Stabilization of Native Protein–Protein Interactions with Molecular Glues: A 14-3-3 Case Study

Protein–protein interactions (PPIs) play a key role in homeostasis and are often dysregulated in disease. PPIs were traditionally considered “undruggable” due to their flat surfaces and disordered domains. Recently, the identification of PPI stabilizers, or molecular glues (MGs), compounds that bind cooperatively to PPI interfaces, has provided a new direction for the field. MGs offer exciting opportunities for chemical biology and drug discovery, particularly for intrinsically disordered domains. To date, many of the fascinating MGs were discovered serendipitously, and their molecular glue mechanism of action was understood retrospectively. Our collaborative contribution has been the development of systematic, rational approaches for the identification, optimization, and validation of MGs.

This Account focuses on the modulation of the native PPIs between the hub protein 14-3-3 and its client proteins. 14-3-3 recognizes specific phospho-serine/threonine motifs on disordered domains of hundreds of clients and, depending on the phospho site, can activate or inhibit signaling pathways. Until recently, only the natural product fusicoccin A and its analogs were known to bind at the structured 14-3-3/client interfaces and modulate cellular pathways. The complexity of the natural products significantly hindered chemical biology approaches and did not provide sufficient insight into the systematic, selective targeting of the client of interest.

Inspired by the natural products, we used fragment-based screens to identify new chemical matter for 14-3-3/client PPIs. Using disulfide-tethering technology, we targeted either engineered cysteines on 14-3-3 or the native cysteine (C38) on 14-3-3σ. Five clients (ERα, C-RAF, FOXO1, USP8, and SOS1), representing varying sequences, binding modes, and physiological roles, were included in the initial screens. We identified both selective and nonselective fragments suitable for medicinal chemistry optimization.

Starting from a fragment that stabilized two 14-3-3 clients, estrogen receptor α (ERα) and C-RAF, we developed cell-active MGs selective for ERα. ERα is a well-validated target in breast cancer, and 14-3-3 is a negative regulator that blocks ERα transcriptional activity. We used structure-guided design to optimize ligand–protein interactions at the composite PPI surface. The molecular glues were validated in biophysical assays, including intact mass spectrometry (MS) and fluorescence anisotropy (FA) assays, allowing the quantification of binding, kinetics, and cooperativity.

We explored alternative strategies for the identification and optimization of MGs. For the 14-3-3/ERα complex, we demonstrated fragment linking to generate non-covalent stabilizers and a scaffold-hopping approach using multicomponent reaction chemistry. For the 14-3-3/C-RAF complex, we used a fragment-merging approach to selectively stabilize the inhibited state of C-RAF. Binding of 14-3-3 to the inhibitory phospho-S259 site prevents C-RAF dimerization and activation, offering an alternative mechanism to block the MAPK pathway. Finally, we validated compounds in cells using pathway-specific assays and a series of proximity-based NanoBRET assays to measure cellular PPIs.

These approaches led to first-in-class MGs for the 14-3-3/ERα and 14-3-3/C-RAF targets. Overall, we have developed a systematic platform for the identification of molecular glues for native PPIs applicable to the broad 14-3-3 interactome and beyond.