Experiments and Analysis of Mitochondrial DNA Segregation, Yeast Polarization and RNA Polymerase Dynamics

Abstract

This thesis presents work done in three different topics in biology with a combination of experiments, mathematical modeling and quantitative data analysis.

All cellular materials are partitioned between daughters at cell division, but by different mechanisms and with different accuracy. Many macromolecules partition in proportion to cytoplasmic volume with significant errors at low numbers, while chromosomes accurately push half the copies to each cell regardless of volume. Little is known about organelles, but in Schizosaccharomyces pombe the mitochondria are pushed to the cell poles by the spindle, along with the chromosomes. Here we find that mitochondria spatially re-equilibrate just before division, and that the mitochondrial volume and DNA-containing nucleoids instead segregate in proportion to the cytoplasm inherited by each daughter. However, in contrast to other macromolecules, nucleoid partitioning errors are still strongly suppressed. This is ensured by control at two levels: mitochondrial volume is actively distributed throughout a cell in a process involving the microtubule associated protein Mmb1p, and nucleoids are spaced out in semi-regular arrays within mitochondria. The data also suggest that nucleoid replication control is passive and Poisson, and that low concentration noise is achieved by accurate segregation rather than corrective feedback control.

Patterning in organisms is both beautiful and functional, but we have limited understanding of the general principles that cause patterns to emerge from isotropic initial conditions. Alan Turing proposed a minimal set of conditions that would allow reaction-diffusion systems to spontaneously form stable patterns. Since then many biological systems have been hypothesized to be of the Turing type, but with limited evidence that they actually rely on this mechanism. One of the conditions of Turing's theory is that reactions must have non-linear sensitivities (cooperativity) in their positive feedback cycles. We use a minimal model of yeast bud site selection to show that the amount of cooperativity achievable is naturally limited by the kinetic parameters achievable by proteins and the type of reaction mechanisms commonly used to achieve cooperativity. Further, higher levels of cooperativity necessarily cause pattern formation to slow down, likely causing a fitness defect for the organism. Then, because cooperativity is also used in many other biology systems, we show examples in TF-promoter binding and cooperative ligand binding where the trade-off between cooperativity speed of response is also present. These considerations bring into question how biological systems respond to these trade-offs. We then explore a conceptual model for yeast budding where an additional actin-based positive feedback loop uses a time-delay, not cooperativity, to ensure that yeast picks a single bud site.

RNA polymerase II (RNAP) pauses and backtracks during transcription elongation to regulate gene expression, control transcriptional efficiency and ensure transcriptional accuracy. After a pause, RNAP often backtracks and, in the presence of TFIIS, cleaves off the 3’ end of nascent RNA, allowing productive transcription to restart. However, the precise determinants of RNAP pausing, backtracking, and restarting remain unclear. Native elongating transcript sequencing (NET-seq) visualizes RNA polymerase pausing and backtracking with nucleotide resolution. Applying NET-seq to wild-type S. cerevisiae gives locations where cleavage occurs after backtracking. In contrast, NET-seq data from a strain where TFIIS is deleted reveal the primary points of pausing. Using these data, we show that RNAP pausing is largely controlled by the relative strength of the RNA:DNA and DNA:DNA hybrids in and downstream of the transcription bubble. In addition, backtracking is likely determined by the stacking interactions between the 3’-end of nascent RNA and Tyr769 of the Rpb2 subunit of RNAP. Though other factors beyond structural energetics of the transcription elongation complex certainly play a role in RNAP backtracking, this work suggests that the ubiquity of RNAP backtracking is largely controlled by sequence elements.