PhD Dissertation Defense: Eliel Akinbami

Title: Reading Enzyme Function from Sequence and Structure

Abstract: Enzymes are among the most proficient catalysts known – the bacterial phosphatase PafA, for example, accelerates its reaction by roughly 1027-fold relative to the uncatalyzed reaction. Yet how a linear amino-acid sequence folds into a structure that achieves this, and how mutations to individual residues tune catalytic activity, is still only partly understood. High-throughput platforms such as HT-MEK now measure the catalytic consequences of thousands of mutations in parallel, but these functional data reveal what mutant changes activity, rarely how or why. Here I use PafA as a model system to close part of that gap through two complementary studies. 

Initial PafA experiments revealed that some mutations far from the active site (i.e D473V) had surprisingly large impacts on substrate specificity. Here, I explain this finding by acquiring high-resolution (1.6 Å) cryo-EM structures that show PafA forms a tetrameric complex in which this distal residue is closer to the active site of the neighboring protomer. Combining structural and functional data further reveals D473V alters substrate specificity by changing the conformation of a substrate-contacting residue R164.

Second, I confront a neglected confounder present in many in vitro high-throughput measurements: misfolded proteins can dramatically distort measured kinetic constants, but most high-throughput assays do not test for or quantify misfolding. To address this, I developed deFOLD (Degradation Enabled Fraction folded of On-Bead Library Displays), a bead-based proteolysis assay that can quantify the misfolded fraction of thousands of variants at once, enabling corrected kinetic parameters and a direct, scalable readout of folding that most high-throughput methods lack.

Together, these two parts of my PhD connect high-throughput functional and structural data: deFOLD explains and corrects the kinetic distortions caused by misfolding, while the cryo-EM work reveals how a mutation propagates its effects through structure, together yielding mechanistic insight into why single residues can change function. In future work, we will push deFOLD to true high-throughput, profiling misfolding across 100,000s of PafA orthologs and mutants to pinpoint the structural regions most prone to it. Ultimately, this work aims to help build a predictive framework for protein engineering, one that can distinguish sequences that fail because they misfold from those that fold but function differently.

Please contact Sofia Rakicevic-More for the Zoom link.