Enzymes that employ cofactors to perform catalysis are ubiquitous across the tree of life and are necessary to assist half of the enzymatic reactions in nature. Among cofactors, nucleotide derivatives are believed to be the catalytic fossils of a hypothetical RNA-based world. They may have become associated with early peptides, which evolved towards the enormous catalysis we observe today. Given their importance, it is fundamental to understand how their binding modes emerged within protein scaffolds. This understanding will allow us to switch cofactor specificities and therefore functions across protein families and superfamilies. To date is widely accepted that protein scaffolds may have evolved through amplification and combination of ancient peptides. In addition, random mutagenesis has proven to alter protein functions. However, the role of insertions and deletions (InDels) in the emergence of protein functions remains underexplored. Not only are InDels highly deleterious, but also difficult to implement in randomized libraries, compared to amino acid substitutions.
We systematically searched for InDels within coenzyme binding pockets to study the hypothesis that insertions and deletions were involved in the divergence of coenzyme-binding in early evolution, and as a proof of principle, we engineered InDels to remodel the nicotinamide dinucleotide (NAD) binding pocket of oxidoreductases into the S-adenosylmethionine (SAM) binding pocket of Rossmann methyltransferases. The elucidation of the mutants’ crystal structures and binding studies corroborated the orthogonal binding switch from NAD to SAM, providing the first example where the implementation of InDels in rational design leads to a coenzyme-specificity switch that links two distinct enzymatic chemistries (redox and methylation) in Rossmann proteins. We believe this strategy can be generalized and will constitute a handy asset to the current protein engineering methods.