Department of Chemistry , Texas A&M University , P.O. Box 30012, College Station , Texas 77842 , United States.
Department of Chemistry , John Brown University , 2000 West University Street , Siloam Springs , Arkansas 72761 , United States.
J Chem Inf Model. 2018 Oct 22;58(10):2085-2091. doi: 10.1021/acs.jcim.8b00417. Epub 2018 Sep 12.
Human infection by Mycobacterium tuberculosis (Mtb) continues to be a global epidemic. Computer-aided drug design (CADD) methods are used to accelerate traditional drug discovery efforts. One noncovalent interaction that is being increasingly identified in biological systems but is neglected in CADD is the anion-π interaction. The study reported herein supports the conclusion that anion-π interactions play a central role in directing the binding of phenyl-diketo acid (PDKA) inhibitors to malate synthase (GlcB), an enzyme required for Mycobacterium tuberculosis virulence. Using density functional theory methods (M06-2X/6-31+G(d)), a GlcB active site template was developed for a predictive model through a comparative analysis of PDKA-bound GlcB crystal structures. The active site model includes the PDKA molecule and the protein determinants of the electrostatic, hydrogen-bonding, and anion-π interactions involved in binding. The predictive model accurately determines the Asp 633-PDKA structural position upon binding and precisely predicts the relative binding enthalpies of a series of 2-ortho halide-PDKAs to GlcB. A screening model was also developed to efficiently assess the propensity of each PDKA analog to participate in an anion-π interaction; this method is in good agreement with both the predictive model and the experimental binding enthalpies for the 2-ortho halide-PDKAs. With the screening and predictive models in hand, we have developed an efficient method for computationally screening and evaluating the binding enthalpy of variously substituted PDKA molecules. This study serves to illustrate the contribution of this overlooked interaction to binding affinity and demonstrates the importance of integrating anion-π interactions into structure-based CADD.
人感染结核分枝杆菌（Mtb）仍然是一个全球性的流行。计算机辅助药物设计（CADD）方法被用来加速传统的药物发现工作。在生物系统中越来越多地被识别出来但在 CADD 中被忽视的一种非共价相互作用是阴离子-π相互作用。本文报道的研究支持这样的结论，即阴离子-π相互作用在指导苯二酮酸（PDKA）抑制剂与苹果酸合酶（GlcB）结合中起着核心作用，GlcB 是结核分枝杆菌毒力所必需的酶。使用密度泛函理论方法（M06-2X/6-31+G(d)），通过比较 PDKA 结合的 GlcB 晶体结构，为预测模型开发了 GlcB 活性位点模板。活性位点模型包括 PDKA 分子和涉及结合的静电、氢键和阴离子-π相互作用的蛋白质决定因素。预测模型准确地确定了 Asp 633-PDKA 结构位置的结合，并精确预测了一系列 2-邻卤代 PDKA 与 GlcB 的相对结合焓。还开发了一个筛选模型，以有效地评估每个 PDKA 类似物参与阴离子-π相互作用的倾向；该方法与预测模型和 2-邻卤代 PDKA 的实验结合焓非常吻合。有了筛选和预测模型，我们开发了一种有效的方法来计算筛选和评估各种取代的 PDKA 分子的结合焓。本研究说明了这种被忽视的相互作用对结合亲和力的贡献，并证明了将阴离子-π相互作用纳入基于结构的 CADD 的重要性。
