Mechanism of MEK1 phosphorylation by the N-terminal acidic motif-mediated asymmetric BRAF dimer

The RAS->RAF->MEK->ERK cascade carries mutations in most human cancers. Interestingly, although we have three RAF paralogues (A, B and C), it is the BRAF that is predominantly mutated in cancer patients.

In 1999, Marais lab showed that this is because of the 4-residue N-terminal acidic (NtA) motif. In CRAF, the NtA must be phosphorylated to become acidic and the kinase to become active. In BRAF the NtA is already acidic, making BRAF one step closer to (over)active.

15 years later, Susan Taylor et al proposed a model: Active RAFs are dimers and one subunit (the “Activator”) uses its NtA to activate the other subunit (the “Receiver”). It is only the Receiver that takes the fully active kinase conformation.

The problem? Multiple BRAF and CRAF structures were solved since, and … dimers were symmetric, kinases looked active, and the NtA looked disordered.

However, to paraphrase the late Cyrus Chothia, if the data don’t fit a beautiful model, you need more data. You can’t just forget about the NtA – cancer genetics and the early Marais experiments don’t allow it!

This work started by our collaborator, Yasushi Kondo from the Paul Scherrer Institute, solving an X-ray crystal structure of BRAF in complex with its substrate MEK1 where the BRAF dimeric interface was asymmetric, with the NtA motif of only one of the subunits making interface contacts. Yes, like in that beautiful model!

But what is the difference between active asymmetric and active symmetric structures? To answer that, we made BRAF mutants
(i) AAAA mutant = Receiver but not Activator
(ii) Kinase-dead mutant = Activator but not Receiver
(iii) Double mutant = neither

If we titrate those mutants into cells with wild type RAFs in the background, we should see a dose dependent effect. And we did. The beautiful model must be correct!

One thing took us by surprise, though. It seems that the asymmetric dimer comes at the end of the RAF activation cycle. After the membrane localisation and after the removal of the inhibitory pS365 phosphorylation which is necessary for the rearrangement of the BRAF:14-3-3 interface.
We then got more structures and saw that the Receiver looks the most like a canonical active kinase. Is the asymmetric conformation then the final catalytically active state of RAF?

One thing is for sure, fun with RAF never ends!