©Eye of Science/Photo Researchers

Yeast is the oldest domesticated microbe. Its potable fermentation products have sparked feuds, ended wars, instigated romance, and wrecked many a morning after. The organism's mark on science is no less notable. In the 1950s, mapping 26 genes was a challenge. Fifty years later, researchers have identified all 6,000, and they have extracted from this single-celled organism clues to the workings of all eukaryotic life. Scientists know more about cell cycle control from yeast than from any other organism.

Yeast primarily eat, divide, sporulate, and mate. These highly ordered and regulated functions appear in some form across all eukaryotic life, including mammals, and they look shockingly similar. "It's beyond similarity. I mean, it's almost identity," says yeast geneticist Leland Hartwell, whose genetic approach to the cell cycle won him the 2001 Nobel Prize.

The complex molecular cascades that control cell division across evolution were discovered with elegant simplicity in yeast. One researcher, Susan Forsburg, compares it to disassembling a lawn mower to understand a Mercedes Benz. Scientists have only a glimpse of what drives yeast, and they continue to suss it out, piece by piece. Others are taking a broader scope, applying genome-scale investigations, hoping that many integrated-parts lists will inevitably enable scientists to draw a complete diagram of the cell. APYG, the "awesome power of yeast genetics," still underscores these modern investigations.

A FOUL-WEATHER FRIEND In 1942 Herschel Roman arrived at the University of Washington to study corn genetics. The dreary Seattle clime wilted his plants, but inside his lab his interest in yeast blossomed, says Rochelle E. Esposito. She joined Roman's lab in 1962 after he had secured a grant with the National Institutes of Health to further develop Saccharomyces cerevisiae, the budding yeast, as a model organism.

Esposito, now a University of Chicago professor, says Roman's work attracted people worldwide. Scientists studying the fly, phage, and bacteria frequented weekly yeast meetings, Esposito reports. "Everyone could appreciate that the organism had aspects of its life cycle that were so flexible, and so easy to work with."

About the same time that Esposito was joining Roman's lab, Hartwell, at the University of California, Irvine, was searching for a eukaryote with good genetics. He found yeast. "[It] was really the only game in town," he says.

In 1968, Hartwell moved to the University of Washington. The big break came, he says, when a grad student's project required time-lapse photography. "As soon as we started looking at the photographs, we realized how much there was there about the cell cycle," says Hartwell, who directs the Fred Hutchinson Cancer Research Center in Seattle. "We immediately dropped everything and started screening all the mutants we had by photomicroscopy, and about one in ten mutants was a cell-cycle mutant." Hartwell's temperature-sensitive mutants consistently arrested at uniform stages in the cell cycle when warmed to 37°C. The evidence suggested that the mutated genes drive cell division. Hence he dubbed the mutants cdc for cell division cycle.

Hartwell identified about 50 CDC genes in this fashion, notably CDC28, which, in mutant form, halted progression into S phase. CDC28 was recognized as powering START, the transition point at which the cell is committed to division. The gene later would be found to encode a cyclin-dependent kinase. Subsequently, homologs were pulled out across multiple species.

A FIELD DIVIDED, A CYCLE UNITED By the mid-1970s, a young UK researcher was finding inspiration in Hartwell's work. Sir Paul Nurse, who would later share Hartwell's Nobel Prize, had a background neither in yeast nor in genetics when he first approached Murdoch Mitchison, the foremost leader of cell cycle research in Schizosaccharomyces pombe, the fission yeast originally derived from East African millet beer. (Indeed, the Swahili word for beer is pombe.)

As genetically divergent from budding yeast as both are from humans, fission yeast divides not by budding, but by medial fission, a process that looks similar to higher eukaryotic division. Also, while S. cerevisiae prefers to be in a diploid state, S. pombe occurs as a haploid in the wild. Regarding cell cycle, some hypothesize that these preferences explain why S. cerevisiae spends most of its time in G1 phase prior to DNA synthesis, and S. pombe spends most of its time in G2 phase, where haploid cells have a "sister chromatid" from which to repair damage. Says Nurse, who is now president of Rockefeller University, "That made study of the G2-to-mitosis transition, which is where I did my initial work, easier."

Nurse, like Hartwell, honors science's most basic mandate: observation. "I noticed, under the microscope, these cells dividing at a small size, and really it immediately clicked [about] what that might mean," Nurse recalls. By investigating division rate and DNA content, Nurse found a checkpoint gene, which he called wee1⁺, that controls entry to the mitotic stage from G2.

Later, Nurse screened a mutant yeast with a human cDNA library and rescued the organism, pulling out a human homolog to cdc2⁺, which encodes a cyclin-dependent kinase like S. cerevisiae's CDC28 and the human CDK1 and 2. This cross-species complementation hammered home the idea that cell cycle was a global phenomenon. "The whole thing came together in the late '80s when all of a sudden people had this amazing epiphany that these were the same molecules all the way up the evolutionary scale," says Forsburg, a researcher at Salk Institute for Biological Studies who worked with Nurse at the time.

Images courtesy of W.H. Freeman and Company

THE OTHER YEAST: The fission yeast cell cycle differs from budding yeast in that it has a clear G2 phase, but no morphological marker for START.

A FUNCTIONAL FUTURE Nurse, Hartwell, and many others shook out major players in the cell cycle. Now, the rise of genome-wide studies is flooding the field. Functional genomic and proteomic approaches are being applied to this organism faster than any other model because of its facile, well-characterized genetics. Mark Johnston at Washington University, St. Louis, Mo., predicts that some function for each of S. cerevisiae's 6,000 genes could be determined in roughly six to eight years, essentially "solving" the organism. Yet, he admits the prediction's facetious nature. "Once you know something about every single protein, you're going to want to know how they interact with each other and how they interact with the environment, and that will come," he says.

Johnston, with a consortium of other yeast scientists, moved in this direction by creating a gene-deletion mutant set, comprising just under 6,000 strains of S. cerevisiae, and representing practically every gene deletion. Each open- reading frame was replaced with a traceable "molecular bar code," creating a powerful tool for large-scale genomic and drug-response studies.¹ The University of Toronto's Charlie Boone is using the set to systematically observe every possible mix of two nonlethal mutations. By screening for so-called synthetic lethals, in which a combination of relatively harmless mutations kills the organism, researchers should uncover redundant genes as well as functional interactions between gene families that might have gone untested with traditional techniques. "The strength of it is that it's no different from the genetics we've always been doing. But, it's automated. Because classical genetics is extremely powerful, this is extremely powerful," says Boone.

New knowledge of the cell cycle has emerged as well. Another Canadian group screened 4,812 viable haploid deletion strains for differences in cell size, which is a characteristic often related to START. Roughly 500 small (whi) and large (lge) mutants were revealed, and additional analyses tightened their relationships to START.² According to Paul Jorgenson, a graduate student at Samuel Lumenfeld Research Institute, University of Toronto, the field has ignored the observations made on cell size and growth that followed much of Nurse's and Hartwell's seminal work. "Their logic stands up so well over time." Jorgenson says.

Elsewhere, a massive screen of synchronously dividing cells showed that approximately 800 yeast genes oscillate in expression throughout the cell cycle. And, it is estimated that many more play a role. Richard Young and a team at the Whitehead Institute for Biomedical Research and Massachusetts Institute of Technology investigated the activity of nine proteins known to regulate transcription of cell cycle genes; they found interaction with up to 80% of that oscillating set.²

More recently, the team looked at an entire set of known regulators and found additional transcriptional regulators involved in the cell cycle. Young stresses the importance of having an interdisciplinary team. "In the same way we're standing on the shoulders of giants like Hartwell, we're also standing on the shoulders of computer scientists."

Data analysis takes the bulk of the time in such efforts, and determining what is significant re-mains a major problem, says Yale University's Michael Snyder, who has been working on similar studies and other proteomic approaches to yeast. Some complain that the fast and furious collection leaves data often less than perfect and always less than complete. And while systematic efforts give a biological problem an idea of scale and perspective, the data are "woefully incomplete," says Snyder.

Great strides and incomplete information are the hallmarks of a strong model organism. The discovery of cyclin-dependent kinases earned the 2001 Nobel Prize, but "we have almost no clue as to what they do," says Kim Nasmyth, at the Institute for Molecular Pathology in Vienna. Mike Tyers, who is Jorgensen's principal investigator, says that he is not ready to court other models. "I'll be happy to keep working on yeast for the rest of my career. I don't think it's going to dry up at all."

Brendan A. Maher can be contacted at bmaher@the-scientist.com.