Editor's evaluation: Pharmacological hallmarks of allostery at the M4 muscarinic receptor elucidated through structure and dynamics

Abstract

Full text Figures and data Side by side Abstract Editor's evaluation Introduction Results Discussion Materials and methods Appendix 1 Data availability References Decision letter Author response Article and author information Metrics Abstract Allosteric modulation of G protein-coupled receptors (GPCRs) is a major paradigm in drug discovery. Despite decades of research, a molecular-level understanding of the general principles that govern the myriad pharmacological effects exerted by GPCR allosteric modulators remains limited. The M4 muscarinic acetylcholine receptor (M4 mAChR) is a validated and clinically relevant allosteric drug target for several major psychiatric and cognitive disorders. In this study, we rigorously quantified the affinity, efficacy, and magnitude of modulation of two different positive allosteric modulators, LY2033298 (LY298) and VU0467154 (VU154), combined with the endogenous agonist acetylcholine (ACh) or the high-affinity agonist iperoxo (Ipx), at the human M4 mAChR. By determining the cryo-electron microscopy structures of the M4 mAChR, bound to a cognate Gi1 protein and in complex with ACh, Ipx, LY298-Ipx, and VU154-Ipx, and applying molecular dynamics simulations, we determine key molecular mechanisms underlying allosteric pharmacology. In addition to delineating the contribution of spatially distinct binding sites on observed pharmacology, our findings also revealed a vital role for orthosteric and allosteric ligand–receptor–transducer complex stability, mediated by conformational dynamics between these sites, in the ultimate determination of affinity, efficacy, cooperativity, probe dependence, and species variability. There results provide a holistic framework for further GPCR mechanistic studies and can aid in the discovery and design of future allosteric drugs. Editor's evaluation This important work advances our understanding of the structural basis of allosteric modulation of the M4 muscarinic receptor but has broad implications for GPCRs. The evidence supporting the conclusions is exceptional, with multiple cryo-EM structures that are complemented by excellent pharmacological and dynamics studies. https://doi.org/10.7554/eLife.83477.sa0 Decision letter Reviews on Sciety eLife's review process Introduction Over the past 40 y, there have been major advances to the analytical methods that allow for the quantitative determination of the pharmacological parameters that characterize G protein-coupled receptor (GPCR) signaling and allosteric modulation (Figure 1A and B). These analytical methods are based on the operational model of agonism (Black and Leff, 1983) and have been extended or modified to account for allosteric modulation (Leach et al., 2007), biased agonism (Kenakin, 2012), and even biased allosteric modulation (Slosky et al., 2021). Collectively, these models and subsequent key parameters (Figure 1B) are used to guide allosteric drug screening, selectivity, efficacy, and ultimately, clinical utility, and provide the foundation for modern GPCR drug discovery (Wootten et al., 2013). Yet, a systematic understanding of how these pharmacological parameters relate to the molecular structure and dynamics of GPCRs remains elusive. Figure 1 with 2 supplements see all Download asset Open asset Pharmacological characterization of the positive allosteric modulators (PAMs), LY298 and VU154, with acetylcholine (ACh) and iperoxo (Ipx) at the human M4 muscarinic acetylcholine receptor (mAChR). (A) Schematic of the pharmacological parameters that define effects of orthosteric and allosteric ligands on a G protein-coupled receptor (GPCR). (B) A simplified schematic diagram of the Black–Leff operational model to quantify agonism, allosteric modulation, and agonist bias with pharmacological parameters defined (Black and Leff, 1983). (C) 2D chemical structures of the orthosteric and allosteric ligands used in this study. (D–G) Key pharmacological parameters for interactions between orthosteric and allosteric ligands in [3H]-N-methylscopolamine ([3H]-NMS) binding assays. (D) Equilibrium binding affinities (pKi and pKB) and (E) the degree of binding modulation (α) between the agonists and PAMs resulting in the modified binding affinities (F) α/KA and (G) α/KB. (H–K) Key pharmacological parameters relating to Gαi1 activation for interactions between orthosteric and allosteric ligands measured with the TruPath assay (Figure 1—figure supplement 1). (H) The signaling efficacy (τA and τB) and (I) transduction coupling coefficients (log (τ/K)) of each ligand. (J) The functional cooperativity (αβ) between ligands and (K) the efficacy modulation (β) between ligands. All data are mean ± SEM of three or more independent experiments performed in duplicate or triplicate with the pharmacological parameters determined using a global fit of the data. The error in (F, G, K) was propagated using the square root of the sum of the squares. See Table 1. Concentration–response curves are shown in Figure 1—figure supplement 1. Figure 1—source data 1 Related to Figure 1D–K. https://cdn.elifesciences.org/articles/83477/elife-83477-fig1-data1-v1.xlsx Download elife-83477-fig1-data1-v1.xlsx The muscarinic acetylcholine receptors (mAChRs) are an important family of five Class A GPCRs that have long served as model systems for understanding GPCR allostery (Conn et al., 2009). The mAChRs have been notoriously difficult to exploit therapeutically and selectively due to high-sequence conservation within their orthosteric binding domains (Burger et al., 2018). However, the discovery of highly selective positive allosteric modulators (PAMs) for some mAChR subtypes has paved the way for novel approaches to exploit these high-value drug targets (Chan et al., 2008; Gentry et al., 2014; Marlo et al., 2009). X-ray crystallography and cryo-electron microscopy (cryo-EM) have been used to determine inactive state structures for all five mAChR subtypes (Haga et al., 2012; Kruse et al., 2012; Thal et al., 2016; Vuckovic et al., 2019) and active state structures of the M1 and M2 mAChRs (Maeda et al., 2019). For the M2 mAChR, this includes structures co-bound with the high-affinity agonist iperoxo (Ipx) and the PAM LY2119620 in complex with a G protein mimetic nanobody (Kruse et al., 2013) and the transducers Go (Maeda et al., 2019) and β-arrestin1 (Staus et al., 2020). These M2 mAChR structures were foundational to validating the canonical mAChR allosteric site but are limited to only one agonist (iperoxo) and one PAM (LY2119620) and do not account for the vast pharmacological properties of ligands targeting mAChRs. A recent nuclear magnetic resonance (NMR) study of the M2 mAChR revealed differences in the conformational landscape of the M2 mAChR when bound to different agonists, but no clear link was established between the properties of the ligands and the conformational states of the receptor (Xu et al., 2019). The M4 mAChR subtype is of major therapeutic interest due to its expression in regions of the brain that are rich in dopamine and dopamine receptors, where it regulates dopaminergic neurons involved in cognition, psychosis, and addiction (Bymaster et al., 2003; Dencker et al., 2011; Foster et al., 2016; Tzavara et al., 2004). Importantly, these findings have been supported by studies utilizing novel PAMs that are highly selective for the M4 mAChR (Bubser et al., 2014; Chan et al., 2008; Leach et al., 2010; Suratman et al., 2011). Among these, LY2033298 (LY298) was the first reported highly selective PAM of the M4 mAChR and displayed antipsychotic efficacy in a preclinical animal model of schizophrenia (Chan et al., 2008). Despite LY298 being one of the best characterized M4 mAChR PAMs, its therapeutic potential has been limited by numerous factors, including its chemical scaffold, which has been difficult to optimize with respect to its molecular allosteric parameters (Figure 1C) and variability of response between species (Suratman et al., 2011; Wood et al., 2017b). In the search for better chemical scaffolds, the PAM, VU0467154 (VU154), was subsequently discovered. VU154 showed robust efficacy in preclinical rodent models; however, it also exhibited species selectivity that prevented its clinical translation (Bubser et al., 2014). Collectively, LY298 and VU154 are exemplar tool molecules that highlight the promises and the challenges in understanding and optimizing allosteric GPCR drug activity for translational and clinical applications. Herein, by examining the pharmacology of the PAMs LY298 and VU154 with the agonists ACh and Ipx across radioligand binding assays and two different signaling assays and analyzing these results with modern analytical methods, we determined the key parameters that describe signaling and allostery for these ligands. To investigate a structural basis for these pharmacological parameters, we used cryo-EM to determine high-resolution structures of the M4 mAChR in complex with a cognate Gi1 heterotrimer and ACh and Ipx. We also determined structures of receptor complexes with Ipx co-bound with the PAMs LY298 or VU154. Moreover, because protein allostery is a dynamic process (Changeux and Christopoulos, 2016), we performed all-atom simulations using the Gaussian accelerated molecular dynamics (GaMD) enhanced sampling method (Draper-Joyce et al., 2021; Miao et al., 2015; Wang et al., 2021a) on the M4 mAChR using the cryo-EM structures. The structures and GaMD simulations, in combination with detailed molecular pharmacology and receptor mutagenesis experiments, provide fundamental insights into the molecular mechanisms underpinning the hallmarks of GPCR allostery. To further validate these findings, we investigated the differences in the selectivity of VU154 between the human and mouse receptors and established a structural basis for species selectivity. Collectively, these results will enable future GPCR drug discovery research and potentially lead to the development of next generation M4 mAChR PAMs. Results Pharmacological characterization of M4 mAChR PAMs with ACh and Ipx The pharmacology of LY298 or VU154 interacting with ACh has been well characterized in binding and functional assays at the M4 mAChR (Bubser et al., 2014; Chan et al., 2008; Gould et al., 2016; Leach et al., 2010; Suratman et al., 2011; Thal et al., 2016). However, their pharmacology with Ipx has not been reported. Therefore, we characterized both PAMs with ACh and Ipx in binding and in two different functional assays to provide a thorough foundational comparative characterization of the pharmacological parameters of these ligands from the same study. We first used radioligand binding assays (Figure 1—figure supplement 1A) to determine the binding affinities (i.e., equilibrium dissociation constants) of ACh and Ipx (KA) for the orthosteric site and of LY298 and VU154 (KB) for the allosteric site of the unoccupied human M4 mAChR (Figure 1D), along with the degree of binding cooperativity (α) between the agonists and PAMs when the two are co-bound (Figure 1E). Analysis of these experiments revealed that LY298 and VU154 have very similar binding affinities for the allosteric site with values (expressed as negative logarithms; pKB) of 5.65 ± 0.07 and 5.83 ± 0.12, respectively (Table 1), in accordance with previous studies (Bubser et al., 2014; Leach et al., 2011). Both PAMs potentiated the binding affinity of ACh and Ipx (Figure 1E), with the effect being greatest between LY298 and ACh (~400-fold increase in binding affinity). Comparatively, the positive cooperativity between VU154 and ACh was only 40-fold. When Ipx was used as the agonist, the binding affinity modulation mediated by both PAMs was more modest, characterized by an approximately 72-fold potentiation for the combination of Ipx and LY298, and 10-fold potentiation for the combination of Ipx and VU154. These results indicate probe-dependent effects (Valant et al., 2012) with respect to the ability of either PAM to modulate the affinity of each agonist (Figure 1F and G). A probe-dependent effect was also observed with the radioligand, [3H]-NMS, evidenced by a reduction in specific radioligand binding due to negative cooperativity between the antagonist probe and LY298, which has been previously reported (Chan et al., 2008; Leach et al., 2010; Suratman et al., 2011; Thal et al., 2016). It is important to note that binding affinity modulation is thermodynamically reciprocal at equilibrium, and the affinities of LY298 and VU154 were thus also increased in the agonist bound state (Figure 1—figure supplement 1A). This results in LY298 having a fivefold higher binding affinity than VU154 when agonists are bound (Table 1). Table 1 Pharmacological parameters from radioligand binding and functional experiments. [3H]-NMS saturation binding on stable M4 mAChR CHO cellsConstructsSites per cell*pKD†Human WT M4 mAChR598,111 ± 43,067 (7)9.76 ± 0.05 (7)Mouse WT M4 mAChR21,027 ± 2188 (3)9.76 ± 0.05 (3)Human D432E M4 mAChR126,377 ± 10,066 (3)9.60 ± 0.07 (3)Human T433R M4 mAChR157,442 ± 36,658 (6)9.64 ± 0.09 (6)Human V91L, D432E, T433R M4 mAChR205,771 ± 20,975 (4)9.58 ± 0.08 (4)[3H]-NMS interaction binding assays between ACh or Ipx and LY298 or VU154 on stable M4 mAChR constructs in Flp-In CHO cellsConstructsPAMpKi ACh ‡pKi Ipx ‡pKB PAM ‡log αACh §log αIpx §Human WT M4 mAChRLY2984.50 ± 0.06 (4)8.30 ± 0.06 (4)5.65 ± 0.07 (8) ¶2.59 ± 0.10 (4)1.86 ± 0.10 (4)VU1544.40 ± 0.09 (4)8.19 ± 0.06 (8)5.83 ± 0.11 (12) ¶1.61 ± 0.13 (4)1.03 ± 0.10 (8)Mouse WT M4 mAChRLY2984.52 ± 0.07 (4)8.55 ± 0.06 (4)5.74 ± 0.07 (8) ¶1.78 ± 0.10 (4)1.30 ± 0.11 (4)*VU1544.59 ± 0.06 (4)8.57 ± 0.06 (3)6.07 ± 0.09 (7) ¶2.43 ± 0.10 (4)1.75 ± 0.12 (3)*Human D432E M4 mAChRLY298N.T.8.28 ± 0.04 (5)5.86 ± 0.07 (5)N.T.1.59 ± 0.06 (5)VU154N.T.8.27 ± 0.06 (6)6.21 ± 0.12 (6)N.T.1.04 ± 0.09 (6)Human T433R M4 mAChRLY298N.T.8.05 ± 0.08 (5)5.04 ± 0.04 (5)*N.T.1.91 ± 0.11 (5)VU154N.T.7.88 ± 0.04 (5)5.50 ± 0.08 (5)N.T.1.67 ± 0.07 (5)*Human V91L, D432E, T433R M4 mAChRLY298N.T.7.95 ± 0.10 (4)5.29 ± 0.26 (4)N.T.1.80 ± 0.22 (4)VU154N.T.7.89 ± 0.12 (4)6.34 ± 0.16 (4)*N.T.1.35 ± 0.16 (4)Gαi1 activation (TruPath) interaction assays between ACh or Ipx and LY298 or VU154 on transiently expressed M4 mAChR constructs in HEK293A cellsConstructsPAMlog τ ACh**log τ Ipx**pKB PAM ‡log τ PAM**log αβACh††log αβIpx††Human WT M4 mAChRLY2982.71 ± 0.14 (4)1.49 ± 0.12 (4)= 5.651.02 ± 0.03 (8) ¶2.01 ± 0.14 (4)1.96 ± 0.16 (4)VU154= 5.83–0.55 ± 0.08 (8) ¶1.22 ± 0.13 (4)0.20 ± 0.13 (4)pERK1/2 interaction assays between ACh or Ipx and LY298 or VU154 on stable M4 mAChR constructs in Flp-In CHO cellsConstructsPAMlog τ ACh**log τ Ipx**pKB PAM ‡log τC PAM ‡ ‡log αβACh††log αβIpx††Human WT M4 mAChRLY2983.27 ± 0.06 (8) ¶1.74 ± 0.03 (16) ¶= 5.651.19 ± 0.05 (12)**2.29 ± 0.22 (4)1.08 ± 0.28 (8)VU154= 5.830.11 ± 0.05 (12)**0.88 ± 0.23 (4)0.66 ± 0.15 (8)Mouse WT M4 mAChRLY298N.T.N.D.= 5.741.32 ± 0.07 (5)N.T.1.24 ± 0.12 (4)VU154N.T.N.D.= 6.071.47 ± 0.08 (5) § §N.T.2.08 ± 0.15 (5) § §Human D432E M4 mAChRLY298N.T.N.D.= 5.861.34 ± 0.08 (5)N.T.1.37 ± 0.28 (5)VU154N.T.N.D.= 6.210.78 ± 0.08 (5) § §N.T.1.02 ± 0.15 (5)Human T433R M4 mAChRLY298N.T.N.D.= 5.041.73 ± 0.13 (5) § §N.T.1.85 ± 0.28 (5)VU154N.T.N.D.= 5.500.95 ± 0.12 (5) § §N.T.1.18 ± 0.14 (5)Human V91L, D432E, T433R M4 mAChRLY298N.T.N.D.= 5.291.62 ± 0.09 (5) § §N.T.1.64 ± 0.30 (5)VU154N.T.N.D.= 6.340.68 ± 0.06 (5) § §N.T.1.34 ± 0.11 (5) § § Values represent the mean ± SEM with the number of independent experiments shown in parenthesis. N.T.: not tested; N.D.: not determined; Ach, acetylcholine; Ipx: iperoxo; PAM: positive allosteric modulator. * Number of [3H]-NMS binding sites per cell. † Negative logarithm of the radioligand equilibrium dissociation constant. ‡ Negative logarithm of the orthosteric (pKi) or allosteric (pKB) equilibrium dissociation constant. § Logarithm of the binding cooperativity factor between the agonist (ACh or Ipx) and the PAM (LY298 or VU154). ¶ Parameter was determined in a shared global analysis between agonists. ** Logarithm of the operational efficacy parameter determined using the Operational Model of Agonism. †† Logarithm of the functional cooperativity factor between the agonist (ACh or Ipx) and the PAM (LY298 or VU154). ‡ ‡ logτC = logarithm of the operational efficacy parameter corrected for receptor expression (methods in Appendix 1). § § Values from pKB PAM, log αIpx, log τC PAM, and log αβIpx that are significantly different from human WT M4 mAChR (p<0.05) calculated by a one-way ANOVA with a Dunnett’s post-hoc test. We subsequently used the BRET-based TruPath assay (Olsen et al., 2020) as a proximal measure of G protein activation with Gαi1 (Figure 1—figure supplement 1B). We also used a more amplified downstream signaling assay, extracellular signal-regulated kinases 1/2 phosphorylation (pERK1/2), that is also dependent on Gi activation (Figure 1—figure supplement 2A), to measure the cell-based activity of each PAM with each agonist. These signaling assays allowed us to determine the efficacy of the agonists (τA) and the PAMs (τB) (Figure 1H, Figure 1—figure supplement 2B). Importantly, efficacy (τ), as defined from the Black–Leff operational model of agonism (Black and Leff, 1983), is determined by the ability of an agonist to promote an active receptor conformation, the receptor density (Bmax), and the subsequent ability of a cellular system to generate a response (Figure 1B). Notably, in both signaling assays, the rank order of efficacy was ACh > Ipx > LY298 > VU154. We subsequently calculated the transducer coupling coefficient (τ/K) (Figure 1I, Figure 1—figure supplement 2C), a parameter often used as a starting point to quantify biased agonism (Kenakin et al., 2012) and that is specific to the intact cellular environment in which a given response occurs. Thus the dissociation constant (K) in the transduction coefficient subsumes the affinity for the ground state (non-bound) receptor, in addition to any isomerization states of the receptor that ultimately yield cellular responses (Kenakin and Christopoulos, 2013). Consequently, in both assays, the rank order of transducer coupling was Ipx >> ACh ~ LY298 > VU154 due to Ipx having a higher binding affinity for the receptor. Overall, these results indicate that although ACh is a more efficacious agonist than Ipx, it has lower transducer coupling coefficient. In contrast, LY298 has both better efficacy and transducer coupling coefficient than VU154 (Table 1). The signaling assays and use of an operational model of allosterism also allowed for the determination of the functional cooperativity (αβ) exerted by the PAMs (Figure 1J, Figure 1—figure supplement 2D), which is a composite parameter accounting for both binding (α) and efficacy (β) modulation. Notably, VU154 displayed lower positive functional cooperativity with ACh than LY298. Strikingly, VU154 had negligible functional modulation with Ipx, in contrast to the cooperativity observed with ACh in the TruPath assay. The tenfold difference in αβ values for VU154 between ACh and Ipx highlights the dependence of the orthosteric probe used in the assay (i.e. probe dependence); on this basis, VU154 would be classified as a ‘neutral’ allosteric ligand (not a PAM) with Ipx in the TruPath assay, that is, VU154 still binds to the allosteric site, but displays neutral cooperativity (αβ = 1) with Ipx (Table 1). The degree of efficacy modulation (β) that the PAMs have on the agonists can be calculated by subtracting the binding modulation (α) from the functional modulation (αβ) (Figure 1K, Figure 1—figure supplement 2E). A caveat of this analysis is that errors for β are higher due to the error being propagated between calculations. Ideally, the degree of efficacy modulation would be determined in an experimental system where the maximal efficacy of system is not reached by the agonists alone (Berizzi et al., 2016). Nevertheless, our analysis shows the PAMs LY298 and VU154 appear to have a slight negative to neutral effect on agonist efficacy in the Gi1 TruPath and pERK1/2 assays (Table 1), suggesting that the predominant allosteric effect exerted by these PAMs is mediated through modulation of binding affinity. Collectively, our extensive analysis on the pharmacology of LY298 and VU154 with ACh and Ipx offers detailed insight into the key differences between these ligands across a range of pharmacological properties: ligand binding, probe dependence, efficacy, agonist–receptor–transducer interactions, and allosteric modulation (Figure 1, Table 1). We hypothesized that structures of the human M4 mAChR in complex with different agonists and PAMs combined with molecular dynamic simulations could provide high-resolution molecular insights into the different pharmacological profiles of these ligands. Determination of M4R-Gi1 complex structures Similar to the approach used in prior determination of active-state structures of the M1 and M2 mAChRs (Maeda et al., 2019), we used a human M4 mAChR construct that lacked residues 242–387 of the third intracellular loop to improve receptor expression and purification, and made complexes of the receptor with Gi1 protein and either the endogenous agonist, ACh, or Ipx. Due to the higher affinity of Ipx compared to ACh (Schrage et al., 2013), we utilized Ipx to form additional M4R-Gi1 complexes with or without the co-addition of either LY298 or VU154. In all instances, complex formation was initiated by combining purified M4 mAChR immobilized on anti-FLAG resin with detergent solubilized Gi1 membranes, a single-chain variable fragment (scFv16) that binds Gαi and Gβ, and the addition of apyrase to remove guanosine 5′-diphosphate (Maeda et al., 2018). For this study, we used a Gi1 heterotrimer composed of a dominant negative form of human Gαi1, and human Gβ1 and Gγ2. (Liang et al., 2018b). Vitrified samples of each complex were imaged using conventional cryo-TEM on a Titan Krios microscope (Danev et al., 2021). The structures of ACh-, Ipx-, LY298-Ipx-, and VU154-Ipx-bound M4R-Gi1 complexes were determined to resolutions of 2.8, 2.8, 2.4, and 2.5 Å, respectively (Figure 2A, Figure 2—figure supplement 1, Table 2). For the ACh-bound M4R-Gi1 complex, an additional focus refinement yielded an improved map of the receptor and binding site (2.75 Å) for modeling (Figure 2—figure supplements 2 and 3). The cryo-EM density maps for all complexes were sufficient for confident placement of backbone and sidechains for most of the receptor, Gi1, and scFv16, and the bound ligands with exception of the alkyne bond of Ipx, which was consistent with prior cryo-EM studies (Maeda et al., 2019; Figure 2B, Figure 2—figure supplement 3). Figure 2 with 5 supplements see all Download asset Open asset Cryo-electron microscopy (cryo-EM) structures of the M4R-Gi1-scFv16 complexes. (A) Cryo-EM maps of Ipx-bound M4R-Gi1-scFv16 complex with views from the membrane and the extracellular surface. Cryo-EM maps of the other ligand-bound structures are shown in Figure 2—figure supplement 1. (B) Representative EM density around the ligands in this study. EM-maps of Ipx-, LY298-Ipx-, and VU154-Ipx were set to a contour level of 0.011 and the receptor-focused map of ACh- was set to 0.32. (C–E) Comparison of the receptor models with bound ligands and views from the (C) membrane, (D) extracellular surface, and (E) intracellular surface. Table 2 Cryo-electron microscopy (cryo-EM) data collection, refinement, and validation statistics. M4R-Gi1-IpxM4R-Gi1-Ipx-LY298M4R-Gi1-Ipx-VU154M4R-Gi1-AChData collection & refinementEMD code26,09926,10026,10126,102Micrographs5056512160215913Electron dose (e-/A2)666659.553.6Voltage (kV)300300300300Pixel size (Å)0.830.830.830.83Spot sizeExposure time4435Movie frames76767571K3 CDS modeNoNoNoYesDefocus range (µm)0.5–1.50.5–1.50.5–1.50.5–1.5Symmetry imposedC1C1C1C1Particles (final map)415,743617,793677,392315,595Resolution @0.143 FSC (Å)2.82.42.52.8RefinementCCmap–model0.870.870.880.82Map sharpening B factor (Å2)–80.9–60.8–46.6–85.1Model qualityPDB code7TRK7TRP7TRQ7TRSR.M.S. deviations Bond length (Å)0.0040.0040.0050.006 Bond angles (o)0.8490.8110.8260.773Ramachandran Favored (%)98.3899.1498.0298.10 Outliers (%)0000Rotamer outliers (%)0.110.2100C-beta deviations (%)0000Clashscore2.692.622.264.08MolProbity score1.061.051.001.19 mAChR: muscarinic acetylcholine receptor; ACh: acetylcholine; Ipx: iperoxo; FSC: Fourier shell correlation. In all four structures, EM density beyond the top of transmembrane helix 1 (TM1) and the third intracellular loop (ICL3) of the receptor was poorly observed and not modeled. Similarly, the EM density of the α-helical domain of Gαi1 was poor and not modeled. These regions are highly dynamic and typically not modeled in many class A GPCR-G protein complex structures. Apart from these regions, most amino acid side chains were well resolved in the final EM density maps (Figure 2—figure supplement 3). Structure and dynamics of agonist binding Recently, cryo-EM structures of M4R-Gi1 complexes bound to Ipx, Ipx, and the PAM, LY2119620, and a putative novel allosteric agonist, c110, were determined (Wang et al., 2022). Surprisingly, comparison of the M4R-Gi1 complex structures revealed larger differences in the position of key orthosteric and allosteric site residues than the M1R-G11 and M2R-GoA complex structures (Figure 2—figure supplement 4). Unfortunately, the quality of density in the EM maps around the orthosteric and allosteric sites of these M4R-Gi1 structures (Wang et al., 2022) was poor, resulting in several key residues being mismodeled in each site (Figure 2—figure supplement 5). Therefore, differences between the M4R-Gi1 structures described herein and those by Wang et al., 2022 are highly likely to not be due to genuine differences and, as such, we compared the prior M1R-G11 and M2R-GoA complex structures (Maeda et al., 2019) in this study. Overall, our M4R-Gi1 complex structures are similar in architecture to that of other activated class A GPCRs, including the M1R-G11 and M2R-GoA complexes (Figure 2—figure supplement 4). Superposition of the M4R-Gi1 complexes revealed nearly identical structures with root mean square deviations (RMSD) of 0.4–0.5 Å for the full complexes and 0.3–0.4 Å for the receptors alone (Figure 2C). The largest differences occur around the extracellular surface of the receptors (Figure 2D) along with slight displacements in the position of the αN helix of Gαi1 and Gβ1, Gγ2, and scFv16 with respect to the receptor (Figure 2—figure supplement 1D). The EM density of side chains surrounding the ACh and Ipx binding sites (Figure 3A and B) was well resolved providing the opportunity to understand structural determinants of orthosteric agonist binding. The orthosteric site of the M4 mAChR, in common with the other mAChR subtypes, is buried within the TM bundle in an aromatic cage that is composed of four tyrosine residues, two tryptophan residues, one phenylalanine residue, and seven other polar and nonpolar residues (Figure 3C). Notably, all 14 of these residues are absolutely conserved across all five mAChR subtypes, underscoring the difficulty in developing highly subtype-selective orthosteric agonists (Burger et al., 2018). Both ACh and Ipx have a positively charged trimethyl ammonium ion that makes cation-π interactions with Y1133.33, Y4166.51, Y4397.39, and Y4437.43 (Figure 3C; superscript refers to the Ballesteros and Weinstein scheme for conserved class A GPCR residues; Ballesteros and Weinstein, 1995). Likewise, both ACh and Ipx have a polar oxygen atom that can form a hydrogen bond to the indole nitrogen of W1644.57 with the oxygen of Ipx also being in position to interact with the backbone of N1173.37 (Figure 3D). Mutation of any of these contact residues reduces the affinity of ACh, validating their importance for agonist binding (Leach et al., 2011; Thal et al., 2016). The largest chemical difference between ACh and Ipx is the bulkier heterocyclic isoazoline group of Ipx that makes a π-π interaction with the conserved residue W4136.48 (Figure 3D). The residue W4136.48 is part of the CWxP motif, also known as the rotamer toggle switch, a residue that typically undergoes a change in rotamer between the inactive and active states of class A GPCRs (Shi et al., 2002). Figure 3 with 1 supplement see all Download asset Open asset Interactions of acetylcholine (ACh) and iperoxo (Ipx) with the receptor. (A, B) Cryo-electron microscopy (cryo-EM) density of the (A) ACh- and (B) Ipx-bound structures. (C, D) Interactions at the orthosteric binding site comparing the active state ACh- and Ipx-bound structures with the inactive state tiotropium-bound structure (PDB: 5DSG). Arrows denote relative movement of residues between the inactive and active states. (D) Detailed interactions of ACh and Ipx. Hydrogen bonds are shown as black dashed lines. (E, F) Time courses from Gaussian accelerated molecular dynamics (GaMD) simulations of the ACh- and Ipx- bound M4R-Gi1 cryo-EM structures, each performed with three separate replicates. Individual replicate simulations are illustrated with different colors. The heading of each plot refers to the specific model used in the simulations. Root mean square deviations (RMSDs) of (E) ACh and (F) Ipx from simulations of the cryo-EM structures. (G, H) Cross-sections through the ACh- and Ipx-bound structures denoting the relative size of the binding pockets outlined in black. To investigate the structural dynamics of the M4 mAChR, we performed three independent 500 ns GaMD simulations on the ACh- and Ipx-bound M4R-Gi1 cryo-EM structures (Table 3). GaMD simulations revealed that ACh undergoes higher fluctuations in the orthosteric site than Ipx (Figure 3E and F, Video