Inhibition of Nonfunctional Ras

Abstract

Intuitively, functional states should be targeted; not nonfunctional ones. So why could drugging the inactive K-Ras4B^G12Cwork-but drugging the inactive kinase will likely not? The reason is the distinct oncogenic mechanisms. Kinase driver mutations work by stabilizing the active state and/or destabilizing the inactive state. Either way, oncogenic kinases are mostly in the active state. Ras driver mutations work by quelling its deactivation mechanisms, GTP hydrolysis, and nucleotide exchange. Covalent inhibitors that bind to the inactive GDP-bound K-Ras4B^G12C conformation can thus work. By contrast, in kinases, allosteric inhibitors work by altering the active-site conformation to favor orthosteric drugs. From the translational standpoint this distinction is vital: it expedites effective pharmaceutical development and extends the drug classification based on the mechanism of action. Collectively, here we postulate that drug action relates to blocking the mechanism of activation, not to whether the protein is in the active or inactive state.

Keywords: K-Ras4B; KRAS; dimer; drug discovery; energy landscape; inhibitor; kinases.

Figure 1.

The free energy landscape ΔΔG of a kinase conformation (here EGFR), the effect of mutations and an allosteric inhibitor. Oncogenic driver mutations (L858R and T790M), destabilize the inactive state (left plot in the middle panel), stabilize the active state (center plot) or both (right plot). The outcome is a shift of the ensemble toward the constitutively active EGFR conformation. In the bottom panel, an allosteric drug binding EGFR’s mutant surface stabilizes this conformation, with a population shift toward this conformation. The drug modulates the active site shape.

Figure 2.

Examples of Ras inhibitors. The first row shows crystal structures of the GDP-bound H-Ras in complex with NS1 monobody (PDB: 5E95), farnesyltransferase (FTase) containing the α and β subunits in complex with farnesyl diphosphate (FPP) and the inhibitor Tipifarnib (R115777) (PDB 1SA4), and phosphodiesterase δ subunit (PDEδ) inhibited by Deltasonamide (PDB: 5ML3). Covalent drugs to the G12C mutation are shown in the second row; crystal structures of the GDP-bound K-Ras4BG12C with MRTX849 (PDB: 6UT0), the GDP-bound K-Ras4BG12C with ARS-1620 (PDB: 5V9U), and K-Ras4BG12C with the covalent GTP-competitive inhibitor SML-8-73-1 (PDB: 5KYK). The third row shows crystal structures of active Ras inhibited by non-covalent, small molecular drugs for the GCP-bound K-Ras4BG12D with DCAI (PDB: 4DST), the GNP-bound K-Ras4BQ61H with Abd-7 (PDB: 6FA4), and the GNP-bound K-Ras4BQ61H with 3344 (Abd-8) (PDB: 6F76).

Figure 3.

Ras inhibitor with nanomolar affinity. Crystal structure of the GCP-bound K-Ras4B^G12D with BI-2852 (PDB: 6GJ8) (upper left panel). The compound 1 drug, BI-2852, exhibits nanomolar affinity to a pocket between Switch I and II. The BI-2852-induced Ras dimerization (right panel). Symmetric rotation of the K-Ras4B crystal structure reveals a β-sandwich dimer aligning its β1, β2, and γ3 strands by two BI-2852 molecules. The first BI-2852 (ligand-1) binds the Switch I and II pocket, the deep pocket formed by Lys5, Val7, Asp54, and Leu56 on the first Ras (Ras-1). The second BI-2852 (ligand-2), a copy of the first, binds the shallow pocket formed by Glu3, Leu52, and Asp54 on Ras-1. The complex is stabilized by several salt-bridges. The molecular dynamics (MD) simulations of Ras molecules at the anionic membrane (Jang et al., 2016b) previously reported the unfunctional β-sandwich dimer for K-Ras4B-GTP and H-Ras-GTP (lower left panel).