ANALYTICAL
BIOCHEMISTRY
197,
47-51
(1991)
Octadecyl Silica: A Solid Phase for Protein Purification by Immunoadsorption’ Mario Chiong,* Sergio Lavandero,* *Departamento TDepartamento
Rodrigo Ramos,* Juan Carlos Aguillbn,t
de Bioquimica y Biologi’a Molecular, Facultad de Ciencius Quimicas de Medicina Preventiva Animal, Facultad de Ciencias Veterinurias
Immunoaffinity chromatography involves binding of an antigen or antibody to a solid matrix, usually agarose, frequently using the cyanogen bromide method. These methods are laborious, rather expensive, and their use has been mostly restricted to immunopurifications on the microscale. We propose here the use of octadecyl silica (SiCl8) beads, a matrix for HPLC, as an alternative solid phase for protein immunopurification and immunoadsorption. Antibodies or antigens are strongly bound to Sic18 by a simple incubation; radiolabeled antibodies can only be eluted from Sic18 by detergent-containing solutions. After the remaining free binding sites have been saturated with bovine serum albumin, Sic18 is incubated with the antigenor antibody-containing crude preparations and is then poured into a minicolumn. The nonspecifically bound proteins are removed by washing; specific proteins are eluted by disruption of the antigen-antibody complexes with a low pH buffer. With this methodology, we have obtained high purity preparations of proteins in single steps, even when these proteins are present in trace amounts (picograms) in a complex mixture such as human serum. Similarly, specific antibodies against an intracellular parasite (!&ypunosoma cruzi) were completely absorbed from human serum with Sic18 coated with parasite antigens. o 1991 Academic PMMI, IIIC.
The unique high specificity and affinity of antibodies for antigens makes immunoaffinity chromatography a valuable method for the isolation of any protein for which specific antibodies are available. Briefly, the antibodies immobilized on a solid support are used seleci This work supported by Grants 88198 UNDP/WORLD BANK/ WHO Special Programme for Research and Training in Tropical Diseases; 191188 FONDECYT CHILE; M2901-8813 DTI, Universidad de Chile; Swedish Agency for Research and Cooperation (SAREC). S.L. and J.C.A. are fellows of Fundacion Andes-Chile and CONICYTChile, respectively. Correspondence should be addressed to M. Chiong, Casilla 233, Santiago 1, Chile. 0003-2697191
Copyright All
rights
$3.00 1991 by of reproduction
0
Academic
Press, in any form
and Arturo Ferreira**t
y Farmace’uticas, Universidud de Chile, and y Pecuarias, Universidud de Chile
tively to adsorb antigen from a complex protein mixture. Contaminant molecules can be removed by washing, and the antigen is then eluted from the immunoabsorbent under conditions that disrupt the antigen-antibody complex (1,2). We describe here the use of octadecyl silica (SiC18)’ as an alternative solid support for protein immunochromatography and immunoadsorption. Binding of proteins to Sic18 requires only incubation for a few hours. Because other polypeptides (including antibodies) cannot be readily eluted from Sic& the method is generally applicable for immunopurifications and immunoadsorptions. MATERIALS
AND
METHODS
Reagents
Goat anti-rabbit immunoglobulin G (IgG), goat antihuman IgG (both reagents affinity-purified) and octadecyl silica (particle size 15-40 brn) were obtained from Sigma Chemical Co. (St. Louis, MO). Radioactive iodine ([ 1251]) was purchased from the Chilean Nuclear Energy Commission. Radiolabeling of proteins was performed by the Iodogen method (3). Polyvinyl chloride (PVC) microtiter plates were obtained from Falcon, Becton Dickinson Co. (Oxnard, CA). Homogeneous potato apyrase and rabbit antiapyrase serum were a kind gift of M. Mantilla (4,5). Recombinant human (Ytumor necrosis factor (aTNF) was generously donated by Genentecht (San Francisco, CA). Human serum, positive and negative for the presence of antibodies (IgG) against Trypanosornu cruzi (by immunoradiometric and immuno-Western-blotting assays), were obtained from the blood bank, J. J. Aguirre University Hospital. A whole sonicated extract of T. cruzi was obtained from ’ Abbreviations used: SiC18, octadecyl silica; IgG, immunoglobulin G; PVC, polyvinyl chloride; aTNF, recombinant human a tumor necrosis factor; PBS, phosphate-buffered saline; BSA, bovine serum albumin; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis. 47
Inc. reserved.
48
CHIONG
ET
AL. TABLE
Effect of Buffer
2
Composition on the Antigen-Antibody Complex Coupled to Sic18 Percentage residual radioactivity
Treatments
=
c 0
6
12
TIME
(HOURS)
4
16
20
FIG. 1.
Kinetics of radiolabeled protein binding to SiC18. [1261]IgG was incubated with Sic18 previously treated (close circles) or nontreated (open circles) with methanol. See Materials and Methods for details. Results represent the average of three experiments and bars indicate the standard deviations.
cultured forms (epimastigote) of the Tulahuen strain, kindly donated by Dr. Antonio Morello, Department of Biochemistry, Faculty of Medicine, University of Chile. All other common reagents were of analytical grade. Assessment of the Conditions for Immunoafinity Chromatography Five milligrams of Sic18 was suspended in 100 ~1 of phosphate-buffered saline (PBS: 10 InM sodium phos-
TABLE Effect
of Different
Buffers
0.1 M Sodium phosphate, 0.15 M NaCl, pH 7.3 0.1 M Sodium phosphate, 1.00 M NaCl, pH 7.3 0.1 M Sodium phosphate, 0.15 M NaCl, 10% dioxan, pH 7.3 0.15 M NH,OH/HCl, 0.15 M NaCl, pH 0.1 M Glycine, 0.15 M NaCl, pH 2.6 None
99.8 f 1.2 80.3 + 1.2
11.0
95.2 + 0.8 89.7 + 2.8 49.1 + 5.2 100
Note. Apyrase-Sic18 was incubated with antiapyrase rabbit serum and then washed with PBS. The pellets were resuspended in 100 ~1 of different buffers, shaken 30 min, and centrifuged. After washing the pellets with PBS, the Silica beads were incubated with [“‘I]-labeled goat anti-rabbit for 2 h, centrifuged, washed, and counted. See Materials and Methods for details. Results are shown as the average of three experiments + standard deviations.
phate, 0.15 M NaCl, pH 7.3) or 100 ~1 of methanol in l&ml polypropylene tubes. After the contents of the tubes had been mixed vigorously for 60 s, 900 ~1 of PBS -labeled affinity-purified containing 2 X lo6 cpm of [ 1251] goat anti-rabbit IgG was added. The tubes were shaken thoroughly at 12°C. At different times, the incubation was stopped by centrifugation for 1 min at 16,000g. The pellets, washed three times with 1 ml PBS, and super-
1
on Protein-Sic18
Treatments 0.1 M Sodium phosphate, 0.15 M NaCl, pH 1.3 0.1 M Sodium phosphate, 1.00 M NaCl, pH 1.3 0.1 M Sodium phosphate, 0.15 M NaCl, 10% (v/v) methanol, pH 7.3 0.1 M Sodium phosphate, 0.15 M NaCl, 20% dioxan, pH 7.3 0.1 M Sodium phosphate, 0.15 M NaCl, 0.05% (v/v) NP-40, pH 7.3 0.1 M Glycine, 0.15 M NaCl, pH 2.6 0.2 M Sodium borate, 0.15 M NaCl, 0.05% (v/v) NP-40, pH 8.3 0.15 M NH,OH/HCl, 0.15 M NaCl, pH 11.0 None
Binding
Percentage residual radioactivity
99.2 f 3.3 99.1 f 0.7 95.3 t 2.5 98.9 f 1.1 43.5 k 8.7 99.5 + 1.5 35.2 + 9.3 98.8 + 0.8 100
Note. [‘2SI]-labeled anti-rabbit IgG was incubated 18 h at 12°C with methanol-treated Sic18 and then washed with PBS by centrifugation. The pellets were resuspended with 1 ml of different buffers and, after 20 min of shaking, the samples were centrifuged and supernatants counted. Results are shown as the average of three experiments + standard deviations.
FIG. 2. Autoradiography of immuno-Western blots using SiClSimmunopurified antibodies. Lane 1 contains a pure apyrase standard, and lane 2 a crude protein extract containing apyrase. (A) Blot developed with a polyspecific rabbit anti-apyrase antiserum; (B) Blot developed with a monospecific antibody affinity-purified by the Sic18 method. See Materials and Methods for details.
OCTADECYL
SILICA
IN
IMMUNOADSORPTIVE
PROTEIN
49
PURIFICATION
apyrase rabbit immunoglobulins was added. After being shaken overnight at lZ”C, the suspension was poured into a l-ml tuberculin syringe plugged with glass wool. The bed volume of this column was 200 ~1. The column was washed with 2 ml of PBS and nonspecifically bound proteins were removed by washing with 1 ml PBS-5% dioxan followed by 1 ml PBS. The specifically bound antibodies were eluted with 1 ml 0.1 M glycine, 0.15 M NaCl, pH 2.6, and collected in two 1.5-ml polypropylene tubes (0.5 ml each) containing 20 ~1 2 M Tris-HCl, pH 8.0. Immunoafinity
FIG. 3. Immunopurification of apyrase from a crude protein extract. The figure shows a 12% polyacrylamide-SDS gel stained with Coomassie blue. Lane A: molecular weight standards, lane B: pure apyrase standard, lane C: crude protein extract containing apyrase, lane D: apyrase immunopurified by the Sic18 method.
natants were counted. The stability of the radioactive IgG attachment to Sic18 was assessed by resuspending the pellets with 1 ml of different buffers. After 20 min of shaking at 12”C, the samples were centrifuged, and the pellets and supernatants counted. The choice to perform batchwise experiments was influenced by the fact that columns packed with the matrix have poor flows. Apyrase was bound to 5 mg of Sic18 and the remaining binding sites were saturated with BSA (see below). The beads were incubated for 2 h at 12°C with IgG, an ammonium sulfate precipitate (O-45%) from antiapyrase rabbit serum (40 pg/ml in PBS 1% (w/v)-BSA). After centrifugation at 4°C the pellets were washed twice with PBS. In order to disrupt the complex, the pellets were resuspended in 100 ~1 of buffers with different disrupting conditions, shaken for 30 min, centrifuged for 1 min, washed, resuspended in 1 ml PBS-l% BSA (containing lo5 cpm of [‘251]-labeled goat anti-rabbit IgG), shaken for 2 h, centrifuged, washed three times with PBS, and counted. Immunoafinity
Purification
of Antibody
One hundred milligrams of Sic18 was suspended with vigorous mixing in 150 ~1 of methanol. After a 5min incubation at room temperature, 100 pugof apyrase was added in a volume of 50 ~1 followed by 950 ~1 of PBS and shaken for 6 h at 12°C. The pellet obtained after centrifugation for 3 min in a benchtop centrifuge was resuspended in 3 ml of PBSl%BSA and shaken for 6 h at 12°C. The suspension was centrifuged and the pellet resuspended in 6 ml of PBS, l%BSA, and 400 pg of (NH,),SO,-precipitated anti-
Purification
of Antigen
Five hundred milligrams of Sic18 was suspended in 800 ~1 of methanol and, after incubation at room temperature for 5 min, 500 pg of ammonium sulfate-precipitated antiapyrase IgG was added. The volume was then adjusted to 6 ml with PBS. After being shaken for 6 h at 12’C, the beads were quantitatively recovered bycentrifugation, resuspended, and incubated for 6 h in 6 ml, of PBS-l% BSA. The suspension was then centrifuged for 3 min and the pellet resuspended in 6 ml of PBS containing lo-30 mg of potato protein extract containing the antigen. After being shaken overnight, the beads were poured into a 5-ml tuberculin syringe plugged with glass wool. The column was washed successively with 2 ml PBS, 1 ml PBS-5% dioxan and 1 ml PBS. The antigen was then eluted with 1 ml 0.1 M glycine, 0.15 M NaCl, pH 2.6, and collected in a 1.5-ml polypropylene pH 8.0. The tube containing 40 ~1 of 2 M Tris-HCl, eluates were analyzed by SDS-PAGE (6) and immunoWestern blotting (7) followed by radioautography. Specific Immunoabsorption Sera
of Antibodies
from
Whole
Fifty milligrams of Sic18 was suspended in 100 ~1 of methanol, incubated with 50 pg of a sonicated extract of T. cruzi in a final volume of 1 ml for 6 h at 4°C; after the centrifugation the pellets were washed five times, 15 min each time, with 1 ml PBS-l%BSA and separated by centrifugation for 1 min. The final pellets were resuspended in l-ml portions of human serum (positive and negative for T. cruzi, diluted l/2000 in PBS-0.01% BSA) and shaken overnight. The Sic18 was removed by centrifugation for 1 min and the supernatants were analyzed by immunoradiometric assay (8,9). RESULTS
AND
DISCUSSION
Eighteen-carbon lipid chains, chemically bound to silanol groups of silica particles (octadecyl silica (SiCl8)), form the commonly used stationary phase for HPLC. This material has a strong binding capacity for molecules with hydrophobic domains, such as proteins which interact with nonpolar groups attached to the solid ma-
50
CHIONG
FIG. 4. Capture of trace levels (pg) of a protein from human serum. Autoradiography of a 20% SDS-Polyacrylamide gel. Lane 1: human serum containing 2 pg/ml of recombinant [“‘I]-oTNF, lane 2: supernatant of human serum after incubation with anti-crTNF rabbit antibodies coupled to Sic& lane 3: fraction eluted from the Sic18 beads after treatment with 0.1 M glycine, 0.15 M NaCl, pH 2.6.
trix (10). Bound proteins can be differentially separated by reverse-phase hydrophobic chromatography (11). Making use of this property, we have recently described the use of Sic18 as an alternative solid phase in immunoradiometric assays (8). All of the immunoadsorbents presently available for immunoaffinity chromatography are prepared by covalently coupling antibodies or antigens to a solid matrix, usually agarose. The matrix is usually previously activated with cyanogen bromide or a similar reagent (1,2,12). Although substantially improved, these methods are still laborious, time consuming, rather expensive, and difficult to handle on the microscale level (13). Optimal conditions for protein binding to Sic18 were determined by using a [‘251]-labeled anti-rabbit IgG. Maximum binding of [‘251]-labeled IgG was practically achieved after 6 h of incubation. When Sic18 was pretreated with methanol the binding capacity was increased twice compared to that obtained in the absence of the organic solvent (Fig. 1). However, when methanol was added at the end, an enhanced binding capacity was not observed, suggesting that the solvent interacts with the Sic18 rather than with the protein. Conditions that increase hydrophobic interactions, such as high ionic strength and changes in the solvent
ET
AL.
dielectric constant (ll), were examined. In all cases, the resulting binding was similar to that obtained by a simple pretreatment of the Sic18 with methanol (results not shown). One hundred milligrams of Sic18 can potentially bind about 3.0 mg of protein (8). We have found that, for antigen purification, the optimal amount of immunoglobulin is about 300400 pg. On the other hand, for antibody immunopurification, 50-100 pg pure antigen is adequate. Higher levels of protein binding to solid matrixes may not be beneficial since some interference by steric blocking of antigen-antibody binding sites may occur (1). The stability of the SiC18-protein complex was studied by incubating SiC18-[1251]-labeled IgG with different buffers and determining the residual radioactivity. Extreme pH conditions, ionic strengths up to 1 M NaCl, and dielectric constant alterations did not affect the protein binding to the solid matrix. However, detergentcontaining buffers eluted 60-65% of the total SiClS-associated radioactivity (Table 1). These results support the proposal that the protein-Sic18 interaction is, to a great extent, of hydrophobic nature. We cannot discount, however, the possibility that unblocked silanol groups may play a certain role in the binding capacity of SiC18. To test conditions for disruption of the complex, apyrase was coupled to Sic18 and incubated with antiapyrase antibodies in PBS. The resulting complexes were
4c
3c
70 r; I a 0
2c
1c
POSITIVE SERA
NEGATIVE
SERA
FIG. 5. Removal of anti-l: cruzi antibodies from positive and negative serum. Sic18 coated with T. cruzi antigens were used to absorb the corresponding antibodies from human sera. The remaining activity present in the supernatant was measured by immunoradiometric assay. The experiment was performed in sextuplicate with two positive and two negative sera. Slashed bars: sera treated in the absence of SiCl8. Closed bars: sera treated with T. cruai-coated SiCl8. Open bars: sera treated with BSA-coated SiCl8. Lines represent the standard deviations.
OCTADECYL
SILICA
IN
IMMUNOADSORPTIVE
subjected to a single batch treatment with high ionic strength and changes in pH and dielectric constant (2). The most effective treatment was 0.1 M glycine, 0.15 M NaCl, pH 2.6 (Table 2). To assess the nonspecific binding, rabbit antibodies bound to Sic18 were incubated with [‘251]-labeled antirabbit IgG and then poured into a column and washed as described below. After low pH treatment, less than 4% of the total bound radioactivity remains adsorbed to the column. For antibody purification, apyrase-Sic18 was incubated with a polyclonal, polyspecific, rabbit antiapyrase immunoglobulin precipitated with ammonium sulfate. The beads were poured into a column and nonspecifically bound proteins were removed by washing. The specific antibodies were quantitatively eluted with 0.1 M glycine, 0.15 M NaCl, pH2.6, and assayed by immunoradiametric assay against the pure antigen (results not shown). In a crude extract containing apyrase, the polyspecific serum, as analyzed by immuno-Western blotting, recognized several proteins (Fig. 2A, lane 2) some of them different from those present in the purified apyrase standard (Fig. 2A, lane 1). However, the affinity-purified immunoglobulins recognized only the apyrase isozymes, all contaminant antibodies being eliminated (Fig. 2B, lanes 1 and 2). For immunopurification of antigen, the specific antibodies were coupled to Sic18 and, by a procedure similar to that described above, were used to purify apyrase in a single step (Fig. 3). The nature of the immunopurified protein was demonstrated by immunoradiometric assay and immuno-Western-blot analysis. Sic18 coupled with both antigens and antibodies can be reused at least three times without losing its binding capacity (results not shown). Purification of antigens present in trace levels in complex protein mixtures was also attempted; 10 pg of [ lz51]-labeled recombinant L~TNF (specific activity 2 X lo7 cpm/Fg) was added to 5 ml of human serum. The radioactive antigen was captured with rabbit aTNF antibody precipitated with (NH,),SO, coupled to SiC18. The bound proteins were eluted with low pH buffer and analyzed by SDS-PAGE, followed by autoradiography. Most of the [‘251]-labeled aTNF present in the serum (>90%) was recovered as judged by autoradiography (Fig. 4). The versatility of this solid phase is further demonstrated by the specific removal of anti-T. cruzi antibodies present in human serum. Thus, Sic18 coated with epimastigote proteins was able specifically to adsorb
PROTEIN
51
PURIFICATION
these antibodies; albumin-coated Sic18 consistently failed to do so (Fig. 5). This simple adsorption system may be extremely useful for a quick and easy removal of unwanted antigens or antibodies from complex protein mixtures. Moreover, Sic18 may be a practical solid phase to adsorb antigens present at low concentrations in biological fluids (i.e., urine), particularly if the matrix is coupled to specific antibodies. We have shown that Sic18 may be used as a multipurpose matrix. It efficiently binds proteins from different sources (vegetable, protozoan, and mammal) and with different molecular weights. In all these casesthe bound proteins could be immunodetected. The possibility exists however, that a protein with a predominance of hydrophilic domains may bind less strongly, a problem that could decrease the overall performance of the system. The system should therefore be characterized for each particular protein. ACKNOWLEDGMENTS We are indebted to Professor Dr. Mario Sapag for support and helpful advice and to Mrs. Patricia Vallejos for excellent technical assistance. We also thank Dr. C. I. Pogson and Professor Fresia Perez for the critical reading of the manuscript.
REFERENCES 1. Livingston,
D. M. (1974) W. B., and Wilckek, M., Press, New York.
2. Mayes,
mun.
in Enzymology. 34, pp. 723-731,
E. L. V. (1984) in Methods in Molecular (Walker, J. M. Ed.), pp. 13-20, Humana,
Proteins
3. Fraker,
in Methods Eds.), Vol.
P. J., and Speck, 80,849-857.
J. C. (1978)
Biochem.
4. Fanta,
N., Anich, M., Mantilla, M., Kettlun, M. A., and Traverso Cori, A. (1988) Arch. 129-133.
(Jakoby, Academic
Biology, Vol. New Jersey.
Biophys.
1.
Res. Com-
A. M., Valenzuela, Biol. Med. Exp. 21,
5. Mantilla,
M., Kettlun, A. M., Valenzuela, M. A., and Traverso Cori, A. (1984) Phytochemistry 23,1397-1400. Laemmli, U. K. (1970) Nature (London) 227, 680-685.
6. 7. Towbin,
Sci. USA
H., Stahelin, 76,4350-4354.
T., and Gordon,
J. (1979)
Proc.
Natl.
8. Lavandero, 9.
S., and Ferreira, A. (1988) J. Immunol. Methods 261-265. Catt, K., and Treagar, G. W. (1967) Science X8,1570-1572.
Acad.
114,
10. Hancock, W. S. (1985) in Handbook of HPLC for the Separation of Aminoacids, Peptides and Proteins (Hancock, W. S., Ed.), p. 3, CRC Press, Boca Raton, FL. 11. Shaltiel, S. (1974) in Methods in Enzymology (Jakoby, W. B., and Wilckek, M., Eds.), Vol. 34, pp. 126-140, Academic Press, New York. 12. Ax&n, R., Porath, J., and Ernback, S. (1967) Nature (London)
214,1302-1304. 13. March, &em.
S. C., Parikh,
60, 149-152.
I., and
Cuatrecasas,
P. (1974)
Anal.
Bio-
BIOCHEMISTRY
197,
47-51
(1991)
Octadecyl Silica: A Solid Phase for Protein Purification by Immunoadsorption’ Mario Chiong,* Sergio Lavandero,* *Departamento TDepartamento
Rodrigo Ramos,* Juan Carlos Aguillbn,t
de Bioquimica y Biologi’a Molecular, Facultad de Ciencius Quimicas de Medicina Preventiva Animal, Facultad de Ciencias Veterinurias
Immunoaffinity chromatography involves binding of an antigen or antibody to a solid matrix, usually agarose, frequently using the cyanogen bromide method. These methods are laborious, rather expensive, and their use has been mostly restricted to immunopurifications on the microscale. We propose here the use of octadecyl silica (SiCl8) beads, a matrix for HPLC, as an alternative solid phase for protein immunopurification and immunoadsorption. Antibodies or antigens are strongly bound to Sic18 by a simple incubation; radiolabeled antibodies can only be eluted from Sic18 by detergent-containing solutions. After the remaining free binding sites have been saturated with bovine serum albumin, Sic18 is incubated with the antigenor antibody-containing crude preparations and is then poured into a minicolumn. The nonspecifically bound proteins are removed by washing; specific proteins are eluted by disruption of the antigen-antibody complexes with a low pH buffer. With this methodology, we have obtained high purity preparations of proteins in single steps, even when these proteins are present in trace amounts (picograms) in a complex mixture such as human serum. Similarly, specific antibodies against an intracellular parasite (!&ypunosoma cruzi) were completely absorbed from human serum with Sic18 coated with parasite antigens. o 1991 Academic PMMI, IIIC.
The unique high specificity and affinity of antibodies for antigens makes immunoaffinity chromatography a valuable method for the isolation of any protein for which specific antibodies are available. Briefly, the antibodies immobilized on a solid support are used seleci This work supported by Grants 88198 UNDP/WORLD BANK/ WHO Special Programme for Research and Training in Tropical Diseases; 191188 FONDECYT CHILE; M2901-8813 DTI, Universidad de Chile; Swedish Agency for Research and Cooperation (SAREC). S.L. and J.C.A. are fellows of Fundacion Andes-Chile and CONICYTChile, respectively. Correspondence should be addressed to M. Chiong, Casilla 233, Santiago 1, Chile. 0003-2697191
Copyright All
rights
$3.00 1991 by of reproduction
0
Academic
Press, in any form
and Arturo Ferreira**t
y Farmace’uticas, Universidud de Chile, and y Pecuarias, Universidud de Chile
tively to adsorb antigen from a complex protein mixture. Contaminant molecules can be removed by washing, and the antigen is then eluted from the immunoabsorbent under conditions that disrupt the antigen-antibody complex (1,2). We describe here the use of octadecyl silica (SiC18)’ as an alternative solid support for protein immunochromatography and immunoadsorption. Binding of proteins to Sic18 requires only incubation for a few hours. Because other polypeptides (including antibodies) cannot be readily eluted from Sic& the method is generally applicable for immunopurifications and immunoadsorptions. MATERIALS
AND
METHODS
Reagents
Goat anti-rabbit immunoglobulin G (IgG), goat antihuman IgG (both reagents affinity-purified) and octadecyl silica (particle size 15-40 brn) were obtained from Sigma Chemical Co. (St. Louis, MO). Radioactive iodine ([ 1251]) was purchased from the Chilean Nuclear Energy Commission. Radiolabeling of proteins was performed by the Iodogen method (3). Polyvinyl chloride (PVC) microtiter plates were obtained from Falcon, Becton Dickinson Co. (Oxnard, CA). Homogeneous potato apyrase and rabbit antiapyrase serum were a kind gift of M. Mantilla (4,5). Recombinant human (Ytumor necrosis factor (aTNF) was generously donated by Genentecht (San Francisco, CA). Human serum, positive and negative for the presence of antibodies (IgG) against Trypanosornu cruzi (by immunoradiometric and immuno-Western-blotting assays), were obtained from the blood bank, J. J. Aguirre University Hospital. A whole sonicated extract of T. cruzi was obtained from ’ Abbreviations used: SiC18, octadecyl silica; IgG, immunoglobulin G; PVC, polyvinyl chloride; aTNF, recombinant human a tumor necrosis factor; PBS, phosphate-buffered saline; BSA, bovine serum albumin; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis. 47
Inc. reserved.
48
CHIONG
ET
AL. TABLE
Effect of Buffer
2
Composition on the Antigen-Antibody Complex Coupled to Sic18 Percentage residual radioactivity
Treatments
=
c 0
6
12
TIME
(HOURS)
4
16
20
FIG. 1.
Kinetics of radiolabeled protein binding to SiC18. [1261]IgG was incubated with Sic18 previously treated (close circles) or nontreated (open circles) with methanol. See Materials and Methods for details. Results represent the average of three experiments and bars indicate the standard deviations.
cultured forms (epimastigote) of the Tulahuen strain, kindly donated by Dr. Antonio Morello, Department of Biochemistry, Faculty of Medicine, University of Chile. All other common reagents were of analytical grade. Assessment of the Conditions for Immunoafinity Chromatography Five milligrams of Sic18 was suspended in 100 ~1 of phosphate-buffered saline (PBS: 10 InM sodium phos-
TABLE Effect
of Different
Buffers
0.1 M Sodium phosphate, 0.15 M NaCl, pH 7.3 0.1 M Sodium phosphate, 1.00 M NaCl, pH 7.3 0.1 M Sodium phosphate, 0.15 M NaCl, 10% dioxan, pH 7.3 0.15 M NH,OH/HCl, 0.15 M NaCl, pH 0.1 M Glycine, 0.15 M NaCl, pH 2.6 None
99.8 f 1.2 80.3 + 1.2
11.0
95.2 + 0.8 89.7 + 2.8 49.1 + 5.2 100
Note. Apyrase-Sic18 was incubated with antiapyrase rabbit serum and then washed with PBS. The pellets were resuspended in 100 ~1 of different buffers, shaken 30 min, and centrifuged. After washing the pellets with PBS, the Silica beads were incubated with [“‘I]-labeled goat anti-rabbit for 2 h, centrifuged, washed, and counted. See Materials and Methods for details. Results are shown as the average of three experiments + standard deviations.
phate, 0.15 M NaCl, pH 7.3) or 100 ~1 of methanol in l&ml polypropylene tubes. After the contents of the tubes had been mixed vigorously for 60 s, 900 ~1 of PBS -labeled affinity-purified containing 2 X lo6 cpm of [ 1251] goat anti-rabbit IgG was added. The tubes were shaken thoroughly at 12°C. At different times, the incubation was stopped by centrifugation for 1 min at 16,000g. The pellets, washed three times with 1 ml PBS, and super-
1
on Protein-Sic18
Treatments 0.1 M Sodium phosphate, 0.15 M NaCl, pH 1.3 0.1 M Sodium phosphate, 1.00 M NaCl, pH 1.3 0.1 M Sodium phosphate, 0.15 M NaCl, 10% (v/v) methanol, pH 7.3 0.1 M Sodium phosphate, 0.15 M NaCl, 20% dioxan, pH 7.3 0.1 M Sodium phosphate, 0.15 M NaCl, 0.05% (v/v) NP-40, pH 7.3 0.1 M Glycine, 0.15 M NaCl, pH 2.6 0.2 M Sodium borate, 0.15 M NaCl, 0.05% (v/v) NP-40, pH 8.3 0.15 M NH,OH/HCl, 0.15 M NaCl, pH 11.0 None
Binding
Percentage residual radioactivity
99.2 f 3.3 99.1 f 0.7 95.3 t 2.5 98.9 f 1.1 43.5 k 8.7 99.5 + 1.5 35.2 + 9.3 98.8 + 0.8 100
Note. [‘2SI]-labeled anti-rabbit IgG was incubated 18 h at 12°C with methanol-treated Sic18 and then washed with PBS by centrifugation. The pellets were resuspended with 1 ml of different buffers and, after 20 min of shaking, the samples were centrifuged and supernatants counted. Results are shown as the average of three experiments + standard deviations.
FIG. 2. Autoradiography of immuno-Western blots using SiClSimmunopurified antibodies. Lane 1 contains a pure apyrase standard, and lane 2 a crude protein extract containing apyrase. (A) Blot developed with a polyspecific rabbit anti-apyrase antiserum; (B) Blot developed with a monospecific antibody affinity-purified by the Sic18 method. See Materials and Methods for details.
OCTADECYL
SILICA
IN
IMMUNOADSORPTIVE
PROTEIN
49
PURIFICATION
apyrase rabbit immunoglobulins was added. After being shaken overnight at lZ”C, the suspension was poured into a l-ml tuberculin syringe plugged with glass wool. The bed volume of this column was 200 ~1. The column was washed with 2 ml of PBS and nonspecifically bound proteins were removed by washing with 1 ml PBS-5% dioxan followed by 1 ml PBS. The specifically bound antibodies were eluted with 1 ml 0.1 M glycine, 0.15 M NaCl, pH 2.6, and collected in two 1.5-ml polypropylene tubes (0.5 ml each) containing 20 ~1 2 M Tris-HCl, pH 8.0. Immunoafinity
FIG. 3. Immunopurification of apyrase from a crude protein extract. The figure shows a 12% polyacrylamide-SDS gel stained with Coomassie blue. Lane A: molecular weight standards, lane B: pure apyrase standard, lane C: crude protein extract containing apyrase, lane D: apyrase immunopurified by the Sic18 method.
natants were counted. The stability of the radioactive IgG attachment to Sic18 was assessed by resuspending the pellets with 1 ml of different buffers. After 20 min of shaking at 12”C, the samples were centrifuged, and the pellets and supernatants counted. The choice to perform batchwise experiments was influenced by the fact that columns packed with the matrix have poor flows. Apyrase was bound to 5 mg of Sic18 and the remaining binding sites were saturated with BSA (see below). The beads were incubated for 2 h at 12°C with IgG, an ammonium sulfate precipitate (O-45%) from antiapyrase rabbit serum (40 pg/ml in PBS 1% (w/v)-BSA). After centrifugation at 4°C the pellets were washed twice with PBS. In order to disrupt the complex, the pellets were resuspended in 100 ~1 of buffers with different disrupting conditions, shaken for 30 min, centrifuged for 1 min, washed, resuspended in 1 ml PBS-l% BSA (containing lo5 cpm of [‘251]-labeled goat anti-rabbit IgG), shaken for 2 h, centrifuged, washed three times with PBS, and counted. Immunoafinity
Purification
of Antibody
One hundred milligrams of Sic18 was suspended with vigorous mixing in 150 ~1 of methanol. After a 5min incubation at room temperature, 100 pugof apyrase was added in a volume of 50 ~1 followed by 950 ~1 of PBS and shaken for 6 h at 12°C. The pellet obtained after centrifugation for 3 min in a benchtop centrifuge was resuspended in 3 ml of PBSl%BSA and shaken for 6 h at 12°C. The suspension was centrifuged and the pellet resuspended in 6 ml of PBS, l%BSA, and 400 pg of (NH,),SO,-precipitated anti-
Purification
of Antigen
Five hundred milligrams of Sic18 was suspended in 800 ~1 of methanol and, after incubation at room temperature for 5 min, 500 pg of ammonium sulfate-precipitated antiapyrase IgG was added. The volume was then adjusted to 6 ml with PBS. After being shaken for 6 h at 12’C, the beads were quantitatively recovered bycentrifugation, resuspended, and incubated for 6 h in 6 ml, of PBS-l% BSA. The suspension was then centrifuged for 3 min and the pellet resuspended in 6 ml of PBS containing lo-30 mg of potato protein extract containing the antigen. After being shaken overnight, the beads were poured into a 5-ml tuberculin syringe plugged with glass wool. The column was washed successively with 2 ml PBS, 1 ml PBS-5% dioxan and 1 ml PBS. The antigen was then eluted with 1 ml 0.1 M glycine, 0.15 M NaCl, pH 2.6, and collected in a 1.5-ml polypropylene pH 8.0. The tube containing 40 ~1 of 2 M Tris-HCl, eluates were analyzed by SDS-PAGE (6) and immunoWestern blotting (7) followed by radioautography. Specific Immunoabsorption Sera
of Antibodies
from
Whole
Fifty milligrams of Sic18 was suspended in 100 ~1 of methanol, incubated with 50 pg of a sonicated extract of T. cruzi in a final volume of 1 ml for 6 h at 4°C; after the centrifugation the pellets were washed five times, 15 min each time, with 1 ml PBS-l%BSA and separated by centrifugation for 1 min. The final pellets were resuspended in l-ml portions of human serum (positive and negative for T. cruzi, diluted l/2000 in PBS-0.01% BSA) and shaken overnight. The Sic18 was removed by centrifugation for 1 min and the supernatants were analyzed by immunoradiometric assay (8,9). RESULTS
AND
DISCUSSION
Eighteen-carbon lipid chains, chemically bound to silanol groups of silica particles (octadecyl silica (SiCl8)), form the commonly used stationary phase for HPLC. This material has a strong binding capacity for molecules with hydrophobic domains, such as proteins which interact with nonpolar groups attached to the solid ma-
50
CHIONG
FIG. 4. Capture of trace levels (pg) of a protein from human serum. Autoradiography of a 20% SDS-Polyacrylamide gel. Lane 1: human serum containing 2 pg/ml of recombinant [“‘I]-oTNF, lane 2: supernatant of human serum after incubation with anti-crTNF rabbit antibodies coupled to Sic& lane 3: fraction eluted from the Sic18 beads after treatment with 0.1 M glycine, 0.15 M NaCl, pH 2.6.
trix (10). Bound proteins can be differentially separated by reverse-phase hydrophobic chromatography (11). Making use of this property, we have recently described the use of Sic18 as an alternative solid phase in immunoradiometric assays (8). All of the immunoadsorbents presently available for immunoaffinity chromatography are prepared by covalently coupling antibodies or antigens to a solid matrix, usually agarose. The matrix is usually previously activated with cyanogen bromide or a similar reagent (1,2,12). Although substantially improved, these methods are still laborious, time consuming, rather expensive, and difficult to handle on the microscale level (13). Optimal conditions for protein binding to Sic18 were determined by using a [‘251]-labeled anti-rabbit IgG. Maximum binding of [‘251]-labeled IgG was practically achieved after 6 h of incubation. When Sic18 was pretreated with methanol the binding capacity was increased twice compared to that obtained in the absence of the organic solvent (Fig. 1). However, when methanol was added at the end, an enhanced binding capacity was not observed, suggesting that the solvent interacts with the Sic18 rather than with the protein. Conditions that increase hydrophobic interactions, such as high ionic strength and changes in the solvent
ET
AL.
dielectric constant (ll), were examined. In all cases, the resulting binding was similar to that obtained by a simple pretreatment of the Sic18 with methanol (results not shown). One hundred milligrams of Sic18 can potentially bind about 3.0 mg of protein (8). We have found that, for antigen purification, the optimal amount of immunoglobulin is about 300400 pg. On the other hand, for antibody immunopurification, 50-100 pg pure antigen is adequate. Higher levels of protein binding to solid matrixes may not be beneficial since some interference by steric blocking of antigen-antibody binding sites may occur (1). The stability of the SiC18-protein complex was studied by incubating SiC18-[1251]-labeled IgG with different buffers and determining the residual radioactivity. Extreme pH conditions, ionic strengths up to 1 M NaCl, and dielectric constant alterations did not affect the protein binding to the solid matrix. However, detergentcontaining buffers eluted 60-65% of the total SiClS-associated radioactivity (Table 1). These results support the proposal that the protein-Sic18 interaction is, to a great extent, of hydrophobic nature. We cannot discount, however, the possibility that unblocked silanol groups may play a certain role in the binding capacity of SiC18. To test conditions for disruption of the complex, apyrase was coupled to Sic18 and incubated with antiapyrase antibodies in PBS. The resulting complexes were
4c
3c
70 r; I a 0
2c
1c
POSITIVE SERA
NEGATIVE
SERA
FIG. 5. Removal of anti-l: cruzi antibodies from positive and negative serum. Sic18 coated with T. cruzi antigens were used to absorb the corresponding antibodies from human sera. The remaining activity present in the supernatant was measured by immunoradiometric assay. The experiment was performed in sextuplicate with two positive and two negative sera. Slashed bars: sera treated in the absence of SiCl8. Closed bars: sera treated with T. cruai-coated SiCl8. Open bars: sera treated with BSA-coated SiCl8. Lines represent the standard deviations.
OCTADECYL
SILICA
IN
IMMUNOADSORPTIVE
subjected to a single batch treatment with high ionic strength and changes in pH and dielectric constant (2). The most effective treatment was 0.1 M glycine, 0.15 M NaCl, pH 2.6 (Table 2). To assess the nonspecific binding, rabbit antibodies bound to Sic18 were incubated with [‘251]-labeled antirabbit IgG and then poured into a column and washed as described below. After low pH treatment, less than 4% of the total bound radioactivity remains adsorbed to the column. For antibody purification, apyrase-Sic18 was incubated with a polyclonal, polyspecific, rabbit antiapyrase immunoglobulin precipitated with ammonium sulfate. The beads were poured into a column and nonspecifically bound proteins were removed by washing. The specific antibodies were quantitatively eluted with 0.1 M glycine, 0.15 M NaCl, pH2.6, and assayed by immunoradiametric assay against the pure antigen (results not shown). In a crude extract containing apyrase, the polyspecific serum, as analyzed by immuno-Western blotting, recognized several proteins (Fig. 2A, lane 2) some of them different from those present in the purified apyrase standard (Fig. 2A, lane 1). However, the affinity-purified immunoglobulins recognized only the apyrase isozymes, all contaminant antibodies being eliminated (Fig. 2B, lanes 1 and 2). For immunopurification of antigen, the specific antibodies were coupled to Sic18 and, by a procedure similar to that described above, were used to purify apyrase in a single step (Fig. 3). The nature of the immunopurified protein was demonstrated by immunoradiometric assay and immuno-Western-blot analysis. Sic18 coupled with both antigens and antibodies can be reused at least three times without losing its binding capacity (results not shown). Purification of antigens present in trace levels in complex protein mixtures was also attempted; 10 pg of [ lz51]-labeled recombinant L~TNF (specific activity 2 X lo7 cpm/Fg) was added to 5 ml of human serum. The radioactive antigen was captured with rabbit aTNF antibody precipitated with (NH,),SO, coupled to SiC18. The bound proteins were eluted with low pH buffer and analyzed by SDS-PAGE, followed by autoradiography. Most of the [‘251]-labeled aTNF present in the serum (>90%) was recovered as judged by autoradiography (Fig. 4). The versatility of this solid phase is further demonstrated by the specific removal of anti-T. cruzi antibodies present in human serum. Thus, Sic18 coated with epimastigote proteins was able specifically to adsorb
PROTEIN
51
PURIFICATION
these antibodies; albumin-coated Sic18 consistently failed to do so (Fig. 5). This simple adsorption system may be extremely useful for a quick and easy removal of unwanted antigens or antibodies from complex protein mixtures. Moreover, Sic18 may be a practical solid phase to adsorb antigens present at low concentrations in biological fluids (i.e., urine), particularly if the matrix is coupled to specific antibodies. We have shown that Sic18 may be used as a multipurpose matrix. It efficiently binds proteins from different sources (vegetable, protozoan, and mammal) and with different molecular weights. In all these casesthe bound proteins could be immunodetected. The possibility exists however, that a protein with a predominance of hydrophilic domains may bind less strongly, a problem that could decrease the overall performance of the system. The system should therefore be characterized for each particular protein. ACKNOWLEDGMENTS We are indebted to Professor Dr. Mario Sapag for support and helpful advice and to Mrs. Patricia Vallejos for excellent technical assistance. We also thank Dr. C. I. Pogson and Professor Fresia Perez for the critical reading of the manuscript.
REFERENCES 1. Livingston,
D. M. (1974) W. B., and Wilckek, M., Press, New York.
2. Mayes,
mun.
in Enzymology. 34, pp. 723-731,
E. L. V. (1984) in Methods in Molecular (Walker, J. M. Ed.), pp. 13-20, Humana,
Proteins
3. Fraker,
in Methods Eds.), Vol.
P. J., and Speck, 80,849-857.
J. C. (1978)
Biochem.
4. Fanta,
N., Anich, M., Mantilla, M., Kettlun, M. A., and Traverso Cori, A. (1988) Arch. 129-133.
(Jakoby, Academic
Biology, Vol. New Jersey.
Biophys.
1.
Res. Com-
A. M., Valenzuela, Biol. Med. Exp. 21,
5. Mantilla,
M., Kettlun, A. M., Valenzuela, M. A., and Traverso Cori, A. (1984) Phytochemistry 23,1397-1400. Laemmli, U. K. (1970) Nature (London) 227, 680-685.
6. 7. Towbin,
Sci. USA
H., Stahelin, 76,4350-4354.
T., and Gordon,
J. (1979)
Proc.
Natl.
8. Lavandero, 9.
S., and Ferreira, A. (1988) J. Immunol. Methods 261-265. Catt, K., and Treagar, G. W. (1967) Science X8,1570-1572.
Acad.
114,
10. Hancock, W. S. (1985) in Handbook of HPLC for the Separation of Aminoacids, Peptides and Proteins (Hancock, W. S., Ed.), p. 3, CRC Press, Boca Raton, FL. 11. Shaltiel, S. (1974) in Methods in Enzymology (Jakoby, W. B., and Wilckek, M., Eds.), Vol. 34, pp. 126-140, Academic Press, New York. 12. Ax&n, R., Porath, J., and Ernback, S. (1967) Nature (London)
214,1302-1304. 13. March, &em.
S. C., Parikh,
60, 149-152.
I., and
Cuatrecasas,
P. (1974)
Anal.
Bio-
