Mechanisms of Protein Sequence Divergence and Incompatibility

Abstract

Alignments of orthologous protein sequences convey a complex picture. Some positions are utterly conserved whilst others have diverged to variable degrees. Amongst the latter, many are non-exchangeable between extant sequences. How do functionally critical and highly conserved residues diverge? Why and how did these exchanges become incompatible within contemporary sequences? Our model is phosphoglycerate kinase (PGK), where lysine 219 is an essential active-site residue completely conserved throughout Eukaryota and Bacteria, and serine is found only in archaeal PGKs. Contemporary sequences tested exhibited complete loss of function upon exchanges at 219. However, a directed evolution experiment revealed that two mutations were sufficient for human PGK to become functional with serine at position 219. These two mutations made position 219 permissive not only for serine and lysine, but also to a range of other amino acids seen in archaeal PGKs. The identified trajectories that enabled exchanges at 219 show marked sign epistasis - a relatively small loss of function with respect to one amino acid (lysine) versus a large gain with another (serine, and other amino acids). Our findings support the view that, as theoretically described, the trajectories underlining the divergence of critical positions are dominated by sign epistatic interactions. Such trajectories are an outcome of rare mutational combinations. Nonetheless, as suggested by the laboratory enabled K219S exchange, given enough time and variability in selection levels, even utterly conserved and functionally essential residues may change.

Author Summary

Orthologs are proteins in different species sharing the same function and structure. However, the mechanisms that underline the divergence of different sequences from a single ancestor remain unclear, particularly because many amino acid exchanges between orthologs result in loss of function (incompatibility). We aimed at disentangling an ancient divergence event within the active-site of a universally spread enzyme that mediates ATP synthesis. Using laboratory evolution experiments, we found that an exchange in a functionally critical active-site residue that is incompatible within contemporary orthologs is enabled by few mutations. These mutations lead to transition sequences in which, unlike the extant sequences, a wide range of amino acids is tolerated. Our experiment reveals the properties of these transition sequences that may resemble the historical ancestral states that underlined this divergence event, and the mechanisms that led to incompatibility within the contemporary orthologs. Our results support theoretical predictions and reshape our understanding of protein structure-function. That a given position is entirely conserved and essential for function does not indicate that it will never exchange, but rather, that the exchange may depend on changes in many other positions.