A simple and efficient method for direct gene deletion in Saccharomyces cerevisiae.

The systematic sequencing of the yeast S.cerevisiae genome has revealed a profusion of Open Reading Frames (ORFs). Although some of them have been previously studied, a large majority represents new genes (1, 2). Deletion of these ORFs is a convenient tool for their functional analysis, Here we describe a new approach for generating null alleles of a gene. Usually, a gene inactivation requires the in vitro creation of a construction in which a selectable marker is sandwiched by the 5' and the 3' flanking sequences of the target ORF. Classically, this strategy requires several cloning steps. In contrast, our approach generates such a construction by one step PCR amplification. Each oligodeoxynucleotide used contains two distinct regions, one which allows homologous recombination at the target locus and will be named the deleting sequence, the second part which permits the PCR amplification of a selectable marker. The deleting sequences, which are respectively the 5' (oligopro) and 3' (oligoterm) flanking sequences of the ORF, range from 35 to 51 nucleotides in length and are followed by a short stretch of 17 nucleotides homologous to the HI S3 selectable marker. Table 1 shows the composition of the deleting sequences, the sequence used for HIS3 amplification always being the same (5'-TCGTTCAGAATGACACG-3' for oligoterm and 5'-CTCTTGGCCTCCTCTAG-3' for oligopro. Following the PCR amplification, the crude mix was directly used to transform yeast by standard procedures (3). In a first set of experiments, we used W3O3-1B (MATa, ura3-l, trpl-1, ade2-l, leu2-3, 112, his3-ll, 15) as recipient strain. All the His transformants tested (a total of 30) presented the same pattern when analysed by Southern blot, corresponding to the insertion of the wild-type HIS3 gene at its own locus (data not shown). We thus used a recipient strain carrying a complete deletion of the HIS3 gene. With the diploid strain BMA 1 (a diploid from cross FY 1679-18B and FY 1679-28C, kindly provided by B.Dujon) containing the His3A200 allele (4), we routinely obtained more than 10 His transformants per plate. As the procedure appeared to be efficient enough, we tested by