I like to say that when I was first working on G-protein coupled receptors (GPCRs) as drug targets, that after I while I thought I really understood a lot of what was going on. But that was the peak. Further experience eroded that confidence, and even though I’ve learned a lot more about them over the years, I think that my awareness of their complexities has always been outpacing my understanding of them. Keep in mind, we’re talking about an extraordinarily important and successful class of drug targets. You’d think that we would have a better grasp on their mechanistic details, but although we know a great deal about them, their landscape just keeps stretching out in front of us as we move on. Here’s another example of that: in this paper from earlier this year, a joint Belgian/German team looked at a key detail of GPCR signaling and peeled back yet another layer.

For those who are outside the field completely, let me furnish a few definitions up front. GPCRs are proteins out of the surfaces of cells, and they extend through the membrane with surfaces on both the outside and inside. The outside region binds all sorts of ligands, signaling molecules that range from very small (acetylcholine) to large proteins. When that binding occurs, the middle part of the GPCR changes conformation – there are seven helices that span the membrane in most of them, and they slide and shift around each other according to what’s bound to the outer (and inner) surfaces.

On the inner side are various cellular proteins, most famously the “G-proteins” themselves, which respond to those conformational shifts by setting off (or inhibiting) binding with their own partners, enzymatic activity, etc. This is the coupling mechanism between signals coming in from outside the cell to intracellular changes: the GPCRs are transducers. Inside the cell, there are “second messengers” whose concentration changes in response to GPCR binding events (calcium ions, various small phosphorylated molecules, and others), and these are coupled to a huge range of different processes in other cellular systems. The same GPCR can be mostly associated with a specific second messenger, or it can couple with various ones in different cell types or under different conditions.

Now, those definitions: an “agonist” is a molecule that binds to a GPCR and sets off those second-messenger signals. Examples are dopamine, serotonin, and (as an example of an outside substance) morphine. An “antagonist” is a molecule that binds to the GPCR in a way that blocks agonist signaling. As an example, for morphine and other opioid agonists, an antagonist is naloxone. Simple, eh? Har, har. Just to give you a peek, there are “partial agonists”, compounds that bind, set off the agonist response, but never to the level that a “full” agonist can reach. And there are “inverse agonists” – compounds that bind, and set off a response, and whose activity is blocked by antagonists – but the response they trigger is the opposite of the “normal” agonists. To add to the confusion, these sometimes get classed as antagonists, since they block the regular agonists as well.

The complication that this team is looking at is the presence or absence of a G-protein. They’re not present and bound at all times, and just that act of binding (as you’d imagine) shifts the conformation of the GPCR’s outer regions. But determining what effect this has on agonist binding has been tricky. What they’ve done here, though, is to come up with a nanobody that binds like the G-protein does, at its same site on the inside of the cell, using the extremely well-studied beta-adrenoceptor as a model system. They then engineered the receptor to have this nanobody region attached at the intracellular end, connected by a flexible linker, so basically it always exists in the G-protein-bound conformation, since this thing is constantly right on top of it. As a control, they expressed another beta-receptor with a nanobody that wouldn’t mimic the G-protein (but would keep the endogenous ones from getting in position). That gives you your choice of always-G-protein-bound receptors and never-bound ones, which is a lot more clear than the usual cellular situation of a constantly shifting mix of everything in between.

When they tried a list of beta-receptor ligands against these two types, the differences were striking. Epinephrine (the natural agonist) binds to the G-protein-on form 2000x better than the GP-off form. But a plain-vanilla antagonist (alprenolol) binds both of them equally well. Meanwhile, an inverse agonist (ICI 118,551) binds the GP-off form better than it does the GP-on one. The authors went on to run a 1000-member fragment library past these two receptors, and binned the hits into those three categories. Elaborating some of the full-agonist hits gave new compounds that had sub-nanomolar affinity for the GP-on beta receptors, up to 10,000 fold less affinity for the GP-off ones, and were very active in regular cellular assays (second messenger concentration). Interestingly, some of these compounds did not recruit one of the major competitors to the G-proteins (beta-arrestin) which suggests that they were completely biased towards the G-protein-bound form of the receptor.

And this work suggests that you can separate agonists, antagonists, and inverse agonists with just a binding assay, without having to measure effects on second messengers, competition with known ligands, or the other things that normally would be considered essential for making that call. It would be interesting to see how this idea (nanobody activation) works with other GPCRs, and also to see if you can do the same trick with other non-G-protein binding proteins like the beta-arrestin just mentioned. (I realize that each of these proposals is a pretty serious research project all its own!) But the only way to an understanding of these things is through all these complications. Onward!