How to increase binding energy in molecular docking in Bioinformatics?

The high-throughput virtual screening provided is built using molecular docking, which is essential in predicting binding conformations and energies of ligands to receptors. AutoDock 4.2, a widely used docking tool with outstanding accuracy, is one of several molecular docking programmes developed so far. It has been established that partial charge calculation is critical for accurate conformation search and binding energy estimation. However, no rigorous evaluation of the effects of electrostatic potentials on AutoDock 4.2 docking accuracy has been made. Using AutoDock4, nine various charge-assigning approaches were sufficiently studied for their molecular docking performance, including AM1-BCC, Del-Re, formal, GasteigerHückel, GasteigerMarsili, Hückel, Merck molecular force field (MMFF), and Pullman, as well as the ab initio Hartree-Fock charge. When applying semi-empirical, empirical, and even ab initio charge approaches to evaluate binding energy estimation, a somewhat poor association was found between the docking score and in vitro measured ligand activity.

To some extent, all nine charge techniques used in our research could not accurately compute binding energies. It should be noted that partial charges are difficult to calculate since, despite their intuitive nature, partial atomic charges do not exist. Partial charges are just simplified representations of genuine electrostatic potential energy in molecules.

The empirical Gasteiger Hückel charge is the most applicable in virtual screening for large databases. According to the findings, the semi-empirical AM1-BCC charge can be used in lead compound optimization and accurate virtual screening for small datasets. In general, the semi-empirical charge technique AM1-BCC, ab initio charge method Hartree-Fock, and empirical charge method MMFF performed better correlations than other charge methods, with considerable superiorities. The results of predicting binding energy using crystal structure revealed that all nine charge methods, particularly the AM1-BCC, MMFF, and Hartree-Fock, aimed to underestimate the binding energy. Because atomic charges play a significant role in the computation of desolvation potential (Edesolv) and electrostatic interactions in the AutoDock 4.2 scoring function, different charge techniques result in varied computed binding energies (Eelec). It’s worth noting that excluding Eelec from overall scores resulted in minor improvements in all nine charge techniques estimated binding energies.

References

  1. Tatu Pantsar1 and Antti Poso1,2,*, Binding Affinity via Docking: Fact and Fiction, Molecules. 2018 Aug; 23(8): 1899.
  2. Xuben Hou,1 Jintong Du,2 Jian Zhang,3 Lupei Du,4 Hao Fang,*1 and Minyong Li*1; How to Improve Docking Accuracy of AutoDock4.2: A Case Study Using Different Electrostatic Potentials; dx.doi.org/10.1021/ci300417y | J. Chem. Inf. Model. 2013, 53, 188−200.