In previous papers we described an extra recombination mechanism in T4 phage, which contributed to general recombination only when particular mutations were used as geneticmarkers (high recombination or HR markers), whereas it was practically inactive towards other rIIB mutations (low recombination or LR markers). This marker-dependent recombination pathway was identified as a repair of mismatches in recombination heteroduplexes. We suggested that the first step in this pathway, recognition and incision of the mismatch, is performed by endonuclease VII (endo VII) encoded by the T4 gene 49. In the present paper, we tested this hypothesis in vivo. We used an experimental model system that combines site-specific double-strand breaks with the famous advantages of the recombination analysis of bacteriophage T4 rII mutants. We compared recombination of homoallelic HR and LR markers in the S17 and S17 E727 background (amber mutations in the uvsX and in the uvsX and 49 genes, respectively). In S17-crosses, the HR and LR markers retain their respective high-recombination and low-recombination behavior. In S17 E727-crosses, however, the HR and LR markers show no difference in the recombination frequency and both behave as LR markers. We conclude that endo VII is the enzyme that recognizes mismatches in recombinational heteroduplexes and performs their incision. This role for endo VII was suggested previously from biochemical studies, but this is its first in vivo demonstration.

The double-strand break (DSB) repair via homologous recombination is generally construed as a high-fidelity process. However, some molecular genetic observations show that the recombination and the recombinational DSB repair may be mutagenic and even highly mutagenic. Here we developed an effective and precise method for studying the fidelity of DSB repair in vivo by combining DSBs produced site-specifically by the SegC endonuclease with the famous advantages of the recombination analysis of bacteriophage T4 rII mutants. The method is based on the comparison of the rate of reversion of rII mutation in the presence and in the absence of a DSB repair event initiated in the proximity of the mutation. We observed that DSB repair may moderately (up to 6-fold) increase the apparent reversion frequency, the effect of being dependent on the mutation structure. We also studied the effect of the T4 recombinase deficiency (amber mutation in the uvsX gene) on the fidelity of DSB repair. We observed that DSBs are still repaired via homologous recombination in the uvsX mutants, and the apparent fidelity of this repair is higher than that seen in the wild-type background. The mutator effect of the DSB repair may look unexpected given that most of the normal DNA synthesis in bacteriophage T4 is performed via a recombination-dependent replication (RDR) pathway, which is thought to be indistinguishable from DSB repair. There are three possible explanations for the observed mutagenicity of DSB repair: (1) the origin-dependent (early) DNA replication may be more accurate than the RDR; (2) the step of replication initiation may be more mutagenic than the process of elongation; and (3) the apparent mutagenicity may just reflect some non-randomness in the pool of replicating DNA, i.e., preferential replication of the sequences already involved in replication. We discuss the DSB repair pathway in the absence of UvsX recombinase.

The experimental system combining double-strand breaks (DSBs), produced site-specifically by SegC endonuclease, with the famous advantages of the bacteriophage T4 rII mutant recombination analysis was used here to elucidate the origin of the recombination bias on two sides of the DSB, especially pronounced in gene 39 (topoisomerase II) and gene 59 (41-helicase loader) mutants. Three sources were found to contribute to the bias: (1) the SegC endonuclease may remain bound to the end of the broken DNA and thus protect it from exonuclease degradation; (2) in heteroduplex heterozygotes (HHs), arising as the recombinant products in the left-hand crosses, the transcribed strands are of rII mutant phenotype, so they, in contrast to the right-hand HHs, do not produce plaques on the lawn of the λ-lysogenic host; and (3) the intrinsic polarity of T4 chromosome, reflected in transcription, may be a cause for discrimination of promoter-proximal and promoter-distal DNA sequences. It is shown that the apparent recombination bias does not imply one-sidedness of the DSB repair but just reflects a different depth of the end processing. It is inferred that the cause, underlying the “intrinsic” bias, might be interference between strand exchange and transcription. Topoisomerase and helicase functions are necessary to turn the process in favor of strand exchange. The idea is substantiated that the double-stranded to single-stranded DNA transition edge (not ss-DNA tip) serves as an actual recombinogenic element.

Coordination of DNA ends during double-strand break (DSB) repair was studied in crosses of bacteriophage T4 in which DSBs were induced site-specifically by SegC endonuclease in the DNA of only one of the parents. Coupling of the genetic exchanges to the left and to the right of the DSB was measured in the wild-type genetic background as well as in T4 strains bearing mutations in several recombination genes: 47, uvsX, uvsW, 59, 39 and 61. The observed quantitative correlation between the degree of coupling and position of the recombining markers in relation to the DSB point implies that the two variants of the splice/patch-coupling (SPC) pathway, the “sequential SPC” and the “SPC with fork collision”, operate during DSB repair. In the 47 mutant with or without a das suppressor, coupling of the exchanges was greatly reduced, indicating a crucial role of the 47/46 complex in coupling of the genetic exchanges on the two sides of the DSB. From the observed dependence of the apparent coupling on the intracellular ratio of breakable and unbreakable chromosomes in different genetic backgrounds it is inferred that linking of the DNA ends by 47/46 protein is the mechanism that accounts for their concerted action during DSB repair. A mechanism of replicative resolution of D-loop intermediate (RR pathway) is suggested to explain the phenomenology of DSB repair in DNA arrest and uvsW mutants. A “left”–“right” bias in the recombinogenic action of two DNA ends of the broken chromosome was observed which was particularly prominent in the 59 (41-helicase loader) and 39 (topoisomerase) mutants. Phage topoisomerase II (gp39-52-60) is indispensable for growth in the DNA arrest mutants: the doubles 47− 39−, uvsX 39− and 59− 39− are lethal.

A model system for studying double-strand-break (DSB)-induced genetic recombination in vivo based on the ets1 segCDelta strain of bacteriophage T4 was developed. The ets1, a 66-bp DNA fragment of phage T2L containing the cleavage site for the T4 SegC site-specific endonuclease, was inserted into the proximal part of the T4 rIIB gene. Under segC(+) conditions, the ets1 behaves as a recombination hotspot. Crosses of the ets1 against rII markers located to the left and to the right of ets1 gave similar results, thus demonstrating the equal and symmetrical initiation of recombination by either part of the broken chromosome. Frequency/distance relationships were studied in a series of two- and three-factor crosses with other rIIB and rIIA mutants (all segC(+)) separated from ets1 by 12-2100 bp. The observed relationships were readily interpretable in terms of the modified splice/patch coupling model. The advantages of this localized or focused recombination over that distributed along the chromosome, as a model for studying the recombination-replication pathway in T4 in vivo, are discussed.