Structural and Functional Analysis of J Chain-Deficient IgM

Cell lines

Cell lines are based on the Sp6 hybridoma cell line, which secretes IgM specific for DNP and TNP³ (21). All J⁺ cell lines producing mutant IgM, except T/μ-S414(0), were obtained by transfecting μ gene constructs into the X10 cell line, which expresses the κ-chain but not the μ-chain (22). The transfectoma, T/μ-S414(0), which produces IgM-S414 in the absence of J chain, has been previously described (20). We have used two Sp6-derived mutant cell lines, igm482 and igm43, for size markers and other controls. The cell line igm482 secretes IgM monomer lacking the Cμ4 domain (23), and igm43 secretes polymer lacking the Cμ1 domain from which the κ light chain dissociates in SDS (21). IgME10 is derived from the myeloma MOPC-315 and produces IgM bearing λ-chain (24).

Protein analysis

Biosynthetic labeling of cells was performed as described previously (4), except that [³⁵S]methionine (SJ.204, Amersham, Arlington Heights, IL) was used at a concentration of 50 to 100 μCi/ml for labeling.

³⁵S-labeled or unlabeled IgM was prepared from supernatants either by affinity purification using DNP-Sepharose (17) or by immunoprecipitation using anti-IgM Abs and protein G-agarose (4).

IgM was fractionated by sucrose density gradient centrifugation at 23,000 rpm for 16.75 h in an SW41 rotor as previously described (17).

Nonreduced IgM was fractionated by SDS-PAGE using a modification of Laemmli’s method (4, 25). To obtain optimal resolution of various IgM species, the concentrations of acrylamide and agarose were varied, as indicated in the figure legends. The sample buffer for SDS-PAGE sometimes contained 25 mM iodoacetic acid to prevent spontaneous reduction, but its inclusion was found not to affect the result. Prestained molecular mass markers (Life Technologies, Grand Island, NY; 14–228 kDa) were included in most experiments. Separated proteins were visualized either by autoradiography and PhosphorImager analysis (Molecular Dynamics, Sunnyvale, CA) or using a Western blot procedure (see below).

Radiolabeled J chain was assayed by AU-PAGE (26). To assess individual IgM species, bands were excised from the SDS-PAGE gel, extracted, and then analyzed by AU-PAGE (27). After extraction, IgM was reduced by 50 mM DTT in 0.1 M Tris (pH 8.0), alkylated with an equal volume of 0.5 M iodoacetic acid in 1.5 M Tris-HCl (pH 8.8), and precipitated with 10 vol of acetone at −20°C. Acetone-precipitated samples were dissolved in AU-PAGE sample buffer (0.2 M Tris (pH 8.0), 8.0 M urea, and bromophenol blue) and loaded onto the gel.

J chain was also detected by Western blot analysis (20). Thus, after SDS-PAGE, the nitrocellulose filter was incubated with rabbit anti-human J chain Abs (1/100 dilution; Biogenex, San Ramon, CA), followed by alkaline phosphatase-conjugated mouse anti-rabbit IgG (1/2000 dilution; Jackson ImmunoResearch, West Grove, PA). Buffer for blocking and Ab incubation contained PBS, 0.2% I-Block (Tropix, Bedford, MA), and 0.1% Tween-20. The wash buffer was 0.1% Tween-20 in PBS. The substrate buffer for enhanced chemiluminescence detection was 0.1 M diethanolamine, 1 mM MgCl₂ (pH 10.0), 1% CSPD (Tropix), and 5% N-Block (Tropix).

Complement-mediated hemolysis

Complement-mediated hemolysis of IgM was tested in V-shaped 96-well plates. Samples were diluted in balanced salt solution containing 1% calf serum. Serial dilutions of IgM samples (either from SDG fractions or standard supernatant, from the J⁺ parental hybridoma Sp6/HL) were incubated with SRBC-TNP at a final concentration of 0.12% for 1 h at 37°C. To remove sucrose, this mixture was centrifuged, and the pelleted IgM-SRBC were resuspended in buffer and recentrifuged. The washed IgM-SRBC were then resuspended at 0.12% in the presence of guinea pig complement at a dilution of 1/250 and incubated at 37°C. After 1-h incubation, unlysed SRBC-TNP were pelleted by centrifugation. Supernatants were transferred to flat-welled microtiter plates, and lysis was measured as adsorbance at 405 nm. In some experiments, hemolysis was also tested by the method of Davis et al. (17).

Northern blot analysis

Isolation of cytoplasmic RNA and Northern blot analysis were performed as described previously. Transferred, immobilized RNA was hybridized to end-labeled oligonucleotides specific for the Cμ4 exon of the μ gene (4), the V-region of the TNP-specific κ-chain (22), or the J chain (20).

A transfection recipient cell line for production of J chain-deficient IgM

We have previously described an expression system based on the hybridoma Sp6, which produces IgM-κ specific for the hapten TNP. Several mutant cell lines lacking the μ gene but retaining the functional κ gene have been derived from Sp6 (22, 28). These μ⁻, κ⁺ cell lines serve as recipients for transfected normal and mutant μ genes, which are present as a genomic expression unit in the pSV2neo vector. The mutant μ genes, resulting transfectants, and corresponding IgM are denoted according to the substituted amino acid, e.g., μ-S414, T/μ-S414, and IgM-S414, respectively.

To produce IgM bearing normal and mutant μ-chains in the absence as well as the presence of J chain, we sought to isolate a hybridoma mutant cell line that had ceased to produce J chain but which would nevertheless express the transfected μ genes. In earlier work we fortuitously isolated a transfectoma, T/μ-S414(0), expressing the mutant protein IgM-S414, but not J chain (19, 20). This cell line produces no detectable J chain protein or mRNA. The cause of this J chain deficiency is not known. To obtain a cell line that had lost the μ-S414 pSV2neo vector, the T/μ-S414(0) cell line was subjected to 300 to 400 rad from a ¹³⁷Cs source and subcloned. Approximately 800 single precursor colonies were tested by ELISA for IgM production. One negative colony was obtained; it was G418 sensitive and lacked both μ mRNA (Fig. 1⇓) and μ DNA (not shown). We then confirmed that this colony produced the κ-chain and lacked J chain mRNA (Fig. 1⇓). This cell line, denoted G10, was used to express normal and mutant μ genes.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

mRNA production by the G10 recipient hybridoma cell line. RNA was isolated from the indicated cell lines and analyzed by Northern blot using ³²P end-labeled oligonucleotides. The membrane was first hybridized to J chain probe (upper panel), then stripped and reprobed simultaneously for μ and κ mRNA (lower panel). X10, IgME10, and T/μ-S414(0), which lack expression of μ-chain, κ-chain, and J chain, respectively, were included for comparison.

Efficient production of IgM pentamer in the absence of J chain

To study the role of J chain in IgM structure and function, we transfected the G10 cell line with plasmids encoding either wild-type μ-chain or μ-S337. From each transfection, two independent transformants were chosen for further study. Their IgM, denoted J⁻IgM-wt and J⁻IgM-S337, was compared with IgM produced by previously established (J⁺) transformants expressing J⁺IgM-wt or J⁺IgM-S337 (4). Each transformant produced a level of IgM expression corresponding to 10 to 60% of that of the parental Sp6/HL hybridoma, with overlapping ranges for each pair.

We initially investigated how secreted J⁻IgM-wt was assembled using sucrose density gradient (SDG) centrifugation in combination with nonreducing SDS-PAGE to assess the state of polymerization, as hexamer is expected to sediment more rapidly than pentamer in SDG centrifugation and to migrate more slowly than pentamer in SDS-PAGE. For this analysis, J⁺IgM-wt and J⁻IgM-wt were metabolically labeled with [³⁵S]methionine, and secreted IgM was purified by its affinity to DNP-Sepharose. This IgM was then separated by (nondenaturing) SDG centrifugation into various polymeric and monomeric species. As shown in Figure 2⇓A, J⁻IgM-wt had a more prominent monomer peak than J⁺IgM-wt. To obtain a more complete separation of the polymeric species, the faster and slower polymeric fractions of the gradient were separately pooled and resedimented (Fig. 2⇓A). Analysis of selected fractions from the secondary gradients indicated that for both J⁺IgM-wt and J⁻IgM-wt, the faster sedimenting material was greatly enriched for species that have lower mobility in SDS-PAGE, and that, conversely, the slowly sedimenting material had a higher mobility in SDS-PAGE (Fig. 2⇓B). Moreover, the slowly sedimenting J⁻IgM polymers have the same mobility in SDS-PAGE as the slowly sedimenting J⁺IgM polymers, which have been previously identified as pentamer (17, 18). We conclude from these results that both J⁺IgM-wt and J⁻IgM-wt include both pentameric and hexameric species. The sequential gradients yielded pentameric material that contained no detectable hexamer band on SDS-PAGE. The hexameric material usually contained detectable pentamer (see below).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Analysis of pentameric and hexameric J⁺IgM-wt and J⁻IgM-wt. A, Separation of pentamer- and hexamer-enriched polymers by consecutive SDG.³⁵S-labeled, DNP-purified IgM was analyzed by ultracentrifugation using SDG. J⁺IgM-wt and J⁻IgM-wt are indicated by circles and triangles, respectively. Proteins sedimented from left to right in the figure. An aliquot of each fraction was counted by beta scintillation. From the primary SDG separation, pentamer- and hexamer-enriched fractions were separately pooled as indicated and recentrifuged (open and filled symbols, respectively). Two secondary gradients are shown together for each of the J⁺IgM (middle panel) and J⁻IgM (bottom panel), where the filled and open symbols represent hexamer and pentamer, respectively. B, SDS-PAGE analysis of fractions from secondary SDG. The indicated fractions from pentamer- and hexamer-enriched polymers prepared in A were analyzed following immunoprecipitation. DNP-purified, but unseparated J⁺IgM-wt and J⁻IgM-wt were included to indicate the multiplicity of species. Samples were separated in the upper phase of a two-phase gel containing 3% acrylamide and 0.4% agarose.

In light of previous studies in which hexameric IgM was the predominant species produced in the absence of J chain (14, 15), we were surprised to find that production of hexamer and pentamer did not change much in the absence of J chain. To confirm that secreted J⁻IgM-wt protein actually lacked J chain, we compared its J chain content with that of J⁺IgM-wt. Individual hexamer, pentamer, and monomer species were extracted from the SDS-PAGE gel as shown in Figure 3⇓A. After reduction, alkylation, and acetone precipitation, the samples were analyzed by AU-PAGE for J chain. To compare the different samples for J chain content, several concentrations of J⁺IgM-wt pentamer were loaded onto the gels. As expected from previous reports, J chain was found in J⁺IgM-wt pentamer but not in monomer. The preparation of J⁺IgM-wt hexamer contained about 2% as much J chain as did the preparation of pentamer, but this is probably due to contaminating pentamer (Fig. 3⇓B). By contrast, we detected no J chain in J⁻IgM-wt pentamer or hexamer (<2% of J⁺IgM-wt pentamer). This experiment was repeated for IgM-wt and IgM-S337 transformants without prior SDS-PAGE separation of its DNP-purified secreted IgM. In these experiments (not shown), J⁻IgM-wt and J⁻IgM-S337 proteins also had no detectable J chain (<3% of the content of their J⁺ counterparts). These experiments indicated that IgM pentamer was efficiently assembled even in the absence of J chain.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Analysis of J chain content. A, Preparation of IgM by SDS-PAGE. ³⁵S-labeled IgM was purified by immunoprecipitation. The upper gel is 3% acrylamide with 0.4% agarose; the lower gel is 7.5% acrylamide. The indicated IgM species were identified by comparison with the markers (IgM lacking Cμ1 domain and κ-chain (mutant igm43), 560 kDa; commercial protein marker, 226 kDa). B, Analysis of J chain by AU-PAGE. AU-PAGE was performed on bands extracted from the SDS-PAGE gel (A). The numbers indicate the amount of radioactivity (counts per minute) loaded in each lane. The positions of the μ, κ, and J chains are indicated.

Purification of IgM by binding to DNP-Sepharose and elution with DNP-alanine has the potential to enrich preferentially for some species. Therefore, we have also examined the species present in IgM that was prepared by immunoprecipitation, as this procedure was previously found to better measure the composition of secreted IgM (4). ³⁵S-labeled IgM-wt was prepared from two μ-wt-J⁺ and two μ-wt-J⁻ transformants by immunoprecipitation, and individual SDS-PAGE bands were quantified by PhosphorImager analysis. For J⁻IgM-wt the following distribution of species was found (mean ± SD): hexamer, 20 ± 9%; pentamer, 49 ± 10%; and monomer, 31 ± 3%. For J⁺-IgM-wt the corresponding values were: hexamer, 16 ± 14%; pentamer, 77 ± 14%; and monomer, 7 ± 3%. This comparison suggests that J chain increases the production of secreted pentamer at the expense of monomer. There was approximately threefold more covalent dimer and trimer in J⁻IgM-wt than in J⁺IgM-wt; each species constituted ≤4% of the total secreted J⁻IgM-wt protein.

J chain-deficient IgM pentamer activates complement poorly

Previous reports have shown that pentameric J⁺IgM is much less cytolytic than hexameric J⁻IgM, but it has been unclear whether this difference is due to the presence of J chain in pentamer or to some other structural difference between the two types of polymer. To distinguish these possibilities we compared the cytolytic activity of the pentameric and hexameric species of J⁺IgM and J⁻IgM prepared by sequential SDG centrifugations. To gauge whether cytolytic activity was derived from a single species, we assayed three fractions from different positions in the pentamer and hexamer peaks. As described in Materials and Methods, IgM from the indicated SDG fractions was serially diluted and adsorbed to TNP-coupled erythrocytes. The IgM-TNP-erythrocytes were then incubated in the presence of guinea pig serum as a complement source, and lysis was assessed spectrophotometrically (Fig. 4⇓). As a measure of cytolytic activity we have calculated the amount of IgM that lysed 50% of the erythrocytes (Table I⇓). Pentamer from both J⁺IgM-wt and J⁻IgM had lower lytic activity than unfractionated IgM-wt, and this activity was nearly the same for each of the three fractions assayed from each gradient, thus arguing that this activity was not due to contaminating hexamer. Hexamer from both J⁺IgM and J⁻IgM had higher lytic activity than unfractionated IgM-wt, and this activity increased for the faster sedimenting fractions. SDS-PAGE analysis of these fractions indicates that the fractions denoted hexamer contained some pentameric IgM (Fig. 2⇑B). The level of contaminating pentamer was less in faster migrating fractions, which had greater hemolytic activity. To estimate the activity of pure hexamer, we assumed that contaminating pentamer did not significantly contribute to the total lytic activity. Pentamer contamination was calculated from the intensity of the SDS-PAGE bands, and the measured lytic activity was divided by the fraction of hexamer to give the activity of pure hexamer.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Hemolytic activities of purified pentamer and hexamer. Pentamer- and hexamer-enriched fractions from secondary SDG were tested. As indicated, J⁺ pentamer was either titrated alone or in the presence of a constant amount of hexamer. Titration of hexamer is indicated by a filled triangle. The single opposite-pointing triangle corresponds to the amount (∼2 ng/ml) and activity (A₄₀₅ = ∼0.2) of hexamer that was mixed with pentamer.

In our analyses of pentameric IgM, we sometimes found that the extent of lysis plateaued at <100%, even at still increasing IgM concentrations (Fig. 4⇑ and Table I⇑). IgM agglutinates TNP-erythrocytes, and this effect might have inhibited lysis at high levels of pentameric IgM. To test for such an effect, we assumed that an inhibitory level of pentameric IgM should also inhibit lysis by hexamer. Therefore, a fixed concentration of hexamer, previously determined to give partial hemolysis, was mixed with graded concentrations of J⁺ pentamer. The lysis of these samples was then compared with that of an identical dilution series of pentamer that did not contain hexamer. As shown in Figure 4⇑, the difference in lysis between the two sets of samples corresponds closely to the amount of lysis seen by hexamer alone (A₄₀₅ = 0.19) up to a pentamer concentration of approximately 1000 ng/ml. A similar result was seen for J⁻ pentamer (not shown). We conclude that our assay accurately reflects the activity of pentameric IgM, at least up to 1000 ng/ml.

Data for three independent SDG experiments are summarized in Table I⇑. For both J⁺IgM-wt and J⁻IgM-wt, we found that pentamer was much less active than hexamer, indicating that its J chain content was not the only reason that pentamer was less lytic than the hexamer. As noted above, our finding that the activity of pentameric J⁻IgM did not differ much between the faster and slower fractions argues that this activity was not due to contamination by a small amount of hexameric IgM. We are unsure whether the two- to threefold differences seen between J⁺ and J⁻ pentamers and between J⁺ and J⁻ hexamers are significant (see Discussion).

A puzzling feature was that the activity of J⁺ pentamer compared with that of hexamer (∼0.8%) was much lower than the 5% value seen in earlier studies from this and another laboratory (17, 18). We have tested whether this difference reflects changes in the method of assay, with variable results. That is, in repeated testing using the method of Davis et al. (17) we found that J⁺ pentamer ranged from 1.4 to 12% of the cytolytic activity of hexamer. We have not identified the source of this variation.

Analysis of disulfide bonding pattern in J chain-deficient IgM-S337 polymers

Inter-μ-chain disulfide bonding in polymeric IgM-wt is mediated by three cysteines, C337, C414, and C575, which are believed to form homo-pairs, e.g., C337-C337 (Fig. 5⇓A). However, the connectivity of the μ-chains, i.e., the series and parallel arrangement of disulfide bonds (see Fig. 5⇓ for definitions), is not known. To gain information about the connectivity and to ascertain whether it is affected by J chain, we examined IgM-S337. Several observations argue that IgM-S337 closely resembles IgM-wt, except for the absence of the C337-C337 disulfide bond. Thus, IgM-S337 efficiently assembles into pentamer and hexamer that have normal sedimentation rates in SDG. IgM-S337 pentamer and hexamer are comparable to the corresponding wild-type proteins in their cytolytic activities. Pentameric IgM-S337 contains a normal amount of J chain (19). IgM-S337 polymers are assembled from noncovalent subunits, whose size and composition are determined by the extent to which C414-C414 and C575-C575 join μ-chains in series. In the model presented in Figure 5⇓B, an IgM-S337 pentamer is represented as a structure that would dissociate in SDS to yield one μ4κ4 subunit and three μ2κ2 subunits. In the μ4κ4 subunit, two C414-C414 bonds join the μ-chains in series with one C575-C575 bond. As illustrated, the μ2κ2 subunits are of three types: a subunit bonded by only C414-C414, a subunit bonded by only C575-C575, and a subunit bonded by C414-C414 and C575-C575 in parallel. These examples illustrate how the size of the subunits (here μ2κ2 and μ4κ4) indicates whether C414-C414 and C575-C575 are in series, but not whether they are parallel.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Models of mutant IgM pentamers. A, Ig domains and cysteines residues available for inter-μ-chain bonding. B, Possible arrangement of disulfide bonds and subunits of IgM-S337 pentamer. In the μ4κ4 unit (μ-chains 5, 6, 7, and 8), the C414-C414 and C575-C575 bonds are in series. In the μ2κ2 unit, composed of μ-chains 3 and 4, C414-C414 and C575-C575 bonds are in parallel, i.e., the two bonds bridge the same μ-chains. C, Model for IgM-S(337, 414) pentamer. In this model the pentamer is maintained by C575-C575 bonds in series with noncovalent interactions between Cμ2 domains. For simplicity, J chain is not included in these models.

We have tested for differences in the covalent assembly of pentameric and hexameric J⁺IgM-S337 and J⁻IgM-S337 by examining the species that are produced from these polymers upon denaturation by SDS. Pentameric and hexameric J⁺IgM-S337 and J⁻IgM-S337 were prepared by sequential SDG centrifugations (Fig. 6⇓), as described above for J⁺IgM-wt and J⁻IgM-wt. As was seen for IgM-wt, the fraction of IgM-S337 sedimenting as monomer in the primary gradients was about threefold higher in preparations of J⁻IgM-S337 than in the case of J⁺IgM-S337, further supporting the idea that IgM-S337 assembles like IgM-wt. As in the analysis of IgM-wt (Fig. 3⇑A), pentamer-enriched IgM-S337 sedimented more slowly than did hexamer-enriched polymers in the secondary gradients. The fact that IgM-S337 polymers had similar sedimentations in secondary and primary gradients and did not release a detectable level of monomer indicates that these polymers are stable despite their noncovalent structure. In a separate preparation of IgM-S337 (not shown), we found that pentamer was much less cytolytic than hexamer, in agreement with an earlier analysis of IgM-S337 (19); for pentamer from both J⁺ and J⁻IgM-S337, >700 ng/ml was required for 50% lysis, while hexamer from J⁺ and J⁻ transformants gave 50% lysis at 12 and 6.5 ng/ml, respectively.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Analysis of pentameric and hexameric J+IgM-S337 and J-IgM-S337. A, Separation of pentamer- and hexamer-enriched polymers by consecutive SDG. This experiment was performed as described in Figure 3⇑. J⁺IgM-S337 is indicated by circles, and J⁻IgM-S337 is indicated by triangles. Pools of pentamer- and hexamer-enriched polymers, which were used for secondary SDG, are indicated. B, SDS-PAGE analysis of fractions from secondary SDG. Fractions from the secondary SDG in A were analyzed by SDS-PAGE following immunoprecipitation. H1 through H4 and P1 and P2 indicate SDS-PAGE bands associated specifically with either hexamer or pentamer, respectively. μ4κ4 and μ2κ2 units are also indicated. For comparison, affinity-purified IgM-S337, not separated by SDG, was included. Due to different preparative techniques, the bands from these preparations did not align exactly with corresponding bands from SDG-separated samples. As molecular mass standards, the following samples were run: DNP-purified IgM-wt (1080-kDa hexamer, 900-kDa pentamer, 180-kDa momer), pentameric IM lacking the Cμ1 domain and κ chain (mutant igm43; 560 κDa) and 226-κDa commercial marker. Samples were separated in the upper phase of a two-phase gel containing 3% acrylamide and 0.4% agarose.

Selected fractions from the secondary gradients of J⁺IgM-S337 and J⁻IgM-S337 were analyzed for subunit composition by SDS-PAGE (Fig. 6⇑B). As reported previously (19) monomer (μ2κ2) and dimer (μ4κ4) species were the predominant subunits for all polymeric fractions tested, each accounting for 30 to 50% of the total IgM, as quantified by PhosphorImager. Extended exposure of the SDS-PAGE gel revealed many bands. Nevertheless, we discerned no major differences in band pattern between J⁺IgM-S337 and J⁻IgM-S337 pentamers. As will be considered in Discussion, these results suggest that J chain does not greatly alter the extent to which C414-C414 and C575-C575 link μ-chains in series.

Like hexameric IgM-wt, hexameric IgM-S337 was highly heterogeneous, and here also the J⁺IgM-S337 and J⁻IgM-S337 gave comparable patterns (Fig. 6⇑B). As in the case of pentameric IgM-S337, the predominant species were μ2κ2 and μ4κ4. Hexamer lacked several of the minor species that were present in pentameric IgM-S337 (designated P1 and P2) and that migrated slower than μ4κ4. Moreover, unlike pentameric IgM-S337, hexameric IgM-S337 included several bands (H1–H4) that migrated more slowly than either pentameric or hexameric IgM-wt. The identities of these very slowly migrating species are considered further in Discussion.

We were concerned that some of the subunits present in both the pentamer- and hexamer-enriched fractions might truly be associated with only one polymeric species or the other, but may exist in all fractions because SDG centrifugation had yielded incomplete separation. It was not possible to assess the purity of the noncovalent pentamer and hexamer in the same direct manner as we did for IgM-wt. Nevertheless, our data suggest that cross-contamination was not a serious problem. Thus, the fact that pentamer-enriched fractions from IgM-S337 were much less cytolytic than hexamer-enriched fractions indicates that pentamer fractions were not severely contaminated. Also, very pure pentamer was obtained under the same separation conditions used for IgM-wt. IgM-S337 hexamer fractions might have been contaminated to some extent with pentamer, but this also was probably not severe. Thus, IgM-S337 pentamer fractions contained two species with unique SDS-PAGE mobility (P1 and P2 in Fig. 6⇑B). These species could not be detected in the slowly sedimenting hexamer fractions, although other bands that are common to pentamer and hexamer fractions had similar intensities. These observations indicate that hexamer obtained from the faster shoulder of the peak included very little pentamer contamination.

Disposition of J chain within IgM subunits

As shown above, IgM-S337 yielded predominantly species with the composition μ2κ2 and μ4κ4. The double mutant, IgM-S(337, 414), also assembles efficiently into polymers, but the polymers dissociate in SDS into μ2κ2 subunits held together by the C575-C575 disulfide bond (4) (Fig. 5⇑C). As indicated in Figure 7⇓, the overall J chain content of secreted IgM-S337 and IgM-S(337, 414) was similar to that of IgM-wt, consistent with earlier findings that C575, but not C414 or C337, is necessary for J chain incorporation (20). To examine how J chain was distributed among the various substructures present in IgM-S337 and IgM-S(337, 414), ³⁵S-labeled polymeric IgM was purified by SDG centrifugation (not shown). IgM from the polymer-containing fractions was immunoprecipitated and then analyzed by SDS-PAGE (Fig. 8⇓A). Some of the minor bands shown in Figure 6B⇑ were not detected here, probably because of the shorter exposure time. The various bands were extracted from the gel and analyzed by AU-PAGE (Fig. 8⇓B). For IgM-S337, J chain was detected only in species B, not in species A or C. The relative mobility of these species thus indicates that J chain was in μ2κ2 subunits, but not in μ4κ4. The same result was also obtained in separate experiments in which polymeric subunits, separated by SDS-PAGE, were probed for J chain by Western blot (not shown). The significance of the restricted distribution of J chain is considered further in Discussion.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

J chain content of wild-type and mutant IgM. ³⁵S-labeled, affinity-purified secreted IgM was separated by AU-PAGE, and bands were visualized by autoradiography. IgM-S(414, 575), which is secreted as monomer and half-mer, was included as a negative control for J chain (4, 19).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

Analysis of subunit structure of polymeric IgM-S337 and IgM-S(337, 414). A, Fractionation of polymeric IgM by SDS-PAGE. ³⁵S-labeled polymers of IgM-S337 and IgM-S(337, 414) were recovered by immunoprecipitation from SDG. Labeled IgM-wt and IgM lacking Cμ4/tail (mutant igm482) were prepared from culture supernatants. Samples were separated on a two-phase, nonreducing SDS-PAGE gel (upper phase, 3% acrylamide and 0.4% agarose; lower phase, 4.8% acrylamide and 0.4% agarose). The left panel is an autoradiograph of the gel; the right panel indicates the strong bands (A, B, D, and E) and weak bands (C, F, G, and H) that were extracted for analysis by AU-PAGE. B, AU-PAGE analysis of subunits from IgM-S337 and IgM-S(337, 414) polymer. Species A through H were extracted from the SDS-PAGE gel in A. Half of each extracted sample was loaded onto AU-PAGE (leftgel). The gel to the right shows positive and negative controls for J chain content: IgM-wt pentamer and IgM lacking Cμ4/tail (mutant igm482).

As noted in the introduction, there is direct evidence that an individual μ-chain binds via C575 to at most one J chain and that an individual J chain binds via its C14 and/or C68 to at most two μ-chains. These observations have left open the possibility that some J chains are bound to μ only via C14, while other J chains are bound only via C68. To test whether an individual J chain in fact bridges two μ-chains by binding to C575, we have examined the disposition of J chain in IgM-S(337, 414). The major bands migrated approximately as μ2κ2 (species D, 21% of the total polymer; species E, 69% of the total polymer) as well as various minor bands of faster mobility. J chain was detected in species D and to a much lesser extent in species E, but not in species F, G, or H. Considering that only C575 was available for intermolecular disulfide bonding in IgM-S(337, 414), the presence of J chain in this μ2κ2 unit indicates that J chain bridged the μ-chains via a C575-J-C575 linkage.

IgM-S337 and IgM-S(337, 414) polymers yielded μ2κ2 subunits (species B, D, and E) that had slightly different mobilities. Nevertheless, under reducing conditions, μ-chains from these three species had the same SDS-PAGE mobility, identical with that of μ-chain from wild-type polymers (not shown). This suggests that the mobility difference between species D and E seen under nonreducing conditions reflects the presence of J chain, and that differences in intrasubunit disulfide bonding might be responsible for the different mobilities of species B and E.