A Microfluidics-Based Assay for Mapping Connectivity in Highly Proficient Enzymes Reveals Functional Modularity

Abstract

Enzymes are some of the most proficient and most specific catalysts known; however, a complete understanding of how these enormous rate enhancements are achieved is lacking. Enzymes are large and highly cooperative, containing functional connections throughout their entire structures. However, due to the cost and labor required to prepare and assay the number of mutants required to characterize this connectivity, it has been difficult to obtain the large sets of data necessary for improving rational enzyme design; as a result, the proficiencies of designed enzymes fall well short of their natural counterparts. To address these experimental limitations, we have developed a microfluidic assay, HT-MEK (High-Throughput Microfluidic Enzyme Kinetics), in which we express, purify, and quantitatively characterize >1000 enzyme mutants in a single experiment across multiple kinetic parameters. This platform greatly reduces both time and cost compared to traditional assays, while providing data of comparable precision. Using this technology, we performed systematic mutagenesis over the entire scaffold of the highly proficient phosphatase PafA. We measured the activity of each mutant against multiple substrates and inhibitors, obtaining the same quantitative biochemical parameters as obtained from traditional assays: kcat, KM, kcat/KM, and KI. Together, these measurements define a quantitative functional ‘fingerprint’ for each mutation. Mutants having similar fingerprints distinctly cluster around the specific active site residues that display similar mutational fingerprints when themselves mutated. These comparisons reveal functional connections between individual active site residues and those throughout the structure, and suggest a high degree of modularity within the enzyme. The ability to rapidly characterize thousands of mutants with the quantitative precision of traditional assays will allow us to efficiently test models from computation and phylogeny, while providing the necessary data to improve future design efforts.