MOLECULAR DOCKING ANALYSIS OF CLINICALLY DEPLOYED RESPIRATORY DRUGS ON BETA-2 ADRENERGIC AND MUSCARINIC ACETYLCHOLINE RECEPTORS: TOWARDS THE RATIONAL DESIGN OF NOVEL DUAL-TARGET LIGANDSAbstract
Respiratory diseases such as asthma, chronic obstructive pulmonary disease (COPD), and cystic fibrosis remain critical global health challenges, necessitating the development of novel therapeutics targeting key airway receptors. This study employed structure-based virtual screening (SBVS) and molecular docking to investigate ligand interactions with beta-2 adrenergic receptors (β2-AR) and muscarinic acetylcholine receptors (mAChRs), pivotal regulators of bronchial tone and mucus secretion. Four high-resolution crystal structures of β2-AR (Protein Data Bank (PDB) IDs: 2RH1, 5D5A, 6PS2, 8GG0) and mAChRs (PDB IDs: 5CXV, 6OIJ, 6OL9, 6WJC) were prepared by removing non-protein moieties, optimizing protonation states, and energy minimization. Six clinically relevant ligands formoterol, ipratropium bromide, levalbuterol, propranolol, salbutamol, and tiotropium were energy-minimized and docked using GOLD (Genetic Optimization for Ligand Docking) software centered on active sites derived from co-crystallized ligands. Key interactions, including hydrogen bonds with conserved residues (e.g., β2-AR: Asp113, Ser203; mAChRs: Tyr506, Asn507), were identified using molecular docking analyses. Tiotropium and ipratropium bromide exhibited high binding affinities for mAChRs, aligning with their established anticholinergic efficacy. For β2-AR, long-acting agonists formoterol and salbutamol showed strong binding, consistent with their bronchodilatory roles. Propranolol, a β-blocker, displayed competitive antagonism at β2-AR, highlighting subtype selectivity challenges. These findings pave the way for optimizing dual-target ligands to address multifactorial respiratory pathologies while minimizing off-target effects.
Keywords:
structure-based virtual screening; molecular docking; beta-2 adrenergic receptors (β2-AR); muscarinic acetylcholine receptors (mAChRs).
INTRODUCTION
Respiratory diseases, such as asthma, chronic obstructive pulmonary disease (COPD), and cystic fibrosis, represent a significant global health burden.1 These conditions are characterized by airway inflammation, bronchoconstriction, and excessive mucus production, leading to impaired lung function and reduced quality of life. Despite advancements in therapeutic interventions, there remains an urgent need for more effective and targeted treatments. The β2-adrenergic receptor (β2-AR) and muscarinic acetylcholine receptors (mAChRs) are two critical classes of G-protein-coupled receptors (GPCRs) that play pivotal roles in regulating airway smooth muscle tone and bronchial responsivenes.2 β2 receptor drugs are designed to mimic the natural effect of epinephrine and norepinephrine hormones on the body. The short-acting β2 agonists (SABAs) approved by the FDA (Food and Drugs Administration) are albuterol, levalbuterol, metaproterenol, and terbutaline. They provide almost instant relief of asthma symptoms and are used as rescue medicines for treating acute asthma attacks. Long-acting β2 agonists (LABAs), such as salmeterol and formoterol cause cardiac dysrhythmias and muscle cramps in normal doses in long-term use.3,4 Short-acting β2-agonists (SABAs; e.g., albuterol, levalbuterol) and long-acting β2-agonists (LABAs; e.g., formoterol, salmeterol) bind to β2-AR, activating Gs protein-coupled adenylate cyclase. This increases cyclic adenosine monophosphate (cAMP) production, leading to airway smooth muscle relaxation and bronchodilation. Conversely, anticholinergics like ipratropium bromide and tiotropium antagonize mAChRs (primarily M3 subtypes), inhibiting Gq-mediated phospholipase C activation. This reduces intracellular calcium, suppressing bronchoconstriction and mucus hypersecretion. While β-blockers (e.g., propranolol) antagonize β2-AR and may exacerbate asthma, they serve as antidotes for β2-agonist overdose. β-blockers can cause exacerbation of symptoms in patients with asthma because they restrict the protective bronchodilatory effect of the natural epinephrine.5 β2 agonists are mostly given in inhaled forms, either as a spray or nebulizer, to limit the systemic adverse effects and reduce the delivered dose to target only the respiratory system.6 Propranolol and esmolol were considered antidotes for β2-agonist overdose as they antagonize the β2-receptor activation caused by them.7,8
Muscarinic acetylcholine receptors (mAChRs) can be classified as five subtypes M1-M5, in which the five kinds of receptors M1, M3, and M5 function with the Gq family and receptors M2 and M4 with the Gi/Go family of G proteins.9-11 M3 receptors were detected in submucosal glands and airway smooth muscle cells and M4 is distribute in alveolar walls.12-14 The subtype M3 is related to airway smooth muscle contraction.15,16 The M1 and M3 receptors played a significant role in regulating mucus secretion through vagal nerve stimulation.17 Sputum production is otherwise known as airway mucus hypersecretion, that is usually found in COPD and asthma patients. The cholinergic system mainly controlled the production of mucus.18 COPD and asthma lead to bronchoconstriction, which can be alleviated by appropriate treatment.
Structure-based virtual screening (SBVS) leverages 3D protein structures to computationally screen ligand libraries for binding affinity. Molecular docking predicts ligand-receptor binding geometries and scores interactions using empirical/scoring functions (e.g., GoldScore). While efficient for initial screening, SBVS has limitations: static protein models ignore dynamics, scoring functions may misrank ligands, and solvation/entropy effects are often simplified.19-24
This study focuses on the detailed preparation for each selected target; four high-resolution crystal structures were retrieved to advance respiratory therapeutic research. Specifically, crystal structures of β2-AR (Protein Data Bank (PDB) IDs: 2RH1, 5D5A, 6PS2, 8GG0) and mAChRs (PDB IDs: 5CXV, 6OIJ, 6OL9, 6WJC). Clinically established drugs were selected to (i) validate our SBVS/docking protocol against known binding profiles, (ii) establish a baseline for future novel compound screening, and (iii) explore off target effects (e.g., unexpected affinity of tiotropium for the β2-AR receptor). While confirmatory, this approach ensures methodological rigor prior to probing unexplored chemical space. In conclusion, the integration of SBVS and molecular docking with advances in structural biology and computational methodologies holds great promise for the discovery of novel respiratory therapeutics targeting β2-AR and mAChRs.
METHODOLOGY
Protein preparation
The molecular docking studies were conducted using four protein structures of the β2-AR and four structures of mAChRs obtained from the PDB. The PDB IDs for β2-AR were 2RH1 (inactive, antagonist-bound), 5D5A (active, agonist-bound), 6PS2 (nanobody-stabilized) and 8GG0 (partial agonist-bound),while the PDB IDs for mAChRs were 5CXV (M1 inactive), 6OIJ (M2 inactive), 6OL9 (M4 antagonist-bound), 6WJC (M3 active-like). These structures were selected based on their high resolution and relevance to respiratory therapeutics.
Protein retrieval and cleaning: the 3D structures of the receptors were downloaded from the PDB. All non-protein molecules, such as water molecules, ions, and co-crystallized ligands, were removed using molecular visualization software Discovery Studio Visualizer, version 2024 (Dassault Systèmes, Waltham, USA, 2024). Missing residues and loops were modeled using homology modeling tools if necessary.
Protein optimization: hydrogen atoms were added, and the protonation states of amino acid residues were optimized at physiological pH (7.4) using GOLD (Genetic Optimization for Ligand Docking, Cambiaso Risso, United Kingdom) docking software, version 4.0. The energy of the protein structures was minimized using molecular mechanics force fields to ensure stable conformations.
Ligand preparation
Six compounds were selected for docking studies based on their pharmacological relevance to respiratory therapeutics: formoterol, ipratropium bromide, levalbuterol, propranolol, salbutamol, and tiotropium (Figure 1). These ligands were prepared as follows:
2D chemical structure of the drug molecules formoterol, ipratropium bromide, levalbuterol, propranolol, salbutamol, and tiotropium
Ligand retrieval: the 3D structures of the ligands were obtained from the PubChem database or generated using chemical drawing tools (e.g., ChemDraw).
Ligand optimization: the ligands were energy-minimized using molecular mechanics force fields to ensure stable conformations. Tautomers and ionization states were generated at physiological pH (7.4) using tools like LigPrep, v3.5.9 (Schrödinger, LLC, USA) or Open Babel v3.1.1.
File format conversion: the optimized ligands were converted into the appropriate file formats for docking simulations. Docking simulations were performed using the widely used GOLD docking software to explore ligand conformations and orientations. The docking poses were scored using Gold Score scoring functions to rank the ligands based on their predicted binding affinities. The top ranked poses were visually inspected for their interactions with key residues in the binding pocket.
In silico prediction of toxicity
The toxicities of the six respiratory drugs were predicted using ProTox-3.0, a freely available in silico toxicity prediction web server. The chemical structures were developed using the smiles into the ProTox-3.0 web server for the analysis.25,26 The predictions involved are of the machine learning algorithm-based models. The platform is classified into four different groups namely, organ toxicity (one model), toxicity endpoints (four models), Tox21 nuclear receptor signaling pathways (seven models) and Tox21 stress response pathways (four models). Acute toxicity prediction for oral administration within 24 h was presented as lethal dose 50 or LD50 (mg kg-1 body weight (mg kg-1 bw)). LD50 was used for further analysis of toxicity classes, which are defined according to the globally harmonized system (GHS) of classification of labelling of chemicals (LD50 values are given in mg kg-1): class I: fatal if swallowed (LD50 ≤ 5); class II: fatal if swallowed (5 < LD50 ≤ 50); class III: toxic if swallowed (50 < LD50 ≤ 300); class IV: harmful if swallowed (300 < LD50 ≤ 2000); class V: may be harmful if swallowed (2000 < LD50 ≤ 5000); class VI: non-toxic (LD50 > 5000).
RESULTS AND DISCUSSION
Docking results for beta-2 adrenergic receptors (β2-AR)
The results of docking studies for six compounds targeting beta-2 adrenergic receptors (β2-AR), performed using four different PDB structures (Figure 2), are summarized in Table 1. The docking analysis was performed using the Gold fitness score (in kcal mol-1), which indicates the binding affinity between the ligand (compound) and the receptor. Lower (more negative) scores suggest stronger binding affinity. The key observations of β2-AR reveals that tiotropium is primarily known as a muscarinic receptor antagonist used in COPD management. Other known binders to the same target, especially those with greater specificity, raise some interesting possibilities, with scores ranging from 51.16 to 58.45 kcal mol-1. This suggests that tiotropium has a strong interaction with β2-AR, despite being primarily a mAChR antagonist. Formoterol, a known β2-AR agonist, also demonstrates strong binding affinity, with scores ranging from 47.06 to 54.90 kcal mol-1. This aligns with its pharmacological role as a β2-AR agonist. Propranolol, a β-blocker, shows moderate binding affinity, with scores ranging from 35.93 to 49.06 kcal mol-1. This is consistent with its mechanism of action as a competitive antagonist at β-adrenergic receptors. Levalbetrol and salbutamol, both β2-AR agonists, show moderate binding affinity, with scores ranging from 32.94 to 38.34 kcal mol-1 and 33.32 to 37.40 kcal mol-1, respectively. Ipratropium bromide shows the lowest binding affinity across all PDB structures, with scores ranging from 22.83 to 29.59 kcal mol-1. This outcome is expected, as ipratropium mainly acts as a mAChR antagonist and not as a β2-AR ligand.
3D superimpose and interaction of the six drug molecules in the active region of the β2-AR protein
Binding mode analysis
The binding mode analysis involves understanding how each compound interacts with the active site of β2-AR (Figures 3-6). Key interactions include hydrogen bonding, hydrophobic interactions, and electrostatic interactions. The compound tiotropium shows the strongest binding affinity across all PDB structures. It likely forms hydrogen bonds with residues like Ser203 and Ser207 in the active site. It also engages in hydrophobic interactions with residues like Phe290 and Trp109. Formoterol shows strong binding affinity, consistent with its role as a β2-AR agonist. It likely forms hydrogen bonds with Ser203 and Asn293. It may interact with hydrophobic residues like Phe290 and Val114. Propranolol shows moderate binding affinity, consistent with its role as a β-blocker. It likely forms hydrogen bonds with Ser203 and Asn293. This may engage in hydrophobic interactions with Phe290 and Val114. Levalbetrol and salbutamol show moderate binding affinity, consistent with their roles as β2-AR agonists. They likely form hydrogen bonds with Ser203 and Asn293. Ipratropium bromide has the weakest binding affinity, as it is not a β2-AR ligand. It may exhibit non-specific interactions with the receptor, such as weak hydrophobic interactions.
Ramachandran plot analysis
The Ramachandran plot analysis provided focuses on β2-AR, a class of G protein-coupled receptors (GPCRs) involved in mediating the effects of catecholamines like adrenaline and noradrenaline. The analysis includes four specific protein structures identified by their PDB IDs: 2RH1, 5D5A, 6PS2, and 8GG0 (Figure 7). Below is a detailed explanation of each protein based on the Ramachandran plot data and general knowledge about β2-AR.25 The 2RH1 structure represents an early high-resolution crystal structure of the human β2 AR, stabilized with an inverse agonist. The plot shows a majority of residues in favored regions, indicating a stable fold. Clusters in glycine and proline regions suggest flexibility, particularly in loops or turns, which is typical for GPCRs. Outliers may reflect regions affected by the fusion protein or crystallization conditions. The 5D5A entry corresponds to a β2-AR structure in complex with a nanobody and an agonist, captured in an active-like state. The plot shows that most residues lie in favored regions, reflecting a well-defined structure. The glycine and proline clusters are more pronounced, possibly due to conformational changes induced by agonist binding. Outliers might indicate strain from the nanobody interaction.
Ramachandran plot analysis for β2-adrenergic receptors (β2-AR) PDB IDs: 2RH1, 5D5A, 6PS2, and 8GG0
The 6PS2 structure features β2-AR bound to a G protein and an agonist, representing a fully active state. The plot shows a high proportion of residues in favored regions, consistent with a stable active conformation. Increased clustering in flexible regions (glycine/proline) may reflect dynamic movements during G protein coupling. Outliers could highlight areas under stress from this interaction. The 8GG0 entry is a more recent β2-AR structure, likely involving a novel ligand or stabilization method. Most residues fall in favored regions, indicating structural integrity. The glycine and proline clusters suggest flexibility, potentially due to a unique ligand or experimental condition. Outliers may point to regions of interest for further refinement.
Docking results for muscarinic acetylcholine receptors (mAChRs)
The structure-based docking studies for mAChRs used four different PDB IDs: 5CXV, 6OIJ, 6OL9, and 6WJC (Figure 8). The docking analysis was performed using the Gold fitness score (in kcal mol-1), which is a measure of the binding affinity between the ligand (compound) and the receptor. Lower (more negative) scores indicate stronger binding affinity (Table 2).
The Gold fitness scores of six compounds against four muscarinic acetylcholine receptor (mAChR) structures (5CXV, 6OIJ, 6OL9, 6WJC)
3D superimpose position and interaction of the six drug molecules in the active site region of the mAChRs protein
Key observations: the compound tiotropium consistently shows the highest Gold fitness scores (strongest binding affinity) across all four PDB structures, with scores ranging from 62.46 to 72.00 kcal mol–1. This suggests that tiotropium has a strong interaction with mAChRs. Ipratropium also demonstrates relatively high binding affinity, with scores ranging from 48.76 to 58.80 kcal mol–1.
Formoterol, levalbuterol, propranolol, and salbutamol show moderate to lower binding affinities compared to tiotropium and ipratropium. Their scores range from 37.40 to 56.65 kcal mol-1. Among the four PDB structures, 6WJC generally shows higher Gold fitness scores for most compounds, indicating stronger binding interactions with this particular receptor conformation.
Binding mode analysis
The binding mode images (Figures 9-12) illustrate how each compound interacts with the active site of the mAChRs. Key interactions might include hydrogen bonds, hydrophobic interactions, and π-π stacking, depending on the compound and receptor conformation. Tiotropium and ipratropium, being mAChR antagonists, likely occupy the orthosteric binding site, preventing acetylcholine binding. Formoterol, levalbetrol, and salbutamol are β2-adrenergic agonists and may show weaker or non-specific interactions with mAChRs. Propranolol, a β-blocker, might exhibit moderate binding due to its structural similarity to some mAChR ligands. Tiotropium shows the strong hydrogen bonds interactions with Asp103 and Tyr506. Hydrophobic interactions with Trp613 and Phe197. Also possible π-π stacking with aromatic residues.
2D interaction of drug molecules with key active site residues of the mAChR (PDB ID: 6OIJ)
2D interaction of drug molecules with key active site residues of the mAChR (PDB ID: 6OL9)
2D interaction of drug molecules with key active site residues of the mAChR (PDB ID: 6WJC)
The compound ipratropium bromide forms hydrogen bonds with Asp103 and Asn382 and hydrophobic interactions with Val113 and Leu116. The main interactions of formoterol show that it forms hydrogen bonds with Ser107 and Thr192. Hydrophobic interactions with Phe197 and Val113. The possibilities of the binding site show that it can bind to an allosteric site or a secondary pocket. Levalbetrol shows the strong hydrogen bonds interactions with Ser107 and Asn382. Hydrophobic interactions with Phe197 and Val113. The binding site is similar to that of formoterol, likely non-specific. Propranolol forms hydrogen bonds with Ser107 and Thr192, as well as hydrophobic interactions with Phe197 and Val113, and its binding mode suggests possible interaction with a secondary pocket or allosteric site. Salbutamol similarly forms hydrogen bonds with Ser107 and Thr192 and hydrophobic interactions with Phe197 and Val113 with a binding site similar to other β2-agonists, likely non-specific.
Structural insights from PDB IDs 5CXV, 6OIJ, 6OL9, 6WJC: these structures represent different conformations or subtypes of mAChRs. The binding modes may vary slightly depending on the receptor subtype (e.g., M1, M2, M3). Structure based virtual screening suggest that 6WJC shows the strongest binding for most compounds, suggesting a more favorable conformation for ligand binding. To perform a binding mode analysis for the compounds listed in our files (formoterol, ipratropium bromide, levalbetrol, propranolol, salbutamol, and tiotropium) with the mAChRs using the provided PDB IDs (5CXV, 6OIJ, 6OL9, 6WJC), we need to consider the steps listed bellow.
Ramachandran plot analysis
The Ramachandran plot analysis for mAChRs, a family of G protein-coupled receptors (GPCRs) that mediate the effects of acetylcholine, provides detailed structural insights into four protein structures identified by PDB IDs: 5CXV, 6OIJ, 6OL9, and 6WJC (Figure 13). Below is a detailed explanation of each protein based on the Ramachandran plot data and general knowledge about mAChRs. The 5CXV structure represents the M2 muscarinic receptor in an inactive state, stabilized with an antagonist. The plot shows most residues in favored regions, indicating a stable fold. Clusters in glycine and proline areas suggest flexibility, likely in extracellular loops or intracellular regions. Outliers may reflect distortions from the fusion protein or crystallization artifacts.27 The 6OIJ entry corresponds to the M2 receptor in an active-like state, bound to an agonist and a G protein mimetic nanobody. Most residues lie in favored regions, reflecting a well-defined active structure. Enhanced clustering in glycine and proline regions may indicate dynamic changes due to agonist binding. Outliers could highlight strain from the nanobody or G protein mimic interaction. The 6OL9 structure features the M2 receptor with a different agonist and a nanobody, offering another view of the active state. The majority of residues are in favored regions, with noticeable glycine and proline clusters suggesting flexibility. Outliers may reflect unique conformational adjustments induced by the specific agonist. The 6WJC entry represents the M4 muscarinic receptor, stabilized with an antagonist. The plot shows a high proportion of residues in favored regions, indicating structural stability. Clusters in flexible regions (glycine/proline) may reflect subtype-specific loop dynamics. Outliers could point to areas affected by the antagonist or crystallization conditions.
Ramachandran plot analysis for muscarinic acetylcholine receptors (mAChRs) PDB IDs: 5CXV, 6OIJ, 6OL9, and 6WJC
Pharmacokinetic properties and toxicity predictions
In silico pharmacokinetic properties and toxicity predictions offer insights into the characteristics of these well-known drug compounds (Table 3). Formoterol has a moderate molecular weight of 344.41 and six hydrogen bond acceptors with four donors, suggesting a good balance between solubility and permeability. Its nine rotatable bonds indicate flexibility in binding, and the molecular refractivity of 97.5 reflects moderate polarizability. The topological polar surface area (TPSA) of 90.82 supports reasonable solubility, while its octanol/water partition coefficient (logP) of 3.32 suggests a lipophilic nature, aiding membrane permeability. Ipratropium bromide is larger, with a molecular weight of 412.38, three hydrogen bond acceptors, and one donor, indicating lower hydrogen bonding capacity compared to formoterol. With six rotatable bonds, it is moderately flexible. Its molecular refractivity (108.91) and TPSA (46.53) suggest limited solubility, and a negative logP (-0.18) reflects a hydrophilic profile, which may limit cell membrane penetration. Levalbuterol shows a molecular weight of 371.52 and a notably low number of hydrogen bond acceptors (2) and donors (0), pointing to a more hydrophobic structure. With eight rotatable bonds and a molecular refractivity of 119.72, it has moderate flexibility and polarizability. The very low TPSA of 12.47 suggests poor solubility, and its high logP (6) indicates extreme lipophilicity, which may affect drug distribution.
The drug propranolol is comparatively small, with a molecular weight of 259.34, and offers a balanced profile with three hydrogen bond acceptors and two donors. It is less flexible; having six rotatable bonds, and its molecular refractivity (78.44) and TPSA (41.49) support decent solubility and permeability. Its logP of 2.97 shows a balanced lipophilic profile, making it well-suited for membrane penetration. Salbutamol (albuterol) is the lightest among these compounds, with a molecular weight of 239.31. It has four hydrogen bond acceptors and donors, suggesting higher hydrogen bonding potential for solubility. With five rotatable bonds and a refractivity of 67.6, it is less flexible but still effective in binding. Its TPSA (72.72) supports good solubility, while a logP of 1.7 indicates moderate hydrophilicity. Tiotropium is the heaviest compound (392.51), with four hydrogen bond acceptors and one donor, reflecting limited hydrogen bonding capacity. Its five rotatable bonds signify moderate flexibility, and molecular refractivity of 104.66 shows good polarizability. A high TPSA (115.54) supports solubility, while a logP of 2.3 indicates a balanced lipophilic nature suitable for drug delivery.
The acute toxicity prediction results for oral administration within 24 h have been evaluated in terms of lethal dose-50 (LD50), toxicity class, average similarity, and prediction accuracy (Table 4). Formoterol exhibits a high LD50 value of 3130 mg kg-1, indicating a relatively low acute toxicity. It falls into toxicity class 5, which is considered “practically non-toxic”. Its predictions have an average similarity of 100% and perfect prediction accuracy of 100%. Ipratropium bromide has a much lower LD50 of 380 mg kg-1, signifying moderate acute toxicity. It is categorized under toxicity class 4, which represents “slightly toxic”. The average similarity is 98.68%, with a prediction accuracy of 72.9%. Levalbuterol shows an LD50 of 1190 mg kg-1, indicating moderate toxicity. It is classified as toxicity class 4, similar to ipratropium. Both its similarity and prediction accuracy are at 100%, demonstrating high confidence in the prediction. Propranolol shares an LD50 of 380 mg kg-1, marking moderate acute toxicity. However, it is categorized under toxicity class 3, suggesting it is “moderately toxic”. Both the similarity and accuracy values are 100%. Salbutamol (albuterol) shows an LD50 of 660 mg kg-1, indicative of moderate toxicity. It falls under toxicity class 4, with a high similarity and prediction accuracy of 100%. Tiotropium presents the lowest LD50 of 263 mg kg-1, signaling higher acute toxicity compared to the others. It is classified as toxicity class 3 (“moderately toxic”) and shows the lowest average similarity (81%) and prediction accuracy (70.97%).25
Acute toxicity prediction for oral administration within 24 h is presented as lethal dose-50 (LD50), toxicity class, average similarity, and accuracy calculated for the drugs
The toxicity predictions from ProTox-3.0 for the formoterol drug compound values presented in Table 5 clearly explain that the drug Formoterol is largely safe, with inactive predictions for hepatotoxicity, nephrotoxicity, carcinogenicity, and cardiotoxicity. Radar plot, active and inactive clusters of the drugs were clearly shown (Figures 14 and 15). However, it shows active predictions for neurotoxicity, respiratory toxicity, immunotoxicity, and clinical toxicity. The drug demonstrates moderate interactions with the blood-brain barrier (BBB) and environmental concerns (ecotoxicity).
ProTox-3.0 radar plot for drugs, illustrating the confidence of positive toxicity results compared to the average of their classes
ProTox-3.0 active and inactive clusters of the drugs (a) formoterol; (b) ipratropium bromide; (c) levalbuterol, (d) propranolol; (e) salbutamol; and (f) tiotropium, illustrating the confidence of the positive toxicity results compared to the classes average
Ipratropium bromide: quite safe overall, with inactive predictions for most toxicities, including hepatotoxicity, nephrotoxicity, and carcinogenicity. However, it exhibits active neurotoxicity and respiratory toxicity predictions, and moderate BBB activity is highlighted.
Levalbuterol: displays active neurotoxicity and respiratory toxicity. It has inactive predictions for carcinogenicity, hepatotoxicity, and nephrotoxicity. Additionally, it has some interaction with CYP enzymes, suggesting possible metabolic implications.
Propranolol: mixed profile, with an active prediction for respiratory toxicity and BBB interactions. Most other categories, including hepatotoxicity, neurotoxicity, and cardiotoxicity, are predicted inactive.
Salbutamol (albuterol): generally safe, with inactive results for most toxicity. Active respiratory toxicity and mild immunotoxicity are flagged, while it also demonstrates minor environmental and BBB interactions.
Tiotropium: mostly safe with inactive predictions for hepatotoxicity, nephrotoxicity, and clinical toxicity. However, it shows active predictions for respiratory toxicity and immunotoxicity, along with interactions with metabolic pathways like CYP enzymes.
This summary provides an overview of potential toxicity and safety profiles for each drug, along with their probabilities.
Structural insights into β2-AR and mAChR antagonism and implications for dual-acting therapeutics
In the treatment of obstructive airway diseases such as asthma and COPD, clinical success has been achieved with β2-AR agonists, not antagonists. These agonists promote bronchodilation by binding to and activating β2-AR on airway smooth muscle cells, leading to muscle relaxation and improved airflow. On the other hand, β2-AR antagonism results in bronchoconstriction, which is harmful in these respiratory conditions and should be avoided. In contrast, antagonists of mAChRs are therapeutically useful because they reduce bronchoconstriction and mucus secretion by blocking the cholinergic pathway. Thus, it is crucial to accurately state that β2-AR agonists mediate bronchodilation beneficial to asthma and COPD therapy, while mAChR antagonists serve to prevent bronchoconstriction.
β2-adrenergic receptor (β2-AR) antagonism
Clinically, β2-AR is targeted by agonists for bronchodilation. However, understanding antagonism is key for dual-acting drugs that might block one pathway while activating another. The orthosteric binding site of β2-AR is a deeply buried pocket formed by seven transmembrane helices (TM3, TM5, TM6, and TM7 are primary contributors). The main interactions for catecholamine-like ligands, with which antagonists must compete, include: (i) an ionic lock between the protonated amine of the ligand and a conserved aspartate residue (Asp113) on TM3, (ii) hydrogen bonding between the catechol hydroxyl groups of the ligand and two serine residues (Ser203 and Ser207) on TM5, and (iii) aromatic stacking interactions with residues like Phe290 on TM6. Therefore, a high-affinity β2-AR antagonist must possess a cationic nitrogen that is protonatable at physiological pH to form this essential ionic bond, and a robust aromatic or heteroaromatic system to engage the hydrophobic subpockets and displace the endogenous agonist.26-30
Muscarinic acetylcholine receptor (mAChR) antagonism
The frontline clinical anticholinergics used for COPD, such as tiotropium, act as potent and long-acting mAChR antagonists. The orthosteric site of mAChRs is deeply embedded and notably hydrophobic. The primary anchoring interaction involves the positively charged ammonium group of the ligand with a conserved aspartate residue on TM3. Unlike β2-AR, the mAChR binding pocket lacks complementary serine residues for hydrogen bonding, with selectivity and affinity predominately governed by van der Waals contacts and hydrophobic packing involving aromatic residues on transmembrane helices 3, 5, 6, and 7. A distinctive feature of many successful mAChR antagonists is the presence of a large, bulky aromatic moiety, which deeply penetrates the receptor and confers slow dissociation kinetics underlying their long duration of action.10,11,16
Structural requirements for dual-acting candidates
The development of dual-acting molecules capable of antagonizing the mAChR pathway while simultaneously activating or modulating the β2-AR pathway represents a complex structural challenge. Despite varied pharmacophoric requirements, both receptors share a critical conserved feature: a cationic, protonatable nitrogen that binds with high affinity to an aspartate residue on TM3, providing an anchor point for ligand binding. Effective dual target ligands must bifurcate functionality extending groups that form hydrogen bonds with TM5 serine residues to enable β2-AR agonism, and incorporating bulky hydrophobic aromatic structures to occupy the extensive mAChR hydrophobic pocket, promoting antagonism and slow off-rate kinetics. Achieving this balance requires sophisticated modeling and fragment-based design to optimize spatial orientation and physicochemical properties, which could lead to novel therapeutics for complex respiratory diseases.2,10
Critical analysis of toxicity limitations and future directions
Current pharmacotherapies for airway diseases like asthma and COPD, including formoterol, salbutamol, levalbuterol, ipratropium bromide, and tiotropium, face challenges posed by mechanism-related adverse effects and limited receptor subtype selectivity. β2-agonists such as salbutamol show dose-dependent cardiovascular side effects due to residual β1-adrenergic receptor activity, necessitating strategies like the development of enantiomerically pure levalbuterol to reduce off-target effects.4,13 Conversely, non-selective β-blockers like propranolol exacerbate bronchoconstriction and are contraindicated in asthma, reaffirming the necessity for β2-AR agonism in respiratory indication.
Muscarinic antagonists, including ipratropium bromide and tiotropium, carry anticholinergic side effects; the quaternary ammonium structure in these agents limits systemic absorption and central nervous system penetration, reducing systemic toxicity.12,16 However, prolonged receptor residence time of tiotropium can lead to local adverse effects such as dry mouth and rare cardiac effects through M2 receptor blockade.30 Future drug design should prioritize subtype selectivity (e.g., β2 over β1, and M3/M1 over cardiac M2 receptors), inhaled delivery optimization to reduce systemic exposure, and stereochemical purity for improved therapeutic index.
Enhancing subtype selectivity
For the β2-AR component, the goal is absolute selectivity over β1 AR. This requires exploiting subtle differences in the ligand-binding pockets (e.g., the so-called “selectivity ring” of residues) through computational modeling and structure-based design to eliminate cardiovascular effects. For the mAChR component, the challenge is achieving M3/M1 selectivity over cardiac M2 receptors. This may involve designing bitopic ligands that engage both the orthosteric site (via the essential cationic amine) and a less-conserved allosteric site to drive subtype specificity.
Optimizing for inhaled delivery and reduced systemic exposure
The success of ipratropium and tiotropium underscores the safety conferred by quaternary ammonium motifs. Future molecules should be designed with high molecular weight, high plasma protein binding, and low oral bioavailability to ensure rapid clearance and minimal systemic activity if absorbed. The “soft drug” strategy, exemplified by investigational agents like espedrine, involves incorporating metabolically labile esters (e.g., isosteric ester replacements of carboxylic groups) that ensure rapid hydrolysis to inactive metabolites in the systemic circulation, thereby confining activity to the lung.
Stereochemical purity
The lesson learned from levalbuterol is that enantiomeric purity is not a mere refinement but a critical determinant of the therapeutic index. Future synthesis must aim for enantiomerically pure compounds to avoid the off-target pharmacologic and toxicologic contributions of inactive or antagonistic stereoisomers.
The historical toxicities of salbutamol, propranolol, and even tiotropium are not inevitable artifacts but predictable consequences of molecular properties. They serve as crucial guideposts. The path forward requires a deliberate design philosophy that prioritizes kinetic and subtype selectivity, metabolic vulnerability, and targeted delivery over mere binding affinity. By learning from these clinical benchmarks, the next generation of inhibitors and dual-acting agents can aspire to dissociate powerful efficacy in the lung from the dose-limiting systemic effects that have long been the Achilles heel of respiratory pharmacotherapy.
Rational design of novel candidate compounds for enhanced safety and efficacy
Building upon the critical analysis of the limitations inherent to current clinical agents, we propose a series of novel candidate compounds designed based on docking based virtual screening to transcend these challenges. This rational design strategy is firmly grounded in the structural pharmacophore requirements for β2-AR and mAChR engagement and is explicitly engineered to incorporate the optimization strategies necessary for a superior therapeutic index. The primary focus is on achieving dual-acting molecules and highly selective antagonists with inherent safety features.
Docking based virtual screening
Docking analysis of ChEMBL ligands with β2-AR (PDB: 2RHI) (Figure 16) showed strong binding affinities, with scores ranging from 80.95 to 87.27 kcal mol-1 (Table 6). CHEMBL2031012 bound at 81.38 / 84.28 kcal mol-1, stabilized by hydrogen bonds (Asn293, Ser204, Thr195), van der Waals contacts (Val114, Asp113, Ile94), and aromatic π-π interactions (Phe193, Phe194, Tyr308). CHEMBL2031018 showed slightly stronger binding (85.00 kcal mol 1), supported by hydrogen bonds (Asn312, Ser203, Asp113), π-alkyl / π-π stacking with Phe194, Tyr199, Phe289, and carbon hydrogen bonds (His296, Ser200). CHEMBL3940231 had the highest affinity (87.27 kcal mol-1), with dense hydrogen bonding (Ser204, Thr195, Asn293) and extensive π-stacking with Phe289, Phe290, and Tyr316, suggesting potent β2-AR binding. CHEMBL3942482 showed moderately strong binding (80.95 kcal mol-1), stabilized by hydrogen bonds (Ser204, Asn293, Tyr199), π-alkyl / π-π contacts (Phe193, Phe194, Tyr308), and hydrophobic interactions (Val114, Ile309, Trp109), though with fewer stacking networks compared to CHEMBL3940231.
2D interaction of the four compounds (a) CHEMBL2031012, (b) CHEMBL2031018, (c) CHEMBL3940231, and (d) CHEMBL3942482, with key active site residues of the β2-AR (PDB: 2RHI)
Docking studies of the four ChEMBL compounds with β2 adrenergic receptor (PDB: 5D5A) showed strong binding affinities (Figure 17), with docking scores of 84.28, 87.01, 99.96, and 87.41 kcal mol-1, respectively. CHEMBL2031012 formed key hydrogen bonds with Asn312, Asn293, and Asp192, along with π-anion and π-alkyl interactions, though an unfavorable clash with Lys308 was observed. CHEMBL2031018 displayed a broader interaction profile involving hydrogen bonds with Asn293, Asp113, and Ser204, π-sulfur with CYS191, and π-anion with Asp300, supported by multiple hydrophobic contacts. CHEMBL3940231, with the strongest score (99.96 kcal mol-1), showed rich stabilization through hydrogen bonds with Asn293, Asp113, and Ser204, plus unique π-cation and π-sulfur interactions, along with extensive hydrophobic stacking. CHEMBL3942482 also exhibited diverse contacts, including hydrogen bonds with Asn312, Asp113, and Ser204, π-anion with ASP300, π-lone pair with ASN301, and multiple π-π interactions with PHE residues. Overall, all compounds anchored well in the binding pocket, with CHEMBL3940231 demonstrating the most stable and favorable interaction network.
2D interaction of the four compounds: (a) CHEMBL2031012, (b) CHEMBL2031018, (c) CHEMBL3940231, and (d) CHEMBL394248, with key active site residues of the β2-AR (PDB: 5D5A)
The docking analysis of the four ligands with mAChR (6WJC) showed distinct predicted binding energy (Figure 18). CHEMBL2031012 exhibited the strongest interaction (92.49 kcal mol-1), supported by multiple hydrogen bonds and extensive π-π stacking with aromatic residues, indicating high binding stability (Table 6). CHEMBL3942482 followed with a favorable docking score (82.00 kcal mol-1), where hydrogen bonding and van der Waals contacts contributed to receptor stabilization, though with fewer aromatic stacking interactions compared to CHEMBL2031012. CHEMBL2031018 displayed moderate affinity (66.23 kcal mol-1), forming a smaller network of stabilizing hydrogen bonds and π-interactions. In contrast, CHEMBL3940231 showed the weakest binding (46.97 kcal mol-1), characterized by limited hydrogen bonding and reduced hydrophobic / π-stacking interactions. Collectively, these results suggest that CHEMBL2031012 is the most promising mAChR binder, followed by CHEMBL3942482, while CHEMBL3940231 shows the least potential.
2D interaction of the four compounds: (a) CHEMBL2031012, (b) CHEMBL2031018, (c) CHEMBL3940231, and (d) CHEMBL394248, with key active site residues of the mAChR (6WJC)
The docking analysis of the four ligands with the muscarinic acetylcholine receptor (mAChR, PDB: 5CXV) showed varied binding strengths (Figure 19). CHEMBL3942482 emerged as the strongest binder with a docking score of 86.19 kcal mol-1, stabilized by multiple hydrogen bonds with residues such as Asn and Thr, along with extensive π-π stacking and hydrophobic interactions that reinforced ligand accommodation. CHEMBL2031012 also showed favorable binding (80.57 kcal mol-1), where hydrogen bonding with Tyr and Glu, supported by π-alkyl and van der Waals interactions, contributed to receptor stabilization. In contrast, CHEMBL2031018 demonstrated a moderate affinity (68.76 kcal mol-1), mainly through van der Waals and limited aromatic interactions, suggesting reduced stability compared to the top ligands. Interestingly, CHEMBL3940231 showed the weakest interaction (15.65 kcal mol-1), lacking extensive stabilizing contacts, which accounts for its poor binding efficiency. Overall, CHEMBL3942482 and CHEMBL2031012 stood out as strong candidates due to their rich interaction profiles, while CHEMBL3940231 appeared to be the least favorable binder for mAChR.
2D interaction of the four compounds: (a) CHEMBL2031012, (b) CHEMBL2031018, (c) CHEMBL3940231, and (d) CHEMBL394248, with key active site residues of the mAChR (5CXV)
CONCLUSIONS
Structure-based virtual screening (SBVS) and molecular docking were employed to examine ligand-binding preferences toward two therapeutically relevant protein targets: the β2-adrenergic receptor (β2-AR) and the muscarinic acetylcholine receptor (mAChR). The results provided clear insights into the molecular interactions, selectivity, and potential pharmacological implications of both clinically established drugs and novel CHEMBL-derived compounds. For clinically approved ligands, formoterol and salbutamol demonstrated strong and selective affinity toward β2-AR, aligning with their well-known roles as bronchodilatory agonists. Conversely, tiotropium and ipratropium displayed the highest binding affinities for mAChRs, consistent with their use as muscarinic antagonists in the management of respiratory conditions such as COPD and asthma. Interestingly, tiotropium, though primarily targeting mAChRs, also showed significant binding affinity toward β2-AR, suggesting potential cross-reactivity. Toxicological evaluation revealed important differences: formoterol exhibited the lowest toxicity profile (high LD50, low toxicity classification), while tiotropium appeared more acutely toxic, indicating the need to balance efficacy with safety in therapeutic applications.
The docking-based screening of four CHEMBL compounds (CHEMBL2031012, CHEMBL2031018, CHEMBL3940231, and CHEMBL3942482) further highlighted distinct binding patterns. Among them, CHEMBL3940231 displayed the strongest binding toward β2-AR (87.27 kcal mol-1), stabilized by multiple hydrogen bonds and π-π aromatic stacking, while its binding to mAChRs was comparatively weak (15.65 kcal mol-1), suggesting a high degree of target selectivity. In contrast, CHEMBL3942482 was the most effective binder for mAChRs (86.19 kcal mol-1), forming extensive hydrogen bonds and hydrophobic contacts, though its affinity toward β2-AR was moderate. CHEMBL2031012 showed balanced interactions across both receptors (84.28 kcal mol-1 for β2-AR and 80.57 kcal mol-1 for mAChRs), indicating potential as a dual-target agent. Meanwhile, CHEMBL2031018 demonstrated slightly weaker but consistent binding at both sites (85.00 and 68.76 kcal mol-1), reflecting a more moderate binding profile.
Taken together, these results emphasize the complementary strengths of approved drugs and novel CHEMBL-derived ligands. Clinically used compounds confirm established therapeutic mechanisms, while CHEMBL molecules, particularly CHEMBL3940231 and CHEMBL3942482, present opportunities for selective or dual-target drug development. This dual approach leveraging both known pharmacology and computationally identified leads can guide the rational design of next-generation therapeutics for respiratory diseases, balancing efficacy, selectivity, and safety.
DATA AVAILABILITY STATEMENT
All data are available in the text.
REFERENCES
-
1 Levy, M. L.; Bacharier, L. B.; Bateman, E. D.; Boulet, L. P.; Brightling, C. E.; Buhl, R.; Reddel, H. K.; NPJ Primary Care Respiratory Medicine 2023, 33, 18. [Crossref]
» Crossref -
2 Sriram, K.; Insel, P. A.; Mol. Pharmacol. 2018, 93, 251. [Crossref]
» Crossref -
3 Dolen, W. K.; Allergy Proc. 1994, 15, 283. [Crossref]
» Crossref -
4 Barisione, G.; Baroffio, M.; Crimi, E.; Brusasco, V.; Pharmaceuticals 2010, 3, 1016. [Crossref]
» Crossref -
5 Allon, M.; Dunlay, R.; Copkney, C.; Ann. Intern. Med. 1989, 110, 426. [Crossref]
» Crossref -
6 Cook, P.; Scarfone, R. J.; Cook, R. T.; Ann. Emerg. Med. 1994, 24, 316. [Crossref]
» Crossref -
7 Jarvie, D. R.; Thompson, A. M.; Dyson, E. H.; Clin. Chim. Acta 1987, 168, 313. [Crossref]
» Crossref -
8 Prior, J. G.; Cochrane, G. M.; Raper, S. M.; Ali, C.; Volans, G. N.; Br. Med. J. 1981, 282, 1932. [Crossref]
» Crossref -
9 Dehu, L.; Wu, J.; Xiong, X.; Molecules 2023, 28, 1139. [Crossref]
» Crossref -
10 Maeda, S.; Qu, Q.; Robertson, M. J.; Skiniotis, G.; Kobilka, B. K.; Science 2019, 364, 552. [Crossref]
» Crossref -
11 Caulfield, M. P.; Birdsall, N. J.; Pharmacol. Rev. 1998, 50, 279. [Link] accessed in November 2025
» Link -
12 Whitsett, J. A.; Hollinger, B.; Pediatr. Res. 1984, 18, 1136. [Crossref]
» Crossref -
13 Cazzola, M.; Page, C. P.; Calzetta, L.; Matera, M. G.; Pharmacol. Rev. 2012, 64, 450. [Crossref]
» Crossref -
14 Hislop, A. A.; Mak, J. C.; Reader, J. A.; Barnes, P. J.; Haworth, S. G.; Eur. J. Pharmacol. 1998, 359, 211. [Crossref]
» Crossref -
15 Dixon, W. E.; The Journal of Physiology 1903, 29, 97. [Crossref]
» Crossref -
16 Barnes, P. J.; Thorax 1989, 44, 161. [Crossref]
» Crossref -
17 Gater, P. R.; Alabaster, V. A.; Piper, I.; Pulm. Pharmacol. 1989, 2, 87. [Crossref]
» Crossref -
18 Jeffery, P. K.; Proc. Am. Thorac. Soc. 2004, 1, 176. [Crossref]
» Crossref -
19 Chtita, S.; Belaidi, S.; Qais, F. A.; Ouassaf, M.; AlMogren, M. M.; Al-Zahrani, A. A.; Bakhouch, M.; Belhassan, A.; Zaki, H.; Bouachrine, M.; Lakhlifi, T.; J. King Saud Univ., Sci. 2022, 34, 102226. [Crossref]
» Crossref -
20 Rossafi, B.; Outahar, F.; Hammoudan, I.; Moumou, M.; Touti, R.; Benharref, A.; Chtita, S.; J. Mol. Struct. 2025, 1320, 139664. [Crossref]
» Crossref -
21 Khedraoui, M.; Guerguer, F. Z.; Karim, E. M.; Errougui, A.; Chtita, S.; Curr. Pharm. Anal. 2025, 21, 238. [Crossref]
» Crossref -
22 Bouribab, A.; Karim, E. M.; Khedraoui, M.; Abchir, O.; Errougui, A.; Raouf, Y. S.; Samadi, A.; Chtita, S.; Pharmaceuticals 2024, 17, 1528. [Crossref]
» Crossref -
23 Guerguer, F. Z.; Bouribab, A.; Karim, E. M.; Khedraoui, M.; Amegrissi, F.; Raouf, Y. S.; Samadi, A.; Chtita, S.; PLoS One 2025, 20, e0313411. [Crossref]
» Crossref -
24 El Bouamri, L.; Bouachrine, M.; Chtita, S.; Curr. Top. Med. Chem. 2025, 25, 657. [Crossref]
» Crossref -
25 Banerjee, P.; Kemmler, E.; Dunkel, M.; Preissner, R.; Nucleic Acids Res. 2024, 52, W513. [Crossref]
» Crossref -
26 Rasmussen, S. G. F.; Choi, H. J.; Fung, J. J.; Pardon, E.; Casarosa, P.; Chae, P. S.; Devree, B. T.; Rosenbaum, D. M.; Thian, F. S.; Kobilka, T. S.; Schnapp, A.; Konetzki, I.; Sunahara, R. K.; Gellman, S. H.; Pautsch, A.; Steyaert, J.; Weis, W. I.; Kobilka, B. K.; Nature 2011, 469, 175. [Crossref]
» Crossref -
27 Ramachandran, G. N.; Ramakrishnan, C.; Sasisekharan, V.; J. Mol. Biol. 1963, 7, 95. [Crossref]
» Crossref -
28 Rasmussen, S. G. F.; Choi, H. J.; Rosenbaum, D. M.; Kobilka, T. S.; Thian, F. S.; Edwards, P. C.; Burghammer, M.; Ratnala, V. R. P.; Sanishvili, R.; Fischetti, R. F.; Schertler, G. F. X.; Weis, W. I.; Kobilka, B. K.; Nature 2007, 450, 383. [Crossref]
» Crossref -
29 McGraw, D. W.; Liggett, S. B.; Proc. Am. Thorac. Soc. 2005, 2, 292. [Crossref]
» Crossref -
30 Kolb, P.; Rosenbaum, D. M.; Irwin, J. J.; Fung, J. J.; Kobilka, B. K.; Shoichet, B. K.; Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 6843. [Crossref]
» Crossref
Edited by
-
Editor handled this article:
Nelson H. Morgon
