Mismatch repair in recombination of bacteriophage T4

Biomol Concepts. 2012 Dec;3(6):523-34. doi: 10.1515/bmc-2012-0021.

Abstract

Abstract The review focuses on the mechanism of mismatch repair in bacteriophage T4. It was first observed in T4 as an extra recombination mechanism, which contributed to the general recombination only when particular rII mutations were used as genetic markers (high-recombination markers), whereas it was inactive toward other rII mutations (low-recombination markers). This marker-dependent recombination pathway was identified as a repair of mismatches in recombinational heteroduplexes. Comparison of the structure of markers enabled us to make several specific conclusions on the nature of the marker discrimination by the mismatch repair system operating during T4 crosses. First, heteroduplexes with one mismatched base pair (either of transition or of transversion type) as well as single-nucleotide mismatches of indel type are not efficiently repaired. Second, among the repairable mismatches, those with two or more contiguous mismatched nucleotides are the most effectively repaired, whereas insertion of one correct pair between two mismatched ones reduces the repairability. Third, heteroduplexes containing insertion mutations are repaired asymmetrically, the longer strand being preferentially removed. Fourth, the sequence environment is an important factor. Inspection of the sequences flanking mismatches shows that runs of A:T pairs directly neighboring the mismatches greatly promote repair. The mismatch is recognized by T4 endonuclease VII and nicked on the 3' side. The nonpaired 3' terminus is attacked by the proofreading 3'→5' exonuclease of T4 DNA polymerase that removes the mismatched nucleotides along with several (~25) complementary nucleotides (the repair tract) and then switches to polymerization. The residual nick is ligated by DNA ligase (gp30). Most probably, the T4 system repairs replication and other mismatches as well; however, it might not discriminate old and new DNA strands and so does not seem to be aimed at repair of replication errors, in contrast to the most commonly studied examples of mismatch repair.