Optimizing non-natural protein function with directed evolution

Abstract

Developing technologies such as unnatural amino acid mutagenesis, non-natural cofactor engineering, and computational design are generating proteins with novel functions; these proteins, however, often do not reach performance targets and would benefit from further optimization. Evolutionary methods can complement these approaches: recent work combining unnatural amino acid mutagenesis and phage selection has created useful proteins of novel composition. Weak initial activity in a computationally designed enzyme has been improved by iterative rounds of mutagenesis and screening. A marriage of ingenuity and evolution will expand the scope of protein function well beyond Mother Nature's designs.

Figure 1

Directed evolution of novel UAA-dependent binding proteins using phage display [22•,23••,25•]. (a) General phage display protocol for selection of HIV gp120 specific ScFv proteins. X – E. coli refers to bacterial strains engineered for the incorporation of UAA into proteins in response to the TAG stop codon. (b) CDR3 sequences of gp120 specific sulfotyrosine (sY)-containing antibody, 412d, and new evolved variants are shown. Replacing sY with natural tyrosine reduces binding of the wild type (412d-YY) and evolved variants. (c) Structure of sulfotyrosine. (d) Novel reactive groups such as boronic acids can provide chemical warheads for binding molecules such as hydroxyl-rich sugars.

Figure 2

Directed evolution of a computationally designed Kemp eliminase [10••,47••]. (a) Kemp elimination reaction mechanism (left), transition state (middle), and product (right) are shown. (b) A schematic of the minimal computationally designed active site of parent Kemp eliminase Ke07 is shown on the left. In the designed model, the distance between Glu101 and Lys222 is approximately 3.6 Å. The observed distance between these two residues is significantly shorter, according to the crystal structure (right, PDB: 2RKX). (c) 7 rounds of directed evolution starting from Ke07 improved catalytic efficiency more than 200 fold. A crystal structure for a variant from round 4 of evolution (4-1E/11H, PDB 3IIO) is shown on the right. The interaction between Glu101 and K222 is significantly diminished in this structure.

Figure 3

Developing technologies such as unnatural amino acid (UAA) mutagenesis, non-natural cofactor engineering and computational design can facilitate the introduction of new function, distinct from those demonstrated in Nature, within a protein scaffold. Fitness is a measure of how well protein sequences perform the target function. Most proteins possess no activity (shown in grey on the fitness landscape). Successful designs (shown as circles) may show weak initial activities that can be starting points for further optimization by directed evolution. Directed evolution leading only to small improvements may indicate that an initial design lies near low fitness peaks (red trajectory) or that further improvements are limited by the presence of local optima (blue trajectory). An ideal design would yield a sequence that lies on a large fitness peak (green trajectory), where multiple rounds of mutagenesis and screening produce a highly functional protein.