Kinetics of homologous recombination in Xenopus oocytes: effects of homology length and substrate concentration.

We use oocyte nuclei from the South African frog, Xenopus laevis, as an in vivo system for the study of extrachromosomal homologous recombination. Previous experiments from our lab have best supported an annealing mechanism which we refer to as resection-annealing to emphasize the role of the 5'? 3' exonuclease activity. In the work presented here, predictions of the resection-annealing mechanism were tested by comparing the kinetics and overall recombination efficiency of several related substrates with increasing homology lengths. After incubation in oocyte nuclei, DNAs with longer homologous repeats were resected more extensively before they appeared as joint molecules and had longer single-stranded tails on the recombination intermediates. Substrates with longer homologies were slower to form recombination intermediates and slower to form complete recombination products. These findings provide further support for the resection-annealing mechanism and indicate that the 5'? 3' exonuclease activity is rate-limiting. When coinjected, the expected length dependence of recombination kinetics was not apparent. Kinetic differences were minimal and only visible in early time points. We propose that the length dependence is obscured by the cumulative effects of end competition--the longer homology act as a sponge for the binding of the repartitioning exonuclease released when the shorter homology is resected. When 10[7] substrate ends were injected, the formation of recombination products can still be accounted for by the resection-annealing mechanism. However, this 250-fold decrease in substrate concentration resulted in a dramatic decrease in reaction efficiency. The simplest explanation for this shift in reaction balance is that competing activities, or other rate-limiting steps, are saturated when the substrate is in vast excess (10[9] ends), and that these activities or factors are uncovered when fewer substrates (10[7] ends) are injected. There is evidence to suggest that rate of bimolecular annealing and the removal rate of unpairable nucleotides on the 3' ends may contribute to the overall reaction rate and affect the reaction balance. We propose that the extension of repair synthesis primed by the completely annealed 3' ends is responsible for the removal of the 5'? 3' exonuclease activity from the annealed intermediate.