Abstract
We perform 2-D capillary electrophoresis on fluorescently labeled proteins and peptides. Capillary sieving electrophoresis (CSE) was performed in the first dimension and MEKC was performed in the second. A cellular homogenate was labeled with the fluorogenic reagent FQ and separated using the system. This homogenate generated a pair of ridges; the first had essentially constant migration time in the CSE dimension, while the second had essentially constant migration time in the MEKC dimension. In addition, a few spots were scattered through the electropherogram. The same homogenate was digested using trypsin, and then labeled and subjected to the 2-D separation. In this case, the two ridges observed from the original 2-D separation disappeared and were replaced by a set of spots that fell along the diagonal. Those spots were identified using a local-maximum algorithm and each was fit using a 2-D Gaussian surface by an unsupervised nonlinear least squares regression algorithm. The migration times of the tryptic digest components were highly correlated (r=0.862). When the slowest migrating components were eliminated from the analysis, the correlation coefficient improved to r=0.956.