The Emergence of an Engineered Enzyme

Ask a chemist or any scientist in general what would be one of the top currently-nascent technologies they would want to mature quicker, and they would surely name protein engineering as one of them. A recent paper in Science brings us closer to this reality.

Just quickly, for the non-scientists: proteins are little molecules in your body that catalyze many reactions (called enzymes) that help you digest food, get energy, and just LIVE in generally, they also make up a lot of the building blocks for various parts of your body.

To be able to create proteins with various non-natural functions, as well as create enzymes with better activity, it is sometime necessary to predict the conformations that these synthetic proteins will take. The interactions of the amino acids (protein building blocks) on various parts of the proteins can be mindbogglingly varied depending on the length of the protein and environment. Computer programs are often used to predict the lowest-energy configurations and the final products. Not only do you need powerful computers to run the simulations, but you need efficient and realistic simulator programs, of which there is a whole slew out there (with no clear winners).

The quest towards engineering any protein on demand is one that is slowly but surely being achieved with the help of better predictive computational algorithms. Here's part of the abstract of the paper:

The Diels-Alder reaction is a cornerstone in organic synthesis, forming two carbon-carbon bonds and up to four new stereogenic centers in one step. No naturally occurring enzymes have been shown to catalyze bimolecular Diels-Alder reactions. We describe the de novo computational design and experimental characterization of enzymes catalyzing a bimolecular Diels-Alder reaction with high stereoselectivity and substrate specificity.

This reaction is widely used in synthetic chemistry, so this means that much time and money could potentially be saved by using an engineered enzyme such as this.The team chose to build an enzyme using a design method called Rosetta. I don't pretend to know anything about it, but in the case that someone is interested, I just put that up here. The researchers then prepared an in silico model of the shape that they'd need to accommodate the transition state for this reaction (high energy intermediate - remember chemistry class?). They then added amino acids to their protein that would hold the reactants in place.

Their first calculation showed that 10¹⁹ theoretical active sites matched several parameters they specified for the two molecules they wanted to join. They then used the RosettaMatch program to screen all these possibilities against already-known protein scaffolds and narrow down to 10⁶ possibilities that were stable.

This was further modeled and the number of enzymes was filtered down to 84, all of which the researchers expressed and purified. Of those, 50 turned out to be soluble and two had Diels-Alderase activity. The enzymes were also tuned to increase their activity.

That's the gist behind the rational design approach for protein engineering. This is of course very laborious, but is state of the art at the moment. There are certainly other approaches, such as protein evolution using single amino acid mutations, but these have their own problems and an in-depth look is beyond the scope of this post (though of course not the blog as a whole).

Protein engineering has a myriad of uses besides making enzymes for chemical synthesis. Imagine piecing together parts of different proteins to make them work quicker, attach more strongly to certain surfaces, play novel roles in signalling, etc. I personally cannot wait until this field matures and we can realize all it's possibilities.