Molecular Docking Analysis of Aspirin with Cyclooxygenase-2 (COX-2) Using Schrödinger Software

Inflammation is a complex biological response initiated by tissues against harmful stimuli such as pathogens, chemical injury, oxidative stress, or trauma. Among the major mediators involved in the inflammatory cascade, prostaglandins play a critical role in pain, fever, vasodilation, and tissue edema. These mediators are biosynthesized from arachidonic acid by the action of cyclooxygenase (COX) enzymes, mainly cyclooxygenase-1 (COX-1) and cyclooxygenase-2 (COX-2). COX-1 is constitutively expressed in normal physiological tissues and is involved in gastric mucosal protection, platelet aggregation, and renal homeostasis, whereas COX-2 is an inducible isoform that is overexpressed during inflammation, cancer, and several chronic disorders [1,2]. Therefore, COX-2 has emerged as an important pharmacological target for anti-inflammatory drug discovery.

Aspirin (acetylsalicylic acid) is one of the oldest and most widely used nonsteroidal anti-inflammatory drugs (NSAIDs). Its pharmacological activity is mainly attributed to the irreversible acetylation of a serine residue within the active site of cyclooxygenase enzymes, thereby inhibiting prostaglandin synthesis [3]. Although aspirin traditionally inhibits both COX isoforms, its interaction with COX-2 is particularly significant because acetylation of COX-2 leads to altered catalytic behavior and generation of anti-inflammatory lipid mediators. The crystal structure of aspirin-acetylated human COX-2 (PDB ID: 5F19) provides valuable structural insight into the binding orientation and molecular interactions responsible for its inhibitory action, making it an ideal receptor model for computational docking studies. Molecular docking has become a fundamental in silico drug design strategy for predicting ligand–protein interactions, estimating binding affinity, and identifying key residues involved in molecular recognition. Schrödinger Glide is among the most reliable docking platforms due to its robust receptor grid generation, conformational sampling, and extra precision (XP) scoring algorithms, which improve the prediction of biologically relevant poses . In the present study, Schrödinger Glide was selected to evaluate the interaction profile of aspirin with COX-2, focusing on docking score, binding orientation, hydrogen bonding, hydrophobic interactions, and possible acetylation-related conformational changes.Several previous studies have highlighted the molecular basis of aspirin binding to cyclooxygenase enzymes. Orlando and Malkowski demonstrated the structural basis of aspirin-mediated acetylation of human COX-2 and explained how aspirin modifies Ser530, leading to inhibition of prostaglandin biosynthesis [4]. Earlier crystallographic investigations by Kurumbail et al. established the selective active site differences between COX-1 and COX-2, paving the way for structure-based NSAID development [5]. Additional docking and simulation studies have further validated the importance of residues such as Tyr385, Ser530, Arg120, and Tyr355 in stabilizing aspirin and related NSAIDs within the COX-2 catalytic channel [6,7]. These studies collectively support the rationale for employing docking approaches to understand aspirin–COX-2 interaction in detail.The working hypothesis of the present investigation is that aspirin exhibits stable and energetically favorable binding within the active site of human cyclooxygenase-2 (PDB: 5F19), mediated by key hydrogen bonding and hydrophobic interactions involving catalytically important amino acid residues. It is further hypothesized that Schrödinger Glide docking can reliably reproduce the experimentally known binding orientation of aspirin and provide mechanistic insights into its anti-inflammatory activity.The present in silico molecular docking study of aspirin with cyclooxygenase-2 (COX-2) is highly significant in the field of rational anti-inflammatory drug design because it computationally validates the molecular basis of aspirin–enzyme interaction. By using the crystal structure of aspirin-acetylated human COX-2 (PDB ID: 5F19) and Schrödinger Glide docking, the study provides valuable insights into the binding orientation, docking score, and interaction profile of aspirin within the catalytic domain of COX-2. Such computational evidence strongly supports the well-established mechanism of aspirin-mediated inhibition of prostaglandin biosynthesis through acetylation of the active site serine residue. This not only strengthens the mechanistic understanding of one of the most widely used NSAIDs but also demonstrates the usefulness of modern molecular modeling tools in validating classical pharmacological concepts. Another important significance of this work lies in its application as a reference structural model for lead optimization studies. Since aspirin remains a prototype molecule for anti-inflammatory drug development, understanding its exact interaction pattern with COX-2 can help medicinal chemists design novel aspirin analogs, esterified prodrugs, hybrid NSAIDs, and selective dual inhibitors with improved efficacy and reduced gastrointestinal side effects. The docking-based interaction map generated in this study can serve as a template for modifying the salicylate scaffold, introducing substituents that enhance hydrophobic occupancy, hydrogen bond stabilization, and selectivity toward COX-2 over COX-1. This approach may contribute to the development of next-generation NSAIDs with superior therapeutic profiles.From a pharmaceutical research perspective, this study also highlights the importance of computational screening as a cost-effective preliminary drug discovery tool. Compared with conventional wet-lab screening, molecular docking significantly reduces time, material requirements, and experimental costs by enabling virtual evaluation of ligand–protein interactions before synthesis. Therefore, the current work establishes a computational workflow that can be readily extended to screen larger libraries of salicylates, NSAID derivatives, or phytochemical anti-inflammatory compounds against COX-2. This enhances its translational value in early-stage drug discovery and repurposing studies.The future scope of this work is extensive and scientifically promising. A logical extension of the present docking study would be the use of molecular dynamics (MD) simulation to evaluate the time-dependent stability and conformational flexibility of the aspirin–COX-2 complex under physiological conditions. While docking provides a static binding pose, MD simulation can reveal the dynamic behavior of key amino acid residues, solvent effects, hydrogen bond persistence, and structural fluctuations over nanosecond to microsecond timescales, thereby improving the biological relevance of the predictions.Further refinement of binding affinity predictions can be achieved through MM-GBSA (Molecular Mechanics–Generalized Born Surface Area) free energy calculations, which provide a more accurate estimation of ligand–receptor binding energies than docking scores alone. Such calculations would help quantify the energetic contribution of van der Waals interactions, electrostatic forces, solvation effects, and nonpolar interactions, thereby enabling better ranking of aspirin analogs or screened derivatives.Another promising direction is QSAR-guided structural optimization of aspirin derivatives. By integrating docking descriptors with molecular physicochemical parameters, machine learning, and quantitative structure–activity relationship models, it is possible to predict structural features responsible for improved COX-2 affinity and selectivity. This can accelerate the rational design of safer NSAIDs with minimized ulcerogenic effects.The present study also opens opportunities for the virtual screening of novel salicylate analogs and hybrid molecules. Structural modification of aspirin through bio-isosteric replacement, linker insertion, conjugation with antioxidants, nitric oxide donors, or phytochemicals may produce multifunctional molecules with enhanced anti-inflammatory and chemo-preventive potential. Such molecules can first be prioritized using Glide XP docking and later validated through higher-level computational methods.In addition, ADMET prediction and pharmacokinetic profiling represent an essential future step. Evaluating parameters such as oral bioavailability, blood–brain barrier permeability, plasma protein binding, CYP450 metabolism, hepatotoxicity, and renal clearance can significantly improve the translational relevance of newly designed aspirin derivatives. Integrating docking with ADMET filtering can lead to more drug-like candidates suitable for further preclinical studies.Finally, the most important future direction is experimental validation of the computational findings. The predicted binding interactions should be confirmed through in vitro COX-2 enzyme inhibition assays, prostaglandin quantification studies, cell-based anti-inflammatory models, and in vivo animal models of acute and chronic inflammation. Such validation would bridge the gap between computational prediction and pharmacological reality, strengthening the evidence for aspirin scaffold optimization.Overall, the present study provides a strong computational platform and mechanistic framework for future rational design, optimization, and biological validation of safer and more effective COX-2-targeted anti-inflammatory agents.

2. Methodology

2.1. Protein Retrieval and Preparation

The three-dimensional crystal structure of human cyclooxygenase-2 acetylated with aspirin (PDB ID: 5F19) was retrieved from the RCSB Protein Data Bank in .pdb format. The selected protein structure was imported into the Schrödinger Maestro molecular modeling suite for preprocessing. Protein preparation was carried out using the Protein Preparation Wizard, which included assignment of bond orders, addition of missing hydrogen atoms, correction of incomplete side chains, optimization of ionization states, and generation of proper protonation states of amino acid residues at physiological pH (7.0 ± 0.2). Water molecules located beyond the active site region were removed to avoid unwanted interference during docking calculations. The hydrogen-bonding network was optimized, followed by restrained energy minimization using the OPLS force field until the heavy atom root mean square deviation (RMSD) reached the default convergence threshold. This step ensured structural stability and removal of steric clashes before docking [8,9].

2.2. Ligand Preparation

The chemical structure of aspirin (acetylsalicylic acid) was drawn using ChemDraw and exported in MOL format before importing into Maestro. Ligand preparation was performed using the LigPrep module of Schrödinger, which generated the optimized three-dimensional geometry of the ligand. During this step, possible ionization states, tautomeric forms, stereochemical corrections, and low-energy conformers were generated at near physiological pH conditions. Partial atomic charges were assigned using the OPLS_2005/OPLS3e force field, and the ligand structure was subjected to geometry minimization to obtain the most stable conformer for docking analysis [9,10].

2.3. Active Site Identification and Receptor Grid Generation

The active binding region of cyclooxygenase-2 was defined based on the coordinates of the co-crystallized aspirin binding pocket in PDB 5F19. The Receptor Grid Generation module of Glide was used to create the docking grid around the active site residues. The centroid of the co-crystallized ligand was selected as the grid center to ensure accurate sampling of the biologically relevant binding site. The outer box dimensions were adjusted to completely cover the catalytic channel and surrounding amino acid residues, including Arg120, Tyr355, Tyr385, and Ser530, which are known to be critical for NSAID binding and acetylation. Van der Waals scaling factors and partial charge cutoffs were maintained at default values for optimal ligand accommodation [10,11].

2.4. Molecular Docking Using Glide

Prepared aspirin ligand was docked into the receptor grid of COX-2 using the Glide docking module available in Schrödinger Maestro. Docking was performed using Standard Precision (SP) mode for initial binding assessment, followed by Extra Precision (XP) docking to improve pose discrimination and scoring reliability. Glide uses a hierarchical filtering algorithm that evaluates ligand flexibility, shape complementarity, electrostatic interactions, hydrophobic enclosure, hydrogen bonding, and desolvation penalties to predict the most favorable binding pose [8]. The ligand was allowed full rotational and torsional flexibility during docking, while the receptor was kept rigid. Multiple poses were generated, and the best docking pose was selected based on the lowest Glide score, best Emodel value, and favorable interaction geometry [8,11].

2.5. Docking Pose Analysis

The top-ranked docking pose of aspirin with COX-2 was analyzed using the Maestro pose viewer and ligand interaction diagram tool. Key intermolecular interactions such as hydrogen bonding, π–π stacking, salt bridges, hydrophobic contacts, and van der Waals interactions were examined in detail. Special attention was given to the interaction of aspirin with Ser530, which is the experimentally known acetylation site responsible for irreversible COX inhibition. The docking score obtained from the Glide XP run was recorded and compared with reported literature values of known COX-2 inhibitors to assess the reliability of the predicted binding affinity [11,12].

2.6. Validation of Docking Protocol

To validate the docking methodology, the co-crystallized ligand orientation in the crystal structure was compared with the predicted docked pose of aspirin. The protocol was considered reliable if the docked pose reproduced the experimentally known binding orientation with minimal deviation and conserved key active-site interactions. In standard docking validation studies, an RMSD value below 2.0 Å between the redocked and crystal pose is considered acceptable for confirming the robustness of the docking setup [11].

3. RESULTS

3.1 Protein and Ligand Preparation

The crystal structure of human cyclooxygenase-2 acetylated with aspirin (PDB ID: 5F19) was successfully retrieved from the RCSB Protein Data Bank and prepared using the Schrödinger Protein Preparation Wizard. The protein preparation process involved correction of bond orders, assignment of protonation states, optimization of the hydrogen-bonding network, and removal of steric clashes, which ensured structural stability for molecular docking studies. The aspirin ligand was prepared and energy minimized using LigPrep, resulting in a stable low-energy three-dimensional conformer with appropriate ionization and charge distribution. These preprocessing steps improved the reliability and reproducibility of the docking workflow.
As shown in Figure 1(a), the prepared COX-2 protein structure is represented in ribbon form, highlighting the compact catalytic cavity, while Figure 1(b) shows the optimized 3D structure of aspirin used for docking.

3.2 Receptor Grid Generation and Active Site Selection

The active binding site of COX-2 was selected based on the position of the co-crystallized aspirin molecule in the 5F19 crystal structure. A receptor grid was generated around the catalytic channel to ensure accurate ligand sampling during docking. The grid box completely enclosed the important active-site residues, including Arg120, Tyr355, Tyr385, and Ser530, which are known to play essential roles in ligand stabilization and irreversible acetylation of the enzyme.
The successful generation of the receptor grid confirmed that the selected catalytic channel was structurally suitable for aspirin docking. As illustrated in Figure 2, the grid box clearly surrounds the active pocket and catalytic residues, validating the docking setup.

3.3 Molecular Docking Score and Binding Affinity

Molecular docking of aspirin into the active site of COX-2 was carried out using Glide SP/XP mode, which generated multiple possible binding poses. Among these, the best-ranked pose exhibited a Glide docking score of –3.591199 kcal/mol, indicating a favorable and energetically stable interaction between aspirin and the COX-2 catalytic channel. The negative docking score suggests spontaneous binding affinity and supports the experimentally known anti-inflammatory mechanism of aspirin.The docking pose shown in Figure 4A demonstrates that aspirin occupies the catalytic cavity nearArg120, Tyr355, Tyr350, and Ser530, confirming biologically relevant binding. The 2D interaction map in Figure 4B further reveals the involvement of the acetyl group of aspirin in stabilizing interactions with active-site residues, particularly near Ser530, which is the known acetylation site responsible for irreversible COX-2 inhibition.Overall, the obtained docking score and interaction profile strongly support the stable binding and inhibitory potential of aspirin against COX-2, thereby validating the docking protocol and confirming the suitability of Schrödinger Glide for predicting aspirin–protein interactions

Figure 1. Molecular docking of aspirin with COX-2 (PDB ID: 5F19) showing (A) prepared protein and ligand structures, (B) receptor grid and active-site residues with Glide score (−3.591), (C) 3D binding pose in the catalytic pocket, and (D) 2D interaction map highlighting key contacts with Arg120, Tyr355, and Ser530.