The J chain is essential for polymeric Ig receptor-mediated epithelial transport of IgA.

Local production of secretory (S)IgA provides adaptive immunologic protection of mucosal surfaces, but SIgA is also protective when administered passively, such as in breast milk. Therefore, SIgA is a potential candidate for therapeutic administration, but its complex structure with four different polypeptide chains produced by two distinct cell types complicates recombinant production. The J chain is critical in the structure of SIgA because it is required for efficient polymerization of IgA and for the affinity of such polymers to the secretory component (SC)/polymeric (p)IgR. To better understand the role of the J chain in SIgA production, we have generated various mutant forms of the human J chain and analyzed the function of these mutants when coexpressed with IgA. We found that the C terminus of the J chain was not required for the formation of IgA polymers, but was essential for the binding of pIgA to SC. Likewise, we found that two of the intrachain disulfide bridges (Cys(13):Cys(101) and Cys(109):Cys(134)) were also required for the binding of pIgA to SC but, interestingly, not for IgA polymerization. Conversely, the last intrachain disulfide bridge (Cys(72):Cys(92)) was not essential for either of these two J chain functions. Finally, we demonstrated that the presence of only Cys(15) or Cys(69) was sufficient to support polymerization of IgA, but that these polymers were mostly noncovalently stabilized. Nevertheless, these polymers bound free SC with nearly the same affinity as pIgA containing wild-type J chain, but were transcytosed by pIgR-expressing polarized epithelial cells at a reduced efficiency.     

Secretory component: interactions with intracellular and surface immunoglobulins of human lymphoid cells.

Unstimulated and PWM-stimulated lymphocytes from normal human peripheral blood, cord blood, peripheral blood of patients with panhypogammaglobulinemia and selective IgA deficiency, as well as human lymphoblastoid cell lines were examined for their ability to bind secretory component (SC) on the surface and in the cytoplasm. SC binding was not detected on the cell surface at any stage of differentiation in these cells. However, binding of SC was detected in the cytoplasm of 2.3% of normal peripheral blood lymphocytes cultured in the presence of PWM for 6 to 7 days, and in two IgA producing lymphoblastoid cell lines. The capability of lymphoid cells to bind SC was not concurrent with J chain production. Although IgA was detected in the cytoplasm of PWM-stimulated lymphocytes from IgA-deficient patients, these cells did not bind SC. The failure to detect surface receptors indicates that SC is not a probable factor determining the homing of IgA precursor cells into exocrine tissues.     

Cellular origins of human polymeric and monomeric IgA: intracellular and secreted forms of IgA

Intracellular and secreted IgA from pokeweed mitogen (PWM)-stimulated normal peripheral blood lymphocytes, from 12-O-tetradecanoylphorbol-13-acetate (TPA)-stimulated peripheral blood lymphocytes of a patient with chronic lymphocytic leukemia (CLL), or from an IgA-producing human Epstein Barr virus (EBV)-transformed lymphoblastoid cell line were analyzed by molecular-sieve chromatography, electrophoresis in sodium dodecyl sulfate, and sucrose density ultracentrifugation. Fluorochrome-labeled anti-human IgA and secretory component (SC) were used as probes for the detection of polymeric IgA in individual cells. These methods demonstrated that the majority of intracellular IgA occurred in monomeric form, even when the predominant form of secreted IgA was polymeric. Sequential analyses of the IgA secreted by PWM-stimulated normal peripheral blood lymphocytes revealed that the proportion of polymeric IgA increased with the time of culture and that polymers represented the prevalent form of secreted IgA from the fifth day of culture. Although approximately one-half of TPA-stimulated CLL cells bound fluorochrome-labeled SC, only trace amounts of extracellular and intracellular polymeric IgA were detected in both culture supernatants and lysates. Culture supernatants of an IgA-secreting EBV-transformed cell line contained predominantly polymeric IgA. However, intracellular IgA was largely represented by monomers. The predominance of intracellular monomers in polymeric IgA-secreting cells suggested that the pathway of the assembly of human IgA molecules is analogous to that described for mouse IgA synthesis.     

Biosynthesis of immunoglobulin A (IgA) and immunoglobulin M (IgM). Control of polymerization by J chain

Cell suspensions of mouse plasma-cell tumours secreting IgA (immunoglobulin A) and IgM (immunoglobulin M) were incubated with radioactive leucine for various periods of time. The secreted immunoglobulins were precipitated from the culture medium with specific rabbit antisera to determine the relative distribution of radioactivity among the different molecular species, and to estimate the fraction of total radioactivity in the J chain. For IgM-secreting cells there is a balanced synthesis of 7S subunits and J chains, and the secreted product is uniformly assembled to the pentamer. In cells secreting IgA, however, the results demonstrate that the pool of intracellular J chain is less than the intracellular IgA pool. The concentration of J chain is therefore limiting and is less than the requirement for complete polymerization. The major factor that determines whether an intracellular monomer is secreted as such or is polymerized with the addition of J chain is therefore the amount of intracellular J chain. When this is limiting, as it is in cells secreting IgA, then monomer will be secreted.     

Ig isotypes produced by EBV-transformed B cells as a function of age and tissue distribution

EBV can transform human B cells giving rise to lymphoblastoid cell lines that produce and secrete Ig. Herein B cells from various tissues of newborns and adults were transformed by EBV and their Ig products were analyzed with isotype-specific mAb. Although IgG- and IgA-bearing B cells were present in the newborn, EBV transformed IgM-producing cells almost exclusively in both newborn blood and breast milk. IgM-secreting cells were derived from IgM+ B cells and IgM- pre-B cells present in neonatal blood, but only from IgM+ cells in adult blood. Whereas in adults most EBV-transformed cells produced IgM, producers of IgG and of IgA were present in frequencies that varied according to the tissue source. Precursors of IgG-producing cells were relatively abundant in blood, spleen, and tonsil, and relatively infrequent in bone marrow and appendix. EBV-inducible IgA producers were relatively concentrated in the appendix and to a lesser extent in tonsils and blood. Differences in the subclass composition of EBV-transformed populations of IgG- and IgA-producers were also observed for the various adult lymphoid tissues. IgG1-producing cells predominated in most tissues, and precursors of IgG2 were largely confined to the circulation. Whereas IgA1-producing cells were predominant in all tissues, a marked enrichment in IgA2-producers was observed in the appendix. These results indicate a remarkable heterogeneity in the isotype distribution pattern of EBV-transformable B cells that is determined both by developmental age and tissue localization. We propose that EBV selectively transforms primed B cells, the isotype commitment of which varies according to tissue origin and age.     

Effect of a disulfide-interchange enzyme on the assembly of human secretory immunoglobulin A from immunoglobulin A and free secretory component.

A disulfide-interchange enzyme from rat liver microsomes was found to promote binding in vitro of human free secretory component (SC) to dimeric serum-type IgA containing J chain, as assessed by immune precipitation and gel filtration. This effect was greater withe native than with partially reduced SC. Most of the bound SC was covalently linked, as determined by electrophoresis in polyacrylamide gels in detergent. The enzyme did not promote binding of native or partially reduce SC to IgG, IgA monomer, IgA dimer without J chain, or IgM. In the case of IgM, the enzyme did, however, promote covalent bonding of previously non-covalently linked SC. The results overall suggest that a disulfide-interchange enzyme could play a role in vivo in the cell-associated assembly of secretory IgA by promoting the covalent attachment of SC to a dimer of serum-type IgA and that the J chain in the IgA dimer contributes to the enzyme effect.     

Rat secretory component binds poorly to rodent IgM.

Our previous studies and those of others indicated that human secretory component (SC), the five domain extracellular portion of the poly Ig receptor, binds avidly to both pIgA and IgM. In this study we report that in rodents, SC binds primarily to pIgA. Rat secretory component was isolated from bile and radiolabeled to known specific activity with 125I. Radiolabeled rat SC was incubated with rat and mouse monoclonal proteins for 1 h at room temperature and overnight at 4 degrees D. Binding of 125I-rat SC to Ig was determined in two ways: 1) immunoprecipitation of putative 125I-rat SC-Ig complexes with anti-L chain antibodies; 2) HPLC gel filtration on an analytical TSK 4000 column that separated free 125I-rat SC from 125I-rat SC bound to Ig. Both methods of analysis yielded similar results. Rat and mouse polymeric (p) IgA bound rat SC with high avidity, although the binding activity of the IgM from either species was virtually nil. The number of SC-binding sites on rat polymeric Ig was determined by immunoprecipitation of mixtures of rat pIg with saturating concentrations of 125I-rat SC and yielded values of 1.0 and 0.05 for rat pIgA and IgM, respectively. The significance of these findings with respect to the biologic function of the pIg R in rodents and the nature of the pIg R-binding site on pIg is discussed.     

J-chain expression in human cells producing IgG subclasses

Previous studies have demonstrated that J chain is expressed not only in cells that produce polymeric immunoglobulins, but also in those engaged in synthesis of monomers including IgG and IgD. The presence of J chain in these cells suggested that its role may not be restricted to the formation of polymers. For the present study, fluorochrome-labeled polyclonal anti-J-chain and monoclonal antibodies to IgG subclasses were used to determine the distribution of J chain in IgG plasma cells from normal human tissues and from pokeweed mitogen (PWM)-stimulated human peripheral blood lymphocytes. The results indicate that J chain is not equally distributed among cells producing different IgG subclasses. The percentages of PWM-stimulated cells containing J chain were: 22 +/- 5 (SE) for IgG1, 49 +/- 6 for IgG2, 17 +/- 7 for IgG3, and 64 +/- 11 for IgG4. Examination of sections of various human lymphoid tissues revealed that the frequency of IgG cells that coexpressed J chain was lower than that observed in the PWM system and displayed variable distribution among IgG subclasses. The frequency of J-chain expression in IgG-producing cells may be related to the degree of cellular maturation and may differ according to the origin of cells.     

J chain is covalently bound to both monomer subunits in human secretory IgA

Previous work has established that the secretory component (SC) in human secretory IgA is covalently linked to only one of the two IgA monomer subunits, but it has not been clear whether the J chain is covalently linked to one or to both of these subunits. In view of the asymmetry in the disulfide bonding between SC and the IgA subunits, an arrangement which follows disulfide interchange, several models for the disulfide linkage of J chain and the bonds between IgA subunits were envisaged and investigated. When sIgA was gel filtered through Sephadex G-200 in acetic acid, a single major symmetrical peak eluted at the front. This material contained SC, alpha and L chains, and all of the J chain. The greater resolution afforded by polyacrylamide gel electrophoresis in detergent confirmed that human sIgA contains no major noncovalently linked components in the 150,000-200,000 molecular weight range. In another series of experiments the Fc monomer, which is not covalently attached to SC, isolated after treatment of sIgA with IgA protease and cyanogen bromide, was investigated to learn whether alpha chain COOH-terminal octapeptides could be released by reduction. The results were negative. The available data thus favor a model in which J chain is disulfide-bonded to both IgA monomer subunits in sIgA.     

Enhancement and suppression of immunoglobulin G-producing B cells in the presence of immune T cells.

Mice were immunized one to three times with sheep red blood cells. Four to seven days after the last immunization, the spleens were removed and the cells were cultured in vitro in the absence of antigen. Removal of most T cells by anti-θ serum treatment prior to culture could increase the number of IgG-producing B cells without affecting the number of specific or nonspecific IgM-producing B cells detected after 2 days of culture. Addition of graded numbers of immune cells to pure immune B cells enhanced the number of IgG-producing B cells, whereas addition or higher number of immune cells caused suppression. Since removal of T cells could also enhance the proliferation of IgG-producing B cells induced by lipopolysaccharide (LPS), a polyclonal B-cell activator, it is suggested that the suppressive effects of high numbers of immune T cells are exerted directly on the B cells.     

Binding sites for measles virus antigens on human B and T lymphocytes

The present study examined the differences in the binding of measles virus antigens to human peripheral blood lymphocyte (PBL) subpopulations. PBL binding sites for measles antigens were detected by an assay involving the rosetting of PBL to measles-infected HeLa cells (HeLa-K11). Three approaches were employed to examine whether measles virus antigen binding sites were present on restricted subpopulations of PBL. First, no significant difference in the proportion of HeLa-K11 forming clusters was observed with unfractionated cells in comparison with enriched B- or T-lymphocyte suspensions. Second, the profile of lymphocyte surface markers before and after adherence of PBL suspensions to HeLa-K11 cells was measured. No difference in the proportion of PBL forming E-rosettes or lymphocytes with Fc-IgG receptors, surface immunoglobulin, or complement receptors was observed. Finally, the percentage of B (Raji, B-35M, Bristow-7B) and T (Molt-3) cell human lymphoid cells which adhered to HeLa-K11 versus noninfected HeLa cells was compared. In all cases, a highly significant adherence of the lymphoid cell suspensions to HeLa-K11 cells was observed in comparison with uninfected HeLa cells. This is the first direct demonstration of binding sites for measles virus antigens present on both human B and T lymphocytes.     

Synergistic effect of IL-4 and IFN-gamma on the expression of polymeric Ig receptor (secretory component) and IgA binding by human epithelial cells

The expression of secretory component (SC), the epithelial receptor for polymeric Ig, was enhanced by the addition of human rIFN-gamma or rIL-4, as revealed by the binding of radiolabeled polymeric, J chain-containing IgA or anti-SC antisera to the human colonic adenocarcinoma epithelial cell line HT-29. In combination, these cytokines exhibited a synergistic effect, and the potentiating effect of IL-4 was inhibitable by polyclonal anti-IL-4 antisera. Because the binding of radiolabeled polymeric IgA (pIgA) to HT-29 cells was inhibited by unlabeled pIgA or a polyclonal anti-SC reagent, but not by IgG, monomeric IgA, or Fab alpha fragments, we conclude that the receptor involved in the increased binding of pIgA is indeed SC. These data suggest that the expression of SC on human epithelial cells and the subsequent binding of pIgA (produced in mucosal tissues and glands by subepithelial plasma cells) is regulated by lymphokines such as IL-4 and IFN-gamma that are presumably derived from T cells found in abundant numbers in these tissues. These findings demonstrate a novel pathway of interaction between T cell products and epithelial cells that may result in enhanced translocation of large amounts of locally produced pIgA through epithelial cells into external secretions.     

Phenotype, frequency, and EBV responsiveness of human marrow B and pre-B cells

We have purified subpopulations of B lineage cells from human adult (rib) bone marrow by cell sorting and panning. Limiting dilution analysis was then used for a clonal analysis of cells able to secrete IgG, IgA, or IgM spontaneously or after infection with EBV. Nonproliferating, high rate IgG or IgA producers occurred at frequencies of about one per 1000 marrow mononuclear cells. Their frequency and Ig production was unaffected by EBV, and they appeared not to express EBNA after exposure to EBV. These cells were Ia+, B1+, and over 85% expressed sIg of the IgM/D (up to 75%) and/or IgG/A isotypes (40 to 60%). B cells committed to the secretion of IgM represent 2 to 10% of marrow B lymphocytes. They were found to be Ia+/B1+/B2+/CALLA- and C3b receptor (CR3)-cells, and most (greater than 90%) required infection with EBV and proliferation to develop into IgM-producing lymphocytes. Thirty to 40% of these cells did not express Ig (H or L chain) on their surface, and therefore resembled pre-B cells at the beginning of the 4- to 5-wk culture period. Proliferating pre-B cells from adult human marrow have been described, but their conversion into IgM-producing cells has not been formally demonstrated. Although EBV induces IgM production, the expression of EBNA, and several rounds of cell division in these cells, the induction of stable (greater than 5 wk) growth transformation represents a rare event in these pre-B cells: in several thousand limiting dilution wells, not a single culture of sIg-cells showed stable growth transformation. The dichotomy between EBV-induced high-rate IgM responses and absent growth transformation discriminates activation and transformation as distinct aspects of EBV-induced B cell "responses", and suggests that cellular properties play critical roles for viral transformation. We propose a model in which cellular target genes for transforming sequences in the EBV genome are transiently expressed during B cell differentiation.     

Heterogeneity of EBV-transformable human B lymphocyte populations

Although most human B cells express receptors for Epstein Barr virus (EBV), few (usually less than 1%) are readily transformed into B lymphoblastoid cell lines after exposure to EBV. Transformable cells previously have been found to be mostly resting B lymphocytes. We recently developed a limiting dilution culture system which permits the growth of EBV-transformed B lymphocytes with high efficiency. Because in this system up to over 30% of peripheral blood- or tonsil-derived B cells respond to EBV, we re-examined the properties of EBV-transformable cells. Frequencies of transformable lymphocytes were determined by Poisson analysis. EBV-susceptible B cells committed to IgM, IgG, or IgA secretion were found to occur in the range of 3 to 27, 0.1 to 6, and 0.1 to 5 per 100 B cells, respectively. Under our culture conditions, a major proportion of the IgM-committed cells derived from large lymphocytes which appeared to have entered the cell cycle. This population contains most of the EBV-responsive cells detected and, therefore, most of the additional cells responding in our culture system. In contrast, precursors of IgG- or IgA-producing lymphoblast lines were small cells with DNA contents typical for the G0/G1 phases of the cell cycle. Anti-immunoglobulin antibodies were used in panning experiments to separate B cell subpopulations which expressed different immunoglobulin isotypes on their surface. In limiting dilution cultures of these purified B lymphocytes subsets, it was found that virtually all precursors of IgM-producing cell lines expressed surface IgM (sIgM) before their infection and transformation by EBV. The "cloning efficiency" of positively selected, large sIgM+ cells approached 100%. In contrast, sIgG or sIgA were found only on cells committed to the production of IgG or IgA, respectively. The expression of sIgD was examined by using sequential panning procedures. Virtually all IgM-committed lymphocytes and a subset of cells committed to IgA secretion were found among the sIgD+ cells. The majority of cells committed to IgA production and all IgG-committed cells were found in the sIgD- B cell population. Our findings indicate that the EBV-susceptible B cell subset of normal lymphocytes is heterogeneous with respect to cell size, cell cycle, sIg determinants, and efficiency of transformation. On the basis of our findings and previous reports, we propose a model in which transformability is a B cell-inherent property. Factors unrelated to the virus but present in our culture system appear responsible for the enhanced vulnerability to viral transformation in some cells which entered into the cell cycle.(ABSTRACT TRUNCATED AT 400 WORDS)     

Human IgA as a heterovalent ligand: switching from the asialoglycoprotein receptor to secretory component during transport across the rat hepatocyte

Asialoglycoproteins are taken up by the rat liver for degradation; rat polymeric IgA is taken up via a separate receptor, secretory component (SC), for quantitative delivery to bile. There is negligible uptake of these ligands by the converse receptor, and only a low level of missorting of ligands to opposite destinations. The two pathways are not cross-inhibitable and operate independently (Schiff, J.M., M. M. Fisher, and B. J. Underdown, 1984, J. Cell Biol., 98:79-89). We report here that when human IgA is presented as a ligand in the rat, it is processed using elements of both pathways. To study this in detail, different IgA fractions were prepared using two radiolabeling methods that provide separate probes for degradation or re-secretion. Behavior of intravenously injected human polymeric IgA in the rat depended on its binding properties. If deprived of SC binding activity by affinity adsorption or by reduction and alkylation, greater than 80% of human IgA was degraded in hepatic lysosomes; radioactive catabolites were released into bile by a leupeptin-inhibitable process. If prevented from binding to the asialoglycoprotein receptor by competition or by treatment with galactose oxidase, human IgA was cleared and transported to bile directly via SC, but its uptake was about fivefold slower than rat IgA. Untreated human IgA was taken up rapidly by the asialoglycoprotein receptor, but depended on SC binding to get to bile: the proportion secreted correlated 1:1 with SC binding activity determined in vitro, and the IgA was released into bile with SC still attached. These results demonstrate that human IgA is normally heterovalent: it is first captured from blood by the asialoglycoprotein receptor, but escapes the usual fate of asialoglycoproteins by switching to SC during transport. Since the biliary transit times of native human and rat IgA are the same, it is probable that the receptor switching event occurs en route. This implies that the two receptors briefly share a common intracellular compartment.     

Assembly, secretion, and vacuolar delivery of a hybrid immunoglobulin in plants

Secretory immunoglobulin (Ig) A is a decameric Ig composed of four alpha-heavy chains, four light chains, a joining (J) chain, and a secretory component (SC). The heavy and light chains form two tetrameric Ig molecules that are joined by the J chain and associate with the SC. Expression of a secretory monoclonal antibody in tobacco (Nicotiana tabacum) has been described: this molecule (secretory IgA/G [SIgA/G]) was modified by having a hybrid heavy chain sequence consisting of IgG gamma-chain domains linked to constant region domains of an IgA alpha-chain. In tobacco, about 70% of the protein assembles to its final, decameric structure. We show here that SIgA/G assembly and secretion are slow, with only approximately 10% of the newly synthesized molecules being secreted after 24 h and the bulk probably remaining in the endoplasmic reticulum. In addition, a proportion of SIgA/G is delivered to the vacuole as at least partially assembled molecules by a process that is blocked by the membrane traffic inhibitor brefeldin A. Neither the SC nor the J chain are responsible for vacuolar delivery, because IgA/G tetramers have the same fate. The parent IgG tetrameric molecule, containing wild-type gamma-heavy chains, is instead secreted rapidly and efficiently. This strongly suggests that intracellular retention and vacuolar delivery of IgA/G is due to the alpha-domains present in the hybrid alpha/gamma-heavy chains and indicates that the plant secretory system may partially deliver to the vacuole recombinant proteins expected to be secreted.     

Biosynthesis of immunoglobulin A (IgA) and immunoglobulin M (IgM). Requirement for J chain and a disulphide-exchanging enzyme for polymerization

Mouse myeloma cells secreting 19S IgM (immunoglobulin M) (MOPC 104E and TEPC 183) or monomer and polymer IgA (immunoglobulin A) (MOPC 315) were incubated with radioactive leucine and the intracellular and secreted immunoglobulins and immunoglobulin subunits were prepared by preparative sucrose-density-gradient centrifugation. Samples were reduced in the presence or absence of isolated J chain, passed over Sephadex G-25 and then incubated at 37 degrees C for 30min with or without a source of disulphide-interchange enzyme. The extent of reassembly of reduced subunits was then evaluated by electrophoresis in polyacrylamide gels. Provided that J chain and the disulphide-interchange enzyme were supplied, both IgM and IgA could be assembled from their respective subunits, obtained by reductive cleavage of polymeric forms. Under similar conditions, assembly of polymeric forms from intracellular or secreted 7S monomer subunits also occurred. Under these conditions polymerization was total, there being no residue of the monomeric form. Reassembly did not occur in the absence of either J chain or the enzyme. All of the J chain released from IgM by reductive cleavage was incorporated back into the reassembled polymer. The J chain is therefore likely to be an essential structural requirement for polymeric immunoglobulins. A variety of controls ruled out non-specific interactions, and further suggested that the amino acid sequence of polypeptide chains determines the specificity of polymerization. The fact that intracellular IgA and IgM monomer subunits known to be deficient in galactose and fucose can be completely polymerized suggests that the addition of carbohydrate does not control polymerization.     

Immunohistochemical studies on various aspects of glandular immunoglobulin transport in man

Summary Gland-associated immunocyte populations have been characterized in human tissue specimens from which extracellular immunoglobulins have been removed by saline extraction. There is a striking preponderance of IgA-producing immunocytes adjacent to glands of the gastro-intestinal and respiratory tracts, in minor and major salivary glands, and in lactating mammary glands. Immunohistochemically, these cells have been found to contain dimeric IgA with incorporated J chain. Despite this local IgA production, immunohistochemical tests on alcohol-fixed specimens demonstrate that the glandular stroma is normally permeated predominantly by IgG, most of which is obviously serum-derived. However, the serous glandular cells selectively transmit dimeric IgA, which appears along their lateral borders and apically in the cytoplasm, whereas the epithelial occurrence of IgG is less conspicuous and is restricted to the interstices.The same epithelial cells produce a glycoprotein called the secretory component (SC) which exhibits specific affinity for J chain-containing dimeric IgA and pentameric IgM. In saline-extracted tissue, IgA, but IgG, is retained regularly along the borders of SC-producing cells; this probably reflects complexing between locally formed IgA and SC in the epithelial cell membranes. SC apparently functions as a glandular receptor for dimeric IgA which thus most likely enters the epithelial cells by adsorptive pinocytosis. After covalent stabilization, the IgA-SC complexes are extruded to the gland lumen. Immunohistochemically the Golgi zone has been found to contain free SC but no IgA, whereas SC occurring more apically in the epithelial cell exhibits characteristics of being IgA-associated. Pentameric IgM is handled by the glands in a way similar to dimeric IgA, but local synthesis of IgM is normally negligible, except in the gut.     

Structural requirements for polymeric immunoglobulin assembly and association with J chain

Both IgM and IgA exist as polymeric immunoglobulins. IgM is assembled into pentamers with J chain and hexamers lacking J chain. In contrast, polymeric IgA exists mostly as dimers with J chain. Both IgM and IgA possess an 18-amino acid extension of the C terminus (the tail-piece (tp)) that participates in polymerization through a penultimate cysteine residue. The IgM (mutp) and IgA (alphatp) tail-pieces differ at seven amino acid positions. However, the tail-pieces by themselves do not determine the extent of polymerization. We now show that the restriction of polymerization to dimers requires both C(alpha)3 and alphatp and that more efficient dimer assembly occurs when C(alpha)2 is also present; the dimers contain J chain. Formation of pentamers containing J chain requires C(mu)3, C(mu)4, and the mutp. IgM-alphatp is present mainly as hexamers lacking J chain, and mumugammamu-utp forms tetramers and hexamers lacking J chain, whereas IgA-mutp is present as high order polymers containing J chain. In addition, there is heterogeneous processing of the N-linked carbohydrate on IgA-mutp, with some remaining in the high mannose state. These data suggest that in addition to the tail-piece, structural motifs in the constant region domains are critical for polymer assembly and J chain incorporation.