Recessive screens in mice require breeding of heterozygous mutations, which is both time-consuming and expensive. Knowing that screens in mammalian cell lines would be cheaper and quicker, both sets of authors have exploited properties of mouse stem cells that are homozygous for mutations in
Blm
(Bloom syndrome homologue). With each generation, these cells are more likely to make the mutations that occur on one chromatid homozygous, presumably through increased mitotic recombination. Guo and colleagues mutagenized the Blm−/− ES cells with a retroviral gene-trap vector to create a library of homozygous mutant cell lines. They first looked for genes that are involved in the mismatch repair (MMR) pathway, which is often mutated in cancer. The use of 6-thioguanine allowed for the selection of cells that are defective in MMR, and the gene-trap method allowed for quick mutant gene identification. Indeed, one of the famous Msh genes was picked up, as was
Dnmt1
— a surprise because it is normally thought of as a methylase. But re-examination of Dnmt1 knockout cells confirmed that they had a higher level of microsatellite instability, qualifying Dnmt1 as a new MMR pathway member.
Yusa et al. used a tetracycline-induced Blm ES-cell knockout line. They mutagenized these cells with N-ethyl-N-nitrosourea (ENU) and used doxycycline selection to make the mutations biallelic. As a proof of principle, they screened for cell lines that were deficient in one pathway — glycosylphosphatidylinositol (GPI)-anchor biosynthesis — which is easier than it might sound, because exposing cells to aerolysin kills those with GPI anchors and leaves the mutants behind. Unlike in the Guo et al. library, here expression cloning is needed to find the genes; Yusa et al. achieved this through complementation analysis of the known genes in the pathway. A total of 12 of the 23 known genes in this pathway were mutated in both alleles after just one round of screening.
This is a preview of subscription content, access via your institution
