Serine proteases found in vertebrates today come in various flavors: trypsin-like enzymes cleave after basic residues, but nontrypsin-like proteases favor nonbasic residues. Husain's group used phylogenetic inference to predict the structure of ancient serine proteases. They find that, although the most ancient proteases were specialized with trypsin-like qualities, the increased diversity that later spawned nontrypsin-like qualities was achieved by recreating a less specialized intermediate.
An in vitro–produced enzyme based on the predicted sequence of this less specialized ancestor did indeed have broad substrate specificities—a feature not found in its descendants. This promiscuity seems to be due to a wider entrance to the substrate pocket that would allow diverse side chains to bind in the cleavage site.
The despecialized intermediate could be mutated in its substrate-binding pocket so that its substrate preference more closely resembled modern proteases. In the modern proteases, attempts to change binding specificities to that of other classes only kill the enzyme. So the intermediate was uniquely able to tolerate mutations that might lead to diversification.
The intermediate is an ancestor of serine proteases that are important in immune defense responses. According to Husain, “respecialization in duplicated daughter genes would have allowed the evolutionary narrowing of specificities … thereby increasing the repertoire of efficient armaments necessary for efficient host defense.”
Husain hopes to determine the structural basis of respecialization using crystallography. “The answers gained would not only have evolutionary significance,” he says, “but could also allow us to predict and make designer proteases with dial-in specificities—[proteins like] restriction enzymes for cutting polypeptides exactly where we want.” ▪