Anal. Chem. 1998, 70, 5272-5277
Stable Cationic Capillary Coating with Successive Multiple Ionic Polymer Layers for Capillary Electrophoresis 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 coated capillary modified with a cationic polymer was developed by using a novel coating procedure, successive multiple ionic-polymer (SMIL) coating. The SMIL coating was achieved by first attaching the cationic polymer to the capillary inner wall, and then the anionic polymer to the cationic polymer layer, and finally the cationic polymer to the anionic polymer layer. The stability of Polybrene (PB)-modified capillary made by SMIL coating was remarkably improved in comparison with a conventional PB-modified capillary. It endured during 600 replicate analyses and also showed strong stability against 1 M NaOH and 0.1 M HCl. The relative standard deviation of the run-to-run, day-to-day, and capillary-tocapillary coating was all below 1%, and good reproducibilities were obtained. The PB-modifed capillary made by SMIL coating was applied to the basic protein analyses. It gave good performances for the protein analyses even when the pH of the electrolyte was near the isoelectric point (pI) of the protein. In addition, 0.1 M NaOH rinse prior to the sample injection allowed the reproducible analysis of a highly adsorptive sample such as plasma because the adsorbed sample could be flushed out of the capillary. Besides protein analyses, an efficient analysis of the cationic drugs by capillary electrophoresis/mass spectrometry (CE/MS) was also possible. Capillary electrophoresis (CE) has been utilized as an important tool for biochemical sample analytes such as proteins and peptides.1,2 Its ability to perform analyses in an open-tubular capillary with aqueous electrolytes is quite useful for obtaining information on the native samples. However, when those samples were analyzed by using an uncoated fused silica capillary, adsorption of the samples to the capillary wall occurred. Therefore, minimization of the protein adsorption to the capillary wall is required. To overcome these problems, we previously developed the novel coating procedure, successive multiple ionic-polymer layer (SMIL) coating.3 Stable anionic-modified capillary was produced by sandwiching cationic Polybrene (PB) between the anionic * To whom correspondence should be addressed. Fax: (+81-298) 47-2037. Tel.: (+81-298) 47-5763. E-mail:
[email protected]. (1) Jorgenson, J. W.; Lukacs, K. D. Anal. Chem. 1981, 53, 1298-1302. (2) Jorgenson, J. W.; Lukacs, K. D. Science 1983, 222, 266-272. (3) Katayama, H.; Ishihama, Y.; Asakawa, N. Anal. Chem. 1998, 70, 2254-2260.
5272 Analytical Chemistry, Vol. 70, No. 24, December 15, 1998
polymer dextran sulfate (DS) and the capillary wall. The coating was achieved by rinsing the opposite charge of the ionic polymer solution alternatingly, and reproducible coating could be achieved even though the capillary sources were different. The characteristic of the anionic SMIL coating is electroosmotic flow (EOF) because coating layers were formed by ionic polymers. The SMIL coating exhibited good performance for acidic proteins analyses because of the ionic repulsion between the negatively charged proteins and negatively charged DS. The general modification of the capillary wall can be classified into the following categories: to perform dynamic coating by adding the cationic or neutral modifier to the electrolytes,4-6 to adsorb the cationic modifier to the capillary wall permanently by physical adsorption,7-11 and to fix the hydrophilic layer permanently by covalent bonding and/or cross-linking.12-20 The dynamic coating had a severe problem when CE was combined with mass spectrometry (CE/MS) because the presence of the nonvolatile buffer constituents may deteriorate the ionization of the analytes.21 Hence, permanent modification would be preferable. However, neither physical adsorption nor covalent bonding and/ or cross-linking are ideal procedures. The physical adsorption has a short lifetime, while it has a simple coating procedure and (4) Gilges, M.; Kleemiss, M. H.; Schomburg, G. Anal. Chem. 1994, 66, 20382046. (5) Yao, Y. J.; Li, S. F. Y. J. Chromatogr. A 1994, 663, 97-104. (6) Cifuentes, A.; Poppe, H.; Kraak, J. C.; Erim, F. B. J. Chromatogr. B 1996, 681, 21-27. (7) Wilktorowicz, J. E.; Colburn, J. C. Electrophoresis 1990, 11, 769-773. (8) Erim, F. B.; Cifuentes, A.; Poppe, H.; Kraak, J. C. J. Chromatogr. A 1995, 708, 356-361. (9) Chiu, R. W.; Jimenez, J. C.; Monnig, C. A. Anal. Chim. Acta. 1995, 307, 193-201. (10) Liu, Q.; Lin, F.; Hartwick, R. A. J. Chromatogr. Sci. 1997, 36, 126-130. (11) Assi, K. A.; Altria, K. D.; Clark, B. J. J. Pharm. Biomed. Anal. 1997, 15, 1041-1049. (12) Bruin, G. J. M.; Huisden, R.; Kraak. J. C.; Poppe, H. J. Chromatogr. 1989, 480, 339-349. (13) Herren, B. J.; Shafer, S. G.; Alstine, S. V.; Harris, J. M.; Snyder, R. S. J. Colloid Interface Sci. 1987, 115, 46-55. (14) Bruin, G. J. M.; Chang, J. P.; Kuhlman, R. H.; Zegers, K.; Kraak, J. C..; Poppe, H. J. Chromatogr. 1989, 471, 429-436. (15) Hjerten, S. J. Chromatogr. 1985, 347, 191-198. (16) Hjerten, S.; Johansson, M. K. J. Chromatogr. 1991, 550, 811-822. (17) Cobb, K. A.; Dolnik, V.; Novotony, M. Anal. Chem. 1990, 62, 2478-2483. (18) Swedberg, S. A. Anal. Biochem. 1990, 185, 51-56. (19) Schmalzing, D.; Piggee, C. A.; Foret, F.; Carrilho, E.; Karger, B. L. J. Chromatogr. A 1993, 480, 149-159. (20) Huang, X.; Horvath, C. J. Chromatogr. A 1997, 788, 155-164. (21) Niessen, W. M. A.; Tjaden, U. R.; Geef, J. J. Chromatogr. 1993, 636, 3-19. 10.1021/ac980522l CCC: $15.00
© 1998 American Chemical Society Published on Web 10/31/1998
Figure 1. Scheme of the capillary profile.
Figure 2. EOF of uncoated and SMIL-PB(3) capillaries. 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 8-13 (I ) 0.05); capillary 50 µm i.d. × 27 cm (20 cm effective length).
Figure 4. Basic protein analyses by SMIL-PB capillaries and neutral capillaries. (a) Electropherogram. (b) Theoretical plates. Conditions were as described in Figure 3. Symbols: (A) R-chymotripsinogen A, (B) ribonuclease A, (C) lysozyme, and (D) cytochrome c.
Figure 3. Endurance of SMIL-PB(1) and SMIL-PB(3) capillaries. Conditions: detection, 214 nm; applied voltage, 7 kV; buffers, phosphate buffer at pH 3.0 (I ) 0.05); capillary, 50 µm i.d. × 27 cm (20 cm effective length).
good reproducibility.22 On the other hand, some of the covalent bonding and/or cross-linking requires a relatively complicated coating procedure, while it has a long lifetime. Therefore, the stable coated capillary made by the simple procedure of SMIL (22) Cordova, E.; Gao, J.; Whitesides, G. M. Anal. Chem. 1997, 69, 1370-1379.
coating is quite attractive because the method could be expected as an alternative coating procedure to covalent bonding. In this study, we further modified the SMIL coating and developed a stable cation-modified capillary in order to extend its application to the basic analytes. Triple layers of the ionic polymers were formed by attaching PB as a third layer. The performances of PB-modified capillary by SMIL coating were evaluated, and its applicability to CE/MS was also investigated. EXPERIMENTAL SECTION Apparatus. CE was performed with a Beckman P/ACE 2100 system (Fullerton, CA). An uncoated fused silica capillary Analytical Chemistry, Vol. 70, No. 24, December 15, 1998
5273
Table 1. Chemical Stability of SMIL-PB(1) and SMIL-PB(3) Capillary (n ) 5) SMIL-PB(1)
1 M NaOH 0.1 M HCl 5 M urea CH3OH CH3CN 1% (v/v) HCOOH SCAT
SMIL-PB(3)
EOF1a (10-8 m2 V-1 s-1)
EOF1b (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 (%)
3.431 3.461 3.439 3.459 3.453 3.443 3.509
ndd 1.300 3.330 1.910 2.864 2.728 2.991
e 62.40 3.17 1.91 17.05 20.78 14.76
3.431 3.461 3.453 3.459 3.439 3.443 3.520
3.371 3.421 3.427 3.428 3.391 3.383 3.484
1.74 1.14 0.76 0.91 1.40 1.76 1.02
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. Conditions: detection, 214 nm; applied voltage, 7 kV; buffers, phosphated buffer at pH 3.0 (I ) 0.05); capillary, 50 µm i.d. × 27 cm (20 cm effective length).
Table 2. Reproducibilities of SMIL-PB(3) Capillary Coating (n ) 5)a
run-to-run day-to-day capillary-to-capillary
EOF average (min)
RSD (%)
3.570 3.535 3.481
0.17 0.69 0.85
a Conditions: detection, 214 nm; applied voltage, -7 kV; buffers, phosphated buffer at pH 3.0 (I ) 0.05); capillary, 50 µm i.d. × 27 cm (20 cm effective length).
Figure 5. Comparison of the electropherogram of the proteins obtained from SMIL-PB(3) and PEG capillary at pH 9. Conditions were as described in Figure 2. Symbols: (A) R-chymotripsinogen A, (B) ribonuclease A, (C) lysozyme, and (D) cytochrome c.
(Beckman) of 27 cm (20 cm effective length) × 50 µm i.d., 360 µm o.d., was used for the evaluation of the cation-modified capillary. The poly(ethylene glycol) (PEG)-modified capillary (µSIL DB-WAX capillary, J&W Scientific, Folsom, CA), linear polyacrylamide-modified capillary (LPA-capillary) (BioCAP LPA coated capillary, Bio-Rad, Hercules, CA), and poly(vinyl alchol)modified capillary (PVA capillary) (Hewlett-Packard, Waldbronn, Germany) of 50 µm i.d., 27 cm, were used as the neutral capillaries. The capillaries were thermostated at 25 °C by using liquid coolant. The applied voltages were 7 kV, and the samples were injected by pressure (0.5 psi; 1 psi ) 6894.76 Pa) for 2 s. CE/MS was performed by the combination of HCZE-30PNO.25LDS (Matsusada Precision, Shiga, Japan) as a voltage supplier for CE with the Perkin-Elmer Sciex API III (Thronhill, Ontario, Canada) as a detector. A Harvard model 22 syringe pump (South Natick, MA) was used for the sheath flow to the coaxial interface 5274 Analytical Chemistry, Vol. 70, No. 24, December 15, 1998
of the CE/MS. An uncoated fused silica capillary (GL Sciences, Tokyo, Japan) of 40 cm × 50 µm i.d., 150 µm o.d., was used for CE/MS analysis. Reagents and Material. Polybrene (PB) (Fluka, Buchs, Switzerland) was used as a cationic coating reagent. Protein markers, R-chymotripsinogen A, ribonuclease A, lysozyme, and cytochrome c were from Sigma (St. Louis, MO). Detergent SCAT 20-X was from Nakarai Tesque (Kyoto, Japan). Formamide (Pharmacia Biotech, Uppsala, Sweden) was used as an EOF marker. Plasma was obtained from a rat. β-Blockers of alprenolol, acebtolol, (()-metoprolol, and nadolol (Sigma) were used as the test samples for CE/MS. Other reagents were analytical grade. Procedure for Successive Multiple Ionic-Polymer Layer (SMIL) Coating. SMIL coating was achieved by the protocol described below. All of the rinsing procedure employed the rinsing function of the Beckman P/ACE 2100. The capillary was rinsed with 0.1% SCAT for 3 min, 1 M NaOH for 30 min, and then deionized water for 15 min in order to clean the capillary and enhance the dissociation of the silanol groups. After preconditioning, the capillary was rinsed with 10% PB solution for 15 min to form the first cationic layer. An anionic polymer such as dextran sulfate, alginic acid, or hyaruronic acid, which was expected to work as an adhesion material for PB, was next rinsed for 15 min to form a second layer. Finally, 10% PB solution was rinsed over the anionic layer for 15 min. The scheme of the SMIL-coated capillary is shown in Figure 1. The capillary on which PB was directly adsorbed to the capillary wall was named SMIL-PB(1) capillary, and this corresponds to the conventional PB capillary. The capillary on which PB was attached as a third layer was named SMIL-PB(3) capillary, and this capillary is the novel cationmodified capillary. RESULTS AND DISCUSSION Measurement of EOF. The EOF of uncoated capillary and SMIL-PB(3) capillary over the pH range of 2 to 13 was investigated. The results are shown in Figure 2. The EOF of the uncoated capillary was generated from anode to cathode. The EOF was stable above pH 6; however, it started to decrease from pH 6 and was almost suppressed below pH 3. On the other hand, the EOF of the SMIL-PB(3) capillary was reversed from cathode to anode and exhibited pH-independent EOF. Endurance of SMIL-PB Capillary. An endurance test of the SMIL-PB capillary against continuous analysis was performed, and
Figure 6. Direct injection of the rat plasma sample performed by SMIL-PB(3) capillary. (a) Without 0.1 M NaOH rinse between runs. (b) With 0.1 M NaOH rinse between runs. Conditions were as described in Figure 3.
the results are shown in Figure 3. The endurance was evaluated by measuring the EOF at pH 3.0, where EOF would not be generated if the coating were detached. Each run was performed for 10 min, and if the EOF could not be detected within 10 min, the analysis was extended to 60 min. If the EOF could not be detected within 60 min, the EOF was defined as 0 (m2 V-1s-1). The PB layer of the SMIL-PB(1) capillary endured only 20 runs. On the other hand, the SMIL-PB(3) capillary endured 600 runs (99 h). The severe disadvantage of the conventional SMIL-PB(1) capillary was its low endurance. However, we overcome this problem by SMIL coating. It was suggested that the interaction between PB and the anionic polymer was stronger than the interaction between the capillary wall and PB. Concerning the slight change in the EOF of the SMIL-PB(3) capillary, the EOF was stable at the beginning but slightly decreased around 20 runs and subsequently stabilized. Although the same tendency was observed for the SMIL-PB(1) capillary, the PB layer was detached at the 21st run. It can be estimated from these results that the slight decrease in EOF was due to the detachment of the minute amount of mono-PB layer existing in the SMIL-PB(3) capillary. Chemical Stability. The chemical stability of the SMIL-PB(1) and SMIL-PB(3) capillaries was investigated, and the results are shown in Table 1. The EOF was first measured when the SMIL-PB(1) and SMIL-PB(3) capillaries were produced. The capillary was then rinsed with the solvent for 15 min, and the EOF was measured again. The chemical stability was evaluated on the basis of the degradation ratio described in Table 1. The SMIL-PB(1) capillary was unstable after 0.1 M HCl, CH3CN, 1% HCOOH, SCAT rinse, and the coating was perfectly detached after 1 M NaOH rinse. On the other hand, the degradation ratio obtained from the SMIL-PB(3) capillary was all
below 2%, and the chemical stability was improved. This stability of the coating was due to the attachment of the triple ionicpolymer layers, and the tolerance to both NaOH and HCl assured the wide pH range analysis. Reproducibilities. The reproducibilities of SMIL coating were investigated. The reproducibilities were evaluated on the basis of the relative standard deviation (RSD) of EOF obtained from five replicate analyses. The results are shown in Table 2. The run-to-run, day-to-day, and capillary-to-capillary RSDs were all below 1%, and good reproducibilities were obtained. The good reproducibilities are one of the great advantages of SMIL coating, because the coating protocol was almost the same as the conventional physical adsorption. Protein Analyses. The ability to perform the protein analysis was evaluated by comparing the separation efficiency of the protein test samples between SMIL-PB capillaries and the neutral capillaries such as LPA, PVA, and PEG capillaries. Four basic proteins, R-chymotripsinogen A, ribonuclease A, lysozyme, and cytochrome c, were used as the test samples. The isoelectric point (pI) of the R-chymotripsinogen A, ribonuclease A, lysozyme, and cytochrome c were 8.8, 8.7, 11.0 and 10.2, respectively.23 Figure 4 a shows the electropherograms of the basic proteins at pH 3.0. In the case of the uncoated capillary, the peaks of the proteins could not be obtained because the positively charged proteins were irreversibly adsorbed to the negatively charged capillary wall (data not shown). However, the peaks could be detected when PB was attached to the uncoated capillary. Comparing SMIL-PB capillaries with the neutral capillaries, the migration order of the proteins was reversed because the EOF of the SMIL-PB capillaries was generated from cathode to anode. The theoretical plates (N) (23) Cohen, N.; Grushka, E. J. Cap. Electrophor. 1994, 1, 112-115.
Analytical Chemistry, Vol. 70, No. 24, December 15, 1998
5275
Figure 7. CE/MS analysis of β-blockers. (a) Selected ion monitoring. (b) Total ion chromatogram. Conditions: applied voltage for CE, -5 kV; ion spray voltage for MS, +5 kV; buffer, 0.5% (v/v) formate buffer, pH 2.26; sheath liquid, mixture of 0.5% (v/v) formate buffer and acetonitrile (25/75); capillary, 50 µm i.d. × 40 cm; injection; the sample was hydrodynamically injected by manual procedure.
of the proteins peaks obtained from SMIL-PB capillaries and neutral capillaries were calculated in order to compare the separation efficiency. The results are shown in Figure 4b. The N obtained from SMIL-PB(1) capillary was apparently lower than that of the neutral capillaries. On the other hand, the separation efficiency obtained from the SMIL-PB(3) capillary was almost the same in comparison with the neutral capillaries. It could be estimated from the difference in the apparent mobility of the proteins obtained from the SMIL-PB(1) capillary and the SMILPB(3) capillary that the SMIL-PB(3) capillary has a higher density of PB than the SMIL-PB(1) capillary. Besides the comparable performance of SMIL-PB(3) and the neutral capillaries under acidic conditions, the SMIL-PB(3) capillary had the advantage of analyzing basic proteins under alkaline conditions. The comparison of the electropherogram between the SMIL-PB(3) capillary and PEG capillary at pH 9 is shown in Figure 5. When the PEG capillary was used, only broad peaks of lysozyme and cytochrome c were obtained. It was impossible to detect R-chymotripsinogen A and ribonuclease A within 3 h because the pH of the electrolyte was close to the pI of these two proteins. On the other hand, all of the peaks could be detected within 5 min when the SMIL-PB(3) capillary was used. The 5276
Analytical Chemistry, Vol. 70, No. 24, December 15, 1998
presence of the EOF enabled efficient analysis, even when the pH of the electrolyte was in the alkaline region. The SMIL-PB(3) capillary was useful for real biological sample analysis as well as purified test sample analysis. A plasma sample obtained from a rat was directly injected into the SMIL-PB(3) capillary, and its performance was evaluated. The electropherograms of the plasma at pH 3.0 are shown in Figure 6. In the case of Figure 6a, the capillary was flushed with the buffer for only 5 min prior to the sample injection. The peak was well detected during the first run. Neither LPA, PVA, nor PEG capillary could detect the peak at pH 3.0, 7.0, and 10.0 because the plasma was strongly adsorbed to the capillary wall (data not shown). On the other hand, it was also possible to detect the peak at pH 7.0 and 10.0 when anionic dextran sulfate-modified (SMIL-DS) capillary was used (data not shown). The Coulombic repulsion between ionic-polymer layer can further avoid the interaction between the proteins and the capillary wall in comparison with the neutral capillaries. However, the separation efficiency was decreased when five replicate analyses were performed. This is because it was impossible to remove the adsorbed sample simply by rinsing with the buffer prior to analysis. To overcome this problem, the SMIL-PB(3) capillary was first rinsed with 0.1 M NaOH for 5 min and next rinsed with the buffer for 5 min prior to the sample injection. The adsorbed sample was flushed out to the capillary by 0.1 M NaOH, and hence, reproducible peaks were obtained during the five runs (Figure 6b). It was impossible to perform a NaOH rinse when the neutral capillaries were used. The strong stability of the SMIL-PB(3) capillary against alkaline conditions assured the NaOH rinsing. These results demonstrated the effectiveness of the SMIL-PB(3) capillary for direct injection of real biological samples. CE/MS Analysis by SMIL-PB(3) Capillary. Use of the cation-coated capillary is essential to perform CE/MS analysis of cationic solutes because the interaction between the capillary wall and analytes must be reduced.24,25 The SMIL-PB(3) capillary was applied to the CE/MS system where the sheath flow was needed for the ionization. Four cationic β-blockers, alprenolol, acebtolol, (()-metoprolol, and nadolol, were used as the test samples. The selected ion monitoring (SIM) and total ion chromatogram (TIC) of the test samples are shown in Figure 7a and b, respectively. All of the test samples could be detected in the positive mode (Figure 7a). In addition, tailing of the cationic solutes was found to be suppressed because of the Coulombic repulsion between the positively charged PB layer and the positively charged test samples (Figure 7b). These results showed that the SMIL-PB(3) capillary could perform efficient analysis of cationic drugs by CE/MS. The neutral capillary, such as the PVA capillary, could perform both acidic and basic protein analyses;4 however, it could not be applied to micro- or nanospray CE/MS, where the EOF is the only driving (24) Mosley, M. A.; Deterding, L. J.; Tomer, K. B.; Jorgenson, J. W. Anal. Chem. 1991, 63, 109-114. (25) Thibault, P.; Paris, C.; Pleasance, S. Rapid. Commun. Mass Spectrom. 1991, 5, 484-490. (26) Figeys, D.; Oostveen, I. V.; Ducret, A.; Aebersold, R. Anal. Chem. 1996, 68, 1822-1828. (27) Figeys, D.; Aebersold, R. J. Chromatogr. B 1997, 695, 163-168. (28) Kelly, J. F.; Ramaley, L.; Thibault, P. Anal. Chem. 1997, 69, 51-60.
force to introduce the analytes to MS.26-28 Therefore, the SMILPB(3) capillary would be useful for micro- or nanospray CE/MS analyses. CONCLUSIONS A stable cation-modified capillary was developed by a novel coating procedure, SMIL coating. SMIL coating could be achieved by a simple procedure with good reproducibility. The SMIL-PB(3) capillary exhibited strong endurance and chemical stability. Highly efficient peaks of the basic proteins could be obtained because Coulombic repulsion exists between the positively charged PB layer and the positively charged proteins. The presence of pH-independent EOF of the SMIL-PB(3) capillary will enable the performance of fast analyses, even in
the alkaline region, where the pH of the electrolyte was near the pI of the proteins. In addition, strong chemical stability assured the performance of the 0.1 M NaOH rinse, and as a result, reproducible analysis of a highly absorptive sample such as plasma was achieved. The SMIL-PB(3) capillary was also applicable to CE/MS analysis. An efficient MS electropherogram of the cationic drugs could be obtained. In the future, we will apply the SMILPB(3) capillary to micro- or nanospray CE/MS where the presence of EOF is essential.
Received for review May 13, 1998. Accepted September 12, 1998. AC980522L
Analytical Chemistry, Vol. 70, No. 24, December 15, 1998
5277
