Mechanism of Hormone Peptide Activation of a GPCR: Angiotensin II Activated State of AT₁R Initiated by van der Waals Attraction

Abstract

We present a succession of structural changes involved in hormone peptide activation of a prototypical GPCR. Microsecond molecular dynamics simulation generated conformational ensembles reveal propagation of structural changes through key “microswitches” within human AT₁R bound to native hormone. The endocrine octa-peptide angiotensin II (AngII) activates AT₁R signaling in our bodies which maintains physiological blood pressure, electrolyte balance, and cardiovascular homeostasis. Excessive AT₁R activation is associated with pathogenesis of hypertension and cardiovascular diseases which are treated by sartan drugs. The mechanism of AT₁R inhibition by sartans has been elucidated by 2.8 Å X-ray structures, mutagenesis, and computational analyses. Yet, the mechanism of AT₁R activation by AngII is unclear. The current study delineates an activation scheme initiated by AngII binding. A van der Waals “grasp” interaction between Phe8^AngII with Ile288^7.39 in AT₁R induced mechanical strain pulling Tyr292^7.43 and breakage of critical interhelical H-bonds, first between Tyr292^7.43 and Val108^3.32 and second between Asn111^3.35 and Asn295^7.46. Subsequently changes are observed in conserved microswitches DRY^TM3, Yx7K(R)^TM5, CWxP^TM6, and NPxxY^TM7 in AT₁R. Activating the microswitches in the intracellular region of AT₁R may trigger formation of the G-protein binding pocket as well as exposure of helix-8 to cytoplasm. Thus, the active-like conformation of AT₁R is initiated by the van der Waals interaction of Phe8^AngII with Ile288^7.39, followed by systematic reorganization of critical interhelical H-bonds and activation of microswitches.

Graphical Abstract

INTRODUCTION

Nearly 5% of the human genome encodes G-protein coupled receptors (GPCRs), which are targets of ~34% of therapeutic drugs currently in use.¹ GPCRs are seven-transmembrane helical proteins that modulate vital cell-signaling responses to a variety of natural and pharmaceutical ligands.^1–5 Extracellular agonists bind and initiate receptor conformational changes which propagate across the trans-membrane (TM) α-helical domain to intracellular “microswitches” formed by highly conserved sequence motifs in GPCRs. These events result in the transient formation of a cytosolic cleft for binding and activating a cognate G protein. By comparing the antagonist-bound inactive and [agonist + G protein/antibody] bound active state structures of GPCRs, the activating structural changes were identified.^6–11 Structure−function studies have shown that mutating a microswitch may still allow the activating motion of TM-helices but impair cell signaling.¹² Intramolecular events that form the electromechanical basis of signal transduction are predominantly documented for amine activated GPCR prototypes. Although three agonist-peptide bound GPCR structures are available (PDB ids: 6DDF, 5UNF, 5T04), an ensemble view of the conformational dynamics of peptide hormone activation of a GPCR is absent to date.^13–16

The angiotensin type 1 receptor (AT₁R) is a clinically important GPCR that binds the renin−angiotensin system hormone angiotensin II (AngII). Cell signaling by activated AT₁R involves stimulation of the Gq_11/12 leading to intra-cellular production of calcium, lipid messengers, and activation of protein kinases.^17–20 In addition, G protein-independent signaling pathways are also activated.^21–25 AngII-activated AT₁R is responsible for homeostatic regulation of blood pressure and water−electrolyte balance in the body.^26,27 AT₁R is widely expressed in vascular and myocardial tissue, the liver, the kidney, the adrenal cortex, and specific brain centers. Chronic AT₁R activation produces disease states including hypertension, inflammation, endothelial dysfunction, athero-sclerosis, congestive heart failure, and end organ damage. AT₁R blocking drugs (ARBs) such as olmesartan are used as a first line of therapy in fighting these diseases. Previous studies have suggested a transition scheme for functional states of AT₁R [Figure 1A] upon antagonist and agonist binding.^28,29 Two high-resolution inactive state structures of AT₁R are reported but attempts made to crystallize the AngII bound AT₁R did not succeed.^13,14 To bridge the knowledge gap with regard to the AngII-activated state of AT₁R, we performed microsecond molecular dynamics simulation (MDS) analysis in a membrane bilayer environment.

Figure 1.

(A) Scheme of proposed conformational states governing AT₁R function. (B) 3D structure of AT₁R showing orthosteric pocket and important motifs in G-protein and non-G-protein signaling. (C) Multiple sequence alignment of AT₁R with Class A GPCRs highlighting conserved motifs.

In this first report delineating conformational ensemble of a clinically relevant peptide hormone—GPCR—we carried out MDS starting with the 2.8 Å X-ray crystallographic structure of human AT₁R.^13,14 We corroborate the MDS results [Figure 1B] by (i) comparing to active-state structures of established GPCRs;^30–34 (ii) using stereochemical features of AngII-analogs defined by structure activity relationship (SAR),^35–38 and (iii) by combining with site-directed mutagenesis findings on pharmacology of AT₁R.^37,39–49 As in all GPCRs, several conserved residues form functional motifs within AT₁R [Figure 1C]. During AT₁R-AngII signaling, these motifs may act as microswitches that change from an inactive to an active conformation. Hence, inspecting perturbation in CWxP^TM6, DRY^TM3, Yx7K(R) ^TM5, and NPxxY^TM7 motifs in comparison to different established activated GPCRs^39–42 will further validate our MDS. SAR analysis of the AngII peptide, Asp¹-Arg²-Val³-Tyr⁴-Ile⁵-His⁶-Pro⁷-Phe⁸ has established that the charged property of Asp¹, Arg², Phe⁸-COO⁻ and bulk-hydrophobicity of Tyr⁴ determine binding affinity and specificity for AT₁R [Figure S1A]. The aromatic Phe⁸ side chain and cumulative influence of Tyr4, His6, and Pro7 are responsible for agonist potency of AngII in AT₁R activation.^50–52 Mutagenesis analysis showed that AngII binding requires residues, Tyr35^1.39, Trp84^2.60, Arg162^ECL2, Lys199^5.42, and Asp281^7.32 [Figure S1B]. In addition, the critical role of Asp74^2.50, Asn111^3.35, Arg126^3.50, Trp253^6.48, His256^6.51, Ile288^7.39, Tyr292^7.43, and Asn295^7.46 in receptor activation are documented.^{37,44,46,48,49,53,54} How these features of the receptor and the hormone cooperatively produce an AngII-activated state of AT₁R is unknown.

Our working hypothesis assumes that microswitches constrain spontaneous activation in the basal state of AT₁R. Distinct interactions of AngII with the AT₁R binding pocket break the prime constraint instigating the microswitches to change from inactive to active conformation in the course of receptor activation. Understanding the details of AngII activation of AT₁R could further help in designing more potent ARBs to prevent AngII associated pathogenesis.

RESULTS

Global Conformational Changes.

In the 1 μS MDS following flexible docking of AngII to membrane embedded AT₁R (PDB id: 4ZUD),¹³ the trajectory plot for AngII and the receptor backbone reached equilibrium (Figure S2A). Significant changes in root-mean-square fluctuation (RMSF) was observed in DRY, ECL2, and ICL3 motifs as shown in Figure S2B in the AT₁R-AngII compared to AT₁R-Olmesartan. High fluctuation in these motif may correlate with the activation of the receptor. The RMSD plot indicated significant movement in TM-helices, ECL2, ICL3, and helix-8 (Figure S3). The AT₁R conformation ensembles produced from MDS closely match the crystal structure of peptide-agonist bound ET_BR (PDB id: 5GLH)¹⁶ (Figure 2A and B). The volume of the AngII bound pocket of AT₁R is ~1203.11 ± 335.83 ų, comparable to the pocket of ET_BR which has a volume of 895.92 ų (PDB id: 5GLH).¹⁶ The extracellular portions of TM1, TM3, TM4, TM5, and TM7 move in the AngII-bound state (Figure S4). These structural changes are associated with the orthosteric pocket adopting a compact “closed” configuration enabling strong interactions with AngII. Together the N-terminal tail (NT) and EC-loops (ECL) including the β-strands of ECL2 form a lid-like architecture that covers the orthosteric pocket (Figure 2C and D). Mutagenesis and accessibility mapping studies have predicted ligand-specific movements in NT and ECL regions of AT₁R, especially in the ECL2.^55–57 The water accessible entrance in the apo-AT₁R might facilitate the entry of ligands to the orthosteric pocket.

Figure 2.

(A) Representative conformational ensembles from 1 μs MDS of AngII bound AT₁R. (B) Superimposed structure of AT₁R-AngII and ET_BR-Endothelin. Superimposed structure of AT₁R-AngII (from MDS) and AT₁R-Olmesartan (crystal structure) bound states showing transmembrane view (C), extra-cellular view (D), and intra-cellular view (E).

By contrast, in the intracellular region, the TM1, TM3, TM5, TM6, and TM7 move outward (Figure S4). The movement of ICL3 associated with TM3, TM6, and TM7 opens a crevice that is supposed to bind the G protein (Figure 2C and E), but the opening is not complete as seen in the G_s protein bound β₂AR structure.⁴⁰ The outward movement of TM6 is comparatively less than observed in the active state crystal structures of GPCRs.⁴⁰ Functional studies in AT₁R have indicated that exposure of the DRY^TM3 motif, the membrane proximal segments of ICL3, and the NPxxY^TM7 motif are necessary and sufficient to recruit the G protein.²⁰ A well-defined helix-8 observed in the inactive state AT₁R shows ~15 Å shift toward the cytoplasmic space with decreased helical stability in a small number of the ensemble frames (Figure S4). Observed conformational change in helix-8 is relevant because functional studies have shown that helix-8 of AT₁R is a hotspot for recruitment of non-G protein signaling molecules such as Jak2, Cam kinase, and Filamin A.^20,58 These AngII-induced receptor conformational changes are not observed when olmesartan is used as the ligand for MDS and the ensemble frames are indicative of agonist bound structures of β₂AR, ET_BR, NT1R, and μ-opioid receptor in which structural changes linking the orthosteric pocket and the cytoplasmic G-protein binding pocket is observed.

AngII Binding Interactions.

Supporting the mutagenesis findings, AngII residues 3−8 are embed within the orthosteric pocket (Figure 3A). Asp1^AngII and Arg2^AngII extend to the NT and ECL2 and ECL3.^{55,56,59–61} Binding of AngII involves ~17 AT₁R residues forming ionic, H-bond, van der Waals, and π−π stacking interactions (Figure 3B). The conformation of bound AngII observed in our MDS is similar to NMR,⁶² flexible docking,³⁸ or modeling studies.³⁵ Superimposition of AngII and Olmesartan indicates overlap of their aromatic−acidic pharmacophore groups surrounded by shared residues, Tyr35^3.91, Trp84^2.60, Tyr87^2.63, Lys199^5.42, Arg167^ECL2, Ile288^7.39, and Tyr292^7.43 which validate competitive antagonism of AngII and ARBs (Figure S5). Mutagenesis studies indicate that these are critical residues for binding both AngII and ARBs (see Table S1).

Figure 3.

(A) Binding mode of AngII in the orthosteric pocket of AT₁R. (B) Interaction of AngII with AT₁R residues >25% in at least 3 out of 6 independent MDS experiments are shown. A single residue can have >100% value due to multiple interactions (e.g., weak, ionic, hydrophobic, etc.).

The octapeptide is oriented so that the Phe8^AngII α-carboxyl interaction with Lys199^5.42 is robust (Figure 3). MDS trajectory analysis of Phe8^AngII α-carboxyl H-bond and ionic interaction with Lys199^5.42 estimated to be >75% in multiple trials. It is known that the Phe8^AngII α-carboxyl, if blocked by methylation or amination causes tachphylaxis in vivo. Lys199^5.42 mutation to Ala/Gln in AT₁R completely abolished binding of AngII and its analog [Sar¹, Ile⁸]AngII. Noda et al.⁴⁵ suggesting that Lys199^5.42 interaction with Phe8^AngII α-carboxyl may be important for the orientation of His6-Pro7-Phe8AngII in mediating receptor agonism. After 200 ns the guanidinium group of Arg167^ECL2 forms multiple interactions including water mediated bonding with backbone carbonyl groups of Phe8^AngII, His6^AngII, and Ile5^AngII amounting to 250% of interaction. When Arg167^ECL2 was mutated, AngII binding is abolished. Therefore, projecting Arg167^ECL2 into the depth of the orthosteric pocket to support interaction with AngII backbone carbonyls might be critical for shaping the receptor-bound conformation of AngII. We have explored every critical binding pocket residue of AT₁R involved in AngII binding. We did not find significant change in the conformation of Arg167^ECL2 in AT₁R bound with AngII or olmesartan. Supporting result was observed in our mutagenesis and crystal structure experiments (Table S1).

The NH₂-backbone amide of Asp1^AngII interacts with Asp263^ECL3 and Ser186^ECL2 and the Asp1^AngII carboxyl group maintains interaction with Lys20^NT. Deletion of Asp1^AngII reduces the binding affinity slightly and substitution with sarcosine enhances the binding affinity. The Arg2^AngII interacts with Asp281^7.32 validating the SAR studies, which established its importance for the hormone’s binding affinity. The Asp281^7.32 mutation was shown to reduce AngII affinity as well as efficacy.⁶⁰ Previous modeling studies also indicate that additional stabilization of this interaction may involve Asp2787.29. During simulation, Arg2^AngII appears to form a strong interaction with Asp1^AngII as well as Asp17NT. These interactions contributed ≈130% each, implicating their role in binding Arg2^AngII.

Phe8^AngII reaches deep toward the core of the receptor, while Tyr4 ^AngII is positioned near the extracellular entrance of the orthosteric binding-cavity.Tyr4^AngII and Phe8^AngII are important for agonistic potency, [Ile4, Ile8]-AngII functions as a low affinity antagonist and [Ile4]AngII and [Ile8]AngII are low efficacy partial agonists for AT₁R. Hence, Tyr4^AngII and Phe8^AngII interactions are critical for activation of the AT₁R (Figure 3). In MDS we observed that dynamic state of Tyr4^AngII precludes a stable interaction of this side chain with the receptor.

Phe8^AngII is oriented toward Trp84^2.60 forming a strong π−π stacking interaction. Also Trp84^2.60 forms intramolecular interactions with Leu81^2.57, Tyr87^2.63, Thr88^2.64, Ala104^3.28, and Arg167^ECL2. Trp84^2.60 mutation abolished AngII binding because the smaller aliphatic residues introduced alter conformational geometry and disrupt interactions with Tyr87^2.63 and Arg167^ECL2. Tyr92^ECL1 mutation to Ala completely abolished AngII binding. Tyr92^ECL1 forms hydrophobic interaction with Pro7^AngII, which is important for the active conformation of the AngII. Pro7^AngII substitution with a positively or uncharged residue completely abolishes binding affinity of AngII.

The Ile288^7.39 side chain is oriented toward Phe8^AngII making a van der Waals interaction. Mutation with Ala in this position abolished AngII binding, since the shorter side chain of Ala could not reach Phe8^AngII to establish a van der Waals interaction. Mutagenesis further showed that a bulkier Phe288^7.39 side chain does not significantly affect AngII binding. This observation is consistent with modifications of Phe8^AngII such as [Sar1, Ile8]AngII and des-Phe8-heptapeptide, which are devoid of AT₁R agonism.

AngII-Induced Conformational Changes.

Systematic changes observed in AT₁R conformation upon AngII binding in our MDS mimic structures of the active-like state GPCRs. In the following sections, we describe the time sequence of structural changes connecting AngII binding site to different cytoplasmic microswitches (Table 1).

Table 1.

Time Course of Molecular Events Observed in MD Simulation^a

	Molecular Events
time	AT1R-AngII	AT1R-olmesartan
0 ns	strong interaction of Phe8AngII-Ile2887.39	strong interaction of alkyl tail of omesartan-Ile2887.39
~200 ns	Phe8AngII pulls up Ile2887.39	position of Ile2887.39 remain same, strong interaction with alkyl tail of Omesartan
~400 ns	breakage of Val1083.35 Cα-OH-Tyr2927.34 H-bond (TM3-TM7)	Val1083.32 Cα-OH-Tyr2927.34 (TM3-TM7) H-bond is consistent
~600 ns	breakage of Asn1113.32 N2957.46 H-bond (TM3-TM7)	Asn1113.32 N2957.46 H-bond (TM3-TM7) is consistent
~700	opening of intracellular domain for Gprotein binding	intracellular domain are closed for Gprotein binding
~1000 ns	open intracellular domain with active like conformation	closed intracellular domain with inactive conformation

^aAll these events are captured and shown in MDS trajectory videos1−4.

Breaking Primary Constraint (Asn111^3.35-Asn295^7.46 H-Bond).

The most important agonism determinant is Phe8^AngII side chain which orientates toward a cluster formed by Tyr35^1.39, Trp84^2.60, Pro285^7.36, Ile288^7.39, and Tyr292^7.43 throughout the 1000 ns MDS. The Phe8^AngII aromatic side chain “grasps” Ile288^7.39 in van der Waals attraction within ~100 ns (Figure 4), and this “grasp” pulls the aromatic side chain of Tyr292^7.43, resulting in breakage of the interhelical H-bond between Tyr292^7.43−OH and the backbone carbonyl of Val108^3.32 in ~200 ns which in turn breaks interhelical H-bonds between Asn111^3.35-Asn295^7.46 in ~500 ns. Interestingly, equivalent intrahelical H-bonds between TM3-TM7 are absent in AT₂R.⁶³The AT₂R structure (PDB id: 5unf) resembles active like GPCRs and it has a Phe residue at TM^7.43 position that is equivalent to Tyr292^7.43 in AT₁R. This evidence further supports the conclusion that MDS produced an active-like state of AT₁R.

Figure 4.

(A) Superimposed Phe8^AngII orientation changes at time intervals indicated. Orientation of Phe8^AngII captured at 0 (B), 0.5 (C), and 1 μs (D).

In contrast, binding olmesartan establishes independent interaction with both Ile288^7.39 and Tyr292^7.43 and consequently stabilizes the TM7-TM3 interhelical H-bonds between Tyr292^7.43-OH-Val108^3.32 and Asn111^3.35-Asn295^7.46 (Figures 5, S6, and S7). In support of this finding, independent studies have shown that nonpeptide AT₁R antagonists show reduced binding affinity to active like state of AT₁R.^63,64 MD simulations with -Ile8^AngII, -Ala8^AngII, and -Gly8^AngII indicated that the aliphatic side chains replacing Phe8^AngII systematically decrease interaction with Ile288^7.39 which is also validated by the agonist-efficacy experimental data shown in Figure S8.

Figure 5.

Representation of Phe8^AngII induced motion of side chains leading to breakage of N111^3.35-N295^7.46 H-bond.

The disruption of the Asn111^3.35-Asn295^7.46 bond being central to AT₁R activation is well-documented: (i) stable interhelical H-bonds between Asn111^3.35-Asn295^7.46 are confirmed in ARB-bound AT₁R crystal structures;^13,14 (ii) Asn111^3.35 mutation to smaller residues Gly, Ala, Ser, and Cys constitutively activate AT₁R;^{47,60,65–67} (iii) Asn295^7.46 to Ser mutation mimics this effect;⁶⁴ and (iv) long-range effects of mutations destabilizing Asn111^3.35-Asn295^7.46 bond also cause constitutive activation.⁵⁷ To further examine, we modeled Asn111Gly^3.35 mutant of AT₁R and found that the model coordinates matched the active-like-state AT₂R structure better than inactive state AT₁R structure. Substitution of larger side chains, such as the Trp111^3.35 produced the most inactive AT₁R mutant.⁴⁹ The residues involved are conserved in AT₂R, suggesting that the same mechanism may be involved in AngII activation of AT₂R.

Mechanism of TM-Bundle Entropy.

Interhelical polar interaction networks involving conserved residues, Asn46^1.50, Asp74^2.60, Asn295^7.46, and Asn298^7.49 are rearranged. The Tyr292^7.43 side chain forms π−π stacking with Phe77^2.53 and the Asn111^3.35-Asn295^7.46 H-bond interacts with Asp74^2.50 in the AngII bound state. When the Asn111^3.35-Asn295^7.46 H-bond breaks, the Asn295^7.46 interacts strongly with Asp74^2.50 and Asn46^1.50. The weakening of TM7-TM3 interaction sways the side chain of Leu112^3.36 toward Lys199^5.42. In a previous MDS study, Cabana et al.⁴⁹ compared constitutively active Gly111^3.35 mutant with the inactive Trp111^3.35 mutant. Although the Asn111^3.35-Asn295^7.46 H-bond is broken in both of these mutants, the activation switch to Asn295^7.46-Asp74^2.50-Asn46^1.50 H-bonds was not observed in the Trp111^3.35 mutant. Instead, Asp74^2.50 stably interacted with Trp111^3.35. In the constitutively active Gly111^3.35 mutant, they observed H-bond between Asp74^2.50-Asn46^1.50. Therefore, orderly rearrangement of interhelical network coordinated with Asn111^3.35-Asn295^7.46 interaction in the AT₁R are critical steps in activation. AngII binding also perturbed another important hydrogen bonding network involving Asp74^2.50, with Ser109^3.33, Tyr113^3.37 and Cys296^7.47 that is conserved among class A GPCRs (Figure S9). Biochemical experiments showed that agonist-activation of AT₁R alters the orientation of TM2, reducing its interaction with TM7. The Asp74^2.50 and Phe77^2.53 side chains buried in AT₁R become solvent exposed when AngII-binds.^61,68,69

In addition to the networks in the AngII binding pocket, extensive H-bond network linking the orthosteric ligand pocket and the G-protein-binding cleft also rearrange in the active-like state. In this network, polar residues are conserved as documented in family A GPCR structures and may involve water molecules and a sodium ion as discovered in the high-resolution inactive-state structure of a few GPCRs.^10,11,40 Since involvement of allosteric sodium is not experimentally confirmed in AT₁R, we did not include a sodium ion in our MDS analysis. However, the cytoplasmic H-bond networks can be traced to the side chains of Asp74^2.50, Asn111^3.35, Trp253^6.48, and Asn295^7.46 that could be involved in coordination of a sodium ion. As discussed below the polar networks extend from CWxP^TM6 motif to the cytoplasmic surface involving microswitches, the DRY^TM3, Yx7K(R)^TM5, and NPxxY^TM7 sequences.

Motion of Trp253^6.48.

Trp253^6.48 is part of the CWxP motif, a rotamer toggle switch which leads to the opening of G protein cleft in class A GPCRs. This highly conserved amino acid residue is located on the floor of the ligand pocket and adjacent to the conserved Pro2556.50. No significant change in orientation of Pro255^6.50 is observed in our MD simulation. In the AngII activated state the orientation of Trp253^6.48 side chain tilts intracellular toward His256^6.51 and Phe249^6.44_, which are in close interactions with the cluster of aromatic residues Phe204^5.47, Phe208^5.51, and Tyr292^7.43 (Figure 6). Of these, His256^6.51 is a key coupling residue as mutations in this residue uncouple AT₁R agonist binding and G protein activation.^44,45 While the C_α torsion remained constant, the Trp253^6.48 side chain torsion was stabilized in ~200 ns in the AngII-bound state and in 500 ns in the Olmesartan-bound states (Figure S10). The nature of CWxP conformational changes is not identical in eight agonist or G-protein bound active state GPCR structures in PDB (PDB ids: 2YDV, 3SN6, 5GLH, 4MQS, 5C1M, 3PQR, 5YNF). However, the motion observed in our simulation best matched with ET_BR and A₂ARs (Figure S10B). The hydrophobic interaction between Trp253^6.48 and Phe249^6.44 is also observed in the active-like states of NTSR1 and μ-opioid receptor.

Figure 6.

Conformation of toggle switch Trp253^6.48 of AT₁R at different time intervals.

Motion of Tyr215^5.58.

Tyr215^5.58 is located closer to the cytoplasmic end of TM5 and is part of the Yx7K^TM5 motif. Highly conserved throughout the GPCR family, the Tyr side chain at this location may serve the need for the polarity and the bulky hydrophobic ring system to ensure requirements for an efficient stabilizing motif in the active states of GPCRs.^70,71 In crystal structures of GPCRs, Tyr215^5.58 forms an H-bond with Arg126^3.50 side chain of the DRY motif in the active state. In AT₁R, this bonding is not observed throughout the simulation. Instead we observed a strong and continuous interaction of Tyr215^5.58 with Leu119^3.43 which helps TM3 and TM5 remain closer. But in the active state, the Tyr215^5.58 side chain moves toward the surface of the protein and forms π−π interaction with Trp2195.62 which makes the intracellular region of TM5 less flexible (Figure 7 and S11).

Figure 7.

Conformation of Tyr215^5.58 projecting inward in AngII-bound AT₁R and outward in olmesartan-bound AT₁R (A). The same pattern is also observed in active and inactive states of β₂AR (B).

DRY^3.51 Switch.

The Asp125^3.49-Arg126^3.50-Tyr127^3.51 located at the cytoplasmic end of TM3 in AT₁R is a highly conserved motif in GPCRs and often called an “ionic-lock”, restricting G protein interaction. In the current simulation we observed a distinct difference in orientation and H-bond networking of Arg126^3.50. In the inactive state, Arg126^3.50 is involved in stable intrahelical salt bridge with Asp125^3.49 which stabilizes the intracellular portion of TM3 to remain compact. This observation was also reported by Matsoukas et al.³⁸ Figure 8 depicts the conformation of this microswitch in the active state. The side chain of Arg126^3.50 moves toward TM6 and forms an H-bond with Asp237^6.32 (Figure S12). These rearrangements could be the beginning of opening of the groove for binding G-protein.

Figure 8.

Breakage of H-bond between D125^3.49-R126^3.50 and formation of H-bond between R126^3.50-D237^6.32 in AngII bound AT₁R. The D125^3.49-R126^3.50 H-bond is retained in the olmesartan-bound state.

NPxxY^TM7.53 Motif and Helix-8.

The Asn298^7.49-Pro299^7.50-Leu^7.51-Glu^7.52-Tyr302^7.53 motif is also highly conserved in GPCRs. Mutations targeting this motif reduce G protein activation by AT₁R.^61,72 During AngII activation, the NPxxY^TM7 moves toward the conserved Tyr215^5.58 forming a hydrogen-bond network with Asn298^7.51_, Tyr302^7.53, and Leu119^3.43 (Figure 9). A similar hydrogen-bond network was reported in activated structures of rhodopsin, β₂AR, and M₂R. The movements associated with NPxxY also contribute to opening the groove for G-protein binding.

Figure 9.

Secondary structure of H8 region of AT₁R-AngII (A) and AT₁R-ZD7511 (B). Loss of α helix (red color) from residue number 302 to 316 in panel A. Ensembles of structures of AT₁R-AngII and AT₁R-ZD7511 (C).

A ~15 Å movement of helix-8 toward the membrane is observed in AngII-bound AT₁R, but not in the olmesartan-bound AT₁R. The secondary structure of helix-8 changes from α helix to unstructured in AngII bound state (Figure 9). Another dynamic feature observed in most activated GPCRs including AT₁R is the increase of the Tyr302^7.53-Tyr112^8.53 distance, indicating TM7-helix-8 conformational rearrangement. This distance increased gradually in MD simulations after the first 20 ns and remained stable for the rest of the trajectory. Large changes in helix-8 conformation are reported in agonist-bound μ-opioid receptor.⁷³

DISCUSSION

Delineating the molecular mechanics involved in the opening of G protein binding cleft when AT₁R engages AngII is an important advancement. Without MDS analysis, visualizing the dynamics of receptor activation would be difficult while accounting for AngII-analog SAR data, molecular pharmacology data, receptor mutagenesis data, and 3D-structure data. MDS analysis indicated that AT₁R conforms to the general paradigm of conformational modulation by ligands. The agonists promote accumulation of the activated state and the inverse agonists depleted the basal state with minor structural changes (Figure 1). The activated state accumulated spontaneously in the constitutively active Asn111Gly^3.35 mutant suggesting that the wild-type AT₁R activation must involve agonist-dependent repositioning of Asn111^3.35 side chain. The Asn111^3.35-Asn295^7.46 bond is the prime constraining interaction in AT₁R. When AngII binds this bond is broken indirectly through strain induced by the van der Waals attraction of Ile288^7.39 toward the well-known agonism-specifying Phe8^AngII (Figures 3−5). This is a novel GPCR activation mechanism unlike the agonist trigger mechanisms in well-known GPCRs. For instance, rhodopsin photoactivation which serves as a prototype includes two triggering events, protonation breakage of the buried salt-bridge linking TM7 (retinal-Lys Schiff-base) to TM3 (counterion) and the resulting mechanical motion of functional microdomains.^12,74 In other ligand activated GPCRs different variations of a Lewis acid/base interaction chemical mechanisms operate. For instance, in β-adrenergic receptor, TM3-TM7 lock is a hydrogen bond betweenTyr^7.35-Asp^3.29 which interacts with protonated quaternary nitrogen in adrenergic ligands and networked with an Asn^7.32 residue.^75,76 In activating S1P1R, the lyso-phospholipid sphingosine 1-phosphate receptor, the zwitterionic phosphonate and primary amine groups respectively bond with Arg^7.35 and Glu^3.29.⁷⁷ Thus, release of the TM3-TM7 ionic-lock is suggested to be the first agonist activation step in GPCRs.⁷⁴ However, a general GPCR mechanism for breaking prime constraint remains unclear especially for peptide GPCRs until now.

The chemical nature of interactions of AngII and ARBs with the orthosteric pocket is fundamental to pharmacological outcomes, i.e., agonism and inverse-agonism. MDS analysis indicated that inverse-agonist (olmesartan) binding stabilizes multiple TM3-TM7 interactions and dampens intrinsic motion of helices and motifs of AT₁R (Table 1, Figures 6−9 and S6− S12). In addition, insights regarding the differing roles of critical residues are revealed. For example, in the activated state Arg167^ECL2 establishes multiple interactions with the backbone carbonyl groups of AngII, morphing its receptor bound conformation. However, Arg167^ECL2 binds the acidic pharmacophore groups at either ends of inverse-agonists, inhibiting residue rearrangements locally in the orthosteric pocket as well as globally in TM and cytoplasmic domains in the inactive-state (Figures 3 and S5).

Receptor bound conformation or the preferred conformation of a hormone peptide provides critical structural information for designing selective pharmaceutical scaffolds for a GPCR. The AngII structure in solution shows a bioactive γ-turn. Cyclic AngII analogs which restrain this conformation increase potency.^78,79 Specific spatial alignment of aromatic residues in AngII, Tyr-hydroxyl (Y4), His-imidazole (H6), and Phe-benzyl (F8) in this conformation maybe important.^80,81 Hormone peptides, AngII (DRVY⁴IH⁶PF⁸), AngIII (RVY⁴IH⁶PF⁸), and AngIV (VY⁴IH⁶PF⁸) which preserve the turn-motif are capable of modulating blood pressure via activating AT₁R. The pose of AngII in our MDS analysis is consistent with spatial alignment of Tyr (Y4), His (H6), and Phe(F8). Additionally, Phe8^AngII substitution with Ile and Gly, as well as its deletion (des-Phe8^AngII) affect binding and activation. As shown in Figure S7, Ang1–7(DRVY⁴IH⁶P) with either unstable or absence of the bioactive γ-turn conformation binds poorly and may activate AT₁R at nonphysiological concentrations which is also consistent with its poor MM/GBSA scores from IFD and MDS. Aromatic side chain of Phe8^AngII helps to project toward the Ile288^7.39. Substitution with nonaromatic residues (Ile, Gly, or Ala) produce a different conformation in which the side chain moves toward the ECL2 region.

The design of currently used ARBs to mimic the Y⁴IH⁶PF⁸ motif in AngII is confirmed by MDS analysis as well as in previous studies.^19,28,82 This knowledge will be useful in designing better AT₁R-selective scaffolds. Development of next generation AT₁R agonists and antagonists in terms of differences in functional selectivity and tissue tropism has important clinical implications. Given that a significant number of patients are currently prescribed ARBs for indications beyond blood pressure reduction, it is important to know the mechanism of action of these drugs at the receptor level.

To our knowledge, this is the first report detailing activating structural dynamics of a clinically important peptide hormone GPCR. There are ~135 human GPCRs responding to endogenous peptide/protein ligands. Static structures of peptide-activated NTR1, ET_BR, GLPR1 and opioid receptors^2,83,84 are documented. Despite this progress the specific interactions that trigger activation of peptide GPCRs are not easy to decipher. Hence MDS analysis starting with human AT₁R crystal structure described here should prove a useful model for gaining insights into the activation mechanism of peptide GPCRs and drug action. Energy-based models allow characterization of side-chain and backbone shift in the pocket as well as provide a 3D framework of determinants of agonism or inverse agonism essential for rational drug discovery. Important insights into agonist induced changes, establishing the path from ligand binding to G-protein activation can be gained.

In functioning GPCRs, thermodynamic motions within the protein produce a conformational landscape with discernible functional substates.^85,86 The MDS ensemble frames provide an integrated view of substate dynamics and time evolution of receptors from inactive to basal to activated states (Figures 2−9). We observed that the orthosteric pocket and cytoplasmic domain where signal transducer molecules converge are semi-independent. The cytoplasmic domain harbors conserved motifs DRY^TM3, Yx7K^TM5, NPxxY^TM7, and helix-8. These microswitches were originally described in rhodopsin,⁷¹ and functional studies in AT₁R demonstrate that a high degree of stability of functional motifs in the inactive state prevents spontaneous receptor activation. Mutagenesis of microswitches either constitutively activate the receptor or interferes with overall signaling. The impetus for activation of cytoplasmic microswitches is transmitted through a network of hydrogen bonds from the conserved CWxP^TM6 toggle-switch motif. Our MDS study shows critical positioning of this TM6-motif, sharing a hydrogen-bonded network between TM1-TM2-TM3-TM7 which includes the highly conserved Asp74^2.50 in AT₁R. The CWxP motif is part of the proposed sodium-ion binding site formed with Asn111^3.35 and Asn295^7.46.^12,13 Further, it interacts with a cluster of residues, His256^6.51, Phe204^5.47, Phe208^5.51, and Tyr292^7.43 which are associated with TM6 movement.^36,37 Because of high level of conservation of residues at TM6.44, TM^6.48, and TM^2.50, the role of the CWxP motif in AT₁R may be same as in other GPCRs. Recent experimental dynamics study in adenosine A₂R found that the interplay between TM^2.50 and TM^6.48 regulated structural dynamics localized at cytoplasmic motifs associated with receptor activation⁸⁷ which mimics similar observations made in rhodopsin activation.^88,89

Time-evolution of changes involving residues, Ile288^7.39, Tyr292^7.43, V108^3.32, Asn111^3.35, Asn295^7.46, Asn46^1.50, Asp74^2.50, and Trp253^6.48 suggests a sequential conformational transition model in which disruption of the TM3-TM7 interaction leads to movement of TM6 followed by reorientation of the TM1-TM2-TM3-TM7 interhelical network. Alternative nonlinear transition model as proposed earlier²⁷ with multiple kinetic scheme to account for native cellular and system-wide conditions is more realistic. Nonlinear transition models do include elements of a linear transition model as has been shown for rhodopsin.^90–92

CONCLUSION

Phe8^AngII engages Ile288^7.39 of AT₁R in van der Waals interaction which by pulling Tyr2927.43 and breaking H-bond with Val108^3.32 initiates AT₁R activation. This is a novel mechanism unlike those described in prototypical GPCR activation models. Breaking the Asn111^3.35-Asn295^7.46 bond relaxes the prime constraint; subsequently Asn295^7.46 interacts with Asp74^2.50 and Asn46^1.50 leading to AngII activation of AT₁R. This change in the molecular network leads to systematic change in the intracellular domain, and these changes may be a necessary for creation of a pocket for G-protein binding. In the olmesartan-bound inactive state, we observed a stable Asn111^3.35-Asn295^7.46 H-bond interaction and in addition Asn295^7.46 interacts with Asp74^2.50. Capturing and understanding the molecular details of activation of AT₁R by AngII and inhibition by ARBs could lead to better design of novel and more potent drugs targeting AT₁R which could effectively prevent AngII pathogenesis.

MATERIALS AND METHODS

Protein Preparation.

The starting coordinates of AT₁R PDB IDs: 4ZUD, 4YAY] were retrieved from Protein Data Bank (www.rcsb.org) and further modified for Glide docking calculations (Glide, Schrodinger, LLC, NY, USA).⁹³ For these calculations, energy minimization was applied to the protein using the Protein Preparation Wizard by applying an OPLS3 (Optimized Potentials for Liquid Simulations) force field.⁹⁴ Missing segments of loops were filled using Prime, Schrodinger, LLC, during the protein preparation.⁹⁵ Progressively weaker restraints applied to only the non-hydrogen atoms and structure refinement was done based on the default protocol of Schrodinger, LLC, NY, USA, as Glide uses the full OPLS3 force field at an intermediate docking stage and is claimed to be more sensitive to geometrical details than other docking tools. The most likely positions of hydroxyl and thiol hydrogen atoms, protonation states, and tautomers of His residues, and Chi “flip” assignments for Asn, Gln, and His residues were selected. Finally, energy minimizations of protein were performed until the average root-mean-square deviation of the non-hydrogen atoms reached 0.3 Å.

Ligand Preparation.

We retrieved AngII crystal structure from Protein Data Bank [PDB ID: 1N9V] and the structure was prepared using LigPrep package from Schrodinger, LLC, NY, USA (LigPrep, 2018)⁹⁶ by assigning appropriate bond order. AngII PDB structure was converted to mae format (Maestro, Schrodinger, LLC, NY, USA), geometrically optimized and partial atomic charges were computed. Then, at most, 32 poses were generated with different steric features for the subsequent docking study. Olmesartan structure was also prepared in Ligprep using the protocol described earlier.²⁸

Induced Fit Docking (IFD).

Induced fit docking (IFD) (Schrodinger, LLC, NY, USA) was used to dock AngII and olmesartan inside the ligand binding pocket of AT₁R. First, the ligand was docked into a rigid receptor model with scaled-down vdW radii. A vdW scaling of 0.5 was used for both the protein and ligand nonpolar atoms. A constrained energy minimization was carried out on the protein structure, keeping it close to the original crystal structure while removing bad steric contacts. Energy minimization was carried out using the OPLS3 force field with implicit solvation model. The Glide XP mode was used for the initial docking and ligand poses were retained for protein structural refinements. Prime (Schrodinger, LLC, NY, USA) was then used to generate the induced-fit protein−ligand complexes (Prime, Schrodinger, LLC, NY, USA).⁹⁷ Each of the structures from the previous step were subjected to sidechain and backbone refinements.⁹⁸ All residues with at least one atom located within 4.0 Å of each corresponding ligand pose were included in the Prime refinement. The refined complexes were ranked by Prime energy, and the 20 receptor structures within 30 kcal/mol of the minimum energy structure were passed through for a final round of Glide docking and scoring. In the final step, each ligand was redocked into the top 20 refined structures using Glide XP.

MM/GBSA.

Prime/MM-GBSA was used to predict the free energy of binding between the receptor and the AngII or olmesartan. The binding free energy (ΔG_bind) was calculated using Prime.⁹⁹ The binding free energy (G_bind) is then estimated using the following equation:

$$G_{\text{bind}}=\text{ER:L}-(\text{ER}+\text{EL})+G_{\text{solv}}+G_{\text{SA}}$$

where ER:L is energy of the complex, ER + EL is the sum of the energies of the ligand and the apo protein, using the OPLS3 force field, and ΔG_solv (ΔG_SA) is the difference between GBSA solvation energy (surface area energy) of complex and sum of the corresponding energies for the ligand and apoprotein.^100–102

MD Simulation.

Molecular dynamics (MD) simulations were carried out for AT₁R-AngII and AT₁R-olmesartan complexes using Desmond MD code and the OPLS3 force field⁹² for minimization of the system. We ran the MD simulation for 1 μS for each of the complexes. For AT₁R-AngII complex, we repeated the MD simulation six times for both the crystal structures solved for AT₁R (PDB id’s: 4ZUD and 4YAY) and two times for the AT₁R-olmesartan complex. Using the Desmond system builder, a 10 Å buffered orthorhombic system with periodic boundary conditions was constructed using a POPC lipid membrane^103–106 and an SPC explicit water solvent. The overall charge was neutralized by 0.15 mol/L NaCl. The simulations were performed in the NPT ensemble with number of elements, pressure, and temperature controlled. The temperature of 300 K and pressure of 1.013 bar were kept constant by coupling the system to a Berendsen thermostat and barostat, which are normal temperature and pressure mimicking the real life environment. An integration step of 2.0 was used, Coulombic interactions were calculated using a cutoff radius of 9.0 Å and long-range electrostatic interactions were calculated using the smooth particle mesh Ewald method.¹⁰⁷ Before each MD simulation, a default Desmond membrane protein relaxation protocol was applied.¹⁰⁸ To monitor protein stability and conformational fluctuations throughout the simulations, we computed backbone RMSDs after least-squares fitting to the starting structure. This was done for the entire protein and transmembrane α-helices. The trajectory was selected when the system is fully equilibrated and the RMSD reached its plateau. Statistical analyses are based on the structural ensemble from each cluster. Clustering of trajectory was performed using RMSD clustering tool of Schrodinger software package. Five clusters were generated and average structures were generated from each cluster. Additional structural analyses were further performed using these average structures. All the figures were generated using PyMol (The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC). The volume of the binding pocket was calculated using SiteMap, Schrödinger Release 2018–2: SiteMap, Schrödinger, LLC, New York, NY, 2018. Five representative structure of AT₁R-AngII were generated from the MDS trajectory and average value was reported.

¹²⁵I-AngII Competition Binding Analysis.

The detailed procedure is described in previous reports from our laboratory.^12,55 Ligand binding was analyzed using membranes prepared from HEK293T cells expressing wild-type HA-AT₁R as described earlier and suspended in membrane binding buffer (140 mMNaCl, 5.4 mMKCl, 1 mM EDTA, 0.006% bovine serum albumin, 25 mMHEPES, pH 7.4). A 10 μg of homogeneous cell membrane was used per well. Competition binding assays were performed under equilibrium conditions, with 2 nM radioligand (¹²⁵I-AngII purchased from Dr. Robert Speth, University of Mississippi) and concentrations of the competing ligand ranging between 0.04 and 1000 nM. Binding reaction was terminated by filtering the binding mixture through Whatman GF/C glass fiber filters, washed with buffer (20 mMsodium phosphate, 100 mMNaCl, 10 mM MgCl2, 1 mMEGTA, pH 7.2). The bound ligand concentration was determined as the counts/min (MicroBeta2 Plate Counter, PerkinElmer Life Sciences). Nonspecific binding was measured in the presence of 10⁵ M ¹²⁷I-AngII (Bachem). The binding kinetics were analyzed by the nonlinear curve-fitting program GraphPad Prism 5. The means ± SE for the IC₅₀ values were calculated.

Comparative Agonism by Intracellular Calcium Levels.

Calcium levels were measured using FLIPR_ calcium 5 assay kit (Molecular Devices) as described previously.^12,13,55 The HA-AT₁R expressing cells were seeded at a density of 100 000 cells/well in a 96-well clear bottom black plate. Following serum starvation, calcium-sensitive dye was added to the cells. The Flipper instrument was programmed in FLEX mode to add ligands (0.04 and 1000 nM concentration) to the cells and to monitor the fluorescence before and after adding the ligands. The dose−response curves were plotted and percent agonism was calculated assuming 100% stimulation by AngII.

Supplementary Material

Supplementari Information

Supporting Information

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jcim.8b00583.

Tables showing the effect of amino acid mutation on AngII binding, structure of AngII, snake plot of AT₁R showing important residues for AngII, RMSD of TM helices throughout MDS, superimposition of TM helices with active and inactive structures, superimposition of AngII and olmesartan in the orthosteric site of AT₁R, time scale of breakage of V108^3.32-Y292^7.43 H-bonding, time scale of breakage of N111^3.35-N295^7.46 H-bonds, experimental binding affinity of Ile8-AngII, Ala8-AngII, Gly8-AngII and its binding pose, comparison of W253^6.48 torsion angle of AT₁R-AngII and AT₁R-olmesartan, movement of Tyr215^5.58 in the AT₁R activation and movement of Arg126^3.50 in the AT₁R activation (PDF)

2D display of interactions of AngII with AT₁R(PDF) MDS trajectory videos 1, 2, 3, and 4 (MP4) PDB file of AT₁R bound with AngII (PDB)