Here’s an interesting paper that combines two fields that normally don’t intersection much: protein evolution and organometallic catalysis. That Venn diagram overlaps so sparsely, of course, because the sorts of organometallic chemistry that gets done in labs is largely incompatible with living systems – if you’re handling things in a glove box or with Schlenk-line technique, it’s probably not going to hold up so well in a fermentation flask.
Or maybe it will. This new paper, from the groups of Sven Panke at the ETH and Thomas Ward at Basel, described a ruthenium-containing enzyme that can catalyze olefin metathesis. This sort of enzyme has been reported before, and that was quite a feat in itself. Catalytic metalloenzymes have been worked on for many years, but it’s a tough field. And when you’ve worked one out, as with many other sorts of modified enzymes, there may be a feeling of being left standing on the other side of the glass: harnessing the extraordinary power of biological evolution is clearly the way to go for optimizing a protein, but when you modify them to that extent, you often have to close that world off. Over the decades, we’ve hijacked the evolutionary toolbox of molecular biology for a lot of uses, but there are still limits.
One key to this paper is how and where the enzyme is expressed: in the periplasm of E. coli bacteria. That’s a really good idea, because it’s a lot less hostile to the metal centers than the cellular environment itself. One of the main reasons for this seems to be inactivation by glutathione, which is, after all, just the sort of thing glutathione is supposed to be doing for the cell, but there’s not much of it out in the periplasm. The organometallic part was brought in (as it has been in other experiments of this general type) by a biotin/streptavidin route. This paper shows the generation of an E. coli strain that expresses streptavidin into the periplasm, and the team then introduced a biotinylated 2nd-generation metathesis catalyst, which did get taken up by the bacteria (about a threefold increase in their Ru content), and did react with the streptavidin. (Those are some not-dismissable hurdles to clear right off).
It was already known (from work in the Ward group) that the biotin-streptavidin-Grubbs/Hoyveda species would catalyze olefin metathesis in water, so the team targeted formation of (fluorescent) umbelliferone as a marker for the reaction in bacteria. And it does happen – none of the controls showed any of the desired fluorescence, except the ones that had the specific combination of periplasmic streptavidin and the biotinylated Ru catalyst.
This would actually be enough of a result for a good paper right there. But this one goes further, seeking mutants of the periplasmic streptavidin species that might perform better. The team set up a high-throughput mutant assay to search for higher-fluorescence strains, and was able to run through thousands of variations in a few days. At the end of the process, a variant with five mutated residues near the Ru end of the conjugate showed about five-fold greater activity.
Now the question comes up about what you’d do with this metalloenzyme. Had it been tuned up just for umbelliferone formation, or would it be a general catalyst? Some of each, apparently: when tried out in water against standard ring-closing metathesis substrates, two different mutant forms did show some varying preferences in how well they catalyzed the reactions as compared to the original complex. But all of them outperformed the commercial catalysts. (Adding glutathione to the reactions killed them, by the way).
So it looks like the periplasmic expression technique can probably be used to find mutants that will catalyze other non-biological metal-based reactions as well – perhaps evolved against specific substrates, or perhaps to deliver some wide-ranging catalysts at the same time. I look forward to seeing what use people will make of this, because the possibilities are quite interesting!