Toward a Comprehensive Model of Yeast Stress Response Pathway

Cells rely on cyto-protection programs to survive exposure to external stressors. In eukaryotes, many of these programs are regulated by Mitogen-Activated Protein Kinases (MAPK). Aberrant signaling in MAPK pathways is associated with pathologies including cancer and neurodegenerative disease. Thus understanding how MAPKs are dynamically regulated is critical for human health. We leverage the simplicity and genetic tractability of S. cerevisiae (yeast) to experimentally investigate and mathematically model MAPK activity in the prototypical High Osmolarity Glycerol (HOG) pathway. The HOG pathway transmits osmostress to Hog1, a MAPK which is homologous to the p38 and JNK kinases, via two distinct signaling branches (Sho1, Sln1). Once phosphorylated, Hog1 induces activation of the hyper-osmotic stress adaptation program. Our goal was to identify and characterize the feedback network that regulates Hog1 activity during hyper-osmotic stress. We previously demonstrated that Hog1 phosphorylation is encoded via positive feedback, and in agreement with previous studies, we showed that Hog1 dephosphorylation is encoded via negative feedback. We used mathematical modeling to define simplified MAPK feedback networks that featured various combinations of negative and positive feedback loops. We then used a modified Approximate Bayesian Computation (ABC) parameter estimation and model selection algorithm to rank each feedback network based on its ability to reproduce the Hog1 phosphorylation training data. We then used the top models to design a set of experiments that could be used to further differentiate the likelihood of each model to be selected as the consensus model. Our simulations suggested that the likelihood of each model could be further differentiated if yeast were exposed to dynamic hyper-osmotic stress conditions. Our simulations also suggested that the HOG pathway would be most sensitive to hyper-osmotic stress if the positive feedback operates at the level of the MAPK. To test our model predictions, we used live-cell fluorescence microscopy and a microfluidic device to measure nuclear accumulation of a Hog1-GFP fusion protein as a function of the dynamic hyper-osmotic stress conditions suggested by our simulations. We observed that nuclear accumulation of Hog1-GFP was most consistent with the model that suggested positive feedback operates at the level of the MAPK. These findings further suggest that positive feedback directly amplifies Hog1 phosphorylation. Furthermore our study exemplifies the power of integrating mathematical models into an existing quantitative experimental framework.Doctor of Philosoph