Read on Twitter
The latest addition to the story of the insulin-like growth factors and their receptors is out today in Structure! My contribution was (as usual) in helping with the model building - this was driven by Mike Lawrence's team at @WEHI_research
https://www.cell.com/structure/fulltext/S0969-2126(20)30171-4 (1/9)
https://www.cell.com/structure/fulltext/S0969-2126(20)30171-4 (1/9)
This family is both the holder of a special place in my heart as my entry-point into the world of structural biology, and the source of a certain level of angst in being the trigger for my largely giving up on attempting to *predict* receptor-ligand interactions... (2/9)
Let me try to give you some idea of why. IGF-II looks like this. The residues highlighted in green (particularly the Phe-Tyr-Phe on the left) are critical for receptor binding but appear stable here, in insulin, in IGF-I, free or in complex with carrier proteins. (3/9)
So naturally you'd expect it to just "dock" onto the receptor fairly rigidly, right? Wrong. Here's what actually happens. These residues pull away, opening up the "C loop" (pointing up), and then the tail end of the receptor *threads itself through it*. I mean, what?? (4/9)
This isn't just a fluke - the same essential behaviour has now been shown for IGF-I and -II in complex with the IGF-I receptor, and for IGF-I in complex with the insulin receptor, in cryo-EM and crystallographic experiments. (5/9)
Incidentally, it neatly explains (in the weirdest possible way) why one splice variant of the insulin receptor binds IGF-I well while the other doesn't: the poorly-binding variant has an extension on this "tail", making it simply too long to thread itself through the loop. (6/9)
And then we come to the overall conformational change of the receptor itself. The obvious point of flexibility is the "hinge" connecting the cysteine-rich region to the L2 domain (red)... my prediction was that L1 (bottom) would swing across to engage the opposite leg. (7/9)
Instead, this is what happens. The receptor undergoes a truly *massive* rearrangement to effectively touch its own head, to the point where you have to spend a while looking at it just to convince yourself you're looking at the same protein. Start: 5u8r; final: 6vwi/j. (8/9)
The upshot that I can't escape: even if I'd managed to successfully predict what actually happens here, my results would have been rejected out of hand by any reputable journal as much too far-fetched. And to be perfectly honest, I wouldn't blame 'em. Biology is *weird*. (fin)
... oh yeah. Images and movies in this thread were generated in @UCSFChimeraX. #WeNeedChimeraX
