Biophysical and structural analysis of KRAS switch-II pocket inhibitors reveals allele-specific binding constraints.

RAS mutations are observed in 20% of all cancers, with the KRAS isoform highly mutated in colorectal, lung and pancreatic cancers. The last several years have seen the development of clinical compounds that target KRAS G12C mutations, with other compounds under clinical development. In this study, a series of KRAS small-molecule inhibitors were compared for their binding affinity against a panel of KRAS mutant alleles. These inhibitors either covalently target the G12C mutant or reversibly target other mutants by binding in a transient pocket known as the switch-II pocket. Covalent inhibitors bound KRAS-GDP with K_D values ranging from 10^-9-10^-3 M, whereas reversible inhibitors bound in the low nM range. A loss of affinity was observed for KRAS-GppNHp, due in part to rearrangements in switch-II, where the hydrogen bond between G60 and the γ-phosphate needs to break to form the switch-II pocket. Interestingly, these inhibitors had reduced affinity to KRAS Q61R-GppNHp, but not to WT and other mutants. The crystal structure of KRAS Q61R-GppNHp reported here revealed that access to the switch-II pocket was restricted due to R61 forming an additional hydrogen bond with the backbone carbonyl of T35 in switch-I. The restricted access to the switch-II pocket caused a decrease in the association rate of inhibitor binding and resulted in a loss of affinity. These findings across KRAS mutants provide valuable insights into the conformational adaptability of the switch-II pocket and may prove useful in developing the next generation of allele-specific and pan-KRAS small molecule inhibitors.