Immunodepletion of albumin from human serum samples

Abstract

A frequent problem for those performing two-dimensional (2-D) gel electrophoresis [l] analyses of human serum, urine and cerebrospinal fluid samples has been the high relative abundance of albumin [2,3]. This type of electrophoresis is an exceptionally high-resolution technique for peptides and proteins, combining isoelectric focusing and SDS-PAGE. Since albumin typically accounts for 40-70% of the total protein in a serum sample, it can obscure many of the proteins near it in a 2-D gel pattern (see Fig. 1A). This observation is due to the inability of such a relatively large quantity of albumin to focus sharply in the isoelectric focusing dimension, and because large quantities of single proteins tend to create their own pH gradients. Typically a sample is electrophoresed in an acrylamide tube gel, separating proteins on the basis of their net charges. Following this step, the tube gel is incubated in an SDS-containing equilibration buffer, and placed atop an SDS-polyacrylamide slab gel. Electrophoresis in the presence of SDS, at right angles to the first (isoelectric) separation, further segregates proteins, this time on the basis of their molecular weights. We have been interested in quantitative and ratio analyses of serum protein isoforms, and their possible correlation with disease states. Unless specifically designed to do so, antibody-based quantitative assays do not discriminate among different isoforms of a protein; hence the advantage of 2-D gel methodology over ELISA and RIA protocols in this instance. We chose to stain our gels with Coomassie blue rather than the more sensitive silver stains, because it binds stoichiometrically and reproducibly to proteins over a