A systematic analysis of protein staining in polyacrylamide gels with Coomassie Brilliant Blue (CBB) R-250 and G-250 using a high resolution densitometer allowing for quantitative measurements during staining and destaining has revealed that none of the published procedures allows quantitative measurements. Protein staining with CBB R-250 in methanol/water/acetic acid is poor, as is staining with CBB G-250 in trichloroacetic acid or perchloric acid, the latter two, however, allowing for a weak background staining. Consequently using the colloidal properties of the CBB dyes, stronger for G-250 than for R-250, it is possible to increase the sensitivity of protein staining to a detection limit of 0.7 ng bovine serum albuminlmm' gel. In addition, sensitive protein staining on a clear background is possible. Recipes are described (Section 3.11) for intensified protein staining with CBB G-250 using trichloroacetic acid or perchloric acid on a clear background. Optimal staining of proteins on a clear background can be performed with phosphoric acid and CBB G-250 in the presence of ammonium sulfate since under these conditions the colloidal state ofthe dye is optimized. Furthermore, conditions are described which allow the stable fixation ofthe protein-dye complex. Combining the optimized staining conditions with the stable fixation in 20 % ammonium sulfate allows for stepwise staining fore. g. detection of weak spots in addition to intense protein spots. The dependence of different staining procedures on gel thickness, gel concentration and compounds routinely used in polyacrylamide gel electrophoresis is also analysed. Calibration curves and application of the new procedure to biological material demonstrate its wide applicability. Convincing arguments for the colloidal properties of the CBB dyes are presented, formulating the rationale for intensified protein staining with CBB dyes in polyacrylamide gels without background staining. Abbreviations: BSA, bovine serum albumin; CBB, Coomassie Brilliant Blue; DMSO, dimethylsulfoxide; EDTA, ethylenediaminetetraacetic acid djsodium salt; MeOH, methanol; ~p -4 0 , id^^ p.40; PAG, polyacrylamide gel: PAGE, polyacrylamide gel electrophoresis; PCA, perchloric acid: PEG, polyethylene glycol; SDS, sodium dodecyl sulfate; TCA, trichloroacetic acid; Tris, tris(hydroxymethy1)aminomethane 'C'VCH Verlagsgesellschaft mbH, D-6940 Weinheim. 1385 01 73-0835/85/0909-0427 $02.50/0

Quantitative determination of stained proteins following polyacrylamide gel electrophoresis (PAGE) is of increasing interest especially since computer-aided densitometers have become available as well as recipes for sensitive and background-free staining with Coomassie Brilliant Blue dyes. However, avoidance of separation artifacts is not the only essential prerequisite for quantitative evaluation. The local particle density of a protein in a given gel is of critical importance since it determines its stainability. Depending on local protein concentration, the dye binding to the same amount of a given protein differs considerably. Since the stainability of proteins using colloidal staining procedures, as with Coomassie Brilliant Blue dyes, is time-dependent and, in addition, also dependent on the pore size of a given polyacrylamide gel used for PAGE, calibration curves for quantitative determinations have to be prepared in polyacrylamide gels of the same composition as used for PAGE. Staining conditions also have to be identical for calibration gels and gels under analysis. If, however, a set of calibration curves is prepared for different staining times, it is possible to calculate a generalized calibration curve, allowing for quantitative evaluation with flexible staining time. Furthermore, and in consequence of the implications due to particle density, quantitative determination via densitometry is only possible by determining the protein amount of each single measuring point (pixel) via its absorbance on the basis of a calibration curve. Since the particle density is inherent in a calibration curve, the final summation of the protein amount per pixel will give values close to reality.

No abstract