Anal. Chem. 1999, 71, 115-118
Membrane Supports as the Stationary Phase in High-Performance Immunoaffinity Chromatography Dongmei Zhou, Hanfa Zou,* Jianyi Ni, Li Yang, Lingyun Jia, Qiang Zhang, and Yukui Zhang
National Chromatographic R & A Center, Dalian Institute of Chemical Physics, The Chinese Academy of Sciences, Dalian 116011, China
The membrane with a composite of cellulose grafted with acrylic polymers formed by polymerizing a glycidyl methacrylate in the presence of dispersed cellulose fiber was prepared as the stationary phase; the column (40 × 4 mm i.d.) which was compatible with the HPLC instrument was packed with pieces of the cut membrane. Protein A and human IgG were immobilized on the membrane stationary phase. The column based on the membrane support provided us good reproducibility, high efficiency, and low back pressure. High-performance immunoaffinity chromatographic analysis of human IgG in serum and polyclonal antibody to human IgG raised in goat was performed within 2.5 min. The fast-speed immunoaffinity analysis was developed by increasing the flow rate of the mobile phase and decreasing the duration time for the switch of the mobile phase; an operation for immunoaffinity analysis of human IgG could be finished within 30 s. High-performance liquid chromatography (HPLC) revolutionized analytical chemistry by facilitating very rapid and efficient separations and the detection and determination of the components of virtually any mixture. The separation, isolation, and purification of biopolymers is very important for their effective application. The analytical and preparative HPLC separations of individual biological macromolecules from their mixtures with both low- and high-molecular-mass compounds has been reviewed several times.1-3 At present, most chromatographic separations are carried out in column packed almost exclusively with beadshaped particles,4 as the technology of bead preparation has been known for more than two decades. Recently, a new technique called “perfusion chromatography” was invented,5-7 which involves the flow of liquid through a porous chromatographic particle for (1) Frenz, J.; Horvath, Cs. In High-Performance Liquid Chromatography: Advances and Perspectives; Horvath, Cs., Ed.; Academic Press: New York, 1988; Vol. 5, p 211. (2) Gooding, K. M.; Reginer, F. E. Eds. HPLC of Biological Molecules, Methods and Applications; Marcel Dekker: New York, 1990. (3) Mant, C. T.; Hodges, R. S. Eds. High-Performance Liquid Chromatography of Peptides and Proteins: Separation, Analysis and Conformation; CRC Press: Boca Raton, FL, 1991. (4) Unger, K. K., Ed. Packings and Stationary Phases in Chromatographic Techniques; Marcel Dekker: New York, 1990. (5) Afyan, N. B.; Gordon, N.; Maszaroff, I.; Varady, L.; Fulton, S. P.; Yang, Y.; Regnier, F. E. J. Chromatogr. 1990, 519, 1. (6) Afeyan, N. B.; Fulton, S. P.; Regnier, F. E. J. Chromatogr. 1991, 544, 267. (7) Fulton, S. P.; Afeyan, N. B.; Gordon, N. F.; Regnier, F. E. J. Chromatogr. 1992, 547, 452. 10.1021/ac980613i CCC: $18.00 Published on Web 12/03/1998
© 1998 American Chemical Society
reducing resistance to stagnant mobile-phase mass transfer without sacrificing adsorbent capacity or necessitating extremely high-pressure operation. Many applications of perfusion chromatography to separations of biopolymers have been reported.8-10 Traditional membranes were introduced into affinity chromatography in 1988.11-13 Ion-exchange cellulose membranes stacked in a cartridge also gave good results in the separation and purification of proteins.14,15 The recently introduced high-performance membrane chromatography (HPMC) combines the advantages of both membrane technology (simple scale-up, lowpressure drop across a membrane, and low cost) and column chromatography (high selectivity and efficiency of separation, high loading capacity).16-19 The detailed studies of individual effects of variable parameters in ion-exchange, hydrophobic interaction, and reversed-phase protein separations confirmed that HPMC obeys the rules typical of HPLC.20 With comparison of HPMC to perfusion chromatography, both perfusion chromatography and HPMC can be easily scale-up for purification of biopolymers, but the cost of perfusion chromatography is much higher than that of HPMC. In this work, high-performance immunoaffinity chromatographic columns with immobilization of ligands of protein A and human IgG on the membrane support have been prepared; analyses of human IgG in serum and polyclonal antibody to human IgG raised in goat serum have been performed. (8) Fulton, S. P.; Meys, M.; Varady, L.; Jansen, R.; Afeyan, N. B. Biochromatography 1991, 11, 226. (9) Zou, H. F.; Zhang, Y. K.; Lu, P. C.; Krull, I. S. Biomed. Chromatogr. 1996, 10, 78. (10) Zou, H. F.; Zhang, Y. K.; Lu, P. C.; Krull, I. S. Biomed. Chromatogr. 1996, 10, 122. (11) Huang, S. H.; Roy, S.; Hou, K. C.; Tsao, G. T. Biotechnol. Prog. 1988, 4, 159. (12) Hou, K. C.; Zaniewski, R. J. Chromatogr. 1990, 525, 159. (13) Langlotz, P.; Kroner, K. H. J. Chromatogr. 1992, 591, 107. (14) Tan, L. U. L.; Yu, E. K. C.; Luis-seize, G. W.; Saddler, J. N. Biotechnol. Bioeng. 1987, 30, 96. (15) Wang, H.; Li, T.; Zou, H.; Zhang, Y.; Chao, J.; Chao, L. Biomed. Chromatogr. 1996, 10, 139. (16) Tennikova, T. B.; Bleha, M.; Sevc, F.; Almazova, T. V.; Belenkii, B. G. J. Chromatogr. 1991, 555, 59. (17) Abou-Rebyeh, H.; Korber, F.; Schubert-Rehberg, K.; Reusch, J.; Josic, Dj. J. Chromatogr. 1991, 566, 341. (18) Josic, Dj.; Reusch, J.; Loster, K.; Baum, O.; Reutter, W. J. Chromatogr. 1992, 590, 59. (19) Josic, Dj.; Lim, Y. P.; Strancar, A.; Reutter, W. J. Chromatogr. 1994, 662, 217. (20) Tennikova, T. B.; Sevc, F. J. Chromatogr. 1993, 646, 279.
Analytical Chemistry, Vol. 71, No. 1, January 1, 1999 115
glutaraldehyde was removed by washing the column with 40 mL of borate buffer. This chemical reaction can be described by following equations:
(2)
Figure 1. Chemical structure of GMA-cellulose composite membrane.
EXPERIMENTAL SECTION Preparation of Membrane Media. The chromatographic solid matrix is a composite of cellulose grafted with acrylic polymers formed by polymerizing a glycidyl methacrylate in the presence of dispersed cellulose fiber, followed by an in situ covalent binding of the acrylic polymer to the cellulose as described by Hou et al.,21 which is named arcylic membrane here. The reaction processes are shown in Figure 1. The bicomponent fiber thus formed consisted of a cellulosic core as the mechanical support and the acrylic sheath as a chemical functional group carrier. The composite fiber can be further derivatized to the required functional groups. The media with amino groups were prepared by reacting the glycidyl groups grafted on the fiber surface with hexyldiamine monomers at 80 °C for an additional hour; it is named amino membrane here. Preparation of Immunoaffinity Column. The composite fiber carrying specific functional groups was then fabricated in paper form by a conventional paper-forming machine. Pieces of membrane with 4.0-mm diameter were cut off; the column (40 × 4 mm i.d.), which was compatible with the HPLC instrument, was packed with the cut membrane pieces until full. Two approaches were used for the immobilization of protein A on the acrylic and amino membranes, respectively. (1) The aldehyde groups on the acrylic membrane were generated directly from the glycidyl groups by acid hydrolysis with HCl at pH 0.6 for 6 h to form vicinal hydroxyl groups, followed by periodate oxidation with NaIO4 (12%) for 30 min. Those chemical reactions can be described by following equation:
(1)
M in eq 1 is the membrane. (2) Aldehyde groups on the amino membrane were generated from the amino matrixes by recirculating 0.1 mol/L sodium borate and HCl buffer, pH 8.2, containing 0.25% glutaraldehyde for 2 h at room temperature. Excess (21) Hou, K. C.; Zanieweski, R.; Roy, S. Biotechnol. Appl. Biochem. 1991, 13, 257.
116
Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
The coupling of protein A with the aldehyde group-generated membranes was performed by recirculating the solutions of 4 mg/ mL protein A in 0.1 mol/L borate buffer, pH 8.2, in the presence of 0.025 mol/L NaCNBH3 for 15 h at room temperature. The uncoupled ligand was removed by rinsing the column with buffer solution until the ligand is no longer detectable in the eluent. The excess active groups were deactivated by recirculating 1.0% glycine ethyl ester hydrochloride in 0.1 mol/L sodium phosphate, pH 6.5, containing the reduction agent for 4 h. The immunoaffinity column was washed extensively and then equilibrated in the loading buffer for the evaluation and analysis of human IgG. The immunoaffinity column by immobilization of human IgG was also prepared by the same method reported above and used for analysis of goat anti-HIgG. Chromatography. The HPLC system used was two Waters 510 pumps (Waters, Milford, CT) controlled by a WDL-95 workstation (National Chromatographic R&A Center, Dalian, China). Eluates were detected at wavelength of 280 nm by a Spectra-200 UV detector (Spectra-Physics, San Jose, CA). Human IgG and bovine serum albumin (BSA) were purchased from Sigma Chemical Co. (St. Louis, MO). Human serum and a crude powder of goat-polyclonal antibody to human IgG (goat anti-HIgG) were obtained from the Dalian Hospital (Dalian, China) and Tianjin Central Hospital (Tianjin, China), respectively. All other chemicals were analytical-reagent grade, and the solutions were made in double-distilled water. The loading buffer for HPIAC was 10 mmol/L phosphate buffer containing 0.15 mol/L NaCl, and the pH value of loading buffer was adjusted to 7.2 with NaOH. The elution buffer was 0.15 mol/L of NaCl in water adjusted to pH 2.6 with HCl. RESULTS AND DISCUSSION Figures 2 and 3 showed the chromatograms to test the nonspecific adsorption of proteins on the protein A columns based on the coupling of protein A on membrane supports through the chemical reactions shown as in eqs 1 and 2, respectively. Curve 3 in Figures 2 and 3 represents the real chromatograms, the difference between curves 1 and 2 shown in Figures 2 and 3, where curves 1 and 2 were obtained by injections of 20 µL of loading buffer containing 100 µg of BSA and only loading buffer itself, respectively. It can be seen that there was a small peak observed in Figure 3 when the eluent was changed from the loading buffer to the elution buffer, which was caused by the nonspecific adsorption of BAS on the protein A column due to the long coupling arm prepared by the chemical reaction of eq 2. But no such a nonspecific adsorption was observed for the protein A column on the stationary phase prepared by the chemical reactions of eq 1, because no peak was observed when the eluent was changed from the loading buffer to the elution buffer. All
a
Figure 2. Chromatograms of BSA solution on the protein A column by its immobilization on the membrane through the chemical reactions shown in eq 1. Experimental conditions: 20 µL of sample solution was injected with loading buffer as the eluent, and 2 min after injection of sample solution, the elution buffer was used as the eluent. Flow rate 1.0 mL/min; detection wavelength 280 nm. Curves: (1) loading buffer containing 100 µg of BSA; (2) loading buffer; (3) difference between curves 1 and 2.
b
Figure 4. Chromatographic analysis of human IgG on the protein A column. Experimental conditions: (a) 20 µL of human IgG solution (2.5 mg/mL) and (b) 10-fold diluted solution of human serum were injected. Other experimental conditions were the same as in Figure 2. Chromatographic peaks: (1) nonretained solutes; (2) human IgG.
Figure 3. Chromatograms of BSA solution on the protein A column by its immobilization on the membrane through the chemical reactions shown in eq 2. Experimental conditions: 20 µL of sample solution was injected with loading buffer as the eluent, and 2.3 min after injection of sample solution, the elution buffer was used as the eluent. Other experimental conditions were the same as in Figure 2. Curves: (1) loading buffer containing 100 µg BSA; (2) loading buffer; (3) difference between curves 1 and 2.
experiments below were performed on the protein A column based on the membrane support prepared by the chemical reactions of eq 1. The pressure drop of the column was determined by changing the flow rate of the mobile phase, and it was observed that the column packed with the membrane support has a very low pressure drop, which is important for high-speed analysis in highperformance immunoaffinity chromatography. For example, the pressure drop of the column was only about 4 MPa even at a flow rate of 3.0 mL/min. Figure 4 showed the typical chromatograms of human IgG standard sample and diluted human serum on the protein A column, which indicated that an immunoaffinity analysis of the human IgG sample could be finished within 2.5 min at a flow rate of 1.0 mL/min, and the peak width at baseline for the retained human IgG is less than 0.2 min. The reproducibility of the column was evaluated by five injections of 10-fold diluted human serum, and it was observed that the protein A immunoaffinity column
based on membrane support gave very good reproducibility of peak areas with RSD less than 1.0% for the nonretained impurities and the retained human IgG. The calibration curve for quantitation of human IgG on the protein A column by injection of the standard human IgG samples was determined, and a good linearity of the peak area of retained human IgG versus the injection amount of human IgG in the range 3-75 µg was obtained with a regression coefficient higher than 0.999. From the average peak areas of the retained IgG measured, the amount of human IgG in serum can be calculated by the calibration curve as 14 mg/mL. Fast-speed analysis of human IgG on the protein A immunoaffinity column with membrane support was developed by increasing the flow rate of the mobile phase and decreasing the duration time for switching of the mobile phase. Figure 5 showed typical chromatograms for injection of the standard sample and the 10fold diluted human serum. It can be seen that one operation for the immunoaffinity analysis of human IgG can be finished in even less than 30 s, and the peak width at baseline for the retained human IgG was about 5 s. The amount of human IgG in serum can be quantitated as 13.6 mg/mL by comparison of peak areas for the retained human IgG in the standard sample and diluted human serum. This result means that the amount of human IgG in serum measured by fast-speed analysis is about 3% lower than that measured by the utilization of a calibration curve measured at a flow rate of 1 mL/min, which may be caused from the impurities that existed in the standard sample of human IgG shown in peak 1 in Figure 5a. Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
117
a
a
b
b
Figure 5. Fast-speed analysis of human IgG on the protein A column. Experimental conditions: (a) 20 µL of human IgG solution (2.5 mg/mL) and (b) 10-fold diluted solution of human serum were injected, and 20 s after injection of sample solution, the elution buffer was used as the eluent. Flow rate 3.0 mL/min; other experimental conditions were the same as in Figure 4. Chromatographic peaks: (1) nonretained solutes; (2) human IgG.
Furthermore, the immunoaffinity analysis of goat-polyclonal antibody to human IgG (goat anti-HIgG) in crude powder was carried out on the protein A and human IgG columns; the latter, the human IgG, was immobilized on the membrane support. The obtained chromatograms were shown in Figure 6. It can be seen that the peak for the retained goat-anti HIgG on the protein A column is much smaller than that on the human IgG column, which means that the immunoaffinity interaction between protein A and goat anti-HIgG is much weaker than that between HIgG and goat anti-HIgG. The reproducibility of the HIgG column for analysis of goat anti-HIgG was tested by five injections of goat anti-HIgG solution, made from the crude powder dissolved in the loading buffer; it was observed that the HIgG column also gave quite good reproducibility with a RSD of 3.2% for the peak areas of retained goat anti-HIgG. It is assumed that the goat anti-HIgG has the same UV absorpitvity as the human IgG at 280 nm; then we can roughly estimate that the amount of goat anti-HIgG in crude powder is about 105 µg/mg. CONCLUSION Columns with membrane support as the stationary phases in high-performance immunoaffinity chromatography have been
118 Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
Figure 6. Chromatograms of goat anti-HIgG solution on the (a) protein A and (b) HIgG columns. Experimental conditions: 20 µL of solution by dissolving crude powder of goat-anti HIgG in loading buffer (5 mg/mL) was injected. Other experimental conditions were the same as in Figure 3. Chromatographic peaks: (1) nonretained solutes; (2) goat anti-HIgG.
developed. The column developed provided us good reproducibility, high efficiency, and low back pressure. High-performance immunoaffinity chromatographic analysis of human IgG in serum and polyclonal antibody to human IgG raised in goat were performed on the protein A and human IgG columns, respectively; usually an immunoaffinity analysis of the human IgG and goat anti-polyclonal antibody could be finished within 2.5 min at a flow rate of 1.0 mL/min. Furthermore, the fast-speed analysis of the human IgG within 30 s on the immunoaffinity column was developed by increasing the flow rate of mobile phase and decreasing the duration time for switching of the mobile phase. ACKNOWLEDGMENT Financial support from the National Natural Science Foundation of China (No. 29635010) is gratefully acknowledged. H.Z. is recipient of the excellent young scientist award from the National Natural Science Foundation of China (No. 29725512). Received for review June 4, 1998. Accepted October 15, 1998. AC980613I