Improving the quality of immunoblots by chromatography of polyclonal antisera on keratin affinity columns

ANALYTICALBIOCHEMISTRY

182,193-196

(I9891

Improving the Quality of lmmunoblots by Chromatography of Polyclonal Antisera on Keratin Affinity Columns Jean-Antoine

Girault,‘,”

Fred S. Gorelick,

and Paul Greengard

Laboratory of Molecular and Cellular Neuroscience, The Rockefeller Box 296,123O York Avenue, New York, New York 10021

Received

May

lo,1989

Unwanted reactivity of polyclonal antisera against keratins (“fingerprint proteins”) is a problem commonly encountered when proteins transferred to nitrocellulose are studied by immunoblotting. Immunoreactivity against keratins is generally accompanied by a spotted background. This antikeratin immunoreactivity could be removed by adsorption of the antisera to human keratin bound to nitrocellulose. Larger amounts of antisera were purified from contaminant antikeratin antibodies by a single passage over a column of human keratin coupled to activated CH-Sepharose 4B. In contrast to nonpurified antisera and their IgG fractions, the column effluent no longer recognized the M, 55,000-70,000 keratin proteins and exhibited a marked decrease in background labeling. We propose this simple method as a valuable alternative when affinity purification of polyclonal antisera on antigen columns is not practical. 0 1989 Academic PWS, 1~.

Epidermal keratins represent a family of water-insoluble proteins that form intermediate filaments in epiderma1 cells (1). Polyclonal rabbit antisera raised against various antigens often contain antibodies against keratins. On immunoblots these antibodies label immunoreactive bands in the M, range between 55,000 and 70,000, generally accompanied by spotted background artifacts. In addition to being aesthetically unpleasant, these artifacts interfere with the study of low abundance proteins,

especially

University,

when they have MT values in the range of

I Present address: INSERM Ull4, College de France, 75005, Paris, France. ’ To whom correspondence should be addressed. 3 Present address: Gastrointestinal Research Laboratory, West Haven Veterans’ Administration Hospital, West Haven, CT 06516. 0003-2697/89 $3.00 Copyright 0 1989 by Academic Press, Inc. All rights of reproduction in any form reserved.

55,000-70,000. High background may also prevent accurate quantification of the amount of antibody bound to the antigen of interest. In the present report we describe a simple procedure to eliminate antikeratin antibodies from antisera, using an affinity column prepared with human keratin. This procedure dramatically decreases the background staining of immunoblots. MATERIALS

AND

METHODS

Materials

Chemicals were obtained from the following sources: phenylmethylsulfonyl fluoride (PMSF),* Calbiochem; leupeptin and pepstatin A, Chemicon; aprotinin (Trasylol), Mobay Chemical Corp.; [lz51]iodoprotein A, Amersham; Nonidet-P40, Particle Data Laboratories, Ltd.; bicinchoninic acid (BCA) protein assay reagent, Pierce; acrylamide, Serva; activated CH-Sepharose 4B and protein A-Sepharose, Pharmacia; nitrocellulose membranes (pore size 0.2 pm), Schleicher & Schuell; Aquatide II, Calbiochem; nonfat dry milk, Carnation Co. Rabbit antiserum G134 was raised against the a-subunit of Ca2+/calmodulin-dependent protein kinase II purified from rat brain (2). Rabbit antiserum G 153, raised against purified bovine ARPP-16 (CAMP-regulated phosphoprotein, M, = 16,000) (3), was a gift of A. Horiuchi. Rabbit antiserum G 187, raised against protein phosphatase inhibitor-l prepared from rabbit skeletal muscle (4), was a gift of H. C. Hemmings, Jr. and A. C. Nairn. The IgG fractions were prepared by affinity chromatography of these sera on a protein A-Sepharose column (Pharmacia). [32P]Phosphosynapsin was a gift of A. J. Czernik. 4 Abbreviations used: PMSF, phenylmethylsulfonyl fluoride; BCA, bicinchoninic acid; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; ARPP-16, CAMP-regulated phosphoprotein, M, = 16,000. 193

194

GIRAULT,

GORELICK,

AND

GREENGARD

Preparation of Brain Extracts and Transfer to Nitrocellulose

A

Male Sprague-Dawley rats were stunned and decapitated. The cerebral cortex and the caudate-putamen were dissected on ice, frozen, and stored at -70°C. Samples were homogenized by sonication in 2% (w/v) boiling sodium dodecyl sulfate (SDS) and the homogenate was boiled in a water bath for 10 min. Protein concentration in the homogenate was measured with BCA reagent (5), using bovine serum albumin as a standard and adjusted to 2.5 mg/ml. “Stop solution” containing a trace of pyronine Y was added to obtain the final concentrations of 3% (w/v) SDS, 62 InM Tris-HCl, pH 6.8, 5% (v/v) glycerol, 0.3 M ,&mercaptoethanol, and 2 mg/ml protein. After boiling for 2 min, 50-~1 aliquots of samples (100 pg protein) were separated by polyacrylamide gel electrophoresis in the presence of SDS (SDS-PAGE) (6). Proteins were transferred electrophoretically to nitrocellulose membranes (7) at 200 mA for 12 h in a buffer (pH 8.6) containing 12.5 mM Tris-base, 96 mM glycine, and 20% (v/v) methanol. Immunoblotting Immunoblotting was performed as described previously (8). Following transfer, the membranes were fixed in a mixture of isopropanol (lo%, v/v) and acetic acid (lo%, v/v) in water and washed thoroughly with water. Excess protein binding sites were quenched in 100 ml TBS buffer (200 mM NaCl, 0.02% (w/v) NaN3, 50 mM Tris-Cl, pH 7.4) containing 5% (w/v) nonfat dry milk (blocking solution) at room temperature for 1 h (9). Membranes were incubated for 2 h in blocking solution containing a l/500 to l/200 dilution of various antibodies and washed in blocking solution. Antibodies were radiolabeled by incubation with [1251]iodoprotein A (0.1 &i/ml). Membranes were then washed extensively in TBS and dried. Autoradiography was performed by exposing Kodak X-Omat AR films at -70°C using DuPont Lightning Plus intensifying screens. Affinity Adsorption of Antikeratin Antibodies Epidermis Keratin Bound to Nitrocellulose

Using

Human epidermis keratin from the hands (2 mg) was prepared as described (l), solubilized in boiling stop solution and subjected to SDS-PAGE on a 9% (w/v) polyacrylamide gel (6). Proteins were transferred to nitrocellulose membranes and stained with 0.1% (w/v) amido black. A 6 X 0.5-cm nitrocellulose strip corresponding to the stained proteins of IV, 55,000-70,000 was excised, incubated in blocking solution for 2 h, and used for the adsorption of antibodies (10). Rabbit antiserum (10 ~1) containing 20 mM benzamidine and 0.01% (w/v) NaN3 was incubated with the nitrocellulose strip at room temperature for 12 h with continuous shaking before being

B

I

I SERUM

6134

FIG. 1. Effects on the quality of immunoblots of antiserum preadsorption on keratin bound to nitrocellulose. Extracts of rat cerebral cortex were prepared as described under Materials and Methods. Aliquots of the homogenate (100 pg protein) were separated by SDSPAGE (9% polyacrylamide, w/v) and transferred electrophoretically to nitrocellulose membranes. Individual lanes were cut from the membrane and each piece was incubated with antiserum (G134) against the a-subunit of Ca*+/calmodulin-dependent protein kinase (l/500, lane A), with serum preadsorbed on keratin bound to nitrocellulose strips (lane B), or with the eluate from these nitrocellulose strips (lane C). Labeling was performed with [‘251]iodoprotein A and immunoreactive bands were revealed by autoradiography. The marker indicates the position of the a-subunit of Ca’+/calmodulin-dependent protein kinase.

used for immunoblotting. Bound antibodies were eluted by immersion of the nitrocellulose strips in 0.1 mM acetic acid for 15 s and the eluate was immediately brought to pH 8 with 1 M Tris-Cl. Preparation

of a Keratin

Afinity

Column

Keratins are highly insoluble proteins which require 8 urea to remain in solution. Since the presence of urea was found not to affect significantly the efficiency of protein coupling to activated CH-Sepharose 4B, as assessed with a radiolabeled tracer protein ( [32P]phosphosynapsin, data not shown), the following procedure was used. Human epidermis keratin (Sigma, 10 mg) in solution in 1.2 ml of 8 M urea, 50 mM Tris, 0.1% (w/v) NaN3 was added to 14 ml of buffer A (0.5 M NaCl, 8 M urea, and 0.1 M NaHC03, pH 8.0). The volume of this solution was reduced to 1 ml by centrifugation in a Centriprep 30 (Amicon) at 2000g. This dilution-concentration procedure was repeated twice to avoid the presence of Tris during the coupling reaction. The coupling reaction was carried out according to the manufacturer’s instructions. Activated CH-Sepharose 4B (400 mg) was suspended M

CHROMATOGRAPHIC

9*

PURIFICATION

OF

ANTISERA

ON

E

F

KERATIN

195

COLUMNS

G

H

I

96-

‘; (3 Iii

55-

3 2

3629-

i 2

19-

6 x

14-

SERUM

G153

SERUM

G167

FIG. 2. Effects on the quality of immunoblots of antisera preadsorption on a keratin affinity column. Extracts of rat caudate-putamen were prepared as described under Materials and Methods. Ahquots of the homogenate (100 pg protein) were separated by SDS-PAGE (13% polyacrylamide, w/v) and transferred electrophoretically to nitrocellulose membranes. Pairs of lanes were cut from the membranes and each piece was incubated with a different preparation of antibodies as indicated below. Labeling was performed with [iz51]iodoprotein A and immunoreactive bands were revealed by autoradiography. Serum G153: (A) crude serum (l/500); (B) IgG fraction; (C) serum passed over the keratin affinity column (l/500); (D) same (l/200). Serum G153 raised against purified bovine ARPP-16, a protein enriched in the caudate-putamen, is known to react with two closely related antigens of M, 16,000 and A4,19,000, respectively (3). Serum G187: (E) crude serum (l/500); (F) IgG fraction; (G) serum passed over the keratin affinity column (l/500); (H) same (l/200); (I) antibodies eluted from the keratin affinity column. Serum G13.7, raised against purified rabbit protein phosphatase inhibitor-l, reacted specifically with rat inhibitor-l (M, 29,000).

and washed in 1 mM HCl. The resin was then briefly washed in coupling buffer (NaCl, 0.5 M, NaHC03, 0.1 M, pH 8.0) and resuspended in 1 ml of the same buffer. Keratin (10 mg in 1 ml of buffer A) was added to the resin and the mixture was rotated end-over-end for 10 h at 4°C. Blocking of the remaining active groups was achieved by incubation in 0.5 M NaCI, 0.1 M Tris-Cl, pH 9. The resin was poured into a column (1 X 5 cm) and washed three times with 5 ml of 0.5 M NaCl, 0.1 M formic acid, pH 4, alternating with 5 ml of 0.5 M NaCI, 0.1 M Tris-Cl, pH 9. The resin was then washed with 10 ml of sodium phosphate buffer (0.1 M, pH 7.4) containing 0.02% (w/v) NaN3 and stored in the same buffer at 4°C. Purification

of Antisera

on the Keratin

Affinity

Column

Before use, the column was equilibrated with 10 ml of Burridge’s buffer (0.15 M NaCl, 0.05% (v/v) Tween 20, 50 mM Tris, pH 7.4). Immune serum (5 ml) containing EDTA (5 mM final), PMSF (1 mM final), and leupeptin (10 pg/ml final) was loaded onto the column. The flowthrough was collected and stored in frozen aliquots. The column was washed with 10 ml of borate-buffered saline (1 M NaCI, 0.1% Tween 20, 25 mM Na borate, pH 8.3) and bound antibodies were eluted with 10 ml of 4.5 M MgC& (adjusted to pH 6.0 with 5 N NaOH). The eluate was diluted 1:2 in Tris-buffered saline, dialyzed extensively against the same buffer, concentrated by placing the dialysis bag in Aquacide II, and tested for the presence of antibodies by immunoblotting. The column was

washed with 10 ml of Burridge’s solution, 10 ml of 0.1 M glycine-HCl, pH 2.5, and 10 ml of borate-buffered saline with Tween 20 and stored in Burridge’s solution. RESULTS

AND

DISCUSSION

The efficiency of affinity removal of antikeratin antibodies from serum using keratin bound to nitrocellulose was tested with a rabbit antiserum against the a-subunit of rat brain Ca’+/calmodulin-dependent protein kinase II (G134). Although the crude antiserum reacted strongly with the a-subunit of the enzyme (iW, 50,000) present in homogenates of rat cortex, it also labeled species of M, 55,000-70,000 and produced a spotted background (Fig. 1A). Following affinity adsorption of 10 ~1 of this antiserum on human epidermis keratin bound to nitrocellulose, staining of the a-subunit persisted, while staining of the M, 55,000-70,000 species and the background spots were markedly reduced (Fig. 1B). Antibodies eluted from the keratin bound to nitrocellulose strips labeled the M, 55,000-70,000 species but not the a-subunit (Fig. lC), confirming that the contaminant antigen recognized in the M, 55,000-70,000 range was keratin. Although the keratin-covered nitrocellulose strips could be regenerated for repeated affinity adsorption by incubation in 0.1 mM acetic acid, no more than 10 ~1 of antiserum could be cleared at a time. To process large amounts of antisera a procedure using a keratin affinity column was developed. The effi-

196

GIRAULT,

GORELICK,

ciency of the purification of antisera on the keratin affinity column was tested with two rabbit antisera against two different antigens. These antisera gave, in addition to the specific immunoreactivity toward the antigens of interest, strong labeling of bands in the M, 55,000-70,000 range corresponding to keratins (Figs. 2A and 2E) and a spotted background. The antibodies against keratins and the activity responsible for the spotted background could not be separated from the immunoglobulins specific for the antigens of interest by chromatography of the antisera on a protein A affinity column (Figs. 2B and ZF). In contrast, both the immunoreactive M, 55,000-70,000 proteins and the background artifacts disappeared when antiserum passed over the keratin affinity column (i.e., the effluent) was used (Figs. ZC, ZD, ZG, and ZH). In the case of the serum against protein phosphatase inhibitor-l, a minor decrease in the intensity of the specific signal was observed with the “purified” serum (Fig. 2G). However, the intensity of specific staining could be increased by the addition of larger amounts of purified serum without increasing the background staining (Fig. 2H). Both the immunoreactivity against keratin and the activity responsible for the background artifacts were absent from preimmune serum (data not shown) and were recovered by elution from the keratin bound to nitrocellulose and from the keratin affinity column (Figs. 1C and 21), indicating that they were probably due to the same antibodies. Therefore the likely explanation for the punctuate, scattered artifacts, which usually accompany the strong antikeratin immunoreactivity, is the contamination of transfer buffers and/or nitrocellulose membranes by keratin derived from human epithelium. Purification of antisera is optimally performed by chromatography on columns of immobilized antigen. However, this approach is not possible if the antigen is only available in limiting amounts or if elution of the antibodies without denaturation is impossible. As demonstrated in the present study, preadsorption of poly-

AND

GREENGARD

clonal antibodies on a human keratin affinity column provides a simple and inexpensive way of improving the quality of immunolabeling of proteins transferred onto membranes. It will be of interest to determine whether this procedure will also be useful in decreasing artifacts in other immunological methods in which unwanted reactivity against keratin might interfere, such as immunocytochemistry or solid phase immunoassays. ACKNOWLEDGMENTS We thank Drs. H. C. Hemmings, Jr., A. C. Nairn, and A. Horiuchi for kindly providing some of the antisera used in this study. We are also grateful to Dr. A. J. Czernik for the gift of 32P-labeled synapsin I and Mrs. G. Bertuzzi for excellent technical help. J.-A.G. was the recipient of a fellowship from Institut National de la Sante et de la Recherche Medicale (France) and F.S.G. was the recipient of a Morton Grossman Research Award. This work was supported by United States Public Health Service Grants MH-40899 (P.G.) and DM31506 (F.S.G.).

REFERENCES 1. Eichner, R., Bonitz, P., and Sun, T. T. (1984) J. Cell Biol. 98, 1388-1396. 2. Gorelick, F. S., Wang, J. K. T., Lai, Y., Nairn, A. C., and Greengard, P. (1988) J. Biol. Chem. 263,17,209-17,212. 3. Horiuchi, A., Nairn, A. C., and Greengard, P. (1987) Sot. Neurosci. Abstr. 13,901. [Abstract] 4. Cohen, P., Foulkes, J. G., Holmes, C. F., Nimmo, G. A., and Tonks, N. K. (1988) in Methods in Enzymology (Corbin, J. D., and Johnson, R. A., Eds.), Vol. 159, pp. 427-437, Academic Press, San Diego, CA. 5. Smith, P. K., Krohn, R. I., Hermanson, G. T., Mallia, A. K., Gartner, F. H., Provenzano, M. D., Fujimoto, E. K., Goeke, N. M., Olson, B. J., and Klenk, D. C. (1985) And. Biochm. 150,76-85. 6. Laemmli,

U. K. (1970)

7. Towbin, H., Staehlin, Sci. USA 76,4350-4354.

Nature

(London)

T., and

Gordon,

227,680-685. J. (1979)

Proc.

Natl.

J.-A., Raisman-Vozari, R., Agid, Y., and Greengard, 8. Girault, (1989) Proc. Natl. Acad. Sci. USA 86,2493-2497. D. A., Gautsch, J. W., Sportsman, J. R., and Elder, 9. Johnson, (1984) Gene Anal. Tech. 1,3-8. 10. Olmsted,

J. B. (1981)

J. Biol.

Chem.

256,

11,955-11,957.

Acad. P. J. H.