Dominant Autoimmune Epitopes Recognized by Antibodies Map to the N-Terminal Adhesive Region of

This information is current as Maiko Sekiguchi, Yuko Futei, Yoshiko Fujii, Toshiro of October 4, 2021. Iwasaki, Takeji Nishikawa and Masayuki Amagai J Immunol 2001; 167:5439-5448; ; doi: 10.4049/jimmunol.167.9.5439 http://www.jimmunol.org/content/167/9/5439 Downloaded from

References This article cites 31 articles, 6 of which you can access for free at: http://www.jimmunol.org/content/167/9/5439.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on October 4, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2001 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Dominant Autoimmune Epitopes Recognized by Pemphigus Antibodies Map to the N-Terminal Adhesive Region of Desmogleins1

Maiko Sekiguchi,*† Yuko Futei,* Yoshiko Fujii,* Toshiro Iwasaki,† Takeji Nishikawa,* and Masayuki Amagai2*

Desmoglein (Dsg) is a -type adhesion molecule found in . Dsg1 and Dsg3 are the target Ags in the autoimmune blistering pemphigus foliaceus (PF) and (PV), respectively. To map conformational epitopes of Dsg1 and Dsg3 in PF and PV, we generated Dsg1- and Dsg3-domain-swapped molecules and point-mutated Dsg3 molecules with Dsg1-specific residues by baculovirus expression. The swapped domains were portions of the N-terminal extracellular domains of Dsg1 (1–496) and Dsg3 (1–566), which have similar structures but distinct epitopes. The binding of to the mutant Downloaded from molecules was assessed by competition ELISAs. Domain-swapped molecules containing the N-terminal 161 residues of Dsg1 and Dsg3 yielded >50% competition in 30/43 (69.8%) PF sera and 31/40 (77.5%) PV sera, respectively. Furthermore, removal of Abs against the 161 N-terminal residues of Dsg1 by immunoadsorption eliminated the ability of PF sera to induce cutaneous blisters in neonatal mice. Within these N-terminal regions, most of the epitopes were mapped to residues 26–87 of Dsg1 and 25–88 of Dsg3. Furthermore, a point-mutated Dsg3 molecule containing Dsg1-specific amino acid substitutions (His25, Cys28, Ala29) reacted with anti-Dsg1 IgG, thus defining one of the epitopes of Dsg1. Using the predicted three-dimensional structure of classic as http://www.jimmunol.org/ a model, these findings suggest that the dominant autoimmune epitopes in both PF and PV are found in the N-terminal adhesive surfaces of Dsgs. The Journal of Immunology, 2001, 167: 5439–5448.

esmoglein (Dsg)3 is a desmosomal transmembrane gly- have circulating anti-Dsg3 IgG alone; those with mucocutaneous coprotein that belongs to the cadherin superfamily of PV have both anti-Dsg3 and anti-Dsg1 IgG; and those with PF D - molecules (1). Dsg has three isotypes: have anti-Dsg1 IgG alone (11, 12). Dsg1, Dsg2, and Dsg3. Expression of Dsg1 and Dsg3 is predom- Characterizing the Dsg binding sites of the pathogenic pemphi- inantly restricted to stratified squamous epithelia, whereas Dsg2 is gus autoantibodies is an essential step in understanding the patho- expressed in all -bearing cells, including simple epi- physiology of blister formation in pemphigus, as well as the basic by guest on October 4, 2021 thelial and myocardial cells (2, 3). Dsg1 and Dsg3 are the targets molecular mechanism of Dsg-mediated cell-cell adhesion. This of pemphigus, a -threatening autoimmune blistering of characterization is hindered by the fact that binding of autoanti- the skin and mucous membranes (4, 5). Compelling evidence has bodies to Dsgs is dependent not only on amino acid sequence, but accumulated that IgG autoantibodies against Dsgs play a patho- also on molecular conformation (9, 13Ð15). This dependence on genic role in blister formation in pemphigus (6Ð10). Pemphigus molecular conformation is shown by the observation that recom- has two major subtypes: pemphigus vulgaris (PV) and pemphigus binant Dsg1 and Dsg3, when expressed in baculovirus as secreted foliaceus (PF). PV can be further divided into two clinical forms; , immunoadsorbs heterogeneous autoantibodies from PF mucosal dominant PV is characterized by predominant mucosal and PV patients’ sera, and that this immunoadsorptive activity is involvement with minimal skin lesions, and mucocutaneous PV is lost upon denaturation by Ca2ϩ chelation, acid or alkaline treat- characterized by extensive lesions both in the skin and mucous ment, or boiling. Furthermore, the importance of conformational membranes. These three forms of pemphigus have different anti- epitopes was also demonstrated in the production of pathogenic Dsg profiles. Patients with mucosal dominant PV Abs by mouse immunization (16Ð18). Therefore, a conventional approach using variously truncated Dsg molecules is inappropriate *Department of Dermatology, Keio University School of Medicine, Tokyo, Japan; for definition of the conformational epitopes of Dsgs in and †Department of Internal Medicine, Faculty of Veterinary Medicine, Tokyo Uni- versity of Agriculture and Technology, Tokyo, Japan pemphigus. Dsg1 and Dsg3 have similar structures, but distinct epitopes. Received for publication November 27, 2000. Accepted for publication August 31, 2001. Recently, we have shown that Dsg1- and Dsg3-domain-swapped The costs of publication of this article were defrayed in part by the payment of page molecules are useful for characterization of the conformational charges. This article must therefore be hereby marked advertisement in accordance epitopes of Dsg3 in PV (19). In this work, we extend the study, with 18 U.S.C. Section 1734 solely to indicate this fact. using 10 domain-swapped molecules and six point-mutated mol- 1 This work was supported by a Health Sciences Research Grant for Research on Specific Diseases from the Ministry of Health, Labour and Welfare of Japan (to M.A.) ecules to map conformational epitopes of Dsg1 and Dsg3. Com- and a grant-in-aid of Scientific Research from the Ministry of Education, Culture, petition ELISAs with these domain-swapped molecules reveal that Sports, Science and Technology, Japan (to M.A. and T.N.). dominant epitopes map to amino acid residues 26Ð87 of Dsg1 and 2 Address correspondence and reprint requests to Dr. Masayuki Amagai, Department 25Ð88 of Dsg3. Furthermore, within those regions we have iden- of Dermatology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku- ku, Tokyo 160-8582, Japan. E-mail address: [email protected] tified important conformational epitopes at which amino acid sub- 3 Abbreviations used in this paper: Dsg, ; PF, pemphigus foliaceus; PV, stitutions alter the binding specificity of pemphigus autoantibod- pemphigus vulgaris. ies. These findings are valuable for understanding the molecular

Copyright © 2001 by The American Association of Immunologists 0022-1767/01/$02.00 5440 CONFORMATIONAL EPITOPE MAPPING OF PEMPHIGUS ANTIGENS

Table I. Primers used for domain-swapped moleculesa

Primers

Primers for Dsg1 Primers for Dsg3

Constructs 5Ј primers 3Ј primers 5Ј primers 3Ј primers

Dsg11Ð24/Dsg325Ð566 Primer 1 Primer 2 Primer 3 Primer 4 Dsg11Ð64/Dsg365Ð566 Primer 1 Primer 5 Primer 6 Primer 4 Dsg11Ð87/Dsg387Ð566 Primer 1 Primer 7 Primer 8 Primer 4 Dsg31Ð26/Dsg126Ð496 Primer 9 Primer 10 Primer 11 Primer 12 Dsg31Ð63/Dsg163Ð496 Primer 13 Primer 10 Primer 11 Primer 14 Dsg31Ð88/Dsg189Ð496 Primer 15 Primer 10 Primer 11 Primer 16

a Primer 1, 5Ј-gaagatctcctataaatATGGACTGGAGTTTCTTCAGAG-3Ј; primer 2, 5Ј-cggactagtAATTTTGGCGATT GGGTT-3Ј; primer 3, 5Ј-gccactagtGATTACCAAGCAACCCAG-3Ј; primer 4, 5Ј-cctgctcgagCCTCCCTGAGTGCGGCCT-3Ј; primer 5, 5Ј-cggactagtTATATTAATTTCACCAGT-3; primer 6, 5Ј-gccactagtATAGTCGACCGGGAGGAA-3Ј; primer 7, 5Ј- cgggacgtcTTGGCCCATTGAGTTCAG-3Ј; primer 8, 5Ј-gccgacgtcGAGAAACCACTTATACTA-3Ј; primer 9, 5Ј-gccac tagtGATTGTGCTGCAAACCAG-3Ј; primer 10, 5Ј-cttgtcgacATGTACATTGTCTGATAACAAATC-3Ј; primer 11, 5Ј- gaagatctCCTATAAATATGGGGCTCTTCCCCAG-3Ј; primer 12, 5Ј-cggactagtAATCTTGGCAATTGGGTT-3Ј; primer 13, 5Ј-gccactagtATAGTTGATCGAGAGGTC-3Ј; primer 14, 5Ј-cggactagtTATGTTAATATCTCCAGT-3Ј; primer 15, 5Ј- gccgacgtcGAGAGGCCTCTAGAGCTC-3Ј; primer 16, 5Ј-cgggacgtcTAGTCCTTGGGCATTTAG-3Ј. Downloaded from

mechanism of Dsg-mediated cell-cell adhesion, as well as for de- mid template coding the extracellular domain of Dsg3 (pEVmod-Dsg3- velopment of epitope-specific therapeutic strategies for pemphigus. His). This procedure eliminated a unique ScaI restriction site (shown in boldface) in the ampr gene (23). The megaprimer was then used to amplify the entire plasmid with the same plasmid template in a second round of Materials and Methods PCR. After DpnI digestion, the PCR product was used for transformation.

Human sera All mutants were sequenced to verify the presence of mutation and to http://www.jimmunol.org/ ensure that no other mutations were introduced. Sera were obtained from 43 patients with PF and 40 patients with PV, whose diagnoses were confirmed by clinical, histologic, and immunopatho- Production of proteins in baculovirus logic findings. All PF sera were positive for anti-Dsg1 IgG and were neg- ative for anti-Dsg3 IgG as determined by ELISA analysis using recombi- The plasmids were cotransfected with BaculoGold baculovirus DNA (BD nant Dsg3 and Dsg1 (20, 21). Two PF (PF#73 and PF#2284) sera obtained PharMingen, San Diego, CA) into cultured insect Sf9 cells, and recombi- by plasmapheresis were used for passive transfer study with neonatal mice. nant were collected from culture supernatant as previously de- All PV sera were positive for anti-Dsg3 IgG autoantibodies and 14 of the scribed (20, 22). High-Five cells cultured in serum-free EX cell 405 me- PV samples were also positive for anti-Dsg1 IgG. In the PF samples, re- dium (JRH Biosciences, Lenexa, KS) were infected with high-titer activity against Dsg1 was not reduced by immunoadsorption by Dsg3-His. stock and incubated for 3 days. The culture supernatant contained ϳ5Ð10 ␮ Similarly, in the PV samples, reactivity against Dsg3 was not significantly g/ml recombinant . These proteins were purified on TALON af- by guest on October 4, 2021 altered by immunoadsorption by Dsg1-His. finity resin (Clontech, Palo Alto, CA) according to the manufacturer’s recommendation. Plasmid constructs Immunoblot analysis We previously constructed forms of recombinant Dsg1 and Dsg3 that are secreted upon expression in baculovirus (13, 14, 20, 22). These recombi- The production of recombinant protein was confirmed by immunoblot nant proteins contain the entire extracellular domain of Dsg1 or Dsg3 fused analysis. Mouse anti-human Dsg3 mAb 5H10 (25) and anti-His-tag mAb with the constant region of human IgG1 and/or a His-tag. In this study, 10 domain-swapped molecules and six Dsg3 point mutants were constructed (see Fig. 1). a Dsg1 and Dsg3 domain-swapped molecules were constructed by sub- Table II. Primers used for point-mutated Dsg3 molecules stitution of regions of Dsg1 or Dsg3 with the corresponding Dsg3 or Dsg1 regions, respectively. The cDNAs for various regions of Dsg1 and Dsg3 Amino Acid Amino Acid were PCR-amplified with appropriate primers (Table I) using pEVmod- Constructs Primers No. Change PVIg (GenBank accession number M76482) (13) and pKS-Dsg1 (Gen- Bank accession number X56654) as templates, respectively. For gross Dsg3-M1 17 25 T to H epitope mapping, four domain-swapped constructs were produced that con- 28 YtoC tained the constant region of human IgG1 and a His-tag at the C terminus, 29 QtoA as previously described (19) (Fig. 1A). To narrow the pemphigus autoan- Dsg3-M2 18 31 T to N tibody epitopes, another six domain-swapped constructs were produced, 33 KtoQ each containing an E-tag (Pharmacia Biotech, Uppsala, Sweden) and a 34 ItoV His-tag at their C termini (Fig. 1B). The PCR products for the Dsg1 frag- Dsg3-M3 19 53 V to I ments comprising residues 1Ð87, 1Ð64 and 1Ð24, 89Ð496, and 63Ð496 54 DtoN and 26Ð496 were digested with BglII/AatII, BglII/SpeI, AatII/SalI, and 55 KtoQ SpeI/SalI, respectively. The PCR products for Dsg3 fragments comprising 56 NtoK residues 1Ð88, 1Ð63 and 1Ð26, 87Ð566, and 65Ð566 and 25Ð566 were Dsg3-M4 20 70 E to V digested with BglII/AatII, BglII/SpeI, AatII/XhoI, and SpeI/XhoI, respec- 73 StoF tively. The resulting products were ligated to BglII/XhoI-digested vector pE- Dsg3-M5 21 75 L to I Vmod-His (20). These constructs were designated Dsg11Ð87/Dsg387Ð566, 77 TtoY Dsg11Ð64/Dsg365Ð566, Dsg11Ð24/Dsg325Ð566, Dsg31Ð88/Dsg189Ð496, Dsg31Ð63/ Dsg3-M6 22 83 A to S Dsg163Ð496, and Dsg31Ð26/Dsg126Ð496. 84 QtoM To introduce point mutations in Dsg3, we performed site-directed mu- a Ј Ј tagenesis according to the unique restriction site elimination method (23) Primer 17, 5 -gaaacccaattgccaagattCAttcagattGcGCagcaacccagaaaatcacctac-3 ; primer 18, 5Ј-acttcagattaccaagcaaAccagCaaGtcacctaccgaatctctggag-3Ј; primer 19, 5Ј- and the PCR-based megaprimer method (24). To perform this procedure, ccgccttttggaatctttgttAttAacCaaaaAactggagatattaacataacag-3Ј; primer 20, 5Ј-gctatag we prepared six mutagenic primers (Table II). In brief, the megaprimer tcgaccgggaggTaactccaTTcttcctgatcacatgtcgg-3Ј; primer 21, 5Ј-gaggaaactccaagcttcA containing the mutation was generated by a first round of PCR with a tCatcTATtgtcgggctctaaatgcc-3Ј; primer 22, 5Ј-catgtcgggctctaaatAGcATGggactagatg mutagenic primer (5Ј-GACTTGGTTGAATACTCACCAG-3Ј) and a plas- tagag-3Ј. The Journal of Immunology 5441

(R&D Systems, Minneapolis, MN) were used as primary Abs, and a 1/1000 dilution of alkaline phosphatase-conjugated goat anti-mouse IgG Abs (Zymed Laboratories, San Francisco, CA) was used as the secondary Ab. Competition ELISA analysis Sera from pemphigus patients were preincubated at 4¡C overnight either with the designated amount of purified recombinant protein in phosphate-

buffered saline containing 1% BSA and 0.5 mM CaCl2 (PBS-Ca), or with culture supernatant containing recombinant protein. When culture super- natant was used, sera were diluted 200-fold. The sera were then subjected to ELISA against the entire extracellular domain of Dsg1 or Dsg3 (21).

When necessary, sera were diluted to keep A450 below 1.2. Competition by immunoadsorption during preincubation was calculated using the formula: ϭ Ϫ Ϫ Ϫ ϫ competition (%) (1 (A450exp A450pos) / (A450neg A450pos)) 100. A450pos values for Dsg1 and Dsg3 are the measurements obtained for sera preincubated with recombinant Dsg1-His and recombinant Dsg3-His,

respectively; A450neg is the measurement obtained for sera preincubated in PBS-Ca or in culture supernatant of uninfected High-Five cells; A450exp is the measurement obtained for sera preincubated with the recombinant pro- tein of interest. Passive transfer study with neonatal mice Downloaded from To evaluate the pathogenic activity of PF sera that were immunoadsorbed with domain-swapped molecules, we performed a passive transfer study with neonatal mice, as previously described (13, 22). In brief, 5 ml of two PF sera (PF#73 and PF#2284) were immunoadsorbed with either Dsg1- His, Dsg3-His, Dsg11Ð161/Dsg3163Ð566, or Dsg31Ð161/Dsg1164Ð496 purified on TALON affinity resin (Clontech) from ϳ200 ml of culture supernatant. IgG was prepared from the pass-through fraction by precipitation with 40%

ammonium sulfate, dialyzed against PBS, and concentrated down to ϳ600 http://www.jimmunol.org/ ␮l with a microconcentrator, Centriprep 30 (Amicon, Beverly, MA). Then, 150 ␮l of IgG solution were injected s.c. into neonatal ICR mice (12Ð24 h of age; body weight, 1.5Ð1.8 g). The neonatal mice were examined and biopsied 18Ð24 h after the injection to evaluate blister formation. At least three neonatal mice were tested for each immunoadsorbed IgG fraction. Results Competition ELISA with domain-swapped and point-mutated molecules by guest on October 4, 2021 A series of Dsg1- and Dsg3-domain-swapped molecules and point- mutated Dsg3 molecules were expressed in baculovirus as secreted proteins (Figs. 1 and 2). Dsg11Ð24/Dsg325Ð566, Dsg11Ð64/Dsg365Ð566, and Dsg11Ð87/Dsg387Ð566 were detected as doublets of 85 and 91 kDa; Dsg31Ð26/Dsg126Ð496, Dsg31Ð63/Dsg163Ð496, and Dsg31Ð88/ Dsg189Ð496 were detected as doublets of 81 and 85 kDa; and Dsg3- M1-His to Dsg3-M6-His were detected as doublets of 77 and 82 kDa. The lower band of each doublet is most likely the result of proteolytic processing of the prosequence present in the upper band, as previously described (13, 14). To determine appropriate conditions for competition ELISA, we first examined the effect of various concentrations of purified re- combinant proteins on the observed level of competition (repre- sentative data are shown in Fig. 3). Competition increased in a dose-dependent fashion, reaching a plateau at 2.5Ð10 ␮g/ml com- petitor protein. Competition levels obtained using culture super- natant containing recombinant proteins before purification were nearly equal to or higher than the maximum competition level obtained with purified proteins. Therefore, we used culture super- natant as the source of competitor protein for subsequent compe- tition ELISA. The reactivity of the PF sera against Dsg1 was completely abolished by competition for the extracellular domain of Dsg1 FIGURE 1. Structures of the recombinant Dsg1 and Dsg3 molecules used in this study. Each of these constructs has a His-tag (His) at its C terminus. A, Domain-swapped molecules used in gross epitope mapping. The extracellular domains of Dsg1 and Dsg3 were divided into three parts, epitope mapping in the N-terminal regions of Dsg1 and Dsg3. The N- and four Dsg1- and Dsg3-domain-swapped molecules were constructed. terminal domains (1Ð161) were divided into various parts. C, Point-mu- These constructs have the constant region of human IgG1 (C) at their tated Dsg3 with Dsg1-specific residues. Six clusters of Dsg3 exhibiting C-termini. Amino acid residues are numbered from the N terminus of the little conservation with Dsg1 were chosen, and Dsg1-specific residues were mature form of Dsg1 and Dsg3. B, Domain-swapped molecules used in fine introduced into Dsg3 by site-directed mutagenesis. 5442 CONFORMATIONAL EPITOPE MAPPING OF PEMPHIGUS ANTIGENS

were used for gross mapping of conformational epitopes on Dsg1 and Dsg3. For epitope mapping of anti-Dsg1 IgG in PF sera, each PF se- rum sample was preincubated with baculovirus culture supernatant containing one of the four recombinant proteins described above, Dsg1-IgHis (positive control), or no recombinant protein (negative control). The depleted samples were then subjected to ELISA anal- ysis against Dsg1-His and the level of competition for each do- main-swapped molecule was calculated. The results for the 43 PF sera tested are shown in the same order in each plot (Fig. 4). The order is determined by the level of competition found for Dsg11Ð161/ Dsg3163Ð566. Residues 1Ð401 of Dsg1 (Dsg11Ð401/Dsg3405Ð566) FIGURE 2. Recombinant proteins produced by baculovirus expression yielded an average competition level of 98.2% (Fig. 4A), whereas in insect cells. The products were visualized by immunoblotting with anti- residues 404Ð496 of Dsg1 (Dsg31Ð403/Dsg1404Ð496) showed an av- His-tag mAb. A, Six domain-swapped molecules were produced. Lane 1, erage level of only 5.8% (Fig. 4C), indicating that the critical epitopes 1Ð24 25Ð566 1Ð64 65Ð566 1Ð87 Dsg1 /Dsg3 ; lane 2, Dsg1 /Dsg3 ; lane 3, Dsg1 / for PF sera are in residues 1Ð401 of Dsg1. Residues 1Ð161 of Dsg1 87Ð566 1Ð26 26Ð496 1Ð63 63Ð496 Dsg3 ; lane 4, Dsg3 /Dsg1 ; lane 5, Dsg3 /Dsg1 ; (Dsg11Ð161/Dsg3163Ð566) yielded an average competition level of lane 6, Dsg31Ð88/Dsg189Ð496. B, Six point-mutated Dsg3 proteins. Lane 1, 62.5%; this level exceeded 50% in 30 (69.8%) of the 43 samples (Fig. Dsg3-M1; lane 2, Dsg3-M2; lane 3, Dsg3-M3; lane 4, Dsg3-M4; lane 5, 1Ð161 164Ð496 Dsg3-M5; lane 6, Dsg3-M6. Molecular mass standards are 205, 120, and 4B). In contrast, residues 164Ð496 (Dsg3 /Dsg1 ) yielded Downloaded from 84 kDa from top to bottom. an average competition level of 16.5%, and this level exceeded 50% in only two (4.7%) samples (Fig. 4D). This finding indicates that most (Dsg1-IgHis or Dsg1-His), but it was not significantly affected by of the epitopes recognized by anti-Dsg1 IgG in PF sera are present in competition for the extracellular domain of Dsg3 (Dsg3-IgHis or residues 1Ð161 of Dsg1. PF sera that were not significantly competed 1Ð161 163Ð566 Dsg3-His) (Fig. 3). Similarly, the reactivity of the PV sera against by residues 1Ð161 (Dsg1 /Dsg3 ) tended to display signif- 1Ð161 164Ð496 Dsg3 was abolished by the extracellular domain of Dsg3, but was icant competition by residues 164Ð496 (Dsg3 /Dsg1 ), http://www.jimmunol.org/ not altered by that of Dsg1. These findings indicated that there was suggesting that the epitopes for these PF sera are present in residues no significant cross-reactivity between Dsg1 and Dsg3 in the pem- 164Ð401 of Dsg1. phigus patients’ sera tested in this study, and additionally showed To map Dsg3 epitopes recognized by anti-Dsg3 IgG in PV sera, that the Dsg1- and Dsg3-domain-swapped molecules could be competition levels for each of the above domain-swapped mole- used for epitope mapping. However, some original epitopes could cules were measured for the 40 PV sera samples (Fig. 5). Residues be lost by making swapping molecules due to the difference be- 1Ð403 of Dsg3 (Dsg31Ð403/Dsg1404Ð496) yielded an average com- tween Dsg1 and Dsg3 molecules, which is a limitation of our petition level of 85.8% (Fig. 5A), whereas residues 405Ð566 of approach. Dsg3 (Dsg11Ð401/Dsg3405Ð566) yielded an average level of only by guest on October 4, 2021 We characterized epitopes recognized by anti-Dsg1 IgG in PF 12.7% (Fig. 5C). Residues 1Ð161 of Dsg3 (Dsg31Ð161/Dsg1164Ð496) sera and by anti-Dsg3 IgG in PV sera. In addition, epitopes rec- yielded an average competition level of 67.4%; this level exceeded ognized by anti-Dsg1 IgG in PV sera were also examined, because 50% in 31 (77.5%) of the 40 samples (Fig. 5B). In contrast, resi- mucocutaneous PV sera contain anti-Dsg1 IgG as well as dues 163Ð566 (Dsg11Ð161/Dsg3163Ð566) gave an average competi- anti-Dsg3 IgG. tion level of 17.8%, and this level exceeded 50% only in four (10%) samples (Fig. 5D). PV sera that displayed no significant Most conformational epitopes recognized by pemphigus Abs competition by residues 1Ð161 (Dsg31Ð161/Dsg1164Ð496) tended to map to the N-terminal 161 amino acids of Dsg1 and Dsg3 show significant competition by residues 163Ð566 (Dsg11Ð161/ Four domain-swapped molecules, Dsg11Ð401/Dsg3405Ð566, Dsg11Ð161/ Dsg3163Ð566). These findings indicate that most of the epitopes of Dsg3163Ð566, Dsg31Ð403/Dsg1404Ð496, and Dsg31Ð161/Dsg1164Ð496 anti-Dsg3 IgG in PV sera are present in residues 1Ð161 of Dsg3

FIGURE 3. Competition ELISA of PF and PV sera using domain-swapped molecules as competitors. A, The reactivity of PF sera against Dsg1 was effectively depleted by recombinant proteins containing Dsg1 fragments in a dose-dependent fashion. B, The reactivity of PV sera against Dsg3 was similarly depleted by recombinant proteins containing Dsg3 fragments in a dose-dependent fashion. The level of competition obtained for recombinant proteins in culture supernatant (sup) is nearly equal to or higher than the maximum level of competition obtained overall. The Journal of Immunology 5443 Downloaded from http://www.jimmunol.org/

FIGURE 4. Gross conformational epitope mapping of Dsg1 with PF sera. Competition levels (%) against Dsg1-His were calculated for each recombinant molecule. PF sera were competed with Dsg11Ð401/Dsg3405Ð566(A), Dsg11Ð161/Dsg3163Ð566 (B), Dsg31Ð403/Dsg1404Ð496 (C), and Dsg31Ð161/Dsg1164Ð496 (D). The 43 PF sera tested are shown in the same order in each graph. Avg. indicates the average competition level for the 43 PF sera tested. and that some minor epitopes are present in the middle region of Results of epitope mapping for anti-Dsg1 IgG in PV sera were Dsg3 (residues 163Ð403). A comparison of mucosal dominant and essentially the same as those found for anti-Dsg1 IgG in PF sera (data mucocutaneous PV sera showed no significant differences in Dsg3 not shown). Dsg11Ð401/Dsg3405Ð566 and Dsg11Ð161/Dsg3163Ð566 epitope mapping (data not shown). displayed Ͼ50% competition in 14 (100%) and 12 (85.7%) of the by guest on October 4, 2021

FIGURE 5. Gross conformational epitope mapping of Dsg3 with PV sera. PV sera were competed with Dsg31Ð403/Dsg1404Ð496 (A), Dsg31Ð161/Dsg1164Ð496 (B), Dsg11Ð401/Dsg3405Ð566 (C), and Dsg11Ð161/Dsg3163Ð566 (D). The 40 PV sera tested are shown in the same order in each graph. Avg. indicates the average competition level obtained for the 40 PV sera tested. 5444 CONFORMATIONAL EPITOPE MAPPING OF PEMPHIGUS ANTIGENS

14 PV sera containing anti-Dsg1 IgG, respectively. In contrast, N-terminal 161 residues, as well as minor epitopes on the middle neither Dsg31Ð161/Dsg1164Ð496 nor Dsg31Ð403/Dsg1404Ð496 dis- residues 164Ð401. played significant competition in any of these samples. PF#73 and PF#2284 were immunoadsorbed with either Dsg11Ð161/ When taken together, our results show that the epitopes on Dsg1 Dsg3163Ð566 or Dsg31Ð161/Dsg1164Ð496 and injected into neonatal and Dsg3 have a very similar distribution. Although the localiza- mice (Figs. 6 and 7). Mice injected with PF#73 adsorbed with tion of epitopes was slightly different for each patient, most of the residues 1Ð161 of Dsg1 showed markedly diminished in vivo IgG epitopes of PF and PV sera were located in the N-terminal 161 deposition on keratinocyte cell surfaces and no apparent blister residues of Dsg1 and Dsg3, and some minor epitopes were found formation (data not shown). In contrast, mice injected with PF#73 in the middle regions of Dsg1 (residues 164Ð401) and Dsg3 (res- adsorbed with residues 164Ð496 demonstrated strong in vivo IgG idues 163Ð403). There were no apparent epitopes in the C-terminal deposition and extensive blister formation (data not shown). Mice extracellular domains of Dsg1 (residues 404Ð496) and Dsg3 (res- injected with PF#2284 adsorbed with residues 1Ð161 of Dsg1 did idues 405Ð566). not show apparent blisters, while mice injected with PF#2284 ad- sorbed with residues 164Ð496 showed extensive blisters, although The N-terminal 161 residues of Dsg1 contain critical pathogenic both mice showed in vivo IgG deposition (Fig. 7). The finding for epitopes for blister formation in PF PF#2284 indicated that the removal of Abs against the epitopes on To show the relevance of the above epitope mapping to the patho- residues 1Ð161 was sufficient to eliminate the pathogenic activity genesis of PF, we performed a passive transfer study with neonatal of this serum. mice using immunoadsorbed PF sera by domain-swapped mole- These findings indicate that the N-terminal 161 residues of Dsg1 contain critical epitopes recognized by pathogenic autoantibodies

cules. PF#73 showed 100 and 74.8% competition by residues Downloaded from 1Ð401 of Dsg1 (Dsg11Ð401/Dsg3405Ð566) and residues 1Ð161 of in the PF sera tested. Dsg1 (Dsg11Ð161/Dsg3163Ð566), respectively, but no significant com- petition by residues 164Ð496 of Dsg1 (Dsg31Ð161/Dsg1164Ð496)or Dominant conformational epitopes map to amino acids 26–87 residues 404Ð496 of Dsg1 (Dsg31Ð403/Dsg1404Ð496) (Fig. 6A). of Dsg1 and 25–88 of Dsg3 PF#2284 showed 96.3, 43.5, 12.5, and 5.9% competition by residues To further narrow down critical conformational epitopes on the 1Ð401, 1Ð161, 164Ð496, and 404Ð496, respectively (Fig. 6B). N-terminal regions of Dsg1 and Dsg3, six additional domain-swapped http://www.jimmunol.org/ Therefore, PF#73 contains the dominant epitopes on the N-terminal molecules (Dsg11Ð24/Dsg325Ð566, Dsg11Ð64/Dsg365Ð566, Dsg11Ð87/ 161 residues of Dsg1, while PF#2284 contains some epitopes on the Dsg387Ð566, Dsg31Ð26/Dsg126Ð496, Dsg31Ð63/Dsg163Ð496, and Dsg31Ð88/Dsg189Ð496) were used in competition ELISA. These studies were performed using the 26 PF sera samples and the 30 by guest on October 4, 2021

FIGURE 6. Competition ELISA of two PF sera using domain-swapped molecules in a passive transfer study. Two PF sera, PF#73 (A) and FIGURE 7. Removal of Abs against the N-terminal 161 residues of PF#2284 (B), were competed with Dsg1-His, Dsg3-His, Dsg11Ð401/ Dsg1 eliminated the pathogenic activity of PF sera. PF#2284 was immu- Dsg3405Ð566 (Dsg1(1Ð401)), Dsg11Ð161/Dsg3163Ð566 (Dsg1(1Ð161)), Dsg31Ð noadsorbed with Dsg11Ð161/Dsg3163Ð566 (A-C) or Dsg31Ð161/Dsg1164Ð496 164Ð496 1Ð403 404Ð496 161/Dsg1 (Dsg1(164Ð496)), and Dsg3 /Dsg1 (DÐF) and injected into neonatal mice. Mice injected with IgG adsorbed (Dsg1(404Ð496)) and their competition rates were calculated. The ability with residues 1Ð161 of Dsg1 did not show apparent blisters (A, B), while of each immunoadsorbed sera to induce superficial cutaneous blisters was mice injected with IgG adsorbed with residues 164Ð496 showed extensive determined in a passive transfer study of neonatal mice (Blister formation). blisters (D) with superficial blisters seen histologically (E), although both ND, not determined. mice showed in vivo IgG deposition (C and F). The bars indicate 50 ␮m. The Journal of Immunology 5445

PV sera samples that showed Ͼ50% competition with the N-ter- The general competition pattern observed for Dsg3 with PV sera minal 161 residues of Dsg1 and Dsg3, respectively. The compe- was quite similar to that of Dsg1 with PF sera. Residues 1Ð26 of tition levels in this set of experiments were calculated relative to Dsg3 (Dsg31Ð26/Dsg126Ð496) yielded an average of only 10.7% the results for residues 1Ð161 of Dsg1 or Dsg3. competition (Fig. 9A), whereas residues 25Ð161 yielded an aver- Residues 1Ð24 of Dsg1 (Dsg11Ð24/Dsg325Ð566) displayed an av- age of 80.8% competition. Twenty-six (86.7%) of the 30 PV sam- erage of only 14.9% competition (Fig. 8A), whereas residues 26Ð ples yielded Ͼ50% competition with residues 25Ð161 (Fig. 9D), 161 (Dsg31Ð26/Dsg126Ð496) displayed an average of 55.1% com- indicating that most of the Dsg3 epitopes are C-terminal to residue petition. Of the 26 PF sera samples, 13 (50%) gave Ͼ50% 25. In contrast, residues 1Ð88 had an average competition level of competition with residues 26Ð161 (Fig. 8D), indicating that the 53.3%, with 18 (60%) of the samples yielding Ͼ50% competition most of the epitopes are present in the region C-terminal to residue (Fig. 9C). Residues 87Ð161 gave an average of only 14.2% com- 26 of Dsg1. In contrast, residues 1Ð87 yielded an average com- petition (Fig. 9F), indicating that most of the epitopes are N-ter- petition level of 41%, and this level was Ͼ50% for 10 (38.5%) of minal to residue 88. Residues 1Ð63 (Dsg31Ð63/Dsg163Ð496) and the 26 samples (Fig. 8C). Residues 89Ð161 gave an average of 65Ð161 (Dsg11Ð64/Dsg365Ð566) yielded an average of 22.1 and only 18.4% competition (Fig. 8F), indicating that most of the 38.2% competition, respectively (Fig. 9, B and E). These findings epitopes are N-terminal to residue 87. Residues 1Ð64 (Dsg11Ð64/ indicate that most of the epitopes recognized by anti-Dsg3 IgG in Dsg365Ð566) and 63Ð161 (Dsg31Ð63/Dsg163Ð496) of Dsg1 displayed PV sera are within residues 25Ð88 of Dsg3. average competition levels of 38.5 and 39.3%, respectively (Fig. 8, 25 28 29 B and E). Results for mucocutaneous PV sera were similar (data Dsg1 residues His , Cys , and Ala define a Dsg1-specific not shown). These findings indicate that most of the epitopes rec- epitope Downloaded from ognized by anti-Dsg1 IgG in PF and PV sera are in the Dsg1 region Comparison of the Dsg1 and Dsg3 amino acid sequences reveals comprised of residues 26Ð87. that the N-terminal EC1 or EC2 domains are more conserved than http://www.jimmunol.org/ by guest on October 4, 2021

FIGURE 8. Residues 26Ð87 of Dsg1 contain dominant conformational epitopes. PF sera which demonstrated Ͼ50% competition with Dsg11Ð161/ Dsg3163Ð566 were further characterized with Dsg11Ð24/Dsg325Ð566 (A), Dsg11Ð64/Dsg365Ð566 (B), Dsg11Ð87/Dsg387Ð566 (C), Dsg31Ð26/Dsg126Ð496 (D), Dsg31Ð63/ Dsg163Ð496 (E), and Dsg31Ð88/Dsg189Ð496 (F). The competition levels were calculated against Dsg11Ð161/Dsg3163Ð566. The 26 PF sera tested are shown in the same order in each graph. 5446 CONFORMATIONAL EPITOPE MAPPING OF PEMPHIGUS ANTIGENS Downloaded from http://www.jimmunol.org/ by guest on October 4, 2021

FIGURE 9. Residues 25Ð88 of Dsg3 contain dominant conformational epitopes. PV sera which demonstrated Ͼ50% competition with Dsg31Ð161/ Dsg1164Ð496 were further characterized with Dsg31Ð26/Dsg126Ð496 (A), Dsg31Ð63/Dsg163Ð496 (B), Dsg31Ð88/Dsg189Ð496 (C), Dsg11Ð24/Dsg325Ð566 (D), Dsg11Ð64/Dsg365Ð566 (E), and Dsg11Ð87/Dsg387Ð566 (F). The competition levels were calculated against Dsg31Ð161/Dsg1164Ð496. The 30 PV sera tested are shown in the same order in each graph. the C-terminal EC3 or EC4 domains (2). The regions shown above (data not shown), indicating that the Dsg1 epitope defined by to contain the dominant conformational epitopes of Dsg1 and Dsg3 Dsg3-M1 is conformation-dependent (15). Similar studies with 12 are in the N-terminal EC1 domains. Residues 26Ð87 of Dsg1 in- PV sera samples containing anti-Dsg1 IgG gave similar results; clude several amino acids that are not conserved in Dsg3. Those Dsg3-M1 protein exhibited a significantly higher competition level nonconserved residues are likely to define the specific epitopes of than did other point-mutated Dsg3 molecules (data not shown). Dsg1 recognized by anti-Dsg1 pemphigus autoantibodies. Six non- Taken together, these findings demonstrate that one of the Dsg1- conserved Dsg3 clusters were chosen, and site-directed mutagen- specific epitopes is composed of residues His25, Cys28, and Ala29. esis was used to replace the residues in these clusters with Dsg1- specific residues (Fig. 1C and Table II). The six point-mutated Dsg3 proteins (Dsg3-M1 through -M6) Discussion were used as competitors in ELISA with 26 PF sera, and compe- In this study, we characterized the conformational epitopes of the tition levels were measured against the entire extracellular domain pemphigus autoantigens Dsg1 and Dsg3. We used a strategy based of Dsg1 (Fig. 10). Interestingly, Dsg3-M1, which contains His25, on domain swapping and point mutations to map regions within Cys28, and Ala29 of Dsg1, gave an average of 19.7% competition, Dsg1 and Dsg3 that constitute the conformational epitopes for PF with Ͼ50% competition in three (11.5%) of the 26 samples and and PV autoantibodies. Dsg1 and Dsg3 were used as swapping Ͼ30% competition in seven (27%) of the 26 samples. In contrast, partners for each other because they share a similar structure but no significant competition was observed for the five other point- no significant cross-reactivity (2, 20). These domain-swapped mol- mutated Dsg3 proteins, although Dsg3-M2 yielded a slightly ecules were used as competitors for ELISA against the entire ex- higher average competition level (8.8%) than the remaining tracellular domain of Dsg1 or Dsg3, allowing us to measure au- M3-M6 proteins (averages of 3Ð4.1%). This competition by toantibodies against specific regions in a quantitative fashion (21). Dsg3-M1 was abolished by acid-treatment (0.1 M citrate, pH 3.0) We tested 40 PV sera and 43 PF sera in this analysis. The epitopes The Journal of Immunology 5447

and EC2) of classic cadherin molecules have two dimer interfaces. One of these interfaces is formed between molecules in a parallel or lateral pair to form a lateral “strand dimer” believed to serve as a functional unit. The other interface is formed between molecules in an antiparallel pair and it is postulated that this interface corre- sponds to the adhesive interface between cadherins emanating from opposing cells. In addition to this structural evidence, bio- logical evidence supports lateral dimerization or clustering as an essential step in mediating cell-cell adhesion by classic cadherins (30Ð32). It is postulated that each individual strand dimer interacts with an opposing strand dimer on another cell through the adhesive interfaces. According to the predicted crystallographic structure (27), side chains directly involved in adhesive interfaces come from residues 35, 37, 39, 44, 45, 53, 54, 55, 56, 79, 81, 83, 84, and 86. If we superimpose our mapping results of Dsg1 and Dsg3 on the predicted three-dimensional structure of classic cadherins, res- idues 26Ð87 of Dsg1 and 25Ð88 of Dsg3 correspond to the region forming the adhesive interfaces. Therefore, these findings suggest that the pathogenic autoantibodies in PF and PV may be domi- Downloaded from nantly raised against the adhesive interfaces of Dsg1 and Dsg3. In contrast, Dsg epitopes recognized by some PF and PV auto- antibodies did not map to the N-terminal regions, but rather to the middle regions of the molecules (Figs. 4 and 5). These sera showed no significant competition with Dsg11Ð161 or Dsg31Ð161, but did show competition with Dsg1164Ð496 or Dsg3163Ð566. No significant clinical differences were noted between patients with autoantibod- http://www.jimmunol.org/ ies against the N-terminal domains vs those with autoantibodies against the middle regions (data not shown). Although these latter cases constitute a minor portion of the total, the existence of such cases suggests that the pathogenic IgG autoantibodies do not nec- essarily recognize the N-terminal region of Dsg1 or Dsg3, and that autoantibodies against the middle portion of the molecule may also be able to block the adhesive function of Dsg1 or Dsg3 and induce FIGURE 10. A point-mutated Dsg3 molecule containing Dsg1-specific

His25, Cys28, and Ala29 reacts with anti-Dsg1 IgG in PF sera. The 26 PF the loss of keratinocyte adhesion. by guest on October 4, 2021 sera which showed Ͼ50% competition with Dsg11Ð161/Dsg3163Ð566 were It is intriguing that pemphigus autoantibodies are raised against competed with six point-mutated Dsg3 molecules (M1 to M6). The com- the N-terminal rather than the C-terminal region of Dsg molecules. petition levels were calculated against Dsg1-His. When the extracellular domains of Dsg1 and Dsg3 are compared, the homology is greater in the N-terminal regions than in the C- terminal regions. The identities between Dsg1 and Dsg3 are 73% of Dsg1 were studied with PF sera and anti-Dsg1 IgG-positive PV in EC1, 65% in EC2, 57% in EC3, and 28% in EC4, with no sera, and the epitopes of Dsg3 were studied with PV sera. significant homology for the EC5 domain (2). Therefore, specific We first performed gross mapping of the epitopes of Dsg1 and recognition of Dsg1 or Dsg3 as “non-self” by the immune system Dsg3, using four domain-swapped molecules, which divide the would seem most likely to occur through Abs against the C-ter- entire extracellular domain of Dsg1 and Dsg3 into three parts (Fig. minal regions, which contain more isotype-specific residues. How- 1A). In both PF and PV patients, most epitopes were mapped to the ever, pemphigus autoantibodies are not raised against the C-ter- respective N-terminal 161 residues of Dsg1 and Dsg3 (Figs. 4 and minal domains. This finding may reflect some aspect of the 5). These N-terminal 161 residues contained critical epitopes rec- pathophysiological mechanism that triggers the autoimmune reac- ognized by pathogenic IgG Abs because the immunoadsorption of tion against Dsgs. two PF sera with residues 1Ð161 of Dsg1 removed their activity to Our study is the first comprehensive conformational epitope induce blisters in neonatal mice (Figs. 6 and 7). We then used mapping analysis of Dsg1 and Dsg3 in PV and PF. These findings another six domain-swapped molecules to narrow down the dom- provide new knowledge useful for elucidating the molecular mech- inant epitopes within this region (Figs. 8 and 9). Although there anism of adhesion by Dsg1 and Dsg3, as well as the pathophysi- was no single major epitope on Dsg1 and Dsg3, dominant epitopes ological mechanism of blister formation in pemphigus. In addition, were mapped to residues 26Ð87 of Dsg1 and 25Ð88 of Dsg3, both our finding provides a basis to develop epitope-specific therapeutic of which are in the N-terminal EC1 domain. Furthermore, within strategies for pemphigus. these residues, we found that one of the epitopes of Dsg1 was 25 28 29 defined by His , Cys , and Ala , using point-mutated Dsg3 mol- Acknowledgments ecules containing Dsg1-specific residues (Fig. 10). We thank Dr. Atsushi Takayanagi for helpful discussions regarding site- Although the three-dimensional molecular structures and pre- directed mutagenesis. We also thank Minae Suzuki for performing the cise functional domain maps of Dsg1 and Dsg3 remain to be elu- immunofluorescence characterization of sera. cidated, nuclear magnetic resonance spectroscopy and radio- graphic characterization have revealed the three-dimensional References molecular structure of the N-terminal domains of classic cadherins 1. Koch, P. J., and W. W. Franke. 1994. Desmosomal cadherins: another growing (26Ð29). It is generally agreed that the N-terminal domains (EC1 multigene family of adhesion molecules. Curr. Opin. Cell Biol. 6:682. 5448 CONFORMATIONAL EPITOPE MAPPING OF PEMPHIGUS ANTIGENS

2. Amagai, M., V. Klaus-Kovtun, and J. R. Stanley. 1991. Autoantibodies against a 17. Fan, J. L., O. Memar, D. J. McCormick, and B. S. Prabhakar. 1999. BALB/c mice novel epithelial cadherin in pemphigus vulgaris, a disease of cell adhesion. Cell produce blister-causing antibodies upon immunization with a recombinant human 67:869. desmoglein 3. J. Immunol. 163:6228. 3. Schafer, S., P. J. Koch, and W. W. Franke. 1994. Identification of the ubiquitous 18. Amagai, M., K. Tsunoda, H. Suzuki, K. Nishifuji, S. Koyasu, and T. Nishikawa. human desmoglein, Dsg2, and the expression catalogue of the desmoglein sub- 2000. Use of autoantigen knockout mice to develop an active autoimmune dis- family of desmosomal cadherins. Exp. Cell Res. 211:391. ease model of pemphigus. J. Clin. Invest. 105:625. 4. Stanley, J. R. 1993. Cell adhesion molecules as targets of autoantibodies in pem- 19. Futei, Y., M. Amagai, M. Sekiguchi, K. Nishifuji, Y. Fujii, and T. Nishikawa. phigus and pemphigoid, bullous diseases due to defective epidermal cell adhe- 2000. Conformational epitope mapping of desmoglein 3 using domain-swapped sion. Adv. Immunol. 53:291. molecules in pemphigus vulgaris. J. Invest. Dermatol. 115:829. 5. Amagai, M. 1995. Adhesion molecules. I. Keratinocyte-keratinocyte interactions, 20. Ishii, K., M. Amagai, R. P. Hall, T. Hashimoto, A. Takayanagi, S. Gamou, cadherins and pemphigus. J. Invest. Dermatol. 104:146. N. Shimizu, and T. Nishikawa. 1997. Characterization of autoantibodies in pem- 6. Schiltz, J. R., and B. Michel. 1976. Production of epidermal acantholysis in phigus using antigen-specific ELISAs with baculovirus expressed recombinant normal human skin in vitro by the IgG fraction from pemphigus serum. J. Invest. desmogleins. J. Immunol. 159:2010. Dermatol. 67:254. 21. Amagai, M., A. Komai, T. Hashimoto, Y. Shirakata, K. Hashimoto, T. Yamada, 7. Anhalt, G. J., R. S. Labib, J. J. Voorhees, T. F. Beals, and L. A. Diaz. 1982. Y. Kitajima, K. Ohya, H. Iwanami, and T. Nishikawa. 1999. Usefulness of en- Induction of pemphigus in neonatal mice by passive transfer of IgG from patients zyme-linked immunosorbent assay (ELISA) using recombinant desmogleins 1 with the disease. N. Engl. J. Med. 306:1189. and 3 for serodiagnosis of pemphigus. Br. J. Dermatol. 140:351. 8. Hashimoto, K., K. M. Shafran, P. S. Webber, G. S. Lazarus, and K. H. Singer. 22. Amagai, M., T. Nishikawa, H. C. Nousari, G. J. Anhalt, and T. Hashimoto. 1998. 1983. Anti-cell surface pemphigus autoantibody stimulates plasminogen activator Antibodies against desmoglein 3 (pemphigus vulgaris antigen) are present in sera activity of human epidermal cells. J. Exp. Med. 157:259. from patients with and cause acantholysis in vivo in neonatal mice. J. Clin. Invest. 102:775. 9. Amagai, M., S. Karpati, R. Prussick, V. Klaus-Kovtun, and J. R. Stanley. 1992. 23. Deng, W.-P., and J. A. Nickoloff. 1992. Site-directed mutagenesis of virtually any Autoantibodies against the amino-terminal cadherin-like binding domain of pem- plasmid by eliminating a unique site. Anal. Biochem. 200:81. phigus vulgaris antigen are pathogenic. J. Clin. Invest. 90:919. 24. Barik, S. 1993. Site-directed mutagenesis by double polymerase chain reaction. 10. Mahoney, M. G., Z. Wang, K. L. Rothenberger, P. J. Koch, M. Amagai, and In PCR Protocols, Vol. 15. B. A. White, ed. Humana Press, Totowa, NJ, p. 277. J. R. Stanley. 1999. Explanation for the clinical and microscopic localization of 25. Proby, C. M., T. Ohta, H. Suzuki, S. Koyasu, S. Gamou, N. Shimizu, Downloaded from lesions in pemphigus foliaceus and vulgaris. J. Clin. Invest. 103:461. M. J. Wheelock, T. Nishikawa, and M. Amagai. 2000. Development of chimeric 11. Ding, X., V. Aoki, J. M. Mascaro, A. Lopez-Swiderski, L. A. Diaz, and molecules for recognition and targeting of antigen-specific B cells in pemphigus J. A. Fairley. 1997. Mucosal and mucocutaneous (generalized) pemphigus vul- vulgaris. Br. J. Dermatol. 142:321. garis show distinct autoantibody profiles. J. Invest. Dermatol. 109:592. 26. Overduin, M., T. S. Harvey, S. Bagby, K. I. Tong, P. Yau, M. Takeichi, and 12. Amagai, M., K. Tsunoda, D. Zillikens, T. Nagai, and T. Nishikawa. 1999. The M. Ikura. 1995. Solution structure of the epithelial cadherin domain responsible clinical phenotype of pemphigus is defined by the anti-desmoglein autoantibody for selective cell adhesion. Science 267:386. profile. J. Am. Acad. Dermatol. 40:167. 27. Shapiro, L., A. M. Fannon, P. D. Kwong, A. Thompson, M. S. Lehmann, 13. Amagai, M., T. Hashimoto, N. Shimizu, and T. Nishikawa. 1994. Absorption of G. Grubel, J. F. Legrand, J. Als-Nielsen, D. R. Colman, and W. A. Hendrickson.

pathogenic autoantibodies by the extracellular domain of pemphigus vulgaris 1995. Structural basis of cell-cell adhesion by cadherins. Nature 374:327. http://www.jimmunol.org/ antigen (Dsg3) produced by baculovirus. J. Clin. Invest. 94:59. 28. Nagar, B., M. Overduin, M. Ikura, and J. M. Rini. 1996. Structural basis of 14. Amagai, M., T. Hashimoto, K. J. Green, N. Shimizu, and T. Nishikawa. 1995. calcium-induced E-cadherin rigidification and dimerization. Nature 380:360. Antigen-specific immunoadsorption of pathogenic autoantibodies in pemphigus 29. Tamura, K., W.-S. Shan, W. A. Hendrickson, D. R. Colman, and L. Shapiro. foliaceus. J. Invest. Dermatol. 104:895. 1998. Structure-function analysis of cell adhesion by neural (N-)cadherin. Neuron 15. Amagai, M., K. Ishii, T. Hashimoto, S. Gamou, N. Shimizu, and T. Nishikawa. 20:1153. 1995. Conformational epitopes of pemphigus antigens (Dsg1 and Dsg3) are cal- 30. Brieher, W. M., A. S. Yap, and B. M. Gumbiner. 1996. Lateral dimerization is cium dependent and glycosylation independent. J. Invest. Dermatol. 105:243. required for the homophilic binding activity of C-cadherin. J. Cell Biol. 135:487. 16. Memar, O., B. Christensen, S. Rajaraman, R. Goldblum, S. K. Tyring, 31. Yap, A. S., W. M. Brieher, M. Pruschy, and B. M. Gumbiner. 1997. Lateral M. M. Brysk, D. J. McCormick, H. Zaim, J. L. Fan, and B. S. Prabhakar. 1996. clustering of the adhesive ectodomain: a fundamental determinant of cadherin Induction of blister-causing antibodies by a recombinant full-length, but not the function. Curr. Biol. 7:308. extracellular, domain of the pemphigus vulgaris antigen (desmoglein 3). J. Im- 32. Vasioukhin, V., C. Bauer, M. Yin, and E. Fuchs. 2000. Directed polymer- munol. 157:3171. ization is the driving force for epithelial cell-cell adhesion. Cell 100:209. by guest on October 4, 2021