Principal Investigator Gerald Fink
Most of the genes in the genome with internal tandem repeat sequences encode cell wall proteins and the length of the tandem repeats are highly variable from strain to strain. To identify the S. cerevisiae open reading frames (ORFs) that contain intragenic tandem repeats, we scanned all 6591 open reading frames of Saccharomyces for the presence of long (>40 nt) tandem repeats. The search yielded 29 ORFs with repeats longer than 40 nt. 22 of the 29 ORFs (75 %) with conserved long repeats encode cell surface proteins, whereas only 1.3 % of all ~6000 S. cerevisiae ORFs are cell surface proteins. A comparison of lab and wild strains for the size of the repeat sequences in each of the 22 cell surface genes using PCR primers flanking the repeats showed that all 22cell surface genes with internal tandem repeats showed size variation from one strain to another.
Gene expansion and contraction of intragenic repeats is a frequent event in real time. The FLO1 gene is 4.6 kb long and contains 18 repeats of about 100 nt, separated by a less conserved 45-nt sequence. A colony grown from a single cell generates progeny that have increased or decreased numbers of repeats in FLO1 at a frequency of 10-4. The new FLO1 genes generated during mitotic growth were as large as 8 kb and as small as 2.5 kb, the sizes varying by an integral number of repeat units. The change in gene size has profound phenotypic consequences: There is a striking, linear correlation between gene size and the extent of adhesion: as the FLO1 proteins become longer (carrying more repeats), theadhesion properties gradually become stronger. Flocculation (i.e. adhesion to other yeast cells) shows the same relationship to the repeat number: the more repeats, the greater the fraction of flocculating cells.
Taken together, the recombination frequencies observed in the various mutants indicate that recombination between the cell wall gene intra-genic repeats is caused by a replication slippage process similar to that observed in inter-genic repeats. Loss of the RAD27-encoded flap endonuclease, which causes the formation of double-stranded breaks during replication, increases the instability of FLO1 repeats almost 40-fold. Furthermore, the decrease in recombination observed in rad50É¢, rad52É¢ and rad1É¢ rad52É¢ mutants suggests that the process depends on break repair by single-strand annealing, a conclusion further supported by the decrease in FLO1 recombination in the rad59É¢ mutant, which is known to be deficient in this type of DNA repair.