Stable Capillary Coating with Successive Multiple Ionic Polymer Layers

Anal. Chem. 1998, 70, 2254-2260

Stable Capillary Coating with Successive Multiple Ionic Polymer Layers Hiroyuki Katayama,* Yasushi Ishihama, and Naoki Asakawa

Department of Analytical Chemistry, Analytical Research Laboratories, Eisai Co., Ltd., 5-1-3 Tokodai, Tsukuba, Ibaraki 300-2635, Japan

A stable modification of the inner wall of a fused silica capillary was established by a simple coating procedure, successive multiple ionic-polymer layer (SMIL) coating. An anionic polymer was tightly fixed to the capillary wall by the SMIL coating, in which a cationic polymer was sandwiched between the anionic polymer and the uncoated fused silica capillary by noncovalent bonding. The SMIL-coated capillary showed a long lifetime. The endurance of the SMIL-coated capillary was more than 100 runs, and it was also tolerant to organic solvents, 1 M NaOH, and a surfactant. The coating efficiency did not depend on capillary sources, and the relative standard deviation of capillary-to-capillary reproducibility was less than 1%. In this study, dextran sulfate (DS) was used as the anionic polymer, and Polybrene was used as the cationic polymer for SMIL modification. The DS-modified capillary (SMIL-DS capillary) exhibited a pH-independent electroosmotic flow (EOF) from anode to cathode in the pH range of 2-11. The SMIL-DS capillary showed good performance for acidic protein analyses under physiological conditions (pH 7.4). Also, the presence of EOF under acidic conditions permitted new applications. Simultaneous separations of cationic, anionic, and neutral amino acids were achieved by capillary zone electrophoresis, and separations of cresol isomers were achieved by micellar electrokinetic chromatography under the acidic conditions. The SMIL-DS capillary was also useful for fast and precise determination of the pKa of acidic functional groups.

by covalent bonding. Procedure d can be further divided into two methods, fixing a hydrophilic layer to the capillary wall by covalent bonding18-25 or fixing a stable, cross-linked hydrophilic layer with another stable layer between the capillary wall and the hydrophilic layer, such as poly(vinylmethylsiloxanediol),26 epoxy resin,27,28 or highly cross-linked poly(styrene-divinylbenzene).29 Procedures a and b were not suitable for the native protein analyses, because choosing the extreme pH condition or adding modifier in the electrolytes may denature the proteins. However, procedures c and d avoid this problem because analyses can be done under moderate conditions. Therefore, coating the capillary permanently is the better choice from the standpoint of the natural protein analyses. The advantage of capillary modification based on covalent bonding is its moderate lifetime, especially when cross-linked double-layer coating was performed. However, it has disadvantages of complicated coating procedure and having variance between the capillaries because several steps of chemical reaction

* To whom correspondence should be addressed. Fax: (+81-298) 47-2037. Tel.: (+81-298) 47-5763. E-mail: [email protected]. (1) McCormic, R. M. Anal. Chem. 1988, 60, 2322-2328. (2) Green, J. S.; Jorgenson, J. W. J. Chromatogr. 1989, 478, 63-70.

(3) Cifuents, A.; Poppe, H.; Kraak, J. C.; Erim, F. B. J. Chromatogr. B 1996, 681, 21-27. (4) Muijselaar, W. G. H. M.; Bruijn, C. H. M. M.; Everaerts, F. M. J. Chromatogr. 1992, 605, 115-123. (5) Swedberg, S. A. J. Chromatogr. 1990, 503, 449-452. (6) Towns, J. K.; Regnier, F. E. Anal. Chem. 1991, 63, 1126-1132. (7) Strege, M. A.; Lagu, A. L. Anal. Biochem. 1993, 210, 402-410. (8) Yao, Y. J.; Li, S. F. Y. J. Chromatogr. A 1994, 663, 97-104. (9) Hult, E. L.; Emmer, Å.; Roeraade, J. J. Chromatogr. A 1997, 757, 255-262. (10) Preisler, J.; Yeung, E. S. Anal. Chem. 1996, 68, 2885-2889. (11) Iki, N.; Yeung, E. S. J. Chromatogr. A 1996, 731, 273-282. (12) Towns, J. K.; Regnier, F. E. J. Chromatogr. 1990, 516, 69-78. (13) Erim, F. B.; Cifuentes, A.; Poppe, H.; Kraak, J. C. J. Chromatogr. A 1995, 708, 356-361. (14) Chiu, R. W.; Jimenez, J. C.; Monning, C. A. Anal. Chim. Acta 1995, 307, 193-201. (15) Liu, Q.; Lin, F.; Hartwick, R. A. J. Chromatogr. Sci. 1997, 36, 126-130. (16) Cohen, N.; Grushka, E. J. Capillary Electrophor. 1994, 1, 112-115. (17) Morand, M.; Blass, D.; Kenndler. J. Chromatogr. B 1997, 691, 192-196. (18) Hjerten, S.; Kiessling-Johansson, M. J. Chromatogr. 1991, 550, 811-822. (19) Hjerten, S. J. Chromatogr. 1985, 347, 191-198. (20) Cobb, K. A.; Dolnik, V.; Novotony, M. Anal. Chem. 1990, 62, 2478-2483. (21) Nashabeh, W.; Rassi, Z. E. J. Chromatogr. 1991, 559, 367-383. (22) Swedberg, S. A. Anal. Biochem. 1990, 185, 51-56. (23) Chiari, M.; Dell’Orto, N.; Gelain, A. Anal. Chem. 1996, 68, 2731-2736. (24) Bruin, G. J. M.; Huisden, R.; Kraak, J. C.; Poppe, H. J. Chromatogr. 1989, 480, 339-349. (25) Bruin, G. J. M.; Chang, J. P.; Kuhlman, R. H.; Zegers, K.; Kraak. J. C.; Poppe, H. J. Chromatogr. 1989, 471, 429-436. (26) Schmalzing, D.; Piggee, C. A.; Foret, F.; Carrilho, E.; Karger, B. L. J. Chromatogr. A 1993, 652, 149-159. (27) Liu, Y.; Fu, R.; Gu, J. J. Chromatogr. A 1996, 723, 157-167. (28) Ren, X.; Shen, Y.; Lee, M. J. Chromatogr. A 1996, 741, 115-122. (29) Haung, X.; Horvath, C. J. Chromatogr. A 1997, 788, 155-164.

2254 Analytical Chemistry, Vol. 70, No. 11, June 1, 1998

S0003-2700(97)00875-5 CCC: $15.00

Recently, capillary electrophoresis (CE) has become a powerful tool to analyze biopolymers. However, in the protein analysis by CE, significant tailing or irreversible adsorption of the proteins on the inner wall of a fused silica capillary was observed, because a Coulombic and/or a hydrophobic interaction exists between the proteins and the inner wall of the capillary. To overcome these problems, several approaches have been proposed so far: (a) to control the pH or ionic concentration of the background electrolytes;1,2 (b) to modify the capillary by dynamic coating;3-9 (c) to modify the capillary permanently by physical adsorption;10-17 or (d) to modify the capillary permanently

© 1998 American Chemical Society Published on Web 04/24/1998

were essential for the capillary modification. On the other hand, modification by noncovalent bonding, such as physical adsorption, offers advantages in its simple procedure and good reproducibilities, while it has disadvantages of limited pH range and short lifetime. It would be an ideal capillary modification if the stable coating could be achieved in a simple procedure. In this study, we developed a novel permanent coating procedure, successive multiple ionic-polymer layer (SMIL) coating. We immobilized the negatively charged polymer dextran sulfate (DS) tightly to the inner surface of the capillary by sandwiching cationic polymer Polybrene (PB) between the capillary wall and the DS without covalent bonding or cross-linking. The stable coating was obtained due to the double layer of the ionic polymer. The simplicity of SMIL coating is almost as the same as that of conventional physical adsorption modification, because it could be achieved by just rinsing the oppositely charged ionic polymer successively at room temperature. Therefore, it could be widely applied to CE analyses. EXPERIMENTAL SECTION Reagents. Polybrene (PB) (Aldrich, Milwaukee, WI) and dextran sulfate (DS), MWav ) 500 000 (Sigma, St. Louis, MO) were used as coating reagents. Protein markers, R-lactalbumin from bovine milk, β-lactoglobulin A from bovine milk, and β-lactoglobulin B from bovine milk were from Sigma. Detergent SCAT 20-X was from Dai-ichi Kogyo Seiyaku (Kyoto, Japan). Formamide and urea were from Pharmacia Biotech (Uppsala, Sweden). Quinine hydrochloride was from Wako (Osaka, Japan). Bovine serum albumin (BSA), pepsin, and trypsin inhibitor were from Sigma. DHFR was kindly donated by Dr. Iwakura (National Institute of the Bioscience and Human Technology). Amino acids were from Sigma. o-, m-, and p-cresol were from Wako. Other reagents were analytical grade. Apparatus and Conditions. CE was performed with a Beckman P/ACE 2100 instrument (Beckman Instruments, Fullerton, CA). The uncoated fused silica capillaries of 75 and 50 µm i.d., 27-67 cm, were from GL Sciences (Tokyo, Japan), Otsuka Electronics (Osaka, Japan), Beckman, J&W Scientific (Folsom, CA) and Hewlett-Packard (Palo Alto, CA). The poly(ethylene glycol)-modified capillary (PEG capillary) (µSIL DB-WAX capillary, J&W Scientific), linear polyacrylamide-modified capillary (LPA capillary) (BioCAP LPA-coated capillary, Bio-Rad, Hercules, CA), and poly(vinyl alchol)-modified capillary (PVA capillary) (HewlettPackard, Waldbronn, Germany) of 50 µm i.d., 27 cm were used for the neutral capillaries. Capillaries were thermostated at 25 °C during both the coating and analyses. Sample injections were performed by pressure (0.5 psi; 1 psi ) 6894.76 Pa) for 2.0 s. All rinsing steps for capillary coating were performed by using the rinse function of the Beckman CE system. The buffers used for analyses were H3PO4-NaHPO4 (I ) 0.05) in pH 2-3, CH3COOHCH3COONa (I ) 0.05) in pH 4-5, NaHPO4-Na2HPO4 (I ) 0.05) in pH 6-7, and H3BO3-Na2B4O7 (I ) 0.05) in pH 8-11. Formamide was used as the electroosmotic flow (EOF) marker. Measurement of the UV absorbance of the coating solution was performed with a Hitachi U-3500 instrument (Hitachi, Tokyo, Japan). Procedure of SMIL Coating. The procedure of SMIL coating is shown in Figure 1. Step a shows the activation of silanol groups. The capillary was rinsed with 0.1% SCAT for 3 min and 1 M NaOH

Figure 1. SMIL coating procedure. (a) Activation of the silanol groups; (b) first layer coating; (c) second layer coating.

for 30 min. The capillary was then rinsed with H2O for 15 min and left for 30 min. Step b shows the first layer coating. After preconditioning (step a), the capillary was rinsed with 5% PB solution, in which PB was dissolved in water for 15 min and left for 15 min. Step c shows the second layer coating. The capillary was rinsed with the 3% DS solution in which DS was dissolved in water for 15 min and left for 30 min. All these procedures were performed at room temperature. RESULT AND DISCUSSION Evaluation of SMIL-DS Capillary. 1. Measurement of EOF. The electrophoretic mobility of EOF was expressed by the following equation:30

µEOF ) ( ζ)/(4π η) where  is the dielectric constant, ζ is the zeta potential of the capillary wall, and η is the viscosity of the solution. ζ would be reversibly changed during SMIL coating because ionic polymer PB was attached as a first layer and DS was attached as a second layer. Therefore, the EOF could be used to check the attachment of the SMIL coating. Figure 2 shows the EOF of uncoated, SMILPB, and SMIL-DS capillaries. The EOF of the uncoated capillary decreased below pH 6, and it was especially suppressed below pH 3. When cationic polymer PB was attached as a first layer (Figure 1b), the EOF was reversed from cathode to anode. Next, when anionic polymer DS was attached as a second layer (Figure 1c), the EOF was reversed again from anode to cathode. The velocity of the EOF was almost the same through pH 2-11. A (30) Hjerten, S. J. Chromatogr. Rev. 1997, 9, 122.

Analytical Chemistry, Vol. 70, No. 11, June 1, 1998

2255

Figure 2. EOF of uncoated (0), SMIL-PB (]), and SMIL-DS (x) capillaries (n ) 5). Conditions: detection, 214 nm; applied voltage, 7 kV; buffers, phosphate buffer at pH 2-3 (I ) 0.05), acetate buffer at pH 4-5 (I ) 0.05), phosphate buffer at pH 6-7 (I ) 0.05), and borate buffer at pH 8-11 (I ) 0.05); capillary, 75 µm i.d. × 27 cm (20 cm effective length).

pH-independent EOF was obtained because the inner surface of the SMIL-DS capillary had sulfonic groups. 2. Comparison of the Background. The background of the SMIL-DS capillary should be investigated because UV transparency could be changed accompanied by successive polymer attachment. UV spectra of the coating solutions are shown in Figure 3 a. The DS solution showed weak absorbance, while the PB solution showed strong absorbance below 220 nm. To estimate the absorbed substance, PB and DS were both analyzed by cathodic and anodic detection using a LPA capillary. The peak was detected from the PB solution when it was analyzed by 214nm anodic detection (data not shown). It is estimated that only bromide ion, which is the counterion of PB, showed strong UV absorbance. Next, as is shown in Figure 3b, comparison of the background between uncoated and SMIL-DS capillaries was investigated. Quinine was used as a reference, and pH 4.4 was selected because the apparent mobility of quinine was almost the same under this condition. A signal-to-noise ratio (S/N) was obtained from the average values of 10 replicate analyses. The S/N values of quinine at 200, 214, and 254 nm were compared, but a significant difference could not be obtained. This is probably because the bromide ion was flushed out of the capillary during the SMIL coating. These results showed that the background of SMIL-DS capillary was almost the same as that of an uncoated capillary. 3. Coating Efficiency. Coating efficiency might depend on the silanol groups of the fused silica capillary in the case of SMIL modification. Comparison of the coating efficiencies between the different capillary sources is shown in Figure 4. The average values of EOF were obtained from five replicate analyses at each pH. When the EOF values of the uncoated capillaries were measured, each capillary had a different pH profile (Figure 4a). This is because the source of fused silica and/or the manufacturing process of the capillaries differed. The EOF still varied when the PB was attached. Capillary brand E, which had the smallest EOF under acidic conditions, still had the smallest EOF when the PB was adsorbed (Figure 4b). This result suggests that the attachment of the PB layer was deeply affected by the difference 2256 Analytical Chemistry, Vol. 70, No. 11, June 1, 1998

Figure 3. Investigation of the background. (a) UV spectra. (b) Comparison of the background. Conditions: applied voltage, +7 kV; buffer, acetate buffer at pH 4.4 (I ) 0.05); S/N marker, quinine 0.005 mg mL-1; capillary, 75 µm i.d. × 27 cm (20 cm effective length).

in the fused silica. On the other hand, the variance decreased with the attachment of the DS (Figure 4c). The quantitative data of these results are shown in Tables 1 and 2. Table 1 shows the run-to-run reproducibilities during five replicate analyses at pH 4.0. The RSD of EOF of the uncoated and the SMIL-PB capillaries varied. The differences in the bare capillary wall still affect the reproducibility and velocity of the EOF of the SMIL-PB capillary, because the amount of attached PB may depend on the density of the silanol groups. However, once the DS layer was formed, the variance of the silanol groups and the PB layer could be negligible because the DS strongly interacts with the PB and perfectly covers the PB layer. Table 2 shows the capillary brand-to-capillary brand reproducibilities. The RSDs of the EOF obtained from the five different sources at each pH were compared. The RSD of the EOF improved with the PB modification and finally convergenced to less than 3% when the SMIL-DS capillaries were produced. These results showed that the DS could be uniformly attached by SMIL coating, even though the capillary sources were different. 4. Chemical Stabilities. To evaluate the stability of the SMIL-coated capillary under practical conditions, chemical stability should be investigated. Table 3 shows the chemical stability of SMIL-PB and SMIL-DS capillaries. EOF was first measured when the SMIL-PB and the SMIL-DS capillaries were produced. The

Figure 4. Comparison of the coating efficiencies between the different capillary sources (n ) 5). (a) Uncoated capillary; (b) SMILPB capillary; (c) SMIL-DS capillary. Conditions were as described in Figure 2. Table 1. Run-to-Run Reproducibilities at pH 4.0 (n ) 5)a RSD (%) brand

uncoated capillary

SMIL-PB

SMIL-DS

A B C D E

5.98 2.64 0.51 0.55 11.11

1.26 1.81 2.26 1.03 3.55

0.24 0.25 0.21 0.41 0.50

a

Conditions were as described in Figure 2.

Table 2. Capillary Brand-to-Capillary Brand Reproducibilites (n ) 5)a RSD (%) pH

uncoated capillary

SMIL-PB

SMIL-DS

3.0 4.0 5.0 6.0 7.0

64.44 13.77 15.64 9.19 1.58

3.00 6.01 3.84 4.64 3.05

2.48 2.77 2.27 2.61 2.93

a

Conditions were as described in Figure 2.

capillary was then rinsed with the solvent for 15 min, and the EOF was measured again. Each EOF was obtained from five replicate analyses. The coating stability was evaluated on the basis of the change in EOF, which was expressed as the degradation ratio defined in Table 3. If the EOF could not be detected within an hour, the coating was defined as degraded. The SMIL-PB capillary was unstable after 0.1 M HCl and CH3CN rinsing, because the degradation ratios were both more than 10%, and the coating was

detached after 1 M NaOH rinse. On the other hand, the DS layer of the SMIL-DS capillary was detached after 0.1 M HCl rinse, but it was stable to other solvents. The SMIL-DS capillary exhibits stronger chemical stability compared to that of the SMIL-PB capillary. This is because PB was used as an adhesion material for the SMIL coating, and once DS was attached and multiple layers were formed, short-term stability of the PB layer could be ignored. Next, the adsorption mechanism in the SMIL coating should be discussed. Both SMIL-PB and SMIL-DS capillaries were stable after 5 M urea rinse. If the major interaction of the SMIL coating was hydrogen bonding, it would be unstable in urea because urea causes disorder of the hydrogen-bonding networks.31 The DS layer of the SMIL-DS capillary was detached in 0.1 M HCl, although it was stable in 1 M NaOH; therefore, the main interaction of the SMIL coating would be ionic interaction. The stability of the SMIL-DS capillary under extremely alkaline conditions is a remarkable property because the other modified capillaries, whether the coating procedure was based on physical adsorption or covalent bonding, were mostly unstable in alkaline environment. From these results, it is possible to use the SMILDS capillary under practical conditions. 5. Endurance and Regeneration of SMIL Coating. One of the most important problems the noncovalent bonding must overcome is its short-term endurance. One-layer adsorption of a cationic polymer such as PB or polyethylenimine endures only 25 runs.32 The SMIL coating was intended to achieve strong endurance rather than the one-layer adsorption. Figure 5 shows the endurance and regeneration of the SMIL-DS coating. EOF, R-lactalbumin, β-lactoglobulin B, and β-lactoglobulin A were used as the test samples. The SMIL-DS capillary endured more than 100 runs. After 200 runs were performed, 60% of the EOF was still generated from anode to cathode at pH 2.8, where the EOF would not be generated if the coating was detached (data not shown). This strong endurance is due to the multiple attachment of the ionic polymer. The SMIL-DS capillary showed strong endurance as well as strong tolerance to the organic solvents and alkaline solvent. As soon as the coating was detached, regeneration of the coating was possible with 0.1 M HCl rinsing and SMIL coating. If the capillary was modified by covalent bonding, regeneration of the coating would be impossible when the capillary lifetime was finished. Therefore, strong endurance and easy regeneration of the coating by the SMIL coating assured long coating lifetime. 6. Coating Reproducibilitiy. Coating reproducibilitiy was investigated next. Five SMIL-DS capillaries were prepared from capillary brand E by SMIL coating, and the coating reproducibilitiy was evaluated. EOF, R-lactalbumin, β-lactoglobulin B, and β-lactoglobulin A were used as the test samples. Table 4 shows the reproducibility of the capillary-to-capillary coating. The relative standard deviation (RSD) of the EOF and the test samples was less than 1%, and excellent reproducibility was obtained. Application. 1. Protein Analyses. The SMIL-DS capillary was applied to acidic protein analyses. At first, separation of the heterogeneous acidic protein was evaluated by comparing the results obtained from the SMIL-DS capillary and the neutral (31) Terabe, S. J. Microcolumn Sep. 1993, 5, 23-33. (32) Cordova, E.; Gao, J.; Whitesides, M. Anal. Chem. 1997, 69, 1370-1379.

Analytical Chemistry, Vol. 70, No. 11, June 1, 1998

2257

Table 3. Chemical Stability of SMIL-PB and SMIL-DS Capillary (n ) 5) SMIL-PB

0.1 M HCl 1 M NaOH CH3CN CH3OH 5 M urea

SMIL-DS

EOF1a (10-8 m2 V-1 s-1)

EOF2b (10-8 m2 V-1 s-1)

degradation ratioc (%)

EOF1a (10-8 m2 V-1 s-1)

EOF2b (10-8 m2 V-1 s-1)

degradation ratioc (%)

3.770 3.762 3.765 3.431 3.776

3.180 ndd 3.216 3.374 3.763

15.6 e 14.6 1.7 0.3

3.750 3.758 3.765 3.748 3.765

ndd 3.735 3.762 3.745 3.775

e 0.6 0.1 0.1 0.3

a EOF was measured before rinsing with the solvents. b EOF was measured after rinsing with the solvents. c Degradation ratio ((EOF 1 2 1 EOF2)/EOF1) × 100 %. d EOF marker could not be detected within an hour. e Degradation ratio could not be determined. f Conditions: detection, 214 nm; applied voltage, 7 kV; buffers, phosphate buffer at pH 3.0 (I ) 0.05); capillary, 75 µm i.d. × 27 cm (20 cm effective length).

Table 4. Reproducibility of the Capillary-to-Capillary Coatings (n ) 3)a capillary

EOF R-lactalubumin β-lactoglobulin B β-lactoglobulin A

no. 1 (min)

no. 2 (min)

no. 3 (min)

no. 4 (min)

no. 5 (min)

RSD (%)

2.995 4.044 4.548 4.873

2.964 3.999 4.540 4.813

2.991 3.993 4.493 4.867

2.958 3.998 4.533 4.924

2.991 4.008 4.594 4.869

0.58 0.51 0.80 0.81

a Conditions: detection, 214 nm; applied voltage, 7 kV; buffers, phosphate buffer at pH 7.0 (I ) 0.05); capillary, 75 µm i.d. × 27 cm (20 cm effective length).

Figure 5. Endurance and regeneration of SMIL coatings. Conditions: applied voltage, +7 kV; buffer, phosphate buffer at pH 7.0 (I ) 0.05); capillary, 75 µm i.d. × 27 cm (20 cm effective length).

capillary. Bovine serum albumin (BSA) and pepsin were used as the test samples because these samples were heterogeneous.33,34 Figure 6 shows the electropherogram of BSA and pepsin obtained from the SMIL-DS and the neutral PEG capillary under physiological conditions (pH 7.4). The obvious four peaks of BSA were obtained from the SMIL-DS capillary, while four broad and small peaks were obtained from the PEG capillary (Figure 6a). Separation was probably improved when PEG capillary was used because EOF was suppressed. However, the ability to avoid the adsorption of the samples to the capillary wall of the SMIL-DS capillary was stronger than that of PEG capillary. Similarly, Figure 6b shows (33) Kubo, K. Anal. Biochem. 1996, 241, 42-46. (34) Landers, J. P.; Oda, R. P.; Madden, B. J.; Spelsberg, T. C. Anal. Biochem. 1992, 205, 115-124.

2258 Analytical Chemistry, Vol. 70, No. 11, June 1, 1998

the separation of pepsin under the same conditions. Although pepsin could be separated into five peaks at pH 8.5 by neutral PVA capillary,35 the peak shape obtained from the PEG capillary was degraded when a lower pH was used. On the other hand, pepsin could still be well detected and separated at pH 7.4 when SMIL-DS capillary was used. This is because DS is the hydrophilic polymer, and the sulfonic groups of the DS layer showed strong repulsion between the protein and the capillary wall. Next, number of theoretical plates (N) and symmetry factor (S) of the acidic proteins were compared between the SMIL-DS capillary and the neutral capillaries (Table 5). The acidic proteins of R-lactalbumin, trypsin inhibitor, and DHFR were used as the test samples because these samples, which are not heterogeneous, are preferable for the comparison of the efficiency. All of the peaks obtained from SMIL-DS capillary showed higher values of N than other neutral capillaries. In addition, symmetrical peaks were obtained especially from the SMIL-DS capillary. The comparison of the electrophrogram of the test samples between SMIL-DS capillary and PEG capillary is shown in Figure 7. The peaks obtained from the SMIL-DS capillary were apparently more efficient than those from the PEG capillary. These results showed the effectiveness of the SMIL-DS capillary not only for separation efficiency but also for the separation of heterogeneous samples under physiological conditions. 2. Analyses under Acidic Conditions. The pH-independent EOF of the SMIL-DS capillary allowed new CZE applications, especially under acidic conditions. Figure 8 shows the separations of basic, neutral, and acidic amino acids (AAs) using the SMIL(35) Gilges, M.; Kleemiss, M. H.; Schomburg, G. Anal. Chem. 1994, 66, 20382046.

Figure 6. Comparison of the heterogeneous acidic protein analyses between SMIL-DS and PEG capillary at pH 7.4. (a) BSA and (b) Pepsin. Conditions: applied voltage, +7 kV; buffer, phosphate buffer at pH 7.4 (I ) 0.05); capillary, 50 µm i.d. × 27 cm (20 cm effective length). Table 5. Comparison of the Separation Efficiency of the Acidic Proteins between SMIL-DS and Neutral Capillaries (n ) 3) R-lactalbumin

SMIL-DS PEG LPA PVA

trypsin inhibitor

DHFR

N (m-1)a

Sb

N (m-1)a

Sb

N (m-1)a

Sb

518 000 335 000 238 000 291 000

0.90 1.34 1.02 1.32

323 000 144 000 134 000 138 000

0.78 1.81 1.94 2.27

191 000 168 000 141 000 170 000

0.73 2.04 1.79 1.83

N, Number of theoretical plates; N ) 5.54(t/W0.5 t, Migration time of the protein. W0.5, Width of the peak at half peak height. b S, Symmetry factor; S ) W0.05/(2f). W0.05, Width of the peak at 1/20 of the peak height. f, Distance between the perpendicular from the peak maximum and the leading edge of the peak at 1/20 of the peak height. c Conditions: detection, 214 nm; applied voltage, 7 kV; buffers, phosphate buffer at pH 7.4 (I ) 0.05); capillary, 50 µm i.d. × 27 cm (20 cm effective length). a

)2.

DS capillary. The cationic, anionic, and neutral AAs were simultaneously separated at pH 2.4, while this could not be done at pH 7.0. The neutral AAs were separated at pH 2.4 because the pKa values of the acidic functional groups were slightly different. This result showed that a mixture of cationic, anionic, and neutral analytes could be detected simultaneously at the cathodic end at every pH if the SMIL-DS capillary was used. Therefore, optimization of the CZE separation could be achieved by optimizing the pH of the electrolyte. The SMIL-DS capillary could be also applied to MEKC. Figure 9 shows the separation of cresol isomers by MEKC. o-, m-, and p-cresol were separated at pH 2.7. Although it was also possible to separate these isomers under neutral conditions, the presence

Figure 7. Comparison of the pure acidic protein analyses between SMIL-DS and PEG capillary at pH 7.4. (a) R-Lactalbumin, (b) trypsin inhibitor, and (c) DHFR. Conditions: applied voltage, +7 kV; buffer, phosphate buffer at pH 7.4 (I ) 0.05); capillary, 50 µm i.d. × 27 cm (20 cm effective length).

of pH-independent EOF enabled the same separation to be obtained under acidic conditions. These results also support the strong chemical stability of the SMIL-DS capillary, because it was tolerant to SDS. 3. pKa Determination. The pH-independent EOF of the SMIL-DS capillary was useful for pKa determination of acidic functional groups. The pKa determination by CE could be achieved by analyzing the effective mobility obtained at each pH.36 However, measurement of the effective mobility under acidic conditions was quite time-consuming or impossible because the EOF was suppressed. Figure 10 a shows the electropherogram of the EOF and Trp at pH 3.1. It took more than 20 min to obtain the EOF when the uncoated capillary was used, while the EOF could be obtained within 5 min when the SMIL-DS capillary was used. Below pH 3.1, the EOF of the uncoated capillary was further decreased. However, the EOF of the SMIL-DS capillary was unchanged at every pH. The time needed for pKa determination was remarkably reduced by use of the SMIL-DS capillary. Figure 10b shows the dependence of the effective mobility on pH. The observed value was well in accordance with the fitting (36) Ishihama, Y.; Oda, Y.; Asakawa, N. J. Pharm. Sci. 1994, 83, 1500-1507.

Analytical Chemistry, Vol. 70, No. 11, June 1, 1998

2259

Figure 8. Separations of amino acids (AAs). (a) pH 7.0 and (b) pH 2.4. Migration order of AAs at pH 2.4: basic AA, Arg; neutral AAs, Gly, Ala, Thr, Ser, Phe, Pro; acidic AA, Glu. Conditions: applied voltage, +14 kV; buffer, phosphate buffer at pH 2.4 and 7.0 (I ) 0.05); capillary, 75 µm i.d. × 67 cm (60 cm effective length).

Figure 9. Separation of cresol isomers by MEKC (pH 2.8). Conditions: applied voltage, +10 kV; buffer, phosphate buffer at pH 2.8 (I ) 0.05) containing 100 mM SDS; capillary, 75 µm i.d. × 47 cm (40 cm effective length).

curve. The pKa of Trp was calculated to be 2.38 by multiline fitting, and the result agreed with the literature value for the acidic functional groups.37 The pKa values below pH 3 still may include some errors; however, pKa determination by use of the SMIL-DS capillary was well in accordance with that of other methods.37 Using the SMIL-DS capillary, fast and precise pKa determination of acidic functional groups could be achieved.

Figure 10. pKa measurement of Trp by CE method. (a) Electrophgerogram and (b) effective mobility. Conditions: detection, 214 nm; applied voltage, +7 kV; buffers, phosphate buffer at pH 2-3 (I ) 0.05), acetate buffer at pH 4-5 (I ) 0.05); capillary, 75 µm i.d. × 27 cm (20 cm effective length).

inner surface of the capillary without using covalent bonding, cross-linking, or thermal immobilization. The reproducible coating could be achieved in a simple procedure, and regeneration was possible when the coating was detached. The ability to perform the protein analyses was compared between the SMIL-DS capillary and the neutral capillary. The neutral capillary has the advantage of being useful for both acidic and basic proteins. Although the SMIL-DS capillary is not suitable for basic protein analyses because the capillary wall was negatively charged through pH 2-11, the ability to prevent the adsorption of the acidic proteins to the capillary wall was superior to that of the neutral capillaries because the sulfonic groups of the DS layer showed strong repulsion between the protein and the capillary wall. Therefore, efficient analyses of acidic proteins could be possible even under physiological conditions (pH 7.4). The SMILDS capillary also showed good performances for CZE, MEKC analysis, and pKa determination under acidic conditions, where the EOF was suppressed when the uncoated fused silica capillary or the neutral capillary was used. The positive use of pH-independent EOF of the SMIL-DS capillary expanded the utility of CE analyses. ACKNOWLEDGMENT We appreciate the gift of DHFR from Dr. Iwakura of the National Institute of Bioscience and Human-Technology.

CONCLUSIONS The novel capillary coating procedure, SMIL coating, was developed. The anionic polymer DS was strongly fixed to the

Received for review August 13, 1997. Accepted March 18, 1998.

(37) Dean, J. A. Lange’s Handbook of Chemistry; McGraw-Hill: New York, 1973.

AC9708755

2260 Analytical Chemistry, Vol. 70, No. 11, June 1, 1998