Utilizing natural diversity to evolve protein function: applications towards thermostability

Abstract

Protein evolution relies on designing a library of sequences that capture meaningful functional diversity in a limited number of protein variants. Several approaches take advantage of the sequence space already explored through natural selection by incorporating sequence diversity available from modern genomes (and their ancestors) when designing these libraries. The success of these approaches is, partly, owing to the fact that modern sequence diversity has already been subjected to evolutionary selective forces and thus the diversity has already been deemed 'fit to survive'. Five of these approaches will be discussed in this review to highlight how protein engineers can use evolutionary sequence history/diversity of homologous proteins in unique ways to design protein libraries.

Figure 1. Approaches and their Use of Evolutionary Sequence Space

Theoretical phylogenetic trees are depicted for each of the five approaches with circled regions representing the sequence space sampled by each when designing variants. DNA shuffling uses recombination between extant sequences. Consensus sequence uses extant sequences to design a single consensus variant. ASR resurrects ancestral sequences at internal nodes of the phylogeny. AMM replaces some residues in an extant sequence with ancestral amino acids. REAP replaces residues in an extant sequence with ancestral amino acids associated with functional divergence between two subfamilies in a phylogeny (depicted by yellow and purple branches).

Figure 2. Thermostability of EF-Tu variants

(a) A phylogenetic tree of the EF-Tu family of proteins. Black circle denotes the E. coli node, red indicates nodes used to construct the consensus sequence, green indicates the ancestral nodes leading back to a proteobacteria ancestor (PBA) circled in grey, and blue indicates the node for the last bacterial common ancestor (LBCA). (b) A multiple sequence alignment of EF-Tu variant sequences. Colored boxes indicate residues differing between the variant and extant E. coli sequences. (c) The protein variants were cloned, purified and their melting temperatures (in degrees Celcius) were measured by circular dichroism.

Figure 3. Exploiting evolutionary sequence information can increase functional information content of variant libraries

Diagram depicting how each of the approaches may relate to the functional information content of their variant library designs. Approaches such as REAP and AMM that can target a small set of mutations to design proteins with desired properties have the potential to have very high functional information content while approaches such as DNA shuffling that require large library sizes generally have lower functional information content.