Synthesis of Peptide–Nonsteroidal Anti-Inflammatory Drug Hybrid Compounds and Their Selectivity in Inhibiting the Dual Cyclooxygenase-2 (COX)–5-Lipoxygenase (LOX) System: A Docking and Molecular Dynamics Study ^†

Abstract

In this paper, the synthesis of compounds based on a hybrid structure combining a fenbufen analog (a nonsteroidal anti-inflammatory drug, NSAID) with amino acids or peptides is described. Additionally, a detailed molecular docking analysis was performed to investigate their drug–receptor interactions with inflammation-related enzymes COX-2 and 5-LOX. The hybrid compound that demonstrated the best interaction was further evaluated through molecular dynamics simulations, and its binding free energy (ΔG) was calculated using the Molecular Mechanics Poisson–Boltzmann Surface Area (MMPBSA) method.

1. Introduction

Inflammation is an immune response to harmful agents, biochemically characterized by the production of pro-inflammatory cytokines such as interleukins (IL) IL-1, IL-2, IL-6, and IL-7 and tumor necrosis factor-alpha (TNFα) [1]. These interleukins are also overexpressed in globally relevant infections, such as COVID-19, where enzymes like cyclooxygenase-2 (COX2) [2] and 5-lipoxygenase (5-LOX) [3] are likewise upregulated, contributing to increased severity. Another condition where these enzymes are also overexpressed is rheumatoid arthritis, where the resulting metabolites stimulate the production of pro-inflammatory cytokines such as IL-1, IL-6, and IL-8, among others [4]. Given the global prevalence of rheumatoid arthritis, estimated at 0.46% between 1980 and 2018 [5], recent research has focused on treating inflammation through the use of selective COX-2/5-LOX inhibitors [6], yielding promising results in diseases like cancer [7]. One approach to achieving dual-inhibition systems has been the selection of nonsteroidal anti-inflammatory drugs (NSAIDs) conjugated with amino acids (AAs) or peptides. In this context, small peptides have shown potent inhibitory activity against the COX-2 isoenzyme, particularly the tripeptides reported by Rajakrishnan et al., which exhibit inhibition values comparable to celecoxib [8]. Furthermore, peptides isolated from wheat have been shown to have a protective effect, reducing intestinal damage and cytokine levels when administered alongside NSAIDs, as described by Yin et al. [9]. The ionic salts of NSAID-AA conjugates [10] have also been explored for their potential to inhibit inflammation and viral replication in virus SARS-CoV-2 [11]. These conjugates have been synthesized either with linkers [12] or through covalent bonds [13]. Moreira et al. [14] reported a series of naproxen (Npx)–dehydrodipeptide hybrids, where AAs served as central linkers or C-terminal residues, with inhibition values against 5-LOX comparable to naproxen. Among these, the Npx-L-Ala-Z-∆Phe-OH hybrid was identified as a lead compound. Li et al. [15], using naproxen as the base, coupled aromatic D-AAs to achieve selectivity ratios of 20:1. Additionally, they conjugated other NSAIDs such as (R)-furbiprofen, (R,S)-furbiprofen, and (R,S)-ibuprofen, forming supramolecular gel aggregates that exhibited strong COX inhibition [16]. Finally, naproxen has been conjugated with other amino acids and their methyl esters, yielding hydrophobic compounds designed to enhance interaction with the enzyme’s binding pocket, thereby improving catalytic cavity interactions [17].

2. Results and Discussion

In this study, the synthesis of four novel hybrid NSAID-AA and NSAID-peptide compounds is described, based on the conjugation of the fenbufen analog, 4-(4-methoxyphenyl)-4-oxobutanoic acid, obtained through a Heck cross-coupling reaction between 2,3-DHF and 4-iodoanisole [18]. The synthesis followed is outlined in Scheme 1, starting with the coupling of L-phenylalanine (Phe) and L-tyrosine (Tyr) esters to the aryl keto-ester fenbufen analog, using isobutyl chloroformate as the coupling agent. This coupling reaction yielded compounds 1 and 2, subsequently subjected to basic hydrolysis to produce the free acids 3 and 4. These acids were then coupled with an additional residue of amino esters derived from Phe and Tyr, resulting in the final hybrid NSAID-dipeptide compounds, namely methyl (4-(4-methoxyphenyl)-4-oxobutanoyl)-L-tyrosyl-L-tyrosinate (5), methyl (4-(4-methoxyphenyl)-4-oxobutanoyl)-L-phenylalanyl-L-tyrosinate (6), methyl (4-(4-methoxyphenyl)-4-oxobutanoyl)-L-tyrosyl-L-phenylalaninate (7) and methyl (4-(4-methoxyphenyl)-4-oxobutanoyl)-L-phenylalanyl-L-phenylalaninate (8). As observed, the peptide fragment in hybrids 5–8 existed in the methyl ester form to reduce the compounds’ polarity and improve their solubility and transdermal absorption. Once the hybrid compounds 3–8 were synthesized and elucidated, molecular docking studies were conducted with the COX-2 (3LN1) and 5-LOX (3O8Y) enzymes. For COX-2, a redocking process was performed to validate the molecular docking method, yielding an RMSD of 0.91 Å for celecoxib (Cel), indicating an accurate method, as the obtained pose closely matched the crystal pose. For 5-LOX, since no co-crystallized reference drug was available, it was docked with the selective inhibitor zileuton. The molecular docking binding results for COX-2 and 5-LOX are presented in

Table 1. The binding energy of compounds 3–8 with COX-2 and 5-LOX.

		Compound
Entry	Protein	Reference	3	4	5	6	7	8
1	COX2 (kcal/mol)	−11.01 (celecoxib)	−10.38	−9.49	−10.09	−11.10	−11.01	−10.57
2	5LOX (kcal/mol)	−7.83 (zileuton)	−9.06	−9.05	−6.89	−9.2	−7.3	−8.90

.

A docking analysis with the 3LN1 protein revealed that compounds 3, 6, 7, and 8 showed interaction values similar to the selective COX-2 inhibitor. A noteworthy finding across all assays was the presence of both polar and hydrophobic interactions with the Val-509 residue, which is considered crucial for the selective inhibition of the COX-2 isoenzyme. A detailed analysis indicated that the best interaction energies were observed with compounds 6 (11.10 kcal/mol) and 7 (11.01 kcal/mol). In both cases, the ligands adopted a folded, helix-like conformation with their aromatic rings exposed to the hydrophobic catalytic cavity. Specifically for compound 7 (Tyr-Phe residue), the hydroxyl group of the Tyr residue was oriented toward a region containing polar residues; this group participated in three hydrogen bonds with residues Gln-178 (2.29 Å), Leu-338 (3.07 Å), and Phe-504 (2.49 Å), interactions commonly found in COXIB-type NSAIDs (Figure 1a,b). Moreover, a comparative analysis of the docking poses between compound 7 and celecoxib showed a remarkably similar folded conformation, with the Tyr hydroxyl group directed toward the same region as the celecoxib sulfonamide group (Figure 1c).

For 5-LOX, compounds 3 and 4 exhibited affinity values exceeding −9 kcal/mol, higher than zileuton. In the lipoxygenase complexes, compound 3 displayed two notable hydrogen bonds: the first with His-367 (2.16 Å), an AA chelated to a Fe⁺² atom, and the second with Ala-424 (2.58 Å), a hydrogen bond also associated with zileuton interactions. In the case of compound 4, the ligand was positioned near the Fe atom, approaching it via the hydroxyl group of Tyr, forming a hydrogen bond with the C-terminal carboxylate of Ile-673, which was also coordinated to the catalytic Fe atom through a metal bond. This interaction could potentially disrupt the coordination sphere around the metal, as the COO⁻ anion was inserted into the helical region, possibly inactivating the protein. In addition, three hydrogen bonds were identified along the ligand groove with residues Gln-363, THR-364, and Ala-424. An unconventional hydrogen bond (2.68 Å) was also observed between the π-cloud of the anisole ring and Asn-425 (Figure 1d,e). In this context, the adopted pose was stabilized by nonpolar alkyl and π-alkyl interactions. In the comparative analysis with zileuton, both compounds displayed a characteristic U-shaped conformation, where the anisole ring was oriented similarly to the benzothiophene ring, and the amide bond of compound 4 was positioned similarly to the hydroxyl urea group present in the 5-LOX inhibitor (Figure 1f).

Afterward, the best result from the COX-2 docking study was subjected to a 100 ns molecular dynamics simulation to assess the stability and evolution of the complex over time, calculating the root mean square deviation (RMSD). The average RMSD for the protein backbone was 1.774 Å (Figure 2a), with minimum values of 0.699 Å and maximum fluctuations reaching 2.587 Å. For the ligand, the average RMSD was 1.326 Å, with minimum values of 0.460 Å and maximum values of 2.294 Å (Figure 2b). This finding was significant, as the observed values indicated a high stability of ligand 7 within the catalytic site, achieved after 20 ns of simulation. Part of this stability could be explained by the hydrogen bonds observed during most of the simulation, with H-bonds formed with Ser-516 (occupancy of 52.12% throughout the simulation), as well as Tyr-334 (75.88%) and Val-74 (38.34%). Additional information indicated that the binding free energy (ΔG_bind) was determined using the MMPBSA method, with the complex showing an energy of −33.91 kcal/mol. Although this energy was higher than the previously reported value for celecoxib (−47 kcal/mol or −198 kJ/mol) [19], it was still promising, considering the fact that it was being compared to one of the best COX-2 inhibitors. The wide energetic contributions attributed to Van der Waals (−59.99 kcal/mol) and electrostatic interactions (−34.13 kcal/mol) were both consistent with the largely nonpolar nature of the catalytic site.

3. Experimental Section

3.1. General

Column chromatography was carried out with Merck Sigma-Aldrich (Milwaukee, WI, USA) Silica Gel (70–230 mesh). TLC was performed with Merck 60-F254 plates (Darmstadt, Germany), and spots were visualized using UV light and iodine vapor. Melting points were measured with a Fisher-Johns melting points apparatus model 5007 (Sunnyvale, CA, USA) and were left uncorrected. NMR spectra were recorded on a Varian Mercury plus (100 and 400 MHz) Palo Alto, CA, USA.

3.2. Computational Details

Molecular docking was conducted with the COX-2 (PDB ID: 3LN1) and 5-LOX stable (PDB ID: 3O8Y) proteins. Water molecules and additional ligands were removed from each protein subunit, and the receptors were prepared with appropriate charge assignments and parameter additions for the present metals. Molecular docking was performed using 100 runs in AutoDock 4.2 [20], applying default parameters with a population size of 1500 individuals. For the 3LN1 protein, the grid box size was set to 50 × 50 × 50 and centered at coordinates 30.29, −23.07, and −15.99; for 308Y, the grid box size was set to 60 × 66 × 60, centered at coordinates −2.24, 25.69, and −0.94.

Molecular dynamics simulations were conducted using NAMD 3.0 software [21], following the protocol described by Solis-Hernández et al. [22].

3.3. General Procedure for Coupling

In a flask equipped with magnetic stirring and a dropping funnel, 1.1 equivalent of the corresponding methyl ester was dissolved in a THF/DMSO (9:1) mixture. To the resulting mixture, 2.5 equivalents of DIEA was added, and the reaction was stirred for 1 h at 0 °C. The mixture was then added dropwise via the dropping funnel to a second flask containing 1 equivalent of δ-keto acid or δ-keto acid-AA-OH dissolved in THF, which had been pre-treated with 2 equivalents of DIEA and 1.1 equivalent of iBBCl. The reaction was allowed to proceed for 1 h at 0 °C and then at room temperature for 18 h. After the reaction was complete, the solvent was evaporated, and the product was extracted with AcOEt (3 × 50 mL). The organic phase was washed with HCl 10% solution and a saturated NaHCO₃ solution. The organic layer was dried over anhydrous Na₂SO₄. Purification was carried out as described in each case.

3.4. Spectral Data for Best Results

(4-(4-methoxyphenyl)-4-oxobutanoyl)-L-phenylalanine (3). ¹H NMR (400 MHz, CD₃OD TMS) δ ppm: 7.90 (d, J = 8.8 Hz, 2H), 7.16 (m, 5H), 6.95 (d, J = 8.8 Hz, 2H), 4.53 (dd, J = 8.0, 4.9 Hz, 1H), 3.83 (s, 3H), 3.21 (dd, J = 13.8, 4.8 Hz, 1H), 3.13 (td, J = 6.9, 3.9 Hz, 2H), 2.95 (dd, J = 13.7, 8.1 Hz, 1H), 2.52 (t, J = 6.9 Hz, 2H). ¹³C NMR (100 MHz, CD₃OD) δ 199.1, 174.8, 165.2, 138.4, 131.4, 130.8, 130.3, 129.3, 127.7, 114.8, 56.0, 55.1, 38.5, 34.5, 30.7.

(4-(4-methoxyphenyl)-4-oxobutanoyl)-L-tyrosine (4). ¹H NMR (400 MHz, CDCl₃, TMS) δ ppm: 7.93 (d, J = 8.9 Hz, 2H), 7.05 (d, J = 8.5 Hz, 2H), 6.98 (d, J = 8.9 Hz, 2H), 6.69 (d, J = 8.5 Hz, 2H), 4.61 (dd, J = 8.6, 5.3 Hz, 1H), 3.85 (s, 3H), 3.15 (t, J = 7.0 Hz, 2H), 3.09 (dd, J = 14.0, 5.2 Hz, 1H), 2.86 (dd, J = 14.0, 8.6 Hz, 1H), 2.56 (td, J = 7.0, 1.6 Hz, 1H).¹³C NMR (100 MHz, CD₃OD) δ ppm: 199.1, 174.9, 174.8, 165.2, 157.2, 131.4, 131.3, 130.8, 129.0, 116.1, 114.8, 56.0, 55.2, 37.7, 34.5, 30.7.

methyl (4-(4-methoxyphenyl)-4-oxobutanoyl)-L-tyrosyl-L-phenylalaninate (7). ¹H NMR (400 MHz, CD₃OD) δ 8.23–8.10 (m, 1H), 7.95 (d, J = 8.9, 2H), 7.28–7.15 (m, 4H), 7.04–6.97 (m, 4H), 6.66 (d, J = 8.9, 2H), 4.68–4.61 (m, 1H), 4.57–448 (m, 1H), 3.86 (s, 3H), 3.58 (s, 3H), 3.24–3.11 (m, 3H), 3.07–2.96 (m, 2H), 2.74–2.64 (m, 1H), 2.59–2.37 (m, 2H). ¹³C NMR (101 MHz, CD₃OD) δ 199.6, 175.1, 173.7, 173.0, 165.4, 157.1, 138.0, 131.5, 131.2, 130.7, 130.3, 129.5, 129.1, 127.8, 116.1, 114.8, 56.1, 56.0, 55.2, 52.6, 38.3, 37.7, 34.7, 30.8.

4. Conclusions

The successful synthesis of hybrid compounds 3, 4, and 7 sets the stage for their exploration and potential use as anti-inflammatory agents, given their significant interactions with their pharmacological targets and results comparable to the reference drug. These findings highlight them as promising candidates, offering a foundation for the development of a new generation of drugs with dual action on two key enzymes, COX and LOX, implicated in chronic degenerative diseases.