Lagging strand-biased initiation of red recombination by linear double-stranded DNAs

The recombination functions of the lambda bacteriophage, collectively called Red, can be exploited for simple and specific in vivo genetic manipulations of Escherichia coli, which is of great use in functional genomics, BAC (bacterial artificial chromosome) engineering, and gene therapy. Linear double-stranded DNAs (dsDNAs), such as PCR products flanked by short homologous sequences, can be used as recombination substrates with high efficiency. However, the molecular details of Red recombination must be more clearly understood to enable more effective substrate designs for complex applications. Here, we performed two recombination assays to show that Red recombination between linear dsDNAs and circular bacterial chromosomes takes place at the replication fork by single-strand annealing. By attaching nonhomologous segments to either end of an antibiotic resistance cassette, we measured the effects on recombination efficiency and found that nonhomology at one end was more inhibitory than that at the other end. We then designed linear dsDNAs that harbored three homologous regions and two antibiotic resistance genes to examine which of the two ends was more prone to initiating recombination. We found that the 3' single-stranded end complementary to the lagging-strand template of the bacterial chromosome preferentially annealed to the targets. Our results suggest that Red recombination occurs by an annealing and replication-dependent mechanism that involves the sequential exposure of homologous chromosomal regions as the replication fork advances, with the lagging-strand homologous region unwound first.