Use of HPLC retention to investigate new P descriptors designed to represent ion-[pi] interactions

• Published in: Journal of liquid chromatography & related technologies 2020-02, Vol.43 (3-4), p.83
• Main Authors: Bagwill, Christina S, Wireduaah, Selina, Cusworth, Brian, Korba, Joshua, Kirkpatrick, Charles C, Lewis, Michael, Stalcup, Apryll M
• Format: Article
• Language: English
• Subjects: Anion exchange; Anion exchanging; Benzene; Chromatography; Computer applications; High performance liquid chromatography; Liquid chromatography; Mathematical models; Molecular modelling; Molecular structure; Organic chemistry; Parameters; Physical properties; Retention; Solvation; Stationary phase; Statistical analysis; Structure-activity relationships
• ISSN: 1082-6076; 1520-572X
• DOI: 10.1080/10826076.2019.1667822

Computational methods increasingly provide access for Quantitative-Structure-Activity Relationship (QSAR) models to molecular descriptors that have minimal identifiable or experimentally accessible chemical/physical properties. However, molecular information inherent in chromatographic data suggests chromatography as a convenient experimental platform for legitimizing newly proposed molecular descriptors. In this study, liquid chromatographic retention data collected on a trimethylammonium (TAM) anion exchange stationary phase was used to evaluate two new molecular descriptors, P- and P+ developed to account for contributions from ion-π interactions. In the Linear Solvation Energy Relationship (LSER) model, the P- and P+ parameter replaced E. Incorporation of the new parameters was most impactful on S (sconv= -0.173 ± 0.070 vs. smod -0.115 ± 0.081). The remaining coefficients (v, b, c) were within experimental error of their respective coefficients in the conventional model. Nonparametric statistical analysis revealed comparable goodness of fits for the conventional and modified models. Finally, comparison of the contribution of individual interaction modes for substituted benzenes (log kbenzene-X) relative to benzene (log kbenzene) conducted with the new model was consistent with chemical intuition for the model-attributed ion - π and other individual interaction modes. Thus, adoption of P descriptors into the LSER model provides a convenient means for exploring models for weak ion - π interactions.