Intrachromosomal Gene Conversion in Mouse Cells

With the intent of further exploring the nature of gene conversion in mammalian cells, we systematically addressed the effects of the molecular nature of mutation on the efficiency of intrachromosomal gene conversion in cultured mouse cells. Comparison of conversion rates revealed that all mutations studied were suitable substrates for gene conversion; however, we observed that the rates at which different mutations converted to wild-type could differ by two orders of magnitude. Differences in conversion rates were correlated with the molecular nature of the mutations. In general, rates of conversion decreased with increasing size of the molecular lesions. In comparisons of conversion rates for single base pair insertions and deletions we detected a genotype-directed path for conversion, by which an insertion was converted to wild-type three to four times more efficiently than was a deletion which maps to the same site. The data are discussed in relation to current theories of gene conversion, and are consistent with the idea that gene conversion in mammalian cells can result from repair of heteroduplex DNA (hDNA) intermediates. NITIAL studies of intrachromosomal homologous I recombination in mammalian cells employed recombination substrates that generated products of reciprocal exchange as well as gene conversion (LISKAY and STACHELEK 1983; LISKAY, STACHELEK and LETSOU 1984; LIN and STERNBERG 1984; SMITH and BERG 1984; SUBRAMANI and RUBNITZ 1985). More recent investigations (LISKAY and STACHELEK 1986; LISKAY, LETSOU and STACHELEK 1987; RUBNITZ and SUBRAMANI 1986) have taken advantage of recombination substrates that can be used to select uniquely for conversion or conversion-like events. Clearly, such substrates can simplify experiments designed to study factors affecting gene conversion. In this report we present experiments that employ conversion-restricted substrates, stably integrated in cultured mouse cells, to assess the effect of the molecular nature of mutation on intrachromosomal gene conversion. A similar question has been addressed previously in meiotic fungal systems. Experiments of LEBLON (1 972a,b) and Yu-SUN, WICKRAMARATNE and WHITEHOUSE (1 977) revealed differences in the conversion spectra of single base pair insertions and deletions in the ascomycetes Ascobolus immersus and Sordaria breuicolis. It is generally believed that this disparity in gene conversion results from repair of heteroduplex DNA (hDNA) intermediates (LEBLON and ROSSIGNOL 1979). In the yeast Saccharomyces cerevisiae, in contrast, virtually all mutations convert to wild type with ' Present address: Department of Molecular Biology, Princeton University, Princeton, New Jersey 08544. Genetics 117: 759-769 (December, 1987) similar efficiencies (for review see ORR-WEAVER and SZOSTAK 1985). The yeast data are most consistent with models for gene conversion that postulate an important role for double-strand (ds) gap repair. A shared caveat in the aforementioned analyses is that the nature of most lesions investigated has never been confirmed at the molecular level. Using various well-characterized mutations in the herpes simplex virus (HSV) thymidine kinase (tk) gene, we have designed studies with cultured mouse cells to investigate systematically the effect of the molecular nature of mutation on intrachromosomal gene conversion. We have demonstrated that, in mammalian cells, all mutations are not converted to wild type with equal efficiencies. The observed differences in conversion rates were correlated with the molecular nature of the different mutations investigated. First, rates of gene conversion were inversely correlated with absolute lesion size. Second, a single base pair insertion was converted to wild type three to four times more efficiently than was a single base pair deletion. As discussed below, results obtained in this study demonstrate that gene conversion in mammalian cells is a discriminatory process, and the data support the idea that gene conversion in mammalian cells results from repair of hDNA intermediates. MATERIALS AND METHODS In vitro mutagenesis: Mutants assembled for this investigation are depicted in Figure 1, and the method of their construction is described below. The molecular nature of all the lesions included in this investigation was confirmed