Restoring the brake: Using molecular glues to fix a broken CRAF off-switch in Noonan Syndrome

The RAS Dialogue Blog posts are written by RAS experts sharing the latest research, updates, and scientific RAS news. The content is curated by the RAS Initiative.

Image

Image

When the brakes fail

The MAPK signaling pathway is among the most tightly choreographed relay systems in human biology, governing cell growth, differentiation, and survival in response to extracellular cues. In cancer, this pathway is frequently floored at full throttle, with oncogenic RAS mutations driving 100s-fold increases in pathway activity. But in another class of disorders the pathway runs too hot through a subtler mechanism. RASopathies are developmental syndromes caused by germline mutations that modestly but persistently hyperactivate MAPK signaling.

Noonan syndrome (NS) is the most common RASopathy, occurring in roughly 1 in 2,000 live births. Children with NS can experience cardiac defects, growth delays, craniofacial abnormalities, and an elevated cancer risk. These are profound consequences arising from what can be just 2- to 10-fold increases in MAPK output; the degree of activation varies between studies and with different NS mutations. This sensitivity reflects how finely tuned normal development is to precise pathway thresholds.

For NS patients, the therapeutic landscape has been particularly sparse. Trametinib, a MEK inhibitor used off-label, can help manage some complications, but it broadly suppresses MAPK activity, which carries its own risks in developing tissues. The field needs a more targeted strategy. NS mutations have been found in at least 13 proteins; each is individually rare, but together they represent a significant patient population. A mechanism that could target all, or nearly all, of them would therefore be highly preferable. Our recent work suggests that molecular glues may provide both the selectivity for the NS-activated pathway and the generality to target the large majority of NS mutations.

A natural brake pad: The 14-3-3/CRAF regulatory complex

To appreciate why molecular glues are an appealing strategy here, it helps to understand how CRAF is normally kept in check. Under basal conditions, CRAF is held in an inhibited state by the hub protein 14-3-3. When CRAF is phosphorylated at serine 259 (pS259), 14-3-3 binds to that site and physically occludes the regions of CRAF that interact with RAS and other RAF kinase subunits to form active RAF kinase dimers. Think of 14-3-3 as a brake and pS259 as the pedal that holds the car at the stoplight. 

NS mutations in CRAF cluster around the S259 regulatory site. They disrupt the 14-3-3/CRAF interaction through two compounding mechanisms: they directly weaken the binding affinity for 14-3-3 and reduce phosphorylation at pS259 (likely in part because the lowered affinity leaves pS259 more accessible to dephosphorylation). We quantified these effects across four clinically relevant CRAF mutants (R256S, S257L, P261A, and V263A) and found that binding affinities were impaired 3- to >100-fold compared to wild-type CRAF. The mutations most proximal to pS259 (S257L and P261A) showed the greatest disruption. As a result, the kinase is constitutively more available to RAS and more prone to forming the active dimeric state, producing chronic low-level hyperactivation of ERK (a common marker of MAPK pathway activity).

Image

This regulatory axis, the 14-3-3/CRAF-pS259 inhibitory complex, remains largely unexploited therapeutically. We asked: could we pharmacologically reinforce this complex, even in the context of mutations that have partially broken it?

Molecular glues: Stabilizing what mutation has weakened

Molecular glues are small molecules that stabilize protein-protein interactions (PPIs) by binding cooperatively at the composite interface formed between two proteins. Rather than binding to a single protein, they wedge themselves into the space between two partners and hold them together more tightly. This mechanism makes them well-suited to situations in which a regulatory interaction has been weakened – precisely the scenario presented by NS and the 14-3-3/CRAF regulatory node.

The Arkin, Brunsveld, and Ottmann labs previously developed covalent molecular glues targeting the 14-3-3σ/CRAF-pS259 complex.¹ These compounds engaged a native cysteine (C38) in the 14-3-3σ isoform and bound at the composite interface formed between 14-3-3σ and the CRAF pS259 motif, stabilizing the inhibitory complex by 100s- to 1,000s-fold in biophysical assays. Here, we asked whether these same molecular glues could restore the interaction in NS mutant backgrounds.² This is a stringent test, since the mutations had structurally perturbed both the phospho-recognition site and the binding interface itself.

The answer was largely “yes.” In biophysical binding assays using phosphorylated CRAF mutant peptides, these molecular glues restored apparent binding affinities to or beyond the wild-type affinity for three of the four mutants tested (R256S, S257L, and V263A). The P261A mutant was more resistant to rescue, likely reflecting a larger structural disruption in the local binding motif.

Restoring the brake inhibits the pathway in cells

The same held true in a cellular context. Using NanoBRET assays to measure the full-length 14-3-3σ/CRAF PPI in live HEK293T cells, we confirmed that each NS mutant weakened the cellular interaction and that the molecular glues could restore it to near wild-type levels for R256S, S257L, and V263A. Importantly, stabilizing the 14-3-3/CRAF complex also increased pS259 levels, decreased CRAF recruitment to NRAS, and reduced CRAF dimerization. Molecular glue stabilization sequentially blocked the key steps in CRAF activation.

We also demonstrated that molecular glues could reduce pathway activity in cells harboring NS mutations upstream of CRAF. For example, activating mutations in SOS1, SHP2, and SHOC2 indirectly reduced pS259, driving CRAF into its active state. Here too, the molecular glue increased pS259 levels to reinforce the inhibitory complex with wild-type CRAF. Thus, the more common NS mutations – roughly 65% between SHP2 and SOS1 variants – were also correctable.

Ultimately, when we measured MAPK activity by ERK phosphorylation (pERK), the molecular glue reduced pERK to the same extent as trametinib in three NS mutant backgrounds (SHOC2^S2G, CRAF^S257L, and CRAF^V263A). This comparable efficacy is significant. Trametinib cuts off the engine; we propose that, by contrast, 14-3-3/CRAF molecular glues restore the brake by replacing the brake pad. 

What this means and where we go next

This study exposes several threads worth pulling.

First, it provides proof-of-concept that molecular glues can rescue weakened PPIs in a disease context, expanding their therapeutic potential beyond wild-type interactions. A small molecule binding at a composite interface can compensate for mutations that disrupt one partner's contribution — a principle that should generalize to other rare diseases in which regulatory PPIs are disrupted, such as loss of 14-3-3/USP8 in Cushing syndrome or 14-3-3/LRRK2 in Parkinson's disease.

Second, the molecular glue mechanism offers a biologically meaningful difference from MEK inhibition in NS. Instead of broadly suppressing MAPK output, it acts to re-establish a natural regulatory equilibrium. In a developmental disorder, normal physiology requires carefully tuned pathway activity rather than its complete suppression. 

Hurdles remain for 14-3-3/CRAF molecular glues in NS. First, we do not yet know whether complete or only partial rescue of the 14-3-3/CRAF interaction is needed for clinical benefit in NS. Second, the compounds used here target the 14-3-3σ isoform specifically. The relative expression levels of 14-3-3 isoforms across tissues most affected in NS will matter for translational progress. Finally, the long-term consequences of modulating MAPK pathway homeostasis in the context of heterozygous, autosomal dominant NS genetics remain to be studied.

These open questions point toward a rich space for future work, including developing next-generation non-covalent or pan-isoform 14-3-3/CRAF stabilizers, deeper characterization of mutant-dependent binding kinetics, and evaluation of these compounds in physiologically relevant NS model systems. The 14-3-3/CRAF regulatory axis has long been understood mechanistically but rarely targeted pharmacologically. These studies begin to pave that road.

If you're interested in participating in the dialogue, consider joining RASLab. RAS Lab is an online discussion forum to facilitate the exchange of scientific information among researchers in cancers related to RAS. If you would like to join RAS Lab, send an email to SolveRAS@nih.gov with the subject line "I would like to join RAS Lab."

References

(1) Konstantinidou, M.; Vickery, H. R.; Pennings, M. A. M.; Virta, J. M.; Luo, S. Y.; Visser, E. J.; Bannier, S. D.; Srikanth, M.; Cismoski, S. Z.; Young, L. C.; van den Oetelaar, M. C. M.; McCormick, F.; Ottmann, C.; Brunsveld, L.; Arkin, M. R. Modulation of the 14-3-3σ/C-RAF “Auto”Inhibited Complex by Molecular Glues. J. Am. Chem. Soc.2026, 148 (5), 4951–4965. https://doi.org/10.1021/jacs.5c12622

(2) Virta, J. M.; Vickery, H. R.; Konstantinidou, M.; Crawford, M. C.; Pennings, M. A. M.; Ottmann, C.; Brunsveld, L.; Arkin, M. R. Restoring the 14-3-3/CRAF Regulatory Interaction in Noonan Syndrome Using Molecular Glues. Proc. Natl. Acad. Sci. U.S.A. 2026, 123 (18), e2602101123. https://doi.org/10.1073/pnas.2602101123