In silico method development for the reversed-phase liquid chromatography separation of proteins using chaotropic mobile phase modifiers

Abstract

Recent advances in biomedical and pharmaceutical processes has enabled a notable increase of protein- and peptide-based drug therapies and vaccines that often contain a higher-order structure critical to their efficacy. Hyphenation of chromatographic and spectrometric techniques is at the center of all facets of biopharmaceutical analysis, purification and chemical characterization. Although computer-assisted chromatographic modeling of small molecules has reached a mature stage across the pharmaceutical industry, software-based method optimization approaches for large molecules has yet to see the same revitalization. Conformational changes of biomolecules under chromatographic conditions have been identified as the major culprit in terms of sub-optimal modeling outcomes. In order to circumvent these challenges, we herein investigate the outcomes generated via computer-assisted modeling from using different chaotropic and denaturing mobile phases (trifluoroacetic acid, sodium perchlorate and guanidine hydrochloride in acetonitrile/water-based eluents). Linear and polynomial regression retention models using ACD/Labs software were built as a function of gradient slope, column temperature and mobile phase buffer for eight different model proteins ranging from 12 to 670 kDa (holo-transferrin, cytochrome C, apomyoglobin, ribonuclease A, ribonuclease A type I-A, albumin, y-globulin and thyroglobulin bovine). Correlation between experimental and modeled outputs was substantially improved by using strong chaotropic and denaturing modifiers in the mobile phase, even when using linear regression modeling as typically observed for small molecules. On the contrary, the use of conventional TFA buffer concentrations at low column temperatures required the used of polynomial regression modeling indicating potential conformational structure changes of proteins upon chromatographic conditions. In addition, we illustrate the power of modern computer-assisted chromatography modeling combined with chaotropic agents in the developing of new RPLC assays for protein-based therapeutics and vaccines.