Cálculos de deslocamentos químicos de RMN 1H e 13C e propriedades eletrônicas de canabinoides utilizando cálculos quânticos

Abstract

The present work is in the field of Computational Chemistry combined with Density Functional Theory (DFT) to provide in silico approaches, which assist in experimental planning activities, processing of 1H and 13C Nuclear Magnetic Resonance (NMR) data, conformational search, and identification of compounds with unclear spatial structure. The compounds evaluated were Cannabidiol (CBD), Δ9 Tetrahydrocannabinol (THC) and Canabinol (CBN), which are specialized metabolites of Cannabis sativa and have clinical and forensic relevance. Macromodel® and Gaussian® were used, respectively, in Molecular Mechanics and Quantum Mechanics optimizations. Subsequently, the population distribution (Boltzman) of the conformers found was verified and the most representative ones were selected to simulate NMR spectra at eight levels of theory. It was found that the theoretical level DFT/PBE0/6-311G+(2d,p) was the most accurate for the compounds studied, obtaining MAD and RMSE values close to the experimental chemical shifts (δ). It was found that the use of the internal standard Dioxane significantly improved the accuracy of the theoretical δ for 13C. While the use of tetramethylsilane (TMS) or Dioxane did not show the same significance in the 1H simulations. A possible inversion of the C8 and C9 methyls of Δ9 THC attributed in the literature was investigated. Thus, it was concluded, by means of the MAD values and after an intentional inversion of the theoretical δ values, that these were diastereotopic carbons. Therefore, C8 and C9 methyls have been reported inverted in the literature. The best levels of theory were compared using the DFT method versus the Hartree-Fock method, and the DFT method was found to be the most accurate. Electronic property data was obtained, such as HOMO, LUMO, Hardness, Softness, electronegativity and electrophilicity, and reactivity issues were discussed together with the Electrostatic Potential Map. It was verified that the results were coherent regarding the chemical nature of the compounds under study. Finally, the computational processing time for each stage of this work was verified. It was observed that molecular flexibility is directly proportional to processing time. Therefore, the present work concluded that Computational Chemistry is an important tool in the optimization and prediction of molecular studies.