YEAST AND CANCERNobel Lecture, December 9, 2001byLELAND H. HARTWELLFred Hutchinson Cancer Research Center, 100 Fairview Avenue North, D1-060 Seattle, WA 98109-2024, USA.My research career has been motivated by a desire to understand cancer.Each time I have identified an intriguing aspect of the cancer problem, I have found that it could be approached more effectively in the simpler euka-ryotic cell, Saccharomyces cerevisiae, than the human cell. Each time the yeastcell has revealed some of its secrets. I will relate four vignettes involving stu-dies on the genetics of cell division, the control of genome fidelity, therapeu-tics for cancer and the role of natural genetic variation in disease susceptibi-lity. In the first two instances, yeast has told us something that is relevant tomankind. For the other two, it is too soon to tell.When I was finishing my graduate studies and thinking about what area ofscience to pick for my postdoctoral work, I wanted to study a field that was stilla mystery and would offer ample opportunity for a career. I chose to study thecontrol of growth and proliferation in relation to cancer cells with RenatoDulbecco and Marguerite Vogt. Even in plastic petri dishes, cancer cells grewwhen normal cells did not and I became thoroughly engrossed with the prob-lem. However, while in Renato’s laboratory I had not been able to establish anexperimental system that I thought could lead to fundamental insights. WhenI set up my own laboratory at the University of California at Irvine I was notsure what to work on.I had applied for and received a grant to work on the control of DNA syn-thesis in mammalian cells. However, after much pondering and an influentialconversation with Dan Wulff, I decided to work on a model system that per-mitted genetic analysis. I had been imprinted as an undergraduate with theinsights that Bob Edgar and Bill Wood had attained on the pathway of viralmorphogenesis using genetics. The processes of the cell cycle, chromosomereplication and segregation, seemed like morphogenetic problems of the same nature. Yeast was the obvious choice because it grew as single cells, animportant property for studies of cell division. Moreover, Don Williamsonhad shown that S. cerevisiae had a eukaryotic cell cycle with G1, S, G2 and Mperiods (1) and C.F. Robinow had demonstrated the presence of an intra-nuclear spindle (2). These were important facts since yeast has, at various times, been accused of not being a proper eukaryote. There was even a timewhen people thought that yeast lacked DNA.246GENES THAT CONTROL CELL DIVISIONSince cell division was an essential process, I set out isolating temperature-sen-sitive mutants that could grow at room temperature but not at 36C. We iso-lated about a thousand mutants and characterized each for protein, RNA,and DNA synthesis, cell division, and cell morphology following a shift fromthe permissive to the restrictive temperature. Although most of the mutantshad unremarkable phenotypes, a significant number behaved as if defectivein a specific one of these processes (3). I thought it important to demonstrate that we could identify the defectiveprotein in some mutants. Calvin McLaughlin (Fig. 1) who was an expert inprotein synthesis generously agreed to collaborate and we began studying theprotein synthesis mutants. We did identify mutations in genes that coded fortwo aminoacyl-tRNA synthetases and another mutant defective in the initia-tion of protein synthesis (4).CDC genes. Three years into the project, I moved to the Genetics depart-ment at the University of Washington. I have had a number of very bright un-dergraduates in my lab over the years and one, Brian Reid (Fig. 2), set us onthe path to the cell cycle. Although we had a few mutants that behaved likeDNA synthesis or cell division mutants from the original survey, we didn’t have any clever ideas about how to analyze them. Brian, through a very amus-ing bit of serendipity that I have described elsewhere (5), discovered how use-ful photomicroscopy was for analyzing the S. cerevisiae cell cycle. This wasbecause the cell forms a small bud on its surface at the beginning of a cell cycle and the bud grows in size as the cycle progresses. Thus one can easilytell where a cell is in the cycle merely from its morphology. Mutants with spe-cific defects in the cell cycle were recognized because the asynchronous po-pulation of cells grown at the permissive temperature became synchronouslyarrested in the cell cycle at the restrictive temperature (Fig.3).247Figure 1. Calvin McLaughlin.Figure 2. Brian Reid.When we realized this, we were able to find hundreds of cell cycle mutantsin a few months. C.F. Robinow agreed to visit the lab and teach us how tostain the nuclei of cells, and Joe Culotti (Fig. 4), a new graduate student, wasable to show that the mutants that arrested with a uniform cell morphologyalso arrested with a uniform nuclear morphology. At this point we were ex-cited that we had some very interesting mutants (Fig. 5) and we were confi-dent that they would reveal new insights into the cell cycle (6).Pathways. We ordered the mu-tants in a number of ways. First, byhow far they traversed the cycle be-fore arresting. We analyzed the fol-lowing cell cycle events: budding,DNA synthesis, nuclear division cy-tokinesis and cell division. Theevent that stopped first after a shiftfrom the permissive to the restricti-ve temperature was considered theprimary defect. After the primarydefect, other cell cycle events wouldoccur or not depending upon theparticular mutant. Eventually, thecell would arrest development andgenerate a terminal phenotype, de-pending on which events occurredand which did not. Mutants werefound with primary defects in eachof the cell cycle events that we mo-nitored. The phenotypes of the mu-tants suggested a relatively simplepathway of dependent events lead-ing to cell division. The first event,248Figure 3. Time-lapse photomicroscopy of a cdc mutant cells growing at the permissive temperature(A) and several hours after a shift to the restrictive temperature (B).Figure 4. Joe Culotti.controlled by the CDC28 gene, was required for initiating two pathways, oneof which led to budding, nuclear migration, cytokinesis and cell division (Fig.6). The second led to DNA replication, nuclear division and joined the firstprior to cytokinesis and cell division. We tested our conclusions by construc-ting all possible double mutants and, indeed, the phenotypes of