Inside the cracked kernel: establishing the molecular basis of AMG510 and MRTX849 in destabilising KRASG12C mutant switch I and II in cancer treatment

Abstract

The Kirsten rat sarcoma oncoprotein (KRAS) has been punctuated by drug development failures for decades due to frequent mutations that occur mostly at codon 12 and the seemingly intractable targeting of the protein. However, with advances in covalent targeting, the oncoprotein is being expunged from the ‘undruggable’ list of proteins. This feat has seen some covalent drugs at different stages of clinical trials. The advancement of AMG510 and MRTX849 as inhibitors of cysteine mutated KRAS (KRAS^G12C) to phase-III clinical trials informed the biased selection of AMG510 and MRTX849 for this study. Despite this advance, the molecular and atomistic modus operandi of these drugs is yet to come to light. In this study, we employed computational tools to unravel the atomistic interactions and subsequent conformational effects of AMG510 and MRTX849 on the mutant KRAS^G12C. It was revealed that AMG510 and MRTX849 complexes presented similar total free binding energies, (ΔG_bind), of –88.15 ± 5.96 kcal/mol and –88.71 ± 7.70 kcal/mol, respectively. Gly10, Lys16, Thr58, Gly60, Glu62, Glu63, Arg68, Asp69, Met72, His95, Tyr96, Gln99, Arg102 and Val103 interacted prominently with AMG510 and MRTX849. These residues interacted with the pharmacophoric moieties of AMG510 and MRTX849 via hydrogen bonds with decreasing bond lengths at various stages of the simulation. These interactions together with pi-pi stacking, pi-sigma and pi-alkyl interactions induced unfolding of switch I whiles compacting switch II, which could interrupt the binding of effector proteins to these interfaces. These insights present useful atomistic perspectives into the success of AMG510 and MRTX849 which could guide the design of more selective and potent KRAS inhibitors.

Communicated by Ramaswamy H. Sarma

Keywords:

• KRASG12C
• MRTX849
• AMG510
• switch I
• switch II
• covalent binding
• molecular dynamics simulation

Acknowledgements

The authors acknowledge, the School of Health Science, University of KwaZulu-Natal, Westville campus for financial assistance. The Centre of High-Performance Computing (CHPC, www.chpc.ac.za), Cape Town, South Africa for computational resources.

Disclosure statement

Authors declare no financial and intellectual conflict of interests.

Funding

The author(s) reported there is no funding associated with the work featured in this article.