Chromosome breakage and repair.

FROM time to time one of my colleagues working at a medical school commiserates with me because I spend 40 hr a year lecturing to undergraduates. I always reply that teaching has compelled me to learn a lot of material that I would not have known about had I taught only my specialized subject. These forays into the ‘‘beyond’’ were instrumental in moving my research in new directions. Three articles that I published in Genetics in the early 1980s (McCusker and Haber 1981; Haber and Thorburn 1984; Haber et al. 1984) were the result of learning about, in order to teach, classic genetic experiments in Drosophila and maize. When I arrived at Brandeis I was assigned to teach genetics, a subject I had never studied as either an undergraduate or a graduate student. I was fortunate to team up in teaching with Jeff Hall, a master geneticist, who taught me much of the lore of Drosophila and maize, to add to the yeast genetics that my lab and I were slowly learning. I was particularly interested in Barbara McClintock’s study of the ‘‘Activator (Ac)/Dissociator (Ds)’’ transposable elements whose excisions led to cycles of breakage–fusion–bridge (BFB) of broken chromosome ends that generated chromosomal truncations. Her studies resonated strongly with the behavior of apparently broken chromosomes that we were studying in budding yeast as a consequence of matingtype gene switching. The study of yeast mating-type (MAT) gene switching has provided gainful employment for many scientists interested in cell-type regulation, gene silencing, chromosome architecture, and DNA repair. Haploid cells express either the MATa or the MATa allele and mate with cells of the opposite type, but cells expressing both MATa and MATa are nonmating. But most unusual was that homothallic MATa cells could switch to MATa, or vice versa, as often as every cell division. Takano and Oshima (1970; Oshima and Takano 1971) showed that switching depended on two distant loci that they viewed as ‘‘controlling elements,’’ similar to those defined by McClintock in maize. Hicks et al. (1977) made the insightful suggestion that these two loci were in fact unexpressed copies of mating-type information (now called HMLa and HMRa) that could be transposed to replace the original MAT allele. Extending the mutational analysis of Mackay and Manney (1974) and the sometimes-published work of the inventive Don Hawthorne (Hawthorne 1963; see also Herskowitz 1988), Strathern et al. (1981) proposed that MATa encoded both a repressor of a-specific genes (MATa2) and a positive regulator of a-specific genes (MATa1). Thus a mata1 mata2 mutant proved to be a-like. The transposition/replacement of MATalleles occurs very frequently in homothallic cells, expressing the HO gene encoding a site-specific endonuclease, but is very rare in heterothallic strains where HO is inactive. One way we tried to study these rare events was by mating two heterothallic MATa strains together, on the assumption that if one of them switched toMATa, it would readily conjugate with a MATa cell, as first shown by Hawthorne (1963). We took up this approach (McCusker and Address for correspondence: 415 South St., Rosenstiel Center, Department of Biology, Brandeis University, Waltham, MA 02454-9110. E-mail: [email protected] I did take the first 3-week yeast genetics course at Cold Spring Harbor in 1970, memorably taught by Fred Sherman and Gerry Fink. I still remember being mystified by a lecture on gene conversion; but learning how to dissect tetrads and to score and maintain strains greatly lowered the energy barrier to undertaking genetic studies in my own lab. The Cold Spring Harbor Laboratory yeast course has been previously celebrated by Sherwood (2001). That the lessons I was teaching in class had not penetrated very deeply into my own laboratory practice is seen from the fact that we isolated one allele of the first identified chromosome loss gene, CHL1 (Haber 1974; Liras et al. 1978), and did not think to get many more and do a complementation analysis. Later, Kouprina et al. (1993), using a similar chromosome loss assay, identified 18 such genes.