ANALYTICAL

BIOCHEMISTRY

192,

39-43

(1991)

A Method for Assessing the Reassembly of a Multisubunit Glycoprotein by Western Blotting’ Joyce

W. Lustbader

Department

Received

July

of Medicine,

and Susan

Pollak

College of Physicians

& Surgeons

of Columbia

Academic

Press,

Inc.

The electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose has proven a useful step for the screening of proteins with specific antibodies (l-3). Electrophoresis in conjunction with immunoblotting is a powerful addition to chromatography for 1 This study POl-HD-15454.

New York, New York 10032

19, 1990

The Western blot procedure has been adapted to detect the reassembly of a two-subunit glycoprotein, urinary human chorionic gonadotropin (hCG), directly on the nitrocellulose. This glycoprotein is composed of two nonidentical subunits, (Y and j3. A simple procedure using immunoblotting has been developed to detect reassembly of the monomers to dimer. Three monoclonal antibodies were required for the development of this method: A109, which binds the a subunit or hCG; B105, which binds the fi subunit or hCG; and B107, specific for the intact hCG dimer. The Q subunit and /3 subunit of hCG were each electrophoresed and transferred to nitrocellulose, and the transfer was then incubated with the appropriate complementary subunit; reassembly of the dimer was determined by the binding of the monoclonal antibody B107. Evidence that the reassembly occurs directly on the nitrocellulose comes from the fact that B107 immunoreactivity is detected at the molecular weight position of the subunit and not at the dimer molecular weight. A genetically expressed recombinant form of the a subunit was also tested for its ability to recombine with the opposite subunit to produce the dimer. The recombinant (r subunit was determined to have additional carbohydrate which interfered with the binding of the fi subunit. N-Glycanase digestion of the recombinant (Y subunit produced a form which, when incubated with the /3 subunit, did recombine on the nitrocellulose and could be recognized by B107. 0 1991

University,

was supported

by National

0003-2697/91 $3.00 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

Institutes

of Health

Grant

the detection of specific protein species. In the case of human chorionic gonadotropin (hCG),2 simple gel sieve chromatography cannot distinguish the intact hormone from the free /3 subunit, since both have similar Stoke’s radii, whereas SDS-gel electrophoresis provides excellent resolution as well as greater rapidity, sensitivity, and the capability to study many samples simultaneously. This report demonstrates a system which can detect recombination of a heterodimer directly on nitrocellulose. The method is applicable to any multisubunit system for which reagents recognizing monomeric versus multimeric states have been developed. The heterodimerit protein we used is human chorionic gonadotropin, a glycoprotein hormone composed of two dissimilar subunits, the a! subunit and the P subunit, joined noncovalently. Previous work in this laboratory had generated monoclonal antibodies which were capable of distinguishing intact hCG from the free (Yor /3subunits. These monoclonal antibodies allowed for the development of a system which could be used to screen for intact hCG. Immunoblotting was chosen as the method not only because of its capacity for comparing numerous samples simultaneously but also because of its reproducibility and ease of use. This technique allows visualization of a specific protein even in a complex mixture of proteins. Samples were electrophoresed and transferred, the nitrocellulose-bound cysubunit or /3 subunit was allowed to recombine with its complementary subunit, and reassembly was visualized by a monoclonal antibody specific for intact hCG. The dimer would be visualized at the molecular weight of the protein which was electrophoresed; recombination would take place at that position on the nitrocellulose and not at the dimer weight. In

’ Abbreviations used: hCG, human sodium dodecyl sulfate; BPV, bovine acrylamide gel electrophoresis; BSA, Tris-buffered saline.

chorionic gonadotropin; papillomavirus; PAGE, bovine serum albumin;

SDS, polyTBS,

39

40

LUSTBADER

AND

POLLAK

addition to the urinary proteins, we used hCG recombinant a! subunit and /3 subunit, which were produced in Cl27 cells with the bovine papillomavirus (BPV) by using recombinant DNA methodology (5); the recombinant hCG subunits were screened for their ability to reassemble directly on the nitrocellulose to form dimer. This method has general applicability to the screening of large numbers of samples which are generated from cloning experiments. The procedure could be used with primary plaque lifts to identify the expressed combinable subunits (presumably correctly folded subunits) before additional subcloning experiments. MATERIALS

AND

I2

789101112

B

METHODS

Samplepreparation. The hCGa! (CR-119), the hCG@ (CR-117), and the hCG (CR-119) were purified according to Morgan et al. (4). The recombinant o subunit and /3 subunit were purified according to Lustbader et al. (5). Gel electrophoresis. SDS-polyacrylamide gels were performed according to the procedure described by Laemmli (6). The stacking and resolving gels each contained 0.1% SDS (Biochemical); acrylamide (BDH) stock was used at a concentration of 30% acrylamide:0.8% bisacrylamide (Bio-Rad). Prestained SDSPAGE standards (BRL) were diluted 1:2 in 2~ sample buffer (1X: 0.0625 M Tris-HCl pH 6.8, 2% SDS, 10% glycerol, and 0.001% bromphenol blue). Approximately 70 pmol of protein was diluted in sample buffer at 4°C. Gels (0.75 mm thick) were electrophoresed for 45-60 min, with constant current, 20 mA at 4°C (7). The gels were silver stained by the method of Wray et aZ. (8). Immunoblotting. Following electrophoresis, the gel was soaked for 5 min in transfer buffer-O.025 M Tris base, 20% methanol, and 0.2 M glycine. The electrophoretie transfer of the proteins in the gel to nitrocellulose was accomplished according to Towbin et al. (9) and Burnette (10) at 250 mA for 2 h. The transfer chamber was kept cold by a circulating water pump set at 12’C. Protein visualization. The nitrocellulose paper was incubated in 5% BSA in 0.01 M Tris-HCl, 0.15 M NaCl, pH 7.6 (BSA-TBS), for 45 min at 37°C. The nitrocellulose was then incubated with 5 ml of 5% BSA-TBS containing 200 pg/ml of complementary subunit overnight, at room temperature, with constant rotation and then washed in TBS for 1 h with four changes. The nitrocellulose was then incubated with the primary antibody at a concentration of 1:lOO in BSA-TBS overnight, rotating at room temperature. The nitrocellulose was washed in TBS as above. The secondary antibody, peroxidaseconjugated rabbit anti-mouse antibody (Accurate Biochemical), at a concentration of 1:300 in BSA-TBS, was incubated with the nitrocellulose, rotating at room temperature for 3 h. The nitrocellulose was washed as above. The peroxidase-conjugated antibody was visualized by adding a mixture of 60 ~1 of 30% H,O, in 100 ml of TBS and 60 mg of 3,3’-diaminobenzidine tetrahydrochloride (Bio-Rad) in 20 ml of cold methanol. The

3 456

I 2 3 4 5 6 7 8 9 IO II I2 I3 FIG. 1. (A) Nitrocellulose transfer of a 10% nonreduced low-temperature gel of hCG and subunits. Lanes l-4: B105 immunostained. Lanes 5-8: A109 immunostained. Lanes 9-12: Amido black protein stain. Lanes 2, 6, and 10: Urinary p subunit. Lanes 3, 7, and 11: Urinary (Y subunit. Lanes 4, 8, and 12: Urinary hCG. Lanes 1, 5, and 9: Molecular weight standards shown to the left of the gel. (B) Analyses of the reconstitution of hCG on nitrocellulose. Western blot of a 10% nonreduced low-temperature gel immunostained with B107. Lanes l-5 were preincubated with a subunit (see Materials and Methods). Lanes 6-10 were preincubated with /3 subunit. Lanes 11-13 were not subjected to any preincubation. Lane 1: Recombinant 0. Lanes 2, 7, and 11: Urinary fi subunit. Lanes 3,8,12: Urinary LY subunit. Lanes 4, 9, and 13: Urinary hCG. Lane 6: Recombinant OLsubunit. Lanes 5 and 10: molecular weight standards which are indicated to the left of the gel.

staining reaction was stopped by rinsing the nitrocellulose in deionized water. Monoclonul antibodies. Three monoclonal antibodies (11) were used in these experiments: (1) A109 has an affinity of 3 X lo7 M-’ for hCG and 3.6 X 10” M-l for the (Ysubunit; (2) B105 has an affinity of 1.5 X 10n M-l for both the p subunit and hCG; (3) B107, which is specific for intact hCG, has a binding affinity of 4 X 10” M -’ with cross-reactivity to the (Y or p subunit of less than 0.1%. RESULTS

Figure 1A shows the electrophoretic pattern after SDS-PAGE of markers of known molecular weight and a! subunit, /3subunit, and hCG from urine transferred to nitrocellulose. Lanes lo-12 are stained with amido black and demonstrate apparent molecular weights of 8 subunit, (Y subunit, and hCG of 34,000, 21,500, and

REASSEMBLY

OF

A MULTISUBUNIT

43,000, respectively. Lanes 2-4 are the B105-immunostained profiles of B subunit, cysubunit, and hCG, respectively (note that the (Y subunit sample in lane 3 is positive due to contamination with 0 subunit, which is undetected by the amido black stain in lane 11). A major band is present as /3 subunit, with aggregates of P subunit present as minor bands of higher molecular weight. hCG appears as one major band with less prominent higher and lower molecular weight bands present, which are aggregates and p subunit dissociated from the dimer, respectively. Again, these minor species do not appear in the amido black-stained lanes. Lanes 6-8 are the AlOS-immunostained proteins. In lane 6, no crossreactivity is present with 0 subunit. Lane ‘7 shows a major band of LYsubunit and polymer bands. In lane 8, the hCG is too weakly stained to reproduce well in the photograph; however, there is some a subunit generated by dissociation of the dimer. In Fig. 1B the nitrocellulose was reacted with B107, which is specific for intact hCG (lane 13) and shows no cross-reactivity with either the fi subunit or (Y subunit on Western blots (lanes 11 and 12) or on dose-response curves (data not shown). The nitrocellulose transfer containing lanes l-4 was preincubated with the (Y subunit before incubation with B107. Lane 1 shows recombinant p subunit. This recombinant @subunit is identical to the urinary /3 subunit (lane 2). Both lanes 1 and 2 are positive for B107 immunoreactivity; note that the bands are present at the molecular weight of the fi subunit, and not of hCG (lane 4), indicating that the a subunit recombined with the p subunit directly on the nitrocellulose and was recognized by B107 as intact hCG. When (Y subunit was bound to the nitrocellulose and /3 subunit was added, B107 detected the recombined dimer at the molecular weight of (Y subunit (lane 8) and not of hCG (lane 9). An expressed form of hCGa which did not recombine with urinary or recombinant fl to form hCG has been isolated in this laboratory (5). This molecule proved to be a useful species to test the technique of reassembly as described in this study. A chemical analysis of the recombinant hCGa subunit indicated it was identical to the urinary hCGa except in the NH,-linked carbohydrate moieties. The recombinant (Y subunit was shown to possess additional sugars on these groups (5), whereas the urinary (Y subunit had strictly biantennary carbohydrate branching. This additional carbohydrate probably sterically hinders the binding of @subunit and we hypothesized that its removal should allow recombination to occur. TO test this, recombinant (Y subunit was treated with several glycosidases to remove all or part of the carbohydrate chains. In Fig. 2A a gel was silver stained to indicate the molecular weights of the species to be tested. Recombinant (Y subunit is shown in lanes 4 and 6; the lower molecular weight contaminant in the affinity purified material (lane 4) is not present after final purifi-

PROTEIN

BY

WESTERN

41

BLOTTING

M.W.x 10-3 4325.7-

14.36.23; I 2 3

4

5 6 7 8 9 IO

1234567

C x 10-3

4325.7 l8.4l4.36.23-

I 2 3 4 5 6 7 8 9 IO II I2 I3 14 15 16 17 I8 FIG. 2. (A) SDS-PAGE analyses of urinary hCG, (Y subunit, fi subunit, and recombinant a subunit. The proteins were applied to a 15% Laemmli low-temperature gel and silver stained (8). Lane 1: Urinary hCG. Lane 2: Urinary /3 subunit. Lanes 3 and 7: Molecular weight markers which are indicated to the left of the gel. Lanes 4 and 6: Recombinant a subunit. Lane 5: Recombinant a subunit, N-glycanase treated. Lane 8: Urinary (Y subunit, N-glycanase treated. Lane 9: Urinary a subunit, endoglycosidase F treated. Lane 10: Urinary a subunit. (B) 15% SDS-PAGE Western blot analysis of intact and giycosidase-treated recombinant a subunit and urinary 01 subunit immunostained with A109. Lanes 1 and 3: Recombinant LYsubunit. Lane 2: Recombinant ar subunit treated with N-glycanase. Lane 4: Molecular weight standards which are indicated to the left of the gel. Lane 5: Urinary 01 subunit treated with N-glycanase. Lane 6: Urinary 01 subunit, endoglycosidase F treated. Lane 7: Urinary (Y subunit. (C) Recombination of urinary cy subunit, recombinant (Y subunit, and glycosidase-treated preparations as visualized by Western blot. A 15% nonreduced low-temperature gel immunostained with B107. Lanes l-9 were preincubated with urinary B subunit. Lane 1,5, 10, and 14: Molecular weight markers which are indicated to the left of the gel. Lanes 2, 4, 11, and 13: Recombinant a subunit. Lanes 3 and 12: Recombinant a subunit, N-glycanase treated. Lanes 6 and 15: Urinary 01 subunit, N-glycanase treated. Lanes 7 and 16: Urinary (Y subunit, endoglycosidase F treated. Lanes 8 and 17: Urinary (Y subunit. Lanes 9 and 18: Urinary hCG.

cation (lane 6). The recombinant (II subunit (lanes 4 and 6) can be distinguished from urinary LYsubunit (lane 10) electrophoretically by the increased molecular weight imparted by the additional carbohydrate. When recom-

42

LUSTBADER

binant and urinary (Y subunits were treated with N-glycanase, their molecular weights became very similar (lanes 5 and 8). Lane 9 shows urinary (Y subunit treated with endoglycosidase F (Endo F); there is a small shift in molecular weight between this lane and lane 8, the N-glycanase-treated urinary (Y subunit, because Endo F leaves an N-acetylglucosamine residue attached, whereas N-glycanase cleaves at the point of carbohydrate attachment to the protein. Recombinant cx subunit was resistant to Endo F digestion (data not shown here; see Ref. (5)), which does not cleave triantennary structures. In Fig. 2B recombinant (Y subunit and urinary (Y subunit are shown immunostained with A109. Of particular note is the fact that glycosidase treatment does not perturb the epitope on either species (lanes 2, 5, and 6). In Fig. 2C hCG, recombinant cy subunit, urinary (Y subunit, and the glycosidase-treated (Y subunits are immunostained with B107 (lanes 10-18) or preincubated with p subunit and then reacted with B107 (lanes l-9). Recombinant cr subunit does not combine with /3 subunit (lane 2 and 4) but does combine when the carbohydrate is removed (lane 3). This provides evidence that the carbohydrate present on intact recombinant CY subunit sterically hinders the binding of fi subunit. Alternatively, both native urinary (Y subunit and glycosidasetreated urinary (Y subunit do combine with p subunit (lanes 6-8). Both types of dimers formed are detected at the molecular weight of the (Y subunits. DISCUSSION Western blotting after SDS-PAGE has become a widely used and successful technique for identifying proteins. 1n vitro reconstitution of biological complexes on blots has been demonstrated by others, such as the reconstitution of nucleic acid-histone complexes by Wolff and co-workers (12). The procedure described in this paper combines electrophoresis, transfer to nitrocellulose, complementary subunit incubation, and detection with dimer-recognizing reagent to demonstrate dimer formation. The method here does not use radioactivity, although it could be introduced if the proteins to be tested were available in limited quantity. The advantage of this method over conventional chromatography is that large number of samples can be evaluated simultaneously; the disadvantage is that it is a qualitative measure of recombination and not a quantitative procedure. Precise identification of each protein is required for the success of this method. In the past it was difficult to obtain a significant difference in the migration of the P subunit from the intact hCG molecule due to the large percentage of carbohydrate on the p subunit which, when translated to an electrophoretic profile, showed the p subunit to corn&ate with hCG. Lowering the temperature of the gel to 4°C eliminates this problem. A prerequisite for this study was the availability of monoclonal antibodies specific for both the monomeric

AND

POLLAK

subunits and the dimer. All of these antibodies are directed to conformational epitopes. It is interesting that the SDS-treated proteins are recognized by these antibodies and that the electrophoresis of the proteins in SDS buffer did not prevent reassembly, suggesting some renaturation of the protein had occurred. Others have used blotting to evaluate the modification of proteins by chemical or enzymatic methods (1316), including the removal of carbohydrate side chains from glycoproteins (17). The question we address is whether the constituent monomers of a dimeric protein can recombine to re-form the proper folding expressed as an epitope and subsequently be recognized by a specific antibody. This method is generally applicable to ascertain the proper folding characteristics of multisubunit proteins or other multimeric systems provided that a specific reagent is available to recognize the proper folding of the multimeric structure. For example, our study sought to understand the relationship of the carbohydrate removal to the reconstitution of the intact macromolecule hCG. The modified recombinant a! subunit we studied in this paper was hyperglycosylated compared to native urinary cx subunit. The method described here readily produced three types of information about the recombinant (Y: (1) the apparent molecular weight, (2) the ability to reassemble with the p subunit, and (3) whether the modification interfered with the antibody binding site. Recombinant CYsubunit was identified as having a greater molecular weight than urinary (Y subunit (Fig. 2A) and we found that this increase was due to additional carbohydrate (5). The recombinant (Y subunit did not recombine with /3 subunit (Fig. 2C, lane 2). However, the additional carbohydrate did not prevent recombinant c~from being recognized by the monoclonal antibody A109 (Fig. 2B, lane 1). Treatment with N-glycanase, which liberated the carbohydrate from the (Y subunit, generated a product which could reassemble with /l subunit (Fig. 2C, lane 3). In the future we plan to expand these studies to include subunits altered by site-specific mutagenesis to map the binding site of the p subunit and the epitope to which B107 is directed, i.e., the epitope which includes regions on both the CYsubunit and the /3 subunit. This technique could be extended to identify proteins capable of combining with their complementary subunit from a number of different sources, such as secreted tumor forms. This technique could also help test for conserved contact homology in proteins from different animal species. Blotting has evolved into a routine technique for evaluating antibody-antigen protein complexes. As the research requires, the utility of blotting will widen; recently a procedure which allows for the sequencing of a protein directly from a blotted surface has been established (18). With that tool firmly established, the goal of direct sequencing of membrane-bound or other difficult to purify proteins should be readily realized, provided that an appropriate probe of the blot is available. The

REASSEMBLY

OF

A MULTISUBUNIT

PROTEIN

reconstitution of protein complexes will of course result in an increased understanding of structure-function relationships.

We thank Drs. Robert Canfield and Steven cussions and critical reviews of this report.

Birken

for helpful

dis-

REFERENCES 1. Bers,

G., and Garfin J. M.,

7. Schlaff, 8. Wray, Anal.

WESTERN

S. (1976)

43

BLOTTING

Endocrinology

98,527-533.

W., Boulikas, T., Wray, Biochem. 118,197-203.

V. P., and

Hancock,

9. Towbin, H. T., Staehlin, T., and Fordin, J. (1979) Acad. Sci. USA 76,4350-4354. 10. Burnette, W. N. (1981) Anal. Biochem. 112, 195-203.

ACKNOWLEDGMENTS

2. Gershoni, 1-15. 3. Towbin, 340.

BY

D. (1985)

Biotechniques

Palade,

G. E. (1983)

and

H., and Gordon,

J. (1984)

J. Zmmurwl.

Biochem.

Methods

P. H., Moustafa, Z. A., Krichevsky, E. G., and Canfield, R. E. (1985) Microbial. 8, 48-54.

12. Wolff, P., Gilz, R., Schumacher, Acids Res. 13, 355-367.

3,276-288. Anal.

11. Erlich, strong, munol.

131,

72,313-

13. Bartles,

J. R., and Hubbard,

Proc.

Natl.

A., Birken, S., ArmAnnu. J. Reprod. Zm-

J., and Riesner, L. (1984)

R. (1981)

Anal.

D. (1985)

Biochem.

Nucleic

140,284-

292. 14. Gershoni, J. M., Hawrot, E., and Lentz, Acad. Sci. USA 80,4973-4977.

4. Morgan, F. J., Canfield, R. E., Vaitukaitis, J. L., and Ross, G. T. (1973) in Methods in Investigative and Diagnostic Endocrinology (Berson, S. A., and Yalow, R. S., Eds.), Vol. 2B, pp. 733-742, North-Holland, Amsterdam.

15. Russell, Brown,

5. Lustbader, J. W., Birken, S., Pollak, S., Levinson, L., Bernstine, E., Hsiung, N., and Canfield, R. (1987) J. Biol. Chem. 262, 14,204-14,212. 6. Laemmli, U. K. (1970) Nature (London) 227, 680-685.

17. Woodward, Immunol.

Proc.

D. W., Schneider, W. J., Yamamoto, T., Luskey, M. S., and Goldstein, J. L. (1984) Cell 37,577-585.

16. Wilson, P. T., Gershoni, (1984) Proc. Natl. Acad.

18. Yven, (1986)

T. L. (1983)

J. M., Hawrot, E., and Sci. USA 81, 2553-2557.

M. P., Young, W. W., and Bloodgood, Methods 78,143-153.

S., Hunkapiller, M. Applied Biosystems

W., Wilson, User Bulletin

Lentz,

Natl. K. L., T. L.

R. A. (1985)

K. J., and Yuan, 25.

J.

P. M.