Al Amin Md, Hossain Md Sakhawat, Habib Md Kawsar, Perveen Faria Farzana, Sabuj Sahinur Islam, Imam Hasan, Islam Sirajul, Mahmud Shahin
Department of Biotechnology and Genetic Engineering, Mawlana Bhashani Science and Technology University, Santosh, Tangail-1902, Bangladesh.
Department of Pharmacy, Mawlana Bhashani Science and Technology University, Santosh, Tangail-1902, Bangladesh.
Comput Biol Med. 2025 Mar;186:109727. doi: 10.1016/j.compbiomed.2025.109727. Epub 2025 Jan 25.
The rapid rise of antibiotic-resistant bacterial strains is a significant global health issue, necessitating the development of new and effective antimicrobial agents. This study focuses on designing synthetic de novo models of fluoroquinolone (FQ) descriptors by fusing the quinolone ring-a derivative of FQs-with potential ATPase inhibitors, which is identified through pharmacophore modeling targeting the DNA gyrase B (gyrB) protein of S. aureus. Initially, the pharmacophore model was generated based on the DNA gyrB protein (PDB IDs: 3TTZ, 3U2D, and 3U2K), specifically targeting their co-crystalized ATPase inhibitors to develop a shared feature pharmacophore (SFP) with key features including hydrophobic regions, hydrogen bond acceptors (HBA), hydrogen bond donors (HBD), aromatic moieties (Ar), and halogen bond donors (XBD). The map was further evaluated using the goodness-of-hit (GH) score of 0.2641, indicating the map's strength in capturing potential compounds. The SFP was used for virtual screening against 160,000 compounds from ZINCpharmer and ChEMBL, resulting in 74 hits (48 from ZINCpharmer and 26 from ChEMBL) with similar ATPase features and exhibiting the best-fit scores ranging from 73.50 to 76.80 and RMSD values from 0.1 to 0.5. These identified ligands were fused with the quinolone ring of FQs using genetic algorithms and fragment-based design to create 50 new synthetic models of FQs. Most of these models contain Pyrrole rings, average similarity of more than 55 %, and a Synthetic Accessibility Score (SAScore) ≤ 3.5 for practical syntheses in the lab. These conformers were evaluated through pharmacokinetics and molecular docking, revealing three top compounds-Molecule 13 (-9.1 kcal/mol), Molecule 20 (-9.1 kcal/mol), and Molecule 49 (-9.4 kcal/mol)-which showed greater binding affinity with the DNA gyrase protein (PDB ID: 4PLB) compare to control, Ciprofloxacin (-7.8 kcal/mol). Additionally, 200 ns Molecular Dynamics Simulations (MDS) were conducted using the Schrödinger suite for these three compounds and control, where Molecule 13 showed potential structural stability. To ensure practical feasibility, computational sequence-to-sequence retrosynthesis, and chemical scaffold comparison analysis were employed to design synthetic routes for the most promising compounds, confirming their likelihood of successful synthesis in the lab. The findings of this study contribute to the ongoing efforts to treat antibiotic resistance by providing a framework for designing and evaluating new antimicrobial agents with improved efficacy.
抗生素耐药细菌菌株的迅速增加是一个重大的全球健康问题，因此需要开发新的有效抗菌剂。本研究的重点是通过将喹诺酮环（氟喹诺酮的衍生物）与潜在的ATP酶抑制剂融合，设计氟喹诺酮（FQ）描述符的从头合成模型，该潜在的ATP酶抑制剂是通过针对金黄色葡萄球菌DNA促旋酶B（gyrB）蛋白的药效团模型确定的。最初，基于DNA gyrB蛋白（PDB ID：3TTZ、3U2D和3U2K）生成药效团模型，特别针对其共结晶的ATP酶抑制剂，以开发具有疏水区域、氢键受体（HBA）、氢键供体（HBD）、芳香部分（Ar）和卤键供体（XBD）等关键特征的共享特征药效团（SFP）。使用0.2641的命中优度（GH）分数进一步评估该图谱，表明该图谱在捕获潜在化合物方面的优势。该SFP用于对来自ZINCpharmer和ChEMBL的160,000种化合物进行虚拟筛选，产生了74个命中物（48个来自ZINCpharmer，26个来自ChEMBL），它们具有相似的ATP酶特征，最佳拟合分数范围为73.50至76.80，均方根偏差（RMSD）值为0.1至0.5。使用遗传算法和基于片段的设计将这些鉴定出的配体与氟喹诺酮的喹诺酮环融合，以创建50个新的氟喹诺酮合成模型。这些模型中的大多数含有吡咯环，平均相似度超过55%，并且合成可及性分数（SAScore）≤3.5，以便在实验室中进行实际合成。通过药代动力学和分子对接对这些构象异构体进行评估，发现三种顶级化合物——分子13（-9.1 kcal/mol）、分子20（-9.1 kcal/mol）和分子49（-9.4 kcal/mol）——与对照环丙沙星（-7.8 kcal/mol）相比，它们与DNA促旋酶蛋白（PDB ID：4PLB）的结合亲和力更高。此外，使用薛定谔套件对这三种化合物和对照进行了200 ns的分子动力学模拟（MDS），其中分子13显示出潜在的结构稳定性。为确保实际可行性，采用计算序列到序列逆合成和化学支架比较分析为最有前景的化合物设计合成路线，证实了它们在实验室中成功合成的可能性。本研究的结果为治疗抗生素耐药性的持续努力做出了贡献，为设计和评估具有更高疗效的新型抗菌剂提供了一个框架。
