Fig. 1: The β-lactamase variant design process.
From: Simultaneous enhancement of multiple functional properties using evolution-informed protein design
Strategy applied to generate and test design variants using an evolution-informed statistical model of the β-lactamase protein family. [A, Computationally Model β-lactamase] WT TEM-1 β-lactamase was used to generate a multiple sequence alignment that was used as input to derive an EVcouplings maximum entropy model. The predicted fitness (EVH) for any sequence (σ) can be calculated as the sum of coupling terms ( Jij) between every pair of residues as well as site-wise conservation terms (hi). [B, Computationally Design β-lactamase Variants] Design variants are generated by using Gibbs sampling to iteratively optimize an objective function that takes into account EVH and sequence similarity to WT TEM-1, to natural homologs, as well as to the other designed sequences. [C, Experimentally Test Designs] Designs were synthesized, cloned into plasmids, expressed in E. coli, and several experimental protocols were performed to characterize each design including cell-based activity assays, biochemical kinetics assessment, and structure determination and analysis.
