Determination of membrane antigens by a covalent crosslinking method with monoclonal antibodies

ANALYTICAL BIOCHEMISTRY 160,483-488

Determination

(1987)

of Membrane Antigens by a Covalent Method with Monoclonal Antibodies

Crosslinking

HIROFLJMIHAMADAANDTAKASHITSURUO' Cancer Chemotherapy Center, Japanese Foundation for Cancer Research, Kami-Ikebukuro, Toshima-ku, Tokyo 170, Japan Received June 16, 1986 Monoclonal antibodies that recognize cell surface proteins may serve as very useful tools for the study of the biological functions of membrane proteins. However, solubilization of the antigens with detergents may lead to major conformational changes of the protein, making their determination with monoclonal antibodies by immune blot or ordinary immunoprecipitation methods difficult. This is especially evident when the monoclonal antibodies recognize tertiary structures of the proteins in the membrane. We have generated two monoclonal antibodies which are specific for the cell surface antigens of multidrug-resistant human cell lines. However, the antigens of both monoclonal antibodies were difficult to detect by either immune blot or ordinary immunoprecipitation methods. We used a cleavable crosslinking reagent dithiobis(succinimidy1 propionate) to covalently link the monoclonal antibody with its antigenie determinant in the membrane of intact cells. By this method, we were able to detect the antigens for these two monoclonal antibodies following solubilization, immunoprecipitation, and analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. This method should have wide applicability in determination of membrane antigens recognized by monoclonal antibodies when immune blot or ordinary immunoprecipitation methods are not successful. 0 1987 Academic Press, Inc. KEY WORDS: monoclonal antibodies; protein determination; immunoprecipitation; membrane solubilization; detergents; cancer chemotherapy.

The biochemical natures of cell surface antigen proteins recognized by monoclonal antibodies are usually determined either by immune blotting (Western blotting) or by immunoprecipitation (1). As antigen proteins are denatured by sodium dodecyl sulfate (SDSP in Western blotting, monoclonal antibodies against the native protein may or may not recognize the denatured product. Similarly, in the immunoprecipitation

method, membrane proteins must be solubilized before antigen determination. By far the most widely used detergents in membrane solubilization are Triton X-100 or Nonidet-P40 (1). They are relatively mild detergents and usually do not denature the antigenic determinants recognized by monoclonal antibodies. However, even when relatively mild detergents such as Triton X-100 are used, the binding of the detergents may cause a significant conformational change in the antigen (2). Especially when monoclonal antibodies recognize the tertiary structure of the native protein in the membrane, they cannot recognize the antigen protein which is denatured by solubilization, resulting in failure of antigen determination by the standard immunoprecipitation method.

’ Author to whom correspondence and reprint requests should be addressed. 2 Abbreviations used: SDS, sodium dodecyl sulfate; PAGE, polyactylamide gel electrophoresis; DSP, dithio bis(succinimidy1 propionate); PBS, phosphate-buffered saline (0.02 M sodium phosphate-O. 15 M NaCI, pH 7.4); DMSO, dimethyl sulfoxide; sulfobetaine 14, 3-[tetradecyl dimethylammoniol-I-propanesulfonate.

483

0003-2697187 $3.00 Copyright 0 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

484

HAMADA

AND TSLJRUO

In an effort to elucidate the mechanisms of multidrug resistance in tumor cells, we have generated two monoclonal antibodies by immunization of mice with an intact human leukemia cell line resistant to adriamycin (K562/ADM) (3,4). These monoclonal antibodies, designated MRKI 6 and MRK17, showed unique effects on K562/ADM cells: MRK 16 affected vincristine transport of K562/ADM cells, whereas MRK17 suppressed the growth of K562/ADM cells (5). By means of cell binding radioimmunoassay and indirect immunolluorescent staining, these antibodies were shown to recognize antigenic determinants on the surface of intact cells (5). The determination of the antigens by Western blot analysis or standard immunoprecipitation using Nonidet-P40 as a detergent was not successful. During trials of identification of the antigens of these monoclonal antibodies, we have used a cleavable crosslinking reagent, dithiobis(succinimidyl propionate) (DSP), to covalently link a monoclonal antibody with its antigenie determinant in the membrane of intact K562/ADM cells. By this method, followed by solubilization and immunoprecipitation, we have successfully determined the antigens recognized by these two monoclonal antibodies. MATERIALS

AND METHODS

Cell lines. Multidrug-resistant human leukemia cell line K562/ADM was established in our laboratory (3,4). The cells were prepared and maintained as described previously (4). Antibodies. Monoclonal antibodies used were MRKI 6 (IgG*, isotype) and MRK17 (IgG, isotype), which were obtained as mice ascites fluid as previously described (5,6). The monoclonal antibodies used in this experiment were purified by precipitation with ammonium sulfate and affinity chromatography on Protein A-Sepharose CL-4B (Pharmacia) as described previously (7). Pro-

tein measurement of the purified monoclonal antibody solution was done by Lowry’s method (8). Biosynthetic labeling and standard immunoprecipitation. Biosynthetic incorporation of [14C]leucine and immunoprecipitation were carried out as described ( 1,9). Tens of millions of K562/ADM cells were labeled by metabolic incorporation of 60 PCi of L-[U-‘4C]leucine (New England Nuclear, 344.0 mCi/mmol) per 30 ml of leucine-free RPM1 1640 medium containing 10% dialyzed fetal bovine serum for 16 h at 37 “C in a CO2 incubator. The labeled cells were suspended at 5 X lo6 and solubilized in 2 ml of buffer A (20 mM Tris-HCl, pH 8.0, 140 IIIM NaCl, 0.1 InM phenylmethylsulfonyl fluoride) containing 1.0% Nonidet-P40 for 30 min at 4°C and the solution was clarified by centrifugation at 10,OOOg for 10 min. Standard immunoprecipitations were carried out by incubating the cell extracts (2 ml containing the lysate from 5 X lo6 cells) with 5 ~1 of solution of monoclonal antibody containing 20 pg protein for 2 h at 4°C. Control precipitations were done with 20 ~1 of normal mouse serum containing approximately 100 pg IgG protein. Then 200 ~1 of Protein ASepharose CL-4B (Pharmacia) suspension (25% by volume, in buffer A) was added. After incubation for 30 min at 4°C with constant mixing, the precipitates were washed five times with 10 ml of buffer A containing 0.2% Nonidet-P40, then resuspended in 100 ~1 of Laemmli sample buffer (lo), boiled for 5 min, and centrifuged. The supernatants were analyzed by SDS-PAGE using 5-l 5% linear gradient gels. The gels were placed in Amplify (Amersham) for 30 min and dried. Fluorograms were exposed to Kodak XAR-5 film for 2 days at -70°C and developed. The molecular mass markers obtained from Amersham were myosin, 200,000; phosphorylase b, 92,500; bovine serum albumin, 69,000; ovalbumin 46,000; carbonic anhydrase, 30,000; and lysozyme, 14,300.

ANTIGEN

DETERMINATION

BY A CROSSLINKING

Chemical crosslinking and modified immunoprecipitation. Chemical crosslinking experiments were carried out by the procedure of Brenner et al. (11) with some modifications. Radiolabeled K562/ADM cells (above) were suspended at 5 X lo6 in 2 ml of PBS containing 20 pg of monoclonal antibody. After incubation for 2 h at 4°C the cells were washed with PBS and suspended in 2 ml of crosslinking buffer (PBS containing 1 mM MgClz, 0.02% sodium azide buffered to pH 8.3 with 0.1 N NaOH). A crosslinking reagent, DSP (Pierce Chemical), was dissolved at 12.5-100 mM in DMSO and added to the cell suspensions at a final concentration of 0.125-l mM. The reaction was allowed to occur for 1 h at 23°C with occasional mixing. Then the cells were washed with quenching buffer (100 mM Tris-HCl, pH 8.0, 140 mM NaCl), and solubilized with 2 ml of buffer A containing either 1% Nonidet-P40, 1.5% octyl glucoside (Sigma), or 2% sulfobetaine 14 (Calbiochem-Behring). After incubation for 30 min at 4°C the solution was clarified by centrifugation at 10,OOOg for 20 min. Then 200 ~1 of Protein A-Sepharose CL-4B suspension (25% by volume, in buffer A) was added to the supernatants and they were incubated for 30 min at 4°C with constant mixing. The precipitates were washed five times with 10 ml of buffer A containing either 0.2% NonidetP40, 0.3% octyl glucoside, or 0.2% sulfobetaine 14. The final Protein A-Sepharose-antibody-antigen complexes were resuspended in 100 ~1 of Laemmli sample buffer containing 5% 2-mercaptoethanol (10). The disulfide bonds of the antibody-antigen complexes introduced by DSP were cleaved under these reducing conditions. The suspensions were boiled for 5 min, and then centrifuged. The supernatants were analyzed by SDS-PAGE as above. In each experiment, a fraction of the labeled cells was mock-crosslinked by the addition of the same volume of DMSO.

METHOD

485

RESULTS

To determine the antigens recognized by MRK16 and MRK17, standard immunoprecipitation of [ “C]leucine-labeled K562/ ADM cells was done with Nonidet-P40 for solubilization. However, no antigenic protein band was detected (data not shown). Western blot analysis also yielded no antigenie protein band (data not shown). We attempted to covalently crosslink the monoclonal antibody with the antigen in the intact cell membrane, followed by solubilization and immunoprecipitation. Figure 1 shows

FIG. 1. Immunoprecipitation of [“‘C]leucine-labeled K562/ADM cells with monoclonal antibody MRK16 using various concentrations of a crosslinking reagent, DSP. Labeled K562/ADM cells were incubated with normal mouse serum (lane 1) or with MRK16 (lanes 2-6). After washing with PBS, the cells were suspended in 2 ml of crosslinking buffer and 20 ~1 of various concentrations of DSP in DMSO were added: the final concentrations of DSP were 0.125 mM (lane 3), 0.25 mM (lane 4), 0.5 mM (lane 5), and 1.0 mM (lane 6). In the control experiment, 20 ~1 of DMSO was added without DSP (lanes 1,2). After chemical crosslinking, the cells were solubilized with 1% NP-40 followed by immunoprecipitation and SDS-PAGE analysis (for details see Materials and Methods).

486

HAMADA

AND TSURUO

the result of immunoprecipitation of [ “C]leucine-labeled K562/ADM cells with monoclonal antibody MRK 16 using various concentrations of a cleavable crosslinking reagent, DSP. Control immunoprecipitation was performed with normal mouse serum. The precipitate of normal mouse serum identified a minor background band at M, 80,000 (Fig. 1, lane 1). When the crosslinking reagent DSP was used, the precipitates of monoclonal antibody MRK16 yielded a band at M, 170,000-180,000, and the density of this band increased with increasing amounts of DSP used (Fig. 1, lanes 3-6). While the mock-crosslinked precipitate of MRK16 yielded no antigenic band in the position of M, 170,000- 180,000 (Fig. 1, lane 2). The protein of M, 170,000-180,000 has been reported to be a specific protein of pleiotropic drug-resistant tumor cells such as adriamycin-resistant K562 cells (3,4), colchitine-resistant Chinese hamster ovary cells (12) vinblastine-resistant human lymphoblasts (13), and adriamycin-resistant Chinese hamster lung cells (14). These results indicated that the M, 170,000-l 80,000 membrane protein (a) was specifically crosslinked with MRK16 by DSP, (b) is the antigen of MRK16; and (c) seems to be a specific protein of resistant tumor cells. Nonidet-P40 might have denatured the antigenic determinant of this protein, resulting in failure of antigen recognition by the standard immunoprecipitation method using Nonidet-P40 for solubilization. Next we investigated the applicability of the crosslinking method using various detergents for solubilization. As the degree of crosslinking was enhanced at higher concentrations of DSP (Fig. I), in the subsequent experiments crosslinking was carried out at a final DSP concentration of 1 mM. Figure 2 shows the result of immunoprecipitation of [ “C]leucine-labeled K562/ADM cells with monoclonal antibody MRK16 with or without chemical crosslinking using either Nonidet-P40, octyl glucoside, or sulfobetaine 14

FOG. 2. Immunoprecipitation of [‘4C]leucine-labeled K562/ADM cells with monoclonal antibody MRK16 with or without chemical crosslinking using various detergents for solubilization. Labeled K562/ADM cells were incubated with monoclonal antibody MRK 16, followed by crosslinking with DSP at a final concentration of 1 mM (lanes 2,4,6) or mock-crosslinking with DMSO (lanes 1,3,5). Then the cells were solubilized with Nonide&P40 (lanes 1,2), octyl glucoside (lanes 3,4), or sulfobetain 14 (lanes 5,6) followed by immunoprecipitation and SDS-PAGE analysis (for details see Materials and Methods).

for solubilization. When Nonidet-P40 was used, no antigenic band was detected without crosslinking (Fig. 2, compare lanes 1 and 2). This was also the case when the zwitterionic detergent sulfobetaine 14 was used (Fig. 2, lanes 5 and 6). When octyl glucoside was used for solubilization, a vague protein band of M, 170,000-l 80,000 was detected without crosslinking (Fig. 2, lane 3). The antigenic band of M, 170,000- 180,000 protein became evident when the chemical crosslinking method was applied (Fig. 2, lane 4). In lanes 3 and 5, bands of M, 15,000 can be seen. These bands occasionally appeared even in control precipitations with normal mouse serum and are probably nonspecific products of the precipitation procedure. We further tried to determine the antigen recognized by another monoclonal antibody, MRK17, using the crosslinking method. Fig-

ANTIGEN

DETERMINATION

BY A CROSSLINKING

METHOD

487

DISCUSSION

Monoclonal antibodies that recognize cell surface proteins may serve as very useful Mr tools for the study of the biological functions of the membrane proteins. In such experi200Kments, it is essential to solubilize membrane proteins without affecting their antigenic de92.5terminants ( 15). In some cases, however, rel69 atively mild detergents such as Nonidet-P40 could function as protein denaturants. This 46 may be due to (i) membrane antigens being less resistant to denaturation. The constraints that hold the proteins in their active 14.3conformation could be affected by the disruption of essential protein-lipid or protein-protein interactions during solubilizaof [‘4C]leucine-labeled FIG. 3. Immunoprecipitation tion. (ii) Detergent binds to the functional K562/ADM cells with monoclonal antibody MRK17 using a crosslinking method. Labeled K562/ADM cells (or antigenic determinant) sites on the prowere incubated with normal mouse serum (lane 1) or teins (2). In marked contrast to polyclonal with MRK17 with PBS (lanes 2,3), followed by crossantibodies against linking with DSP at a final concentration of 1 mM (lane antibodies, monoclonal the native protein may or may not recognize 3) or mock-crosslinking without DSP (lanes 1,2). Then the cells were solubilized with Nonidet-P40 followed by the denatured product (1). So far there have immunoprecipitation and SDS-PAGE analysis (for de- been no alternative ways to determine the tails see Materials and Methods). membrane protein recognized by a monoclonal antibody when Western blot or standard immunoprecipitation methods are not successful. The method described here using ure 3 shows the result of immunoprecipitation of [ “C]leucine-labeled K562/ADM cells crosslinking reagents to form covalent linkwith monoclonal antibody MRK17 using ages between the monoclonal antibody and Nonidet-P40 for solubilization. Control im- the antigenic protein in intact cell memmunoprecipitation was performed with nor- branes followed by immunoprecipitation, mal mouse serum (Fig. 3, lane 1). Using the provides a new way to detect the membrane crosslinking method, the precipitate of antigen that is easily denatured when solubilized with detergents. This crosslinking MRK17 yieled a band at M, 170,000method is effective with various detergents 180,000 (Fig. 3, lane 3), while the mockused for membrane protein solubilization, as crosslinking precipitate of MRK 17 yielded is shown in Fig. 2. no antigenic band (Fig. 3, lane 2), indicatIn summary, when immune blot or ordiing that the M, 170,000-l 80,000 protein is nary immunoprecipitation methods result in the antigen specifically crosslinked with MRK17. When the M, 170,000-180,000 pro- failure, the application of this crosslinking tein is solubilized with Nonidet-P40, the anti- method would provide an alternative way to genie determinant of the protein seems to be determine membrane antigens recognized by denatured, and can not be recognized by monoclonal antibodies. MRK17. These results indicate the wide appliREFERENCES cability of the crosslinking method for the de1. Goding, J. W. (1983) Monoclonal Antibodies: Printermination of membrane antigens recognized ciples and Practice, pp. 134-187, Academic by various monoclonal antibodies. Press, London.

488

HAMADA

AND TSURUO

2. Helenius, A., and Simons, K. (1975) Biochim. Biophys. Acta 415,29-79. 3. TSUNO, T. (1985) Proc. Amer. Assoc. Cancer Rex 26, 345. 4. TSUNO, T.,

5. 6. 7. 8.

Iida-Saito, H., Kawabata, H., Oh-ham, T., Hamada, H., and Utakoji, T. (I 986) Japan. J. Cancer Rex 77,682-692. Hamada, H., and TSUNO, T. (1986) Proc. Amer. Assoc. Cancer Res. 27, 1548. Hamada, H., Kawabata, H., Makishima, F., Yamori, T., Sugimoto, Y., and TSUNO, T. (1985) Proc. Japan. Cancer Assoc. 44,95 1. Goding, J. W. (1983) Monoclonal Antibodies: Principles and Practice, pp. 98- 133, Academic Press, London. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and

Randall,

R. J. (1951) J. Biof. Chem. 193,

265-275.

Kessler, S. W. (1981) in Methods in Enzymology (Langone, J. J., and Van Vunakis, H., Eds.), Vol. 73, pp. 442-459, Academic Press, New York. 10. Laemmli, U. K. (1970) Nature (London) 227, 9.

680-685.

11. Brenner, M. B., Trowbridge, I. S., and Strominger, J. L. (1985) CeII40, 183-190. 12. Kartner, N., Rio&n, J. R., and Ling, V. (1983) Science 221, 1285-1288. 13. Beck, W. T., Mueller, T. J., and Tanner, L. R. (1979) Cancer Rex 39,2070-2076. 14. Garman, D., and Center, M. S. (1982) Biochem. Biophys. Res. Commun. 105, 157-163. 15. Racker, E. (1983) Fed. Proc. 42,2899-2909.