Conventional gene targeting protocols lead to loss of targeted cells when applied to a silent gene locus in primary fibroblasts

Gene targeting in livestock fibroblasts has proven difficult to achieve, particularly if the target gene is silent. We first tested whether efficient gene targeting at the transcriptionally active ovine α1(I) procollagen (COL1A1) locus required the use of a promoter trap vector. We compared gene targeting frequencies at the ovine COL1A1 locus using both a promoter trap and a non-promoter trap selection strategy. We demonstrated that targeted cells could be isolated regardless of whether an enrichment step (promoter trap) was used. Next, we used our optimised protocol to target a non-expressed gene, ovine β-casein. We obtained clones that were scored positive by PCR for the targeting event, but were negative after cell expansion and Southern analysis. We propose that targeted cells were initially generated but that they were at a selective growth disadvantage during culture. We suggest modifications to the conventional targeting protocol that would prevent such loss of targeted cells.