J. Biochem. I l l , 633-637 (1992)
Mapping of a Linear Autoantigenic Epitope within the Human Thyroid Peroxidase Using Recombinant DNA Techniques Brigitte Frorath,* Charles C. Abney,* Mirco Scanarini,* Heike Berthold,* Nick Hunt,** and Wolfgang Northemann*1 'Department of Molecular Biology, ELJAS Entwicklungslabor, Obere Hardtstrasse 18, D-7800 Freiburg, F.R.G.; and "Institute for Hormone and Fertility Research, D-2000 Hamburg, F.RG. Received for publication, January 7, 1992
Autoantibodies directed against the thyroid peroxidase (TPO), the thyroid microsomal antigen, are widely used to diagnose human autoimmune thyroid disease. A cloned 3.088 kb cDNA coding for the entire mature human TPO was isolated from a cDNA library derived from a pathological thyroid gland of a Graves' disease patient and used further to generate a so-called TPO epitope cDNA library in order to map linear autoantigenic epitopes involving a recombinant molecular biology approach. The TPO epitope cDNA library consisting of randomly fragmented cDNA sequences inserted in the expression vector pGEX-2T was expressed in Escherichia coli and screened with characterized anti-TPO autoantisera from Hashimoto's disease patients. All the sera were positively tested with a purified thyroid microsomal antigen fraction (TMA/TPO). Only about 1% of examined autoantisera were able to recognize bacterial expressed recombinant TPO representing sequential antigenic determinants. A corresponding autoantigenic epitope with 61 ami no acids in length was located at the C-terminus of human TPO.
Organ specific autoimmune diseases are common in man which include thyroid autoimmune disease, insulin-dependent diabetes mellitus, myasthenia gravis, and others which are all mediated by abnormal generation of autoantibodies (for review see Ref. 1). The autoantibodies directed to the thyroid tissue are very important clinical markers for the human autoimmune thyroid disease which is a relatively common disease affecting about 5% of the population (2). The characteristics of this disease are mainly tissue destruction in lymphocytic thyroiditis (Hashimoto's disease) and the abnormal modulation of the thyrotropin receptor by autoantibodies leading to hyperthyroidism (Graves' disease) (1, 3-5). Thyroid-specific autoantibodies are directed against several components of the thyroid gland, including thyroid peroxidase (TPO) which is one of the major thyroid-specific antigens {6-8). The TPO was recently identified as the previously designated thyroid microsomal antigen (TMA) (9-11). Both the cloning of the human TPO cDNA (12-14) and its expression in mammalian cell lines have been successfully accomplished (1417). The TPO is a thyroid-specific hemoprotein mainly located in the apical plasma membrane of thyrocytes and which is involved in important reactions in the biosynthesis of thyroid hormone. Many techniques have been described for the measurement of anti-TPO autoantibodies as a diagnostic tool for this autoimmune thyroid disease. All the anti-TPO assay systems are based on TPO purified as a microsomal membrane fraction from human thyroid glands and require a consistent source of human TPO since human autoanti1
To whom correspondence should be addressed. Abbreviations: ELISA, enzyme-linked immunosorbent assay; GST, glutathione-S-transferase; IPTG, isopropyl-/9-D-thiogalactoside; TPO, thyroid peroxidase; TMA, thyroid microsomal antigen. Vol. I l l , No. 5, 1992
bodies may be species specific. Therefore, it is necessary to develop expression systems which can provide recombinant human TPO produced either by bacterial or eukaryotic cells as a reproducible antigen source for the immunoassay systems to be independent from patient organs. No reliable and specific assay system is available for the detection of anti-TPO autoantibodies using recombinant human TPO expressed by bacterial cells. In the present manuscript we describe a new approach for a direct mapping of linear autoantigenic epitopes of the human TPO with respect to two objectives: first, we wished to design and establish a new methodology for the mapping of antigenic epitopes using recombinant molecular biology techniques rather than the traditional mapping procedures; second, we wished to demonstrate that recombinant human TPO produced by bacterial cells especially its defined antigenic epitopes might be recognized by anti-TPO autoantibodies, and therefore be useful for the development of specific and reproducible immunoassay systems such as an ELISA for diagnosis of autoimmune thyroid diseases such as the Hashimoto's disease. For expression of recombinant TPO in bacterial cells and mapping of the antigenic epitopes we isolated a human TPO cDNA clone which was used to generate a TPO epitope cDNA library. With a specific anti-TPO autoantiserum we were able to localize and characterize at least one linear TPO-specific autoepitope. MATERIALS AND METHODS Isolation of TPO cDNA Clone and Generation of a TPO Epitope cDNA Library—Poly(A) RNA was isolated from a human thyroid gland of patient with Graves' disease and used for the construction of a cDNA library in LambdaZAPII (Stratagene) with a primary complexity of about 10*
633
634 independent clones according to the standard protocol (18). About 3 X105 clones were screened by plaque hybridization with a 45-mer oligonucleotide which corresponds to the published TPO cDNA sequence (14) between position 2066 and 2110. Nine positive TPO cDNA clones were obtained (pTPO. 1-9) and characterized by DNA sequence analysis. The clone pTPO.9 carried the longest cDNA insert with 3,088 bp in length. For the generation of the TPO epitope cDNA library 25 //g of the cDNA insert of pTPO.9 was isolated and fragmented by sonication with 4 pulses for 30 s at about 250 W. The sonicated cDNA was size-selected into fragments between 300 and 500 bp by agarose gel electrophoresis, end-repaired with T4 DNA polymerase (Pharmacia), and cloned into the Smal site of the expression vector pGEX-2T which codes for a fusion protein with 26 kDa glutathione-Stransferase (19). After transformation of Escherichia coli LE392 (20) the obtained colonies representing the TPO epitope cDNA library (Fig. 1) were collected. Expression and Immunoscreening of the TPO Epitope cDNA Library with Human Anti-TPO Autoantiserum—The TPO epitope cDNA library was plated in a density of about 100-150 cells per 85-mm petri dish and incubated until the colonies have been 1-2 mm in diameter. The colonies were lifted with nitrocellulose niters (Schleicher and Schuell), placed onto prewarmed LB-plates containing 0.5 mM IPTG and incubated for 5 h at 37"C for the expression of the GST fusion proteins. For lysis of the bacteria the niters were exposed to chloroform vapor for 20 min and washed with TBST buffer (10 mM Tris-HCl, pH 8.0,150 mM NaCl, 0.05% Tween-20) by gentle agitation to remove the cell debris. The identification of clones expressing antigenic TPO epitopes were carried out with an immunoreaction using the human anti-TPO autoantiserum in analogy to the Western blotting technique as described below. Western Blotting Analysis of the TPO Epitope Fusion Proteins—For expression of the recombinant TPO epitopes, overnight cultures of the corresponding bacterial stocks were diluted 10-fold with prewarmed LB-medium and incubated for 90 min prior the induction with 0.5 mM IPTG for 5 h at 37'C. After induction and incubation bacterial cells of 100//I culture were sedimented and heat-denatured in SDS-loading buffer. The bacterial proteins were separated in 10% SDS-PAGE under reducing conditions and electrophoretically transferred onto nitrocellulose filters (Amersham) using the trans-blot semi-dry electrophoretically transfer cell (Bio-Rad) with 0.8 mA/ cm2 for 20 min. The unoccupied protein binding sites on the nitrocellulose filters were blocked with 5% nonfat dried milk in TBST-buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.05% Tween-20) for 12-16 h at room temperature. The filters were incubated with a 600-fold dilution of anti-TPO patient sera in TBST buffer for 60 min. The diluted sera were pretreated with 0.5 mg/ml E. coli (LE392) cell lysate, 10 //g/ml SDS-denatured recombinant GST and 1 mM PMSF (phenylmethylsulfonyl fluoride) for 12-16 h at room temperature prior the incubation with antigens. The bound antibodies were visualized with antihuman immunoglobulins conjugated with alkaline phosphatase (Promega) after an incubation for 45 min in TBST buffer followed by a color reaction using the NBT (nitro blue tetrazolium) and BCIP (5-bromo-4-chloro-3-indoyl
B. Frorath et al phosphate) as substrates (Promega). RESULTS Generation of a Human Thyroid Gland cDNA Library and Isolation of a TPO cDNA Clone—A human thyroid gland cDNA library was constructed in the lambda-ZAPII system based on the mRNA isolated from a thyroid gland of a single Graves' disease patient. Nine positive clones were found which were characterized by DNA sequence analysis and compared with the published sequence (14). All TPO cDNA inserts showed a 100% sequence homology and ranged between 990 and 3,088 bp in length. The analysis of the 3'-ends of the TPO clones revealed the existence of two alternative* polyadenylation sites with corresponding polyadenylation signals 65 bases apart (data not shown). The clone with the longest insert designated pTPO.9 possessed an open protein reading frame encoding for the entire 933 amino acids of TPO representing a full-length TPO cDNA clone.
TPO cDNA
sonication
random DNA fragments
gelelectrophoresis sized DNA fragments 1100 - 500 bp)
end- repair subcloning into pGEX-2T
TPO epitope cDNA
library
expression immunoscreening with TPO-specific autoantiserum
positive cDNA clones
characterization
TPO epitope
Fig. 1. Scheme of the mapping of linear autoantigenic TPO epitopes. J. Biochem.
635
Epitope Mapping of Human Thyroid Peroxidase Antigen Fig. 2. Expression and Western blotting analysis of TPO epitope candidates. Candidates carrying sequences coding for potential antigenic TPO epitopes were cultured and induced with 0.5 mM IPTG for 5 h. Total bacterial proteins representing a 100 ft 1 culture were separated in 10% SDS-PAGE and either stained with Coomassie Brilliant Blue (A) or transfered to nitrocellulose filter prior to a Western blotting analysis with a 600-fold dilution of the anti-TPO128 autoantiserum (B). Clones used for analysis: (1) pTPOE.3, (2) pTPOE.7, (3) pTPOE.8, (4) pTPOE.16, (5) pTPOE.18, (6) pTPOE.21, (7) pTPOE.22, and (8) pTPOE.85.
B kDa 00.097.4 68.04 3.0-
29.0-
12
TPO
pTPOS
pTPOE 3 pTPOE 8 pTPOE 18 pTPOE 85 pTPOE.22
—|6loo]—
Fig. 3. Mapping of a linear autoantigenic human TPO epitope. The sequence regions of the positively identified TPO epitope cDNA clones were lined up with the entire TPO cDNA. A stretch of cDNA between positions 2688 and 2871 of the published TPO cDNA sequence (14) coding for 61 amino acids resembled a linear autoantigenic TPO epitope located at the C-terminus of the human TPO (TM, transmembrane spanning region).
Mapping of Linear Autoantigenic Epitopes of the Human TPO with Anti-TPO Specific Patient Sera—To investigate the possibility that patient autoantibodies directed against the TPO are able to recognize linear epitopes within the TPO we constructed first a so-called TPO epitope library as outlined in Fig. 1. The 3.088 kb TPO cDNA insert was isolated, randomly cut in DNA fragments of 300 to 500 bp in length by sonication and subcloned into the prokaryotic expression vector pGEX-2T (19). The pGEX-2T is designed for high-efficient expression of fusion proteins with the glutathione-S-transferase (26 kDa) at their N-termini and the recombinant protein at their C-termini. The recombinant plasmids were used to transfeet the E. coli strain LE392 (20) which was found to be a very suitable host for this vector system (21). For identification of a cDNA clone coding for a specific linear antigenic TPO epitope, the generated TPO epitope cDN A library had to be screened only with classified patient Vol. I l l , No. 5, 1992
3 4 5 6 7 8
12
3 4 5 6 7 8
sera. Therefore, we tested individually about 80 sera from patients with clinically proofed autoimmune thyroid disease by ELJSA using a conventionally purified microsomal membrane-fraction highly enriched in TPO from human thyroid glands as the antigen source (U. Steffens, personal communications). All sera tested reacted positively with this TMA preparation representing specific autoantibodies directed against the human TPO (data not shown). The specificity of this immunoassay was carefully proofed by several control and standard experiments (U. Steffens, personal communications). For further experiments the tested sera were pooled. About 600 colonies of the TPO epitope cDNA library were screened immunologically with the antisera pool which was preabsorbed with bacterial proteins as well as with purified and denatured recombinant GST to minimize unspecific background binding. In all, eight clones could be selected which strongly reacted with anti-TPO antisera. DNA sequence analysis confirmed the correct orientation of the inserted cDNA fragments and their open protein reading frame corresponding to the human TPO amino acid sequence. Western dot blot analysis under reducing conditions with the recombinant TPO proteins and the individual patient sera of the pool used for screening of the epitope cDNA library demonstrated that only one serum designated as anti-TPO128 was able to react with the expressed recombinant TPO (data not shown). Characterization of the cDNA Clones Coding for TPO Autoepitopes—The expressed fusion proteins of the eight positive cDNA clones (pTPOE.3, pTPOE.7, pTPOE.8, pTPOE.16, pTPOE.18, pTPOE.21, pTPOE.22, and pTPOE.85) were characterized by gel electrophoresis (Fig. 2A) and by Western blotting analysis under reducing conditions (Fig. 2B). The expected molecular weights of the expressed fusion proteins were in agreement with the observed molecular weight shown in the stained gel. The GST-TPO autoepitope fusion proteins were highly enriched and were constituted under optimal conditions up to 20% of total cellular proteins. As shown by the Western blotting analysis only five of
636
B. Frorath et al.
the positive candidates, pTPOE.3, pTPOE.8, pTPOE.18, pTPOE.22, and pTPOE.85 (Fig. 2B, lanes 1, 3, 5, 7, and 8) reacted strongly with anti-TPO128 and carried therefore authentic antigenic TPO amino acid sequences. The clones pTPOE.7, pTPOE.16, and pTPOE.21 (Fig. 2B, lanes 2, 4, and 6) were classified immunologically as false positive and represented suitable negative controls. Testing other patient sera of the pool used for the screening also no reaction was observed with these expressed TPO fusion proteins (data not shown). After the alignment of the cDNA sequences derived from the five positive TPO autoepitope clones only one linear epitope was localized between the amino acids #873 and 933 with 61 amino acids in length (Fig. 3). This autoepitope region was located at the C-terminus of the TPO beyond the transmembrane spanning region (amino acids #847-871) within the cisternal domain of the endoplasmic reticulum. Both the ability of expressing of the entire mature TPO in fragments derived statistically from the TPO epitope cDNA library and the inability to detect other antigenic amino acid sequences within the TPO polypeptide were examined (data not shown). All other patient sera tested under these conditions failed completely to recognize linear antigenic epitopes of the TPO. DISCUSSION Thyroid microsomal antibodies are frequently present in serum of the majority of patients with autoimmune thyroid disease and been identified to participate in thyroid cell damage. Reports from several research groups have confirmed that TPO is identical to the thyroid microsomal autoantigen and is one of the major thyroid autoantigens. Many clinical studies have shown that the generation and occurrence of anti-TPO autoantibodies are closely associated with the development of autoimmune thyroid disease (for reviews see Refs. 1-11, 22, and 23). For the diagnosis of this autoimmune thyroid disease several immunoassay systems have been developed for the measurement of anti-TPO autoantibodies (for review see Ref. 6). All these assay systems are using microsomal fraction prepared from human thyroid glands as antigen component designated as thyroid microsomal antigen. However, each human thyroid gland contains varying amounts of TPO antigen which will reflect a wide variability in the purity and quality of obtained membrane fractions (24, 25). It was necessary to develop a new assay system using recombinant TPO as antigen target to eliminate all disadvantages caused by a natural TPO source with its contaminants. For measurements of anti-TPO autoantibodies in serum as a diagnostic tool we wished to develop a suitable immunoassay system based on ELISA technology using recombinant human TPO produced by bacterial cells. Therefore, first we generated a cDNA library from a thyroid gland of a Graves' disease patient and isolated a full-length cDNA clone coding for the entire mature human TPO. Second, we inserted the TPO cDNA into the highefficient prokaryotic expression system pGEX-2T (19) and transfected E. coli cells. Unfortunately, using the fulllength TPO cDNA clone, no expression was observed; maybe either due to non-expressable sequences within the TPO cDNA or due to the exceptional length of the TPO
(data not shown). Therefore, it was necessary to express the large polypeptide of TPO in smaller segments and use defined antigenic epitopes for a reliable immunoassay system. The full-length TPO cDNA was fragmented by sonication and randomly cloned in the expression vector pGEX-2T to generate a TPO epitope cDNA library expressing fusion proteins with GST. The concept of this epitope mapping is based on the ability to select certain antigenic peptide8 among the randomly expressed amino acid sequences by anti-TPO autoantibodies from patients followed by the characterization of the positively identified clones by Western blotting and DNA sequence analysis as shown in Fig. 1. Resulting from the cloning of randomly fragmented cDNA the generated TPO epitope cDNA library should carry all possible autoantigenic regions including adjacent overlapping sequences which were statistically all expressed in bacterial cells. A check of 75 TPO subclones randomly chosen from the epitope library regarding correct protein reading frames and expression efficiency demonstrated that about 98% of the protein coding sequences of TPO in fragments ranging from 60 to 168 amino acids could be expressed (data not shown). Autoantisera from 80 clinical selected patients with autoimmune thyroid disease were used to screen the possible autoepitopes of TPO. All sera were tested either by ELISA or by Western blotting analysis under reducing conditions using conventionally purified microsomal membranes from thyroid glands. Only one serum designated anti-TPO128 positively tested in the ELISA maintained its reactivity with the TMA in the Western blotting analysis under reducing conditions (U. Steffens, personal communications). After analysis of positively characterized TPO cDNA clones coding for antigenic epitopes (Fig. 2) we were able to identify only one autoepitope in a 61 amino acid range located at the C-terminus of the TPO (Fig. 3). It is not immediately apparent how an epitope residing within the cisternal domain of the endoplasmic reticulum might be recognized and presented as an autoantigen. Further investigation explaining the biological relevance of this antigenic epitope is necessary. Interestingly, the anti-TPOl28 autoantiserum was specific to the amino acid sequence of TPO in an almost monoclonal behavior suggesting that the main immunogenic determinant should be located nearby the C-terminus of TPO polypeptide (26). Comparable studies with other autoantisera than anti-TPOl28 from the selected pool failed completely in identifying the same or other autoepitopes in the TPO epitope cDNA library (data not shown). More detailed studies revealed that the majority of sera obtained from patients with autoimmune thyroid disease can only react with the natural conformation of TPO and not with unglycosylated and non-modified recombinant TPO such as expressed by bacterial cells (H. Bertold et al, manuscript in preparation). Less than 1% of sera tested here in this study could reacted with a linear autoantigenic epitope of TPO. In conclusion, we have demonstrated the availability of basic molecular biology techniques for mapping of an antigenic epitope using the human TPO as an antigenic target. In contrary to other's findings (27-31) we could not confirm several linear antigenic epitopes existing throughout the TPO polypeptide except one located at the C-terminus. About 98% of patient sera were not able to recognize J. Biochem.
Epitope Mapping of Human Thyroid Peroxidase Antigen sequential antigenic determinants of TPO polypeptide. Probably they are only specific to so-called conformational determinants (32, 33). Therefore, the approach with recombinant TPO produced by prokaryotic expression systems is not useful for development of a suitable immunoassay system to measure anti-TPO autoantibodies for diagnosis of autoimmune thyroid disease. The authors like to thank Drs. B.E. Wenzel, University Lttbeck, F.R.G., J.P. Banga, King's College, London, and B. Liedvogel (ELJAS, Freiburg) for the helpful discussions. REFERENCES 1. Banga, J.P., Barnett, P.S., Mahadevan, D., & McGregor, A.M. (1989) Ewr. J. Clin. Invest 19, 107-116 2. Weetman, A.P. & McGregor, A.M. (1984) Endocr. Rev. 5, 309355 3. Trotter, W.R., Belyavin, G., & Wadhans, A. (1957) Proc. R Soc. Med 50, 961-964 4. Rees Smith, B., McLachlan, S.M., & Furmaniak, J. (1988) Endocr. Rev. 9, 106-121 5. Carayon, P. & Ruf, J. (1990) in Thyroperoxidase and Thyroid Autoimmunity (Carayon, P. & Ruf, J., eds.) Vol. 207, pp. 1-358, Colloque INSERM/John Libbey Eurotext, Paris 6. Banga, J.P., Barnett, P.S., & McGregor, A.M. (1991) Autoimmunity 8, 353-343 7. McLachlan, S.M., Atherton, M.C., Nakajima, Y., Napier, J., Jordan, R.K., Clarck, F., & Rees Smith, B. (1990) Immunology 79, 182-188 8. Yokoyama, N., Taurog, A., & Klee, G.G. (1989) J. Clin. Endocrinol Metab. 68, 766-773 9. Portmann, L., Hamada, N., Heinrich, G., &De Groot, L.J. (1985) J. Clin. EndocrinoL Metab. 61, 1001-1003 10. Cramocka, B., Ruf, J., Ferrand, M., Carayon, P., & Lissitzky, S. (1985) FEBS Lett. 190, 147-152 11. Okamoto, Y., Hamada, N., Saito, H., Ohno, M., Noh, J., Ito, K., & Morii, H. (1988) J. Clin. Endocrinol. Metab. 68, 730-734 12. Magnusson, R.P., Chazenbalk, G.D., Gestautas, J., Seto, P., Filetti, S., DeGroot, L.J., &Rapoport, B. (1987) MoL Endocrinol. 1, 866-861 13. Libert, F., Ruel, J., Ludgate, M., Swillens, S., Alexander, N., Vassart, G., & Dinsaert, C. (1987) EMBO J. 6, 4193-4196 14. Kimura, S., Kotani, T., McBridge, O.W., Umeki, K., Hirai, K., Nakayama, T., &Ohtaki, S. (1987) Proc. NatL Acad. Sci. U.S.A. 84, 5555-5559
Vol. I l l , No. 5, 1992
637 15. Kaufman, K.D., Filetti, S., Seto, P., & Rapoport, B. (1990)
Mapping of a Linear Autoantigenic Epitope within the Human Thyroid Peroxidase Using Recombinant DNA Techniques Brigitte Frorath,* Charles C. Abney,* Mirco Scanarini,* Heike Berthold,* Nick Hunt,** and Wolfgang Northemann*1 'Department of Molecular Biology, ELJAS Entwicklungslabor, Obere Hardtstrasse 18, D-7800 Freiburg, F.R.G.; and "Institute for Hormone and Fertility Research, D-2000 Hamburg, F.RG. Received for publication, January 7, 1992
Autoantibodies directed against the thyroid peroxidase (TPO), the thyroid microsomal antigen, are widely used to diagnose human autoimmune thyroid disease. A cloned 3.088 kb cDNA coding for the entire mature human TPO was isolated from a cDNA library derived from a pathological thyroid gland of a Graves' disease patient and used further to generate a so-called TPO epitope cDNA library in order to map linear autoantigenic epitopes involving a recombinant molecular biology approach. The TPO epitope cDNA library consisting of randomly fragmented cDNA sequences inserted in the expression vector pGEX-2T was expressed in Escherichia coli and screened with characterized anti-TPO autoantisera from Hashimoto's disease patients. All the sera were positively tested with a purified thyroid microsomal antigen fraction (TMA/TPO). Only about 1% of examined autoantisera were able to recognize bacterial expressed recombinant TPO representing sequential antigenic determinants. A corresponding autoantigenic epitope with 61 ami no acids in length was located at the C-terminus of human TPO.
Organ specific autoimmune diseases are common in man which include thyroid autoimmune disease, insulin-dependent diabetes mellitus, myasthenia gravis, and others which are all mediated by abnormal generation of autoantibodies (for review see Ref. 1). The autoantibodies directed to the thyroid tissue are very important clinical markers for the human autoimmune thyroid disease which is a relatively common disease affecting about 5% of the population (2). The characteristics of this disease are mainly tissue destruction in lymphocytic thyroiditis (Hashimoto's disease) and the abnormal modulation of the thyrotropin receptor by autoantibodies leading to hyperthyroidism (Graves' disease) (1, 3-5). Thyroid-specific autoantibodies are directed against several components of the thyroid gland, including thyroid peroxidase (TPO) which is one of the major thyroid-specific antigens {6-8). The TPO was recently identified as the previously designated thyroid microsomal antigen (TMA) (9-11). Both the cloning of the human TPO cDNA (12-14) and its expression in mammalian cell lines have been successfully accomplished (1417). The TPO is a thyroid-specific hemoprotein mainly located in the apical plasma membrane of thyrocytes and which is involved in important reactions in the biosynthesis of thyroid hormone. Many techniques have been described for the measurement of anti-TPO autoantibodies as a diagnostic tool for this autoimmune thyroid disease. All the anti-TPO assay systems are based on TPO purified as a microsomal membrane fraction from human thyroid glands and require a consistent source of human TPO since human autoanti1
To whom correspondence should be addressed. Abbreviations: ELISA, enzyme-linked immunosorbent assay; GST, glutathione-S-transferase; IPTG, isopropyl-/9-D-thiogalactoside; TPO, thyroid peroxidase; TMA, thyroid microsomal antigen. Vol. I l l , No. 5, 1992
bodies may be species specific. Therefore, it is necessary to develop expression systems which can provide recombinant human TPO produced either by bacterial or eukaryotic cells as a reproducible antigen source for the immunoassay systems to be independent from patient organs. No reliable and specific assay system is available for the detection of anti-TPO autoantibodies using recombinant human TPO expressed by bacterial cells. In the present manuscript we describe a new approach for a direct mapping of linear autoantigenic epitopes of the human TPO with respect to two objectives: first, we wished to design and establish a new methodology for the mapping of antigenic epitopes using recombinant molecular biology techniques rather than the traditional mapping procedures; second, we wished to demonstrate that recombinant human TPO produced by bacterial cells especially its defined antigenic epitopes might be recognized by anti-TPO autoantibodies, and therefore be useful for the development of specific and reproducible immunoassay systems such as an ELISA for diagnosis of autoimmune thyroid diseases such as the Hashimoto's disease. For expression of recombinant TPO in bacterial cells and mapping of the antigenic epitopes we isolated a human TPO cDNA clone which was used to generate a TPO epitope cDNA library. With a specific anti-TPO autoantiserum we were able to localize and characterize at least one linear TPO-specific autoepitope. MATERIALS AND METHODS Isolation of TPO cDNA Clone and Generation of a TPO Epitope cDNA Library—Poly(A) RNA was isolated from a human thyroid gland of patient with Graves' disease and used for the construction of a cDNA library in LambdaZAPII (Stratagene) with a primary complexity of about 10*
633
634 independent clones according to the standard protocol (18). About 3 X105 clones were screened by plaque hybridization with a 45-mer oligonucleotide which corresponds to the published TPO cDNA sequence (14) between position 2066 and 2110. Nine positive TPO cDNA clones were obtained (pTPO. 1-9) and characterized by DNA sequence analysis. The clone pTPO.9 carried the longest cDNA insert with 3,088 bp in length. For the generation of the TPO epitope cDNA library 25 //g of the cDNA insert of pTPO.9 was isolated and fragmented by sonication with 4 pulses for 30 s at about 250 W. The sonicated cDNA was size-selected into fragments between 300 and 500 bp by agarose gel electrophoresis, end-repaired with T4 DNA polymerase (Pharmacia), and cloned into the Smal site of the expression vector pGEX-2T which codes for a fusion protein with 26 kDa glutathione-Stransferase (19). After transformation of Escherichia coli LE392 (20) the obtained colonies representing the TPO epitope cDNA library (Fig. 1) were collected. Expression and Immunoscreening of the TPO Epitope cDNA Library with Human Anti-TPO Autoantiserum—The TPO epitope cDNA library was plated in a density of about 100-150 cells per 85-mm petri dish and incubated until the colonies have been 1-2 mm in diameter. The colonies were lifted with nitrocellulose niters (Schleicher and Schuell), placed onto prewarmed LB-plates containing 0.5 mM IPTG and incubated for 5 h at 37"C for the expression of the GST fusion proteins. For lysis of the bacteria the niters were exposed to chloroform vapor for 20 min and washed with TBST buffer (10 mM Tris-HCl, pH 8.0,150 mM NaCl, 0.05% Tween-20) by gentle agitation to remove the cell debris. The identification of clones expressing antigenic TPO epitopes were carried out with an immunoreaction using the human anti-TPO autoantiserum in analogy to the Western blotting technique as described below. Western Blotting Analysis of the TPO Epitope Fusion Proteins—For expression of the recombinant TPO epitopes, overnight cultures of the corresponding bacterial stocks were diluted 10-fold with prewarmed LB-medium and incubated for 90 min prior the induction with 0.5 mM IPTG for 5 h at 37'C. After induction and incubation bacterial cells of 100//I culture were sedimented and heat-denatured in SDS-loading buffer. The bacterial proteins were separated in 10% SDS-PAGE under reducing conditions and electrophoretically transferred onto nitrocellulose filters (Amersham) using the trans-blot semi-dry electrophoretically transfer cell (Bio-Rad) with 0.8 mA/ cm2 for 20 min. The unoccupied protein binding sites on the nitrocellulose filters were blocked with 5% nonfat dried milk in TBST-buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.05% Tween-20) for 12-16 h at room temperature. The filters were incubated with a 600-fold dilution of anti-TPO patient sera in TBST buffer for 60 min. The diluted sera were pretreated with 0.5 mg/ml E. coli (LE392) cell lysate, 10 //g/ml SDS-denatured recombinant GST and 1 mM PMSF (phenylmethylsulfonyl fluoride) for 12-16 h at room temperature prior the incubation with antigens. The bound antibodies were visualized with antihuman immunoglobulins conjugated with alkaline phosphatase (Promega) after an incubation for 45 min in TBST buffer followed by a color reaction using the NBT (nitro blue tetrazolium) and BCIP (5-bromo-4-chloro-3-indoyl
B. Frorath et al phosphate) as substrates (Promega). RESULTS Generation of a Human Thyroid Gland cDNA Library and Isolation of a TPO cDNA Clone—A human thyroid gland cDNA library was constructed in the lambda-ZAPII system based on the mRNA isolated from a thyroid gland of a single Graves' disease patient. Nine positive clones were found which were characterized by DNA sequence analysis and compared with the published sequence (14). All TPO cDNA inserts showed a 100% sequence homology and ranged between 990 and 3,088 bp in length. The analysis of the 3'-ends of the TPO clones revealed the existence of two alternative* polyadenylation sites with corresponding polyadenylation signals 65 bases apart (data not shown). The clone with the longest insert designated pTPO.9 possessed an open protein reading frame encoding for the entire 933 amino acids of TPO representing a full-length TPO cDNA clone.
TPO cDNA
sonication
random DNA fragments
gelelectrophoresis sized DNA fragments 1100 - 500 bp)
end- repair subcloning into pGEX-2T
TPO epitope cDNA
library
expression immunoscreening with TPO-specific autoantiserum
positive cDNA clones
characterization
TPO epitope
Fig. 1. Scheme of the mapping of linear autoantigenic TPO epitopes. J. Biochem.
635
Epitope Mapping of Human Thyroid Peroxidase Antigen Fig. 2. Expression and Western blotting analysis of TPO epitope candidates. Candidates carrying sequences coding for potential antigenic TPO epitopes were cultured and induced with 0.5 mM IPTG for 5 h. Total bacterial proteins representing a 100 ft 1 culture were separated in 10% SDS-PAGE and either stained with Coomassie Brilliant Blue (A) or transfered to nitrocellulose filter prior to a Western blotting analysis with a 600-fold dilution of the anti-TPO128 autoantiserum (B). Clones used for analysis: (1) pTPOE.3, (2) pTPOE.7, (3) pTPOE.8, (4) pTPOE.16, (5) pTPOE.18, (6) pTPOE.21, (7) pTPOE.22, and (8) pTPOE.85.
B kDa 00.097.4 68.04 3.0-
29.0-
12
TPO
pTPOS
pTPOE 3 pTPOE 8 pTPOE 18 pTPOE 85 pTPOE.22
—|6loo]—
Fig. 3. Mapping of a linear autoantigenic human TPO epitope. The sequence regions of the positively identified TPO epitope cDNA clones were lined up with the entire TPO cDNA. A stretch of cDNA between positions 2688 and 2871 of the published TPO cDNA sequence (14) coding for 61 amino acids resembled a linear autoantigenic TPO epitope located at the C-terminus of the human TPO (TM, transmembrane spanning region).
Mapping of Linear Autoantigenic Epitopes of the Human TPO with Anti-TPO Specific Patient Sera—To investigate the possibility that patient autoantibodies directed against the TPO are able to recognize linear epitopes within the TPO we constructed first a so-called TPO epitope library as outlined in Fig. 1. The 3.088 kb TPO cDNA insert was isolated, randomly cut in DNA fragments of 300 to 500 bp in length by sonication and subcloned into the prokaryotic expression vector pGEX-2T (19). The pGEX-2T is designed for high-efficient expression of fusion proteins with the glutathione-S-transferase (26 kDa) at their N-termini and the recombinant protein at their C-termini. The recombinant plasmids were used to transfeet the E. coli strain LE392 (20) which was found to be a very suitable host for this vector system (21). For identification of a cDNA clone coding for a specific linear antigenic TPO epitope, the generated TPO epitope cDN A library had to be screened only with classified patient Vol. I l l , No. 5, 1992
3 4 5 6 7 8
12
3 4 5 6 7 8
sera. Therefore, we tested individually about 80 sera from patients with clinically proofed autoimmune thyroid disease by ELJSA using a conventionally purified microsomal membrane-fraction highly enriched in TPO from human thyroid glands as the antigen source (U. Steffens, personal communications). All sera tested reacted positively with this TMA preparation representing specific autoantibodies directed against the human TPO (data not shown). The specificity of this immunoassay was carefully proofed by several control and standard experiments (U. Steffens, personal communications). For further experiments the tested sera were pooled. About 600 colonies of the TPO epitope cDNA library were screened immunologically with the antisera pool which was preabsorbed with bacterial proteins as well as with purified and denatured recombinant GST to minimize unspecific background binding. In all, eight clones could be selected which strongly reacted with anti-TPO antisera. DNA sequence analysis confirmed the correct orientation of the inserted cDNA fragments and their open protein reading frame corresponding to the human TPO amino acid sequence. Western dot blot analysis under reducing conditions with the recombinant TPO proteins and the individual patient sera of the pool used for screening of the epitope cDNA library demonstrated that only one serum designated as anti-TPO128 was able to react with the expressed recombinant TPO (data not shown). Characterization of the cDNA Clones Coding for TPO Autoepitopes—The expressed fusion proteins of the eight positive cDNA clones (pTPOE.3, pTPOE.7, pTPOE.8, pTPOE.16, pTPOE.18, pTPOE.21, pTPOE.22, and pTPOE.85) were characterized by gel electrophoresis (Fig. 2A) and by Western blotting analysis under reducing conditions (Fig. 2B). The expected molecular weights of the expressed fusion proteins were in agreement with the observed molecular weight shown in the stained gel. The GST-TPO autoepitope fusion proteins were highly enriched and were constituted under optimal conditions up to 20% of total cellular proteins. As shown by the Western blotting analysis only five of
636
B. Frorath et al.
the positive candidates, pTPOE.3, pTPOE.8, pTPOE.18, pTPOE.22, and pTPOE.85 (Fig. 2B, lanes 1, 3, 5, 7, and 8) reacted strongly with anti-TPO128 and carried therefore authentic antigenic TPO amino acid sequences. The clones pTPOE.7, pTPOE.16, and pTPOE.21 (Fig. 2B, lanes 2, 4, and 6) were classified immunologically as false positive and represented suitable negative controls. Testing other patient sera of the pool used for the screening also no reaction was observed with these expressed TPO fusion proteins (data not shown). After the alignment of the cDNA sequences derived from the five positive TPO autoepitope clones only one linear epitope was localized between the amino acids #873 and 933 with 61 amino acids in length (Fig. 3). This autoepitope region was located at the C-terminus of the TPO beyond the transmembrane spanning region (amino acids #847-871) within the cisternal domain of the endoplasmic reticulum. Both the ability of expressing of the entire mature TPO in fragments derived statistically from the TPO epitope cDNA library and the inability to detect other antigenic amino acid sequences within the TPO polypeptide were examined (data not shown). All other patient sera tested under these conditions failed completely to recognize linear antigenic epitopes of the TPO. DISCUSSION Thyroid microsomal antibodies are frequently present in serum of the majority of patients with autoimmune thyroid disease and been identified to participate in thyroid cell damage. Reports from several research groups have confirmed that TPO is identical to the thyroid microsomal autoantigen and is one of the major thyroid autoantigens. Many clinical studies have shown that the generation and occurrence of anti-TPO autoantibodies are closely associated with the development of autoimmune thyroid disease (for reviews see Refs. 1-11, 22, and 23). For the diagnosis of this autoimmune thyroid disease several immunoassay systems have been developed for the measurement of anti-TPO autoantibodies (for review see Ref. 6). All these assay systems are using microsomal fraction prepared from human thyroid glands as antigen component designated as thyroid microsomal antigen. However, each human thyroid gland contains varying amounts of TPO antigen which will reflect a wide variability in the purity and quality of obtained membrane fractions (24, 25). It was necessary to develop a new assay system using recombinant TPO as antigen target to eliminate all disadvantages caused by a natural TPO source with its contaminants. For measurements of anti-TPO autoantibodies in serum as a diagnostic tool we wished to develop a suitable immunoassay system based on ELISA technology using recombinant human TPO produced by bacterial cells. Therefore, first we generated a cDNA library from a thyroid gland of a Graves' disease patient and isolated a full-length cDNA clone coding for the entire mature human TPO. Second, we inserted the TPO cDNA into the highefficient prokaryotic expression system pGEX-2T (19) and transfected E. coli cells. Unfortunately, using the fulllength TPO cDNA clone, no expression was observed; maybe either due to non-expressable sequences within the TPO cDNA or due to the exceptional length of the TPO
(data not shown). Therefore, it was necessary to express the large polypeptide of TPO in smaller segments and use defined antigenic epitopes for a reliable immunoassay system. The full-length TPO cDNA was fragmented by sonication and randomly cloned in the expression vector pGEX-2T to generate a TPO epitope cDNA library expressing fusion proteins with GST. The concept of this epitope mapping is based on the ability to select certain antigenic peptide8 among the randomly expressed amino acid sequences by anti-TPO autoantibodies from patients followed by the characterization of the positively identified clones by Western blotting and DNA sequence analysis as shown in Fig. 1. Resulting from the cloning of randomly fragmented cDNA the generated TPO epitope cDNA library should carry all possible autoantigenic regions including adjacent overlapping sequences which were statistically all expressed in bacterial cells. A check of 75 TPO subclones randomly chosen from the epitope library regarding correct protein reading frames and expression efficiency demonstrated that about 98% of the protein coding sequences of TPO in fragments ranging from 60 to 168 amino acids could be expressed (data not shown). Autoantisera from 80 clinical selected patients with autoimmune thyroid disease were used to screen the possible autoepitopes of TPO. All sera were tested either by ELISA or by Western blotting analysis under reducing conditions using conventionally purified microsomal membranes from thyroid glands. Only one serum designated anti-TPO128 positively tested in the ELISA maintained its reactivity with the TMA in the Western blotting analysis under reducing conditions (U. Steffens, personal communications). After analysis of positively characterized TPO cDNA clones coding for antigenic epitopes (Fig. 2) we were able to identify only one autoepitope in a 61 amino acid range located at the C-terminus of the TPO (Fig. 3). It is not immediately apparent how an epitope residing within the cisternal domain of the endoplasmic reticulum might be recognized and presented as an autoantigen. Further investigation explaining the biological relevance of this antigenic epitope is necessary. Interestingly, the anti-TPOl28 autoantiserum was specific to the amino acid sequence of TPO in an almost monoclonal behavior suggesting that the main immunogenic determinant should be located nearby the C-terminus of TPO polypeptide (26). Comparable studies with other autoantisera than anti-TPOl28 from the selected pool failed completely in identifying the same or other autoepitopes in the TPO epitope cDNA library (data not shown). More detailed studies revealed that the majority of sera obtained from patients with autoimmune thyroid disease can only react with the natural conformation of TPO and not with unglycosylated and non-modified recombinant TPO such as expressed by bacterial cells (H. Bertold et al, manuscript in preparation). Less than 1% of sera tested here in this study could reacted with a linear autoantigenic epitope of TPO. In conclusion, we have demonstrated the availability of basic molecular biology techniques for mapping of an antigenic epitope using the human TPO as an antigenic target. In contrary to other's findings (27-31) we could not confirm several linear antigenic epitopes existing throughout the TPO polypeptide except one located at the C-terminus. About 98% of patient sera were not able to recognize J. Biochem.
Epitope Mapping of Human Thyroid Peroxidase Antigen sequential antigenic determinants of TPO polypeptide. Probably they are only specific to so-called conformational determinants (32, 33). Therefore, the approach with recombinant TPO produced by prokaryotic expression systems is not useful for development of a suitable immunoassay system to measure anti-TPO autoantibodies for diagnosis of autoimmune thyroid disease. The authors like to thank Drs. B.E. Wenzel, University Lttbeck, F.R.G., J.P. Banga, King's College, London, and B. Liedvogel (ELJAS, Freiburg) for the helpful discussions. REFERENCES 1. Banga, J.P., Barnett, P.S., Mahadevan, D., & McGregor, A.M. (1989) Ewr. J. Clin. Invest 19, 107-116 2. Weetman, A.P. & McGregor, A.M. (1984) Endocr. Rev. 5, 309355 3. Trotter, W.R., Belyavin, G., & Wadhans, A. (1957) Proc. R Soc. Med 50, 961-964 4. Rees Smith, B., McLachlan, S.M., & Furmaniak, J. (1988) Endocr. Rev. 9, 106-121 5. Carayon, P. & Ruf, J. (1990) in Thyroperoxidase and Thyroid Autoimmunity (Carayon, P. & Ruf, J., eds.) Vol. 207, pp. 1-358, Colloque INSERM/John Libbey Eurotext, Paris 6. Banga, J.P., Barnett, P.S., & McGregor, A.M. (1991) Autoimmunity 8, 353-343 7. McLachlan, S.M., Atherton, M.C., Nakajima, Y., Napier, J., Jordan, R.K., Clarck, F., & Rees Smith, B. (1990) Immunology 79, 182-188 8. Yokoyama, N., Taurog, A., & Klee, G.G. (1989) J. Clin. Endocrinol Metab. 68, 766-773 9. Portmann, L., Hamada, N., Heinrich, G., &De Groot, L.J. (1985) J. Clin. EndocrinoL Metab. 61, 1001-1003 10. Cramocka, B., Ruf, J., Ferrand, M., Carayon, P., & Lissitzky, S. (1985) FEBS Lett. 190, 147-152 11. Okamoto, Y., Hamada, N., Saito, H., Ohno, M., Noh, J., Ito, K., & Morii, H. (1988) J. Clin. Endocrinol. Metab. 68, 730-734 12. Magnusson, R.P., Chazenbalk, G.D., Gestautas, J., Seto, P., Filetti, S., DeGroot, L.J., &Rapoport, B. (1987) MoL Endocrinol. 1, 866-861 13. Libert, F., Ruel, J., Ludgate, M., Swillens, S., Alexander, N., Vassart, G., & Dinsaert, C. (1987) EMBO J. 6, 4193-4196 14. Kimura, S., Kotani, T., McBridge, O.W., Umeki, K., Hirai, K., Nakayama, T., &Ohtaki, S. (1987) Proc. NatL Acad. Sci. U.S.A. 84, 5555-5559
Vol. I l l , No. 5, 1992
637 15. Kaufman, K.D., Filetti, S., Seto, P., & Rapoport, B. (1990)
