QSAR studies on benzothiophene derivatives as Plasmodium falciparum N‐myristoyltransferase inhibitors: Molecular insights into affinity and selectivity

Abstract

Malaria is an infectious disease caused by protozoan parasites of the genus Plasmodium and transmitted by Anopheles spp. mosquitos. Due to the emerging resistance to currently available drugs, great efforts must be invested in discovering new molecular targets and drugs. N‐myristoyltransferase (NMT) is an essential enzyme to parasites and has been validated as a chemically tractable target for the discovery of new drug candidates against malaria. In this work, 2D and 3D quantitative structure–activity relationship (QSAR) studies were conducted on a series of benzothiophene derivatives as P. falciparum NMT (PfNMT) and human NMT (HsNMT) inhibitors to shed light on the molecular requirements for inhibitor affinity and selectivity. A combination of Quantitative Structure–activity Relationship (QSAR) methods, including the hologram quantitative structure–activity relationship (HQSAR), comparative molecular field analysis (CoMFA), and comparative molecular similarity index analysis (CoMSIA) models, were used, and the impacts of the molecular alignment strategies (maximum common substructure and flexible ligand alignment) and atomic partial charge methods (Gasteiger‐Hückel, MMFF94, AM1‐BCC, CHELPG, and Mulliken) on the quality and reliability of the models were assessed. The best models exhibited internal consistency and could reasonably predict the inhibitory activity against both PfNMT (HQSAR: q2/r2/r2pred = 0.83/0.98/0.81; CoMFA: q2/r2/r2pred = 0.78/0.97/0.86; CoMSIA: q2/r2/r2pred = 0.74/0.95/0.82) and HsNMT (HQSAR: q2/r2/r2pred = 0.79/0.93/0.74; CoMFA: q2/r2/r2pred = 0.82/0.98/0.60; CoMSIA: q2/r2/r2pred = 0.62/0.95/0.56). The results enabled the identification of the polar interactions (electrostatic and hydrogen‐bonding properties) as the major molecular features that affected the inhibitory activity and selectivity. These findings should be useful for the design of PfNMT inhibitors with high affinities and selectivities as antimalarial lead candidates.

Structural Predictive Model of Presenilin-2 Protein and Analysis of Structural Effects of Familial Alzheimer’s Disease Mutations

“…As there is no crystallized structure, structural biology tools are used to generate the protein model, and two structure predictors, I-Tasser and Phyre2, are chosen for this. These software programs generate their models from an algorithm that allows for creating the hypothetical model, based on the homology comparison and folding recognition through the primary sequence, and these sequences are joined under a threading-type assembly [ 66 – 70 ]. The prediction models used, as a template, the subunit B of the CryoEM model for the PDB with ID 6IYC [ 71 ] and a complete artificial model of the PS1 protein that guides the generation of the missing flexible region.…”

A Molecular Image-Based Novel Quantitative Structure-Activity Relationship Approach, Deepsnap-Deep Learning and Machine Learning

“…Fifth, a 4D-molecular descriptor is calculated through the interaction with other compounds, for example, the interaction energy resulting from molecular dynamics simulation (Kumar andKulkarni, 2017, Ataide Martins et al, 2018;Ma et al, 2019). Since the affinity of chemical substances to protein enzymes and receptors strongly depends on intermolecular interactions, such as hydrogen bonding, electrostatic interactions, and hydrophobic interactions, it has been observed that the complementary positional relationship between the base groups of the chemicals and their binding site of the enzyme or receptor required for the interactions of chemical compounds is important (Yang et al, 2017;Kimani et al, 2018;Pan et al, 2019;Garcia et al, 2020). Therefore, besides molecular descriptors that show physiochemical properties related to intermolecular interactions, QSAR analysis in a 3D space that utilizes descriptors representing the 3D structure of chemical substances has been recently attracting research attention (Chen et al, 2020;Hadni and Elhallaoui, 2020;Kumar et al, 2020;Liu et al, 2020;Zhang et al, 2020).…”

“…Further, in the case of the best subset selection or round-robin method, in which analysis is performed by utilizing all the combinations of the explanatory variables, there are 2 k − 1 combinations of k explanatory variables, resulting in a huge calculation cost (Noon et al, 2011;Saavedra et al, 2020). Additionally, based on the usefulness of each univariate regression coefficient, there are some sequential selection methods, which include forward−backward stepwise selection method, forward stepwise selection method, backward stepwise selection method, and backward−forward stepwise selection method, that sequentially increase or decrease the explanatory variables individually (Goodarzi et al, 2012;Fatima et al, 2018;Fatima et al, 2019;Hrynkiewicz et al, 2019;Fatima and Agarwal, 2020;McCann et al, 2020). However, the effect of unselected explanatory variables is not corrected, so to make variable selections, it is necessary to include technical and academic knowledge to the variable selection criteria.…”

Protein Predictive Modeling and Simulation of Mutations of Presenilin-1 Familial Alzheimer’s Disease on the Orthosteric Site

“…On the other hand, semi-empirical methods seem to be a promising alternative in the study of biological and polyatomic systems as these intend to solve the Schrödinger equation from an approximated perspective, that is, considering an average between the electron interactions and appropriated theory levels, reducing the computational time. Of particular interest, the semi-empirical Austin model 1 is characterized because its parameters are derived from experimental data in order to solve the Schrödinger equation ( Carvalho et al, 2014 ; Christensen et al, 2016 ; Rafique et al, April 2019 ; Garcia et al, 2020 ; Grillo et al, 2020 ). These can be applied efficiently to large molecules to calculate their respective surface potentials ( Foresman and Frisch, 1996 ; Levitt, 2014 ; Silva et al, 2015 ; Marín and Soto-ospina, 2020 ; Cano et al, 2021 ).…”