Towards directed evolution of oligomeric enzymes

Protein complexes are omni-present in nature but mostly un-exploited in directed evolution experiments. In this work, two different strategies for directed evolution of homo-oligomeric enzymes are tested. For both strategies, one copy of the gene are cloned in a plasmid and another copy in a phagemid. Both gene copies are modified by random mutagenesis. The libraries of mutants are combined by infection of the library of plasmids by phage particles prepared from the phagemids library. The first strategy consists in selecting variants of the monomeric circularly permuted β-lactamase TEM-1 (TEM-1cp) oligomerised by domain swapping, based on the presence of the β-lactamase activity. Unfortunately, oligomeric β-lactamases could not be isolated or even observed, probably because of their toxicity. The expression of the wild type TEM-1cp also resulted in cyto- toxicity, associated to the catalytic activity of the enzyme. The second strategy developed consists in the evolution of the homodimeric isocitrate dehydrogenase of Escherichia coli (IDH) into a heterodimer, mimicking natural gene duplication. We could show that the products of the two copies of the gene (further called paralogs) may interact and form heterodimers. Moreover, the introduction of one different mutation in each copy of the gene un-expectedly affected the dimerization properties of the heterodimer, giving an example of epistasis between two interacting paralogs. The heterodimers were produced in vivo in small amounts, positively correlated to protein expression levels and suggesting a co-translational mechanism of dimerization for IDH. A directed evolution experiment has been set up to evaluate the effect of heterodimerization on protein evolvability. Finally, DmlA, a protein from the same family as IDH, has been characterized and features high catalytic versatility.