Discovery of a new generation of angiotensin receptor blocking drugs: Receptor mechanisms and in silico binding to enzymes relevant to SARS-CoV-2

Abstract

The discovery and facile synthesis of a new class of sartan-like arterial antihypertensive drugs (angiotensin receptor blockers [ARBs]), subsequently referred to as "bisartans" is reported. In vivo results and complementary molecular modelling presented in this communication indicate bisartans may be beneficial for the treatment of not only heart disease, diabetes, renal dysfunction, and related illnesses, but possibly COVID-19. Bisartans are novel bis-alkylated imidazole sartan derivatives bearing dual symmetric anionic biphenyl tetrazole moieties. In silico docking and molecular dynamics studies revealed bisartans exhibited higher binding affinities for the ACE2/spike protein complex (PDB 6LZG) compared to all other known sartans. They also underwent stable docking to the Zn^{2 +} domain of the ACE2 catalytic site as well as the critical interfacial region between ACE2 and the SARS-CoV-2 receptor binding domain. Additionally, semi-stable docking of bisartans at the arginine-rich furin-cleavage site of the SARS-CoV-2 spike protein (residues 681-686) required for virus entry into host cells, suggest bisartans may inhibit furin action thereby retarding viral entry into host cells. Bisartan tetrazole groups surpass nitrile, the pharmacophoric "warhead" of PF-07321332, in its ability to disrupt the cysteine charge relay system of 3CLpro. However, despite the apparent targeting of multifunctional sites, bisartans do not inhibit SARS-CoV-2 infection in bioassays as effectively as PF-07321332 (Paxlovid).

Keywords: Angiotensin receptors; Animal AT1 receptor studies; Azil, Azilsartan; BisA, 4-Butyl-N,N0-bis{[[20-(2H-tetrazol-5-yl)]biphenyl-4-yl] methyl}imidazolium bromide; BisB, 4-Butyl-2-hydroxymethyl-N,N0-bis{[20-(2H-tetrazol-5-yl)- biphenyl-4-yl]methyimidazolium bromide; BisC, 2-Butyl-4-chloro-5-hydroxymethyl-N,N0-bis{[20-(2H-tetrazol- 5-yl)biphenyl-4-yl]methyl}imidazolium bromide; BisD, 2-Butyl-N,N0-bis{[20-(2H-tetrazol-5-yl)biphenyl-4-yl]methyl} imidazolium bromide; Bisartan NMR studies; Bisartans tetrazole; Cande, Candesartan; DIZE, deminazene aceturate; Docking RBD/ACE2 studies; EXP3174, losartan carboxylic acid; Epro, Eprosartan; Irbe, Irbesartan; Lo, Losartan; Olme, Olmesartan; SARS-CoV-2 spike/ACE2 complex blockers; Sartans; Telmi, Telmisartan; Val, Valsartan.

Graphical abstract

Fig. 1

Structure (upper left) and electrostatic charge distribution (lower image) of bisartan A (BisA). Negatively charged regions = red; neutral to positive regions = green to blue. Charge distribution calculated using the RM1 semiempircal method (UHF calculation, total charge = −2.0e, geometry-optimized structure @ 0.1 kcal/mol-Å gradient).

Fig. 2

(A) Inhibitory effect of the novel bisartan B salts BV6(K⁺)₂, BV6(Na⁺)₂, and BV6(TFA)₂ on ANG II-induced vasoconstriction responses in rabbit iliac arteries. To determine if bisartans could mimic the effect of ARBs, rabbit iliac artery rings were incubated with BV6(K⁺)₂, BV6(Na⁺)₂, or BV6(TFA)₂ and constricted using an ANG II dose–response. Incubation with Bisartans resulted in potent inhibition of vasoconstriction at: BV6(K⁺)₂ doses [10^−10.5 M to 10^−7.0 M]; BV6(Na⁺)₂ doses [10^−11.0 M to 10^−6.0 M] and BV6(TFA)₂ doses [10^−11.0 M to 10^−6.5 M]. Data represented as mean$\pm$SEM, significance represented at Table 1A. (B) Receptor switching from an agonist-induced state to a desensitized inverse agonist state. The binding of submaximal doses of ANG II to its receptor induces G protein binding and dimerization of the receptor, a concerted mechanism of positive cooperativity (Hill coefficient n_H > 1), resulting in an increase in agonist affinity and consequent amplification of the contractile response. Different analogs may invoke different levels of cooperativity (increasing the affinity) when compared to other analogs (e.g. weak or partial agonists). The maximum response may be limited by the available supply of G protein, without which the mode of receptor binding of ANG II at supramaximal doses changes to inverted state and induces negative cooperativity (n_H < 1) synonymous with inverse agonism (tachyphylaxis). Thus, at high doses, ANG II becomes an inverse agonist, as do other partial agonists . Inverse agonists, such as ARBs, sarilesin, and angiotensin “antipeptides”, are unable to activate the receptor, and bind to inverted state forming a salt bridge with Arg167 (shown as - -X). The resulting insurmountable “inverted” state of the angiotensin receptor engenders smooth muscle relaxation (vasodilation, via an alternative second messenger) for prolonged periods. This receptor lockdown effect may be due to a slow rate of dissociation of the second messenger.

Fig. 3

Left Panel: Results of global docking of 15 ARBs (including the 4 bisartans) to the ACE2-SARS-CoV-2/RBD complex (PDB 6LZG). The docking volume comprised a non-periodic (walled) cuboid cell with boundaries 8 Å distant from any target atom. Docking was carried out using AutoDock VINA (YAMBER3 force field; https://www.Yasara.org) with 100 trials per ARB. Best poses for the docked ARBs are superimposed in this image to show preferential binding regions (domains). Five binding domains (BDs) were observed (I-V), with most ligands occupying BD #IV. Bisartans BisB and BisC docked (though comparatively weakly) into the region (BD #III) corresponding to the interface between the ACE2 receptor and the SARS-CoV-2 RBD, suggesting these two bisartans might serve as potential antagonists of virus adsorption to host tissues expressing the ACE2 receptor. Right Panel: Rotation of the ACE2-RBD complex 90^O.

Fig. 4

Docking of 15 ARBs (including the 4 bisartans) to the zinc pocket of the SARS-CoV-2 RBD-ACE2 complex (PDB 6LZG). AutoDock VINA binding energies are illustrated in the central plot with columns. Green columns = the 4 Bisartans; blue columns = other ARBs. Dissociation constants are represented by the orange line with square markers. Based on the VINA docking metric the order of binding was: BisA > BisC > BisD > Telmi > Olme > Azil > BisB > Lo > Cande > Irbe > Dize > Epro > Exp3174 > Lisinopril > Val.

Fig. 5

(A) Cumulative stacked line graphic depicting the relative frequency of ACE2 residue contacts for each of the 15 ARBs docked into the Zn²⁺ pocket of the ACE2 receptor. Dominant amino acid contacts with ARBs included Asp, Glu, His, Leu, Phe, Ser, and Trp. Residues with the least contacts included Cys, Gln, Ile, Lys, Pro, Val. Bisartans A, B, and D were the only ARBs contacting the Zn²⁺ cofactor. (B) 2D interaction ligand-receptor diagrams of BisA, B, C, and D in the Zn²⁺ pocket of the ACE2 receptor (6LZG). Colour key for chemical interactions: green shading = hydrophobic regions; Blue shading = hydrogen bond acceptor; White dashed arrows = hydrogen bonds; Grey parabolas = accessible surface for large areas; Grey residues = “generic” van der Waals contact (non-hydrophobic, non-H-bond); Broken thick line around ligand = accessible surface; Size of residue ellipse = strength of the contact; 2D distance between residue label and ligand = proximity.

Fig. 6

(A) Docked chlorinated bisartan BisC to the interfacial region between the ACE2 receptor (Van der Waals surface; yellow) and the SARS-CoV-2/RBD (PDB 6LZG). This pose resulted from the global docking of BisC to PDB entry 6LZG using AutoDock VINA. The docking domain comprised cuboidal cells with non-periodic (wall) boundaries 8 Å from any target atom. (B) The BisC binding motif primarily involved pi/pi (red lines), pi/cation (blue lines), and hydrogen bonding (thick dashed lines) interactions with the SARS-CoV-2 RBD residues Tyr505, Arg403, Phe456, and Tyr421 (yellow spheres). The binding of BisC to the ACE2 interfacial region was mainly dominated by hydrophobic interactions (green lines) to Pro389, Leu29, Val93, Lys26, and Asp30; and secondarily by pi/cation interactions (blue lines) to Arg393. (C) The bound BisC molecule was moderately stable and remained bound inside the interfacial zone (ave. RMSD = 2.46 Å) in MD simulations run out to at least 41 ns (NPT ensemble, 0.9% saline, 311°K). The green shading indicates the time period over which the RMSD values were calculated. (D) Upper panel: Frame captures from MD simulations of bisartan-A, B, C, D/ACE2 complexes. Bound bisartans are indicated by red spheres. MD simulations were run with periodic boundaries for approximately 40 ns at 311oK in physiological saline (water and NaCl ions are hidden for clarity). All bisartans remained stable inside the zinc pocket of ACE2, although Bis B, C, and D exhibited greater motion (RMSD[ave] = 3.2 Å, 2.1 Å and 2.6 Å respectively) compared to BisA (RMSD[ave] = 1.22 Å). ACE2 molecular surfaces are shown. Lower panel: RMSD values as a function of MD simulation time. The stability of the complexes is retained for at least 40 ns.

Fig. 7

Docking of bisartans and selected common antiviral drugs to the proteo-catalytic Cys145 site of the main 3CLpro SARS-CoV-2 protease (AutoDock VINA results). BisA, B, C, and D are imidazoles containing anionic dual branching biphenyl tetrazole rings. PF-07331332 is the active antiviral component of Pfizer’s Paxlovid agent. All of the bisartan analogs (BisA-BisD) exhibited stronger binding scores compared to PF-07321332. Note that PF-07321332 remained in the non-covalently bound form for these docking exercises. Results of docking of selected antiviral ligands to the catalytic pocket of the 3CLpro main protease of SARS-CoV-2.

Fig. 8

Comparison of docking motifs of BisA and Pfizer’s PF-07321332 antiviral drug with the Cys145 catalytic site of the main 3CLpro protease of SARS-CoV-2. (A) Details of BisA-3CLpro residue interactions. Ligand Atom Colors: C = cyan; N = blue; O = red. Tetrazole#2 of BisA interacted strongly with the His163 ring via pi-pi resonance (red lines). Tetrazole#1 exhibited no strong interactions, however, the biphenyl groups showed significant hydrophobic interactions (green lines) with several nearby residues, particularly Pro168. BisA did not appear to undergo direct interactions with Cys145, although BisA tetrazole#2 was proximal (3.74 Å) to the sulfur atom (large yellow sphere) and therefore probably interacted electrostatically. (B) As was the case for BisA binding, no direct interactions were observed between PF-07321332 and Cys145. However, it is noted that the –CN nitrile group was located proximal to the Cys145 sulfur, which is a potential site of covalent interaction with the ligand. Ligand binding was also marked by a strong hydrogen bond between the nitrile N and the Ser144 –OH. (C) Superimposed view of bound BisA and PF-07321332. (D) Frame captures at 0, 26, 40 and 68 ns from independent MD simulations of the BisA-3CLpro complex and the PF-07321332-3CLpro complex (drugs bound in the Cys145 catalytic pocket). MD conditions for both runs were: 311°K, NVT ensemble, 0.9 wt% sodium chloride (physiological saline), AMBER14 force field parameters. Bound PF-07321332 was unstable and exited the catalytic pocket beginning at about 20 ns, with complete extraction by about 25 ns. In contrast, BisA remained stably bound in the pocket for the duration of the 70-ns MD run. (E) RMSD values for: (1) BisA-3CLpro (blue line) and (2) PF-07321332-3CLpro complexes as a function of MD simulation time out to approximately 70 ns.

Fig. 9

Comparison the docking behaviour of the hypothetical compound BisNitrile against bisartans A-D, Ombitavir and PF-07321332. Docking was carried out using AutoDock VINA (500 runs per ligand) targeting the 3CLpro catalytic pocket of SARS-CoV-2 (PDB 6Y2F).

Fig. 10

(A) Binding of bisartan B to ACE2 open channel near to zinc-binding motif. The highlighted interactions are: i) π stacking between Tyr202 and an aromatic group of bisartan, ii) H-bond between OH group of Tyr202 and N atom of bisartan and iii) salt bridges between the second tetrazolate with Arg514. The distance between the second tetrazole and Lys 562 is greater than 5.5 A. It is difficult to support any contact between them. (B) Binding of bisartan A by two Arginines in wild type furin cleavage site (PRRARS) of the spike protein receptor-binding domain. One tetrazole interacts with two Arg residues (R682 and R683 which is closest to the tetrazole) located in the cleavage site cavity. The third Arg (R685) is oriented in the opposite direction. (C) Binding of bisartan A by three arginines in P681R mutated furin cleavage site (RRRARS) of the spike protein-binding domain. One tetrazole interacts with the three arginine residues R681, R682, and R683, and the second tetrazole with His655. (D) Structures of ARBs Losartan (surmountable) and EXP3174 (insurmountable). One-dimensional graphs and (E) CRS in ANG II and (F) angiotensin receptor blocker binding to the receptor. (G) Interactions between receptor and Olmesartan were determined from the crystal structure. The critical interactions of Olmesartan are with Arg167, Trp84, Tyr35, and Lys199 residues of AT1 receptor. (H) Amino acids at positions 452 (Leu) and 478 (Thr) in spike RBD were replaced by arginines in the delta variant located at the nearest negatively charged residues (Glu22, Asp38, and Glu35) in ACE2. They also depict the bound bisartan (best bitetrazole binder A) molecule and the zinc atom.

Fig. 11

(A) Binding of ARB to angiotensin receptor. The binding of ARB results in a strong salt bridge (hatched bond) with Arg167 (and Lys199), preventing binding of the receptor to the G protein (which in turn prevents receptor dimerization), and thereby locking the receptor in the inverted state (Fig. 1B). An alternative second messenger (possibly B-arrestin or a different G protein) binds to the intracellular domain of the receptor leading to inverse agonism (relaxation). Binding and activation of ANG II receptor. (B) The conformation of ANG II determined by various spectroscopic techniques is characterized by the presence of a CRS and clustering of the three aromatic rings, including a Phe:His ring:ring interaction. The CRS relays the negative charge at the C-terminus of ANG II to the Tyr hydroxylate, distributing the charge across the CRS, and steering the ANG II molecule to align with two corresponding positive charges (Arg167 and Lys199) on the receptor. (C) Salt bridge formation (hatched bond) with Arg167 and Lys199 enables an induced fit of key peptide and receptor-based groups, allowing Tyr35 to H-bond with the imidazole N of His6 of ANG II. This releases the salt bridge between Tyrosinate4 of ANG II and Arg167. Agonist activity may derive from the exchange of the hydroxylate of Tyr4 of ANG II with that of Tyr35 of the receptor, releasing the intramolecular charge relay interactions in ANG II and replacing it with an intermolecular interaction with the receptor. These dynamic interactions are transduced by a cooperative mechanism involving G protein binding and receptor dimerization leading to amplification (n_H > 1, Table 1B) of the contractile response. Simultaneous interaction of the aromatic ring of Tyr35 with the rings of Phe8 and His6 of ANG II may be a critical factor in this receptor triggering process, perhaps by aligning Tyr35 for bonding to the imidazole N of ANG II [NOTE: an aromatic ring has a quadrupole moment which allows it to form a slipped parallel plate or perpendicular plate electrostatic interaction with another ring]. Accordingly, sarilesin, which lacks the necessary aromatic ring for electrostatic interaction with the Tyr35 ring, is unable to exchange its tyrosinate with that of Tyr35 and elicit the response. Instead, it maintains a strong salt bridge anchor with Arg167 rendering sarilesin insurmountable. In contrast, the nonpeptide losartan and the peptide analog sarmesin are surmountable antagonists because they cannot form this salt bridge and form a weaker ion:dipole bond with Arg167, as does the ANG II TyrOH when released by Tyr35. (D) Binding of bisartan to the active site zinc atom of ACE2 (E) and Arg167 of the AT1 receptor.

Fig. 12

Bisartans are in the form of tetrazole potassium salts. (A) Two-step facile synthesis of bisartan A (B) Facile two-step high yield synthesis of bisartan A, B, and D. Bisartan BisC (bis alkylated Losartan), designed and included in the in silico study for docking comparison, is under preparation.

Fig. 13

Interaction of bisartan negative tetrazoles with positive guanidino group of AT1 receptor (Arg167). 4-butyl imidazole bisartan (A) and 2-butyl imidazole bisartan, dialkylated bisartan (B) interacting with Arg167 of AT1 receptor.

Fig. 14

CRS Protease mechanisms. SARS-CoV-2 spike protein can be cleaved by furin at positions 685–686 and 3CLpro at glutamine positions through CRS mechanisms. The catalytic center of furin is the triad Asp – His – Ser and for 3CL protease is the dyad Cys145 – His41. [James J. Neitzel, Enzyme Catalysis: The Serine Proteases, Nature Education 3(9):21 (2010)].

Fig. 15

COVID-19 assay comparing NTV (PF07321332), 2-BDN (BisD nitrile), 2-BV6 (BisD tetrazole). Vero cells infected with SARS-Cov-2 were treated with the different compounds. The cytopathic effect (CPE) was measured under an inverted microscope and expressed as a percentage to the CPE observed in untreated cells. The 2-BDN and 2-BV6 bisartans bear the butyl group at position 2 of the imidazole ring as in losartan. Details of the study are seen in the supplementary material.