Descriptive structure–separation relationship studies in chiral ligand‐exchange chromatography

Abstract

A new approach to model and interpret the separation of amino acid enantiomers is described. The descriptive structure-separation relationship (DSSR) study of a set of amino acid enantiomer couples analyzed with two chiral selectors S-benzyl-(R)-cysteine ((R)-SBC) and S-trityl-(R)-cysteine ((R)-STC) has been performed with the aim to highlight the molecular properties of the analyte affecting the enantiodiscrimination process. Although with some exceptions, the classification models obtained with (R)-STC indicate that the partial positive surface area (PPSA-1) is the suitable descriptor to account for the separation of both (R)- and (S)-enantiomers. On the other hand, two different descriptors characterize the separation process with (R)-SBC. While fractional negatively charged partial surface area (FNSA-3) results the discriminating descriptor for (R)-enantiomers, the relative polar surface area (RPSA) is able to best discriminate among (S)-enantiomers.

Cysteine-based chiral selectors for the ligand-exchange separation of amino acids☆

“…Davankov [36] reported that there is no thermodynamic without kinetic enantioselectivity, whereas the opposite assertion is not valid. With all the employed sulphur-containing chiral selectors and with the exception of His and AIDA, (S)-enantiomers are more retained than their antipodes [16,37]. Thiophene-derived amino acids do not represent exceptions since the presence of the sulphur atom reverses the usual amino acid priority.…”

“…With all the adopted selectors, aspartic acid (Asp), F94C, O-phospho-serine (OPho-Ser) and ornitine (Orn) enantiomer couples did not experience any enantiorecognition. While a comparable enantioseparation of threonine (Thr) antipodes was achieved with (R)-STC and (R)-SDC, (R)-SBC was not effective thus providing for the enantiomers coelution [16]. With the exception of 5-MATIDA, F94B, cysteine (Cys) and CIP-B enantiomer couples which resulted better discriminated with (R)-SDC, a gain in enantiodiscrimination was obtained for all the remaining compounds when the most lipophilic chiral selector was engaged [(R)-STC].…”

“…A comparison between (R)-STC and S-benzyl-(R)-cysteine [(R)-SBC] (Fig. 1), which differ for the number of the aromatic rings in the side chain, has been also performed with the aim of identifying molecular properties of analytes that affect the enantiodiscrimination process through a descriptive structure-separation relationship (DSSR) study [16]. For such a purpose, the selection was directed toward a cumulative pool of 40 descriptors (Jurs [17] and Shadow descriptors [18]) which relate to the physico-chemical properties and shape of the molecular surface area.…”

Exploring enantioselective molecular recognition mechanisms with chemoinformatic techniques

“…separated or nonseparated racemates, can be easily individuated. This concept has been recently adapted by Natalini et al [148,149] that carried out a descriptive structure-separation relationship (DSSR) in which some set of physicochemical descriptors are directly used with a partition tree or a classification model towards the experimental enantioselectivities. In a quantitative form also Del Rio and Gasteiger [147] applied this idea to devise a simple method based on the atomic counts around the chiral centre to assess whether a compound will be separated or not on a Whelk-O1 CSP.…”

“…It is important not to confuse the discrete or continuous nature of the dependent variable with the building of quantitative or qualitative models. For instance, a classification problem can be both qualitative (see Natalini et al [148]) or quantitative (see Del Rio and Gasteiger [147]) as well.…”