Cryptic pocket formation underlies allosteric modulator selectivity at muscarinic GPCRs

Scott A. Hollingsworth, Brendan Kelly, Céline Valant, Jordan Arthur Michaelis, Olivia Mastromihalis, Geoff Thompson, Aj Venkatakrishnan, Samuel Hertig, Peter J. Scammells, Patrick M. Sexton, Christian C. Felder, Arthur Christopoulos, Ron O. DrorNature Communications 2019 · DOI

TL;DR

This paper reveals that the high selectivity of certain drugs targeting muscarinic acetylcholine receptors (mAChRs) isn't due to static differences in binding sites, but rather the dynamic opening of a "cryptic pocket" in the M1 subtype. This dynamic pocket, not visible in standard crystal structures, allows M1-selective drugs to bind effectively, offering a new paradigm for designing highly specific therapeutics.

What Problem Does This Paper Address?

Developing drugs that selectively target one receptor subtype while sparing others is a major challenge, especially for G-protein-coupled receptors (GPCRs), which are involved in numerous physiological processes. This is particularly true for the muscarinic acetylcholine receptor (mAChR) family (M1-M5), where the orthosteric binding sites (where the natural ligand binds) are nearly identical across subtypes, making selective drug design difficult. While allosteric modulators (drugs that bind to a different site than the natural ligand) have shown promise for selectivity, recent structural data revealed that even these allosteric binding sites are surprisingly similar across mAChR subtypes. This raised a critical question: how do some allosteric modulators achieve such exquisite subtype selectivity when the static structures appear so alike?

Methods

The researchers employed a multi-pronged approach combining computational simulations and experimental validation:

  • Molecular Dynamics (MD) Simulations: They ran extensive microsecond-timescale MD simulations of the M1, M2, M3, and M4 mAChRs under various conditions (active/inactive states, with/without orthosteric ligands). This allowed them to observe the dynamic behavior of the receptors and identify transiently formed pockets.
  • Cryptic Pocket Identification: Through MD simulations, they identified a "cryptic pocket" in the M1 mAChR's allosteric site that is not present in static crystal structures. They characterized the specific residue movements (Y2.64 and E7.36) that lead to its formation and stabilization.
  • Ligand Docking and Pose Prediction: They used the identified cryptic pocket to predict binding poses for M1-selective allosteric modulators, reconciling previously contradictory experimental data.
  • Site-Directed Mutagenesis: To experimentally validate their findings, they designed mutations in the M1 mAChR to disrupt the formation of the cryptic pocket.
  • Radioligand Binding Assays: They measured the binding affinities of various allosteric modulators to wild-type and mutant M1 mAChRs to assess the impact of cryptic pocket disruption.

Datasets

The study utilized existing crystal structures of mAChR subtypes (M1, M2, M3, M4) as starting points for their simulations. No new benchmark datasets were introduced; rather, the focus was on analyzing the dynamic behavior of known receptor structures.

Key Insights and Ideas

  • Dynamic Pockets Drive Selectivity: The core insight is that receptor selectivity for allosteric modulators can arise not from static structural differences, but from the dynamic formation of transient, "cryptic" pockets.
  • The M1 mAChR's Unique Dynamic Behavior: The M1 mAChR subtype exhibits a unique propensity to form a cryptic pocket in its allosteric binding site, which is largely absent in other mAChR subtypes, even those with highly similar static structures.
  • Explaining Non-Planar Ligand Binding: The cryptic pocket provides a space for the non-planar arms of M1-selective positive allosteric modulators (PAMs) like BQZ12 to dock, explaining why these ligands are non-planar and how they achieve high affinity.
  • Reconciling Conflicting Data: This cryptic pocket mechanism elegantly explains previously puzzling mutagenesis data that suggested specific residues were important for M1-selective PAM binding, but which couldn't be accommodated by static crystal structures.
  • A New Design Strategy: The findings suggest a powerful strategy for designing subtype-specific drugs: exploit cryptic pockets that open in the target receptor but remain closed in off-target receptors.

Key Results

  • MD simulations showed that the cryptic pocket in the M1 mAChR opens dynamically in over 50% of simulation frames across various conditions, while it opened in less than 5% of frames for M2, M3, and M4 mAChRs.
  • The formation of the cryptic pocket in M1 mAChR was attributed to specific residue movements (Y2.64 and E7.36) that create a new binding cavity.
  • Mutagenesis experiments confirmed the importance of the cryptic pocket: mutations designed to prevent its opening significantly reduced the binding affinity of M1-selective PAMs (BQZ12 and BQCA) by 10- to 100-fold.
  • In contrast, the affinity of a non-M1-selective PAM (LY2033298) remained largely unaffected by these mutations, supporting the conclusion that the reduced affinity was due to disrupted cryptic pocket formation.
  • A modified M1-selective PAM (MIPS-1519), lacking a key feature for pi-stacking interactions within the cryptic pocket, showed reduced affinity and still demonstrated sensitivity to cryptic pocket-disrupting mutations, further validating the model.

Limitations and Future Work

  • Limited Subtype Simulation: The study did not include simulations for the M5 mAChR due to the lack of an available experimental structure.
  • Generalizability: While the authors suggest the strategy is broadly applicable, further work is needed to determine how conformational dynamics and cryptic pocket formation can be leveraged for selectivity in other receptor families.
  • Future Directions: The paper strongly advocates for exploiting cryptic pocket formation as a key strategy in future therapeutic drug design, particularly for GPCRs and other protein superfamilies. The increasing accessibility of computational resources is expected to facilitate such approaches.

Summary updated

Abstract

Allosteric modulators are highly desirable as drugs, particularly for G-protein-coupled receptor (GPCR) targets, because allosteric drugs can achieve selectivity between closely related receptors. The mechanisms by which allosteric modulators achieve selectivity remain elusive, however, particularly given recent structures that reveal similar allosteric binding sites across receptors. Here we show that positive allosteric modulators (PAMs) of the M1 muscarinic acetylcholine receptor (mAChR) achieve exquisite selectivity by occupying a dynamic pocket absent in existing crystal structures. This cryptic pocket forms far more frequently in molecular dynamics simulations of the M1 mAChR than in those of other mAChRs. These observations reconcile mutagenesis data that previously appeared contradictory. Further mutagenesis experiments validate our prediction that preventing cryptic pocket opening decreases the affinity of M1-selective PAMs. Our findings suggest opportunities for the design of subtype-specific drugs exploiting cryptic pockets that open in certain receptors but not in other receptors with nearly identical static structures.
BibTeX
@inproceedings{hollingsworth2019cryptic,
  author = {Scott A. Hollingsworth and Brendan Kelly and Céline Valant and Jordan Arthur Michaelis and Olivia Mastromihalis and Geoff Thompson and AJ Venkatakrishnan and Samuel Hertig and Peter J. Scammells and Patrick M. Sexton and Christian C. Felder and Arthur Christopoulos and Ron O. Dror},
  title = {{Cryptic pocket formation underlies allosteric modulator selectivity at muscarinic GPCRs}},
  booktitle = {Nature Communications},
  year = {2019},
  doi = {10.1038/s41467-019-11062-7},
}
Figure