In the shadow of rapid microcapillary biochip analysis methods, slab-gel SDS-PAGE still affords researchers high versatility and control. This electrophoresis method better preserves sample integrity and supports coupled methods and varied dyes.

In the almost 70 years since Arne Tiselius first applied moving boundary electrophoresis to fractionate serum proteins, Serono International SA, Geneva, Switzerland (a

click the image to enlarge   Figure 1: The relation between amino acid sequence and formula weight is blindsided by posttranslational modifications. Sh = sham digestion in which protein J was denatured, heated, and incubated overnight at 37 °C in enzyme-free digestion buffer; C = control aliquot of protein J stored overnight in PBS at 4 °C; N = protein J digested with N-glycanase (1,000 dgu/ml); S + N = protein J digested three hours with sialidase (40 U/mL) and then overnight with N-glycanase (1,000 dgu/mL); M = Bio-Rad broad range molecular weight markers; Nb = enzyme-only digestion blank for N-glycanase; (S+N)b = enzymes-only digestion blank for sialidase followed by N-glycanase.

company more than 100 years old) has witnessed widespread "adaptive radiation" of the use of electrophoresis in biochemistry. In the original zonal electrophoretic technique, called native polyacrylamide gel electrophoresis (native PAGE), moving electrophoretic boundaries and the zones they create migrate through a matrix of cross-linked polyacrylamide [L. Ornstein, Ann. N.Y. Acad. Sci., vol. 121, pp. 321-349 (1964); B. J. Davis, Ann. N.Y. Acad. Sci., vol. 121, pp. 404-427 (1964)]. Native PAGE is seldom used today because another method, SDS-PAGE, gives sharper bands and is relatively immune to effects determined by protein shape and charge. The anionic detergent sodium dodecyl sulfate (SDS) strongly associates with most proteins. Interaction with SDS disrupts protein quaternary, tertiary, and secondary levels of structure and renders all proteins highly negatively charged. When used in conjunction with a reducing reagent in order to disrupt disulfide bonds, SDS-PAGE approximates an analysis of randomly-coiled protein molecules in which relative electrophoretic mobility is inversely proportional to the logarithm of polypeptide chain length. In the wake of "labs on a chip" [D. J. Beebe et al., Annual Review of Biomedical Engineering, vol. 4, pp. 261-286 (2002)], it seems appropriate for companies such as Serono to ask whether SDS-PAGE is still worthwhile, and if so, then for what applications.

Estimated Protein Formula Weight is of Limited Value
A secreted glycoprotein (protein J) that Serono researchers worked on nearly 10 years ago is a good example of the problem with protein formula weight estimation by SDS-PAGE. Posttranslational addition of N- and/or O-linked oligosaccharides, even when polypeptide mass can be accurately predicted from genetic information, often results in a formula weight that is different from the genetically predicted value. The expected average polypeptide mass of protein J (Figure 1) is 24,411.7 amu [atomic mass units]. By several SDS-PAGE methods, including the Tris-Tricine gradient gel [H. Schagger, G. von Jagow, Anal. Biochem., vol. 166, pp. 368-379 (1987)] Figure 1 shows that 6 micrograms of protein J (by amino acid analysis) were loaded in each experimental lane.

The amino acid sequence of protein J contains five identifiable N-glycosylation sites. As demonstrated in Figure 1, an apparent molecular weight for protein J, consistent with the expected polypeptide mass, could be obtained by SDS-PAGE following peptide N-glycosidase (PNGase) digestion, but the holoprotein formula weight of protein J is still unknown. MALDI-TOF mass spectrometry has made it possible to determine such formula weights, including the glycosylation, directly (see Figure 1), if a relatively homogeneous preparation is available. Nevertheless, one must still produce that first lot of purified protein. A second lesson taught by glycoproteins such as protein J is to rely on amino acid composition analysis, rather than Coomassie blue dye-binding, to determine protein concentration. Figure 1 shows that 6 mg of protein J (by amino acid analysis) were loaded in each experimental lane.

Burdensome Time for SDS-PAGE
A disadvantage that SDS-PAGE shares with all electrophoretic methods is joule heating. Generating electro-phoretograms very quickly (36 minutes in MES buffer, pH 7.2 or 55 minutes in MOPS, pH 7.8, for example) can have unintended consequences because high-power electrophoretic separations "run hot." The lane labeled H NR

click the image to enlarge   Figure 2: Joule heating can be an issue in electrophoretic separations. Recombinant heterodimeric glycoprotein (H) analyzed in disulfide-reduced (HR) or nonreduced (H NR) form in 1 mm precast 10% acrylamide Nu-PAGE Bis-Tris gels with MOPS (pH 7.8) running buffer. Separations were carried out on the bench at constant voltage gradient (lane labeled 200V/RT) or in the cold room at constant current (lane labeled 30 mA/4°); lane labeled "both" indicates electrophoretograms of reduced sample and MW markers were not affected by the difference between separation protocols.

200V/RT in Figure 2 illustrates denaturation, by joule heating, of a recombinant hetero-dimeric glycoprotein. The lane labeled H NR 30 mA/4° shows the same sample subjected to significantly less joule heat. After electrophoresis at 30 mA/gel and 4 °C, a heterodimer band consistent with the MALDI-TOF formula weight can be routinely recovered, but it takes 3.5 hours to effect such a separation instead of 55 minutes. Is it more important to complete a separation in 55 minutes or to keep the molecules under study intact? Clearly, the joule heat of a high-power electrophoretic separation might less adversely affect other proteins, but ultimately, it is analyte integrity that needs to be valued more highly than the speed of a separation.

Simultaneous Parallel Analyses
Coomassie blue-stained SDS-PAGE slab gels are effective tools by which to evaluate homogeneity during purification. The strength of the method is side-by-side comparison of samples separated in parallel and stained/ destained simultaneously. Digital image analysis of SDS-PAGE gels is a quasi-quantitative technique, at best, unless a reference standard is available for each analyte. Nevertheless, the case in Figure 3 illustrates how Serono scientists use Coomassie blue-stained SDS-PAGE and digital image analysis to guide decision-making during purification of the "first lot" of a new recombinant protein. Panel A shows column fractions loaded to equivalent A280. The bracketed fractions were pooled and moved forward into a gel permeation chromatography polishing step, the results from which appear in panels B and C.

Protein Zones for Coupled Analyses
Slab gel PAGE supports coupled methods like SDS-PAGE/Edman degradation and SDS-PAGE/Western blotting. Electrophoretically separated proteins that have been transferred to PVDF are directly compatible with automated Edman degradation or detection determined by interaction with an antibody or other ligand. Figure 4 illustrates

click the image to enlarge   Figure 3: SDS-PAGE and image analysis assist decision-making in protein fractionation. (A) Fractions from an anion exchange column loaded to equal A280/lane. Gels were stained with Coomassie brilliant blue R-250. (B) the GPC peak heart, 10 mg (by BCA) loaded per lane. (C) Tabulated results derived by expressing each band as a fraction of integrated optical density in the lane, lane baseline having been modeled with a rolling circle of the appropriate diameter. The heterodimeric protein in this figure is not the same protein as appears in Figure 2. (Data courtesy of Dr. T. Brush.)

semiquantitative Western analysis of a protein we can refer to as "R." R was initially expressed and purified in a tagged form. A polyclonal antibody directed against a peptide from the N-terminal domain of R was available commercially. The tagged form of R was used as a calibration standard in lanes 2 through 6, so that we might estimate the volumetric expression of R in culture medium conditioned by cells transfected to express untagged R (lanes 7 through 10). Some results in a previous experiment sent to us from another lab suggested proteolysis in the conditioned medium might be a problem. Therefore, this sample was divided immediately after harvest and 0.2 mm filtration, and a protease inhibitor cocktail was added to half the filtered culture supernatant (labeled +PI). A sixfold-concentrated aliquot of medium, with and without protease inhibitors, was analyzed on the same blot as the 1X samples.

The calibration curve for digital volume analysis and a partial results table are embedded in Figure 4. Both 1X samples were spanned by calibration standards, which lead to a volumetric productivity estimate of 1.4 mg/mL. We conservatively estimated the experimental uncertainty for this semiquantitative western to be about a factor of three (from the 95% confidence interval of prediction of the standard curve for the 20 ng standard) and concluded that the addition of a protease inhibitor cocktail did appear to improve recovery of intact Protein R.

Slab-gel Versus On-chip Separations
At least three instrument manufacturers—Caliper Life Sciences (Lab Chip 90), Agilent Technologies (2100 Bioanalyzer and Protein 200 LabChip kit), and Bio-Rad Laboratories (Experion and Pro260 analysis kit)—market systems that use laser-induced fluorescence detection of dye-protein complexes and capillary microchannel electrophoresis calibrated to internal standards as a substitute for slab-gel PAGE. Clearly, such systems can provide data at high speed. For a company considering analyses by such a system, it would seem worthwhile to look closely at what is to be determined about the analytes and to compare costs and benefits of collecting the information on a chip system to the value of collecting it by classical technology.

• Apparent Protein Formula Weight: This information can be confounded, comparably, by posttranslational modification whether it is obtained by capillary microchannel separation or slab-gel SDS-PAGE. Thus, both provide equally suspect, i.e., low value, information.

• Speed: The slowest chip system offers analysis of 10 samples in 30 minutes. A novice operator can easily analyze 28 samples in less than 4.5 hours, so the time required per sample is about 10 to 30 times smaller on chip systems.

• Sample Integrity: SDS-PAGE technology and chip systems are equally subject to the potential ill effects of joule heating on sample integrity. Although small size might be expected to promise good heat dissipation, the material from which "chips" are fabricated (polydimethylsiloxane) is known to be particularly problematic in this regard [Y. Zhang et al., J. Chromatogr. A, vol. 1057, pp. 247-251 (2004)]. Traditional slab-gel systems put responsibility for temperature control into the user's hands.

click the image to enlarge   Figure 4: Slab electrophoretograms are compatible with western analysis. Proteins were electrophoretically transferred to PVDF for colorimetric western detection using a primary antibody described in the text. The secondary antibody was BA1000 from Vector Labs and detection was by Vector ABC-AP + Vector Black colorimetric substrate. Unstained markers were used to produce a negatively-stained molecular weight ladder on this blot (lane 1).

•Sensitivity: SDS-PAGE and chip systems both depend on a dye-protein interaction. A single dye is available for capillary microchannel separations. Whereas any of hundreds of available dyes could be used with SDS-PAGE, in reality, most investigators rely on Coomassie brilliant blue R-250.

• Versatility: Chip systems introduce sample into the separating zone by electrokinetic injection. This means they load equivalent conductance at each analysis cycle. Thus, if the ionic strength of a sample were lower than the ionic strength of "pure sample buffer" a larger volume of sample would be injected while the injected volume of a sample at ionic strength higher than pure sample buffer would be correspondingly decreased. This means an "equal mass loading" analysis such as shown in Figure 3, panel A could not be made on a chip without the sacrifice of processing time (all samples could be dialyzed to equivalent conductivity) or mass balance. The chip systems do not support coupled methods (SDS-PAGE/ Edman degradation, SDS-PAGE western blotting).

In summary, it is clear that capillary microchannel technology enables a researcher to collect low-value information more than 10 times faster than the same data could be collected by slab-gel SDS-PAGE. Although slab-gel SDS-PAGE appears to require "excess effort" to collect similarly low-value information, it provides a well-characterized parallel sample analysis platform in which the effects of joule heating can be readily controlled to the extent deemed neces sary. Finally, unlike chip systems, slab-gel PAGE is sufficiently flexible that its utility is not restricted to collecting low-value information.

About the Author
Louise Garone is a senior principal investigator at Serono Research Institute, Rockland, Mass.

This article was published in G & P magazine: Vol. 6, No. 2, March, 2006, pp. 26-29.