Molecular Glue refers to small molecules, typically smaller than PROTACs, often with improved physicochemical properties, designed to stabilize interactions between two proteins. In most cases, this approach is used to enhance or induce interactions between a target and an E3 ligase, but it has also been shown to stabilize interactions to increase activity or inhibit other interactions that compete with natural effector molecule binding. This article will describe the induced proximity effects and discuss the current understanding of the mechanisms by which molecular glue functions.
What is Molecular Glue?
Interactions between molecules are essential components of controlling biological processes, offering attractive therapeutic opportunities for modulating interactions between target proteins and other macromolecules. Molecular glue, as small molecules, induces proximity between target proteins and effector macromolecules, thereby generating new interactions or stabilizing existing ones, leading to functional changes in the target protein. Therefore, molecular glue has the potential to alter a range of biological processes, including transcription, translation, protein folding, and degradation. The concept of molecular glue is not novel; the term was used in the late 1980s to refer to small molecules or proteins that induce fusion of protoplasts or allow platelet aggregation. The term was also used for immunosuppressants such as cyclosporin A and FK506, describing their ability to induce proximity between the shared targets calcineurin, cyclophilin, and FKBP. Following the success of PROTACs in inducing proximity between target proteins and E3 ligases for protein degradation, there has been recent effort to expand this approach to identify molecules capable of inducing protein proximity without the need for specific warheads and linkers. So far, most identified glues can stabilize interactions between target proteins and E3 ligases, resulting in degradation. This mechanism is mainly achieved by enhancing pre-existing weak interactions, although it is possible to induce the formation of ternary complexes by binding to a protein partner of a binary complex. When molecular glue enhances pre-existing weak interactions or induces new interactions between molecules that do not typically interact, they may not necessarily require binding pockets on individual target proteins but can improve complementary surfaces between interacting proteins. Therefore, molecular glue holds exciting prospects in improving the druggability of previously challenging targets. The challenge lies in identifying molecular glue, as many known compounds were not identified through rational screening but rather through serendipitous discoveries following traditional phenotypic screening methods. However, current focus is shifting towards approaches allowing rational methods, including diversity screening and rational design, leveraging a range of binding partners, which will enhance our ability to exploit this novel mechanism. Here, we will focus on opportunities for inducing protein proximity and discuss current and future cell-based and cell-free screening methods, highlighting the scope of utilizing molecular glue in drug discovery.
Common Molecular Glue at BOC Sciences
| BP-900057 | Auxin | 87-51-4 |
| BP-900058 | (R)-CR8 | 1786438-30-9 |
| BP-900059 | E 7820 | 289483-69-8 |
| BP-900063 | Thalidomide | 50-35-1 |
| BP-900064 | Cytochalasin J | 56144-22-0 |
| BP-900065 | KB02-JQ1 | 2384184-44-3 |
| BP-900066 | KB02-SLF | 2384184-40-9 |
| BP-900068 | Asukamycin | 61116-33-4 |
| BP-900069 | BI-3802 | 2166387-65-9 |
| BP-900070 | CCT369260 | 2253878-44-1 |
| BP-900071 | Chloroquinoxaline sulfonamide | 97919-22-7 |
| BP-900073 | FPFT-2216 | 2367619-87-0 |
| BP-900074 | HQ461 | 1226443-41-9 |
| BP-900075 | NRX-252114 | 2763260-39-3 |
| BP-900076 | NRX-252262 | 2438637-61-5 |
| BP-900078 | Tasisulam | 519055-62-0 |
| BP-900079 | TMX-4100 | 2367619-63-2 |
| BP-900080 | TMX-4113 | |
| BP-900081 | TMX-4116 | 2766385-56-0 |
| BP-900060 | Indisulam | 165668-41-7 |
Inducing Protein Proximity
Biochemical processes occurring between molecules within cells are controlled by the degree of physical proximity between two interacting species. Interactions between molecules occur only when they are in close proximity to each other to generate the desired biological effect. For many years, the focus of drug discovery has been on finding molecules that disrupt protein-protein interactions (PPIs) or inhibit/antagonize protein-ligand interactions to interrupt interactions between proteins. In the early 1990s, it was recognized that artificial small molecules (molecular glues) could be used to induce protein proximity, alongside the role of avidin and the growing understanding of the function of FK506 in forming the FKBP12-FK506-calcineurin complex, as well as the recognition that natural processes of molecular glue could also be targeted, such as the SH2 domain of tyrosine kinases promoting signal transduction by binding to phosphotyrosine without catalysis.
The action of molecular glue involves inducing or enhancing interactions between two proteins, bridging surfaces to achieve complementarity, thereby forming ternary complexes. It has been suggested that molecular glue may primarily act by binding to only one binding partner or by binding to binary complexes formed between interacting proteins (Figure 1). When both binding partners have a certain level of intrinsic affinity, molecular glue further stabilizes this interaction, leading to what is described as chemical stabilization. This hypothesis distinguishes molecular glue from PROTACs, which are bifunctional molecules with affinity for both protein partners. Thus, molecular glue does not encounter the hook effect (a common problem with PROTACs), where at higher concentrations, both proteins are complexed with the PROTAC, preventing the formation of the ternary complex, known as the hook effect. Molecular glue does not undergo this effect and does not exhibit saturable binding behavior. The rational design of PROTACs initially focuses on identifying ligands for both the target and effector proteins, typically using traditional hit discovery methods, and then optimizing linker between these two warheads to enhance affinity and function. In contrast, the discovery and design of molecular glue require new screening strategies to induce proximity. Most molecular glues are expected to interact between protein molecules (intermolecular), but interactions within individual proteins (intramolecular) have also been described, exhibiting glue-like behavior. In this case, different domains of the protein become very close, altering function. An example is the allosteric SHP2 inhibitor SHP099. The functional outcomes of inducing protein proximity may vary, resulting in stabilization, inhibition, activation, or even degradation. In the following subsections, we describe the effects of induced proximity in several contexts.
Protein Stabilization
Molecular glue can bind to existing protein-protein complexes and enhance the affinity of their interactions. This could offer significant advantages for molecular glue, as it may not need as strict affinity compared to inhibitory compounds that may have to compete with natural binding partners to achieve the desired biological modulation. Therefore, enhancing pre-existing interactions is an attractive strategy. Additionally, molecular glue can bind to transient and specific intermolecular surfaces or pockets formed between interacting proteins. Ternary complexes are formed only when the relevant binding partners are present. This may provide greater selectivity and help reduce off-target effects. Indeed, potential pockets at protein-protein interfaces have been studied, mechanisms of pocket formation proposed, and some attempts have been made to describe the chemical conditions that make good stabilizers. These studies suggest that pockets formed at protein interfaces may bear similarity to more traditional small molecule binding sites, which increases the assumption of chemical promiscuity within existing small molecule libraries, offering initial hits for molecular glue optimization. The development of trametiglue is a good example; it effectively stabilizes the interaction between KSR and MEK1. Here, trametinib, a derivative of the MEK1 inhibitor trametinib, was designed to combine the potency and dissociation kinetics of trametinib with the functional ability of another MEK inhibitor, CH5126766, to capture the inactive state of RAF-bound MEK.
Enzyme Inhibition
Perhaps the most famous example of molecular glue inhibition is the inhibition of calcineurin (CN) by cyclosporin A (CsA)/cyclophilin A (CYPA). CsA induces its biological effects-immunosuppressive activity – by forming an initial complex with CYPA, which then binds to CN and inhibits its phosphatase activity. Competitive inhibition of CN, which is Ca2+-dependent, prevents the dephosphorylation of the cytoplasmic component of the transcription factor NF-AT (nuclear factor of activated T-cells), thereby preventing NF-AT translocation to the nucleus and blocking the transcription of early growth factors. It has been shown that CsA or CYPA alone cannot inhibit CN. Meanwhile, another natural product and its binding complex (FK506-FKBP12) have also been shown to inhibit CN. FKBP12 employs a molecular glue mechanism to inhibit the protein kinase mTOR. The binding of FKBP12 to mTOR is induced by the macrolide rapamycin. Subsequently, a synthetic FKBP12-mTOR molecular glue was discovered through FKBP-centered, target-agnostic library screening, demonstrating the induced proximity between FKBP12 and the FRB domain of mTOR. Thus, besides protein stabilization, enzyme function or inhibition of downstream signaling is also an important molecular glue mechanism.
Activation
Molecular glue may also be used to activate target proteins. For example, asukamycin has been shown to bind to the E3 ligase UBR7, covalently modifying the Cys374 site and activating the new substrate TP53. This increases the transcriptional activity of the tumor suppressor in a UBR7-dependent manner. This activation leads to inhibition of cancer cell growth in 250 cancer cell lines, with IC50 values ranging from 5 to 30 µM. Another molecular glue can activate p53 through a different mechanism. RO-2443, identified through small molecule screening, can induce MDMX dimerization. This can prevent MDMX from binding to TP53 and alleviate the negative regulation of the tumor suppressor, potentially providing effective treatment for cancers overexpressing MDMX by mediating the activation of TP53 apoptotic activity. Thus, molecular glue can activate proteins by either gluing regulatory proteins together or binding them to other binding partners to eliminate negative regulation.
Degradation
Protein proximity-induced degradation of target proteins has emerged as a promising strategy for selective protein degradation and is currently the most abundant category of functional molecular glues. Molecular glues that induce proximity between target proteins and E3 ligases have been identified through serendipitous discovery, rational design, data mining, or traditional screening methods. The close proximity between the target and E3 ligases leads to ubiquitination of the target. The ubiquitination process involves four different types of enzymes: E1-E4 ligases. When ubiquitin is covalently linked to E1 (ubiquitin-activating enzyme) and then transferred to E2 (ubiquitin-conjugating enzyme). Subsequently, E3 (ubiquitin-protein ligase) is responsible for transferring ubiquitin from E2 to the target protein. Once the first ubiquitin is attached (monoubiquitination), E3 can form longer ubiquitin chains by creating ubiquitin-ubiquitin peptide bonds. Ubiquitin has seven lysine residues (K6, K11, K27, K29, K33, K48, and K63) available for chain extension. E4 (chain elongation factors) is a subclass of E3 enzymes that also catalyzes this process. Subsequently, polyubiquitinated proteins are recognized and degraded by the proteasome. Molecular glues typically work by directly modulating protein-protein interaction surfaces to introduce or enhance the affinity of interactions between the target protein and E3 ligase. Examples include enhancing the interaction between cereblon (CRBN) and degradation targets with IMiDs. The structures of compounds binding to the DDB1-CRBN E3 ligase complex were resolved several years ago, and the first molecular glue demonstrated to have this function was thalidomide, which was first marketed under the trade name Contergan in 1957 for the treatment of morning sickness but caused severe birth defects. Thalidomide was shown to interact with the E3 ligase CRBN, leading to the ubiquitination and subsequent degradation of many new substrates. Thalidomide and its derivatives induce cereblon substrates, including Ikaros, SALL4, and CK1α.
The first rational approach to enhance protein degradation using molecular glues is through cell death-phenotypic screening of active compounds, followed by neddylation (ubiquitination-like) cell-coupled multi-omics to identify targets. The method of comparing low ubiquitinated cells with near-ubiquitinated intact cells stems from the fact that ubiquitin ligase activity requires modification of cullin proteins by a ubiquitin-like protein called NEDD8. Ubiquitin-like modification leads to conformational rearrangements within cullin ring ligases, which are necessary for the transfer of ubiquitin to substrates. This led to the identification of several compounds that cause instability in cell cycle proteins K. In targeted protein degradation strategies, four similar compounds were screened using fluorescence anisotropy (FA) (or fluorescence polarization (FP)) and the pSer33/Ser37 peptide, which is part of the β-catenin phosphorylation degradation sequence (DpSGφXpS) bound by β-TrCP. This approach identified four similar compounds, including NRX1532, which demonstrated a 10-fold synergy between β-catenin and β-TrCP. Structural studies of the more soluble analogue NRX-1933 showed that the compound binds at the β-catenin:β-TrCP interface with minimal impact on the conformation of individual β-catenin or β-TrCP.
Regardless of how potential molecular glue degraders are identified, it is crucial to analyze these compounds in assays aimed at monitoring target protein degradation. Fortunately, due to the growing importance of protein degradation in drug discovery in recent years, there is a range of potential and highly sensitive methods for monitoring ubiquitination and degradation that can be applied to molecular glue discovery.
Reference:
Holdgate, Geoffrey A., et al. “Screening for molecular glues–challenges and opportunities.” SLAS Discovery (2023).