Photo by Karen Moore, courtesy of Nancy Jenkins.

Fly geneticists have been performing a nifty trick for some time now in which they generate patches of cells that are homozygous for a mutation or a marker gene in heterozygous flies. These so-called genetic mosaics — organisms of more than one genotype — have been instrumental in cell-lineage and cell-fate determination studies of fly development, and have been created by using the site-specific recombination system, FLP/FRT, to induce recombination during mitosis (see link to animation for more). This approach has also been tested in mice, without much success, but now Pentao Liu and colleagues report a Cre/loxP-based strategy that induces mitotic recombination in mouse embryonic stem (ES) cells at reasonable frequencies and in genomic regions that are less amenable to this type of event.

Liu et al. began by generating two recombination cassettes that each contain complementary, but non-functioning, halves of a human HPRT minigene that are flanked by loxP sites, which they targeted to various allelic chromosomal regions in an Hprt -deficient ES cell line. When Cre recombinase is expressed in doubly targeted cells, the 'floxed' cassettes on non-sister chromatids can recombine to create a functioning HPRT gene, allowing cells with recombinant chromatids to become HAT (hypoxanthine, aminopterin and thymidine) resistant (HATr). When the authors targeted these casssetes to a chromosome 7 locus ( D7Mit178 ), they achieved tenfold higher recombination frequencies than those reported in previous studies. And to check that the recombinant chromatids segregate away from each other in mitosis — so-called X segregation, which produces homozygous mutant cells — the authors assayed the methylation status of a nearby imprinted gene, Snrpn , and found that all HATr cells were uniparental distal to the recombination event at D7Mit178.

But such recombination frequencies were not achieved in other genomic regions. When the cassettes were targeted to Wnt3 and to D11Mit71 on chromosome 11, recombination frequencies dropped considerably, but increased when cre was constitutively expressed for up to eight days. Liu et al. also increased recombination frequencies by replacing each single loxP site in the chromosome-7 recombination cassettes with three lox variants. Although this modification improved mitotic recombination frequencies, it also reduced the apparent incidence of X segregation, probably because the extra lox sites mediate a second exchange between sister chromatids.

Although this approach has yet to be tested in vivo — for example, by introducing recombined mutant clones into wild-type blastocysts or by generating them in vivo by the spatially and temporally controlled expression of cre — it will no doubt help mammalian geneticists to piece together new pictures of gene function.