Mapping and modeling the impact of protein biochemical variation on growth rate phenotype
Individual proteins can be expressed, purified, and exquisitely characterized in terms of their biochemical and biophysical parameters in vitro. However, the quantitative relationship between these parameters and complex phenotypes like growth remains mysterious. For example, what values of protein abundance, thermal stability (ΔGfold) and catalytic activity (kcat, Km) must an enzyme attain to sustain metabolic pathway flux and support cell growth? In many cases, we are missing even orders-of-magnitude level bounds on these fundamental biochemical parameters — we do not have a sense of which protein properties must be precisely tuned and which are robust to variation. To address this knowledge gap, my lab seeks to quantify the intracellular constraints on protein abundance, activity, regulation, and ultimately sequence. We then use this information to engineer new protein systems and build mathematical models relating protein activity and sequence to phenotype. In this talk, I will first discuss our recent study of how variation in the activity of one enzyme constrains the biochemical parameters and sequence of another. Using a combination of deep mutational scanning and mathematical modeling we showed that inter-enzyme biochemical coupling can strongly reshape an enzyme’s sensitivity to mutation. Then, I will introduce a CRISPR-interference based strategy for quantitatively mapping the relationship between protein expression level and cell growth. We used these high throughput measurements to train an interpretable machine learning model that predicts growth rate given combinatorial variation in gene expression and environment. Together these data lay a foundation for defining the biochemical “design specifications” of metabolic pathways and cell systems.
Please use this link to access Zoom.
For any questions, please contact Hyacinth Camillieri at hcamillieri@gc.cuny.edu.