Synthesis, Biological Studies, and In Silico-Driven Design of 8-Aminoquinoline-Based Sulfonamide Derivatives as Potential Antioxidant and Antimicrobial Agents.

Synthesis, Biological Studies, and In Silico-Driven Design of 8-Aminoquinoline-Based Sulfonamide Derivatives as Potential Antioxidant and Antimicrobial Agents - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Comput Struct Biotechnol J . 2026 Apr 17;35(1):0032. doi: 10.34133/csbj.0032 Search in PMC Search in PubMed View in NLM Catalog Add to search Synthesis, Biological Studies, and In Silico -Driven Design of 8-Aminoquinoline-Based Sulfonamide Derivatives as Potential Antioxidant and Antimicrobial Agents Ratchanok Pingaew Ratchanok Pingaew 1 Department of Chemistry, Faculty of Science, Srinakharinwirot University, Bangkok 10110, Thailand. Find articles by Ratchanok Pingaew 1, * , Apilak Worachartcheewan Apilak Worachartcheewan 2 Department of Community Medical Technology, Faculty of Medical Technology, Mahidol University, Bangkok 10700, Thailand. Find articles by Apilak Worachartcheewan 2, * , Veda Prachayasittikul Veda Prachayasittikul 3 Center for Research Innovation and Biomedical Informatics, Faculty of Medical Technology, Mahidol University, Bangkok 10700, Thailand. Find articles by Veda Prachayasittikul 3 , Rungrot Cherdtrakulkiat Rungrot Cherdtrakulkiat 4 Department of Clinical Microbiology and Applied Technology, Faculty of Medical Technology, Mahidol University, Bangkok 10700, Thailand. Find articles by Rungrot Cherdtrakulkiat 4 , Supaluk Prachayasittikul Supaluk Prachayasittikul 3 Center for Research Innovation and Biomedical Informatics, Faculty of Medical Technology, Mahidol University, Bangkok 10700, Thailand. Find articles by Supaluk Prachayasittikul 3 , Somsak Ruchirawat Somsak Ruchirawat 5 Laboratory of Medicinal Chemistry, Chulabhorn Research Institute, Bangkok 10210, Thailand. 6 Program in Chemical Sciences, Chulabhorn Graduate Institute, Bangkok 10210, Thailand. 7 Center of Excellence on Environmental Health and Toxicology (EHT), Commission on Higher Education, Ministry of Education, Bangkok 10400, Thailand. Find articles by Somsak Ruchirawat 5, 6, 7 , Virapong Prachayasittikul Virapong Prachayasittikul 4 Department of Clinical Microbiology and Applied Technology, Faculty of Medical Technology, Mahidol University, Bangkok 10700, Thailand. Find articles by Virapong Prachayasittikul 4 Author information Article notes Copyright and License information 1 Department of Chemistry, Faculty of Science, Srinakharinwirot University, Bangkok 10110, Thailand. 2 Department of Community Medical Technology, Faculty of Medical Technology, Mahidol University, Bangkok 10700, Thailand. 3 Center for Research Innovation and Biomedical Informatics, Faculty of Medical Technology, Mahidol University, Bangkok 10700, Thailand. 4 Department of Clinical Microbiology and Applied Technology, Faculty of Medical Technology, Mahidol University, Bangkok 10700, Thailand. 5 Laboratory of Medicinal Chemistry, Chulabhorn Research Institute, Bangkok 10210, Thailand. 6 Program in Chemical Sciences, Chulabhorn Graduate Institute, Bangkok 10210, Thailand. 7 Center of Excellence on Environmental Health and Toxicology (EHT), Commission on Higher Education, Ministry of Education, Bangkok 10400, Thailand. * Address correspondence to: [email protected] (R.P.); * Address correspondence to: [email protected] (A.W.) Received 2025 Oct 26; Revised 2026 Mar 12; Accepted 2026 Mar 12; Collection date 2026. Copyright © 2026 Ratchanok Pingaew et al. No claim to original U.S. Government Works. Distributed under a Creative Commons Attribution License (CC BY 4.0) . PMC Copyright notice PMCID: PMC13087402  PMID: 42007176 Highlights • 8-Aminoquinoline-based sulfonamide derivatives ( 3 - 13 ) displayed antioxidant and antimicrobial activities. • Significant descriptors influencing bioactivities were explored. • Computational methods were used to construct predictive (QSAR and QSPR) models. • In silico rational design of new 84 structurally modified compounds were generated. • Potential newly designed compounds with the most promising predicted were highlighted. Abstract Eleven 8-aminoquinoline-based sulfonamide derivatives ( 3 - 13 ) were synthesized and experimentally investigated for their antioxidant effects (using superoxide dismutase [SOD] and 2,2-diphenyl-1-picrylhydrazyl [DPPH] assays) and antimicrobial activities (using agar dilution method). Most of the tested compounds were active antioxidants as indicated by SOD assay, affording SOD half-maximal inhibitory concentration in the range of 83.34 to 600.81 μM. However, their DPPH activities were considerably weak (%DPPH = 7.74% to 36.49%). Additionally, 7 compounds ( 3 , 4 , 5 , 6 , 8 , 10 , and 12 ) displayed growth-inhibiting effects against various gram-positive and gram-negative microbes (minimum inhibitory concentration values = ≤4 to 256 μg/ml). Quantitative structure–activity/property relationship (QSAR/QSPR) modeling was performed to obtain predictive models, including antioxidant QSAR (SOD and DPPH) and antimicrobial QSPR models. The constructed models displayed acceptable predictive performance and robustness (QSAR models: R 2 Tr = 0.9980 to 0.9996, Q 2 LOO-CV = 0.9930 to 0.9978, RMSE Tr = 0.0121 to 0.1677, RMSE LOO-CV = 0.0233 to 0.3830 and QSPR model: accuracy = 90.91% for training and 63.64% for leave-one-out cross-validation [LOO-CV] sets). The constructed models were subsequently utilized to guide rational design and predict activities of an additional set of 84 structurally modified compounds. Finally, newly designed compounds with promising predicted antioxidant activity (SOD: 13c , 13f , 8c , 12e , and 11c ; DPPH: 9b , 6f , 9c , 11h , and 3d ) as well as a set of 34 compounds predicted as active antimicrobial agents are summarized for potential investigations. Furthermore, key essential properties influencing bioactivities (electronegativity, polarizability, I-state, van der Waals volume, connectivity, and ionization potential) were also revealed for future beneficial design of the related compounds for medical applications. Graphical Abstract Open in a new tab Introduction Free radicals are unstable and highly reactive molecules endogenously produced or received externally [ 1 – 3 ], These radicals are properly destroyed in our body by antioxidant defense to maintain oxidant–antioxidant balance [ 3 , 4 ]. However, in situations with overproduction of free radicals or impairment of antioxidant defenses, free radicals are excessively accumulated, leading to oxidative damage of the cellular components [ 1 , 2 ]. Oxidative stress is well recognized as a causative factor of multiple oxidative-related diseases and aging conditions (i.e., diabetes mellitus, cardiovascular diseases, cancer, and neurodegenerative diseases) [ 1 – 3 ]. Accordingly, the applications of antioxidant agents in therapeutics, prevention, and health promotion have gained considerable attention [ 4 ]. Infectious diseases are another concerning health problem worldwide [ 5 ], particularly the emergence of resistant bacterial species including gram-positive bacteria (i.e., Staphylococcus aureus and Streptococcus pneumoniae ) and gram-negative bacteria (i.e., Escherichia coli , Salmonella , Shigella , Klebsiella , Enterobacter , and Pseudomonas spp.). Antimicrobial drug resistance in both community and hospital settings has also become an alarming issue [ 5 – 8 ]. This indicates an urgent need for discovery and development of novel potent antimicrobial agents for combating antimicrobial drug resistance crisis [ 9 ]. Heterocyclic compounds (i.e., quinolines) are attractive pharmacophores in medicinal chemistry for the design and development of diverse pharmacologically active compounds as well as the generation of newer derivatives and scaffolds [ 10 , 11 ]. 8-Aminoquinoline (8AQ), containing an amino group (–NH 2 ) at 8-position on the quinoline ring, is a core structure found in antimalarial drugs (i.e., primaquine, pamaquine, and tafenoquine) widely known for their effectiveness against Plasmodium species. (i.e., P. faciparum and P. vivax ) [ 12 ]. Its metabolic process in the mosquito was reported to indicate the promising role of the 8AQ scaffold in antimalarial drug discovery [ 13 ]. Additionally, 8AQ was reported as a ligand capable of coordinating with metal ions to give diverse bioactive metal complexes with medicinal values (i.e., antimicrobial, antioxidant, antimalarial, and anticancer properties) [ 12 , 14 ]. Apart from 8AQ, 4-aminoquinoline was also noted as a scaffold of many reported bioactive compounds displaying antimalarial, antimicrobial, and anticancer activities [ 10 , 15 ]. Sulfonamide is a scaffold widely found in many synthetic antimicrobial agents. It is well known as a versatile pharmacophore in drug design due to its metabolic stability [ 16 ]. Sulfonamide derivatives were also reported for their antioxidant and anticancer activities [ 17 – 19 ]. Molecular hybridization is one of the strategies in drug design for developing multifunctional bioactive compounds [ 20 ]. In recent drug development, great attention has been given to the design of diverse hybrid compounds, including quinoline–sulfonamide hybrids for combating antimicrobial drug resistance [ 21 ]. Cheminformatic approaches have been widely used to effectively facilitate successful drug design and discovery [ 22 ]. Particularly, these tools are employed to reveal structure–activity relationships essential for potent biological activities and preferable drug-like properties [ 23 , 24 ]. Quantitative structure–activity/property relationship (QSAR/QSPR) modeling is one of the computational methods commonly used in drug development to facilitate efficacious virtual design of new candidates [ 25 – 27 ]. The constructed models explored the relationship between the response end points such as biological and chemical activities (represented in numerical values or activities such as active and inactive). The successful stories of QSAR-driven rational design of new derivatives for therapeutics (i.e., antioxidant, neuroprotective, and anticancer agents) have recently been reported by our research group [ 28 – 30 ]. This study demonstrates the combined utilization of in vitro and in silico approaches for the efficacious design of new 8AQ-based sulfonamide analogs. Chemical synthesis was performed to obtain a series of 11 8AQ-based sulfonamide derivatives (compounds 3 - 13 ), and their antioxidant and antimicrobial activities were experimentally investigated. The obtained bioactivity data together with chemical structures of the compounds were further used for constructions of predictive QSAR and QSPR models to reveal key essential structural features required for desirable activities. Finally, the constructed models were further applied to guide the rational design and predict bioactivities of an additional set of 84 structurally modified compounds Materials and Methods Chemistry Analytical thin-layer chromatography was performed on silica gel 60 F 254 aluminum sheets. 1 H- and 13 C-NMR (nuclear magnetic resonance) spectra were recorded on a Bruker AVANCE 300 or a Bruker AVANCE NEO 500 NMR spectrometer. High-resolution mass spectra (HRMS) were recorded on a Bruker Daltonics (micro time of flight [TOF]). Melting points were determined using a Griffin melting point apparatus and are uncorrected. α -Tocopherol, superoxide dismutase (SOD) (bovine erythrocytes), DPPH (2,2-diphenyl-1-picrylhydrazyl), Hepes, nitro blue tetrazolium chloride, L-methionine, riboflavin, and Triton X-100 were purchased from Sigma, USA as well as dimethyl sulfoxide (DMSO, 99.9%) from RCI Labscan, Thailand, and methanol from Merck, Germany. Ampicillin, ciprofloxacin, and tetracycline were procured from Sigma, USA; Mueller–Hinton broth (MHB) and Mueller–Hinton agar from Becton Dickinson, USA; and sodium chloride from Merck, Germany. Solvents are analytical grades. General procedure for the synthesis of quinoline sulfonamides (3-13) Benzenesulfonyl chloride 2 (1 mmol) was added to a solution of 8AQ 1 (1 mmol) in pyridine (5 ml), and the mixture was stirred at room temperature for 20 to 36 h; the completion of the reaction was monitored by thin-layer chromatography (30% acetone:hexane). The reaction mixture was quenched with water (30 ml), and the precipitated product was filtered by vacuum filtration. The obtained sulfonamides were recrystallized from MeOH-CH 2 Cl 2 to give the pure compounds ( 3 - 13 ). 4-Fluoro-N-(quinolin-8-yl)benzenesulfonamide (3) Brown solid. 74% Yield. Mp 130 to 131 °C. 1 H NMR (300 MHz, DMSO-d 6 ): δ (ppm) 7.31 (t, 3 J = 8.8 Hz, 2H, H -3′ and H -5′), 7.53 (t, 3 J = 8.8 Hz, 1H, H -6), 7.58 (dd, 3 J = 8.3 Hz; and 4 J = 4.2 Hz, 1H, H -3), 7.69 (d, 3 J = 7.8 Hz, 2H, H -5 and H -7), 7.96 (dd, 3 J = 8.8 Hz; and 4 J = 5.2 Hz, 2H, H -2′ and H -6′), 8.35 (dd, 3 J = 8.3 Hz; and 4 J = 1.2 Hz, 1H, H -4), 8.84 (dd, 3 J = 4.1 Hz; and 4 J = 1.3 Hz, 1H, H -2), 10.10 (br s, 1H, N H ). 13 C NMR (75 MHz, DMSO-d 6 ) δ 116.0 (d, 2 J CF = 23 Hz, C -3′ and C -5′), 117.2 ( C -7), 122.1 ( C -3), 123.2 ( C -5), 126.4 ( C -6), 128.0 ( C -4a), 129.8 (d, 3 J CF = 10 Hz, C -2′ and C -6′), 133.2 ( C -8), 135.8 (d, 4 J CF = 3 Hz, C -1′) 136.3 ( C -4), 138.8 ( C -8a), 149.2 ( C -2), 164.2 (d, 1 J CF = 252 Hz, C -4′). HRMS-TOF: mass/charge ratio ( m/z ) [M + H] + 303.0599 (calculated for C 15 H 12 FN 2 O 2 S: 303.0598). The spectroscopic data are in accordance with the reported literature [ 31 – 33 ]. 4-Chloro-N-(quinolin-8-yl)benzenesulfonamide (4) Brown solid. 72% Yield. Mp 128 to 129 °C. 1 H NMR (300 MHz, DMSO-d 6 ): δ (ppm) 7.50 to 7.59 (m, 4H, H -3, H -6, H -3′, and H -5′), 7.67 to 7.70 (m, 2H, H -5 and H -7), 7.89 (d, 3 J = 8.5 Hz, 2H, H -2′ and H -6′), 8.35 (dd, 3 J = 8.3 Hz; and 4 J = 1.3 Hz, 1H, H -4), 8.83 (dd, 3 J = 4.2 Hz; and 4 J = 1.4 Hz, 1H, H -2), 10.09 (br s, 1H, N H ). 13 C NMR (75 MHz, DMSO-d 6 ) δ 117.6 ( C -7), 122.2 ( C -3), 123.5 ( C -5), 126.5 ( C -6), 128.1 ( C -4a), 128.8, 129.1 ( C -2′, C -3′, C -5′, and C -6′), 133.3 ( C -8), 136.4 ( C -4), 137.9 ( C -4′), 138.4 ( C -1′), 139.0 ( C -8a), 149.4 ( C -2). HRMS-TOF: m/z [M + H] + 319.0282 (calculated for C 15 H 12 ClN 2 O 2 S: 319.0303). The spectroscopic data are in accordance with the reported literature [ 32 , 34 ]. 4-Bromo-N-(quinolin-8-yl)benzenesulfonamide (5) Brown solid. 60% Yield. Mp 161 to 162 °C. 1 H NMR (500 MHz, DMSO-d 6 ): δ (ppm) 7.53 (t, 3 J = 7.9 Hz, 1H, H -6), 7.58 (dd, 3 J = 8.3 Hz; and 4 J = 4.2 Hz, 1H, H -3), 7.66 to 7.70 (m, 4H, H -5, H -7, H -3′ and H -5′), 7.82 (d, 3 J = 8.7 Hz, 2H, H -2′ and H -6′), 8.35 (dd, 3 J = 8.3 Hz; and 4 J = 1.6 Hz, 1H, H -4), 8.84 (dd, 3 J = 4.2 Hz; and 4 J = 1.7 Hz, 1H, H -2), 10.14 (br s, 1H, N H ). 13 C NMR (125 MHz, DMSO-d 6 ) δ 117.8 ( C -7), 122.3 ( C -3), 123.5 ( C -5), 126.7 ( C -6), 127.0 ( C -4′), 128.2 ( C -4a), 129.0, 132.2 ( C -2′, C -3′, C -5′, and C -6′), 133.5 ( C -8), 136.6 ( C -4), 139.0, 139.2 ( C -8a and C -1′), 149.5 ( C -2). HRMS-TOF: m/z [M + Na] + 384.9625 (calculated for C 15 H 11 BrN 2 NaO 2 S: 384.9617). The spectroscopic data are in accordance with the reported literature [ 31 , 33 ]. N-(quinolin-8-yl)-4-(trifluoromethyl)benzenesulfonamide (6) Brown solid. 70% Yield. Mp 120 to 121 °C. 1 H NMR (500 MHz, DMSO-d 6 ): δ (ppm) 7.54 (t, 3 J = 7.9 Hz, 1H, H -6), 7.56 (dd, 3 J = 8.4 Hz; and 4 J = 4.2 Hz, 1H, H -3), 7.69 (dd, 3 J = 7.7 Hz; and 4 J = 1.2 Hz, 1H, H -5), 7.71 (dd, 3 J = 8.3 Hz; and 4 J = 1.0 Hz, 1H, H -7), 7.87 (d, 3 J = 8.3 Hz, 2H, H -3′ and H -5′), 8.08 (d, 3 J = 8.2 Hz, 2H, H -2′ and H -6′), 8.35 (dd, 3 J = 8.3 Hz; and 4 J = 1.6 Hz, 1H, H -4), 8.80 (dd, 3 J = 4.2 Hz; and 4 J = 1.7 Hz, 1H, H -2), 10.41 (br s, 1H, N H ). 13 C NMR (125 MHz, DMSO-d 6 ) δ 118.8 ( C -7), 122.3 ( C -3), 123.4 (q, 1 J CF = 252 Hz, C F 3 ), 124.0 ( C -5), 126.3 (q, 3 J CF = 3 Hz, C -3′ and C -5′), 126.7 ( C -6), 127.9 ( C -2′ and C -6′), 128.3 ( C -4a), 132.6 (q, 2 J CF = 32 Hz, C -4′), 133.4 ( C -8), 136.5 ( C -4), 139.5 ( C -8a), 143.7 ( C -1′), 149.5 ( C -2). HRMS-TOF: m/z [M + Na] + 375.0377 (calculated for C 16 H 11 F 3 N 2 NaO 2 S: 375.0386). The spectroscopic data are in accordance with the reported literature [ 31 , 32 ]. 4-Acetyl-N-(quinolin-8-yl)benzenesulfonamide (7) Brown solid. 91% Yield. Mp 124 to 125 °C. 1 H NMR (300 MHz, DMSO-d 6 ): δ (ppm) 2.54 (s, 3H, C H 3 ), 7.53 (t, 3 J = 8.3 Hz, 1H, H -6), 7.58 (dd, 3 J = 8.3 Hz; and 4 J = 4.2 Hz, 1H, H -3), 7.68 to 7.71 (m, 2H, H -5 and H -7), 8.00 (d, 3 J = 8.9 Hz, 2H, H -3′ and H -5′), 8.05 (d, 3 J = 8.8 Hz, 2H, H -2′ and H -6′), 8.35 (dd, 3 J = 8.3 Hz; and 4 J = 1.6 Hz, 1H, H -4), 8.84 (dd, 3 J = 4.2 Hz; and 4 J = 1.6 Hz, 1H, H -2), 10.29 (br s, 1H, N H ). 13 C NMR (75 MHz, DMSO-d 6 ) δ 26.9 ( C H 3 ), 117.5 ( C -7), 122.3 ( C -3), 123.5 ( C -5), 126.6 ( C -6), 127.2 ( C -3′ and C -5′), 128.1 ( C -4a), 128.8 ( C -2′ and C -6′), 133.3 ( C -8), 136.5 ( C -4), 139.0 ( C -8a), 139.8 ( C -4′), 143.2 ( C -1′), 149.4 ( C -2), 197.7 ( C =O). HRMS-TOF: m/z [M + Na] + 349.0617 (calculated for C 17 H 14 N 2 NaO 3 S: 349.0617). 4-Cyano-N-(quinolin-8-yl)benzenesulfonamide (8) Brown solid. 74% Yield. Mp 121 to 122 °C. 1 H NMR (300 MHz, DMSO-d 6 ): δ (ppm) 7.54 (t, 3 J = 8.0 Hz, 1H, H -6), 7.57 (dd, 3 J = 8.4 Hz; and 4 J = 4.3 Hz, 1H, H -3), 7.69 (d, 3 J = 7.7 Hz, 1H, H -5), 7.73 (d, 3 J = 8.3 Hz, 1H, H -7), 7.96 (d, 3 J = 8.2 Hz, 2H, H -3′ and H -5′), 8.04 (d, 3 J = 8.3 Hz, 2H, H -2′ and H -6′), 8.36 (dd, 3 J = 8.3 Hz; and 4 J = 1.2 Hz, 1H, H -4), 8.81 (dd, 3 J = 4.2 Hz; and 4 J = 1.2 Hz, 1H, H -2), 10.45 (br s, 1H, N H ). 13 C NMR (75 MHz, DMSO-d 6 ) δ 115.3 ( C -4′), 117.5 ( C≡ N), 118.9 ( C -7), 122.3 ( C -3), 124.1 ( C -5), 126.6 ( C -6), 127.6 ( C -3′ and C -5′), 128.2 ( C -4a), 133.1 ( C -8, C -2′ and C -6′), 136.5 ( C -4), 139.5 ( C -8a), 143.9 ( C -1′), 149.5 ( C -2). HRMS-TOF: m/z [M + Na] + 332.0461 (calculated for C 16 H 11 N 3 NaO 2 S: 332.0464). The spectroscopic data are in accordance with the reported literature [ 31 , 32 ]. 2-Nitro-N-(quinolin-8-yl)benzenesulfonamide (9) Yellow solid. 72% Yield. Mp 144 to 145 °C. 1 H NMR (500 MHz, DMSO-d 6 ): δ (ppm) 7.57 (t, 3 J = 8.0 Hz, 1H, H -6), 7.61 (dd, 3 J = 8.3 Hz; and 4 J = 4.2 Hz, 1H, H -3), 7.74 to 7.76 (m, 2H, H -5 and H -7), 7.77 (td, 3 J = 7.8 Hz; and 4 J = 1.4 Hz, 1H, H -4′), 7.82 (td, 3 J = 7.6 Hz; and 4 J = 1.4 Hz, 1H, H -5′), 7.99 (dd, 3 J = 7.9 Hz; and 4 J = 1.3 Hz, 1H, H -6′), 8.18 (dd, 3 J = 7.8 Hz; and 4 J = 1.4 Hz, 1H, H -3′), 8.39 (dd, 3 J = 8.3 Hz; and 4 J = 1.6 Hz, 1H, H -4), 8.86 (dd, 3 J = 4.2 Hz; and 4 J = 1.6 Hz, 1H, H -2), 10.21 (br s, 1H, N H ). 13 C NMR (125 MHz, DMSO-d 6 ) δ 117.5 ( C -7), 122.6 ( C -3), 124.0 ( C -5), 125.2 ( C -6′), 126.8 ( C -6), 128.2 ( C -4a), 130.6 ( C -3′), 131.4 ( C -1′), 132.7 ( C -8), 132.9 ( C -4′), 135.0 ( C -5′), 136.7 ( C -4), 138.9 ( C -8a), 147.6 ( C -2′), 149.7 ( C -2). HRMS-TOF: m/z [M + Na] + 352.0361 (calculated for C 15 H 11 N 3 NaO 4 S: 352.0362). The spectroscopic data are in accordance with the reported literature [ 35 ]. 4-Nitro-N-(quinolin-8-yl)benzenesulfonamide (10) Yellow solid. 87% Yield. Mp 158 to 159 °C. 1 H NMR (300 MHz, DMSO-d 6 ): δ (ppm) 7.52 to 7.58 (m, 2H, H -3, H -6), 7.71 (dd, 3 J = 7.3 Hz; and 4 J = 1.0 Hz, 1H, H -5), 7.73 (dd, 3 J = 7.9 Hz; and 4 J = 1.0 Hz, 1H, H -7), 8.14 (d, 3 J = 8.8 Hz, 2H, H -3′ and H -5′), 8.29 (d, 3 J = 8.8 Hz, 2H, H -2′ and H -6′), 8.35 (dd, 3 J = 8.3 Hz; and 4 J = 1.5 Hz, 1H, H -4), 8.80 (dd, 3 J = 4.2 Hz; and 4 J = 1.6 Hz, 1H, H -2), 10.56 (br s, 1H, N H ). 13 C NMR (75 MHz, DMSO-d 6 ) δ 118.8 ( C -7), 122.1 ( C -3), 123.9 ( C -5), 124.0 ( C -3′ and C -5′), 126.4 ( C -6), 128.1 ( C -4a), 128.3 ( C -2′ and C -6′), 132.9 ( C -8), 136.3 ( C -4), 139.4 ( C -8a), 145.2 ( C -1′), 149.3 ( C -2), 149.7 ( C -4′). HRMS-TOF: m/z [M + Na] + 352.0359 (calculated for C 15 H 11 N 3 NaO 4 S: 352.0362). The spectroscopic data are in accordance with the reported literature [ 32 , 33 ]. 2,3,5,6-Tetramethyl-N-(quinolin-8-yl)benzenesulfonamide (11) Brown solid. 90% Yield. Mp 206 to 207 °C. 1 H NMR (300 MHz, DMSO-d 6 ): δ (ppm) 2.14 (s, 6H, 2 × C H 3 ), 2.54 (s, 6H, 2 × C H 3 ), 7.16 (s, 1H, H -4′), 7.40 to 7.48 (m, 2H, H -3 and H -6), 7.59 to 7.63 (m, 2H, H -5 and H -7), 8.36 (dd, 3 J = 8.3 Hz; and 4 J = 1.5 Hz, 1H, H -4), 8.89 (dd, 3 J = 4.2 Hz; and 4 J = 1.5 Hz, 1H, H -2), 9.72 (br s, 1H, N H ). 13 C NMR (75 MHz, DMSO-d 6 ) δ 17.5 (2 × C H 3 ), 20.4 (2 × C H 3 ), 115.1 ( C -7), 122.4 ( C -3), 122.5 ( C -5), 126.7 ( C -6), 128.1 ( C -4a), 133.6 ( C -8), 134.7, 135.70 ( C -2′, C -3′, C -5′, and C -6′), 135.72 ( C -4), 136.6 ( C -4′), 137.6 ( C -1′), 138.2 ( C -8a), 149.3 ( C -2). HRMS-TOF: m/z [M + Na] + 363.1136 (calculated for C 19 H 20 N 2 NaO 2 S: 363.1138). N-(quinolin-8-yl)naphthalene-2-sulfonamide (12) Brown solid. 60% Yield. Mp 138 to 139 °C. 1 H NMR (500 MHz, DMSO-d 6 ): δ (ppm) 7.50 (t, 3 J = 8.0 Hz, 1H, H -6), 7.55 (dd, 3 J = 8.3 Hz; and 4 J = 4.2 Hz, 1H, H -3), 7.60 to 7.67 (m, 3H, H -5, H -6′ and H -7′) 7.74 (dd, 3 J = 7.6 Hz; and 4 J = 1.2 Hz, 1H, H -7), 7.92 (dd, 3 J = 8.7 Hz; and 4 J = 1.9 Hz, 1H, H -5′), 7.94 (dd, 3 J = 8.0 Hz; and 4 J = 1.0 Hz, 1H, H -8′), 7.99 (d, 3 J = 8.8 Hz, 1H, H -4′), 8.10 (dd, 3 J = 7.9 Hz; and 4 J = 0.8 Hz, 1H, H -3′) 8.31 (dd, 3 J = 8.3 Hz; and 4 J = 1.6 Hz, 1H, H -4), 8.61 (d, 4 J = 1.6 Hz, 1H, H -1′), 8.84 (dd, 3 J = 4.2 Hz; and 4 J = 1.7 Hz, 1H, H -2), 10.11 (br s, 1H, N H ). 13 C NMR (125 MHz, DMSO-d 6 ) δ 116.8 ( C -7), 122.3 ( C -3), 122.4 ( C -5′), 123.1 ( C -5), 126.7 ( C -6), 127.7 ( C -7′), 127.8 ( C -8′), 128.1 ( C -4a), 128.3 ( C -1′), 129.1 ( C -6′), 129.2 ( C -3′), 129.3 ( C -4′), 131.5 ( C -9′), 133.6 ( C -8), 134.3 ( C -10′), 136.5 ( C -2′), 136.6 ( C -4), 138.8 ( C -8a), 149.4 ( C -2). HRMS-TOF: m/z [M + Na] + 357.0667 (calculated for C 19 H 14 N 2 NaO 2 S: 357.0668). The spectroscopic data are in accordance with the reported literature [ 31 , 32 , 36 ]. 4-(Methylsulfonyl)-N-(quinolin-8-yl)benzenesulfonamide (13) Brown solid. 92% Yield. Mp 178 to 179 °C. 1 H NMR (300 MHz, DMSO-d 6 ): δ (ppm) 3.23 (s, 3H, C H 3 ), 7.55 (t, 3 J = 8.1 Hz, 1H, H -6), 7.57 (dd, 3 J = 8.3 Hz; and 4 J = 4.5 Hz, 1H, H -3), 7.69 to 7.74 (m, 2H, H -5 and H -7), 8.03 (d, 3 J = 8.4 Hz, 2H, H -2′ and H -6′), 8.15 (d, 3 J = 8.5 Hz, 2H, H -3′ and H -5′), 8.36 (dd, 3 J = 8.3 Hz; and 4 J = 1.5 Hz, 1H, H -4), 8.81 (dd, 3 J = 4.2 Hz; and 4 J = 1.5 Hz, 1H, H -2), 10.45 (br s, 1H, N H ). 13 C NMR (75 MHz, DMSO-d 6 ) δ 42.9 ( C H 3 ), 118.3 ( C -7), 122.1 ( C -3), 123.7 ( C -5), 126.4 ( C -6), 127.6, 127.7 ( C -2′, C -3′, C -5′, and C -6′), 128.0 ( C -4a), 133.0 ( C -8), 136.3 ( C -4), 139.2 ( C -8a), 144.2 ( C -1′), 144.4 ( C -4′), 149.2 ( C -2). HRMS-TOF: m/z [M + Na] + 385.0291 (calculated for C 16 H 14 N 2 NaO 4 S 2 : 385.0287). The spectroscopic data are in accordance with the reported literature [ 32 ]. Antioxidant activity assay The compounds were investigated for their antioxidant activities using DPPH and SOD assays as described below. DPPH is a stable free radical that produces a violet solution in methanol. The DPPH radical is reduced by an antioxidant molecule, which donates an electron or a hydrogen atom, to give light-yellow product of DPPH [ 37 ]. The assay was initiated by adding 1 ml of 0.1 mM DPPH (dissolved in methanol) to the tested compounds dissolved in DMSO with the final concentration of 300 μg/ml, and the reaction mixture was incubated at room temperature in the dark for 30 min. The absorbance of the reaction was measured at 517 nm using an ultraviolet (UV)–visible spectrophotometer (UV-1610, Shimadzu). The standard antioxidant α -tocopherol was employed as a control compound, and DMSO solvent was used as the blank reaction. Experiments were performed in triplicate. The percentage of radical scavenging activity or DPPH was calculated using Eq. (1) : DPPH % = 1 − Abs . sample Abs . control × 100 (1) where Abs . control is the absorbance of the control reaction, and Abs. sample is the absorbance of the tested compound. Superoxide anion is a free radical that can be neutralized by SOD and SOD-like compounds. The SOD activity was evaluated using SOD assay [ 38 ]. The stock solution containing 27 ml of Hepes buffer (50 mM, pH 7.8), 1.5 ml of L-methionine (30 mg/ml), 1 ml of nitro blue tetrazolium chloride (1.41 mg/ml), and 750 μl of Triton X-100 (1 wt%) was prepared, and then 1 ml of the solution was added to the tested compounds dissolved in DMSO with the final concentration of 300 μg/ml. The reaction was initiated by adding 10 μl of riboflavin (44 mg/ml) and followed by illumination under a Philips Classic Tone lamp (60 W) in a light box for 7 min. The absorbance of the reaction was measured at 550 nm using a UV–visible spectrophotometer (UV-1610, Shimadzu). SOD from bovine erythrocytes served as a control substance. Experiments were performed in triplicate. The percentage of SOD activity was computed using Eq. (2) : SOD % = 1 − Abs . sample Abs . control × 100 (2) where Abs . control is the absorbance of the control reaction and Abs. sample is the absorbance of the tested compound. The compounds exhibiting antioxidant activities (DPPH and SOD) greater than 50% at 300 μg/ml were further determined for their IC 50 values by the 2-fold dilution method. The IC 50 is half-maximal inhibitory concentration value for measurement of the compound concentration to inhibit a free radical by 50%. Plotting between antioxidant activities (%DPPH or %SOD) against compound concentrations was performed to obtain the corresponding IC 50 values. Antimicrobial activity assay The compounds were determined for antimicrobial activity using the conventional agar dilution method as described by Clinical & Laboratory Standards Institute (CLSI) guidelines [ 39 ] against 29 strains of microorganisms including reference strains and clinical isolates: gram-positive bacteria: Staphylococcus aureus ATCC 29213, Staphylococcus aureus ATCC 25923, Staphylococcus epidermidis ATCC 12228, Enterococcus faecalis ATCC 29212, Enterococcus faecalis ATCC 33186, Micrococcus luteus ATCC 10240, Bacillus subtilis ATCC 6633, Corynebacterium diphtheria NCTC 10356, methicillin-resistant Staphylococcus aureus JCSC 3063, methicillin-resistant Staphylococcus aureus N315, methicillin-resistant Staphylococcus aureus JCSC 4788, Bacillus cereus , and Listeria monocytogenes ; gram-negative bacteria: Escherichia coli ATCC 25922, Klebsiella pneumonia ATCC 700603, Serratia marcescens ATCC 8100, Salmonella typhimurium ATCC 13311, Salmonella choleraesuis ATCC 10708, Shewanella putrefaciens ATCC 8671, Achromobacter xylosoxidans ATCC 27061, Pseudomonas aeruginosa ATCC 27853, Pseudomonas stutzeri ATCC 17587, Salmonella enteritidis , Morganella morganii , Aeromonas hydrophilia , Citrobacter freundii , and Plesiomonas shigellloides ; and diploid fungus (yeast): Candida albicans ATCC 90028 and Saccharomyces cerevisiae ATCC 2601. The tested compounds and the reference antibacterial agents (i.e., ampicillin, ciprofloxacin, and tetracycline) were dissolved in DMSO, and the 2-fold dilution was further performed using MHB. Each dilution (1 ml) was added to the sterile 19-ml Mueller Hinton agar to give the final concentrations in the range of 256 to 4 μg/ml. The microorganisms were cultured in MHB at 37 °C for 24 h and suspended in 0.9% normal saline solution to adjust an optical density at 600 nm of 0.1 for a cell density of 1 × 10 8 CFU/ml compared with the 0.5 McFarland turbidity standard. After that, 1 μl of microorganism suspensions (1 × 10 4 CFU/spot) was inoculated onto the plates having a variety of compound concentrations using a multipoint inoculator and further incubated at 37 °C for 24 to 48 h. The plates containing DMSO and MHB without any antibacterial agents were simultaneously used as controls. The inhibition of microbial cell growth of tested compounds was investigated to obtain a minimum inhibitory concentration (MIC), which is the lowest concentration to inhibit the growth of microorganisms. In addition, the MIC quality control ranges, displayed in micrograms per milliliter of reference antibacterial agents according to CLSI, were investigated for the control system [ 40 ]. QSAR analysis Dataset and data preprocessing Experimentally determined antioxidant activities (%DPPH and SOD IC₅₀) and antimicrobial activities (MIC, μM) were used as response end points for developing the QSAR and QSPR models, respectively. Three datasets were separately prepared to construct 3 models corresponding to each investigated assay. Each dataset contains (a) molecular descriptors (assigned as independent variables, X n ) as structural representatives obtained by calculations and (b) bioactivity values (assigned as dependent variables, Y ) obtained for experimental results. Experimental results obtained from DPPH and SOD assays were in numerical form (%DPPH or SOD IC₅₀), while those from the antimicrobial assays were in category form (active or inactive). Accordingly, 2 QSAR (DPPH and SOD) models and 1 QSPR (antimicrobial) model were constructed. For QSAR model construction, only active compounds displaying numerical %DPPH or SOD IC 50 values were included in the datasets, whereas those inactive ones were discarded. Experimental antioxidant activity of SOD, expressed in IC 50 (μM) values, was converted to pIC 50 (M) by taking negative logarithmic transformation to the base of 10 [−log 10 (IC 50 )], whereas the antioxidant activity of DPPH exhibited in percentage (%) was directly utilized. Therefore, the compound displaying a high pIC 50 (low IC 50 ) value is interpreted to have high antioxidant activity. Regarding antimicrobial activity, the compounds were classified as active or inactive based on their MIC values against the tested microorganisms. Calculation of quantum chemical and molecular descriptors Molecular structures of the compounds were constructed using Chemdraw Pro13 software (PerkinElmer, USA) and subjected to GaussView software [ 41 ]. The structures were initially saved as *.smi files and converted to *.mol files using OpenBabel version 2.3.2. The prepared *.mol files were subsequently used as input files for computing 2-dimensional (2D) descriptors using PaDEL [ 42 ] and Mold 2 [ 43 ] software to obtain a set of 1,444 PaDEL 0D-2D descriptors and 777 Mold2 2D descriptors, respectively. Geometrical optimization was performed using Gaussian 09, Revision A.02 [ 44 ] software at the semiempirical level using Austin Model 1, followed by density functional theory calculation using Becke’s 3-parameter hybrid method and the Lee–Yang–Parr correlation functional (B3LYP) together with the 6-31 g(d) basis set. The low-energy conformers of the compounds were received and were consequently used to compute a set of 13 quantum chemical descriptors including the total energy ( E total ) of the molecule, the highest occupied molecular orbital energy ( E HOMO ), the lowest unoccupied molecular orbital energy ( E LUMO ), the total dipole moment ( μ ) of the molecule, the electron affinity, the ionization potential, the energy difference of HOMO and LUMO (HOMO–LUMO Gap ), Mulliken electronegativity ( χ ), hardness ( η ), softness ( S ), electrophilicity ( ω ), electrophilic index ( ω i ), and the mean absolute atomic charge ( Q m ) [ 17 ]. The optimized structures were further used as input files for calculating a set of 3,224 molecular descriptors using Dragon software, version 5.5 [ 45 ]. Descriptor variables with constant and redundant values were initially removed by Dragon software to give a remaining set of 1,308 molecular descriptors. Finally, a total set of 3,542 descriptors (including 1,308 molecular descriptors, 13 quantum chemical descriptors, 1,444 PaDEL 0D-2D descriptors, and 777 Mold2 2D descriptors) was obtained and subsequently subjected to a feature selection process. Feature selection A final set of 3,542 calculated descriptors underwent a feature selection process to select only a set of significant descriptors that are correlated with bioactivities using stepwise multiple linear regression (MLR) (SPSS Statistics 18.0, SPSS Inc., USA) or by automatic variable selection procedure (CfsSubsetEval combined with the BestFirst) as implemented in WEKA software, version 3.4.5 [ 46 ] or by the UFS algorithm (UFS software package, version 1.8) [ 47 ]. The selected descriptors were further investigated for their intercorrelations. Pearson’s correlation coefficient ( r ) values between each pair of selected descriptors within the same model were calculated using SPSS Statistics 18.0 (SPSS Inc., USA), and a cutoff value of | r | ≤ 0.9 was used to determine their independence. Finally, the selected descriptors were further used to prepare final datasets for model construction. Model construction MLR Two QSAR models (i.e., DPPH and SOD) were constructed using the MLR method. Molecular descriptors were assigned as independent variables ( X ), and biological activity values were assigned as dependent variables ( Y ). The QSAR models were generated according to Eq. (3) : Y = m 1 x 1 + m 2 x 2 + … + m n x n + b (3) where Y is the biological activity values of the compounds, m is the regression coefficient values of the descriptors x , and b is the intercept. The MLR method was performed using Weka software, version 3.4.5 [ 46 ]. Decision tree analysis An antimicrobial QSPR model was constructed using decision tree analysis. The decision tree is a supervised method that generates a set of if-then rules. It finds and explores the most essential independent variables ( X ) to classify compounds with interest activity ( Y ). The tree exhibited the top–down manner starting from the root node through the internal nodes using the independent variables and finally to the terminal leaf nodes for the class prediction [ 48 – 50 ]. Herein, the J48 decision tree algorithm, a derivative of the C4.5 algorithm implemented in Weka software version 3.4.5 [ 46 ], was used to construct the QSPR model for classifying the end points of the compounds of interest as active or inactive. Generation of dataset The dataset was divided into training and testing sets. The training set was used to generate the predictive models, whereas the testing set was employed to validate the models. The testing set was generated by leave-one-out cross-validation (LOO-CV), in which 1 sample was left out from the whole dataset ( N ) to be used as a testing set while the remaining samples ( N − 1) were used as the training set. This principle was iteratively continued until every sample was used as the testing set [ 17 ]. Additionally, 5-fold cross-validation (5-fold-CV) was performed to evaluate and compare the predictive performance of the QSAR/QSPR models with the results obtained from LOO-CV [ 30 ]. Evaluation of model predictive performance Statistical parameters were calculated to validate the predictive performance of the constructed models. For the QSAR models, correlation coefficient parameters (i.e., squared correlation coefficient [ R Tr 2 ] and cross-validated R 2 [ Q L O O − C V 2 and Q 2 5-fold-CV ] for training, LOO-CV, and 5-fold-CV sets, respectively) and root mean square error (RMSE) were calculated to determine the predictive error of the models (i.e., training [ RMSETr ], LOO-CV [ RMSE LOO-CV ], and 5-fold-CV [ RMSE 5-fold-CV ] sets) [ 17 , 30 ]. Furthermore, adjusted R 2 ( R 2 adj ), mean absolute error (MAE), and concordance correlation coefficient (CCC) were computed to evaluate the robustness of the constructed QSAR models [ 51 – 53 ]. For the QSPR decision tree model, the classification model provided descriptor-based cutoff values to classify compounds as active or inactive. Statistical indices including accuracy, precision, recall, and F-measure were used to assess the predictive performance of the model (i.e., training, LOO-CV, and 5-fold-CV sets) [ 54 ]. Application of the constructed QSAR models for guiding design of new derivatives After validating their reliability, the constructed models were used for the rational design of new derivatives, in which descriptor variables presented in the models were used for guiding structural modification on the core structures of the selected prototypes (compounds 3 , 4 , 5 , 6 , 7 , 8 , 9 , 11 , 12 , and 13 ) to finally give an additional set of 84 newly designed derivatives. All newly designed compounds were drawn, geometrically optimized, and calculated for their descriptor values in the same manner as those of prototypes (as described in the “Calculation of quantum chemical and molecular descriptors” section). The obtained key descriptor values were subsequently used for predicting their SOD and DPPH activities using the constructed models. A similar process was performed by applying the constructed antimicrobial QSPR model for predicting antimicrobial classes (active or inactive) of the newly designed compounds. Results and Discussion Chemistry A set of quinoline–sulfonamides ( 3-13 ) was synthesized by N -sulfonylation of 8AQ 1 with the corresponding benzenesulfonyl chloride 2 in pyridine at room temperature in moderate to good yields (60% to 92%) as depicted in Fig. 1 [ 33 ]. The structures of the sulfonamides ( 3-13 ) were confirmed by means of spectroscopic methods, namely, HRMS, 1 H NMR, and 13 C NMR. All sulfonamides had molecular ion peaks consistent with their molecular formulas. 1 H NMR spectra of sulfonamides ( 3-13 ) displayed signals of aromatic protons of both the quinoline ring and the sulfonyl part (R), indicating that the N -sulfonation products were formed. The spectroscopic data of the known compounds 3 - 6 , 8 - 10 , and 12 - 13 are in accordance with those reported in the literature [ 31 – 36 ]. In 1 H NMR, these derivatives showed the NH signal of sulfonamide group at a chemical shift of about 10 ppm. The quinoline protons at the C-2 and C-4 positions typically appeared as doublet of doublets (dd) at the low-field chemical shift around 8.8 and 8.3 ppm, respectively. The title sulfonamides ( 3 - 13 ) have a common R group (benzene) substituted with various substituents (X = F, Cl, Br, CF 3 , CN, COCH 3 , NO 2 , SO 2 CH 3 , and 2,3,5,6-tetramethyl) at ortho ( o ), meta ( m ), and para- ( p ) positions, except for compound 13 , R = naphthalene ring. 1 H and 13 C NMR spectral data of compounds 3 - 13 are provided in Supplementary Materials (Figs. S1 to S22 ). Fig. 1. Open in a new tab Synthesis of 8-aminoquinoline (8AQ)-based sulfonamide derivatives (3-13) [ 33 ]. Antioxidant activities The synthesized 8AQ-based sulfonamides ( 3 - 13 ) were evaluated for their antioxidant activities using DPPH and SOD assays to measure their scavenging capabilities toward DPPH free radicals and superoxide anions (O 2 •- ), respectively. The DPPH assay is a simple method to investigate radical scavenging activities, known as DPPH activity [ 37 ]. All tested compounds ( 3 - 13 ; Table 1 ) exhibit weak DPPH radical scavenging activity, with inhibition values ranging from 7.74% to 36.49% at 300 μg/ml. Since none of the compounds achieved 50% inhibition, IC 50 values were not determined. Among all, o -NO 2 -substituted compound 9 and tetramethyl-substituted compound 11 (%DPPH = 36.49% and 22.43%, respectively) exhibited the most potent radical scavenging effect, whereas a bromo-containing compound 5 displayed the weakest activity (%DPPH = 7.74%). Considering the halogen-containing compounds (Table 1 ), compounds 3 , 4 , and 5 with mono-halogen substitution (X = F [10.09%], Cl [9.26%], and Br [7.74%], respectively) showed lower activity than the trifluoro compound 6 (X = CF 3 , 12.06%). However, all tested compounds exhibited considerably lower DPPH radical scavenging activity than that of the standard antioxidant α -tocopherol (IC₅₀ = 12.82 μM), suggesting their limited potential for effective therapeutics. Table 1. Antioxidant activities of 8AQ-based sulfonamides (3-13). The standard antioxidant α -tocopherol was used as a control in DPPH assay (IC 50 = 12.82 μM), and SOD from bovine erythrocytes was used as a control in SOD assay (IC 50 = 0.002 μM). Compound DPPH activity (%) a SOD activity (IC 50 , μM) 10.09 153.56 9.26 ND b 7.74 212.58 12.06 247.39 9.79 250.37 12.53 96.07 36.49 254.58 19.06 448.75 22.43 530.93 13.88 83.34 17.33 600.81 Open in a new tab 8AQ, 8-aminoquinoline; SOD, superoxide dismutase; ND, not determined a Compounds were tested at 300 μg/ml, and half-maximal inhibitory concentration (IC 50 ) values were not determined because compounds exhibited 2,2-diphenyl-1-picrylhydrazyl (DPPH) activity <50%. b Compound 4 (17.88% SOD) displayed SOD activity <50%; therefore, IC 50 was not determined. The synthesized compounds ( 3 - 13 ) were investigated for their abilities to scavenge superoxide anion (SOD-mimic) activity [ 38 ]. All tested compounds, except for the chloro-containing compound 4 (%SOD = 17.88% at 300 μg/ml), showed SOD-mimic activity greater than 50%. Therefore, their IC₅₀ values were subsequently determined, providing the SOD IC 50 values in the range of 83.34 to 600.81 μM (Table 1 ). Naphthalene-substituted compound 12 displayed the highest SOD activity (IC 50 = 83.34 μM), followed by CN-substituted compound 8 (IC 50 = 96.07 μM). However, these compounds displayed considerably lower activity when compared to that of the control SOD enzyme (IC₅₀ = 0.002 μM), suggesting their modest SOD activity. Comparing the 2 antioxidant assays, the tested compounds ( 3 - 13 ) exhibited more potent SOD activity than DPPH activity. The isomeric effect of NO 2 substition on the benzene ring was noted for 2 NO 2 derivatives 9 and 10 . It was observed that the substitution of NO 2 on the o- position of the benzene ring provided better antioxidant effects (both DPPH and SOD) than those with the p- position. Compound 9 (X = o- NO 2 , %DPPH = 36.49%, SOD IC 50 = 254.58 μM) displayed more potent SOD and DPPH activities than compound 10 (X = p- NO 2 , %DPPH = 19.06%, SOD IC 50 = 448.75 μM), as shown in Table 1 . Antimicrobial activity All synthesized compounds were investigated for their antimicrobial activities against gram-positive and gram-negative bacteria as well as diploid fungi using the agar dilution method [ 39 ]. It is an assay used to evaluate the ability of compounds to inhibit the growth of microorganisms. The mechanism of action of the tested compounds was not investigated because the main objective was to screen for growth-inhibiting activity. The MIC quality control ranges for reference antibacterial agents (including ampicillin, ciprofloxacin, and tetracycline) were performed to evaluate the control system according to CLSI [ 40 ]. It was found that the ampicillin displayed an MIC value of 8 μg/ml to inhibit the growth of E. coli ATCC 25922 and an MIC value of 2 μg/ml for E. faecalis ATCC 29212, in which the acceptable MIC quality control ranges of E. coli ATCC 25922 and E. faecalis ATCC 29212 were 2 to 8 μg/ml and 0.5 to 2 μg/ml, respectively. The ciprofloxacin exhibited MIC values of 0.25 μg/ml against S. aureus ATCC 29213 and MIC values of 0.5 μg/ml against E. faecalis ATCC 29212 and P. aeruginosa ATCC 27583. The acceptable MIC quality control ranges of ciprofloxacin against the microorganism were 0.12 to 0.5 μg/ml, 0.25 to 2 μg/ml, and 0.25 to 1 μg/ml for S. aureus ATCC 29213, E. faecalis ATCC 29212, and P. aeruginosa ATCC 27583, respectively. The tetracycline inhibited the growth of S. aureus ATCC 29213, E. faecalis ATCC 29212, E. coli ATCC 25922, and P. aeruginosa ATCC 27583 at MIC values of 1, 32, 2, and 32 μg/ml with the acceptable MIC quality control ranges of 0.12 to 1 μg/ml, 8 to 32 μg/ml, 0.5 to 2 μg/ml, and 8 to 32 μg/ml, respectively. Furthermore, both DMSO solvent and MHB have no effect on microbial growth, and no contaminations were observed. Collectively, the results from MIC quality control testing confirmed the reliability of the prepared system for compounds’ investigations and indicated the interpretability of the results. The MIC values expressed in micrograms per milliliter (μg/ml) are provided in Table 2 . Most of the 8AQ-sulfonamides exhibited preferable antimicrobial activity, except for 4 inactive compounds ( 7 , 9 , 11 , and 13 ). The MIC values in units of μg/ml were converted (using calculated molecular weight from Dragon software, version 5.5) to micromolar units (μM) for comparing the effective activity of the tested compounds. It was found (Table 2 ) that compounds 3 , 4 , 5 , and 6 with halogen groups displayed excellent antimicrobial activity (MIC = ≤4 to 128 μg/ml or ≤11.01 to 423.35 μM), followed by compounds 8 with nitrile (CN) group (MIC = ≤4 to 128 μg/ml or ≤12.93 to 413.74 μM), compound 10 with nitro group (NO 2 ) (MIC = 128 μg/ml or 388.63 μM), and naphthalene-substituted compound 12 (MIC = 256 μg/ml or 765.50 μM). Most of the active compounds ( 3 , 4 , 5 , 6 , 8 , and 10 ) showed significant growth-inhibiting effects against gram-positive bacteria (i.e., S. aureus ATCC 29213, S. aureus ATCC 25923, S. epidermidis ATCC 12228, M. luteus ATCC 10240, B. subtilis ATCC 6633, MRSA JSCS 4788, MRSA N315, MRSA JCSC 3063, B. cereus , and L. monocytogenes , MIC ≤4 to 128 μg/ml or ≤11.01 to 423.35 μM), whereas compound 12 (R = naphthalene) mainly displayed antimicrobial activity against gram-negative bacteria (i.e., E. coli ATCC 25922, S. typhimurium ATCC 13311, P. stutzeri ATCC 17587, S. enteritidis , M. morganii , A. hydrophila , and C. freundii , MIC = 256 μg/ml or 765.50 μM). Interestingly, most of the tested compounds inhibited growth of the methicillin-resistant Staphylococcus aureus (MRSA) providing the MIC range of ≤4 to 64 μg/ml or ≤11.01 to 211.68 μM. In contrast, compounds 7 , 9 , 11 , and 13 (X = p -COCH 3 , o -NO 2 , 2,3,5,6-tetra-CH 3 , and p -SO 2 CH 3 ) exhibited inactive antimicrobial activities. Table 2. Antimicrobial activity (MIC) of 8AQ-based sulfonamides (3-13). Compounds 7, 9, 11, and 13 displayed no antimicrobial activity at a concentration of 256 μg/ml. Molecular weight (MW) calculated from Dragon software, version 5.5 of compounds 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, and 13 are 302.35, 318.80, 363.25, 352.36, 326.40, 309.37, 329.36, 329.36, 340.48, 334.42, and 362.46 g/mol, respectively. The MW of the compounds was used to convert MIC values in μg/ml to μM. Compound MIC Microorganism μg/ml μM 3 128 423.35 B. cereus 64 211.68 S. aureus ATCC 29213, S. aureus ATCC 25923, M. luteus ATCC 10240, MRSA JCSC 4788 8 26.46 MRSA N315 ≤4 ≤13.23 B. subtilis ATCC 6633, MRSA JCSC 3063 4 64 200.75 S. aureus ATCC 29213, S. epidermidis ATCC 12228 16 50.19 P. shigelloides , L. monocytogenes ≤4 ≤12.55 S. aureus ATCC 25923, M. luteus ATCC 10240 B. subtilis ATCC 6633, B. cereus , MRSA JCSC 3063, MRSA N315, MRSA JCSC 4788 5 8 22.02 MRSA N315 ≤4 ≤11.01 B. subtilis ATCC 6633, MRSA JCSC 3063 6 ≤4 ≤11.35 S. aureus ATCC 25923, B. subtilis ATCC 6633, B. cereus , P. shigelloides , L. monocytogenes , MRSA JCSC 3063, MRSA N315, MRSA JCSC 4788 8 128 413.74 A. hydrophila ≤4 ≤12.93 S. aureus ATCC 25923, B. subtilis ATCC 6633, B. cereus , MRSA JCSC 3063, MRSA N315 10 128 388.63 B. cereus 12 256 765.50 E. coli ATCC 25922, S. typhimurium ATCC 13311, P. stutzeri ATCC 17587, S. enteritidis type C , M. morganii , A. hydrophila , C. freundii Open in a new tab MIC: Minimum inhibitory concentration is the lowest concentration to inhibit the growth of microorganisms; 8AQ, 8-aminoquinoline In overview, the compounds containing halogen atom (F, Cl, Br, and CF 3 ), nitrile (CN), and naphthalenyl (C 10 H 7 ) are active antimicrobial agents. Among others, compound 3 (X = F) and compound 4 (X = Cl) displayed the most effective antimicrobial activity against various micoorganisms (MIC range: compound 3 = ≤4 to 128 μg/ml or ≤13.23 to 423.35 μM and compound 4 = ≤4 to 64 μg/ml or ≤12.55 to 200.75 μM). The halogen atoms, especially F and Cl atoms, are well recognized as key functional groups found in several classes of antimicrobial scaffolds (i.e., β -lactam antibiotics, sulfa drugs, and aminoglycosides). Halogenated antimicrobial agents displayed diverse inhibitory mechanisms against microorganisms via acting as inhibitors of cell wall synthesis, cell proliferation, and protein synthesis [ 55 ]. Although 8AQ itself has been reported as an inactive antimicrobial agent [ 12 ], its notable antimicrobial effect as well as antimalarial and antioxidant properties were observed when it was combined with other compounds (i.e., 5-iodouracil and 5-nitrouracil) to form mixed-ligand metal complexes (i.e., copper [Cu], manganese [Mn], and nickel [Ni] complexes) [ 12 , 14 , 56 ]. Furthermore, 8AQ derivatives displayed antimalarial, anticancer, and antioxidant activities [ 15 , 57 – 59 ]. These previously reported works supported our in vitro results, which suggested that some of the synthesized 8AQ-sulfonamide hybrids could be potentially further developed as antioxidant and antimicrobial agents. The biological activities, including antioxidant and antimicrobial activities, of the 8AQ derivatives have been noted to associate with potential mechanisms such as redox cycling, oxidative stress induction, and metal chelation. Notably, its metal-chelating ability causes the disruption of the key biochemical components within the microorganisms, leading to the generation of reactive oxygen species or the sequestration of essential metal cofactors, ultimately inhibiting microbial growth [ 12 ]. QSAR and QSPR models QSAR and QSPR models are computational methods commonly used to reveal the relationships between structural information and bioactivity values of the studied compounds [ 50 ]. Among others, the MLR algorithm is practically used in the drug design area due to its interpretable nature [ 60 ]. Many QSAR models have been reported for elucidating structure–activity relationships of diverse bioactive compounds toward several bioactivities such as anticancer, antioxidant, and antimicrobial activities [ 61 – 63 ]. Feature selection was performed to select only a set of informative descriptors to be included in the final datasets for model construction. Definitions of the selected descriptors selected for construction of DPPH (4 descriptors: ATS5s, GATS1e, Mor04p, and Mor24u), SOD (4 descriptors: R1v, AATS8p, B08[C–O], and D21), and antimicrobial (2 descriptors: X4sol and VR2_Dzi) models are provided in Table 3 , and their values are provided in Supplementary Materials (Tables S1 to S3 ). Additionally, the intercorrelation matrix calculated by Pearson’s correlation coefficient (cutoff values of | r | > 0.9) indicated that these descriptors are independent predictors (Supplementary Materials, Tables S4 to S6 ). Table 3. Definition of significant descriptors for QSAR modeling Activity Symbol Description Class Software DPPH (%) ATS5s Broto–Moreau autocorrelation of lag 5 (log function) weighted by I-state 2D autocorrelations PaDEL GATS1e Geary autocorrelation of lag 1 weighted by Sanderson electronegativity 2D autocorrelations Dragon Mor04p Signal 04 / weighted by polarizability 3D-MoRSE descriptors Dragon Mor24u 3D-MoRSE - signal 24 / unweighted 3D-MoRSE descriptors Dragon SOD (pIC 50 ) R1v R autocorrelation of lag 1 / weighted by van der Waals volume GETAWAY descriptors Dragon AATS8p Average Broto–Moreau autocorrelation - lag 8 / weighted by polarizabilities Autocorrelation descriptors PaDEL B08[C–O] Presence/absence of C–O at topological distance 8 2D atom pairs Dragon D211 Average vertex connectivity order-5 index 2D descriptor Mold 2 Antimicrobial (MIC) X4sol Solvation connectivity index of order 4 Connectivity indices Dragon VR2_Dzi Normalized Randic-like eigenvector-based index from Barysz matrix / weighted by first ionization potential Barysz matrix descriptor PaDEL Open in a new tab SOD, superoxide dismutase; DPPH, 2,2-diphenyl-1-picrylhydrazyl; MIC, minimum inhibitory concentration; QSAR, quantitative structure–activity relationship; 2D, 2-dimensional; 3D-MoRSE, 3D-molecule representation of structures based on electron diffraction; GETAWAY, geometry, topology, and atom-weights assembly; pIC 50 , the negative base 10 logarithm of the SOD IC 50 (μM) Two QSAR models (DPPH and SOD models) were constructed by MLR algorithms using Weka software, version 3.4.5 (Table 4 ). The models utilized a training set to generate the QSAR models, which were subsequently validated using LOO-CV and 5-fold-CV sets. It was revealed that electronegativity (GATS1e), signal 24 / unweighted (Mor24u), polarizability (Mor04p), and I-state (ATS5s) were noted as key properties influencing DPPH activity, whereas connectivity (D211), van der Waals volume (R1v), polarizability (AATS8p), and the presence/absence of carbon–oxygen (B08[C–O]) were noted for SOD activity (Tables 3 and 4 ). Both constructed models provided preferable predictive performance (Table 4 ) as indicated by high correlation values above 0.99 for training ( R 2 : 0.9980 to 0.9996) and LOO-CV ( Q 2 : 0.9930 to 0.9978) as well as low RMSE values ( RMSE Tr : 0.0121 to 0.1677 and RMSE LOO-CV : 0.0233 to 0.3830). In addition, 5-fold-CV demonstrated robust statistical performance (Table 4 ) as shown by high correlation coefficients and low RMSE values for both training ( R 2 = 0.9980 to 0.9996 and RMSE Tr = 0.0121 to 0.1677) and validation ( Q 2 5-fold-CV = 0.9936 to 0.9978 and RMSE 5-fold-CV = 0.0220 to 0.3997) sets. These statistical values indicated the acceptable reliability of the models ( R 2 > 0.6 and Q 2 > 0.5) [ 24 , 50 ]. To further ensure the reliability and predictive accuracy of the built models, 3 additional statistical parameters were assessed, including R 2 adj , MAE, and CCC. The DPPH model showed the following results: training set with R 2 adj = 0.9993 and MAE = 0.1384; LOO-CV set with R 2 adj = 0.9963, MAE = 0.3147, and CCC = 0.9988; and 5-fold-CV set with R 2 adj = 0.9963, MAE = 0.3171, and CCC = 0.9987. Similarly, the SOD (pIC 50 ) model displayed the following results: training set of R 2 adj = 0.9964 and MAE = 0.0099; LOO-CV set of R 2 adj = 0.9874, MAE = 0.0203, and CCC = 0.9964; and 5-fold-CV set of R 2 adj = 0.9885, MAE = 0.0194, and CCC = 0.9968. The calculated parameters of both built models (i.e., DPPH and SOD) showed preferable high R 2 adj and CCC values along with low MAE values for all evaluated sets (i.e., training, LOO-CV, and 5-fold-CV sets), indicating their preferably high predictive performance and robustness (threshold of MAE < 0.6 and CCC > 0.85). The results of the predicted values and comparative plots of experimental versus predicted activities for LOO-CV and 5-fold-CV sets of both models are provided in Supplementary Materials (DPPH model: Table S1 and Fig. 2 A and B, and SOD model: Table S22 and Fig. 2 C and D). The predictive reliability of the models was also demonstrated by the closeness of the dots displayed in the plots of experimental versus predicted activities (Fig. 2 A to D). This was supported by the low calculated residual values (Tables S1 and S2 ), which is a difference between experimental and predicted values, obtained from LOO-CV and 5-fold-CV sets of both models. Plots of experimental activities versus residual values of both DPPH (Fig. S23 A and B) and SOD (Fig. S23 C and D) models using LOO-CV and 5-fold-CV sets revealed that most of the dots are distributed near the zero axis, indicating that the models are well performed to provide close values to those of actual ones. Table 4. QSAR equations and statistical parameters indicating predictive performance of the antioxidant models (DPPH and SOD) Activity Equation N Training set LOO-CV set 5-Fold-CV set R 2 Tr RMSE Tr Q 2 LOO-CV RMSE LOO-CV Q 2 5-fold-CV RMSE 5-fold-CV DPPH %DPPH = 0.0685(ATS5s) − 74.5372(GATS1e) − 6.0133(Mor04p) − 12.0745(Mor24u) + 20.4924 11 0.9996 0.1677 0.9978 0.3830 0.9978 0.3997 SOD pIC 50 = 7.8664(R1v) − 2.2827(AATS8p) − 0.1794(B08[C–O]) + 22.3828(D211) − 3.3978 10 0.9980 0.0121 0.9930 0.0233 0.9936 0.0220 Open in a new tab SOD, superoxide dismutase; DPPH, 2,2-diphenyl-1-picrylhydrazyl; QSAR, quantitative structure–activity relationship; LOO-CV, leave-one-out cross-validation; 5-fold-CV, 5-fold cross-validation; R 2 Tr , squared correlation coefficient; RMSE Tr , root mean square error for training; Q 2 LOO-CV , cross-validated R 2 for LOO-CV; RMSE LOO-CV , root mean square error for LOO-CV; Q 2 5-fold-CV , cross-validated R 2 for 5-fold-CV; RMSE 5-fold-CV , root mean square error for 5-fold-CV; pIC 50 , the negative base 10 logarithm of the SOD IC 50 (μM) Fig. 2. Open in a new tab Plots of experimental and predicted activities of the 2,2-diphenyl-1-picrylhydrazyl (DPPH) model for (A) LOO-CV and (B) 5-fold-CV sets and the superoxide dismutase (SOD) model for (C) LOO-CV and (D) 5-fold-CV sets. Training set is represented by black squares and solid lines, while the LOO-CV and 5-fold-CV sets are represented by white squares and dotted lines. Furthermore, Y-randomization was performed to verify that the obtained QSAR models were not generated by chance correlation. The dependent variables, DPPH (%) and SOD (pIC₅₀), were randomly shuffled, and new models were generated using the same MLR method. This procedure was repeated 10 times. The R 2 and Q 2 values of the randomized models were then compared with those of the original models to assess model reliability. All randomized models exhibited lower R 2 and Q 2 values than the original models, indicating good model robustness and confirming the absence of chance correlation (Supplementary Materials, Fig. S24 ). In addition, the applicability domain (AD) of the QSAR models (DPPH and SOD) was evaluated using the leverage method and the Williams plot, which combine descriptor-space leverage values with standardized residuals to detect potential structural outliers (high-leverage compounds) and response outliers (large residual deviations). The critical leverage thresholds were calculated as h* = 1.36 for the DPPH model and h* = 1.50 for the SOD model. Compounds exhibiting leverage values h < h* and standardized residuals within the ±3σ interval were considered to fall within the AD. Based on these criteria, all compounds were found within the AD for both models, confirming the robustness and predictive reliability of the QSAR models (Supplementary Materials, Figs. S25 and S26 ). Owing to experimental antimicrobial results, the individual tested compounds showed antimicrobial activity in a nonuniform pattern, in which some of them exhibited inhibitory effects against the particular microorganisms. Due to the characteristic of the experimentally obtained data, the classification QSPR model was constructed to elucidate the influencing key structural features (independent variables, X n ) governing the antimicrobial classes (dependent variable, Y : active or inactive) of the compounds. Accordingly, an antimicrobial QSPR model was generated using the binary data as the input. Compounds displaying antimicrobial MIC values were assigned as actives, whereas those with no MIC values were assigned as inactives. The QSPR model was built using the decision tree analysis to allow the revealing of an interpretable if-then rule used to classify the compounds. The decision tree is a supervised machine learning algorithm successfully used to generate if-then rules to classify various classes of bioactive compounds (i.e., antimicrobial agents, aromatase inhibitors, and neuraminidase inhibitors) [ 64 – 66 ]. Herein, the decision tree algorithm, namely, J48, implemented in Weka software, version 3.4.5 was used to construct the QSPR model to generate the classification rule representing the descriptor-based cutoffs for categorizing the active and inactive compounds (Fig. 3 ). Similar to the antioxidant QSAR modeling, the antimicrobial QSPR model was validated using LOO-CV and 5-fold-CV sets. The model was generated using the training set (Fig. 3 ). It revealed an if-then rule that used 2 decision nodes (X4sol and VR2_Dzi) to identify 3 leaf nodes as antimicrobial classes of the compounds (active or inactive) in a top–down manner. The solvation connectivity index (X4sol: cutoff value of >7.564 or ≤7.564) was used as a root node for initial classification, followed by ionization potential (VR2_Dzi: cutoff value of >11.729 or ≤11.729) as an internal node. It was found that 5 compounds ( 3 , 4 , 5 , 6 , and 8 ) whose X4sol value are ≤7.564 were initially categorized as actives. The rest of compounds ( 7 and 9 - 13 ) with X4sol > 7.564 were sequentially categorized using their ionization potential (VR2_Dzi) as inactives ( 9 , 11 , and 13 with VR2_Dzi ≤11.729) and actives ( 10 and 12 with VR2_Dzi > 11.729). However, an inactive compound 7 (VR2_Dzi > 11.729) was misclassified as an active. Statistical parameters indicated that the constructed QSPR model displayed preferable predictive performance for training and both validation sets as indicated by high values of accuracy (training set = 90.91, LOO-CV set and 5-fold-CV set = 63.64), precision (training set = 0.920, LOO-CV set and 5-fold-CV set = 0.691), recall (training set = 0.909, LOO-CV set and 5-fold-CV set = 0.636), and F-measure (training set = 0.906, LOO-CV set and 5-fold-CV set = 0.642) (Supplementary Materials, Table S7 ). However, it was observed that the antimicrobial classification model shows a training accuracy of 90.91% and a cross-validated accuracy of 63.64%, indicating a notable drop in performance. The confusion matrices for classifying active and inactive compounds of training, LOO-CV, and 5-fold-CV sets are provided in Supplementary Materials (Table S8 ). The distribution of antimicrobial classes (as active and inactive) of the tested compounds ( 3 - 13 ) based on values of their 2 key descriptors (Fig. S27 A) also suggested that these descriptor space can effectively classify the compounds. Fig. 3. Open in a new tab Antimicrobial decision tree classification model of 8-aminoquinoline (8AQ)-based sulfonamides (3-13). The model was constructed using the training set. Decision nodes represent key descriptors used to classify compounds as active or inactive, whereas leaf nodes indicate their antimicrobial classes. The numbers in parentheses in each terminal leaf node indicate the number of correctly classified compounds, followed by the number of incorrectly classified compounds. Using the internal threshold of presenting the MIC values to classify the compound as active, there is a concerning issue regarding the interpretation of the antimicrobial activity of the original dataset used to construct the model. Compound 12 with a high MIC value (765.50 μM or 256 μg/ml) was classified as active, which is considerably weak potency from a drug discovery perspective. Therefore, it is suggested that the built QSPR model is applicable for initial screening, and further experimental investigations are highly required to ensure antimicrobial activities of the compounds. Although all constructed models (i.e., 2 QSAR antioxidant models and 1 QSPR antimicrobial model) were successfully constructed with acceptable predictive performances, some concerning issues should be noted as their limitations. Furthermore, the datasets used for QSAR/QSPR modeling were considered small-sized datasets because of the limited numbers of experimentally synthesized compounds for generating predictive models (i.e., DPPH = 11 compounds, SOD = 10 compounds, and antimicrobial = 11 compounds). Although the issue of the small-sized datasets is generally concerned for several limitations, it has been demonstrated that the small-sized datasets can be practically used to generate QSAR/QSPR models affording acceptable predictive performance, accuracy, and reliability [ 17 , 63 , 67 ]. Herein, the main aim of the QSAR modeling is to reveal a set of key structural properties that govern preferable activity of the compounds that are useful for guiding the effective rational design of the new 8AQ derivatives. The activities of the virtually designed compounds can also be predicted for facilitating the selection of potential compounds for further synthesis and validation. To avoid the overfitting issue, we performed the comparative analysis of statistical parameters using the k-fold cross-validation (CV) and LOO-CV testing sets to assess the predictive performance of constructed models. However, the external test sets are not available for further validating the model’s predictive performance outside the training datasets. Therefore, the reliability and accuracy of the constructed QSAR/QSPR models are not perfectly established. In addition, the 3 constructed models are restricted for their generalizability due to the limited size of the datasets. Another concerning issue was noted regarding the considerably weak DPPH activity of these 8AQ-sulfonamide compounds. The compounds displayed DPPH activity with less than 50%, and their DPPH IC 50 values were not obtained. The use of percentage values, with a narrow activity window, for modeling may limit statistical reliability, which should be noted as a limitation of the constructed DPPH model. However, the constructed DPPH model aimed for its application in guiding the rational design to obtain the modified compounds with improved radical scavenging potency. The knowledge regarding the structure–activity relationships would be beneficial for the future design of the new 8AQ-based compounds. Regarding the unavoidable narrow activity window of the dataset, the constructed DPPH QSAR model should be noted as a hypothesis-generating model for preliminary prediction rather than predictive in a rigorous quantitative sense. Furthermore, the notable drop in accuracy of the cross-validated antimicrobial QSPR model (63.64%) compared to its training (90.91%) suggested possible overfitting, which could be due to the small-sized dataset and imbalanced dataset that further limits the interpretability of classification accuracy. Accordingly, the constructed antimicrobial QSPR model is suitable for preliminary classifying the compounds as active or inactive, and further investigations are highly encouraged to confirm their definite antimicrobial properties. Rational design The facilitating role of the QSAR models for guiding the design of new analogs has been demonstrated for various classes of bioactive compounds (i.e., anticancer [ 27 , 67 ], antioxidant [ 17 , 68 ], antimicrobial [ 69 , 70 ], and aromatase-inhibitory [ 71 ] agents). Herein, the constructed QSAR/QSPR models were employed for guiding the design of new derivatives based on the key descriptors presented in the models (Table 4 and Fig. 3 ). Structural modifications were performed by substitutions of electron-withdrawing and electron-donating groups on the quinoline–sulfonamide core of the prototypes to give an additional set of 84 newly designed compounds (i.e., series 3 , 4 , 5 , 6 , 7 , 8 , 9 , 11 , 12 , and 13 ) (Table S9 and Figs. S28 to S37 ). All newly designed compounds were preprocessed to obtain values of key descriptors (Tables S10 to S12 ) for further predicting their antioxidant activities using the constructed QSAR models (i.e., DPPH and SOD; Table 4 ) or their antimicrobial classes (active or inactive) using the constructed QSPR decision tree model (Fig. 3 ). Predicted antioxidant (i.e., DPPH and SOD) activities and predicted antimicrobial classes of the newly designed compounds are provided in Tables S9 and S12 , respectively. The predictions of antioxidant activities indicated that the new analogs performed better predicted activities than their prototypes (i.e., 32 compounds with higher %DPPH and 31 compounds with higher SOD pIC 50 values; Table S13 ). These suggested that the structural modification strategies guided by the QSAR models could facilitate efficacious rational design to obtain new compounds with improved activities. For antimicrobial class prediction, it was found that 34 compounds were classified as actives, whereas 50 compounds were identified as inactives (Tables S12 and S13 ). The distribution pattern of 84 newly designed compounds in descriptor space indicated that the model can effectively classify the antimicrobial class of the compounds (into active and inactive clusters) based on these 2 key descriptors (X4sol and VR2_Dzi), as shown in Fig. S27 B. Although the reliability of the constructed QSAR/QSPR models used for guiding the design of these new analogs was acceptably ensured, the predicted activities/classes of these virtually designed compounds should be noted as only hypotheses that required further experimental investigations (i.e., synthesis and testing) for validating their potential prior to further development. Structure–activity relationship Experimental antioxidant findings revealed that the tested 8AQ-sulfonamides ( 3 - 13 ) exhibited more potent SOD activity than the DPPH activity. The result showed that the naphthalene-bearing compound 12 displayed the most potent SOD activity (IC 50 = 83.34 μM), followed by compounds 8 (X = CN, IC 50 = 96.07 μM) > 3 (X = F, IC 50 = 153.56 μM) > 5 (X = Br, IC 50 = 212.58 μM) > 6 (X = CF 3 , IC 50 = 247.39 μM). Among others, compound 12 is the only one bearing a naphthalene group. This suggested that the replacement of the sulfonyl benzene ring with the sulfonyl naphthalene ring could enhance the SOD activity of the compounds. When the sulfonyl benzene ring was maintained, the substitution with electron-withdrawing groups such as CN (compound 8 ) and the mono-halogen atom (compounds 3 and 5 with X = F and Br, respectively) was suggested for preferable SOD activity. In contrast, substitutions on the benzene ring with the NO 2 group ( o -NO 2 compound 9 with IC 50 = 254.58 μM and p -NO 2 compound 10 with IC 50 = 448.75 μM), the tetra-CH 3 group (compound 11 , IC 50 = 530.93 μM), and SO 2 CH 3 (compound 13 , IC 50 = 600.81 μM) gave the compounds with considerably weaker SOD activities. These findings were supported by information obtained from the constructed SOD QSAR model (Table 4 ), in which the connectivity index descriptor (D211) was noted as the most significant predictor, followed by van der Waals (R1v), polarizabilities (AATS8p), and the presence/absence of C–O (B08[C–O]), respectively (regression coefficient values: D211 = 22.3828, R1v = 7.8664, AATS8p = −2.2827, and B08[C–O] = −0.1794; Table 4 ). The model indicated that high positive values of D211 and R1v descriptors are required to obtain potent activity. Considering the descriptor profiles of the tested compounds, the most potent SOD-mimic compound 12 bearing naphthalene possessed high D211 (0.074) and the highest R1v (1.188) values. A similar descriptor profile was observed for the second most potent compound 8 bearing CN group (D211 = 0.075 and R1v = 1.164) (Table S2 ). In contrast, the least potent compound 13 (X = SO 2 CH 3 , pIC 50 = 3.221) displayed the lowest R1v (1.060) and low D211 (0.074) along with high B08[C–O] (1) values (Table S2 ). The influences of the polarizability descriptor (AATS8p) were noticed when comparing the mono-halogen-substituted compounds ( 3 , 4 , and 5 ). Among these compounds, compound 3 (X = F) displayed the highest SOD activity (pIC 50 : 3 [3.814] > 5 [3.672] > 4 = inactive), which could be due to its lowest AATS8p value (AATS8p: 3 = 1.419, 5 = 1.622, 4 = 1.551) (Table S2 ). The isomeric effect of NO 2 substitution on the benzene ring was observed, in which the o- NO 2 substitution could give the compound 9 (X = o- NO 2 , IC 50 = 254.58 μM) with better SOD activity when compared to that of the compound 10 bearing the p- NO 2 (X = p- NO 2 , IC 50 = 448.75 μM). This could be due to the position of the substituted NO 2 group that affected the values of 2 main descriptors (D211: 9 = 0.077 > 10 = 0.074 and R1v: 9 = 1.102 > 10 = 1.083; Table S2 ). Additionally, it was observed that the replacement of mono-F of compound 3 (pIC 50 = 3.814) with trifluoromethyl group to give compound 6 (pIC 50 = 3.607) could affect the values of D211 and R1v descriptors, leading to decreased SOD activity ( 3 : D211 = 0.076 and R1v = 1.113, 6: D211 = 0.074 and R1v = 1.085) (Table S2 ). Five most promising newly designed compounds (Fig. 4 ) with the highest predicted SOD activity were listed as 13c (pIC 50 = 4.793) > 13f (pIC 50 = 4.554) > 8c (pIC 50 = 4.355) > 12e (pIC 50 = 4.198) > 11c (pIC 50 = 4.187) (Table S9 ). It should be noted that a replacement of sulfone (SO 2 CH 3 ) group on prototype 13 (experimental pIC 50 = 3.221, R1v = 1.060, and D211 = 0.074; Table S2 ) with a sulfide-linked benzene ring could considerably improve the activities of the compounds as observed for 13c (predicted pIC 50 = 4.793, X = p- SC 6 H 5 ) and 13f (predicted pIC 50 = 4.554, X = m- SC 6 H 5 ). The enhancing SOD activity of these 2 derivatives ( 13c and 13f ) could be due to the increasing values of 2 influential descriptors including R1v ( 13c = 1.305 and 13f = 1.262) and D211 ( 13c = 0.077 and 13f = 0.078) (Table S11 ). Substitution with another o -CN group on the core of p -CN prototype 8 (experimental pIC 50 = 4.017) to give o -, p -di-CN compound 8c with improved SOD activity (predicted pIC 50 = 4.355) (Table S9 ). Similarly, a substitution with OCH 3 on the naphthalene ring of prototype 12 (experimental pIC 50 = 4.079) could give compound 12e with more potent activity (predicted pIC 50 = 4.198). Fig. 4. Open in a new tab Summary of top 5 newly designed compounds with the highest predicted superoxide dismutase (SOD) activity. Experimental DPPH assay indicated that none of the tested 8AQ-sulfonamides displayed DPPH activity greater than 50%. Compound 9 (X = o- NO 2 ) displayed the highest DPPH activity (%DPPH = 36.49%), followed by compound 11 (X = tetra-CH 3 , %DPPH = 22.43%), compound 10 (X = p- NO 2 , %DPPH = 19.06%), and compound 13 (X = SO 2 CH 3 , %DPPH = 17.33%) (Table 1 ). It was noticed that the effects of substitutions on the DPPH activity of this set of 8AQ-sulfonamides are conversely observed from those affecting the SOD activity. Compounds that exhibited preferable SOD IC 50 values (i.e., compounds 12 , 8 , 3 , and 5 ) were ranked as the least potent DPPH compounds, displaying considerably low %DPPH ( 12 = 13.88%, 8 = 12.53%, 3 = 10.09%, and 5 = 7.74%; Table 1 ). From the DPPH QSAR model (Table 4 ), the electronegativity descriptor GATS1e was noted as the most influential predictor, as indicated by its highest regression coefficient value (−74.5372), followed by Mor24u (−12.0745), Mor04p (−6.0133), and ATS5s (0.0685), respectively. To obtain a high %DPPH value, a compound needs low positive GATS1e but high positive ATS5s values. Considering the descriptor profile of the most potent DPPH compound 9 (X = o- NO 2 , %DPPH = 36.49%), the compound displayed the lowest positive GATS1e (0.345) but the highest positive ATS5s (560.981) values, which were aligned with its highest potent experimental DPPH activity. In contrast, the least potent compound 5 (%DPPH = 7.74%) exhibited high positive GATS1e (0.457) but the lowest positive ATS5s (266.058) values (Table S1 ), supporting its low experimentally observed DPPH activity. Moreover, the enhancing effect of the o- NO 2 substitution on the benzene ring ( 9 : X = o- NO 2 , %DPPH = 36.49%; Table 1 ) on DPPH activity compared to the p- NO 2 substitution ( 10 : X = p- NO 2 , %DPPH = 19.06%; Table 1 ) was noticed similarly to that found for the SOD activity. The descriptor profiles of these 2 compounds ( 9 and 10 ) showed that the position of NO 2 substitution affected values of all key descriptors, except for the GATS1e (Table S1 ). The o- NO 2 substitution provided the better descriptor profile (i.e., higher positive ATS5s and Mor24u values and higher negative Mor04p value) that supports the achievement of the higher %DPPH when compared to compound 10 with the o- NO 2 substitution (Table S1 ). The top 5 structurally modified compounds (Fig. 5 ) with the most potent predicted %DPPH were ranked as 9b (40.76%) > 6f (37.22%) > 9c (27.57%) > 11h (25.29%) > 3d (23.21%), as shown in Table S9 . The high predicted %DPPH of compounds 9b (X = o- and p- NO 2 groups) and 9c (X = di m- NO 2 groups) suggested that the presence of di-NO 2 in the molecule is essential for radical scavenging effect of the compounds, and the substituted position could affect DPPH activity of the compounds. The effect of the substitution position was also noted for compound 6f (X = o -CF 3 ), which displayed improved predicted activity compared to its parent compound 6 (X = p -CF 3 , experimental %DPPH = 9.79%; Table 1 ). It was also noted that an addition of the OH group on the 8AQ ring to give compound 11h could improve %DPPH ( 11 : experimental %DPPH = 22.43%; Table 1 ). Although the improved DPPH activities of the modified compounds were achieved, it should be noted that the predicted activities of these newly designed compounds are still inferior to desirable effects (as indicated by predicted %DPPH less than 50%). Fig. 5. Open in a new tab Summary of top 5 newly designed compounds with the highest predicted 2,2-diphenyl-1-picrylhydrazyl (DPPH) activity. The constructed QSPR decision tree model (Fig. 3 ), performed well to correctly identify antimicrobial classes of 10 out of 11 compounds in the training set and 7 out of 11 compounds in the LOO-CV and 5-fold-CV sets, suggested that the antimicrobial activity of the studied compounds is classified orderly based on 2 main properties (i.e., solvation connectivity index [X4sol] and ionization potential [VR2_Dzi]). These findings suggested that different types of functional groups substituted on the 8AQ-sulfonamide core could affect the key properties (i.e., solvation and ionization potential) that govern the antimicrobial effects of the compounds. Thirty-four structurally modified compounds exhibiting antimicrobial activity are presented in Supplementary Materials (Fig. S38 ). From the newly designed compounds, it was suggested that the types and number of the halogen atoms (i.e., F, Cl, Br, and I) substituted on the core of the prototypes affected the antimicrobial activities of their modified analogs. The enhancing effects of fluorine substitutions were observed for modified compounds of series 3 , in which the compounds bearing at least 1 F atom (i.e., mono-F compounds 3a - 3b or di-F compounds 3c - 3d ) were predicted to be active antimicrobial agents. In contrast, the presence of at least a Cl, I, or Br atom seemed to diminish the antimicrobial effects as shown by compounds in series 4 and 5 , except for the di-Cl-F compounds (i.e., 4g and 4h with p- Cl substitution on the sulfonyl benzene ring) (Table S12 and Figs. S28 to S30 and S38 ). This may imply that halogen substitutions play crucial roles in modulating antimicrobial effects of the compounds. Previous related works have reported antioxidant and antimicrobial activities of the 8AQ-based compounds, and their redox potential was noted to be one of the influential key properties linked to their biological activities [ 72 ]. In overview, structural modifications with several types of functional moieties could give new compounds with improved or worsen predicted activities. This could be due to electron-donating and electron-withdrawing effects of the substituents that interfere redox behavior of the compounds. Conclusion The discovery of new classes of antioxidant and antimicrobial agents has gained attention to address current health issues. 8AQ is a pharmacophore well recognized for its antimalarial effect, making it a promising core for the design of hybrid compounds for therapeutic applications. Herein, a set of 11 newly synthesized 8AQ-based sulfonamides ( 3 - 13 ) were experimentally investigated for their antioxidant and antimicrobial activities. These compounds exhibited notable SOD-mimic activity but weak DPPH radical scavenging activity. The naphthalene-bearing compound 12 was noted as the most potent antioxidant agent. Antimicrobial investigation indicated that only 7 compounds (i.e., halogen-containing compounds [ 3 , 4 , 5 , and 6 ], nitrile compound [ 8 ], p- NO 2 compound 10 , and naphthalene-containing compound 12 ) were active antimicrobial agents. Particularly, some of them are active anti-MRSA agents. Two antioxidant QSAR (i.e., SOD and DPPH) models and 1 antimicrobial QSPR model were successfully constructed and validated to ensure their preferable predictive performances. The constructed models were further used to guide the design and predict antioxidant activities or antimicrobial classes of an additional set of 84 newly designed compounds, in which their antioxidant activities or antimicrobial class were predicted using the constructed models. Most of the newly designed compounds display better predicted antioxidant activities when compared to their prototypes, suggesting the effectiveness of the constructed models in guiding the design. A series of 5 compounds with the most preferable predicted antioxidant activities (SOD: 13c , 13f , 8c , 12e , and 11c ; DPPH: 9b , 6f , 9c , 11h , and 3d ) and 34 compounds with active predicted antimicrobial class were suggested to be further developed. Substitutions on the 8AQ-sulfonamide core with naphthalene, halogen atoms (i.e., F and Cl), NO 2 , CN, and SC 6 H 5 groups were revealed to improve activities of the modified compounds. In summary, the study demonstrates the role of computational modeling in guiding the efficacious design of new 8AQ-based sulfonamides. To validate the predictions, these highlighted newly designed compounds should be further synthesized and experimentally investigated for their activities. Further studies (i.e., in vitro and in vivo) regarding their pharmacokinetics (absorption, distribution, metabolism, and excretion), pharmacodynamics (molecular docking), toxicity profiles, and mechanism of action are highly encouraged for successful development. Acknowledgments We would like to thank Dr. Nuttapat Anuwongcharoen for conducting the Y-randomization and applicability domain analyses. We would like to acknowledge Chulabhorn Research Institute for recording the mass spectrometry data. Funding: This work is supported by the annual budget grant from Mahidol University (B.E. 2562-2563). R.P. is supported by the Srinakharinwirot University and National Science, Research and Innovation Fund (NSRF) (grant no. 041/2569). Author contributions: R.P.: Conceptualization, methodology, investigation, validation, formal analysis, resources, writing—review and editing, and funding acquisition. A.W.: Conceptualization, methodology, investigation, validation, visualization, formal analysis, resources, writing—review and editing, writing—original draft, and funding acquisition. Veda P.: Writing—review and editing, visualization, methodology, formal analysis, and data curation. R.C.: Investigation and formal analysis. S.P.: Writing—review and editing and supervision. S.R.: Resources and supervision. Virapong P.: Writing—review and editing and supervision. Competing interests: The authors declare that they have no competing interests. Data Availability The data supporting the findings of this study are available from the corresponding authors upon request. Supplementary Materials Supplementary 1 Figs. S1 to S38 Tables S1 to S13 csbj.0032.f1.docx (11.9MB, docx) References 1. Pooja G, Shweta S, Patel P. Oxidative stress and free radicals in disease pathogenesis: A review. Discov Med. 2025;2(1):104. [ Google Scholar ] 2. Chandimali N, Bak SG, Park EH, Lim H-J, Won Y-S, Kim E-K, Park SI, Lee SJ. Free radicals and their impact on health and antioxidant defenses: A review. Cell Death Discov. 2025;11(1):19. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 3. Al-Madhagi H, Masoud A. Limitations and challenges of antioxidant therapy. Phytother Res. 2024;38(2):5549–5566. [ DOI ] [ PubMed ] [ Google Scholar ] 4. Gulcin İ. Antioxidants: A comprehensive review. Arch Toxicol. 2025;99(5):1893–1997. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 5. Munita JM, Arias CA. Mechanisms of antibiotic resistance. Microbiol Spectrum. 2016;4(2):10.1128/microbiolspec.VMBF-0016-2015. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 6. Breijyeh Z, Jubeh B, Karaman R. Resistance of gram-negative bacteria to current antibacterial agents and approaches to resolve it. Molecules. 2020;25(6):1340. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 7. Loomba PS, Taneja J, Mishra B. Methicillin and vancomycin resistant S. aureus in hospitalized patients. J Glob Infect Dis. 2010;2(3):275–283. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 8. Van Bambeke F, Reinert RR, Appelbaum PC, Tulkens PM, Peetermans WE. Multidrug-resistant Streptococcus pneumoniae infections: Current and future therapeutic options. Drugs. 2007;67(16):2355–2382. [ DOI ] [ PubMed ] [ Google Scholar ] 9. GBD 2021 Antimicrobial Resistance Collaborators. Global burden of bacterial antimicrobial resistance 1990-2021: A systematic analysis with forecasts to 2050. Lancet. 2024;404:1199–1226. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 10. Kabir E, Uzzaman M. A review on biological and medicinal impact of heterocyclic compounds. Results Chem. 2022;4: Article 100606. [ Google Scholar ] 11. Kerru N, Gummidi L, Maddila S, Gangu KK, Jonnalagadda SB. A review on recent advances in nitrogen-containing molecules and their biological applications. Molecules. 2020;25(8):1909. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 12. Phopin K, Sinthupoom N, Treeratanapiboon L, Kunwittaya S, Prachayasittikul S, Ruchirawat S, Prachayasittikul V. Antimalarial and antimicrobial activities of 8-aminoquinoline-uracils metal complexes. EXCLI J. 2016;15:144–152. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Hamerly T, Tweedell RE, Hritzo B, Nyasembe VO, Tekwani BL, Nanayakkara NPD, Walker LA, Dinglasan RR. NPC1161B, an 8-aminoquinoline analog, is metabolized in the mosquito and inhibits Plasmodium falciparum oocyst maturation. Front Pharmacol. 2019;10:1265. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 14. Pingaew R, Worachartcheewan A, Prachayasittikul V, Prachayasittikul S, Ruchirawat S, Prachayasittikul V. Transition metal complexes of 8-aminoquinoline-5-substituted uracils with antioxidative and cytotoxic activities. Lett Drug Des Discov. 2013;10(9):859–864. [ Google Scholar ] 15. Rudrapal M, Chetia D, Prakash A. Synthesis, antimalarial-, and antibacterial activity evaluation of some new 4-aminoquinoline derivatives. Med Chem Res. 2013;22(8):3703–3711. [ Google Scholar ] 16. Aslam AA, Ahmed M, Mughram MHA, Habib-Ur-Rahman Mahmood M, Basheer S, Hussain R, Eiman E, Sanaullah M, Raza H, Saeed A, et al. Sulfonamides as a promising scaffold in drug discovery: An insightful review on FDA-approved molecules, synthesis strategy, medical indication, and their binding mode. Chem Biodivers. 2025;22: Article e202403434. [ DOI ] [ PubMed ] [ Google Scholar ] 17. Worachartcheewan A, Pingaew R, Prachayasittikul V, Apiraksattayakul S, Prachayasittikul S, Ruchirawat S, Prachayasittikul V. Synthesis, biological investigation, and in silico studies of 2-aminothiazole sulfonamide derivatives as potential antioxidants. EXCLI J. 2025;24:60–81. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 18. Pingaew R, Prachayasittikul V, Worachartcheewan A, Thongnum A, Prachayasittikul S, Ruchirawat S, Prachayasittikul V. Anticancer activity and QSAR study of sulfur-containing thiourea and sulfonamide derivatives. Heliyon. 2022;8(8): Article e10067. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Özkan H, Demirci B. Synthesis and antimicrobial and antioxidant activities of sulfonamide derivatives containing tetrazole and oxadiazole rings. J Heterocyclic Chem. 2019;56(9):2528–2528. [ Google Scholar ] 20. Sena Murteira Pinheiro P, Franco LS, Montagnoli TL, Fraga CAM. Molecular hybridization: A powerful tool for multitarget drug discovery. Expert Opin Drug Discov. 2024;19:451–470. [ DOI ] [ PubMed ] [ Google Scholar ] 21. Saifi Z, Ali A, Inam A, Azam A, Kamthan M, Abid M, Abid M, Ali I. Synthesis and antibacterial evaluation of quinoline-sulfonamide hybrid compounds: A promising strategy against bacterial resistance. RSC Adv. 2025;15:1680–1689. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 22. Chetry AB, Ohto K. From molecules to data: The emerging impact of chemoinformatics in chemistry. J Cheminform. 2025;17(1):121. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 23. Sabe VT, Ntombela T, Jhamba LA, Maguire GEM, Govender T, Naicker T, Kruger HG. Current trends in computer aided drug design and a highlight of drugs discovered via computational techniques: A review. Eur J Med Che. 2021;224: Article 113705. [ DOI ] [ PubMed ] [ Google Scholar ] 24. Prachayasittikul V, Worachartcheewan A, Shoombuatong W, Songtawee N, Simeon S, Prachayasittikul V, Nantasenamat C. Computer-aided drug design of bioactive natural products. Curr Top Med Chem. 2015;15(18):1780–1800. [ DOI ] [ PubMed ] [ Google Scholar ] 25. Shah M, Patel M, Shah M, Patel M, Prajapati M. Computational transformation in drug discovery: A comprehensive study on molecular docking and quantitative structure activity relationship (QSAR). Intell Pharm. 2024;2(5):589–595. [ Google Scholar ] 26. Worachartcheewan A, Songtawee N, Siriwong S, Prachayasittikul S, Nantasenamat C, Prachayasittikul V. Rational design of colchicine derivatives as anti-HIV agents via QSAR and molecular docking. Med Chem. 2019;15(4):328–340. [ DOI ] [ PubMed ] [ Google Scholar ] 27. Prachayasittikul V, Pingaew R, Anuwongcharoen N, Worachartcheewan A, Nantasenamat C, Prachayasittikul S, Ruchirawat S, Prachayasittikul V. Discovery of novel 1,2,3-triazole derivatives as anticancer agents using QSAR and in silico structural modification. Springerplus. 2015;4(1):571. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 28. Worachartcheewan A, Prachayasittikul V, Prachayasittikul S, Tantivit V, Yeeyahya C, Prachayasittikul V. Rational design of novel coumarins: A potential trend for antioxidants in cosmetics. EXCLI J. 2020;19:209–226. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 29. Pratiwi R, Prachayasittikul V, Prachayasittikul S, Nantasenamat C. Rational design of novel sirtuin 1 activators via structure-activity insights from application of QSAR modeling. EXCLI J. 2019;18:207–222. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 30. Prachayasittikul V, Mandi P, Pingaew R, Prachayasittikul S, Ruchirawat S, Prachayasittikul V. Substituted 1,4-naphthoquinones for potential anticancer therapeutics: In vitro cytotoxic effects and QSAR-guided design of new analogs. Comput Struct Biotechnol J. 2025;27:3492–3509. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 31. Sen C, Sahoo T, Singh H, Suresh E, Ghosh SC. Visible light-promoted photocatalytic C-5 carboxylation of 8-aminoquinoline amides and sulfonamides via a single electron transfer pathway. J Org Chem. 2019;84(16):9869–9896. [ DOI ] [ PubMed ] [ Google Scholar ] 32. Rouffet M, Oliveira CA, Udi Y, Agrawal A, Sagi I, McCammon JA, Cohen SM. From sensors to silencers: Quinoline- and benzimidazole-sulfonamides as inhibitors for zinc proteases. J Am Chem Soc. 2010;132:8232–8233. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 33. Silva LE, Joussef AC, Pacheco LK, Silva DG, Steindel M, Rebelo RA, Schmidt B. Synthesis and in vitro evaluation of leishmanicidal and trypanocidal activities of N -quinolin-8-yl-arylsulfonamides. Bioorg Med Chem. 2007;15(24):7553–7560. [ DOI ] [ PubMed ] [ Google Scholar ] 34. Diaconu D, Mangalagiu V, Amariucai-Mantu D, Antoci V, Giuroiu CL, Mangalagiu II. Hybrid quinoline-sulfonamide complexes (M 2+ ) derivatives with antimicrobial activity. Molecules. 2020;25(12):2946. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 35. Makida Y, Ohmiya H, Sawamura M. Sulfonamidoquinoline/palladium(II)-dimer complex as a catalyst precursor for palladium-catalyzed γ-selective and stereospecific allyl–aryl coupling reaction between allylic acetates and arylboronic acids. Chem Asian J. 2011;6(2):410–414. [ DOI ] [ PubMed ] [ Google Scholar ] 36. Macías B, García I, Villa MV, Borrás J, Castiñeiras A, Sanz F. Synthesis and structural characterization of zinc complexes with sulfonamides containing 8-aminoquinoleine. Z Anorg Allg Chem. 2003;629(2):255–260. [ Google Scholar ] 37. Blois MS. Antioxidant determinations by the use of a stable free radical. Nature. 1958;181(4617):1199–1200. [ Google Scholar ] 38. Marklund S, Marklund G. Involvement of the superoxide anion radical in the autoxidation of pyrogallol and a convenient assay for superoxide dismutase. Eur J Biochem. 1974;47(3):469–474. [ DOI ] [ PubMed ] [ Google Scholar ] 39. Baron EJ, Peterson LR, Finegold SM. Methods for testing antimicrobial effectiveness. In: Bailey and Scott’s diagnostic microbiology , 9th ed. St. Louis: Mosby-Year Book, Inc.; 1994, p. 168–193. 40. CLSI, Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing; 23rd Informational supplement M100-S23. Wayne (PA): CLSI, 2013. 41. Dennington IIR, Keith T, Millam J, Eppinnett K, Hovell WL, Gilliland R, GaussView, version 3.09, Semichem, Shawnee Mission, KS, USA (2003). 42. Yap CW. PaDEL-descriptor: An open source software to calculate molecular descriptors and fingerprints. J Comput Chem. 2011;32(7):1466–1474. [ DOI ] [ PubMed ] [ Google Scholar ] 43. Hong H, Xie Q, Ge W, Qian F, Fang H, Shi L, et al. Mold 2 , molecular descriptors from 2D structures for chemoinformatics and toxicoinformatics. J Chem Inf Model. 2008;48(7):1337–1344. [ DOI ] [ PubMed ] [ Google Scholar ] 44. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, Scalmani G, Barone V, Mennucci B, Petersson GA, et al. Gaussian 09, Revision A.02, Gaussian Inc.: Wallingford CT. 2009. 45. Talete srl. DRAGON for Windows (software for molecular descriptor calculations), version 5.5. Milano, Italy (2007). http://www.talete.mi.it 46. Witten IH, Frank E, Hall MA. Data mining: practical machine learning tools and techniques . San Francisco (CA): Morgan Kaufmann; 2011. 47. Nantasenamat C, Naenna T, Isarankura Na Ayudhya C, Prachayasittikul V. Quantitative prediction of imprinting factor of molecularly imprinted polymers by artificial neural network. J Comput Aided Mol Des. 2005;19(7):509–524. [ DOI ] [ PubMed ] [ Google Scholar ] 48. Han L, Wang Y, Bryant SH. Developing and validating predictive decision tree models from mining chemical structural fingerprints and high-throughput screening data in PubChem. BMC Bioinform. 2008;9:401. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 49. Worachartcheewan A, Nantasenamat C, Isarankura-Na-Ayudhya C, Pidetcha P, Prachayasittikul V. Identification of metabolic syndrome using decision tree analysis. Diabetes Res Clin Pract. 2010;90(1):e15–e18. [ DOI ] [ PubMed ] [ Google Scholar ] 50. Nantasenamat C, Isarankura-Na-Ayudhya C, Prachayasittikul V. Advances in computational methods to predict the biological activity of compounds. Expert Opin Drug Discov. 2010;5(7):633–654. [ DOI ] [ PubMed ] [ Google Scholar ] 51. Nantasenamat C, Isarankura-Na-Ayudhya C, Naenna T, Prachayasittikul V. Prediction of bond dissociation enthalpy of antioxidant phenols by support vector machine. J Mol Graph Model. 2008;27(2):188–196. [ DOI ] [ PubMed ] [ Google Scholar ] 52. Yarahmadi B, Hashemianzadeh SM. Determining the quality of imprinted polymers using diverse feature selections methods, Ada Boost and Gradient boosting algorithms. Results Mater. 2025;27: Article 100722. [ Google Scholar ] 53. Chitre TS, Hirode PV, Lokwani DK, Bhatambrekar AL, Hajare SG, Thorat SB, Priya D, Pradhan KB, Asgaonkar KD, Jain SP. In-silico studies of 2-aminothiazole derivatives as anticancer agents by QSAR, molecular docking, MD simulation and MM-GBSA approaches. J Biomol Struct Dyn. 2024;42:11396–11414. [ DOI ] [ PubMed ] [ Google Scholar ] 54. Oliveira MD, Ferreira MS, Katchborian-Neto A, Almeida de Oliveira R, Baldim JL, Oliveira TB, Dias DF, Chagas-Paula DA, Soares MG. Machine learning decision tree-based models for predicting the antibacterial activity of Lamiaceae essential oils against Staphylococcus aureus . J Mol Graph Model. 2025;140: Article 109116. [ DOI ] [ PubMed ] [ Google Scholar ] 55. Faleye OS, Boya BR, Lee JH, Choi I, Lee J. Halogenated antimicrobial agents to combat drug-resistant pathogens. Pharmacol Rev. 2023;76(1):90–141. [ DOI ] [ PubMed ] [ Google Scholar ] 56. Ruankham W, Songtawee N, Prachayasittikul V, Worachartcheewan A, Suwanjang W, Pingaew R, Prachayasittikul V, Prachayasittikul S, Phopin K. Promising 8-aminoquinoline-based metal complexes in the modulation of SIRT1/3-FOXO3a axis against oxidative damage-induced preclinical neurons. ACS Omega. 2023;8(49):46977–46988. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 57. Bhat HR, Masih A, Shakya A, Ghosh SK, Singh UP. Design, synthesis, anticancer, antibacterial, and antifungal evaluation of 4-aminoquinoline-1,3,5-triazine derivatives. J Heterocyclic Chem. 2020;57:390–399. [ Google Scholar ] 58. Zhai B, Hao Q, Wang M, Luo Z, Yang R, Yang J, Cao Y. Discovery of new 4-aminoquinoline derivatives containing an amine or hydroxamic acid terminal as multifunctional agents for the treatment of Alzheimer′s disease. Bioorg Chem. 2024;153: Article 107954. [ DOI ] [ PubMed ] [ Google Scholar ] 59. Albayrak F, Cicek M, Alkaya D, Kulu I. Design, synthesis and biological evaluation of 8-aminoquinoline-1,2,3-triazole hybrid derivatives as potential antimicrobial agents. Med Chem Res. 2022;31(4):652–665. [ Google Scholar ] 60. Priya S, Tripathi G, Singh DB, Jain P, Kumar A. Machine learning approaches and their applications in drug discovery and design. Chem Biol Drug Des. 2022;100(3):136–153. [ DOI ] [ PubMed ] [ Google Scholar ] 61. Pingaew R, Prachayasittikul V, Worachartcheewan A, Nantasenamat C, Prachayasittikul S, Ruchirawat S, et al. Novel 1,4-naphthoquinone-based sulfonamides: Synthesis, QSAR, anticancer and antimalarial studies. Eur J Med Chem. 2015;103:446–459. [ DOI ] [ PubMed ] [ Google Scholar ] 62. Brahmbhatt H, Molnar M, Pavić V, Rastija V. Synthesis, characterization, antibacterial and antioxidant potency of N- substituted-2-sulfanylidene-1,3-thiazolidin-4-one derivatives and QSAR study. Med Chem. 2019;15(8):840–849. [ DOI ] [ PubMed ] [ Google Scholar ] 63. Cherdtrakulkiat R, Worachartcheewan A, Tantimavanich S, Lawung R, Sinthupoom N, Prachayasittikul S, Ruchirawat S, Prachayasittikul V. Discovery of novel halogenated 8-hydroxyquinoline-based anti-MRSA agents: In vitro and QSAR studies. Drug Dev Res. 2020;81(1):127–135. [ DOI ] [ PubMed ] [ Google Scholar ] 64. Lira F, Perez PS, Baranauskas JA, Nozawa SR. Prediction of antimicrobial activity of synthetic peptides by a decision tree model. Appl Environ Microbiol. 2013;79(10):3156–3159. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 65. Nantasenamat C, Li H, Mandi P, Worachartcheewan A, Monnor T, Isarankura-Na-Ayudhya C. Exploring the chemical space of aromatase inhibitors. Mol Divers. 2013;17(4):661–677. [ DOI ] [ PubMed ] [ Google Scholar ] 66. Anuwongcharoen N, Shoombuatong W, Tantimongcolwat T, Prachayasittikul V, Nantasenamat C. Exploring the chemical space of influenza neuraminidase inhibitors. PeerJ. 2016;4: Article e1958. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 67. Miladiyah I, Jumina J, Haryana MS, Mustofa M. Biological activity, quantitative structure–activity relationship analysis, and molecular docking of xanthone derivatives as anticancer drugs. Drug Des Devel Ther. 2018;12:149–158. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 68. Erzincan P, Saçan MT, Yüce-Dursun B, Danış Ö, Demir S, Erdem SS, et al. QSAR models for antioxidant activity of new coumarin derivatives. SAR QSAR Environ Res. 2015;26:721–737. [ DOI ] [ PubMed ] [ Google Scholar ] 69. Liu J, Zhou L, Zuo Z. Antibacterial activities of carbapenem derivatives and quantitative structure–activity relationship for drug design. QSAR Comb Sci. 2008;27(10):1216–1226. [ Google Scholar ] 70. Bruijn WJC, Hageman JA, Araya-Cloutier C, Gruppen H, Vincken JP. QSAR of 1,4-benzoxazin-3-one antimicrobials and their drug design perspectives. Bioorg Med Chem. 2018;26:6105–6114. [ DOI ] [ PubMed ] [ Google Scholar ] 71. Prachayasittikul V, Pingaew R, Worachartcheewan A, Sitthimonchai S, Nantasenamat C, Prachayasittikul S, et al. Aromatase inhibitory activity of 1,4-naphthoquinone derivatives and QSAR study. EXCLI J. 2017;16:714–726. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 72. Assary RS, Brushett FR, Curtiss LA. Reduction potential predictions of some aromatic nitrogen-containing molecules. RSC Adv. 2014;4(101):57442–57451. [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Supplementary 1 Figs. S1 to S38 Tables S1 to S13 csbj.0032.f1.docx (11.9MB, docx) Data Availability Statement The data supporting the findings of this study are available from the corresponding authors upon request. Articles from Computational and Structural Biotechnology Journal are provided here courtesy of AAAS Science Partner Journal Program ACTIONS View on publisher site PDF (2.6 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top