We have spent a great deal of time on G-protein coupled receptors in drug discovery, and well we should. GPCRs are vital signaling systems in all organ systems, and what’s more, many of them have small-molecule binding sites that we can target directly with our drug candidates. Dopamine, serotonin, acetylcholine, histamine, epinerphrine, the opioid peptides. . .there are a lot of really important GPCR ligands out there. And while we’ve learned a huge amount about binding sites and signaling mechanisms, it has always been abundantly clear that there are many important features that we don’t understand.
This new paper illustrates that well. It’s investigating the “dry pockets” that have been found in some GPCR structures. These things, which have shown up in other protein classes over the years as well (blog post here about this from 2017), are rather mysterious. As the name implies, they’re outright voids in the protein structure, cavities of various sizes that don’t seem to have any small-molecule ligands associated with them. Some of them have hydrophilic amino acid side chains pointed into them and are thus presumably associated with water molecules – although “filled with water” isn’t a very useful mental picture when you’re talking about three or four water molecules total! But others have just hydrophobic side chains and are thus very unlikely to have favorable interactions with any waters. So what’s in them? Empty space?
Apparently. That earlier blog post link shows an example of that, and it’s also an example of using noble gases like xenon to investigate these things. Those are of course single atoms, since they barely react with anything and certainly not with each other, and they really have no outright polar character at all. They’re one of the few things that actually seem to accumulate in these hydrophobic cavities, and these experiments have been valuable in tracking them down and trying to understand what (if any) function they might have. In this latest paper, it’s the beta-1 adrenergic receptor that’s under investigation. In common with other GPCRs, it’s quite conformationally mobile. There are X-ray structures of these things now (something that didn’t exist when I started out in the drug industry, and would have gotten raised eyebrows if you’d predicted it!), but as always, an X-ray structure is a static picture and doesn’t reveal important protein motions. In fact, some of the X-rays of GPCRs bound to agonist small molecules are more or less indistinguishable from other X-ray structures of the same receptor bound to antagonists, which tells you right there that we’re missing something.
In the case of the beta-1 receptor, NMR studies (which can pick up more dynamic behavior) have shown that the antagonist complexes are pretty fixed, while the agonist ones are an equilibrium between the active conformation and a “pre-active” one that looks a bit more like the antagonist complex. That equilibrium is really on a knife edge and can be shifted around by all sorts of factors – and it turns out that one of them is sheer physical pressure. If you do high-pressure protein NMR (a rather exotic technique that not many people are set up to study), you find that the beta-1 receptor at 600 bar is shifted almost completely to the active conformation, and varying the pressure suggests that this happens because of the collapse of a void of about 100 cubic angstroms. And there are your dry pockets, presumably – the authors were able to locate two of them with xenon atoms occupying them, although it’s hard to say much about the size and behavior of these voids by just their xenon occupancy rate.
This ties in with the interaction of GPCRs and other transmembrane proteins with cholesterol. Cholesterol is firmly fixed in the public imagination as a bad actor in human health, but while it can be just that in the cardiovascular system, it’s also essential for lipid membrane function in every cell in the body. This paper identifies a dry void in the beta-1 receptor that seems to be right on top of a known cholesterol binding site – and cholesterol itself has already been characterized as a negative allosteric regulator of this receptor. The story comes together: a cholesterol molecule can occupy this hydrophobic void (since cholesterol itself is a famously greasy and water-avoiding substance all its own), and when it does so, it acts as a “wedge” that keeps the receptor from switching over to its active state (which would require collapsing that void, just as it happens in the high-pressure experiments). And that’s how it’s a negative regulator. The authors suggest that searching for other such voids in GPCRs (and other proteins) could reveal allosteroic regulatory sites that we haven’t been aware of – ones that are dynamic and more or less disappear completely when the proteins involved are in their active states. Experimentally detected binding of xenon atoms, of all things, would lead the way!
