Application of the model described in J. Chromatogr., 327, 27 (1985) to the separation of protein/peptide mixtures by reversed-phase gradient elution.

A previously reported model describes retention and band width as functions of experimental conditions, for the reversed-phase gradient elution separation of peptides and proteins. The model begins by relating separation in gradient elution to corresponding separations by isocratic elution (same high-performance liquid chromatographic system). The use of certain semi-empirical relationships then allows band width to be predicted for samples for which isocratic data are unavailable. This in turn allows the prediction of band width, peak capacity and relative peak height (or peak volume) as a function of experimental conditions such as column dimensions, gradient conditions, mobile phase flow-rate, sample molecular weight, etc. The next paper [J. Chromatogr., 327 (1985) 93] demonstrates the practical value of this approach for optimizing the separation of peptide and protein mixtures.In the present paper it is first shown that isocratic data for desamido-insulin (6000 daltons) allow the accurate prediction of band widths as a function of experimental conditions in related gradient separations. It is further shown that the present model accurately predicts how band widths vary with experimental conditions for peptides and proteins having molecular weights in the range 600–80 000 daltons.

The presently accepted theory for gradient separations of small molecules has been used to develop a predictive model for peptides and proteins as samples, using reversed-phase high-performance liquid chromatography. Given the experimental conditions (gradient time, flow-rate, temperature, etc.), the molecular weight of the sample, and certain column characteristics (Knox parameters, column dimensions, particle diameter, etc.), it is possible to calculate the overall results of a given separation by gradient elution: peak capacity or average resolution, peak volume or relative peak height, etc. This information can in turn facilitate the optimized separation of any sample. The present model assumes that isocratic and gradient retention are interrelated for peptide molecules, in the same fashion as for small molecules. This assumption has been verified for various peptides and proteins and further used to gain new insight into the control of retention and band-spacing in gradient elution.

Equipment for high-performance liquid chromatographic (HPLC) gradient elution generally distorts the gradient selected by the operator, which in turn affects the retention of solutes separated by gradient elution. A theoretical analysis describes these gradient distortions as a function of equipment design and operating conditions. Comparisons of theory with experimental data show generally good agreement. As a result, it is now possible to select gradient conditions for minimal gradient distortion, or to correct for the effect of gradient distortion on solute retention. This will be shown in later papers to allow the use of gradient elution in new ways for more efficient method development and optimization of separation by HPLC.An analysis of instrumental factors that can limit the accuracy of gradient retention data.

Under “ideal” conditions it is possible to model retention in gradient elution so as to be able to calculate retention times, tG, as a function of isocratic retention in corresponding liquid chromatographic systems. In this paper we consider various “non-ideal” processes that lead to errors in calculated values of tG. 

An HPLC-method for the measurement of blood Cyclosporin A levels (CyA) of renal allograft transplanted patients within 9 min is described. After a simple protein precipitation of the blood the supernatant is transferred to an HPLC-system. The short time of analysis is obtained by a step gradient elution technique and a precolumn separation of the fractions of interest followed by a backflush regeneration step of the precolumn. The analysis of the fraction of interest takes place on a column with high resolution power as long as the precolumn is regenerated. CyA is monitored by UV-absorption at 206 nm. Detection of 20 to 2000 ng/ml Cya allows the use of the method for patient monitoring and for research purposes.