CAPILLARY SIEVING ELECTROPHORESIS WITH A CATIONIC SURFACTANT FOR SIZE

SEPARATION OF PROTEINS

FIELD OF THE INVENTION

The present invention generally relates to capillary electrophoresis of proteins in sieving media, particularly in the presence of one or more cationic surfactants that form charged complexes with the proteins and so allow their size separation and molecular-weight determination. Specifically, the invention is directed at capillary sieving electrophoresis of proteins in the presence of cationic detergents at low pH.

BACKGROUND OF THE INVENTION

SDS polyacrylamide gel electrophoresis (PAGE) (Dunker and Rueckert, 1969; Shapiro and Maizel, 1969; Shapiro, Vinuela, and Maizel, 1967; Weber and Osborn, 1969) has become a popular method (Kresge, Simoni, and Hill, 2006) for size separation of proteins. It has been based on formation of SDS-protein complexes when an equal amount of SDS binds to the proteins, independent of ionic strength (Reynolds and Tanford , 1970). Nevertheless, some proteins exhibit an anomalous migration in SDS PAGE (Shapiro, Vinuela, and Maizel, 1967; Williams and Gratzer, 1971). It has been found that pepsin, papain, and glucose oxidase do not bind measurable amount of SDS (Nelson, 1971). The anomalous migration in SDS PAGE may be an inherent property of acidic proteins as esterifϊcation of carboxyl groups normalize their migration (Williams and Gratzer, 1971).

SDS PAGE was adapted into the capillary format while employing crosslinked polyacrylamide gel as the sieving matrix (Cohen and Karger, 1987; Dolnik, Cobb, and Novotny, 1991). Later, the crosslinked gel was replaced with polymer solutions (Craig, Polakowski, Arriaga, Wong, Ahmadzadeh, Stathakis, and Dovichi, 1998; Ganzler, Greve, Cohen, Karger, Guttman, and Cooke, 1992; Guttman, 1995; Guttman, Horvath, and Cooke, 1993; Guttman, Shieh, Lindahl, and Cooke, 1994; Hu, Ye, Surh, Clark, and Dovichi, 2002; Hunt and Nashabeh, 1999; Karim, Janson, and Takagi, 1994; Nakatani, Shibukawa, and Nakagawa, 1994; Nakatani, Shibukawa, and Nakagawa, 1996; Salas-Solano, Tomlinson, Du., Parker, Strahan, and Ma, 2006;

Page l of 18

Tsuji, 1994; Werner, Demorest, Stevens, and Wiktorowicz, 1993). The method has been also modified for size separation of proteins on microchip (Bousse, Mouradian, Minalla, Yee, Williams, and Dubrow, 2001; Yao, Anex, Caldwell, Arnold, Smith, and Schultz, 1999).

Similarly, as in SDS PAGE, some proteins migrated anomalously in SDS capillary sieving electrophoresis (SDS CSE). The equal SDS biding to proteins was demonstrated to be an approximation only (Guttman and Nolan, 1994; Guttman, Nolan, and Cooke, 1993a; Werner, Demorest, and Wiktorowicz, 1993). A Ferguson plot of protein mobility vs. sieving matrix concentration showed different mobilities at the extrapolated zero sieving matrix concentration (Cohen and Karger, 1987; Guttman, Nolan, and Cooke, 1993). Post-translation modifications of proteins such as glycosylation, phosphorylation, and sulfonation were responsible for deviations from the idealized protein migration (Guttman, 1996; Werner, Demorest, and Wiktorowicz, 1993). Molecular weights of 65 proteins were measured by SDS-PAGE and SDS CSE, and compared to the values obtained by other techniques. For some proteins, significant deviations from the literature data were observed (Guttman and Nolan, 1994). A Ferguson method has been proposed to improve the accuracy of the molecular-weight determination. Although it improved the molecular-weight accuracy significantly, for some proteins, the method provided only mediocre results. E.g., for IgG L-chain, where SDS CSE showed a 87% error, the Ferguson method still exhibited a 17% deviation (Guttman, Shieh, Lindahl, and Cooke, 1994). Deviations in protein migration in SDS PAGE and SDS CSE are so frequent they represent more an inherent property of the method than an anomaly.

Shortly after the development of SDS PAGE, a method to separate proteins by PAGE in the presence of cationic surfactants was described (Williams and Gratzer, 1971). Nevertheless, observations on complexes of proteins with cationic surfactants have been published, predicting the electrophoresis in the presence of cationic detergents would not be suitable for determination of molecular weights (Nozaki, Reynolds, and Tanford, 1974). Later, cetylpyridinium chloride was introduced to separate proteins by PAGE at low pH (Schick, 1975). Cetyltrimethylammonium bromide (CTAB) was used more frequently at various pH values to separate proteins by PAGE: at pH 8.2 (Akins, Levin, and Tuan, 1992; Akins and Tuan, 1994), pH 6 (Akin, Shapira, and Kinkade Jr., 1985), pH 7 (Eley, Burns, Kannapell, and Campbell, 1979), and pH 4.6

(Panyim, Thitipongpanich, and Supatimusro, 1977). Also various protocols were developed for sample preparation, including a protocol without any heating of the sample (Akins, Levin, and Tuan, 1992).

While CTAB has been used in capillary electrophoresis as a dynamic coating for electroosmotic flow reversal (Chiari, Damin, and Reijenga, 1998; Corradini, 1997; Ding and Fritz, 1997; Reijenga, Aben, Verheggen, and Everaerts, 1983; Tsuda, 1987), no cationic surfactants have been combined with a sieving matrix to separate proteins by CSE.

SUMMARY OF THE INVENTION

The present invention is suitable for a fast, quantitative, and highly reproducible size separation of proteins by means of capillary sieving electrophoresis. Disclosed herein are the composition of the sample denaturant, the composition of the sieving matrix, the method of proper sample preparation, and the method of performing capillary sieving electrophoresis in the presence of a cationic surfactant. In a preferred embodiment, the sieving matrix contains a buffer that keeps pH of the sieving matrix acidic (pH < 5), a hydrophilic sieving polymer, and about 0.01 mM - about 100 mM cationic surfactant. The sample denaturant contains about 0.01 mM - about 100 mM cationic surfactant, about 0 mM - about 100 mM KCl or another salt composed of high-mobility ions, and dithiotreitol or 2-mercaptoethanol as a reducing agent.

DETAILED DESCRIPTION OF THE INVENTION

Even in today's proteomic era, proteins are mainly size-separated by a slow, laborious, and cumbersome method of SDS electrophoresis on slab gels. A trial transfer of the method into a fast, automatable, and easy capillary format has not been fully successful: Ionized silanol groups in the fused silica wall cause electroosmotic flow that leads to insufficient reproducibility of qualitative and quantitative analysis as well as to a mediocre separation efficiency. The electroosmotic flow can be suppressed by a neutral hydrophilic coating. In case of SDS capillary electrophoresis, the capillary coating does not help significantly, as SDS binds to the coating and generates a secondary electroosmotic flow.

The solution of this problem is a capillary sieving electrophoresis with a cationic surfactant performed at pH below 5, where ionization of silanol groups is suppressed. However, SDS does not strongly bind proteins at low pH and the proteins have to be complexed with a cationic surfactant.

Suppressed ionization of silanol groups results in significantly diminished adsorption of ionic surfactants on the capillary wall. That translates into a suppression of both the zeta potential and the electroosmotic flow. The practically eliminated electroosmotic flow, which is otherwise deleterious to the reproducibility of capillary electrophoresis and the separation efficiency of analytes, means improved separation efficiency and higher peak capacity of the protein separation. It also results in superior reproducibility of migration times, improving accuracy of qualitative analysis. In capillary electrophoresis, analytes do not travel through the detection cell with the same velocity as in chromatography, but with a velocity dependent on their apparent electrophoretic mobility. Because of that, the elimination of the electroosmotic flow also means an improved accuracy of the quantitative analysis of proteins.

The problem of size separation of proteins in a capillary format is the insufficient reproducibility of qualitative and quantitative analysis as well as the mediocre separation efficiency. The solution to this problem is capillary sieving electrophoresis of proteins with a cationic surfactant at low pH, when ionization of silanol groups is suppressed.

The sieving matrix for this method has to contain a cationic surfactant, a sieving polymer, an acidic buffer, and additives. It is essential the sieving matrix has an acidic pH. Silanol groups of fused silica capillary are not ionized at low pH and as a result, electroosmotic flow, which otherwise deteriorates electrophoretic separation, is suppressed. So are suppressed the adsorption of cationic surfactants on the capillary wall and the reversed electroosmotic flow.

The pH of the sieving matrix requires fine optimization: Below pH 3, the high-mobility H ⁺ ion contributes significantly to the conductivity of the sieving matrix. This may lead to an elevated generation of Joule heat and overheating of the capillary. Above pH 5, the silanol ionization is not negligible and electroosmotic flow becomes a serious issue. Keeping the pH of the sieving matrix at about pH 4 is the best compromise. One possibility is to use a free weak acid, e.g., acetic acid, as the electrolyte. Another option is using low mobility co-ion, e.g., Tris, with a buffering counter-ion, e. g., glutamic acid. pH can be also kept at a proper

level with a buffering co-ion, e.g., β-alanine or γ-aminobutyric acid (GABA) and a low-mobility counter anion. Probably the most attractive buffering option is a combination of a weak base and a weak acid, having close pK's, e.g., GABA (pK 4.0) and glutamic acid (pK 4.2), in an equimolar mixture.

The sieving polymer should exhibit low viscosity to allow a fast replacement of the sieving matrix in the capillary by a low pressure of about 1 bar. Hydrophilic polymers such as linear low-molecular-mass polyacrylamide or low molecular- weight poly(ethylene oxide) (PEO) are suitable sieving polymers. Poly(vinyl pyrrolidone), which absorbs UV light is not suitable for CSE with UV detection but may be used for CSE with laser induced fluorescence detection.

Not all proteins form complexes with cationic surfactants with the same ease. The cationic surfactant used in the sieving matrix should exhibit a sufficient solubility in water and, simultaneously, it should bind proteins. A cetyltrimethylammonium ion having a proper counter-ion does not precipitate at room temperature at a concentration below 1.1% and also binds proteins forming a complex with a positive charge.

The sample denaturant should contain a cationic surfactant, which may but need not be identical with the cationic surfactant in the sieving matrix, a reduction agent, which can disrupt disulfide bridges (β- mercaptoethanol or dithiotreitol, DTT), and an electrolyte with a high-mobility cation that allows a transient isotachophoresis during the injection and helps to focus the analytes into sharp bands.

High-mobility ions present in the sample denaturant have another role during the injection steps as they allow quantitative analysis not only with the pressure injection but also with the electrokinetic injection. As the sieving matrix containing a polymer solution exhibits an increased viscosity, the accuracy of the pressure injection may be compromised and the electrokinetic injection may be preferred to quantitate proteins. However, if EOF is suppressed in the separation capillary, the amount of analytes injected electrokinetically is not necessarily proportional to its concentration in the sample. A non-linear calibration curve is obtained for analytes in low-conductivity samples because the injected amount of the analytes strongly depends on the analyte mobility. Nevertheless, if a high-conductivity salt is added to the sample denaturant, the analytes do not contribute significantly to the overall specific conductivity of the sample and the calibration curves become linear.

Separation efficiency is closely watched in separation methods as it influences the number of analytes that can be analyzed. The separation efficiency depends on many factors. Some of them can be successfully influenced by the experimental conditions. Eddy migration as a result of inhomogeneous residual EOF is one of the most deleterious effects on CE separation. Typically, it is suppressed by employing a capillary coating, which eliminates the overall electroosmotic flow. In CSE with a cationic surfactant, a mild acidic pH suppresses the ionization of the silanol groups in the capillary wall and thus the adsorption of cationic surfactants on the wall. The suppression of EOF then results in exceptional separation efficiency and unmatched run-to run reproducibility.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows separation of model protein mixture. BGE: 100 mM β-alanine (BALA), 100 mM glutamic acid (GLU), 0.1% cetyldimethylethylammonium bromide (CDMEAB), 16 g/L poly(ethylene oxide) (PEO, M ₁ 400k). Guarant™ capillary (Alcor BioSeparations, Palo Alto, CA. U.S.A.): total length= 335 mm, effective length= 250 mm, ID= 75 μm, OD= 360 μm. Voltage: +10 kV. Electrokinetic injection: 6 s at +8 kV. Sample: 1 g/L proteins in 10 g/L CDMEAB, 100 mM KCl, 10 g/L dithiotreitol (DTT) heated 5 min at 95°C (lysozyme), 2 min at 95°C (all other proteins).

FIG. 2 is the plot of protein mobility vs. their logarithmic molecular weight calculated from the electropherogram in Fig. 1.

FIG. 3 displays the separation of BSA oligomers. Pressure injection: 10 s at 50 mbar. Sample: 10 g/L BSA in water. All other experimental conditions were same as in Fig. 1.

FIG. 4 presents the plot of the mobility vs. logarithmic molecular weight for BSA oligomers as calculated from the electropherogram in Fig. 3.

FIG. 5 shows 10 overlaid electropherograms of model proteins from 10 consecutive runs. BGE: 100 mM GABA, 100 mM GIu, 25 mM CTAB, 20 g/L PEO (200k). Capillary: bare capillary, l(total)= 335 mm, l(effective)= 250 mm, ID= 75 μm, OD= 360 μm. Voltage: +10 kV. Electrokinetic injection: 3 s at +3 kV.

Sample: 0.83 g/L each protein in 30 mM CTAB, 60 mM DTT, 5 min incubated at 100 ⁰ C (lysozyme), 2 min at 100 ⁰ C (all other proteins).

FIG. 6 depicts calibration curves of model proteins with electrokinetic injection 30 s at +10 kV. Sample denaturant: 10 g/L CDMEAB, 10 g/L DTT, 100 mM KCl, 5 min incubated at 95°C (lysozyme), 2 min at 95°C (all other proteins). ♦- lysozyme, x- BSA (monomer), ■- β-lactoglobulin, δ- ovalbumin.

BEST MODE FOR CARRYING OUT THE INVENTION

EXAMPLE 1

Preparation and composition of the sieving matrix

The sieving matrix for CSE with a cationic surfactant has been formulated to contain 0.1 % cetyldimethylethylammonium bromide (CDMEAB), 100 mM β-alanine, 100 mM glutamic acid, andlό g/L PEO (M _r 400,000). Alternative compositions of the sieving matrix are a) 0.2 % cetyldimethylethylammonium bromide (CDMEAB), 100 mM γ-aminobutyric acid, 100 mM glutamic acid, and20 g/L PEO (M ₁ 200,000; b) 0.2 % cetyldimethylethylammonium bromide (CDMEAB), 100 mM β- alanine, 100 mM 2-hydroxyisobutyric acid, and20 g/L PEO (M ₁ 200,000), c) 25 mM cetyltrimethylammonium bromide (CTAB), 100 mM γ-aminobutyric acid, 100 mM glutamic acid, and20 g/L PEO (M _r 200,000; b).

The composition of the sample electrolyte for quantitative analysis has been formulated to contain 30 mM CTAB, 100 mM KCl, and 60 mM DTT. Alternatively, a sample electrolyte without KCl has been proposed for qualitative analysis and to measure protein mobilities that contains 30 mM CTAB, and 60 mM DTT

EXAMPLE 2 Sample preparation

The composition of the sample denaturant has been formulated to contain 1% cetyldimethylethylammonium bromide (CDMEAB), 100 mM KCl, and 10 g/L dithiotreitol. Alternative sample denaturants contain a) 30

niM CTAB, 100 niM KCl, and 60 mM DTT; b) 1% CDMEAB, 100 mM KCl, and 10 g/L β- mercaptoethanol. Sample denaturants of the same composition without KCl have been also proposed for qualitative analysis and to measure actual protein mobilities, e.g., 1% CDMEAB and 10 g./L DTT. During the sample preparation, proteins are dissolved in the sample denaturant and the protein solution is incubated at 95°C for 2 min. Some proteins, e.g., lysozyme, are resistant to the thermal denaturation with cationic surfactants and an extended incubation at high temperature is necessary (5 min in case of lysozyme). Proteins such as BSA, on the other hand, do not require any denaturation at all prior to the CSE separation.

EXAMPLE 3

The method of capillary sieving electrophoresis

Capillary sieving electrophoresis with a cationic surfactant is performed in a fused silica capillary, 75 μm ID, 360 μm OD, 335 mm total length, 250 mm effective length. After a run, the capillary is flushed with 100 mM citric acid at pressure of 930 mbar for 7 min to remove the gel from the previous run from the capillary. In the next step, the capillary is prepared for the next run: the fresh sieving matrix is pumped into the capillary with a pressure of 930 mbar for 3 min. The sample is injected either electrokinetically or by pressure. The amount of the injected sample depends on the protein concentration. If the sample is prepared with the sample denaturant containing 1% CDMEAB, 100 mM KCl, and 10 g/L dithiotreitol, the sample containing 0.1 - 1 g/L proteins is typically injected for 8 s at 6 kV. The separation is performed at +10 kV and takes typically 10- 12 minutes. The separation of a model protein mixture is shown in Fig. 1. The electrophoretic mobility of the proteins can be plotted against the logarithmic molecular weight (Fig. 2). A quadratic equation is preferred for interpolation of this relationship.

For the separation of native BSA oligomers, the sample is to be injected by pressure injection, typically for 10 s at 50 mbar. CSE of a high-concentration BSA sample takes less than 12 min. and reveals up to nine peaks (Fig. 3). While the BSA monomer is heavily overloaded, the BSA oligomers from dimer to nonamer are clearly discernable although octamer and nonamer as shoulders only. The electrophoretic mobility of the BSA oligomers can be plotted against the logarithmic molecular weight (Fig. 4). A straight-line interpolation can serve as a calibration curve for molecular- weight determination of large proteins.

CSE with a cationic surfactant provides narrow peaks of high separation efficiency. Table 1 summarizes the data on separation efficiency of a series of 7 runs. The calculation of the separation efficiency from a half- height peak width assumes ideal Gaussian peaks and provides results rather lower than the calculation based on an unrevealed algorithm used in ChemStation software.

Table 1 Separation efficiency of protein peaks (n=7)

- calculated from half-height peak width b- - obtained directly from the ChemStation software (Agilent)

EXAMPLE 4

Reproducibility of migration times

Low pH of the sieving matrix minimizes electroosmotic flow and improves reproducibility of separation. 10 overlaid consecutive electropherograms of a model mixture containing 0.8 g/L of insulin B, lysozyme, β- lactoglobulin, α-chymotrypsinogen A, ovalbumin, and BSA are shown in Fig. 5. Run-to-run reproducibility of the migration times ranged from 0.14% to 0.25% and is summarized in Table 2.

Table 2 Run-to-run reproducibility of migration times (n= 10).

EXAMPLE 5

Quantitative analysis

Capillary sieving electrophoresis with a cationic surfactant allows quantitative analysis with electrokinetic injection. When proteins were denatured in 10 g/L CDMEAB, 100 rtiM KCl, and 10 g/L DTT and injected 30 s at +10 kV, the calibration lines for lysozyme, β-lactoglobulin, ovalbumin, and BSA were linear in the concentration range 0 - 1.0 g/L (Fig. 6). The squared correlation coefficient ranged from 0.99 for β- lactoglobulin to 0.998 for BSA.

Table 3 Reproducibility of the peak area for 0.2 g/L proteins injected 30 s at +10 kV.

REFERENCES CITED

U.S. Patent Documents:

1) 4,481,094 Stabilized polyacrylamide gels and system for SDS electrophoresis

2) 5,213,669 Capillary column containing a dynamically cross-linked composition and method of use

3) 5,275,708 Cetyltrimethylammonium bromide gel electrophoresis

4) 5,370,770 Capillary column containing removable separation gel composition and method of use

5) 20050161329 Multiplexed capillary electrophoresis systems

Other References:

1. Akin D. T., Shapira R., Kinkade Jr. J. M. The determination of molecular weights of biologically

active proteins by cetyltrimethylammonium bromide-polyacrylamide gel electrophoresis. Anal.

Biochem. 145, 170-176 (1985).

2. Akins R. E., Levin P. M., Tuan R. S. Cetyltrimethylammonium bromide discontinuous gel electrophoresis: M(r)- based separation of proteins with retention of enzymatic activity. Anal. Biochem. 202, 172-178 (1992).

3. Akins R. E., Tuan R. S. Separation of proteins using cetyltrimethylammonium bromide discontinuous gel electrophoresis. MoI. Biotech 1, 211-228 (1994).

4. Bousse L., Mouradian S., Minalla A., Yee H., Williams K., Dubrow R. Protein sizing on a microchip. Anal. Chem. 73, 1207-1212 (2001).

5. Chiari M., Damin F., Reijenga J. C. Characterization of poly(dimethylacryl amide) and the combination of poly(vinyl alcohol) and cetyltrimethylanlmonium bromide as dynamic electroosmotic flow suppression agents in capillary electrophoresis. J. Chromatogr. A 817, 15-23 (1998).

6. Cohen A. S., Karger B. L. High-performance sodium dodecyl sulfate polyacrylamide gel capillary electrophoresis of peptides and proteins. J. Chromatogr. 397, 409-417 (1987).

7. Corradini D. Buffer additives other than the surfactant sodium dodecyl sulfate for protein separations by capillary electrophoresis. J. Chromatogr. B 699, 221-256 (1997).

8. Craig D. B., Polakowski P. M., Arriaga E., Wong J. C. Y., Ahmadzadeh H., Stathakis C, Dovichi N. J. Sodium dodecyl sulfate-capillary electrophoresis of proteins in a sieving matrix utilizing two-spectral channel laser-induced fluorescence detection. Electrophoresis 19, 2175-2178 (1998).

9. Ding W. L., Fritz J. S. Separation of Basic-Proteins and Peptides by Capillary Electrophoresis Using a Cationic Surfactant. HRC-J. High Res. Chromatogr. 20, 575-580 (1997).

10. Dolnik V. , Cobb K. A., Novotny M. : Preparation of polyacrylamide gel-filled capillaries for capillary electrophoresis. J. Microcol. Sep. 3, 155-159 (1991).

11. Dunker A. K. , Rueckert R. R. Observations on molecular weight determinations on polyacrylamide gel. J. Biol. Chem. 244, 5074-5080 (1969).

12. Eley M. H., Burns P. C, Kannapell C. C, Campbell P. S. Cetyltrimethylammonium bromide polyacrylamide gel electrophoresis: estimation of protein subunit molecular weights using cationic detergents. Anal. Biochem. 92, 411-419 (1979).

13. Ganzler K., Greve K. S., Cohen A. S., Karger B. L., Guttman A., Cooke N. C. High-Performance Capillary Electrophoresis of SDS Protein Complexes Using UV-Transparent Polymer Networks. Anal. Chem. 64, 2665-2671 (1992).

14. Guttman A. On the separation mechanism of capillary sodium dodecyl sulfate-gel electrophoresis of proteins. Electrophoresis 16, 611-616 (1995).

15. Guttman A. Capillary sodium dodecylsulfate-gel electrophoresis of proteins. Electrophoresis 17, 1333- 1341 (1996).

16. Guttman A., Horvath J., Cooke N. Influence of Temperature on the Sieving Effect of Different Polymer Matrices in Capillary SDS Gel Electrophoresis of Proteins. Anal. Chem. 65, 199-203 (1993).

17. Guttman A., Nolan J. Comparison of the Separation of Proteins by Sodium Dodecyl Sulfate- Slab Gel Electrophoresis and Capillary Sodium Dodecyl Sulfate-Gel Electrophoresis. Anal. Biochem. 221, 285- 289 (1994).

18. Guttman A., Nolan J. A., Cooke N. Capillary sodium dodecyl sulfate gel electrophoresis of proteins. J. Chromatogr. 632, 171-175 (199).

19. Guttman A., Shieh P., Lindahl J., Cooke N. Capillary sodium dodecyl sulfate gel electrophoresis of proteins .2. On the Ferguson method in polyethylene oxide gels. Journal of Chromatography A 676, 227-231 (1994).

20. Hu S., Ye Y. L., Surh G., Clark J. L, Dovichi N. J. Analysis of Proteins by Capillary SDS-DaIt Electrophoresis With Laser-Induced Fluorescence Detection. LC-GC Europe 15, 166-+ (2002).

21. Hunt G., Nashabeh W. Capillary electrophoresis sodium dodecyl sulfate nongel sieving analysis of a therapeutic recombinant monoclonal antibody: A biotechnology perspective. Anal. Chem. 71, 2390- 2397 (1999).

22. Karim M. R., Janson J. C, Takagi T. Size-dependent separation of proteins in the presence of sodium dodecyl sulfate and dextran in capillary electrophoresis: Effect of molecular weight of dextran. Electrophoresis 15, 1531-1534 (1994).

23. Kresge N., Simoni R. D., Hill R. L. SDS-PAGE to Determine the Molecular Weight of Proteins: the Work of Klaus Weber and Mary Osborn . J. Biol. Chem. 281 , el 9- e21 (2006).

24. Nakatani M., Shibukawa A., Nakagawa T. High-Performance Capillary Electrophoresis of SDS- Proteins Using Pullulan Solution as Separation Matrix. J. Chromatogr. A 672, 213-218 (1994).

25. Nakatani M., Shibukawa A., Nakagawa T. Effect of temperature and viscosity of sieving medium on electrophoretic behavior of sodium dodecyl sulfate-proteins on capillary electrophoresis in presence of pullulan. Electrophoresis 17, 1210-1213 (1996).

26. Nelson C. A. The binding of detergents to proteins. J. Biol. Chem. 246, 3895-3901 (1971).

27. Nozaki Y., Reynolds J. A., Tanford C. The Interaction of a Cationic Detergent with Bovine Serum Albumin and Other Proteins. J. Biol. Chem. 249, 4452-4459 (1974).

28. Panyim S., Thitipongpanich R., Supatimusro D. A simplified gel electrophoretic system and its validity for molecular weight determinations of protein cetyltrimethylammonium complexes. Anal. Biochem. 81, 320-327 (1977).

29. Reijenga J. C, Aben G. V. A., Verheggen T. H. A. M., Everaerts F. M. Effect of electroosmosis on detection in isotachophoresis.260, 241-254 (1983).

30. Reynolds J. A., Tanford C. Binding of Dodecyl Sulfate to Proteins at High Binding Ratios. Possible Implications for the State of Proteins in Biological Membranes . Proc. Natl. Acad. Sci. U.S.A. 66, 1002-1007 (1970).

31. Salas-Solano O., Tomlinson B., Du. S., Parker M., Strahan A., Ma S. Optimization and validation of a quantitative capillary electrophoresis sodium dodecyl sulfate method for quality control and stability monitoring of monoclonal antibodies. Anal. Chem. 78, 6583-6594 (2006).

32. Schick M. Influence of a cationic detergent on electrophoresis in polyacrylamide gel. Anal. Biochem.

63, 345-349 (1975).

33. Shapiro A. L., Maizel J. V. Molecular weight estimation of polypeptides by SDS-polyacrylamide gel electrophoresis: further data concerning resolving power and general considerations. Anal. Biochem. 29, 55-514 (1969).

34. Shapiro A. L., Vinuela E., Maizel J. V. Molecular weight estimation of polypeptide chains by electrophoresis in SDS-polyacrylamide gels. Biochem Biophys Res Commun. 28, 815-820 (1967).

35. Tsuda T. Modification of electroosmotic flow with cetyltrimethylammonium bromide in capillary zone electrophoresis. HRC CC, J. High Resolut. Chromatogr. Chromatogr. Commun., 10(11), 622-624 (1987).

36. Tsuji K. Sodium dodecyl sulfate polyacrylamide gel- and replaceable polymer-filled capillary electrophoresis for molecular mass determination of proteins of pharmaceutical interest. J. Chromatogr. B 662, 291-299 ( 1994).

37. Weber K., Osborn M. The Reliability of Molecular Weight Determinations by Dodecyl Sulfate - Polyacrylamide Gel Electrophoresis. J. Biol. Chem. 242, 4406-4412 ( 1969).

38. Werner W. E., Demorest D. M., Stevens J., Wiktorowicz J. E. Size-Dependent Separation of Proteins Denatured in SDS by Capillary Electrophoresis Using a Replaceable Sieving Matrix. Anal. Biochem. 212, 253-258 (1993).

39. Werner W. E. , Demorest D. M., Wiktorowicz J. E. Automated Ferguson analysis of glycoproteins by capillary electrophoresis using a replaceable sieving matrix. Electrophoresis 14, 759-763 (1993).

40. Williams J. G., Gratzer W. B. Limitatons of the detergent polyacrylamide gel elexctrophoresis method for molecular weight determination of proteins. J. Chromatogr. 57, 121-125 (1971).

41.Yao S., Anex D. S., Caldwell W. B., Arnold D. W., Smith K. B., Schultz P. G. SDS capillary gel electrophoresis of proteins in microfabricated channels. Proc. Natl. Acad. Sci. U.S.A. 96, 5372-5377 (1999).