Retention times and bandwidths in reversed-phase gradient liquid chromatography of peptides and proteins.

Due to dramatic effects of even small changes in mobile phase composition on the retention, separations of high-molecular compounds are very difficult, if possible at all, at isocratic conditions and need gradient elution. The theory of gradient elution for small molecules is well established, however its applications to reversed-phase gradient separations of biopolymers are not straightforward because of specific problems, such as slow diffusion, limited accessibility of the stationary phase for larger molecules, or possible sample conformation changes during the elution. Theoretical prediction of gradient data needs the parameters of model retention equations to be known, which however cannot be determined at isocratic conditions. The present work overviews the attempts at implementation of the conventional gradient theory developed for low-molecular compounds to the description and prediction of gradient separations of peptides and proteins on various types of HPLC columns: conventional analytical columns packed with wide-pore fully porous, fused-core superficially porous and non-porous particles; silica-based monolithic columns and organic-polymer poly(alkylmethacrylate) and poly(styrene-divinylbenzene) monolithic columns in capillary and disc formats. The attention is focused on the determination of the parameters necessary to predict gradient retention times (volumes) and bandwidths using the theoretical model equations. The accuracy of the prediction of protein retention on totally porous columns improves if size exclusion effect is taken into account, but this is not necessary with non-porous or superficially porous particles. Band dispersion effects counteracting band compression in gradient elution depend on the type of column, on the protein and on the gradient volume (steepness) and complicate the prediction of band broadening in gradient chromatography of proteins, however the conventional gradient model can be employed to estimate the effects of changing gradient parameters on the bandwidths, as well as on the elution times (volumes) of proteins.