BRAF beyond V600: class II and III mutations and RAF dimers

Almost everything written about BRAF in cancer is really about one residue: V600. But roughly a third of BRAF mutations seen in the clinic fall outside codon 600, and they do not behave like V600E at all. They signal through RAF dimers rather than monomers, some of them are catalytically dead yet still oncogenic, and the V600-selective drugs that transformed melanoma either do nothing against them or actively make them worse. The framework that makes sense of this — the class I / II / III scheme from Zvi Yao and Neal Rosen’s group — is worth understanding before you dock anything against a non-V600 BRAF.

Three classes, defined by how the kinase signals

The classes are not about where the mutation sits in the sequence; they are about the mechanism of ERK activation. Yao et al. (Cancer Cell 2015; Nature 2017) sorted BRAF mutants into three groups:

• Class I — V600 mutants (V600E/K/D/R). Constitutively active as monomers, fully independent of upstream RAS. This is the only class the approved BRAF inhibitors were designed for.
• Class II — non-V600, high or intermediate activity.Examples include K601E, L597Q, G469A/V, and the BRAF fusions (e.g. KIAA1549-BRAF). These signal as constitutive dimers, still RAS-independent, but they need two protomers working together rather than one.
• Class III — kinase-impaired or kinase-dead.Examples include D594G/N, G466V/E, G596R, and N581S/I. These have low or no intrinsic kinase activity, yet they are oncogenic because they bind RAS more tightly than wild type and amplify signaling through CRAF. They are RAS-dependent: they only drive tumors when there is a coexisting source of active RAS, such as an NRAS/KRAS mutation, NF1 loss, or upstream receptor tyrosine kinase signaling.

Why vemurafenib fails outside V600

Vemurafenib, dabrafenib, and encorafenib are monomer-preferringtype-I RAF inhibitors. They bind the active conformation of a single BRAF protomer, which is exactly what a V600E monomer presents. Put one of these drugs in front of a RAF dimer and you hit the defining failure mode of the whole class: paradoxical activation. The inhibitor occupies one protomer of the dimer, which allosterically transactivates the unoccupied partner protomer. Net ERK output goes up, not down.

That is not a theoretical curiosity. It is the mechanism behind the cutaneous squamous-cell carcinomas seen in vemurafenib-treated patients whose normal cells carry RAS mutations, and it is why a class II or class III BRAF tumor will often progress, not respond, on a V600 inhibitor. The structural lesson: if the oncogenic unit is a dimer, a drug that binds only one half of it can do the wrong thing.

Where the drugs are going

The non-V600 problem is fundamentally a dimer problem, so the design goal shifts from “bind the active monomer” to “shut down both protomers of the dimer without triggering transactivation.” Several approaches are in play:

• Tovorafenib (Ojemda) — a type-II pan-RAF inhibitor, FDA accelerated approval April 2024 for relapsed/refractory pediatric low-grade glioma driven by BRAF fusions or V600 mutations. It is the first targeted agent approved for BRAF-fusion (class-II-like) disease, precisely because type-I inhibitors paradoxically activate fusion- driven tumors. In FIREFLY-1 the overall response rate was 53%.
• Plixorafenib (PLX8394) — a “paradox breaker” designed to disrupt RAF dimerization rather than just occupy the active site, blunting the transactivation that drives paradoxical signaling.
• Pan-RAF / dimer inhibitors — belvarafenib and naporafenib (LXH254) target the dimer directly and have been studied in RAS-mutant and class II/III BRAF contexts, often paired with a MEK inhibitor to close the downstream escape route.
• Hitting upstream RAS for class III. Because class III mutants are RAS-dependent, Yao et al. (Nature 2017) showed they are sensitive to anything that lowers active RAS — RTK inhibitors in lung and colorectal disease, and the broader RAS-pathway toolkit (including the newer KRAS and SHP2 inhibitors) where the RAS lesion is drug-accessible.

What this means for a docking workflow

Two practical points. First, a monomer-binding score against a class II or III BRAF is misleading on its own — the biology that matters is dimerization and RAS engagement, which a single-protomer rigid dock does not capture. Second, the meaningful comparison is relative: a V600-selective inhibitor should look comparable against a V600E monomer and noticeably worse, or simply irrelevant, against a dimer-driven non-V600 setup. That ΔΔ is the structural echo of why these drugs split so cleanly in the clinic.

Try the docking yourself

Open Studio and pick BRAF from the target catalog. Dock a V600-selective inhibitor such as vemurafenib against V600E, then against a non-V600 residue, and watch the binding signature change. It is a fast way to build intuition for why the class I / II / III split is a drug-design fault line, not just a genetics footnote.

Liganx is molecular docking online: free, browser-based, no install. Running molecular docking against V600 and non-V600 BRAF side by side is the quickest way to see why one residue reshaped an entire treatment landscape.

Primary sources

• Yao Z, et al. BRAF mutants evade ERK-dependent feedback by different mechanisms that determine their sensitivity to pharmacologic inhibition. Cancer Cell 28, 370-383 (2015). doi:10.1016/j.ccell.2015.08.001
• Yao Z, et al. Tumours with class 3 BRAF mutants are sensitive to the inhibition of activated RAS. Nature 548, 234-238 (2017). doi:10.1038/nature23291
• U.S. FDA. FDA grants accelerated approval to tovorafenib for patients with relapsed or refractory BRAF-altered pediatric low-grade glioma (April 23, 2024). fda.gov