Separation mechanism and chiral-HPLC chromatogram profile of racemic mandelate compounds: a comparative study between experiment and computation using conformer–rotamer ensemble sampling tool (CREST)-XTB

Abstract

Context

Nearly 90% of drugs on the market are racemates. A racemate is a mixture of two enantiomers or substances in equal amounts that have an asymmetric molecular structure that is a mirror image of each other. Despite having the same chemical structure, chiral drug isomers can exhibit very different biological behaviors in terms of pharmacology, toxicity, pharmacokinetics, metabolism, etc. Since racemic drugs have only one bioactive enantiomer while its counterpart enantiomers impart undesirable pharmacological properties, it is necessary to separate these racemic compounds to obtain the desired active enantiomer. Chromatography is one of the approaches for the separation of enantiomers. In this study, we observed the chromatographic profile of racemic mandelic acid compound passed through a chiral HPLC column. The chromatogram profile was then observed computationally to study the separation mechanism. The experimental results are in line with the computational analysis that the S chromatogram eluted first compared to the R-enantiomer. It can be predicted that the binding energy of the R-enantiomer (–108.92 kJ/mol) is stronger than the S-enantiomer (− 67 kJ/mol).

Methods

The chromatogram profile of mandelic acid racemate was observed experimentally using a chiral OD column, and the prediction of column-ligand binding energy was based on computational studies using the conformer–rotamer ensemble sampling tool (CREST). The chromatogram profile was identified using a 0.46 cm × 25 cm chiral OD column HPLC instrument (Daicel Chemical). The samples used were racemic compounds of mandelic acid and standard (S)-mandelic acid. Computational calculations of column capacity factors and binding energies of each enantiomer were performed with a Windows 11 Pro 64-bit operating system, × 64-based processor, equipped with the MGL-Tools program consisting of the ADT (Autodock Tools) application, Avogadro, AutoDock 4.2, Discovery Studio 2020 Client®, and CREST installed as a driver program for the XTB semiempirical quantum chemistry package. For geometry optimization and sampling of DMPC-ligand complexes, we used CREST at https://github.com/grimme-lab/crest.