Gene targeting in mammalian cells plays a crucial role in biotechnology. These experiments are characterised by low rates of homologous recombination and high rates of random integration. Therefore, many fibroblast colonies must be screened to identify a targeting event. To dramatically reduce the survival of random integration events, we have developed a splicing-dependent selectable marker strategy by introducing a mutation in a codon-optimized G418 resistance gene (mNeo). This mutation could be corrected upon homologous recombination. Since the C-terminal region of aminoglycoside phosphotransferase (AphII, Neo/Kan resistance) participates in formation of the active site of this enzyme, we hypothesised that addition of even one amino acid at the C-terminus would render this protein non-functional. To test this hypothesis, a mutation was introduced in an E. coli AphII expression vector that converted the stop codon of AphII to tryptophan (X265W, TGA > TGGTAA). This mutation was confirmed to inactivate AphII by independently characterising the G418 and Kanamycin resistance (or lack thereof) provided by the X265W mutation. To evaluate this mutation in mammalian cells, two intronless mammalian expression vectors were constructed that differed by the presence or absence of the X265W mutation. G418 resistance was only provided by the wildtype sequence, thus confirming that X265W inactivates AphII in mammalian cells. An identical mutation was then introduced into a eukaryotic expression vector based on mNEO. Further, the sequence was extended to create a 5′ intron splice site (TGA > TGGTAAGAGTT). This region was designed to direct splicing between the first and second G residues thus removing the G in the third position of the W codon. The 3′ intron splice sites was then designed to provide an A residue as the first base of the next exon so that successful splicing would correct the mutation by recreating an appropriately positioned stop codon (TGA). To evaluate this strategy in mammalian cells, two plasmids were constructed that harbored the X265W mutation embedded at the 5′ splice site of a downstream intron. In one plasmid (pSC3-G) the first base of the downstream exon begins with a G residue resulting in inactivation of AphII. In the other plasmid (pSC2-A), the first base of the downstream exon begisn with an A residue forming a stop codon that allows for active, wildtype AphII. These plasmids were transfected into porcine fetal fibroblasts and subjected to selection with G418. A positive control plasmid and pSC2-A produced colonies that were too numerous to count. A negative control plasmid and pSC3-G produced no colonies. It can be concluded that the X265W mutation can be corrected by splicing to an exon that begins with an A residue. This splicing-dependent selectable marker may prove useful in gene targeting experiments when the site of modification is followed by an exon that begins with an A.
Target frequency and integration pattern for insertion and replacement vectors in embryonic stem cells
