The key residues in the immunodominant region of human thyroid peroxidase were identified

Transcription

1 International Immunology, Vol. 18, No. 7, pp doi: /intimm/dxl042 ª The Japanese Society for Immunology All rights reserved. For permissions, please The key residues in the immunodominant region of human thyroid peroxidase were identified Sandra A. Rebuffat 1, Damien Bresson 1,2, Brigitte Nguyen 1 and Sylvie Péraldi-Roux 1 1 CNRS UMR 5160, Faculté de Pharmacie, Montpellier Cedex 5, France 2 Present address: Department of Developmental Immunology-3, La Jolla Institute for Allergy and Immunology, San Diego, CA 92121, USA Keywords: auto-antibodies, autoimmune thyroid disease, autoimmunity, immunodominant region, thyroid peroxidase Abstract Auto-antibodies (aabs) to thyroid peroxidase (TPO) interact with a restricted immunodominant region (IDR) divided into two overlapping regions A and B. Among the five major regions structuring the IDR/B, regions , , , and , region constitutes an important anchor point for the binding of TPO-specific aabs in sera from Hashimoto s and Graves patients. We combined site-directed mutagenesis and expression of TPO mutants in stably transfected CHO cells to precisely define the critical residues in that region. By using flow cytometry and ELISA, we identified four amino acid residues, H353, D358, S359 and R361, that contribute to the interaction between human TPO and anti-tpo aabs. This identification of these contributing amino acid residues in the IDR allowed us to more precisely depict contours of the IDR. Introduction Thyroid peroxidase (TPO) is a key enzyme involved in the biosynthesis of thyroid hormones. Localized on the apical membrane of thyrocytes, it catalyzes the iodination and coupling of iodotyrosine residues on thyroglobulin to produce thyroxine and 3,39,5-triiodothyronine (1, 2). Besides its role in the thyroid function, TPO constitutes one of the major autoantigens involved in the autoimmune thyroid disease (AITD) (3). Auto-antibodies (aabs) against TPO are present in high concentrations in 90% of Hashimoto s thyroiditis and 75% of Graves disease patients and are invaluable markers to diagnose such diseases. They are thought to be involved in thyroid cell destruction through cytotoxic mechanisms mediated by effector cells and/or complement activation (4, 5). More importantly, TPO-specific B cells have been found to be involved in antigen presentation to auto-aggressive T cells (3, 6 8) and may favor the maintenance or exacerbation of AITD. Consequently, a delineation of the epitopes recognized by anti-tpo aabs is crucial for the comprehension of the mechanisms involved in the pathogenesis of AITD. Such an information would also allow us to use a rational approach to design therapeutic agents such as peptides that could deviate and/or block the undesirable autoimmune response when used in combination with other immunotherapies (systemic antibody treatment, antigen-specific immunizations or strategies generating antigen-specific regulatory T cells), particularly for Hashimoto s thyroiditis. For several years, it has been known that anti-tpo aabs recognize conformational epitopes that are highly dependent on the three-dimensional structure of the TPO molecule (9). These TPO aabs are known to be restricted to two immunodominant regions (IDRs) named A and B, containing different but adjacent surface epitopes (10 12). Recently, with the aim of localizing the discontinuous IDR on the human thyroid peroxidase (htpo), we have used a new strategy combining two technological advances: (i) selection of mimotopes by screening phage-displayed peptide libraries on an IDR/B-specific human recombinant anti-tpo aab (T13) mimicking anti-tpo aabs from patients sera and (ii) sequence alignment of the selected mimotopes on the primary sequence of htpo (13). Four distinct regions, distributed between the myeloperoxidase-like domain (regions , and ) and the control complement protein-like domain (region ), were identified as being a part of the IDR. Interestingly, mutation of regions , and abrogated specifically the binding of human anti-tpo aabs from Hashimoto s and Graves disease patients sera. Furthermore, using a number of different approaches, some key residues in regions and [K713 to D717 Correspondence to: S. Péraldi-Roux; sylvie.roux@cpbs.univ-montp1.fr Received 11 January 2006, accepted 18 April 2006 Transmitting editor: A. Cooke Advance Access publication 30 May 2006

2 1092 Key residues in the TPO immunodominant region B (14, 15) and Y772 (16)] were identified as being contributing amino acid residues in the interaction of IDR/B-specific human anti-tpo aabs with TPO. However, the region has not yet been studied in detail, even though this region is strongly recognized by anti-tpo aabs present in the serum of patients suffering from AITD. The aim of the present study was to identify the key amino acid residues taking part in the binding of human anti-tpo aabs (recombinant as well as serum antibodies from patients) to htpo in the region For this purpose, we combined a site-directed mutagenesis approach with a stable expression of TPO mutants at the surface of CHO cells. We demonstrated that mutation at position 358 led to a loss of the binding of human anti-tpo aabs, as efficient as a mutation of the entire region We also demonstrated that amino acid residues D358 and R361 participate in the binding of (i) the human recombinant anti-tpo aab, named ICA1, previously obtained from the in-cell phage-displayed library which mimics the variable heavy/variable light (VH/VL) pairing in vivo (17, 18) and (ii) anti-tpo aabs present in the serum of patients suffering from Hashimoto s and Graves diseases. Methods Table 1. Description of sera from patients suffering from autoimmune diseases Patient no. Disease TPO aab titer a (units ml ÿ1 ) Patients sera and anti-tpo aabs Sera from patients suffering from Graves disease, patients with Hashimoto s thyroiditis and healthy donors were obtained from L. Baldet (Lapeyronie University Hospital, Montpellier, France) (Table 1). The anti-tpo aab titers were determined by radioimmunoassay using the TPO-AB-CT Kit (CIS bio, Gyf sur Yvette, France). The patients sera were characterized further for the presence of anti-thyroid-stimulating hormone receptor aabs by radioimmunoassay using the TSH receptor aabs human kit (BRAHMS, Hemigsdorf, Germany). As controls, sera from 10 patients suffering from other autoimmune and non-autoimmune affections, five with human systemic lupus erythematosus and five with multiple myelomas, were obtained from S. Benzaken (Archet Hospital, Nice, France) and J.-L. Delarbre (University Hospital, Nîmes, France). The rabbit polyclonal anti-tpo antibody P14 was provided by J.-P. Banga and A. Gardas (19, 20). The human recombinant aab T13 expressed as IgG1 by using the baculovirus/ insect cells system has been described (13). The human recombinant aab ICA1 was produced in our laboratory using the same protocol. Directed mutagenesis and stable expression of wild-type and mutated htpo The full-length wild-type (wt) htpo was previously cloned in the pcdna5/frt expression vector from the Flp-In TM System (Invitrogen Life Technologies). In the region , amino acid residues (except A354, A362 and G360) were individually mutated to alanine (A). All mutants were constructed by overlap extension PCR (Fig. 1) as described previously (21). The final PCR products were cloned in the full-length htpo cdna by using the unique restriction endonuclease sites SfiI and ClaI (New England Biolabs). All sequences were verified by the dideoxynucleotide termination method (22). Then, the Flp-In TM System was used to generate isogenic stable CHO cell lines expressing wt and mutated htpo according to the manufacturer s instructions. Briefly, mutated and wt htpo were obtained by co-transfection of the hygromycin-resistant pcdna5/frt plasmids each bearing a TPO mutant with the pog44 plasmid, which constitutively expresses the Flp recombinase in the selected zeocin-resistant CHO host cell line already used in our laboratory (13). The clones of interest were selected for hygromycin resistance and zeocin sensitivity; then, the TPO-expressing CHO cells were sorted and cloned by using an anti-tpo antibody on the FACSVantage SE Turbosort (Becton Dickinson). Cloning and expression of TPO fully mutated in the regions (TPO ) and (TPO ) were previously described (13). Transfected CHO cells expressing mutated and wt htpo were routinely grown in DMEM/nutrient mixture F-12 (Invitrogen Life Technologies) supplemented with 10% FCS (PAA Laboratories) TSH (microunits ml ÿ1 ) T3 (pg ml ÿ1 ) T4 (ng dl ÿ1 ) 1 Graves Graves ND ND Graves Graves Graves ND 6 Hashimoto ,61 ND 7 Hashimoto ND ND 8 Hashimoto ND 9 Hashimoto ND 10 Hashimoto ND Normal value < <150 TRAK b (units ml ÿ1 ) TSH, thyroid-stimulating hormone; T3, triiodothyronine; T4, thyroxine; TRAK, TSH receptor aabs; ND, not determined. a The values were obtained by radioimmunoassay and standardized to the WHO/National Institute for Biological Standards and Control 66/387 (see Methods). b The values were obtained by a radio receptor assay using a DYNOtest TSH receptor aabs human kit and standardized to the WHO 90/672 (see Methods).

3 Key residues in the TPO immunodominant region B 1093 Fig. 1. Directed mutagenesis. (A) Strategy of mutagenesis used to produce the specific mutants of the TPO in the region of TPO. A first PCR (1) generates two fragments: a constant fragment by using primers 1 and 3 (PCR1) and different fragments for each mutation by using primers 2 and 4 (PCR2). An overlap PCR allowed us to amplify a third fragment (2). The final PCR product of 1100 bp, containing the mutation, was cloned into the pcdna5/frt vector encoding for the full-length htpo by using two single restriction sites, SfiI and ClaI, which border the region (3). (B) The size of each PCR product was controlled by electrophoresis and visualized on a gel of 1% agarose. (C) Primers used for the mutagenesis are described in the table. containing 2 mm L-glutamine, 100 lg ml ÿ1 penicillin and 100 lg ml ÿ1 streptomycin (Sigma-Aldrich Co.) at 37 C in a humidified incubator with 5% CO 2. Flow cytometry analysis of wt or mutated htpo expressed at the surface of stably transfected CHO cells Cells were removed from the culture flasks by using HEPES EDTA buffer (HEPES 10 mm, EDTA 3 mg ml ÿ1, ph 7.0), rinsed and pelleted (5 min, 1000 r.p.m., 4 C) in Dulbecco s PBS (D-PBS) (CAMBREX Bio Science) containing 2% FCS (FACS buffer). The cells (~10 6 ) were incubated with 200 ll of FACS buffer containing 10 lg ml ÿ1 of human recombinant anti-tpo aab (T13) or 10 lg ml ÿ1 of rabbit anti-peptide P14 for 90 min at 4 C. The cells were washed twice and then incubated in 200 ll of FACS buffer with 10 lgml ÿ1 of fluorescein-conjugated anti-human IgG c-chain-specific (Sigma) or fluoresceinconjugated anti-rabbit IgG (H&L; Rockland) for 60 min at 4 C in the dark. As negative control, cells were incubated with only the secondary antibody. After washings with D-PBS, the cells were analyzed ( events) on a cytofluorometer (FACScan, Becton Dickinson). Membrane protein extraction from CHO cells Stably transfected CHO cells were washed three times with D-PBS and scraped at 4 C. Membrane protein extraction was performed as follows. After centrifugation at 1000 r.p.m. for 5 min at 4 C, membrane proteins were solubilized by adding 500 ll of lysis buffer (50 mm Tris HCl, ph 7.3, 150 mm NaCl, 5 mm EDTA, 10% glycerol, 1% Triton X-100 and a protease inhibitor mixture tablet per 10 ml of lysis buffer; Roche Molecular Biochemicals) and subjected to two freeze thaw cycles. After centrifugation at r.p.m. for 30 min at 4 C, the supernatants containing membrane proteins were recovered and stored at ÿ80 C. The protein concentrations were evaluated by the BCA protein assay reagent (Pierce).

4 1094 Key residues in the TPO immunodominant region B Western blotting analysis Approximately 20 lg of membrane proteins was mixed with protein buffer (0.2 mm Tris HCl, ph 6.8, 50% glycerol, 1% SDS and 0.1% bromophenol blue). Each sample of protein in native condition was loaded into individual wells and electrophoresed through 10% SDS-PAGE. Proteins were transferred to polyvinylidene fluoride membranes which were then saturated with 5% non-fat, powdered milk in Dulbecco s PBS 0.1% Tween 20 (D-PBS-T; blocking buffer) for 60 min at room temperature. The first antibody (rabbit anti-peptide P14) was incubated at a concentration of 10 lg ml ÿ1 overnight at 4 C. After three washings with D-PBS-T for 10 min at room temperature, the membranes were probed with peroxidaseconjugated secondary antibody (anti-rabbit IgG whole molecule) diluted 1:3000 in blocking buffer for 60 min at room temperature. After three washings, the signal was detected by chemiluminescence ECL (Amersham BioScience) on a sensitive film. Binding of membrane proteins containing wt or mutated htpo to anti-tpo antibodies assessed by ELISA Wells from microtiter plates were coated with membrane proteins in 100 mm NaHCO 3, ph 9.0, overnight at 4 C. The plates were washed with D-PBS-T and blocked with 1% nonfat, powdered milk in D-PBS-T for 60 min at 37 C. After three washings, human recombinant anti-tpo (10 lg ml ÿ1 ), rabbit anti-peptide (10 lg ml ÿ1 ) or patient s serum (1:1000) was diluted in 1% non-fat, powdered milk in D-PBS-T for 90 min at 37 C. The plates were washed and peroxidase-conjugated anti-human IgG (diluted 1:1000 in the blocking buffer) or antirabbit antibody (1:3000) was added for 60 min at 37 C. Three washings were performed, and the binding of membrane proteins to anti-tpo antibodies was detected by adding a 4-mg ml ÿ1 2-phenylenediamine solution containing 0.03% (v/v) hydrogen peroxide in 0.1 M citrate buffer, ph 5.0. The reaction was stopped with 2 M H 2 SO 4 and the resulting optical density (OD) was measured at 490 nm. Results Generation of TPO mutants To determine the amino acid residues involved in the binding of aabs to the region of the htpo, we decided to individually mutate all residues in this region, except for residues A354, G360 and A362, to alanine (A). Overlap PCR methodology was used to introduce punctual mutations in the region of the htpo (Fig. 1A). A first set of PCRs using oligonucleotides 1 and 3 and 4 and 2 (Fig. 1A and C) allowed the amplification of two fragments of 250 and 850 bp (Fig. 1B). These two PCR products (PCR1 and PCR2, respectively, of 250 and 850 bp) overlapped in part. PCR2 contained a punctual mutation introduced by oligonucleotide 4 (Fig. 1C). After the overlapping PCR, the resulting 1100-bp fragment was digested and cloned into the pcdna5/frt vector. Unfortunately, with this approach, we failed to obtain the mutant in position 363. All the other mutants were stably transfected in CHO cells and expressed at the cell surface by using the Flp-In System. Identification of amino acid residues in the region interacting with the human anti-tpo aab T13 As previously described (13), the human recombinant anti- TPO aab, named T13, interacts with four regions on the htpo molecule, including region Flow cytometry was used to determine critical amino acid residues in this region involved in the binding of the aab T13. First, we verified that all TPO mutants are expressed at the surface of the CHO cells. By using the polyclonal anti-peptide P14, interacting with the region , as shown in Fig. 2(A), we observed by flow cytometry a similar binding profile between the wt and the TPO mutants, whereas no binding was observed with the control cells (non-transfected CHO). This clearly demonstrates that neither individual mutation of htpo (in positions H353, R355, L356, R357, D358, S359 and R361) affected the folding of the molecule. Next, we evaluated the capacity of our human recombinant T13 anti-tpo aab to interact with the mutants by using an identical approach. As control, CHO cells stably transfected with fully mutated htpo in regions (TPO ) and (TPO ) were used. Whereas anti-tpo aab T13 clearly recognized wt htpo and TPO , a significant decrease in the binding of aab T13 to mutant TPO was observed. These data are in agreement with our previously published data (13). Interestingly, the mutation in position D358 strongly impaired the binding of anti-tpo aab T13, as the fully mutated TPO in region (TPO ) did. This demonstrates that D in position 358 is an important residue for the interaction between aab T13 and htpo. We also observed, by comparison with wt htpo, a decrease in the fluorescence intensity for the mutant H353, leading us to the conclusion that this residue contributes to the T13 epitope. Finally, we conclude that amino acid residues in positions 355, 356, 357, 359 and 361 are not involved in the interaction of htpo with anti-tpo aab T13. Analysis of TPO mutant recognition by human recombinant and patient s serum anti-tpo aabs To place this study in a pathological context, we decided to compare the binding of the human recombinant anti-tpo aab T13 and ICA1 (obtained previously by screening an in-cell antibody library (18)) with anti-tpo aabs from sera of patients suffering from AITD. As control, serum antibodies from patients affected by other autoimmune and non-autoimmune pathologies and healthy donors were used. For this purpose, proteins were extracted from CHO cell membranes expressing either wt or mutated TPO. We first analyzed the TPO content in each protein extract by using a western blotting approach. The rabbit polyclonal anti-peptide P14, recognizing a linear determinant outside the region (minimal epitope ) as previously described (23), was used as a probe to evaluate the expression and the folding of TPO. As shown in Fig. 3, all TPO mutants were recognized by the antipeptide P14, as we observed by flow cytometry, whereas nontransfected CHO cells (control) were not (Fig. 2A). However, we observed some differences among the mutants in the membrane preparations. More precisely, TPO , fully mutated in the region , was under-expressed compared with the wt protein or TPO , fully mutated in region (Fig. 3). Individual mutants revealed modest

5 Key residues in the TPO immunodominant region B 1095 Fig. 2. Flow cytometry analysis of binding of rabbit anti-peptide (P14) and human recombinant anti-tpo antibody (T13) to wt or mutated htpo expressed on the surface of stably transfected CHO cells. (A) Rabbit anti-tpo anti-peptide (P14) was incubated with CHO cells expressing wt or mutated htpo (mutation of amino acids , mutation of amino acids and individual mutations between amino acids 353 and 363) (see Methods). (B) Binding of recombinant human T13 anti-tpo aab on CHO cells expressing wt or mutated htpo was analyzed. Non-transfected CHO cells were used as control. variations in their expression levels, and, more importantly, mutations in positions 353 and 358 on the TPO molecule led to an expression close to that which we observed with the wt TPO, thus confirming the fact that the decrease in the binding of anti-tpo aab T13 observed by flow cytometry with TPO H353A and TPO D358A is specific (Fig. 2B). We next assessed the capability of human recombinant aabs T13 and ICA1 as well as serum anti-tpo aabs from

6 1096 Key residues in the TPO immunodominant region B Fig. 3. Expression level of mutated htpo. Western blotting with native membrane proteins was performed as indicated in Methods. Each sample (10 lg of total protein) was loaded onto SDS-PAGE gels. Native membrane proteins were detected by western blotting using rabbit anti-tpo anti-peptide P14. TPO-specific bands are shown by a black arrow, and the molecular mass in kda is given on the left. Nontransfected CHO cells were used as control. patients affected by AITD to interact with wt TPO or the series of mutants. To this end, we performed an ELISA experiment with the membrane proteins extracted from CHO cells. To avoid any problem during data interpretation due to underexpression of mutated TPO on the cell membrane, we tested different amounts of each membrane extract to obtain approximately the same binding (OD at 490 nm 2) with the wt TPO as with the mutants, when using the rabbit anti-peptide P14 as control (Fig. 4A). After this preliminary standardization, instead of using the polyclonal P14, we used the recombinant human anti-tpo aabs T13 and ICA1 under the same conditions. As shown in Fig. 4(B), full mutation of the region (TPO ) as well as individual mutations of the amino acid residues H353 and D358 significantly impaired the binding of aab T13 (a decrease of 30 to 50% compared with the wt TPO). These results are in agreement and confirm those obtained by flow cytometry (Fig. 2B). Next, we used the human recombinant anti-tpo aab ICA1, where the VL chain and the VH chain pairing is identical to that existing in vivo (17, 18), to address the following question: Are H353 and D358 also involved in epitopes recognized by other anti-tpo aabs? While the mutation of D358 indeed induced a significant decrease in the binding of ICA1 to TPO (~50% as compared with wt TPO), mutation in position H353 only lightly affected its binding (a 10% decrease as compared with wt TPO). Interestingly, mutation of residues S359 and R361 markedly affected the binding of the ICA1, suggesting that these two amino acid residues are also involved in the ICA1 epitope. This observation points out, once again (14, 23), that even if two aabs recognize the same region, their anchor points may differ slightly. Finally, we analyzed in a similar experiment the anti-tpo aabs contained in the sera of patients suffering from Graves disease and Hashimoto s thyroiditis. As expected, the mutant TPO dramatically decreased the binding of sera from both Graves and Hashimoto s patients, whereas sera from healthy donors or patients suffering from other autoimmune diseases did not react (Fig. 4C and D). These data also confirmed the importance of residues H353, D358 and R361 in the binding of anti-tpo aabs from AITD patients (Fig. 4C and D). It should be emphasized that anti-tpo aabs from patients with Graves or Hashimoto s disease demonstrate a similar pattern of reactivity, leading us to the conclusion that the anti-tpo epitopic repertoire in both diseases is very close, if not identical. Taken together, our data demonstrate for the first time that H353, D358, S359 and R361 are amino acid residues contributing to the binding of anti-tpo aabs in the IDR. Discussion Human aabs recognize on the surface of htpo a restricted and discontinuous domain named IDR (11, 24 26). The identification of amino acids taking part in the IDR is extremely difficult owing to the architecture of the htpo (19), and several groups have tried to determine residues in contact with human anti-tpo aabs using the directed mutagenesis approach (10, 27, 28). To date, the major region ( ), belonging to the IDR/A, and five regions ( , , , and ) structuring the IDR/B have been identified by different groups (13, 29). Importantly, we have demonstrated that three regions (amino acid residues , and ) are crucial for the binding of anti-tpo aabs present in sera from patients suffering from AITD (13) and first focalized our studies on the identification of key residues in the region (14). Since the region is the main region recognized by anti-tpo aabs in patients sera (as previously demonstrated (13)) but was never studied in detail, we decided to evaluate the contribution of the amino acid residues in this region to the interaction between the TPO and anti-tpo aabs. Therefore, we demonstrated here, by using wt and punctually mutated htpo extracted from stably transfected CHO cells, that amino acid residues H353, D358, S359 and R361 are key residues for the IDR epitopes recognized by human anti-tpo aabs. Using recombinant htpo mutants with single amino acid replacements, a number of key residues have been attributed to the IDR (14 16, 29, 30). Residues R225, Y772 and K713 to D717 were shown to be involved in the IDR/B-specific human anti-tpo aab epitopes, and one amino acid residue (K627) was assigned to the IDR/A-specific human anti-tpo aab epitopes. Surprisingly, no amino acid residue in the region could be assigned to the IDR, even if the threedimensional model of htpo revealed that the region corresponds to an exposed loop. A plausible explanation for this could be the strategy used. The groups of A. Gardas and J.-P. Banga, who have contributed to the characterization of the IDR, have based the selection of amino acids for mutagenesis essentially upon their charge and accessibility in the structural model of htpo (29). Twenty-five residues located on the surface of the predicted model of htpo were chosen, but only two contributing amino acid residues (R225 and K627) were clearly assigned to the IDR. To identify amino acids constituting the IDR, with a maximum success, we have proceeded in three steps. First, we selected minotopes and localized sequences involved in the binding of anti-tpo aabs on htpo (13), and then we mutated amino acids in the

7 A % of Binding B % of Binding Key residues in the TPO immunodominant region B 1097 T13 ICA1 * * * * * * * C OD 490 D % of Binding TPO wt TPO wt TPO wt TPO * * TPO H353A * * H353A R355A R355A TPO Hashimoto L356A L356A R357A R357A H353A Graves D358A D358A S359A S359A R361A R361A R355A 0 Graves Hashimoto AID Healthy control * * * * Fig. 4. Effect of single amino acid mutations in the region of htpo on the binding of recombinant human anti-tpo and patient s serum aabs. (A) Standardization of membrane protein extracts of wt or mutated htpo CHO cells using rabbit anti-peptide P14. A concentration of extracted membrane protein from the wt and the htpo mutants was used to obtain ~100% binding in ELISA, corresponding to an OD 2. (B) Human recombinant T13 and ICA1 anti-tpo aabs binding to extracted membrane proteins (wt or htpo mutants) expressed as a percentage in standardized ELISA (100% binding corresponds to an OD of 2). (C) Binding on wt htpo of individual sera from patients affected by Graves disease, Hashimoto s thyroiditis, other autoimmune diseases (AID) or healthy controls. The percent binding in (D) corresponds to the average of the five OD values obtained with the Graves disease or Hashimoto s thyroiditis patients sera. (D) Pooled sera from Graves or Hashimoto s patients were tested for their ability to bind wt or mutated TPO molecules (the fully mutated region and individually mutated amino acid residue in this region). Statistical significance was determined using the two-tailed Student s t test. *P < 0.05 (data significantly different). The data shown (mean 6 SD) are representative results from three independent experiments. L356A TPO wt TPO R357A H353A R355A L356A D358A R357A D358A S359A S359A R361A R361A identified regions. In the region , after mutation of all amino acid residues (except A354, G360, A362 and Y363) to alanine, this strategy led us to identify four additional amino acids (H353, D358, S359, and R361) constituting the dominant autoreactive region of the htpo. Interestingly, single amino acid substitutions in the region of htpo strongly reduced recombinant human T13 anti-tpo aab binding by flow cytometry. This was not the case for amino acid substitutions in the region (14), where it was extremely difficult to demonstrate the influence of an individual amino acid on the interaction. This strongly suggests that in the region, previously characterized as crucial for the binding of anti-tpo aabs in the sera of patients suffering from AITD (13), some amino acids are key

8 1098 Key residues in the TPO immunodominant region B residues for the interaction of anti-tpo aabs with the htpo. Furthermore, examination of htpo expression in stable CHO cell lines by western blotting in SDS-PAGE gels showed that wt htpo and all point mutation constructs in the region of TPO were expressed at the cell membrane. After standardization of membrane protein extracts of htpo mutants, we confirmed the participation of H353 and D358 in the binding of recombinant T13 anti-tpo aab observed by flow cytometry. Interestingly, D358 was found to be involved in the interaction of both recombinant human T13 anti-tpo aabs and ICA1 and patients sera with htpo. This demonstrates that D358 is a crucial residue involved in the IDR/B. However, H353, S359 and R361 are variably involved in the epitope. R361 is important for the interaction of ICA1 and patient s serum aabs with the htpo molecule, whereas H353 is involved in T13 and patient s serum aabs binding, and S359 was identified as a critical amino acid only for ICA1 htpo interaction. Probably, the contribution of amino acids and their participation in the epitope differ as a function of the anti-tpo aab involved in the interaction. The variations in the epitopic recognition observed between recombinant anti-tpo and patient s serum aabs could be explained by the polyclonality that exists in patients sera. Other possible explanations for variations observed between the two recombinant human anti-tpo aabs (T13 and ICA1) could be (i) the difference in VL chain genes used for T13 and ICA1 aabs (T13 and ICA1 VL chains are encoded, respectively, by the VL1-40 and VL1-51 germ line gene) or (ii) a VH/VL pairing used for recombinant human anti-tpo aabs that does not reflect the in vivo situation. Human recombinant ICA1 anti-tpo aab was previously selected from an in-cell combinatorial library where the combination of heavy and light chains is identical to the pairing present in vivo in the patient whose thyroid-infiltrating B cells were used to construct the library (17, 18). The human recombinant ICA1 anti-tpo aab inhibits the binding of aabs in patients sera to TPO and consequently is a representative anti-tpo aab that mimics these aabs. The T13 anti-tpo aab was selected from a random combinatorial library, and we do not know if the VL/ VH pairing of this aab is identical to that existing in vivo. However, the T13 anti-tpo aab is able to inhibit the binding of aabs in patients sera on htpo as does the ICA1 anti-tpo aab. Taken together, these findings support the fact that T13 and ICA1 anti-tpo aabs share the same or neighboring epitopes on the htpo molecule as those recognized by serum aabs and that identical amino acid residues could participate in the interaction of T13, ICA1 and patient s serum anti-tpo aabs with htpo. Finally, we must keep in mind that the individual contribution of amino acids in an epitope recognized by an anti-tpo antibody can be different and have a modest or strong participation in the binding of an aab for a given region, even if all these antibodies are characterized as being able to bind this region. In summary, this study demonstrates that the three-step strategy (selection of minotopes, identification of sequences involved in the IDR and directed mutagenesis of single amino acids) is a powerful strategy to position discontinuous epitopes on a highly complex structure like htpo and to identify crucial amino acids. Our results definitively assign key residues from the region of the IDR recognized by anti-tpo aabs in AITD and further emphasize the importance of H353, D358 and R361 and in some cases the contribution of S359 for TPO recognition by human anti-tpo aabs during the onset of diseases. Acknowledgements We thank S. L. Salhi for carefully reading the manuscript. We are indebted to J.-P. Banga and A. Gardas for providing us with the rabbit anti-peptide to TPO. We also thank L. Baldet and A. M. Puech for providing the patient s sera. D.B. is a recipient of a Marie-Curie Outgoing fellowship ( ). S.A.R. is financed by Ligue Nationale Contre le Cancer. Abbreviations aabs auto-antibodies AITD autoimmune thyroid diseases D-PBS Dulbecco s PBS D-PBS-T Dulbecco s PBS 0.1% Tween 20 htpo human thyroid peroxidase IDR immunodominant region OD optical density TPO thyroid peroxidase wt wild type References 1 Zimmer, K. P., Scheumann, G. F., Bramswig, J., Bocker, W., Harms, E. and Schmid, K. W Ultrastructural localization of IgG and TPO in autoimmune thyrocytes referring to the transcytosis of IgG and the antigen presentation of TPO. Histochem. Cell Biol. 107: Taurog, A., Dorris, M. L. and Doerge, D. R Mechanism of simultaneous iodination and coupling catalyzed by thyroid peroxidase. Arch. Biochem. Biophys. 330:24. 3 McLachlan, S. M. and Rapoport, B Autoimmune response to the thyroid in humans: thyroid peroxidase the common autoantigenic denominator. Int. Rev. Immunol. 19: Chiovato, L., Bassi, P., Santini, F. et al Antibodies producing complement-mediated thyroid cytotoxicity in patients with atrophic or goitrous autoimmune thyroiditis. J. Clin. Endocrinol. Metab. 77: Rodien, P., Madec, A. M., Ruf, J. et al Antibody-dependent cell-mediated cytotoxicity in autoimmune thyroid disease: relationship to antithyroperoxidase antibodies. J. Clin. Endocrinol. Metab. 81: Guo, J., Wang, Y., Rapoport, B. and McLachlan, S. M Evidence for antigen presentation to sensitized T cells by thyroid peroxidase (TPO)-specific B cells in mice injected with fibroblasts co-expressing TPO and MHC class II. Clin. Exp. Immunol. 119:38. 7 Guo, J., Quaratino, S., Jaume, J. C. et al Autoantibodymediated capture and presentation of autoantigen to Tcells via the Fc epsilon receptor by a recombinant human autoantibody Fab converted to IgE. J. Immunol. Methods 195:81. 8 Quaratino, S., Badami, E., Pang, Y. Y. et al Degenerate selfreactive human T-cell receptor causes spontaneous autoimmune disease in mice. Nat. Med. 10: Gardas, A., Blottner, A. and Domek, H Comparison of autoantibodies level in thyroid autoimmune diseases estimated by different methods. J. Endocrinol. Invest. 11: Ruf, J., Toubert, M. E., Czarnocka, B., Durand-Gorde, J. M., Ferrand, M. and Carayon, P Relationship between immunological structure and biochemical properties of human thyroid peroxidase. Endocrinology 125: Portolano, S., Chazenbalk, G. D., Seto, P., Hutchison, J. S., Rapoport, B. and McLachlan, S. M Recognition by recombinant autoimmune thyroid disease-derived Fab fragments

9 of a dominant conformational epitope on human thyroid peroxidase. J. Clin. Invest. 90: McLachlan, S. M. and Rapoport, B Genetic and epitopic analysis of thyroid peroxidase (TPO) autoantibodies: markers of the human thyroid autoimmune response. Clin. Exp. Immunol. 101: Bresson, D., Cerutti, M., Devauchelle, G. et al Localization of the discontinuous immunodominant region recognized by human anti-thyroperoxidase autoantibodies in autoimmune thyroid diseases. J. Biol. Chem. 278: Bresson, D., Pugniere, M., Roquet, F. et al Directed mutagenesis in region of human thyroperoxidase assigns 713KFPED717 residues as being involved in the B domain of the discontinuous immunodominant region recognized by human autoantibodies. J. Biol. Chem. 279: Guo, J., Yan, X. M., McLachlan, S. M. and Rapoport, B Search for the autoantibody immunodominant region on thyroid peroxidase: epitopic footprinting with a human monoclonal autoantibody locates a facet on the native antigen containing a highly conformational epitope. J. Immunol. 166: Estienne, V., Duthoit, C., Blanchin, S. et al Analysis of a conformational B cell epitope of human thyroid peroxidase: identification of a tyrosine residue at a strategic location for immunodominance. Int. Immunol. 14: Chapal, N., Chardes, T., Bresson, D. et al Thyroid peroxidase autoantibodies obtained from random single chain FV libraries contain the same heavy/light chain combinations as occur in vivo. Endocrinology 142: Chapal, N., Peraldi-Roux, S., Bresson, D. et al Human antithyroid peroxidase single-chain fragment variable of Ig isolated from a combinatorial library assembled in-cell: insights into the in vivo situation. J. Immunol. 164: Hobby, P., Gardas, A., Radomski, R., McGregor, A. M., Banga, J. P. and Sutton, B. J Identification of an immunodominant region recognized by human autoantibodies in a three-dimensional model of thyroid peroxidase. Endocrinology 141: Gardas, A., Watson, P. F., Hobby, P. et al Human thyroid peroxidase: mapping of autoantibodies, conformational epitopes to the enzyme surface. Redox Rep. 5:237. Key residues in the TPO immunodominant region B Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, J. K. and Pease, L. R Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77: Sanger, F., Nicklen, S. and Coulson, A. R DNA sequencing with chain-terminating inhibitors. Proc. Natl Acad. Sci. USA 74: Bresson, D., Rebuffat, S. A., Nguyen, B., Banga, J. P., Gardas, A. and Peraldi-Roux, S New insights into the conformational dominant epitopes on thyroid peroxidase recognized by human autoantibodies. Endocrinology 146: Finke, R., Seto, P., Ruf, J., Carayon, P. and Rapoport, B Determination at the molecular level of a B-cell epitope on thyroid peroxidase likely to be associated with autoimmune thyroid disease. J. Clin. Endocrinol. Metab. 73: Frorath, B., Abney, C. C., Scanarini, M., Berthold, H., Hunt, N. and Northemann, W Mapping of a linear autoantigenic epitope within the human thyroid peroxidase using recombinant DNA techniques. J. Biochem. 111: Banga, J. P., Barnett, P. S., Ewins, D. L., Page, M. J. and McGregor, A. M Mapping of autoantigenic epitopes on recombinant thyroid peroxidase fragments using the polymerase chain reaction. Autoimmunity 6: Kaufman, K. D., Rapoport, B., Seto, P., Chazenbalk, G. D. and Magnusson, R. P Generation of recombinant, enzymatically active human thyroid peroxidase and its recognition by antibodies in the sera of patients with Hashimoto s thyroiditis. J. Clin. Invest. 84: Chazenbalk, G. D., Portolano, S., Russo, D., Hutchison, J. S., Rapoport, B. and McLachlan, S Human organ-specific autoimmune disease. Molecular cloning and expression of an autoantibody gene repertoire for a major autoantigen reveals an antigenic immunodominant region and restricted immunoglobulin gene usage in the target organ. J. Clin. Invest. 92: Gora, M., Gardas, A., Watson, P. F. et al Key residues contributing to dominant conformational autoantigenic epitopes on thyroid peroxidase identified by mutagenesis. Biochem. Biophys. Res. Commun. 320: Gora, M., Gardas, A., Wiktorowicz, W. et al Evaluation of conformational epitopes on thyroid peroxidase by antipeptide antibody binding and mutagenesis. Clin. Exp. Immunol. 136:137.