1. ¹The Netherlands Cancer Institute, Division of Molecular Biology, Amsterdam, The Netherlands
2. ²Department of Clinical Genetics, Free University Medical Centre, Amsterdam, The Netherlands

Correspondence: Professor H te Riele, The Netherlands Cancer Institute, Division of Molecular Biology, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands. E-mail: h.t.riele@nki.nl

Received 10 June 2005; Revised 27 September 2005; Accepted 17 October 2005; Published online 28 January 2006.

Abstract

We have previously demonstrated that site-specific insertion, deletion or substitution of one or two nucleotides in mouse embryonic stem cells (ES cells) by single-stranded deoxyribo-oligonucleotides is several hundred-fold suppressed by DNA mismatch repair (MMR) activity. Here, we have investigated whether compound mismatches and larger insertions escape detection by the MMR machinery and can be effectively introduced in MMR-proficient cells. We identified several compound mismatches that escaped detection by the MMR machinery to some extent, but could not define general rules predicting the efficacy of complex base-pair substitutions. In contrast, we found that four-nucleotide insertions were largely subject to suppression by the MSH2/MSH3 branch of MMR and could be effectively introduced in Msh3-deficient cells. As these cells have no overt mutator phenotype and Msh3-deficient mice do not develop cancer, Msh3-deficient ES cells can be used for oligonucleotide-mediated gene disruption. As an example, we present disruption of the Fanconi anemia gene Fancf.