Characterization of Soluble, Disulfide Bond-Stabilized, Prokaryotically ...

0013-7227/97/$03.00/0 Endocrinology Copyright © 1997 by The Endocrine Society

Vol. 138, No. 2 Printed in U.S.A.

Characterization of Soluble, Disulfide Bond-Stabilized, Prokaryotically Expressed Human Thyrotropin Receptor Ectodomain* Y. BOBOVNIKOVA, P. N. GRAVES, H. VLASE,

AND

T. F. DAVIES

Division of Endocrinology and Metabolism, Department of Medicine, Mount Sinai School of Medicine, New York, New York 10029 ABSTRACT To study the interaction of TSH receptor (TSHR) autoantibodies with receptor protein, it is necessary first to express the receptor in the proper conformation including the formation of correct disulfide bridges. However, the reducing environment of the Escherichia coli (E. coli) cytoplasm prevents the generation of protein disulfide bonds and limits the solubility and immunoreactivity of recombinant human TSHR (hTSHR) products. To circumvent these limitations, hTSHR complementary DNA encoding the extracellular domain (hTSHR-ecd; amino acids 21– 415) was inserted into the vector pGEX-2TK by directional cloning and used to transform the thioredoxin reductase mutant strain of E. coli (Ad494), which allowed formation of disulfide bonds in the cytoplasm. After induction, the expressed soluble hTSHR-ecd fusion protein was detected by Western blot analysis using a monoclonal antibody directed against hTSHR amino acids 21–35.

This showed that over 50% of the expressed hTSHR-ecd was soluble in contrast to expression in a wild-type E. coli (strain aF9), where the majority of the recombinant receptor was insoluble. The soluble recombinant receptor was affinity purified and characterized. Under nonreducing SDS-PAGE conditions, the soluble hTSHR-ecd migrated as refolded, disulfide bond-stabilized, multimeric species, whose formation was independent of fusion partner protein. This product was found to be biologically active as evidenced by the inhibition of the binding of 125I-TSH to the full-length hTSHR expressed in transfected CHO cells and was used to develop a competitive capture enzymelinked immunosorbent assay for mapping of hTSHR antibody epitopes. Hence, hTSHR-ecd produced in bacteria with a thioredoxin reductase mutation was found to be highly soluble and biologically relevant. (Endocrinology 138: 588 –593, 1997)

R

nant and natural TSHR (4, 5). Because stable disulfide bonds cannot be formed in the cytoplasm of wild-type Escherichia coli (E. coli) strains, bacterially expressed, hTSHR-ecd is mostly insoluble (4, 6 – 8). Recently thioredoxin reductase (trxB) mutants of E. coli have been shown to be useful in addressing this problem (9). In the absence of the oxidized form of trxB, thioredoxin-like proteins accumulated and promoted disulfide bond formation in the cytoplasm. The mutants expressed biologically active forms of alkaline phosphatase and mouse urokinase, where secondary structure depended on disulfide bond formation (9). Therefore, we have used one of these mutants (E. coli Ad494) to express the hTSHR-ecd fused with glutathione S-transferase (GST) and have analyzed receptor folding and disulfide based interactions in vivo. We showed that the hTSHR-ecd expressed in this system was mostly soluble and formed disulfide-linked multimers with or without the fusion partner. This soluble hTSHR-ecd inhibited binding of the 125I-TSH to the hTSHR and was used in a competitive capture enzyme-linked immunosorbent assay (ELISA) for analyses of TSHR-Abs specific for the C-terminal end of the hTSHR-ecd.

EGULATION of the thyroid gland by TSH involves the binding of the hormone to its receptor (TSHR) expressed on thyroid follicular epithelial cell plasma membranes (1, 2). The binding activates adenylate cyclase and stimulates the growth of thyroid cells and the production of thyroid hormones. Stimulating human TSHR autoantibodies (hTSHR-Abs) in the serum of patients with Graves’ hyperthyroidism mimic the action of the hormone, leading to hyperstimulation of the gland. Conversely, hTSHR-Abs in some patients with Hashimoto’s disease block the action of TSH at the receptor and contribute to the cause of hypothyroidism. These hTSHR-Abs interact with the extracellular domain of the hTSHR (hTSHR-ecd) at multiple sites (for review, see Ref. 3). However, mapping of specific hTSHR-Ab binding sites on the receptor has been complicated by difficulties in isolating the natural receptor and soluble, properly folded forms of recombinant receptor protein. The ectodomain of the hTSHR contains 11 cysteine residues, some of which are likely to form intrachain disulfide bridges (3) that stabilize the native structure of the receptor. Additionally, we previously have described interchain disulfide links involved in multimerization of both recombi-

Materials and Methods Cloning and expression of hTSHR-ecd in pGEX-2TK

Received July 1, 1996. Address all correspondence and requests for reprints to: Dr. P. N. Graves, Box 1055, Mount Sinai Medical Center, 1 Gustave L. Levy Place, New York, New York 10029. * This work was supported in part by grants from NIDDKD (DK35764 and DK-45011 to T.F.D.), the David Owen Segal Endowment, the Irvington Foundation for Medical Research (to H.V.), and Training Grant DK-07645 in Cellular and Molecular Endocrinology (to Y.B.).

The signal sequenceless hTSHR-ecd (amino acids 21– 415) was amplified by RT-PCR from human thyroidal complementary DNA using 59 (59-GATCGGATCCGGAATGGGGTGTTCGTCTCC) and 39 (59-CACTTCTGTATTACCCGATGTTCATTCTTAAGCCATGGCTAG) primers, and the resulting product was digested with BamHI and EcoRI (sites underlined) and ligated into the GST fusion vector pGEX-2TK vector (Pharmacia, Piscataway, NJ). The recombinant DNA was used to trans-

588

SOLUBLE TSH RECEPTOR ECTODOMAIN form two competent E. coli strains: wild-type (aF9; Strategene, La Jolla, CA) and trxB mutant Ad494 (9). Ampr colonies were grown overnight in L broth (10) containing 50 mg/ml ampicillin, diluted ten times in the same medium, and grown to a concentration of 1 absorbance U/ml (at 600 nm). Expression of hTSHR-ecd was induced with 1 mm IPTG for 1 h (Ad494; aF9 expresses constitutively), and the bacterial suspension was processed immediately or held on ice for 1 h. The cells were pelleted, resuspended at 1/20 vol in cold 1 3 PBS, and sonicated on ice (Branson Sonifier 450, Branson Ultrasonics Corporation, Danbury, CT) (Method 1). Alternatively, cells were resuspended at 1/100 vol in 10 mm EDTA, 20 mm Tris-HCl (pH 8.0), incubated with chicken egg white lysozyme (2 mg/ml) for 30 min on ice, and lysed by three cycles of freezing in a dry ice-ethanol bath and thawing in a 28 C water bath (Method 2). The viscous suspension then was diluted twice in cold 2 3 PBS and sonicated three times for 30 – 40 sec on ice.

Western blotting and quantitative analysis of recombinant hTSHR-ecd Protein concentration was measured by the bicinchoninic acid method (Pierce, Rockland, IL). One to ten micrograms of total protein from soluble and insoluble fractions was dissolved in 2 3 SDS-PAGE buffer (2% SDS, 20% glycerol, 0.001% Bromphenol blue, and 0.125 m Tris-HCl pH 6.8) with or without subsequent boiling in a presence of 2% b-mercaptoethanol (bME) and fractionated by SDS-PAGE (Miniprotean II, Bio-Rad Laboratories, Richmond, CA). Proteins were transferred onto supported nitrocellulose membranes (0.2 mm; BA-S NC, Schleicher and Schuell, Keene, NH), and blots were developed as described (4) using hTSHR antibodies shown in Table 1. Bound antibodies were detected by peroxidase-labeled antimouse or antirabbit Ig using enhanced chemiluminescence (Amersham, Amersham, UK). The relative amount of hTSHR-ecd in different fractions was measured using densitometric readings of autoradiographs (Molecular Dynamics Personal Densitometer and ImageQuant V3.3 software, Sunnyvale, CA). Scanned peaks were corrected for the total amount of protein in each fraction calculated from the same original volume of cells. Sizes of proteins were estimated using prestained protein molecular weight standards (GIBCO BRL, Bethesda, MD).

Affinity purification of hTSHR-ecd Purification of soluble hTSHR-ecd from a 500-ml culture of E. coli Ad494 was performed on Glutathione Sepharose 4B columns (Pharmacia) using the recommended procedure, except that sonication was applied after lysis of cells by Method 2 and the sonicate filtered through a 0.45-mm filter (Nalgen) and diluted twice in cold 2 3 PBS before applying to the column, which was then washed extensively with 1 3 PBS. The fusion protein was digested overnight on the column with 50 U thrombin (Pharmacia) and eluted with either 1 ml PBS or elution buffer (50 mm Tris-HCl, pH 8.0, 10 mm reduced glutathione). Binding and elution steps were carried out in the cold to decrease proteolytic degradation of hTSHR-ecd.

TSH binding of the affinity-purified hTSHR-ecd Binding of TSH to the soluble purified hTSHR-ecd was tested in a competition assay using a CHO cell system (11). As a control-binding protein, we used similar amounts of an unrelated protein, anthrax toxin protective antigen, which we expressed in E. coli Ad494 and affinity purified in parallel with hTSHR-ecd. A pGEX plasmid containing this

589

anthrax toxin subunit was a kind gift of Dr. A. V. Teixeira, Department of Medicine, Mount Sinai School of Medicine, New York. Affinity-purified hTSHR-ecd or anthrax toxin was dialyzed against binding buffer (NaCl-free HBSS containing 280 mm sucrose) and preincubated for 1 h at room temperature with 8,000 cpm 125I-bovine TSH (bTSH) (Kronus, San Clemente, CA) in 300 ml binding buffer. The mixture was added to CHO JP09 cells stably expressing the wild-type hTSHR (12) (gift of Dr. G. Vassart, Free University of Bruxelles, Belgium). After 1.5 h incubation at 37 C, the cells were washed three times with ice-cold binding buffer, solubilized with 0.5 ml 1N NaOH, and radioactivity was measured. The maximum amount of 125I-TSH bound to the cells in the presence of only the binding buffer was used as a positive control, and the amount bound in the presence of 1026 m unlabeled bovine TSH (Thytropar, Armour, Kankakee, IL) was used as a control for nonspecific binding.

ELISA for epitope mapping The 96-well plates (Immulon 2, Dynatech Laboratories, VA) were coated overnight with A10 or A11 monoclonal antibody (mAb) (Table 1) diluted in carbonate-bicarbonate buffer (pH 9.6). After washing the plate with PBS-0.05% Tween 20, nonspecific binding sites were blocked with 1% BSA (Sigma Chemical Co., St. Louis, MO) in PBS-Tween for 1 h at 37 C. Bacterial sonicate was diluted in PBS-Tween/1 mm phenylmethylsulfonyl fluoride (PMSF) to a concentration 0.5 mg protein/ml. Sonicate (100 ml/well) was added and incubation continued for 2 h at 37 C. After rigorous washing, captured hTSHR-ecd was incubated with either 100 ml test serum (diluted in PBS-Tween) or dilution buffer only. The wells were washed again and incubated for 1 h at 37 C with reporter antibody (for example, R397; Table 1). After washing, the reporter antibody was detected by incubation with alkaline phosphatase-conjugated goat antirabbit IgG (Sigma) using p-nitrophenylphosphate substrate diluted in diethanolamine buffer (pH 9.8). The absorbance at 405 nm was measured using an automated ELISA plate reader (Titerek Multiscan MCC/340 MK II; Flow Laboratories, McLean, VA). The percent of competitive inhibition of all samples was calculated after subtracting background ODs (OD of wells containing host bacterial lysate without the hTSHR-ecd), and inhibition by immune serum was calculated after subtracting nonspecific inhibition using control serum. Analyses of data were performed using the Excel 5.0 program for Windows.

Results Expression of soluble hTSHR-ecd in E. coli Ad494

The expression of soluble hTSHR-ecd in nonmutant (aF9) and txtB mutant (Ad494) strains of E. coli is shown in Fig. 1A. Soluble hTSHR-ecd was almost undetectable in aF9 (lane 1) but was readily detected as a 74-kDa band corresponding to the GST-fusion protein (26 kDa GST plus 48 kDa hTSHR-ecd) in the trxB mutant strain Ad494 (lane 2). The amount of soluble hTSHR was markedly increased by induction (lane 3) and further increased by cooling the bacterial suspension for 1 h (lane 4). The increase in soluble hTSHR-ecd during storage of induced cells on ice indicated that disulfide bond formation was more favorable under these conditions. The low-molecular mass bands visible in lanes 3 and 4 represented products of proteolytic cleavage by host proteases

TABLE 1. Characterization of hTSHR-ecd antibodies hTSHR-Ab

A10/11

Epitope specificity (aa)

21–35

A7

402– 415

R397

397– 415

Source

Antigen

Dilution

References

Mouse mAb (IgG) Mouse mAb (IgG) Rabbit serum

hTSHR-ecd

1:5000

(21)

hTSHR-ecd

1 :500

(21)

Peptide

1:5000

(4)

The murine monoclonal antibodies were generated using insect cell recombinant hTSHR-ecd (2) and were a kind gift from Dr. Paul Banga, University of London, London, UK.

590

SOLUBLE TSH RECEPTOR ECTODOMAIN

FIG. 1. Expression of soluble hTSHR-ecd in E. coli detected by Western blot of reduced SDS-PAGE fractionated proteins. A, Soluble hTSHR-ecd, produced in E. coli aF9 (lane 1) and trxB Ad494 (lanes 2– 4) strains. Single colonies of transformed cells were grown as described in Materials and Methods with (lanes 3, 4) or without (lanes 1, 2) induction, and cells were harvested and lysed directly after induction (lane 3) or kept on ice for 1 h (lane 4). Lysates were cleared by centrifugation at 12,000 rpm for 15 min, and soluble fractions containing 10 mg of total proteins were boiled for 3 min in 2 3 SDS-PAGE sample buffer containing 2% bME and analyzed by Western blotting as described in Materials and Methods. B, Accumulation of hTSHRecd in different fractions of E. coli Ad494. Clones were grown as described earlier, and cells were sonicated on ice for 30 sec. Unbroken cells were collected by brief centrifugation, and the sonicate was separated into soluble and insoluble fractions by centrifugation at 12,000 3 g for 15 min. The pellets were washed once with PBS containing 1% Triton and dissolved in 2 3 SDS-PAGE sample buffer. Proteins from pellets before (lane 1) and after (lane 2) sonication and from soluble (lane 3) and insoluble (lane 4) fractions were separated under reducing conditions by 8% SDS-PAGE and analyzed by Western blotting. The relative amount of hTSHR-ecd present in each fraction was calculated using quantitative densitometry.

during preparation of protein extracts and varied among preparations. The presence of some soluble hTSHR-ecd expression in the total absence of induction (lane 2) indicated that downregulation of basal expression was not complete. The relative amounts of soluble and insoluble hTSHR-ecd were calculated using densitometry of the Western blots (see Materials and Methods) (Fig. 1B). The amount of hTSHR-ecd present in cells before (lane 1) and after (lane 2) sonication showed that most hTSHR-ecd remained in the cell pellets. This revealed that sonication alone (Method 1) was not sufficient to break the Ad494 cells and release hTSHR-ecd. In most experiments, therefore, the cells were opened with lysozyme followed by freeze-thaw, and a brief sonication was applied to decrease viscosity (Method 2). The quantitation of densitometric readings of soluble (lane 3) and insoluble (lane 4) fractions of sonicate relative to the amount of protein loaded (see Materials and Methods) showed that the total amount of soluble hTSHR-ecd was present in quantities 1.4 – 1.9 times higher than the insoluble protein. Affinity purification of recombinant hTSHR-ecd

Soluble hTSHR-ecd expressed in E. coli Ad494 was purified on Glutathione Sepharose 4B. Binding of the 74-kDa GST-hTSHR-ecd fusion protein and release of cleaved (48 kDa) hTSHR-ecd from the matrix by thrombin was monitored as a function of time (Fig. 2A). Partial digestion of the fusion protein (lane 1) occurred after 2 h (lane 2) and was nearly complete after 16 h of digestion (lane 3). Elution of hTSHR-ecd from the column also was monitored. As shown

Endo • 1997 Vol 138 • No 2

FIG. 2. Affinity purification of hTSHR-ecd on a glutathione Sepharose 4B column. E. coli Ad494 cells, transformed with pGEX-2TK containing hTSHR-ecd, were grown, harvested, and lysed as described in Materials and Methods. A, Immunoanalysis of thrombin digestion of hTSHR-ecd bound to Glutathione Sepharose 4B using small-scale bulk purification. Lane 1, undigested hTSHR-ecd; lane 2, 2-h digestion; lane 3, overnight digestion. B, Column purification of hTSHR-ecd. Soluble hTSHR-ecd was eluted with 1 ml 1 3 PBS after first digesting with thrombin (lane 1) and two subsequent elutions with 1 3 PBS after a second digestion with thrombin (lanes 2 and 3) followed by 4 elutions with buffer containing 10 mM of reduced glutathione (lanes 4 –7).

in Fig. 2B, the amount of purified hTSHR-ecd after one round of thrombin cleavage and elution using PBS (lane 1), was inefficient. A second cycle of cleavage and elution with the same buffer did not improve the recovery of the purified receptor (lanes 2 and 3). However, addition of reduced glutathione to the elution buffer greatly enhanced the recovery of bound hTSHR-ecd protein (lanes 4 –7). This suggested that the poor recovery in PBS may have been caused by aggregate formation by cleaved protein. The multiple small bands of recombinant hTSHR-ecd less then 48 kDa were most likely the products of proteolytic cleavage of hTSHR-ecd during preparation. Addition of 1 mm PMSF to the bacterial pellets before sonication was later used to decrease this degradation, and addition of sodium chloride (1–2 m) to PBS during elution further increased the recovery of purified receptor (data not shown). The total amount of purified soluble hTSHR-ecd produced varied from 10 –50 mg/liter of culture. Formation of hTSHR-ecd multimers

Previous data from this laboratory indicated that recombinant prokaryotic hTSHR-ecd, solubilized in urea and purified by PAGE-SDS, formed tetramers under nonreducing conditions (4). To examine the formation of multimeric forms in the disulfide bond-stabilized soluble hTSHR-ecd, Ad494generated soluble and insoluble fractions were analyzed under reducing and nonreducing conditions. Under reducing conditions, the size of the major soluble species (74 kDa) was the same as the insoluble species, in agreement with the estimated molecular mass of the hTSHR-ecd fusion protein, and slower migrating forms were not detected (Fig. 1B). Under nonreducing conditions, the sizes of the major species in the insoluble fraction were 74 and 215 kDa, suggesting some formation of multimeric forms (Fig. 3A, lane 1). In contrast, the soluble receptor formed only a 215-kDa species with no other forms present (Fig. 3A, lane 2), indicating that the soluble hTSHR-ecd proteins formed disulfide-linked


591

under partially reduced conditions with the GST-fusion protein (lane 1) was resolved into a major 74-kDa band after complete denaturation (lane 2). The hTSHR-ecd without the GST partner formed 180-kDa aggregates under partially denaturing conditions (lane 3), which were resolved into 48-kDa proteins during denaturation (lane 4). This suggested that the purified soluble recombinant hTSHR-ecd protein formed stable structures under nondenaturing conditions. The apparent mass of these structures, estimated from migration rates, was compatible with four monomers held together by disulfide bridges. The GST sequence did not contribute to this process. FIG. 3. The sizes of hTSHR-ecd present in soluble and insoluble fractions. A, Soluble and insoluble protein fractions of E. coli Ad494, expressing hTSHR-ecd, were separated under nonreducing conditions by 8% SDS-PAGE and analyzed by Western blotting. Insoluble fraction of sonicate (lane 1) and soluble fraction of sonicate (lane 2) obtained as described in Fig. 1B legend. B, Soluble fractions, containing hTSHR-ecd with (lanes 1, 2) or without (lanes 3, 4) GST fusion partner were diluted in 2 3 SDS-PAGE sample buffer with (lanes 2, 4) or without (lanes 1, 3) boiling. Proteins were separated by 7% SDS-PAGE and analyzed by Western blotting. Thrombin digestion of fusion protein was performed at 4 C overnight.

multimers in solution with much higher efficiency than the insoluble. To assess the possible role of the GST fusion partner in the formation of hTSHR-ecd multimers, we analyzed the sizes of soluble hTSHR-ecd with and without the GST partner (Fig. 3B). Thrombin digestion removed the 26-kDa GST and generated a 48-kDa hTSHR-ecd (lanes 3 and 4). Both digested and undigested recombinant proteins were analyzed under denaturing (boiling of samples in the presence of 2% bME) and partially denaturing conditions (samples were mixed in the same buffer and loaded directly onto the gel without boiling). Partially denaturing conditions were chosen because they allowed us to have a mixture of both forms of proteins (reduced and nonreduced). The band running above the 210-kDa marker

FIG. 4. Binding of TSH to recombinant hTSHR-ecd. Affinity-purified and dialyzed hTSHR-ecd, containing 10 mg/ml of the receptor, was preincubated with 125 I-TSH and then added to confluent CHO cells expressing hTSHR as described in Materials and Methods. As a control we used anthrax toxin, which was expressed and purified in the same system. The insert in the figure shows the control inhibition of 123I-TSH binding in the presence of bTSH.

Binding of TSH to hTSHR-ecd

To determine whether the soluble hTSHR-ecd prepared from E. coli Ad494 cells was biologically active, we tested its ability to compete for binding of radioiodinated TSH to functional hTSHR expressed on CHO JP09 cells (12). Affinity-purified hTSHR-ecd (without the fusion partner) reduced binding of 125I-TSH to the hTSHR expressed in CHO cells in a dose-dependent manner, reaching approximately 40% at 150 ml/well of eluate containing approximately 10 mg/ml hTSHR-ecd (Fig. 4). The same amount of control anthrax toxin protein reduced binding by approximately 12%. Binding data analyzed by Fourier-derived affinity spectrum analysis (FASA, 13) using the MacFASA program (Biomath Explorations, Port Washington, NY) showed that the affinity of binding of labeled TSH to the purified hTSHR-ecd was 4.2 3 1028 m. This affinity was about 100 times less than achieved using porcine thyroid membranes as a natural source of full length TSHR (14). Nevertheless, these data reflected increased biological activity of the soluble hTSHR-ecd compared with the insoluble hTSHR-ecd, where no inhibition of labeled TSH binding was observed using similar concentrations of the receptor (6).

592

Endo • 1997 Vol 138 • No 2


Defining epitopes of antibodies to hTSHR-ecd by capture ELISA

The soluble hTSHR-ecd was then used to develop a capture ELISA assay for the detection of epitope-specific TSHR antibodies in the serum of hTSHR-ecd immunized mice. Soluble hTSHR-ecd was captured by immobilized antihTSHR from bacterial sonicates using plates bound with mAb specific for the N-terminal region of the hTSHR-ecd (either A10 or A11). Bound, purified, receptor ecd was detected by rabbit antibody R397 produced against the C-terminal portion of the ecd (Table 1). Using the capture antibody A11 approximately 100 ng ecd was bound when bacterial sonicate, containing 50 mg of total protein, was added per well; whereas 50 ng was captured using A10 (Fig. 5A). The potential competitive inhibition of the rabbit antibodies binding to the captured hTSHR-ecd was tested initially using, as a competitor, mAb A7 specific for aminoacids 402– 415 of hTSHR ecd (Table 1). As shown in Fig. 5B, detection of the captured hTSHR-ecd by R397 was inhibited by more than 50% by preincubation with mAb A7 (sample 1) but by less than 5% by preincubation with control mAb (sample 2), demonstrating the specificity of detection of R397-binding sites (aa 397– 415). We then analyzed serum from seven CBA/J mice immunized with hTSHR-ecd (15) for the presence of antibodies to the C-terminal residues of the receptor ecd. Incubation of the captured receptor with immune serum reduced specific R397 binding by more than 30% (Fig. 5B, sample 3), compared with no specific inhibition by pooled serum from seven mice immunized with control insect cell extract (15). These results showed that the immunized mice had developed an immune response to the same C-terminal residues of the hTSHR-ecd.

FIG. 5. Inhibition of rabbit R397 antibody binding to the C-terminal regions of the recombinant TSHR-ecd by mouse serum. A, Capture of hTSHR-ecd. hTSHR-ecd, produced in E. coli Ad494, was purified on an ELISA plate using serial dilutions of sonicate containing 5–500 mg/ml total protein as described in Materials and Methods. Either 1:5000 dilution of mAb A11 ascites or 1:1000 dilution of an IgG fraction (protein A purified, 2.0 mg/ml total protein) of mAb A10 was used as capture antibody. The amount of bound receptor was detected by rabbit R397 antibody at 1:5000 dilution. B, Inhibition of the R397 antibody binding by mAb A7. mAb A7 (100 ml at 1:500 dilution), specific for aa 402– 415 (sample 1), or the same dilution of mAb produced against neuraminidase (ND) receptor (sample 2) or serum from 7 mice immunized with hTSHR-ecd (1:200 dilution, sample 3) were incubated overnight at room temperature with hTSHR-ecd, bound to the plate. After washing, available sites were detected with R397 antibody (Materials and Methods). The results were expressed as percentage inhibition of R397 binding; means and SEM of six repeats.

Discussion

The hTSHR-ecd is the major site of TSH binding and also is a major thyroid antigen in patients with Graves’ and Hashimoto’s diseases. Autoantibodies are thought to interact with different sites of the TSHR-ecd, resulting in TSH agonism or antagonism and causing the resulting abnormalities in thyroid function. The precise characterization of these autoantibodies requires a supply of correctly folded hTSHRecd. The low level of the receptor obtained from thyroid cells and mammalian expression systems has not allowed the production of the necessary amounts of conformationally correct and functional receptor. Much higher levels of the hTSHR-ecd are attainable in bacterial and insect cells, but most of the product is insoluble. In this report, we have described the expression of soluble recombinant hTSHR-ecd in the txt mutant strain of E. coli (Ad494), its structural and functional characterization, and application for TSHR-Ab detection and epitope mapping. We showed that the truncated receptor expressed in E. coli Ad494 was mostly soluble. This was in contrast to the expression of the truncated receptor in other E. coli strains (DH5a and aF’), where most of the recombinant protein was insoluble (4, 6). This difference in solubility of recombinant hTSHR-ecd in different E. coli strains reflected the fact that proper folding of the hTSHR-ecd depended on S-S bond formation, which cannot be achieved in the cytoplasm of commonly used E. coli strains (9). However, it was possible to stabilize the structure by allowing disulfide bond formation in a bacterial strain containing mutations in the txt gene (9). The soluble hTSHR-ecd without the GST fusion protein was affinity purified on a glutathione Sepharose column using thrombin digestion of the fusion partner and elution under mild conditions. However, the total amount of hTSHR-ecd produced in E. coli Ad494 was less than reported for the expression of insoluble hTSHR-ecd (9) secondary to proteolytic degradation of the hTSHR-ecd and low efficiency of its elution from the affinity column. The relevance of the structure of hTSHR-ecd, expressed in bacteria as soluble protein, to the native receptor conformation was addressed by analyzing formation of multimers in vivo. The ecd of the hTSHR contains 11 cysteine residues, some of which are likely to be involved in dimerization/ multimerization of subunits. Recently, we detected the multimerization of the natural TSHR and receptor subunits in detergent-solubilized porcine thyroid membranes (5). We also have previously demonstrated that multimers formed in vitro from recombinant hTSHR-ecd proteins purified from E. coli, and insect cells (4, 16) were recognized by autoantibodies from patients with Graves’ disease (16). In this report, we demonstrated that the soluble hTSHR-ecd formed such multimers in vivo. To exclude the possible involvement of the GST fusion partner in multimerization, soluble hTSHR was analyzed with and without the GST portion of the molecule. SDS-PAGE analysis of the soluble fraction showed that the soluble hTSHR-ecd formed multimers under nonreducing conditions, and their formation was independent of the presence of the GST fusion partner. This was in agreement with the formation of hTSHR-ecd multimers using in vitro refolding of urea-solubilized recombinant receptor, expressed as


insoluble aggregates in bacteria (4) and insect cells (16). The relevance of such multimeric organization of recombinant hTSHR-ecd and the natural TSHR was supported by the fact that lutropin-choriogonadotropin and FSH receptors, which have a high degree of similarity with the TSHR, have been shown also to form tetramers (17–20). The function of the expressed hTSHR-ecd was analyzed by TSH binding-assay. Our results showed that 30 pm of the soluble, affinity-purified, recombinant receptor inhibited TSH binding to hTSHR expressed on CHO cells by approximately 40%. Previous studies with the same amount of hTSHR-ecd, produced in E. coli as insoluble protein, and refolded in vitro have shown no inhibition of TSH binding (6). These data indicated that solubility of recombinant hTSHRecd was an important step in producing biologically active protein and that, perhaps, previous solubilization under harsh denaturing conditions (8 m urea) and refolding in vitro did not allow the formation of the same structure as was formed in the present studies in the cytoplasm of the bacteria. At the same time, inhibition of TSH binding by soluble recombinant hTSHR-ecd was much less efficient than inhibition by the TSHR present in porcine thyroid membranes, as reported in the literature (14), indicating that other factors (for example, the presence of the transmembrane domain and the proper posttranslational modification of the hTSHR) are even more important for the high-efficiency binding of the TSH to its receptor. We used the soluble prokaryotic hTSHR-ecd to develop a competitive capture ELISA assay for detection of epitopespecific hTSHR-Abs. The hTSHR-ecd was captured on a plastic surface by antibody generated against the N-terminal end of the receptor, and the amount of hTSHR-Ab specific for the C-terminal end (aa 397– 415) was detected by inhibition of detecting rabbit antibody (R397) binding. To test the system, we used mAb A7 specific for aa 402– 415 of the hTSHR. This mAb blocked binding of rabbit antibody by 50%, compared with only 5% blocking by unrelated antibody. We also analyzed sera from mice immunized with hTSHR-ecd. The 30% of specific inhibition by this serum suggested the presence of hTSHR-Ab to aa 397– 415 in the immunized mice. This region of the receptor previously was shown to be important in hormone binding because the same rabbit antibody inhibited binding of TSH to porcine TSHR by up to 33% (5). However, mice which developed hTSHR(397– 415)-Ab remained euthyroid (15), suggesting that the interaction with mouse TSHR was insufficient to significantly inhibit thyroid function or that such function was compensated for by other mechanisms. By changing the detecting antibody, it should be possible to measure the entire epitope repertoire recognized by the mouse sera. This technique will give us the opportunity to analyze sera for the presence of antibodies to a variety of epitopes on the unglycosylated hTSHR-ecd, thus providing their epitope characterization.

593

Acknowledgments We thank Dr. Jon Beckwith, Department of Microbiology and Molecular Genetics, Harvard Medical School, Boston, MA, for the kind gift of E. coli Ad494 strain and Dr. Jeffrey I. Mechanick for his help with FASA binding analyses.

References 1. Pastan I, Roth J, Macchia V 1996 Binding of hormone to tissue: the first step in polypeptide hormone action. Proc Natl Acad Sci USA 56:1802–1809 2. Magner JA 1990 Thyroid-stimulating hormone: biosynthesis, cell biology, and bioactivity. Endocr Rev 11:354 –385 3. Nagayama Y, Rapoport B 1992 The thyrotropin receptor 25 years after its discovery: new insight after its molecular cloning. Mol Endocrinol 6:145–156 4. Graves PN, Vlase H, Davies TF 1995 Folding of the recombinant human thyrotropin receptor (TSH) extracellular domain: identification of folded monomeric and tetrameric complexes that bind TSH receptor autoantibodies. Endocrinology 136:521–527 5. Graves PN, Vlase H, Bobovnikova Y, Davies TF 1996 Multimeric complex formation by the thyrotropin receptor in solubilized thyroid membranes. Endocrinology 137:3915–3920 6. Huang GC, Collison KS, McGregor AM, Banga JP 1992 Expression of a human thyrotropin receptor fragment in Escherichia coli and its interaction with the hormone and autoantibodies from patients with Graves’ disease. J Mol Endocrinol 8:137–144 7. Harfst E, Johnstone AP, Nussey SS 1992 Characterization of the extracellular region of the human thyrotropin receptor expressed as a recombinant protein. J Mol Endocrinol 9:227–236 8. Costagliola S, Alcalde L, Ruf J, Vassart G, Ludgate M 1994 Overexpression of the extracellular domain of the thyrotropin receptor in bacteria; production of thyrotropin-binding inhibiting immunoglobulins. J Mol Endocrinol 13:11–21 9. Derman AI, Prinz WA, Belin D, Beckwith J 1994 Mutations that allow disulfide bond formation in the cytoplasm of Escherichia coli. Science 262:1744 –1747 10. Maniatis T, Fritsch EF, Sambrook J 1989 Molecular Cloning - a Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor 11. Filetti S, Foti D, Costante G, Rapoport B 1991 Recombinant human thyrotropin (TSH) receptor in a radioreceptor assay for the measurement of TSH receptor autoantibodies. J Clin Endocrinol Metab 72:1096 –1101 12. Costagliola S, Swillens S, Niccoli P, Dumont JE, Vassart G, Ludgate M 1992 Binding assay for thyrotropin receptor autoantibodies using the recombinant receptor protein. J Clin Endocrinol Metab 75:1540 –1544 13. Mechanick JI, Peskin CS 1986 Resolution of steroid binding heterogeneity by Fourier-derived affinity spectrum analysis (FASA). Anal Biochem 157:221–235 14. Saltiel AR, Thomas Jr CG, Nayfeh SN 1982 Thyrotropin binding to porcine thyroid plasma membranes: kinetic and thermodynamic analyses. Mol Cell Endocrinol 28:299 –312 15. Vlase H, Nakashima M, Graves PN, Tomer Y, Morris JC, Davies TF 1995 Defining the major antibody epitopes on the human thyrotropin receptor in immunized mice: evidence for intramolecular epitope spreading. Endocrinology 136:4415– 4423 16. Vlase H, Graves PN, Magnusson RP, Davies TF 1995 Human autoantibodies to the TSH receptor: recognition of linear, folded and glycosylated recombinant extracellular domain. J Clin Endocrinol Metab 80:46 –53 17. Chen W, Bahl OP 1993 High expression of the hormone binding active extracellular domain (1–294) of rat lutropin receptor in Escherichia coli. Mol Cell Endocrinol 91:35– 41 18. Dufau ML, Minegishi T, Buczko ES, Delgado CJ, Zhang R 1989 Characterization and structure of ovarian and testicular LH/hCG receptors. J Steroid Biochem Mol Biol 33:715–720 19. Dattatreyamurty B, Smith RA, Zhang SB, Santa-Coloma TA, Reichert LE 1992 The size of the mature membrane receptor for follicle-stimulating hormone is larger than that predicted from its cDNA. J Mol Endocrinol 9:115–121 20. Dattatreyamurty B, Zhang SB, Reichert Jr LE 1990 Purification of follitropin receptor from bovine calf testes. J Biol Chem 265:5994 –5503 21. Nicholson LB, Vlase H, Graves PN, Nilsson M, Molne J, Huang GC, Morgenthaler NG, Davies TF, McGregor AM, Banga JP 1996 Monoclonal antibodies to the human TSH receptor: epitope mapping and binding to the native receptor on the basolateral plasma membrane of thyroid follicular cells. J Mol Endocrinol 16:159 –170