Genetic and Evolutionary Dissection of the Role of Cell Cycle-Regulated Gene Expression

Abstract

Phase-dependent transcription of cell cycle genes is a ubiquitous feature across eukaryotes. However, this feature is not universal — some organisms, including some animal embryos during their first divisions and kinetoplastid parasites, undergo cell division without cell cycle-dependent transcription. These natural exceptions raise a fundamental question: is cell cycle-dependent transcription essential for the proper execution of the cell cycle, or rather a fail-safe mechanism that some cells use and others do not? A clean test of this question has been difficult. My thesis uses the classic model organism for cell cycle studies, the budding yeast Saccharomyces cerevisiae, to answer this question by targeting the source of the cell cycle transcriptional program: five master transcription factors Swi5, Swi6, Hcm1, Fkh1, and Fkh2. Through simultaneous deletion (5∆) and acute degron-mediated depletion (5-AID) of all five of these transcription factors, followed by experimental evolution for ~1,250 generations, I characterize both the immediate consequences of losing the transcriptional program and how yeast adapt to its absence. Deleting or acutely depleting the five master cell cycle transcription factors abolishes cell cycle-regulated transcription and slows, but does not prevent, proliferation. Doubling times increase from ~90 min in wild type to ~150 min in 5-AID and ~250 min in 5∆. 5∆ cells display multicellular clustering, multinucleate cells, and anaphase bridges, indicating that the transcriptional program is required for normal cell biology even when it is not required for division per se. After ~1,250 generations of experimental evolution, the 5∆ populations recovered most of the fitness defect, yet transcriptional oscillations were not restored. Whole-genome sequencing, filtered against ancestor sequencing, identified recurrent loss-of-function mutations in two chromatin/transcription complexes — the Isw2 chromatin remodeling complex, including a definitive premature stop codon in one of its components, Itc1, and the Cdk8 kinase submodule of Mediator — as well as C-terminal truncations of another cell cycle transcription factor Swi4. Reconstructing the Swi4 truncation alone in the ancestral 5∆ recovered ~80% of wild-type fitness, restored single-cell morphology, and re-established expression of G1/S target genes. Together, these results show that cell cycle-dependent transcription in budding yeast is dispensable for cell division but required for the faithful execution of cell cycle events that distinguish wild-type from perturbed cells. Yeast rapidly evolve substitute regulatory architectures, through loss of chromatin remodeling and derepression of an associated transcription factor, rather than by rebuilding the transcriptional wave. This suggests that cell cycle-dependent transcription is one of several possible regulatory solutions maintained by selection for normal cell biology, rather than being the only regulatory architecture that can support cell division.