Computational MitoTarget Scanning Based on Topological Vacancies of Single-Walled Carbon Nanotubes with the Human Mitochondrial Voltage-Dependent Anion Channel (hVDAC1)

Abstract

We present an in silico approach for modeling the noncovalent interactions between the human mitochondrial voltage-dependent anion channel (hVDAC1) and a family of single-walled carbon nanotubes (SWCNTs) with a defined pattern of topological vacancies (v = 1–16), obtained by removing atoms from the SWCNT surface. The general results showed more stable docking interaction complexes (SWCNT–hVDAC1), with more negative Gibbs free energy of binding affinity values, and a strong dependence on the vacancy number (R 2 = 0.93) and vacancy formation energy (R 2 = 0.96). In addition, for most of the SWCNT vacancies that were analyzed, the interatomic distances for the interactions of the SWCNT–hVDAC1 complex with the functional catalytic residues (i.e., Pro7, Gln199, Gln182, Phe181, Val20, Asp19, Lys15, Gly14, Asp12, Ala11, and Arg18) that form the hVDAC1 active site (i.e., the voltage-sensing N-terminal α-helix segment) were very similar to or shorter than the interatomic distances of these residues for ATP–hVDAC1 interactions. In particular, the hVDAC1 residues that can be phosphorylated like Tyr10, Tyr198, and Se16 were significantly perturbed by the interactions with SWCNT with at least nine vacancies. In addition, the SWCNT vacancy family members can affect the flexibility properties of the hVDAC1 N-terminal α-helix segment inducing different patterns of local perturbations in inter-residue communication. Finally, vacancy quantitative structure–binding relationships (V-QSBRs) were unveiled for setting up a robust model that can predict the strength of docking interactions between SWCNTs with a specific topological vacancy and hVDAC1. The developed V-QSBR model classified properly all of the SWCNTs with a different number of SWCNT vacancies with exceptional sensitivity and specificity (both equal to 100%), indicating a strong potential to unequivocally predict the influence of SWCNT vacancies on the mitochondrial channel interactions.

Computational Drug Repurposing for Antituberculosis Therapy: Discovery of Multi-Strain Inhibitors

“…This means that the LQI descriptor for a molecule will have the same value regardless of the experimental condition cj used to assess the anti-TB activity of that molecule. To solve this inconvenience, we applied the adaptation of the Box–Jenkins approach, which is a distinctive characteristic of all the PTML models [ 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 55 ]: …”

“…To solve the aforementioned limitations, several researchers have emphasized the use of interpretable in silico models focused on a combination of perturbation theory concepts and machine learning techniques (PTML) [ 15 , 16 , 17 ], which can integrate different sources of chemical and biological data, enabling the simultaneous prediction of multiple biological endpoints against many targets of varying degrees of complexity. Seminal works on PTML models have found successful applications in diverse research areas such as infectious diseases [ 18 , 19 ], oncology [ 20 , 21 ], neuroscience [ 22 , 23 , 24 , 25 ], proteomics [ 26 ], metabolomics [ 27 ], nanotechnology [ 28 , 29 , 30 , 31 ], toxicology [ 32 ], and immunology and immunotoxicity [ 33 , 34 ].…”