A Transferable and Robust Computational Framework for Class A GPCR Activation Free Energies.

The activation of G-protein coupled receptors is involved in many biomedically important cellular pathways. However, capturing it with molecular simulations is far from trivial, as it requires capturing both local and global motions. We recently achieved this goal in a specific receptor (the β1-adrenergic receptor, or ADRB1) by combining a multiple replica enhanced sampling approach with tailored collective variables. While that approach can be applied to other receptors, it would require a tedious and error-prone choice and refinement of the collective variables and, in particular, of the main path-like variable. Herein, we introduce an effective and streamlined evolved strategy for defining CVs that reduces user intervention while still achieving a robust free energy convergence. We apply it to two apo-GPCRs of pharmacological relevance, ADRB1 and the μ-opioid receptor. In the first case, we show that the reconstructed free energies agree with those obtained with the previous tailored approach, while for the μ-opioid receptor activation, we gain novel biological insights. The proposed method can be easily applied to other class A GPCRs, paving the way for the systematic elucidation of the activation mechanisms of many crucial drug targets.