| Abstract: | Directly targeting oncogenic V-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (K-Ras) with small-molecule inhibitors has historically been considered prohibitively challenging. Recent reports of compounds that bind directly to the K-Ras G12C mutant suggest avenues to overcome key obstacles that stand in the way of developing such compounds. We aim to target the guanine nucleotide (GN)-binding pocket because the natural contents of this pocket dictate the signaling state of K-Ras. Here, we characterize the irreversible inhibitor SML-8-73-1 (SML), which targets the GN-binding pocket of K-Ras G12C. We report a high-resolution X-ray crystal structure of G12C K-Ras bound to SML, revealing that the compound binds in a manner similar to GDP, forming a covalent linkage with Cys-12. The resulting conformation renders K-Ras in the open, inactive conformation, which is not predicted to associate productively with or activate downstream effectors. Conservation analysis of the Ras family GN-binding pocket reveals variability in the side chains surrounding the active site and adjacent regions, especially in the switch I region. This variability may enable building specificity into new iterations of Ras and other GTPase inhibitors. High-resolution in situ chemical proteomic profiling of SML confirms that SML effectively discriminates between K-Ras G12C and other cellular GTP-binding proteins. A biochemical assay provides additional evidence that SML is able to compete with millimolar concentrations of GTP and GDP for the GN-binding site.Multiple lines of evidence suggest that if achievable, inhibiting V-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (K-Ras) signaling may have therapeutic advantages in cancer. Approximately 30% of all human cancers contain activating Ras mutations, making them one of the most common identifiable molecular drivers of cancer (1, 2). K-Ras mutation–positive tumors tend to be less responsive to current therapy than other biological subtypes, and patients with these tumors have worse cancer outcomes (3, 4). We aim to develop and evaluate direct-acting inhibitors of K-Ras as a therapeutic strategy.Ras proteins are GTPase enzymes that transduce extracellular signals when growth factors bind to extracellular receptors, resulting in cellular responses such as proliferation, apoptosis, and differentiation (5). Activating signals are transmitted when Ras is bound to GTP and cease once the bound GTP is hydrolyzed to GDP. Two structurally dynamic loops, so-called “switch I” (residues 25–40 in K-Ras) and “switch II” (residues 57–75 in K-Ras) form a key portion of the binding interface between K-Ras and a number of regulators and downstream effectors, including Raf kinases, PI3 kinases, and RalGDS (6–8). We hypothesize that to be successful, direct-acting K-Ras inhibitors must alter the conformation of switch I and/or switch II such that it becomes incapable of transmitting activating signals. Because the guanine nucleotide (GN)-binding pocket dictates the switch conformation, we postulate that developing compounds binding to this region will have a high likelihood of modulating K-Ras signaling.Development of GN competitive inhibitors of K-Ras is challenging because GTP and GDP bind with subnanomolar affinity and intracellular concentrations of GTP and GDP are high. Recently, our group (9) and Shokat and coworkers (10) independently developed and reported two classes of compounds that have a direct impact on productive nucleotide binding to the GN-binding site. Both target a common activating mutation, G12C, to achieve irreversible binding. K-Ras G12C is present in an estimated 10–20% of Ras-driven cancers and in roughly 50% of Ras-mutated lung adenocarcinomas (11–13). For lung cancer alone, this means that therapeutics targeting the G12C mutation could treat roughly 25,000 people per year in the United States (14). Mutations at codon 12, along with the other most common cancer-causing Ras mutations at codons 13 and 61, decrease intrinsic GTPase activity to some extent and impair interactions with GTPase-activating proteins that modulate (GTP) hydrolysis. Codon 12 is adjacent to the active site, such that the mutation places a solvent-accessible cysteine near the GN terminal phosphate.We previously reported development of SML-8-73-1 (SML), a GDP analog containing an electrophilic warhead extending from the beta-phosphate that undergoes a Michael reaction addition to Cys-12, forming a stable thioether linkage (9). Even in the presence of large excesses of GDP and GTP, quantitative complete irreversible binding was observed by MS. In biochemical assays, SML prevents K-Ras association with the Ras-binding domain of the downstream effector BRaf. Preliminary cellular tests using a cell-permeable caged prodrug version, SML-10-70-1, demonstrated that treatment of a G12C mutant K-Ras cancer cell line with the SML class of compounds, albeit at a high concentration (100 μM), leads to inactivation of K-Ras and down-regulation of Akt and Erk signaling pathways, demonstrating as a proof of concept that the GN-binding pocket is a viable target for inhibitors of Ras signaling.Here, we provide three lines of evidence to support further the concept of targeting the GN-binding pocket of K-Ras. First, we report high-resolution X-ray crystal structures of K-Ras, including a structure containing the irreversible inhibitor, SML, bound to K-Ras G12C. The models are analyzed with the aim of understanding implications for K-Ras interactions with downstream effectors. Second, we address compound selectivity with MS-based in situ profiling demonstrating that SML preferentially interacts with K-Ras G12C over most other cellular GTP-binding proteins. Finally, we provide additional evidence that in a purified system, SML is able to compete for the GN-binding site of K-Ras G12C in the presence of millimolar concentrations of GTP and GDP, similar to those found in a living cell. The prospects for developing GTPase inhibitors are also broadly considered. |
