalized to the titer of DH5, in which EcoK restriction is defective. The remaining strains are derived from AB1157. a and b, shown in the recE (exonuclease) column, indicate Δ(racC-recE)188 and Δ(racC-recE)191, respectively. In the recT column, 950 is an allele name.

FIG. 8.

Model for type III restriction alleviation. Infected lambda phage genome will be replicated prior to the type III restriction break. Restriction cleavage initiates homologous recombination with the sister chromosome, or the broken chromosome will be degraded by endogenous exonuclease (RecBCD enzyme).