Secretory Formation: Conserved Binding Interactions between and Polymeric Ig Receptor from Humans and Amphibians This information is current as of September 29, 2021. Ranveig Braathen, Valerie S. Hohman, Per Brandtzaeg and Finn-Eirik Johansen J Immunol 2007; 178:1589-1597; ; doi: 10.4049/jimmunol.178.3.1589 http://www.jimmunol.org/content/178/3/1589 Downloaded from

References This article cites 48 articles, 21 of which you can access for free at:

http://www.jimmunol.org/content/178/3/1589.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 by guest on September 29, 2021

*average

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 © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Secretory Antibody Formation: Conserved Binding Interactions between J Chain and Polymeric Ig Receptor from Humans and Amphibians1

Ranveig Braathen,* Valerie S. Hohman,† Per Brandtzaeg,* and Finn-Eirik Johansen2*

Abs of the secretory Ig (SIg) system reinforce numerous innate defense mechanisms to protect the mucosal surfaces against microbial penetration. SIgs are generated by a unique cooperation between two distinct cell types: plasma cells that produce polymers of IgA or IgM (collectively called pIgs) and polymeric Ig receptor (pIgR)-expressing secretory epithelial cells that mediate export of the pIgs to the lumen. Apical delivery of SIgs occurs by cleavage of the pIgR to release its extracellular part as a pIg-bound secretory component, whereas free secretory components are derived from an unoccupied receptor. The joining chain (J chain) is crucial in pIg/SIg formation because it serves to polymerize Igs and endows them with a binding site for the pIgR. In Downloaded from this study, we show that the J chain from divergent tetrapods including mammals, birds, and amphibians efficiently induced polymerization of human IgA, whereas the J chain from nurse shark (a lower vertebrate) did not. Correctly assembled polymers showed high affinity to human pIgR. Sequence analysis of the J chain identified two regions, conserved only in tetrapods, which by mutational analysis were found essential for pIgA-pIgR complexing. Furthermore, we isolated and characterized pIgR from the amphibian Xenopus laevis and demonstrated that its pIg binding domain showed high affinity to human pIgA. These results

showed that the functional site of interaction between pIgR, J chain and Ig H chains is conserved in these species and suggests that http://www.jimmunol.org/ SIgs originated in an ancestor common to tetrapods. The Journal of Immunology, 2007, 178: 1589–1597.

ecretory IgA (SIgA)3 has been recognized as the major Ab with innate defense mechanisms on the mucosal surfaces to pro- class in mammals for Ͼ40 years. The conventional expla- vide protection against pathogens and to restrict commensals to the S nation for this huge SIgA production has been that it must lumen (5–7). protect the large mucosal surfaces against a many different bacte- Plasma cells located in the mucosal lamina propria of the gas- rial and viral pathogens. Recently, however, the role of SIgA in trointestinal and respiratory tracts are preferentially switched to host interactions with the commensal bacteria in the intestine has IgA and coexpress the joining chain (J chain). Therefore, mucosal been revealed (1). Thus SIgA is important both for defense against plasma cells secrete mostly polymeric IgA (pIgA; predominantly by guest on September 29, 2021 mucosal pathogens but also to restrict the indigenous gut flora to dimers) with the capacity to bind to the polymeric Ig receptor the lumen. (pIgR) (8, 9). SIgA production is initiated when pIgA binds pIgR Animals are adapted to live with vast number of bacteria colo- expressed on the basolateral side of epithelial cells lining the mu- nizing their mucosal surfaces. Such mutualistic microorganisms cosal surfaces. Endocytosis and intracellular trafficking of the re- provide benefit to the host, not only as competitors of pathogenic ceptor-ligand complex to the apical plasma membrane is mediated bacteria but also through direct positive tissue interactions (2, 3). by sorting signals in the cytoplasmic tail of the pIgR, and lumenal However, if the mucosal barrier is breached the indigenous micro- delivery of SIgA occurs by cleavage of the receptor near the biota may cause life-threatening infections or a severe tissue-dam- plasma membrane (10). Covalently bound secretory component aging response from the systemic immune system. In mammals, (SC; the extracellular cleaved piece of the pIgR) confers muco- luminal bacteria induce production of IgA Abs (4) that cooperate philic properties and proteolytic resistance to the SIgA molecule (11). Similarly generated SIgM Abs are mainly important in the neonate and in individuals with selective IgA deficiency (12), *Laboratory for Immunohistochemistry and Immunopathology, Institute of Pathol- whereas constitutive transcytosis of unoccupied pIgR leads to the ogy, University of Oslo, Rikshospitalet-Radiumhospitalet Medical Center, Oslo, Nor- secretion of free SC (12). † way; and Department of Biology, University of San Diego, San Diego, CA 92110 The J chain has been characterized in a range of species includ- Received for publication September 22, 2006. Accepted for publication November ing mammals (human (13, 14), mouse (15), bovine (16), rabbit 16, 2006. (17), and brushtail possum (18)), birds (chicken (19)), reptiles (tur- The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance tle (20)), amphibians (Xenopus laevis (21) and Rana catesbeiana with 18 U.S.C. Section 1734 solely to indicate this fact. (22)), and cartilaginous fish (nurse shark (23) and clearnose skate 1 Funding for this project has been provided by the Research Council of Norway. (24)). In humans, this unique 15-kD polypeptide contains eight 2 Address correspondence and reprint requests to Dr. Finn-Eirik Johansen, Laboratory Cys residues that form three intrachain bridges and two bridges to for Immunohistochemistry and Immunopathology, Institute of Pathology, University the tailpieces of IgA or IgM in human pIgs (25). No crystal struc- of Oslo, Rikshospitalet-Radiumhospitalet Medical Center, N-0027 Oslo, Norway. E-mail address: [email protected] ture is available for SIgA or SIgM, but several lines of evidence 3 Abbreviations used in this paper: SIgA, secretory IgA; CHO, Chinese hamster suggest that J chain and pIgR interact directly, although this re- ovary; D1, domain 1; D2, domain 2; J chain, joining chain; NIP, -iodo-4-hydroxy- mains controversial (8, 9, 25, 26). 2-nitrophenylacetyl; pIgA, polymeric IgA; pIgR, polymeric Ig receptor; SC, secretory Monomeric IgA (composed of two H chains and two L chains) component. readily form polymers when coexpressed with a J chain (13). Ear- Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 lier investigations into the function of the J chain showed that www.jimmunol.org 1590 CONSERVATION OF SECRETORY ANTIBODY FORMATION IN TETRAPODS

Table I. Characteristics of J chain variants used in this study

J-chain variant Sequencea Polymersb J Chainc

Mock 16 Human 65 100 Ϯ 9 Mouse 51 Chicken 41 Xenopus 50 Bullfrog 16 Nurse shark 19 Region 1 wild type 25-SSEDPNED-32 65 100 Ϯ 9 SSE25AAA AAADPNED 80 124 Ϯ 20 ED27AA SSAAPNED 67 124 Ϯ 4 DP28AA SSEAANED 64 107 Ϯ 5 NED30AAA SSEDPAAA 67 88 Ϯ 5 Region 2 wild type 45-NNRENISDPTS-55 65 100 Ϯ 9 NN45AA AARENISDPTS 75 115 Ϯ 13 RE47AA NNAANISDPTS 81 103 Ϯ 2 ISD50AAA NNRENAAAPTS 55 100 Ϯ 10 TS54AA NNRENISDPAA 62 122 Ϯ 10 N49A NNREAISDPTS 46 140 Ϯ 6 ISD50ASA NNRENASAPTS 72 107 Ϯ 9 Downloaded from Region 3 wild type 102-YTYDRNK-108 65 100 Ϯ 9 YTY102AAA AAADRNK 70 116 Ϯ 7 RNK106AAA YTYDAAA 68 103 Ϯ 3 YTY102FTF FTFDRNK 70 100 Ϯ 3 T103A YAYDRNK 58 55 Ϯ 8 D105A YTYARNK 84 110 Ϯ 12

R106A YTYDANK 73 111 Ϯ 14 http://www.jimmunol.org/ N107A YTYDRAK 89 119 Ϯ 19 K108A YTYDRNA 67 77 Ϯ 2 D105R YTYRRNK 87 83 Ϯ 3 R106E YTYDENK 44 66 Ϯ 2 DR105RE YTYRENK 73 91 Ϯ 7

a letters in bold font represent a substitution. b The amount of polymer secreted by CHO cells cotransfected with IgA and the indicated J chain calculated as a percentage of total IgA as described in Materials and Methods. c J chain content was determined by ELISA as described in Materials and Methods. J chain content in the polymeric IgA fraction co-expressed with wild-type human IgA was set to 100%. by guest on September 29, 2021 structural requirements for polymerization of IgA were less strin- amino acid sequence of the extracellular domain 1 (D1) of human or gent than the requirements for pIgR/SC binding of the resulting chicken pIgR as the query both identified an expressed sequence tag pu- polymers (27). In this study we show that J chain from several tatively encoding X. laevis pIgR. The clone (IMAGE 4031968) was ob- tained from Geneservice and sequenced in full (GenBank accession no. different tetrapod species is capable of inducing polymerization of EF079076). An expression plasmid for xD1-hD2D5 SC was constructed by human IgA and that the resulting pIgA binds human pIgR/SC. fusing the D1 of X. laevis pIgR with the domain 2 (D2) from human pIgR Furthermore, we identify two conserved regions in the J chain that by splice overlap extension and subcloning as described above (domain are necessary for the latter function only. Finally we characterize boundaries are indicated in Fig. 3). An expression plasmid for hD2D5-SC was constructed by deleting the D1 from the human SC expression plasmid the first amphibian pIgR and show that its pIg binding domain is by QuikChange PCR. capable of binding human pIgA. These findings reveal a remark- able conservation of protein-protein interaction sites in secretory Abs from different species, suggesting an evolutionary significance Production and purification of recombinant IgA and SC of this first-line defense system. The 5-iodo-4-hydroxy-2-nitrophenylacetyl (NIP)-specific IgA1-producing Chinese hamster ovary (CHO) cell line ␣4.2 (27) was cultured in Ham’s- Materials and Methods F12 medium supplemented with 10% FCS, 2 mM glutamine, and 50 ␮g/ml Plasmid construction gentamicin. Semistable transfections (episomal expression) were per- formed as previously described (27). Briefly, cells were seeded at 10–15% The plasmid for episomal expression of wild-type human J chain (pCEP4- confluence in 6-well trays on day 1 and transfected on day 2 with 1.4 ␮g wtJ chain) has been described previously (27); this expression vector con- of DNA and 3 ␮l of FuGENE (Roche Diagnostics) according to the man- tained ori P and encoded EBV nuclear Ag (EBNA) for episomal repli- ufacturer’s protocol. On day 3, the cells received fresh medium and on day cation, HygR for antibiotic selection, and CMV promoter for cDNA 4 they were trypsinized and seeded into 10-cm plates in medium supple- expression. The J chain open reading frame from mouse (15), chicken (28), mented with 300 ␮g/ml hygromycin B (Roche Diagnostics). Cells re- X. laevis (21), bullfrog (22), and nurse shark (23) was isolated by PCR and ceived fresh medium with hygromycin B every 3–4 days; after 10–12 subcloned into pCEP4. Mutagenesis of the human J chain was performed days each 10-cm plate contained at least 500 hygromycin B-resistant by QuikChange PCR (Stratagene) or by splice overlap extension and sub- colonies that were pooled and seeded into 10-cm plates or large flasks cloning into a HindIII- and BamHI-digested pCEP4. The entire open read- for IgA production. The vector pCEP4 without insert was used as neg- ing frame of all plasmids was verified by DNA sequencing ative control (mock). (Medigenomics). IgA Abs secreted from pools of Ͼ500 transfectants were affinity-puri- A plasmid for recombinant expression of human SC was constructed by fied on a NIP-Sepharose column (13). Fractions containing monomeric or PCR with a forward primer introducing a Kozac sequence at the ATG pIgA were separated by ion-exchange (MonoQ) chromatography (27). initiation codon and a reverse primer inserting a stop codon at the cleavage Briefly, IgA was eluted by a linear salt gradient from 2.5–500 mM NaCl in site for SC. The resulting PCR product was ligated into pCDNA3.1 direc- 20 mM Tris-HCl (pH 7.1) and collected in 300-␮l fractions. Peak fractions tional TOPO (Invitrogen Life Technologies). A TBLASTN search with the containing pIgA were pooled and analyzed for molecular size by native gel The Journal of Immunology 1591 electrophoresis and by ELISA for binding to free SC and for J chain con- tent (see below). The J558L transfectoma cell lines expressing human Fc-containing IgA1, IgM, and IgG1 have been described previously (29, 30). Crude cell supernatants were used in ELISA for SC binding analyses. Production of recombinant human free SC and derivatives were per- formed essentially as described (31). Briefly, one 10-cm plate with 293E cells was transiently transfected with 6 ␮g of each expression vectors using FuGENE. Supernatants were harvested after 3–7 days, and the SC was precipitated by ammonium acetate, resuspended, and dialyzed against PBS. Immunoprecipitations, gel electrophoresis, and Western blotting Approximately 500 ng of IgA from CHO cell supernatants were immuno- precipitated with rabbit anti-human IgA (diluted 1/100; DakoCytomation) and sheep anti-rabbit Ig-coated magnetic beads (7 ϫ 106 beads in 1 ml; Dynal). Samples were resolved by nonreducing SDS-PAGE on a 5% (w/v) acrylamide gel and transferred onto a polyvinylidene fluoride membrane (Millipore) that, after air drying, was blocked with PBST (10% skimmed milk in PBS plus 0.05% Tween 20). Subsequent incu-

bations were all performed in PBST. The membrane was incubated with Downloaded from a mouse antiserum to human IgA (diluted 1/3000; gift from T. Lea, Institute of Immunology, Rikshospitalet-Radiumhospitalet Medical Center, University of Oslo, Norway) for 90 min, washed three times for 10 min, and then incubated with HRP-conjugated sheep anti-mouse IgG (diluted 1/3000; Amersham Biosciences) for 90 min. After three washes, the membrane was incubated for 10 min with SuperSignal (Pierce) and the substrate was detected with a light-sensitive camera (Chemidoc; Bio-Rad). http://www.jimmunol.org/ For analysis of NIP-purified IgA, 30 ng were resolved by nonreducing SDS-PAGE on a 5% (w/v) acrylamide gel with the Criterion system (Bio- Rad) and transferred and probed as described above except that the mem- brane was first incubated with a rabbit antiserum to human IgA (diluted 1/3000; DakoCytomation) and next incubated with HRP-conjugated don- key anti-rabbit Ig (diluted 1/3000; Amersham Biosciences). For J chain immunoblotting, the membrane was stripped in buffer (53 mM Tris (pH 6.8), 1.6% SDS, and 14.3 mM 2-ME) at 60°C (15 min) and then blocked again with 10% (w/v) skimmed milk in PBST and incubated with our rabbit antiserum to human J chain (previously absorbed with monomeric by guest on September 29, 2021 IgA) diluted 1/800 (32). Secondary Ab and the revealing reaction were as described above. For native immunoblots, 60 ng of IgA was resolved on a 5% (w/v) acrylamide gel as described above. SDS was omitted from all buffers and the pH of the loading buffer was 8.5. The gel was transferred to a polyvi- nylidene difluoride membrane and the membrane was probed as for anal- ysis of NIP-purified IgA as described above. ELISAs For all Ag-specific ELISAs, microtiter plates were coated with 3 ␮g/ml NIP-BSA and blocked with 1% (w/v) BSA in PBS. For total IgA quanti- fication, the primary Ab was mAb against human IgA (diluted 1/30 000; a gift from T. Lea, Institute of Immunology, Rikshospitalet University Hospital, Oslo, Norway) and the secondary Ab was alkaline phosphatase-conjugated rabbit anti-mouse Ig Ab (diluted 1/1000; DakoCytomation). FIGURE 1. J chains from different vertebrates facilitate polymerization For J chain-specific ELISA (Table I) and the SC-affinity ELISA of human IgA and binding to human SC. A, NIP-specific recombinant IgA (Figs. 1C and 2B), 800 ng/ml, 400 ng/ml, 200 ng/ml, and 100 ng/ml without a J chain (mock) or containing a J chain from different species was purified recombinant pIgA was incubated in triplicate wells overnight. immunoprecipitated with polyclonal anti-human IgA serum, subjected to For J chain-specific ELISA, the microtiter plates were washed and fixed nonreducing SDS-PAGE, and immunoblotted with a mAb Ab against hu- with 2% glutaraldehyde for 30 min then incubated with 6 M urea (pH man IgA. Migration of monomers and dimers are indicated. The percentage 3.0) for an additional 30 min to reveal bound J chain by denaturation of polymers for each sample is indicated above each lane. B, Binding of (27). After washing, the plates were incubated with our rabbit antiserum recombinant human SC to CHO cell-produced IgA with different J chains to human J chain (diluted 1/300) and alkaline phosphatase-conjugated Materials and Methods goat anti-rabbit Ig (diluted 1/2000; DakoCytomation). The J chain con- was analyzed by ELISA as described in . Mean OD ϩ tent of each pIgA sample was normalized to pIgA with a wild-type values SD from quadruplicate wells in one of three similar experiments human J chain. For SC affinity measurement, the microtiter plates were is shown. C, The polymeric fractions shown in A were purified by chro- washed and incubated with 8 ␮g/ml free SC purified from human co- matography and tested for binding to native human secretory component in lostrum (33), then rabbit antiserum to human SC (DakoCytomation; an ELISA as described in Materials and Methods. Mean Ϯ SEM of three diluted 1/1000), and finally goat anti-rabbit Ig as above. The values independent experiments. Binding values that differ from wild-type J chain were expressed as the slope of the linear regression of OD405 vs IgA or from mock control (p Ͻ 0.05; two-tailed Student’s t test) are indicated .or a closed circle (F), respectively (ء) concentration and normalized to the corresponding slope obtained for by a star NIP-specific myeloma pIgA (27). ELISA for recombinant human and chimeric SC was performed by coating microtiter plates with 2 ␮g/ml SpsA surface protein from S. pneumoniae, which requires intact human D3D4 for SC/pIgR binding goat anti-mouse IgG (1/2000; Sigma-Aldrich) and color reaction (31). Detection was with a mAb against human pIgR D2 (present in all with tetramethylbenzidine peroxidase substrate (Kirkegaard & Perry SC variants) (1/5000; Sigma-Aldrich) followed by HRP-conjugated Laboratories). 1592 CONSERVATION OF SECRETORY ANTIBODY FORMATION IN TETRAPODS Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 2. Two conserved regions in the J chain are critical for binding of polymeric IgA to SC. A, The amino acid sequences of the J chains from human, mouse, chicken, bullfrog, X. laevis, and nurse shark were aligned with ClustalW. Dashes denote identity to residues in the human sequence, and dots denote gaps. Amino acid numbering, predicted ␤-strands, and surface accessibility (bold font) of human J chain are indicated. SS marks the three intrachain disulfide bridges and “TP” marks the Cys paired with the secretory tail piece of IgA or IgM. The conserved N-linked glycosylation site is Secondary structure prediction and surface were performed with JPRED (http://www.compbio.dundee.ac.uk/). B, NIP-specific pIgA .(ء) indicated by a star containing mutants of human J chain was purified and analyzed as described in the legend to Fig. 1. Names of mutants include original amino acid (left), the position as indicated in A, and the new amino acid (right). Where several amino acids were altered, the first amino acid in the range is indicated. Mean ϩ SEM of four independent experiments. Binding values that differ from wild-type J chain or from mock control (p Ͻ 0.05; two-tailed Student’s t test) are indicated .or a closed circle (F), respectively (ء) by a star The Journal of Immunology 1593

The binding of recombinant human and chimeric SC to pIgA1 or Table II. Sequence conservation of regions of J chaina pentameric IgM or IgG1 was performed by coating microtiter plates with NIP-BSA as described above and saturating with NIP-specific Region 1 Region 2 Region 3 Overall J558L transfectoma-produced Ig (ϳ1.5 ␮g/ml). Dilution series of SC were added and detected as in ELISA for recombinant SC. For the Mouse 88 100 86 76 binding of recombinant human and chimeric SC to CHO-produced IgA Chicken 38 82 71 58 containing human, mouse, chicken, Xenopus, bullfrog, or nurse shark J Xenopus 63 82 100 56 chain, CHO cell supernatants were precipitated with 40% (w/v) ammo- Bullfrog 50 64 100 48 nium acetate, resuspended in PBS, and the concentration was adjusted Nurse shark 13 82 29 31 to 1.5 ␮g/ml. ELISA for SC binding analysis was performed as de- scribed above. a Percentage of the amino acid identity of the J chain from the indicated species to the human J chain in region 1, region 2, region 3, and overall. See Fig. 2A for alignment and location of each region. Results J chains from divergent tetrapods support pIgA and SIgA formation with human IgA and human secretory component (SC) ants with human IgA. All mutants supported pIgA formation (Ta- ble I). When analyzing pIgA binding to human SC, we found that We coexpressed the J chain from human, mouse, chicken, bull- mutation of two Tyr residues (Tyr102 and Tyr104) to Ala was not frog, X. laevis or nurse shark with human IgA and analyzed the tolerated, suggesting a critical role in pIgR interaction. Neverthe- secreted IgA for molecular size and pIgR affinity (Fig. 1). Al- less, Phe residues were able to substitute in these positions without though the human J chain induced the greatest level of poly- loss of SC binding. Therefore, we concluded that these Tyr resi- Downloaded from merization, the J chain from mouse, chicken, or Xenopus also in- dues are most likely required for a hydrophobic “core” rather than duced significant pIgA formation (Fig. 1A; Table I). We then tested for direct pIgR interaction. Several different mutations that affected the binding of secreted IgA to recombinant human free SC (Fig. 1B). Arg106 significantly reduced SC binding to the pIgA produced. We found that the J chains from different species that induced poly- Substitution of this Arg for Ala or for a negatively charged Glu merization of human IgA efficiently also were part of a polymer with (R106E) had a similar effect (Fig. 2B; compare R106A with high affinity to human SC. Surprisingly, the small amounts of pIgA 105 R106E). Altering the negatively charged Asp to Arg did not http://www.jimmunol.org/ formed in the presence of the bullfrog J chain showed significant compensate for the R106E mutation (DR105RE). Therefore, it ap- affinity to human SC. Conversely, the small amounts of pIgA formed peared that Arg-106 was essential for pIgR interaction. Although in the absence of J chain (or when IgA was coexpressed with the triple mutation (RNK106AAA) tended to produce a slightly nurse shark J chain) showed virtually no affinity to SC. To less functionally active pIgA than R106A, there was no statistical eliminate possible bias in the SC binding experiments due to difference between these mutants. Furthermore, no other individual different levels of polymerization, we purified each pIgA frac- amino acid in this region resulted in a J chain with significant loss tion and analyzed its ability to bind native human free SC pu- of function. Altogether, Arg106 was the only amino acid in region rified from the colostrum (Fig. 1C). These results confirmed the 3 found to be essential for pIgR interaction. binding pattern described above. by guest on September 29, 2021 X. laevis pIgR binds human pIgA but not pentameric IgM Two motifs in human J chain are important for pIgA Having established that the J chains from birds and amphibians affinity to pIgR/SC efficiently substituted for human J chain in the formation of pIgA We identified two regions of high amino acid conservation in with a pIgR docking site, we made the assumption that amphibians the J chain from tetrapods (regions 2 and 3), but only one of encode a pIgR. By BLAST search, we identified an expressed them (region 2) was conserved in the shark (Fig. 2A; Table II). sequence tag encoding X. laevis pIgR and obtained the clone from We tested a series of point mutations of these two regions in the Geneservice. The sequencing of cDNA from Xenopus revealed human J chain as well as a third amino-terminal region pre- that its pIgR contained four extracellular Ig-like domains similar to dicted to be in a surface-exposed loop (region 1) for the effect the pIgR from chicken (Fig. 3) but in contrast to mammalian pIgR, on pIgA formation and SC binding (Fig. 2B). which contains five Ig-like domains (18, 35, 36). The alignment of Several mutations in region 1 showed similar ability to support Xenopus pIgR with the pIgR from human, possum, and chicken pIgA formation and SC binding as the wild-type J chain, suggest- and a sequence comparison of each Ig-like domain indicated that ing no direct involvement of this region in polymerization or pIgR X. laevis pIgR lacks an Ig domain equivalent to mammalian D2 interaction. For region 2, we found that substitution of amino acids (Fig. 3; Table III). Interestingly, a Pro residue within the trans- 50–52 (ISD) with Ala significantly reduced the affinity of pIgA to membrane helix present in mammalian and chicken pIgR was ab- SC. Because glycosylation of the phylogenetically conserved sent in Xenopus pIgR, suggesting that signaling through this re- Asn49 is directed by Ser51 the ISD mutant lacked the glycosylation ceptor may be different in amphibians. However, the strong of Asn49. However, the mutation of Asn49 to Ala (N49A) did not basolateral targeting signal present in all pIgRs was also conserved reduce pIgA affinity to SC, although a slight decrease in pIgA in Xenopus pIgR. formation was seen (Table I) similar to that previously observed The initial noncovalent interaction of pIgA and pentameric IgM with mutagenesis of mouse J chain (34). Furthermore, the I50A/ with pIgR is mediated by its extracellular D1 and the respective Ig D52A mutant showed reduced SC affinity, although glycosylation Fc portions, including the J chain (37, 38). We wanted to test would be maintained. Thus, glycosylation per se was apparently whether mammalian pIgA could bind amphibian pIgR and there- unimportant for the pIgR interaction, although an important motif fore constructed a chimeric SC molecule that contained D1 from for pIgR interaction was located in the highly conserved region 2 Xenopus and D2 through D5 from human pIgR (xD1-hD2D5). The of the human J chain. binding of this chimera to human pIgA or pentameric IgM or IgG A region between the sixth and seventh Cys of the J chain was was compared with that of human SC. The xD1-hD2D5 chimeric found to be highly conserved in all tetrapods examined, but not in SC showed distinct binding to pIgA but no detectable affinity to the nurse shark (Table II; Fig. 2A). Again, we made several J chain pentameric IgM (Fig. 4A). Conversely, human SC bound both mutations in this region and coexpressed the resulting J chain vari- pIgA and pentameric IgM. A human SC lacking D1 showed no 1594 CONSERVATION OF SECRETORY ANTIBODY FORMATION IN TETRAPODS Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 3. Alignment of pIgR from X. laevis with human, possum, and chicken pIgR. Residues identical in all sequences, highly conserved sequences, a colon (:), and a period (.), respectively. Gaps in the alignment are indicated by hyphens. The start of ,(ء) or conserved sequences are indicated by a star each extracellular domain, the cytoplasmic tail, and the cleavage site for human pIgR is indicated. The CDR1, CDR2, and CDR3 in D1 are written in bold font. The predicted signal peptide and transmembrane region are underlined. GenBank accession numbers: Homo sapiens, NP_002635; possum (Tricho- surus vulpecula), AAD41688; and chicken (Gallus gallus), AAW71994.

affinity to either ligand, demonstrating that human pIgA bound the pIgA composed of human ␣ H chains and the Xenopus J chain Xenopus-human chimera because of Xenopus D1. No SC variant would show better binding to Xenopus pIgR than pIgA containing bound IgG (data not shown). the human or mouse J chain. Therefore, we tested the binding of xD1-hD2D5 SC to pIgA containing the human, mouse, or Xenopus J Binding of human pIgA to Xenopus pIgR is improved by chain and did indeed find that pIgA with Xenopus J chain was the substituting human J chain with that of Xenopus better ligand for Xenopus pIgR D1 (Fig. 4B). Furthermore, IgA with- We hypothesized that the J chain contributes directly to the pIgR out J chain showed no binding to Xenopus pIgR D1, demonstrating J interaction site of pIgA. It was therefore reasonable to assume that chain dependence for pIgR-pIg interaction also in this species. The Journal of Immunology 1595

Table III. Sequence comparison of the extracellular Ig-like domains from pIgRa

Domain 1 Domain 2 Domain 3 (2) Domain 4 (3) Domain 5 (4)

HPCHPHPCHPCHPC

XD1 33 36 41 14 22 21 24 21 19 24 24 12 16 13 XD2 28 27 24 30 27 40 40 44 29 28 27 20 21 24 XD3 29 27 27 24 23 18 21 26 35 40 40 18 22 24 XD427221962711202821212247 43 46

a Percentage of amino acid sequence identity between the four extracellular Ig domains (xD1, xD2, xD3, and xD4) of Xenopus laevis pIgR (indicated vertically on the left) compared with each of the five (human and possum) or four (chicken) Ig-like domains of pIgR from human (H), possum (P), and chicken (C). The Ig domains of the human, possum, and chicken most similar to each of the Xenopus Ig domains are indicated in bold font. Chicken and Xenopus have no equivalent of mammalian domain 2.

Discussion skate, suggesting that it is not required for normal J-chain function Knockout mice for the J chain and pIgR have shown that both in Chondrichthyes (24). We therefore speculate that the highly are critical for SIg generation (9, 39). In this study, we conserved J chain region in tetrapods is required for pIgR inter- present a mutational analysis of human J chain supporting a direct action, thereby allowing active export of pIgs to mucosal surfaces. Downloaded from interaction between these two polypeptides. Furthermore, we iden- Because the pIgR interaction site is not conserved in J chain from tify the first amphibian pIgR and demonstrate that protein-protein cartilaginous fish, its function here might solely be to facilitate Ig interaction between pIgR and the J chain is remarkably conserved polymerization. Thus, Abs with higher avidity and better aggluti- between tetrapods. nating capacity would be produced. Dooley and Flajnik (40) have The J chain probably originated in the first jawed vertebrate proposed that absence of J chain expression in B cells that have because it is present in rays and sharks. Intriguingly, it appears to undergone affinity maturation would provide monomers with en- be deleted in bony fish; despite near completion of the genome hanced tissue penetration in a memory response. http://www.jimmunol.org/ project for several fish species, no resembling this unique Because the Xenopus J chain showed high homology to the hu- gene has been identified. Interestingly, we found that a region crit- man J chain in two regions that were required for pIgR binding, we ical for pIgA-pIgR interaction was highly divergent between the J predicted that a pIgR homologue would be found in this species. chains from tetrapods and the nurse shark. Furthermore, this C- Database searches identified such a gene and we deduced the pri- terminal region is absent from the J-chain sequence of clearnose mary sequence of the open reading frame by DNA sequencing of by guest on September 29, 2021

FIGURE 4. D1 of X. laevis pIgR/SC mediates pIgR binding to J chain-containing IgA. A, Recombinant human SC, a chimeric SC with X. laevis D1 fused to human D2 (xD1-hD2D5-SC) and human SC lacking D1 (hD2D5-SC), were tested for binding to J558L transfectoma-produced pIgA or pentameric IgM (mouse J chain and human Fc region). B, Human SC or xD1-hD2D5 SC were tested for binding to CHO cell-produced IgA without J chain (mock) or with J chain from different species in an ELISA. Mean Ϯ SD of quadruplicate (A) or triplicate (B) wells are shown. Each experiment shown is representative of three or more similar experiments. 1596 CONSERVATION OF SECRETORY ANTIBODY FORMATION IN TETRAPODS full-length cDNA. Although the Xenopus pIgR clearly has an over- gut may serve to curtail the commensal microbiota within the mu- all structure similar to that of pIgR from other species, including cosal compartment (7, 49). Thus, mucosal immune exclusion pro- four Ig domains (like chicken pIgR, whereas mammalian pIgR tects against pathogens and promotes luminal habitation of the contain five Ig domains), a transmembrane region, and a cytoplas- beneficial indigenous microbiota, thereby preventing adverse sys- mic tail with a basolateral targeting motif, it was surprisingly di- temic immune reactions. vergent in the critical ligand-binding CDR1-like region in D1 (37, In conclusion, our structural data imply that the J chain has 41): the sequence VNRH, conserved in all pIgRs published to date served two roles in the evolution of a secretory Ab system. Its (resequencing of rabbit pIgR has revealed this sequence in that Ig-joining capacity allowed regulation of polymerization to form species, although it was initially reported as VTRH) is substituted Abs with better agglutinating capabilities, while its interaction with ANKY in Xenopus pIgR (Fig. 3). Additionally, CDR1 is one with the pIgR allowed the export of such Abs to the mucosal amino acid longer in X. laevis than in other pIgRs published to date. surfaces where the bacterial load is high. The functional conser- However, structural modeling (Swiss Model; http://swissmodel. vation between each polypeptide chain specifically involved in the expasy.org/) predicted the CDR1 of Xenopus pIgR would form an ␣ generation of secretory Abs from amphibians, birds, and mammals helix characteristic of the pIgR D1 (Ref. 42 and data not shown). suggests that this important first-line defense mechanism has been Due to the relative divergence of CDR1, it was surprising that the maintained in tetrapod evolution. Xenopus pIgR D1 mediates fairly strong binding to mammalian pIgA. A possible explanation for its failure to bind mammalian Acknowledgments pentameric IgM is that this polymer adopts a different conforma- We thank Prof. Itaru Moro for providing chicken J chain cDNA, Prof. tion than Xenopus J chain-containing IgM, which reportedly is Andrew Macpherson and Prof. Martin Flajnik for critical reading of the Downloaded from hexameric (43). We have previously shown that differences in the manuscript, and Kathrine Hagelsteen and Linda Solfjell for technical as- CDR2 region between human and rabbit pIgR D1 are primarily sistance with ELISA and mutagenesis, respectively. responsible for the disparity of IgM binding between these two Disclosures species; only human pIgR binds pentameric IgM with high affinity The authors have no financial conflict of interest. (38). It is therefore possible that differences between the CDR2 regions in Xenopus and human pIgR contribute to the failure of References http://www.jimmunol.org/ Xenopus pIgR to bind mammalian pentameric IgM. 1. Macpherson, A. J., and N. L. Harris. 2004. Interactions between commensal An interesting difference between the mammalian pIgR on the intestinal bacteria and the immune system. Nat. Rev. Immunol. 4: 478–485. one hand and the pIgR from chicken and Xenopus on the other is 2. Backhed, F., R. E. Ley, J. L. Sonnenburg, D. A. Peterson, and J. I. Gordon. 2005. Host-bacterial mutualism in the human intestine. Science 307: 1915–1920. that the former contain five extracellular Ig-like domains while the 3. Rakoff-Nahoum, S., J. Paglino, F. Eslami-Varzaneh, S. Edberg, and R. Medzhitov. latter contain four, lacking an equivalent to mammalian D2. In 2004. Recognition of commensal microflora by Toll-like receptors is required for intestinal homeostasis. Cell 118: 229–241. mammalian pIgR, D2 and D3 are encoded by a single exon, 4. Cebra, J. J. 1999. Influences of microbiota on intestinal immune system devel- thereby diverging from the rule “one domain, one exon.” Sequence opment. Am. J. Clin. Nutr. 69: 1046S–1051S. comparisons of each individual Ig-like domain within and across 5. Stavnezer, J., and C. T. Amemiya. 2004. Evolution of isotype switching. Semin.

Immunol. 16: 257–275. by guest on September 29, 2021 mammalian and nonmammalian pIgRs showed that mammalian 6. Lamm, M. E. 1997. Interaction of antigens and at mucosal surfaces. D2 is most similar to mammalian D3, suggesting that it arose by Annu. Rev. Microbiol. 51: 311–340. a duplication of the exon encoding D3 (Table III and data not 7. Macpherson, A. J., and T. Uhr. 2004. Compartmentalization of the mucosal im- mune responses to commensal intestinal bacteria. Ann. NY Acad. Sci. 1029: shown). There is no remnant of such a duplication event in the 36–43. chicken pIgR gene (36) or in the pIgR gene of Xenopus tropicalis 8. Brandtzaeg, P., and H. Prydz. 1984. Direct evidence for an integrated function of J chain and secretory component in epithelial transport of immunoglobulins. Na- (data not shown), suggesting that his duplication event occurred ture 311: 71–73. after mammals diverged from birds and amphibians. 9. Hendrickson, B. A., D. A. Conner, D. J. Ladd, D. Kendall, J. E. Casanova, Amphibians express five Ig classes: IgM, IgY, and IgX (5) and B. Corthesy, E. E. Max, M. R. Neutra, C. E. Seidman, and J. G. Seidman. 1995. Altered hepatic transport of in mice lacking the J chain. the recently identified IgF and IgD (44). IgY exists as a monomer J. Exp. Med. 182: 1905–1911. and is the major serum Ig with functions similar to those of both 10. Mostov, K., and C. S. Kaetzel. 1999. Immunoglobulin transport and the poly- IgG and IgE (5). IgM and IgX constitute the polymeric Igs in meric immunoglobulin receptor. In Mucosal Immunology. P. L. Ogra, J. Mestecky, M. E. Lamm, W. Strober, J. Bienenstock, and J. R. McGhee, eds. Xenopus, but only IgM polymers incorporate a J chain (43, 45, 46). Academic Press, San Diego, p. 133. IgX has been suggested to be an analog to IgA because of its 11. Phalipon, A., and B. Corthesy. 2003. Novel functions of the polymeric Ig recep- tor: well beyond transport of immunoglobulins. Trends Immunol. 24: 55–58. ability to form polymers and its expression chiefly in the gut (46). 12. Brandtzaeg, P., and F. E. Johansen. 2005. Mucosal B cells: phenotypic charac- It is an intriguing possibility that IgX could be a J chain-indepen- teristics, transcriptional regulation, and homing properties. Immunol. Rev. 206: dent ligand for this receptor. We found that Xenopus pIgR only 32–63. 13. Johansen, F. E., I. Natvig Norderhaug, M. Roe, I. Sandlie, and P. Brandtzaeg. bound J chain-containing pIgA and that the affinity improved when 1999. Recombinant expression of polymeric IgA: incorporation of J chain and this pIgR contained the Xenopus J chain compared with the human secretory component of human origin. Eur. J. Immunol. 29: 1701–1708. or mouse J chain, suggesting a J chain dependence for pIg-pIgR 14. Max, E. E., and S. J. Korsmeyer. 1985. Human J chain gene. Structure and expression in B lymphoid cells. J. Exp. Med. 161: 832–849. interaction also in this species. Importantly, our results demon- 15. Matsuuchi, L., G. M. Cann, and M. E. Koshland. 1986. Immunoglobulin J chain strated that the interaction site between pIgR and J chain-contain- gene from the mouse. Proc. Natl. Acad. Sci. USA 83: 456–460. 16. Kulseth, M. A., and S. Rogne. 1994. Cloning and characterization of the bovine ing pIgs is conserved between mammals and amphibians. immunoglobulin J chain cDNA and its promoter region. DNA Cell Biol. 13: In humans, SIgA is the dominant Ab on a biosynthetic basis and 37–42. reportedly protects against Vibrio cholerae and enterotoxigenic 17. Hughes, G. J., S. Frutiger, N. Paquet, and J. C. Jaton. 1990. The amino acid sequence of rabbit J chain in secretory immunoglobulin A. Biochem. J. 271: Escherichia coli-induced diarrhea (47), in agreement with the no- 641–647. tion that such Abs mediate mucosal immune exclusion (12). We 18. Adamski, F. M., and J. Demmer. 1999. Two stages of increased IgA transfer have previously generated pIgR knockout mice and have recently during lactation in the marsupial, trichosurus vulpecula (Brushtail possum). J. Immunol. 162: 6009–6015. shown that SIgs protect against horizontal spread of the mucosal 19. Takahashi, T., T. Iwase, T. Tachibana, K. Komiyama, K. Kobayashi, C. L. Chen, pathogen Salmonella typhimurium even in naive animals, thereby J. Mestecky, and I. Moro. 2000. Cloning and expression of the chicken immu- noglobulin joining (J)-chain cDNA. Immunogenetics 51: 85–91. providing herd immunity (39, 48). In addition, recent data have 20. Iwata, A., T. Iwase, Y. Ogura, T. Takahashi, N. Matsumoto, T. Yoshida, N. suggested that the large amounts of SIgA generated in the mouse Kamei, K. Kobayashi, J. Mestecky, and I. Moro. 2002. Cloning and expression The Journal of Immunology 1597

of the turtle (Trachemys scripta) immunoglobulin joining (J)-chain cDNA. Im- regions of the molecule are conserved among five mammalian species. J. Immu- munogenetics 54: 513–519. nol. 154: 1735–1747. 21. Hohman, V. S., S. E. Stewart, C. E. Willett, and L. A. Steiner. 1997. Sequence 36. Wieland, W. H., D. Orzaez, A. Lammers, H. K. Parmentier, M. W. Verstegen, and expression pattern of J chain in the amphibian, Xenopus laevis. Mol. Immu- and A. Schots. 2004. A functional polymeric immunoglobulin receptor in chicken nol. 34: 995–1002. (Gallus gallus) indicates ancient role of secretory IgA in mucosal immunity. 22. Hohman, V. S., and L. A. Steiner. 2000. The sequence of J chain in an amphibian, Biochem. J. 380: 669–676. Rana catesbeiana. Immunogenetics 51: 587–590. 37. Bakos, M. A., A. Kurosky, and R. M. Goldblum. 1991. Characterization of a 23. Hohman, V. S., S. E. Stewart, L. L. Rumfelt, A. S. Greenberg, D. W. Avila, critical binding site for human polymeric Ig on secretory component. J. Immunol. M. F. Flajnik, and L. A. Steiner. 2003. J chain in the nurse shark: implications for 147: 3419–3426. function in a lower vertebrate. J. Immunol. 170: 6016–6023. 38. Roe, M., I. N. Norderhaug, P. Brandtzaeg, and F. E. Johansen. 1999. Fine spec- 24. Cannon, J. P., R. N. Haire, and G. W. Litman. 2002. Identification of diversified ificity of ligand-binding domain 1 in the polymeric Ig receptor: importance of the genes that contain immunoglobulin-like variable regions in a protochordate. Nat. CDR2-containing region for IgM interaction. J. Immunol. 162: 6046–6052. Immunol. 3: 1200–1207. 25. Johansen, F. E., R. Braathen, and P. Brandtzaeg. 2000. Role of J chain in secre- 39. Johansen, F. E., M. Pekna, I. N. Norderhaug, B. Haneberg, M. A. Hietala, tory immunoglobulin formation. Scand. J. Immunol. 52: 240–248. P. Krajci, C. Betsholtz, and P. Brandtzaeg. 1999. Absence of epithelial immu- 26. Brandtzaeg, P. 1974. Presence of J chain in human immunocytes containing noglobulin A transport, with increased mucosal leakiness, in polymeric immu- J. Exp. Med. various immunoglobulin classes. Nature 252: 418–420. noglobulin receptor/secretory component-deficient mice. 190: 27. Johansen, F. E., R. Braathen, and P. Brandtzaeg. 2001. The J chain is essential for 915–922. polymeric Ig receptor-mediated epithelial transport of IgA. J. Immunol. 167: 40. Dooley, H., and M. F. Flajnik. 2005. Shark immunity bites back: affinity matu- 5185–5192. ration and memory response in the nurse shark, Ginglymostoma cirratum. Eur. 28. Takahashi, T., M. Kimura, N. Matsumoto, A. Iwata, Y. Ogura, T. Yoshida, J. Immunol. 35: 936–945. N. Kamei, K. Komiyama, J. Mestecky, and I. Moro. 2002. Cloning of the chicken 41. Coyne, R. S., M. Siebrecht, M. C. Peitsch, and J. E. Casanova. 1994. Mutational immunoglobulin joining (J)-chain gene and characterization of its promoter re- analysis of polymeric immunoglobulin receptor/ligand interactions. Evidence for gion. DNA Cell Biol. 21: 81–90. the involvement of multiple complementarity determining region (CDR)-like 29. Braathen, R., V. Sorensen, P. Brandtzaeg, I. Sandlie, and F. E. Johansen. 2002. loops in receptor domain I. J. Biol. Chem. 269: 31620–31625. The carboxyl-terminal domains of IgA and IgM direct isotype-specific polymer- 42. Hamburger, A. E., A. P. West, Jr., and P. J. Bjorkman. 2004. Crystal structure of Downloaded from ization and interaction with the polymeric immunoglobulin receptor. J. Biol. a polymeric immunoglobulin binding fragment of the human polymeric immu- Chem. 277: 42755–42762. noglobulin receptor. Structure (Camb.) 12: 1925–1935. 30. Claypool, S. M., B. L. Dickinson, J. S. Wagner, F. E. Johansen, N. Venu, 43. Hadji-Azimi, I., and M. Michea-Hamzehpour. 1976. Xenopus laevis 19S immu- J. A. Borawski, W. I. Lencer, and R. S. Blumberg. 2004. Bidirectional transep- noglobulin. Ultrastructure and J chain isolation. Immunology 30: 587–591. ␥ ithelial IgG transport by a strongly polarized basolateral membrane Fc -receptor. 44. Zhao, Y., Q. Pan-Hammarstrom, S. Yu, N. Wertz, X. Zhang, N. Li, J. E. Butler, Mol. Biol. Cell 15: 1746–1759. and L. Hammarstrom. 2006. Identification of IgF, a hinge-region-containing Ig 31. Elm, C., R. Braathen, S. Bergmann, R. Frank, J. P. Vaerman, C. S. Kaetzel, class, and IgD in Xenopus tropicalis. Proc. Natl. Acad. Sci. USA 103:

G. S. Chhatwal, F. E. Johansen, and S. Hammerschmidt. 2004. Ectodomains 3 12087–12092. http://www.jimmunol.org/ and 4 of human polymeric Immunoglobulin receptor (hpIgR) mediate invasion of 45. Hsu, E., M. F. Flajnik, and L. Du Pasquier. 1985. A third immunoglobulin class Streptococcus pneumoniae into the epithelium. J. Biol. Chem. 279: 6296–6304. in amphibians. J. Immunol. 135: 1998–2004. 32. Brandtzaeg, P. 1975. Purification of J chain after mild reduction of human im- Xenopus munoglobulins. Scand. J. Immunol. 4: 309–320. 46. Mussmann, R., L. Du Pasquier, and E. Hsu. 1996. Is IgX an analog of Eur. J. Immunol. 33. Brandtzaeg, P. 1974. Human secretory component. I. Purification of free secre- IgA? 26: 2823–2830. tory component from colostrum. Scand. J. Immunol. 3: 579–588. 47. Sanchez, J., and J. Holmgren. 2005. Virulence factors, pathogenesis and vaccine 34. Krugmann, S., R. J. Pleass, J. D. Atkin, and J. M. Woof. 1997. Structural re- protection in cholera and ETEC diarrhea. Curr. Opin. Immunol. 17: 388–398. quirements for assembly of dimeric IgA probed by site-directed mutagenesis of 48. Wijburg, O. L., T. K. Uren, K. Simpfendorfer, F. E. Johansen, P. Brandtzaeg, and J chain and a cysteine residue of the ␣-chain CH2 domain. J. Immunol. 159: R. A. Strugnell. 2006. Innate secretory antibodies protect against natural Salmo- 244–249. nella typhimurium infection. J. Exp. Med. 203: 21–26. 35. Piskurich, J. F., M. H. Blanchard, K. R. Youngman, J. A. France, and C. S. 49. Macpherson, A. J., and T. Uhr. 2004. Induction of protective IgA by intestinal

Kaetzel. 1995. Molecular cloning of the mouse polymeric Ig receptor. Functional dendritic cells carrying commensal bacteria. Science 303: 1662–1665. by guest on September 29, 2021