- Email: [email protected]
Efficient purification of mouse monoclonal antibodies from ascites fluid by medium-performance anion exchange chromatography
1987, Gene Anal Techn 4:1-4
Efficient Purification of Mouse Monoclonal Antibodies from Ascites Fluid by Medium-Performance Anion Exchange Chromatography M I C H A E L E. A N N U N Z I A T O and D A N T E J. M A R C I A N I
Medium-performance anion-exchange chromatography was applied to the purification of murine lgG class monoclonal antibodies from ascites fluid. The separations were performed under mild conditions at pH 8 using relatively low sodium-chloride concentrations. Recoveries for monoclonal antibodies of subclasses lgG1, lgG2a, and lgG2b were about 90%. The lgG preparations were free of other ascites fluid proteins.
Since the introduction of monoclonal antibody technology [1] its usage has intensified the need for highly purified preparations in larger quantities. Monoclonal antibody preparations obtained as ascites fluid contain in addition to murine immunoglobulins other mouse proteins, such as albumin and transferrin. A reliable purification procedure should, besides separating the monoclonal antibodies from other proteins, be fast, costeffective, and able to be scaled up. A diversity of procedures is used for the purification including affinity chromatography on protein A [2], chromatography on DEAE Am-Gel Blue [3], immunoaffinity chromatography, and anion-exchange chromatography. However, most of the above mentioned procedures do not meet all the requirements mentioned above. Not all murine immunoglobulins exhibit good binding characteristics toward protein A, especially mouse IgG1, which binds very poorly to this ligand [2]. Moreover, the low pH used to elute immunoglobulins from protein A may in some cases alter the antibody activity. Ion-exchange chromatography on DEAE-cellulose [4] involves a combination of difFrom Cambridge BioScience Corporation, Hopkinton, Massachusetts. Address reprint requests to: Dante J. Marciani, Cambridge BioScience Corporation, 35 South Street, Hopkinton, MA 01748. Accepted July 21, 1986.
ferent steps requiring several hours to perform; a somewhat similar situation is also observed with the DEAE Affi-Gel Blue procedure [3]. Significant improvement in anion exchange chromatographic separations of monoclonal antibodies has been achieved recently using high-performance liquid chromatography (HPLC). Excellent analytical separations of mouse IgG-type monoclonal antibodies have been obtained in less than 1 hr using different commercially available HPLC columns [5, 6]. Unfortunately HPLC preparative columns and the high-pressure instrumentation needed are very expensive, potentially limiting the scale-up of analytical separations. For the above reasons we have investigated the use of newly developed medium-performance anion exchangers, which are of large particle size, easy to pack, and inexpensive [7, 8]. Our results show that DEAE-Toyopearl 650 is an ideal support for one-step chromatographic purification of mouse monoclonal IgG, yielding preparative separations comparable to those obtained with the more expensive HPLC systems.
Materials and Methods
Treatment of Ascites Fluid The ascites fluids were routinely centrifuged at 1000g for 20 min at 4°C to remove cells and debris. The clarified fluids were then used immediately or quick frozen and stored at -80°C for further use. Prior to anion-exchange chromatography the ascites fluids were equilibrated with the starting buffer, 20 mM Tris-HCl, pH 8, by gel filtration on Sephadex G-25 medium (Pharmacia Fine Chemicals, Piscataway, N J). Gel preparation, packing, and elution were performed according to established p r o c e d u r e s [9]. The volume of ascites fluid applied to the Sephadex G-25 was equivalent to 30% of the bed volume or less. The effluent eluted b e t w e e n the void volume (27% bed volume) and 66% of the bed volume contained the ascites fluid proteins and was collected. The final concentration of protein in the sample applied was adjusted to 4-8 mg protein per milliliter of sample. Higher concentrations of protein were avoided to eliminate clogging of the chromatographic column. Total protein content of the pool was determined by the procedure of Lowry [10] using bovine serum albumin as standard or by measurement of absorbance at 280 nm.
© 1987 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017 0735-0651/87/$03.50
~J
f ~<
J
2 1987,Gene Anal Techn 4:1-4
Purification of Monoclonal IgG by Medium-Performance Anion-Exchange Chromatography All chromatographic separations were performed on DEAE-Toyopearl 650S (superfine grade), particle d i a m e t e r 2 5 - 4 4 I~m, available from E. Merck. Glass columns of different sizes and with single end fittings were obtained from BioRad L a b o r a t o r i e s ( R i c h m o n d , CA) and A l t e x Beckman (Berkley, CA). Packing of the columns was carried out by the constant-velocity method [11] at flow rates 20-40% higher than those used during the actual chromatographic separations. The dependence of resolution on the packing flow rate was negligible, as reported previously [12]. Preparative chromatographic separations of IgG monoclonal antibodies on DEAE-Toyopearl 650S were performed with a simple set-up using a peristaltic pump or by adaptation of the Pharmacia F P L C System (Pharmacia Fine Chemicals). In the first case, 0.5 ml of ascites fluid equilibrated with starting buffer and diluted fivefold were applied to a I x 3.5 cm column (BioRad Laboratories) and washed with 5 ml of the same buffer. The different proteins, including IgG, were eluted with a 100-ml, 0-0.3 M linear NaCI gradient in starting buffer at room temperature. The flow rate was maintained at 1.7 ml/min by use of a Microperpex peristaltic pump (LKB Instruments, Gaithersburg, MD), and 3.2-ml fractions were collected. The effluent was constantly monitored at 280 nm using a LKB Uvicord S monitor, and the conductivity of each fraction was measured with a YSI model 32 c o n d u c t i m e t e r . Columns were regenerated by washing with onecolumn volume of 6 M guanidine HCI and reequilibrated with starting buffer. In order to achieve faster flow rates and allow gradient manipulations, the Pharmacia FPLC System was used with DEAE-Toyopearl 650S packed in an Altex glass column (1.5 x 10 cm). The pressure was set so as to not exceed 2 MPa, the m a x i m u m allowable for this Altex glass columns. The separations of some monoclonal antibodies that eluted too close to albumin or transferrin were optimized by the gradient manipulation capacity of the FPLC System. Columns were regenerated with guanidine HCI as mentioned above.
Electrophoretic Analysis Ascites fluid and fractions obtained from chromatographic separations were analyzed by SDS-
polyacrylamide gel electrophoresis (SDS-PAGE) according to the method of Laemmli [13]. A 3.5% stacking gel was used to load samples on a 10% separating gel. Gels were stained with Coomassie blue. In many instances, ascites fluid and/or chromatographic fractions were analyzed by electrophoresis on cellulose acetate and stained with Ponceau Red or Amido Black 10B. Briefly 4 txl of ascites fluid or chromatograhic fractions (about 10-40 Ixg of protein) were applied to cellulose acetate strips, and electrophoresis ran at constant voltage (180V) for 20 min using 0.067 M Tris-barbital-Na barbital, pH 8.8, as electrophoresis buffer. The cellulose acetate electrophoresis was performed at room temperature using a Multiphor unit (LKB Instruments).
Enzyme-Linked Immunoassay Determinations of the different monoclonal IgGs were performed by enzyme-linked (ELISA) using Dynatech microtiter plates coated with rabbit antimouse IgG H + L chain specific (Cappel Laboratories, Cochranville, PA). For coating, the antibody was used at 10 txg/ml in 0.1 M Na bicarbonate, pH 8. Column fractions were diluted in phosphate saline buffer (PBS), pH 7.4, containing 0.2% bovine serum albumin; a fixed volume was added per well and incubated. After washing with water the antibody reactivity was determined with a peroxidase-conjugated rabbit antimouse IgG (Cappel) using 2,2 azino-di-3-ethyl-benzothiazoline sulfonic acid (ABTS) as a substrate. The microtiter plates were read at 405 mm using a Dynatech Microplate Reader MR600. Determinations of specific antibody activities were performed according to Engvall [14]. Dynatech microtiter plates were coated with 200 Ixl antigen (20 txg/ml). The antibody reactivity was determined with a peroxidase-conjugated rabbit antimouse IgG (Cappel). The microtiter plates were read as discussed above. Results Several ascites fluids containing monoclonal antibodies of the IgG class were fractionated in a single step by anion-exchange chromatography on DEAE-Toyopearl 650S. The elution profile for ascites fluid containing monoclonal antibody I, an IgG1 subclass, is shown in Figure 1. Three major protein peaks were obtained and further characterized by electrophoresis and ELISA. The first
© 1987 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
1987, Gene Anal Techn 4:1-4
1.0 L-
A
B
-
C
3"%.
,
~
•
-,
i 0.4 ~)
0.4 Z
0.2 0.2
0
0
20
40
FRACTION
protein peak, eluted between 0.05 and 0.08 M NaCI, consisted of mouse transferrin as shown by SDS-PAGE. Immunochemical analysis of the different fractions identified the IgG nature of the second peak, which eluted between 0.09 and 0.12 M NaCI. The third peak contained mostly albumin, as shown by SDS-PAGE analysis (Figure 1). Depending on the hybridoma cell line, the albumin peak revealed some chromatographic heterogeneity, possibly reflecting differential binding of pigments to albumin. Preparative runs were also performed using the Pharmacia FPLC delivery system in combination with a DEAE-Toyopearl 650S column. The separations obtained for monoclonal antibodies of IgG1 and IgG2a subclasses were similar (not shown). Gradient manipulation with the Pharmacia FPLC system allowed satisfactory separations in cases where partial overlap of ascites fluid proteins occurred. Excellent resolution of an IgG2b monoclonal antibody from multiple albumin peaks was obtained by maintaining the NaC1 concentration at 0.12 M for a determined length of time (Figure 2). Elution of albumin was achieved by increasing the salt concentration from 0.12 to 0.25 M NaC1 in a linear fashion. The nature of each protein peak was confirmed by SDSPAGE. A similar optimization was achieved for the separation of a monoclonal antibody from multiple species of putative transferrin, which under a linear gradient partially coeluted with the IgG peak (not shown). Recovery of the IgG from the ascites fluid was calculated by ELISA of the immunoglobulin peak
60
NUMBER
Figure 1. Purification of monoclonal antibody IgG1 by ion-exan-h-'li-@e'~hromatography c on DEAE-Toyopearl 650S. One-half millileter of ascites fluid containing 20 mg of protein and equilibrated with 20 mM Tris-HC1 buffer, pH 8, was applied to a 3.5 x 1 cm I.D. column. After a wash with 5 ml of the same buffer, elution of the different proteins was achieved with a 120 ml linear gradient of NaC1 from 0 to 0.3M in 20 mM Tris-HCl buffer, pH 8. The flow rate was maintained at 1.7 ml/min with a peristaltic pump, and 3.4-ml fractions were collected. The three major protein peaks were analyzed by SDS-PAGE; A = transferrin peak, B = IgG peak, and C = albumin peak.
and found to be 87-91% of the original ascites IgG. In some instances recovery of specific antibodies was measured by ELISA using antigencoated plates. The recoveries of specific activities were approximately 88%. Analysis of proteases in the purified monoclonal antibody preparations failed to detect any activity. This absence of proteases was reflected in the good stability of the monoclonal antibodies or their conjugates for several months at 4°C. Discussion Our data show that the IgG monoclonal antibody purifications obtained with DEAE-Toyopearl 650S, a medium-performance support, are comparable to those reported with HPLC columns such as mono Q and TSK-DEAE 5PW [15]. The medium-performance supports, which are of large particle size, inexpensive, and easy to pack, have all the characteristics desired for preparative separations. In these studies we have purified IgG monoclonal antibodies from ascites in a prepara-
© 1987 Elsevier Science PublishingCo., Inc., 52 Vanderbilt Ave., New York, NY 10017
_--- 2
4 1987, Gene Anal Techn 4:1-4
~-_._._-.
/-
1.0
i
~
B
C
0.8 0.6 z 0.3
0.4
0.2
< 0.2 00
0.1 15
30 TIME (MIN.)
Figure 2. Purification of monoclonal antibodies by FPLC using a DEAE-Toyopearl 650S column (10 x 1.5 cm I.D.). Separation of a IgG2b from partially overlapping albumin was achieved with the following gradient conditions: 50 ml, 0-0.12 M NaC1, 25 ml of 0.12 M NaCI followed by 50 ml, 0.12-0.25M NaCI. Elution solutions were made with the starting buffer, 20 mM Tris-HCl, pH 8. A flow rate of 2 ml/min was maintained with the Pharmacia FPLC system. The sample was I ml ascites fluid equilibrated and diluted tenfold with starting buffer. Protein fractions were analyzed by SDS-PAGE; S = ascites sample, A = transferrin peak, B = IgG peak, and C = albumin peak.
tive mode, with the total protein applied ranging from 10 to 100 mg, or 5 to 56 mg protein per square centimeter of column cross-sectional area. High sample loadings, i.e., over 40 mg/cm 2, resulted in a negligible decrease of resolution and still yielded high purity IgG. This efficient separation of the IgG from other proteins allowed us to eliminate the need for any gel chromatography steps. In the case where higher sample loadings, i.e. over 60 mg of protein per square centimeter, will be applied to the column, the need for columns of larger diameter should be considered. The performance of this separation procedure has been demonstrated for several subclasses of IgG, such as IgG1, IgG2a, and IgG2b. Recoveries of about 90% have been obtained with each subclass. The possibility of using a peristaltic pump with this chromatographic support will allow the
45
60
0
average laboratory to obtain fast and efficient purifications without the need of expensive HPLC systems.
References 1. Kohler, G., and Milstein, C. (1975) Nature 256, 495-497. 2. Godwin, J. W. (1980) J. Immunol. Methods 39, 285-292. 3. Bruck, C., portetelle, D., Glineur, C., and Bollen, A. (1982) J. Immunol. Methods 53,313-319. 4. Parham, P., Androlewicz, M. J., Brodsky, F. M., Holmes, N. J., and Ways, J. P. (1982) J. Immunol. Methods 53, 133-173. 5. Burchiel, S. W., Billman, J. R., and Alber, T. R. (1984) J. Immunol. Methods 69, 33-42. 6. Stanker, L. H., Vanderlaan, M., and Juarez-Salinas, H. (1985) J. lmmunol. Methods 76, 157-169. 7. Kato, Y., Nakamura, K., and Hashimoto, T. (1982) J. Chromatogr. 245, 193-211. 8. Kato, Y., Nakamura, K., and Hashimoto, T. (1982) J. Chromatogr. 253,219-225. 9. Fischer, L. (1969) An Introduction to Gel Chromatography, North Holland, Amsterdam. 10. Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193,265-275. 11. Kato, Y., Komiya, K., lwaeda, T., Sasaki, H., and Hashimoto, T. (1981) J. Chromatogr. 205, 185-190. 12. Kato, Y., Nakamura, K., and Hashimoto, T. (1983) J. Chromatogr. 256, 143-150. 13. Laemmli, U. K. (1970) Nature 227,680-684. 14. Engvall, E. (1980) Meth. Enzymol. 70, 419-439. 15. Deschamps, J. R., Hildreth, J. E. K., Derr, D., and August, J. T. (1985) Anal. Biochem. 147,451-454.
© 1987 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
Efficient Purification of Mouse Monoclonal Antibodies from Ascites Fluid by Medium-Performance Anion Exchange Chromatography M I C H A E L E. A N N U N Z I A T O and D A N T E J. M A R C I A N I
Medium-performance anion-exchange chromatography was applied to the purification of murine lgG class monoclonal antibodies from ascites fluid. The separations were performed under mild conditions at pH 8 using relatively low sodium-chloride concentrations. Recoveries for monoclonal antibodies of subclasses lgG1, lgG2a, and lgG2b were about 90%. The lgG preparations were free of other ascites fluid proteins.
Since the introduction of monoclonal antibody technology [1] its usage has intensified the need for highly purified preparations in larger quantities. Monoclonal antibody preparations obtained as ascites fluid contain in addition to murine immunoglobulins other mouse proteins, such as albumin and transferrin. A reliable purification procedure should, besides separating the monoclonal antibodies from other proteins, be fast, costeffective, and able to be scaled up. A diversity of procedures is used for the purification including affinity chromatography on protein A [2], chromatography on DEAE Am-Gel Blue [3], immunoaffinity chromatography, and anion-exchange chromatography. However, most of the above mentioned procedures do not meet all the requirements mentioned above. Not all murine immunoglobulins exhibit good binding characteristics toward protein A, especially mouse IgG1, which binds very poorly to this ligand [2]. Moreover, the low pH used to elute immunoglobulins from protein A may in some cases alter the antibody activity. Ion-exchange chromatography on DEAE-cellulose [4] involves a combination of difFrom Cambridge BioScience Corporation, Hopkinton, Massachusetts. Address reprint requests to: Dante J. Marciani, Cambridge BioScience Corporation, 35 South Street, Hopkinton, MA 01748. Accepted July 21, 1986.
ferent steps requiring several hours to perform; a somewhat similar situation is also observed with the DEAE Affi-Gel Blue procedure [3]. Significant improvement in anion exchange chromatographic separations of monoclonal antibodies has been achieved recently using high-performance liquid chromatography (HPLC). Excellent analytical separations of mouse IgG-type monoclonal antibodies have been obtained in less than 1 hr using different commercially available HPLC columns [5, 6]. Unfortunately HPLC preparative columns and the high-pressure instrumentation needed are very expensive, potentially limiting the scale-up of analytical separations. For the above reasons we have investigated the use of newly developed medium-performance anion exchangers, which are of large particle size, easy to pack, and inexpensive [7, 8]. Our results show that DEAE-Toyopearl 650 is an ideal support for one-step chromatographic purification of mouse monoclonal IgG, yielding preparative separations comparable to those obtained with the more expensive HPLC systems.
Materials and Methods
Treatment of Ascites Fluid The ascites fluids were routinely centrifuged at 1000g for 20 min at 4°C to remove cells and debris. The clarified fluids were then used immediately or quick frozen and stored at -80°C for further use. Prior to anion-exchange chromatography the ascites fluids were equilibrated with the starting buffer, 20 mM Tris-HCl, pH 8, by gel filtration on Sephadex G-25 medium (Pharmacia Fine Chemicals, Piscataway, N J). Gel preparation, packing, and elution were performed according to established p r o c e d u r e s [9]. The volume of ascites fluid applied to the Sephadex G-25 was equivalent to 30% of the bed volume or less. The effluent eluted b e t w e e n the void volume (27% bed volume) and 66% of the bed volume contained the ascites fluid proteins and was collected. The final concentration of protein in the sample applied was adjusted to 4-8 mg protein per milliliter of sample. Higher concentrations of protein were avoided to eliminate clogging of the chromatographic column. Total protein content of the pool was determined by the procedure of Lowry [10] using bovine serum albumin as standard or by measurement of absorbance at 280 nm.
© 1987 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017 0735-0651/87/$03.50
~J
f ~<
J
2 1987,Gene Anal Techn 4:1-4
Purification of Monoclonal IgG by Medium-Performance Anion-Exchange Chromatography All chromatographic separations were performed on DEAE-Toyopearl 650S (superfine grade), particle d i a m e t e r 2 5 - 4 4 I~m, available from E. Merck. Glass columns of different sizes and with single end fittings were obtained from BioRad L a b o r a t o r i e s ( R i c h m o n d , CA) and A l t e x Beckman (Berkley, CA). Packing of the columns was carried out by the constant-velocity method [11] at flow rates 20-40% higher than those used during the actual chromatographic separations. The dependence of resolution on the packing flow rate was negligible, as reported previously [12]. Preparative chromatographic separations of IgG monoclonal antibodies on DEAE-Toyopearl 650S were performed with a simple set-up using a peristaltic pump or by adaptation of the Pharmacia F P L C System (Pharmacia Fine Chemicals). In the first case, 0.5 ml of ascites fluid equilibrated with starting buffer and diluted fivefold were applied to a I x 3.5 cm column (BioRad Laboratories) and washed with 5 ml of the same buffer. The different proteins, including IgG, were eluted with a 100-ml, 0-0.3 M linear NaCI gradient in starting buffer at room temperature. The flow rate was maintained at 1.7 ml/min by use of a Microperpex peristaltic pump (LKB Instruments, Gaithersburg, MD), and 3.2-ml fractions were collected. The effluent was constantly monitored at 280 nm using a LKB Uvicord S monitor, and the conductivity of each fraction was measured with a YSI model 32 c o n d u c t i m e t e r . Columns were regenerated by washing with onecolumn volume of 6 M guanidine HCI and reequilibrated with starting buffer. In order to achieve faster flow rates and allow gradient manipulations, the Pharmacia FPLC System was used with DEAE-Toyopearl 650S packed in an Altex glass column (1.5 x 10 cm). The pressure was set so as to not exceed 2 MPa, the m a x i m u m allowable for this Altex glass columns. The separations of some monoclonal antibodies that eluted too close to albumin or transferrin were optimized by the gradient manipulation capacity of the FPLC System. Columns were regenerated with guanidine HCI as mentioned above.
Electrophoretic Analysis Ascites fluid and fractions obtained from chromatographic separations were analyzed by SDS-
polyacrylamide gel electrophoresis (SDS-PAGE) according to the method of Laemmli [13]. A 3.5% stacking gel was used to load samples on a 10% separating gel. Gels were stained with Coomassie blue. In many instances, ascites fluid and/or chromatographic fractions were analyzed by electrophoresis on cellulose acetate and stained with Ponceau Red or Amido Black 10B. Briefly 4 txl of ascites fluid or chromatograhic fractions (about 10-40 Ixg of protein) were applied to cellulose acetate strips, and electrophoresis ran at constant voltage (180V) for 20 min using 0.067 M Tris-barbital-Na barbital, pH 8.8, as electrophoresis buffer. The cellulose acetate electrophoresis was performed at room temperature using a Multiphor unit (LKB Instruments).
Enzyme-Linked Immunoassay Determinations of the different monoclonal IgGs were performed by enzyme-linked (ELISA) using Dynatech microtiter plates coated with rabbit antimouse IgG H + L chain specific (Cappel Laboratories, Cochranville, PA). For coating, the antibody was used at 10 txg/ml in 0.1 M Na bicarbonate, pH 8. Column fractions were diluted in phosphate saline buffer (PBS), pH 7.4, containing 0.2% bovine serum albumin; a fixed volume was added per well and incubated. After washing with water the antibody reactivity was determined with a peroxidase-conjugated rabbit antimouse IgG (Cappel) using 2,2 azino-di-3-ethyl-benzothiazoline sulfonic acid (ABTS) as a substrate. The microtiter plates were read at 405 mm using a Dynatech Microplate Reader MR600. Determinations of specific antibody activities were performed according to Engvall [14]. Dynatech microtiter plates were coated with 200 Ixl antigen (20 txg/ml). The antibody reactivity was determined with a peroxidase-conjugated rabbit antimouse IgG (Cappel). The microtiter plates were read as discussed above. Results Several ascites fluids containing monoclonal antibodies of the IgG class were fractionated in a single step by anion-exchange chromatography on DEAE-Toyopearl 650S. The elution profile for ascites fluid containing monoclonal antibody I, an IgG1 subclass, is shown in Figure 1. Three major protein peaks were obtained and further characterized by electrophoresis and ELISA. The first
© 1987 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
1987, Gene Anal Techn 4:1-4
1.0 L-
A
B
-
C
3"%.
,
~
•
-,
i 0.4 ~)
0.4 Z
0.2 0.2
0
0
20
40
FRACTION
protein peak, eluted between 0.05 and 0.08 M NaCI, consisted of mouse transferrin as shown by SDS-PAGE. Immunochemical analysis of the different fractions identified the IgG nature of the second peak, which eluted between 0.09 and 0.12 M NaCI. The third peak contained mostly albumin, as shown by SDS-PAGE analysis (Figure 1). Depending on the hybridoma cell line, the albumin peak revealed some chromatographic heterogeneity, possibly reflecting differential binding of pigments to albumin. Preparative runs were also performed using the Pharmacia FPLC delivery system in combination with a DEAE-Toyopearl 650S column. The separations obtained for monoclonal antibodies of IgG1 and IgG2a subclasses were similar (not shown). Gradient manipulation with the Pharmacia FPLC system allowed satisfactory separations in cases where partial overlap of ascites fluid proteins occurred. Excellent resolution of an IgG2b monoclonal antibody from multiple albumin peaks was obtained by maintaining the NaC1 concentration at 0.12 M for a determined length of time (Figure 2). Elution of albumin was achieved by increasing the salt concentration from 0.12 to 0.25 M NaC1 in a linear fashion. The nature of each protein peak was confirmed by SDSPAGE. A similar optimization was achieved for the separation of a monoclonal antibody from multiple species of putative transferrin, which under a linear gradient partially coeluted with the IgG peak (not shown). Recovery of the IgG from the ascites fluid was calculated by ELISA of the immunoglobulin peak
60
NUMBER
Figure 1. Purification of monoclonal antibody IgG1 by ion-exan-h-'li-@e'~hromatography c on DEAE-Toyopearl 650S. One-half millileter of ascites fluid containing 20 mg of protein and equilibrated with 20 mM Tris-HC1 buffer, pH 8, was applied to a 3.5 x 1 cm I.D. column. After a wash with 5 ml of the same buffer, elution of the different proteins was achieved with a 120 ml linear gradient of NaC1 from 0 to 0.3M in 20 mM Tris-HCl buffer, pH 8. The flow rate was maintained at 1.7 ml/min with a peristaltic pump, and 3.4-ml fractions were collected. The three major protein peaks were analyzed by SDS-PAGE; A = transferrin peak, B = IgG peak, and C = albumin peak.
and found to be 87-91% of the original ascites IgG. In some instances recovery of specific antibodies was measured by ELISA using antigencoated plates. The recoveries of specific activities were approximately 88%. Analysis of proteases in the purified monoclonal antibody preparations failed to detect any activity. This absence of proteases was reflected in the good stability of the monoclonal antibodies or their conjugates for several months at 4°C. Discussion Our data show that the IgG monoclonal antibody purifications obtained with DEAE-Toyopearl 650S, a medium-performance support, are comparable to those reported with HPLC columns such as mono Q and TSK-DEAE 5PW [15]. The medium-performance supports, which are of large particle size, inexpensive, and easy to pack, have all the characteristics desired for preparative separations. In these studies we have purified IgG monoclonal antibodies from ascites in a prepara-
© 1987 Elsevier Science PublishingCo., Inc., 52 Vanderbilt Ave., New York, NY 10017
_--- 2
4 1987, Gene Anal Techn 4:1-4
~-_._._-.
/-
1.0
i
~
B
C
0.8 0.6 z 0.3
0.4
0.2
< 0.2 00
0.1 15
30 TIME (MIN.)
Figure 2. Purification of monoclonal antibodies by FPLC using a DEAE-Toyopearl 650S column (10 x 1.5 cm I.D.). Separation of a IgG2b from partially overlapping albumin was achieved with the following gradient conditions: 50 ml, 0-0.12 M NaC1, 25 ml of 0.12 M NaCI followed by 50 ml, 0.12-0.25M NaCI. Elution solutions were made with the starting buffer, 20 mM Tris-HCl, pH 8. A flow rate of 2 ml/min was maintained with the Pharmacia FPLC system. The sample was I ml ascites fluid equilibrated and diluted tenfold with starting buffer. Protein fractions were analyzed by SDS-PAGE; S = ascites sample, A = transferrin peak, B = IgG peak, and C = albumin peak.
tive mode, with the total protein applied ranging from 10 to 100 mg, or 5 to 56 mg protein per square centimeter of column cross-sectional area. High sample loadings, i.e., over 40 mg/cm 2, resulted in a negligible decrease of resolution and still yielded high purity IgG. This efficient separation of the IgG from other proteins allowed us to eliminate the need for any gel chromatography steps. In the case where higher sample loadings, i.e. over 60 mg of protein per square centimeter, will be applied to the column, the need for columns of larger diameter should be considered. The performance of this separation procedure has been demonstrated for several subclasses of IgG, such as IgG1, IgG2a, and IgG2b. Recoveries of about 90% have been obtained with each subclass. The possibility of using a peristaltic pump with this chromatographic support will allow the
45
60
0
average laboratory to obtain fast and efficient purifications without the need of expensive HPLC systems.
References 1. Kohler, G., and Milstein, C. (1975) Nature 256, 495-497. 2. Godwin, J. W. (1980) J. Immunol. Methods 39, 285-292. 3. Bruck, C., portetelle, D., Glineur, C., and Bollen, A. (1982) J. Immunol. Methods 53,313-319. 4. Parham, P., Androlewicz, M. J., Brodsky, F. M., Holmes, N. J., and Ways, J. P. (1982) J. Immunol. Methods 53, 133-173. 5. Burchiel, S. W., Billman, J. R., and Alber, T. R. (1984) J. Immunol. Methods 69, 33-42. 6. Stanker, L. H., Vanderlaan, M., and Juarez-Salinas, H. (1985) J. lmmunol. Methods 76, 157-169. 7. Kato, Y., Nakamura, K., and Hashimoto, T. (1982) J. Chromatogr. 245, 193-211. 8. Kato, Y., Nakamura, K., and Hashimoto, T. (1982) J. Chromatogr. 253,219-225. 9. Fischer, L. (1969) An Introduction to Gel Chromatography, North Holland, Amsterdam. 10. Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193,265-275. 11. Kato, Y., Komiya, K., lwaeda, T., Sasaki, H., and Hashimoto, T. (1981) J. Chromatogr. 205, 185-190. 12. Kato, Y., Nakamura, K., and Hashimoto, T. (1983) J. Chromatogr. 256, 143-150. 13. Laemmli, U. K. (1970) Nature 227,680-684. 14. Engvall, E. (1980) Meth. Enzymol. 70, 419-439. 15. Deschamps, J. R., Hildreth, J. E. K., Derr, D., and August, J. T. (1985) Anal. Biochem. 147,451-454.
© 1987 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017