Protein Degradation Time

by: Derek Lowe

I’ve written a couple of times about the technique for selectively targeting and destroying proteins by forcing them to be ubiquitinated and hauled off to the proteosome. Both the Bradner lab (at Harvard/Dana-Farber) and the Crews lab at Yale found ways to do this a couple of years ago through broadly similar methods.

You create a hybrid species, one end of which is a ligand or binder of some sort to the target of interest, while the other end has a known bait for some ubiquitination protein complex. Bringing that in close proximity to your target protein sets the complex to ubiquitinating away on said protein, and that targets it for proteosomal degradation.

Here’s a new paper that shows the technique in action, courtesy of a team from Freiburg, Halle, and Budapest. They had previously discovered small-molecule ligands that selectively target Sirt2 (yep, a target of the infamous sirtuin inhibitors). That’s not so easy – getting chemical matter that hits the sirtuin proteins is certainly doable, but making it selective has been more of a job. It’s certainly worth trying, though, since Sirt2 has been implicated as a player in a whole list of important cellular (and disease) pathways, both through its ability to acetylate histones and other substrates, and as part of various protein-protein complexes.

The new compounds (led by the appealingly named SirReal2, at right) have a different binding mode than had been seen before – they induce a rearrangement of the protein, putting it into an open conformation that exposed an entirely new binding surface.

So with ligands in hand, the same group set about adapting them for the proteosome degradation trick. A scan through the SAR of the series had already suggested that the napthyl in SirReal2 could be replaced by a 3-(propargyloxy) phenyl and its resulting triazole “click” products, which led to a classic biotinylated derivative suitable for protein pull-down experiments. Both the Crews and Bradner groups had already shown the utility of thalidomide (and phthalamide derivatives in general) as baits for the ubiquitination end of the needed hybrid molecules, and adapting this to an azido-substituted thalidomide derivative was straightforward.

This seems to be the first appearance of a triazole-linked species for this technique. It most certainly won’t be the last – I feel sure that the azido-thalidomide derivative in this paper will be a catalog item pretty soon.

The resulting species had activity of about 250 nanomolar against Sirt2 in vitro, and in cells it showed a dose-dependent depletion of the protein, just as planned. The downstream effects in cells were notably stronger than those seen by straight pharmacological inhibition of the protein, as you might expect, since in this case, you’re taking it completely out of commission as opposed to blocking one of its binding sites or altering its conformation. This should be a very interesting tool to help unravel more of Sirt2’s functions, which certainly need more unraveling.

A lot of people (and a lot of companies) have been excited about this technique, for just these sorts of reasons. It’s a completely new way to interfere with a protein’s function, working both through a different mechanism and on a different time scale. With a small-molecule inhibitor, you’ve blocked one part of the protein for however long your drug lasts in vivo, but you’ve probably left some of its other functions (as a binding partner for other proteins) intact.

With siRNA or its cousins, you have presumably beaten down the translation of the whole protein, but perhaps at the cost of some off-target effects. CRISPR-Cas9 will also take out the protein entirely, but at a somewhat deeper level in the process, perhaps more selectively, and certainly more permanently. (Indeed, there’s a growing literature showing different effects between RNA-based targeting versus CRISPR).

This new method, though, combines the small-molecule techniques and the ability to completely remove a protein from the equation that the genetic methods have. Soon there will be another field of the literature comparing and contrasting all three or four methods on individual proteins.

There are tricky parts. For one, you need a good binder for the target protein, which is not always easy to come by. I think that what we’ll see most of the time is this current paper’s pathway – a group finds a new small-molecule ligand for a protein of interest and says “Aha! Now to stick a thalidomide on it to degrade the thing” If you want to target p53 or cMyc, though, you’re going to have a rough time on the small-molecule-ligand part of the process.

Even with one in hand, another step is getting the hybrid species into the cells, and that is the usual voodoo. We have broad ideas about what kinds of compounds get into cells, of course, but it’s still an extremely empirical field, to put it gently. Finally, there’s the question of how well the whole ubiquitination/degradation pathway works once you’ve got your hybrid ligand in there. There are informal reports of variability in this step – things that should work better than they do, and other things that work a lot better than anyone thought, and the reasons for these changes are (so far) obscure.

So it’s the frontier, for sure. Watch as the number of papers in this area swells beyond the ability of any of us to read them all!