Toward an understanding of the structural basis of allostery in muscarinic acetylcholine receptors

Abstract

Recent breakthroughs and developments in structural biology have led to a spate of crystal structures for G protein-coupled receptors (GPCRs). This is the case for the muscarinic acetylcholine receptors (mAChRs) where inactive-state structures for four of the five subtypes and two active-state structures for one subtype are available. These mAChR crystal structures have provided new insights into receptor mechanisms, dynamics, and allosteric modulation. This is highly relevant to the mAChRs given that these receptors are an exemplar model system for the study of GPCR allostery. Allosteric mechanisms of the mAChRs are predominantly consistent with a two-state model, albeit with some notable recent exceptions. Herein, we discuss the mechanisms for positive and negative allosteric modulation at the mAChRs and compare and contrast these to evidence offered by pharmacological, biochemical, and computational approaches. This analysis provides insight into the fundamental pharmacological properties exhibited by GPCR allosteric modulators, such as enhanced subtype selectivity, probe dependence, and biased modulation while highlighting the current challenges that remain. Though complex, enhanced molecular understanding of allosteric mechanisms will have considerable influence on our understanding of GPCR activation and signaling and development of therapeutic interventions.

Figure 1.

The many facets of GPCR allosteric modulation. (A) Structure of the M₂ mAChR highlighting allosteric and orthosteric-binding sites. (B) Allosteric modulators are characterized through three different modes of behavior: PAM, NAM, or NAL. This is represented schematically, where increasing concentrations of different allosteric modulators are titrated against a single concentration of an orthosteric ligand. The saturability of the modulatory effect above a certain concentration of allosteric modulator is indicative of the “ceiling effect,” which is a key molecular characteristic of allosteric drugs. (C) Probe dependence is another unique pharmacological characteristic of allosteric modulators whereby the magnitude and direction of the allosteric effect can change depending on the orthosteric ligand used as a probe of receptor function. Shown is the effect on a fixed concentration of orthosteric ligand 1 or orthosteric ligand 2 at the same receptor in the presence of increasing concentrations of the same allosteric modulator. The binding of orthosteric ligand 1 is increased, whereas the binding of orthosteric ligand 2 is decreased. (D) Allosteric modulators can display selectivity in their ability to only enhance the binding of an orthosteric ligand at one particular subtype relative to other related receptor subtypes. (E) Biased modulation occurs when an allosteric ligand can promote more than one type of active state to the relative exclusion of another, such that the observed effect of the same agonist–modulator pair can vary depending on the signaling pathway that is linked to each receptor conformation. In this example, the modulator increases the potency of an agonist for pathway 1 but decreases the potency of the agonist in pathway 2; the two pathways are thus downstream of two different active receptor states. (F) Allosteric interactions can also arise between binding sites (orthosteric or allosteric) located on individual GPCRs that are arranged in conformationaly linked dimeric or oligomeric arrays.

Figure 2.

Inactive-state structures of the M₁–M₄ mAChRs. (A) Alignment of the inactive-state structures of the M₁–M₄ mAChRs reveals high similarity in their structure. (B) The orthosteric site has high similarity in the positions and sequence of the amino acid residues that interact with the orthosteric ligand. Highlighted in red are the three tyrosine residues that separate the ECV from the orthosteric-binding site through the formation of a tyrosine lid. (C) Residue L/F⁺⁵ in ECL2 is numbered relative to the conserved disulfide bond between ECL2 and TM3. Protein Data Bank codes are indicated in B.

Figure 3.

Comparison of the inactive and active states of the M₂ mAChR. (A) Alignment of the inactive (yellow), active (cyan), and active-state, PAM-bound (pink) M₂ mAChR structures reveal reorganization of TMs 5, 6, and 7 in the active state (indicated by arrows). Residue Y⁺¹ in ECL2 is numbered relative to the conserved disulfide bond between ECL2 and TM3. (B and C) Views of the orthosteric- (B) and ECV-binding (C) sites show contraction mediated by the movements of TM5, TM6, ELC3, and TM7. (D) Cross sections of the receptor shown with the interior surface colored black. The tyrosine lid capping the orthosteric site is highlighted in red and completely seals the orthosteric site in the active state. LY2119620 is not shown for clarity.

Figure 4.

Structural insights into allosteric modulation of mAChRs. (A) Structural basis of positive allosteric modulation. In the active-state M₂R•iperoxo•LY2119620 structure, LY2119620 acts as a small “wedge” by binding to a closed ECV conformation that further stabilizes the active state of the receptor. A similar mechanism can be observed for the inactive state, whereby a NAM of an agonist can behave as a PAM of an inverse agonist. At the M₂ mAChR, MD simulations revealed that alcuronium acts a PAM for the inverse agonist NMS by stabilizing the ECV in a more open conformation (relative to not only the active state but also the inactive state in the absence of modulator; Dror et al., 2013). (B) Structural basis of negative allosteric modulation. NAMs of mAChR agonists promote a more open ECV receptor conformation that is detrimental to agonist binding and G protein coupling. In contrast, NAMs of inverse agonists are not large enough to stabilize the open ECV, thus favoring a conformation that precludes inverse agonist binding. Arrows indicate an increase in the dynamics of the ECV that could reflect effects on ligand dissociation.

Figure 5.

Probe dependence as a consequence of a two-state model of allostery. (A) Cooperativity values for BQCA and different orthosteric agonists at the M₁ mAChR from cAMP accumulation assays (y axis values denote the fold increase in agonist potency in the presence of modulator). Acetylcholine (ACh) and carbachol (CCh) are high-efficacy agonists and are potentiated to a greater extent than the low-efficacy agonists pilocarpine and xanomeline. Thus, the degree of cooperativity is correlated with the intrinsic efficacy of each ligand. Data are replotted from Canals et al. (2012). (B) A speculative cartoon for a two-state model of allostery at the M₁ mAChR. High efficacy agonists have high cooperativity with PAMs by promoting a fully active conformation, indicated by a fully closed tyrosine lid. Low-efficacy agonists do not promote a fully active conformation, indicated by a partially closed tyrosine lid, and thus have lower cooperativity with BQCA.

Figure 6.

A potential mechanism of PAM selectivity. (A) The M₂/M₄ mAChR selective PAM LY2119620 binds in the allosteric pocket, as revealed by the active-state M₂ mAChR crystal structure (Protein Data Bank accession no. 4MQT). Residues within 4 Å of LY2119620 (colored orange) are shown as green colored sticks. Residues in ECL2 are numbered relative to the conserved disulfide bond between ECL2 and TM3. (B) Comparison of residues from A across all five human mAChR subtypes. Conserved residues are identified by an asterisk. (C) Nonconserved residues surrounding the LY2119620-binding site may contribute to selectivity.

Figure 7.

An allosteric network at the M₄ mAChR. (A) Previous studies on the M₄ mAChR have identified residues that either disrupt the binding of LY2033298 or alter the cooperativity between the allosteric and orthosteric site. These residues are shown in pink, modeled on to an active-state M₄ mAChR homology model (white) based on the M₂•iperoxo•LY2119620 crystal structure (Protein Data Bank accession no. 4MQT; Nawaratne et al., 2010; Leach et al., 2011; Thal et al., 2016). Views from the side (A) and extracellular surface (B) are shown.

Figure 8.

Probe dependence as a consequence of biased modulation at the M₂ mAChR. Cooperativity values of LY2033298 for different orthosteric agonists at the M₂ mAChR from ERK1/2 phosphorylation assays. Note that the magnitude and direction of the allosteric effect are not related to the degree of intrinsic efficacy of the orthosteric agonist. Data represent mean ± SEM and are replotted from Valant et al. (2012).

Figure 9.

Positive allosteric modulation of monomeric M₄ mAChRs. (A and B) [³H]NMS interaction binding studies between the PAM, LY2033298, and acetylcholine at the M₄ mAChR expressed in either CHO cells (A; replotted from Leach et al., 2010) or purified from Sf9 cells (B; M₄R-mT4L; for methods, see Thal et al., 2016) and reconstituted into rHDL particles, as was previously done for rhodopsin (Whorton et al., 2007). Nearly identical levels of cooperativity were observed, indicating that LY2033298 is able to influence the binding of acetylcholine in a monomeric mAChR and is not dependent on oligomerization status of the receptor. Data from B represent the mean ± SEM of n = 3 experiments performed in duplicate.