Characterization of fragments of the murine Ia-associated invariant chain.

It has been proposed that invariant chain (Ii), a nonpolymorphic, transmembrane glycoprotein found in noncovalent association with Ia molecules, may function to protect the Ia Ag-binding site from association with self-peptides during Ia synthesis. Selective binding of foreign antigenic peptides could then be allowed by the dissociation of Ii molecules from Ia in the appropriate intracellular compartment. In this study, we have examined the structure and intracellular trafficking patterns of a putative proteolytic product of Ii, p25. We found that p25 is a non-membrane-bound fragment of Ii with an N terminus beginning at Met98 of the Ii sequence. p25 is formed at a very early stage of Ii synthesis in the rough endoplasmic reticulum rather than in a post-Golgi Ag-processing compartment. We have also characterized a second Ii-related species, p28, which has not been reported previously. The p28 form of Ii, unlike p25, is generated under acidic conditions similar to those found during Ag processing.

Internalization of Ia molecules by antigen-presenting B cells is limited.

Our data demonstrate that the uptake of surface Ia into an intracellular compartment of B lymphoma or normal spleen cells is limited to about 20% after 2 to 3 h. The extent of internalization does not vary with several types of stimulation, including LPS, phorbol esters, anti-Ig-plus phorbol ester-stimulated EL-4 T cell supernatant, and Con A supernatant. Resting and activated B cells had similar rates of internalization. The rate and extent of uptake of surface Ia molecules into an intracellular compartment was monitored quantitatively through the use of a mAb radiolabeled with 125I. The internalization of Ia molecules was compared to that of transferrin receptor, a receptor that undergoes rapid internalization and recycling and accumulates in a intracellular pool that can be trapped by monensin. The internalization of Ia was not affected by monensin, although its synthetic pathway is disturbed by this drug. The potential use of internalized Ia for formation of T cell-triggering complexes of Ia and Ag fragments is not ruled out by these data, but it appears unlikely that internalization provides the major mechanism permitting Ia interaction with Ag.

Biosynthesis and intracellular transport of MHC class II molecules associated with a mutated, glycosaminoglycan-negative invariant chain.

The MHC Class II molecular complex is composed of polymorphic alpha and beta chains and a non-polymorphic "invariant" chain (Ii) that can be converted to a chondroitin sulfate proteoglycan. To determine whether the proteoglycan form of invariant chain (Ii-CS) had a role in the expression of the Class II complex, we studied the biosynthetic fate of alpha-beta in cells transfected either with normal Ii cDNA or with a site-directed mutant of Ii (IiAla201) that lacked the site of glycosaminoglycan addition. We had reported [Miller J., Hatch J. A., Simonis S. and Cullen S. E. (1988) Proc. Natl. Acad. Sci. U.S.A. 85, 1359-1363] that the mutant protein associated stably with alpha and beta chains, and that it showed charge heterogeneity suggesting some oligosaccharide processing. In this study we demonstrated that the rate and extent of alpha-beta processing and the rate of alpha-beta cell surface expression is the same in the presence of either normal or mutant Ii. Examination of the mutant Ii protein itself revealed normal transport through different processing compartments, as evidenced by the addition of fatty acid and by maturation of N-linked oligosaccharides. Moreover, though the rate of conversion of IiAla201 into more processed forms was slower in the absence of alpha-beta than in its presence, this was also typical of the wild type invariant chain. The ability to alter glycosaminoglycan addition without significantly disturbing other processing events or trafficking should allow a rigorous assessment of the impact of glycosaminoglycan addition on Class II function.

The role of the Ia-invariant chain complex in the posttranslational processing and transport of Ia and invariant chain glycoproteins.

The invariant chain (Ii) is a nonpolymorphic glycoprotein that associates with the Ia alpha- and beta-chains of MHC class II Ag during their transport to the cell surface. Although surface expression of Ia can occur in the absence of Ii, it has not been shown whether the intracellular association of Ia and Ii affects the biosynthetic rate or specificity of posttranslational modifications to the individual molecules. Analysis of transfected cell lines carrying either Ia, Ii, or both Ia and Ii demonstrated efficient assembly of alpha-beta whether or not Ii was present. Pulse-chase studies and two-dimensional gel electrophoresis showed that Ii did not affect the addition of Ia N-linked oligosaccharide chains, or the extent or rate of their conversion to the complex form, nor did it affect the sulfation of alpha and beta glycoproteins. Ii also did not affect the rate of Ia synthesis or appearance of Ia at the cell surface. In contrast, Ia dramatically affected the posttranslational modification of Ii. Although invariant chain was modified by addition of fatty acid, N-linked oligosaccharide, and glycosaminoglycan in the absence of Ia, the processing of Ii-linked oligosaccharide into more acidic, terminally glycosylated forms was significantly less when Ia was absent, and although conversion to proteoglycan did occur, the glycosaminoglycan chains were significantly shorter than normal. The disappearance of radiolabel from the 31,000 Da form of Ii was faster when Ia was present, and the processing of Ii was more rapid when it was associated with Ia. Thus, the rate and manner in which Ii enters and passes through the Golgi is critically affected by Ia.

Identification of the glycosaminoglycan-attachment site of mouse invariant-chain proteoglycan core protein by site-directed mutagenesis.

The invariant chain (Ii), a nonpolymorphic glycoprotein that associates with the immunoregulatory Ia proteins encoded by the major histocompatibility complex, has a proteoglycan form (Ii-CS) that bears a chondroitin sulfate glycosaminoglycan. In this proteoglycan form, Ii may remain associated with Ia at the cell surface. Inhibitors that prevent the addition of glycosaminoglycan to Ii have been found to depress antigen-presenting function. Ii does not have multiple candidate glycosaminoglycan-attachment sites, and we used site-directed mutagenesis to replace a candidate serine glycosaminoglycan-acceptor site with alanine at position 201 in the murine Ii protein. Transfection of the normal or altered gene into Ii-negative COS-7 cells showed that equivalent amounts of core Ii protein and its acidic, terminally glycosylated forms were synthesized, but the Ala-201 mutant Ii did not give rise to Ii-CS. The mutant protein had apparently normal transport through the Golgi compartment and associated stably with Ia molecules. Thus, this mutation directly identifies the site of glycosaminoglycan addition and shows that it can be eliminated without adversely affecting the overall biosynthesis of Ii.

Molecular mapping of murine I region recombinants. II. Crossing over in the E beta gene is restricted to a 4.5 kb stretch of DNA that excludes the beta 1 exon.

The crossing over in five murine I-region recombinants (Is/Ik) was studied by restriction fragment length polymorphism analysis after Southern blot hybridization by using I-region-specific probes. These recombinants included three recently developed strains, B10.ASR1, B10.ASR11, and B10.ASR12; and two strains B10.S(9R) and B10.HTT derived earlier. Although these recombinants were reciprocal in haplotype orientation to the three recombinants we reported recently, these too crossed over within the same 7 kb stretch of DNA in the E beta gene. This 7 kb stretch of DNA included the 3' half of the first intron, the beta 1 exon, the second intron, and the beta 2 exon. A comparison of the cDNA sequences of the two parental E beta alleles revealed that although the beta 2 exons were identical, there were several nucleotide differences between the two beta 1 exons. This allowed us to determine the parental origin of the beta 1 exon in the recombinants at the level of transcription by using S1 nuclease mapping. Thus we were able to show that in each case the 3' portion of the first intron and the beta 1 exon were upstream from the site of crossover. All eight recombinants involving the k and s haplotypes now can be mapped within a 4.5 kb stretch of DNA, which includes only the beta 1-beta 2 intron and the beta 2 exon of the E beta gene. These findings imply that the I-E molecules expressed in these recombinants will probably have conserved sequences and therefore will exhibit identical I-E-restricted immune responses, despite the fact that crossing over could have occurred at different sites within the beta 1-beta 2 intron.

Slower processing, weaker beta 2-M association, and lower surface expression of H-2Ld are influenced by its amino terminus.

To determine why Ld antigens are expressed on the cell surface at levels three to four times lower than Dd or Kd antigens, pulse-chase experiments were used to compare their rates of biosynthesis and processing. Electrophoresis on sodium dodecyl sulfate gradient polyacrylamide gels resolved immunoprecipitates of each of these histocompatibility complex class I molecules into a slower and faster species. During the chase period, the faster migrating species appeared to be converted to the slower migrating species in a time-dependent manner. However, the conversion of Ld from the faster to the slower migrating species proceeded significantly more slowly than did the conversion of either Dd or Kd. Endoglycosidase H sensitivity and cell surface radiolabeling were used to determine the glycosylation state and cell location of each species of Ld and Dd. The results from these experiments, along with the pulse-chase studies and cytofluorometric analyses, suggest that Ld possesses a much slower rate of processing from a faster migrating, high mannose-bearing species to a slower migrating, complex oligosaccharide-bearing species found on the cell surface. Analysis of the beta 2-microglobulin (beta 2-m) association confirmed that Ld is associated with less beta 2-m than Dd. To localize the structures on class I molecules influencing their surface expression, rate of processing, and beta 2-m association, the Ddm1 molecule was analyzed. The Ddm1 molecule of the mutant B10.D2-H-2dm1 has previously been shown to be a chimeric Dd (amino-terminal)/Ld (carboxyl-terminal) polypeptide. The surface expression, processing and beta 2-m association of Ddm1 were found to be similar to Dd rather than Ld, suggesting that each of these phenomena are influenced by protein structure in the amino terminus.

Fatty acylation of murine Ia alpha, beta, and invariant chains.

Labeling of murine spleen cells with [3H]palmitate followed by analysis of immunoprecipitated Ia molecules indicated that Ia alpha- and beta-chains and their associated invariant chain contain covalently bound fatty acid. This modification is present in I-A and I-E molecules and has been found in all haplotypes examined. The 3H label was not dissociated from the glycoproteins by detergents or under the denaturing conditions of SDS-polyacrylamide gel electrophoresis. The fatty acid linked to Ii is released by treatment with neutral hydroxylamine, which indicates thioester linkage. The acylation of alpha- and beta-chains appears to involve attachment of palmitoyl groups via an ester linkage sensitive to alkaline hydrolysis. The radioactive species released from the isolated chains by treating with KOH/methanol co-migrated with palmitic acid and palmitic acid methyl ester on thin-layer chromatography.

Molecular mapping of murine I region recombinants: crossing over in the E beta gene.

The parental origin of genomic DNA from two independently derived murine I-region recombinants, B10.ASR7 [as3] and B10.BASR1 [as4], was determined by Southern blot hybridization by using DNA probes corresponding to A beta, A alpha, 5'-E beta, 3'-E beta, and A alpha genes. New E beta gene probes were specifically constructed to make analysis of the E beta gene region definitive. Although the immune response phenotypes of the recombinants had suggested an I-A subregion cross-over, a number of restriction fragment length polymorphisms distinguishing the k and the s haplotypes showed that both recombinations mapped within a 7-kb segment of the E beta gene. The validity of these results was tested by analysis of two other H-2k/s recombinants. One of them, B10.S(8R) [as1], mapped within the same 7-kb region of the E beta gene, whereas the other, B10.BASR2 [as5], mapped outside the I-region as expected. Including those studied here, there are a dozen I region recombinants whose cross-over positions have been determined at a molecular genetic level, and all of the cross-overs occurred within the E beta gene.

Invariant chain is the core protein of the Ia-associated chondroitin sulfate proteoglycan.

The murine Ia-associated chondroitin sulfate proteoglycan (CSPG) was studied both biochemically and immunochemically to determine the nature of its core protein. Chondroitinase ABC or chondroitinase AC treatment of the CSPG digested the chondroitin sulfate glycosaminoglycan, yielding a core protein that migrated with an apparent molecular weight of 38,000. Comparative V8 protease digestion of the CSPG core protein and conventional invariant glycoproteins yielded homologous peptides, indicating that the core protein and invariant chain were structurally similar. The purified CSPG and its core protein were both shown to react directly with the monoclonal anti-invariant chain antibody, In-1. Comparative tryptic peptide analysis by high performance liquid chromatography demonstrated coelution of the majority of the peptides from the invariant chain and the CSPG core protein. Collectively, these results indicate that the CSPG is an alternatively processed form of invariant chain.

Biosynthetic relationships of the chondroitin sulfate proteoglycan with Ia and invariant chain glycoproteins.

As a first step to elucidating the role of the Ia-associated chondroitin sulfate proteoglycan (CSPG) in the biology of the Ia antigens, we have studied several aspects of the interactions of the CSPG with the Ia/invariant chain glycoproteins. These studies revealed that at any time point, only 2 to 5% of Ia molecules were associated with the CSPG, and that this fraction included Ia molecules that were expressed at the cell surface. Pulse-chase studies indicated that the association of Ia molecules and the CSPG was rapid and short lived. Newly synthesized [35S]sulfate-labeled CSPG molecules were detected in association with Ia molecules immediately after a 15-min pulse, but were barely detectable after a 30-min chase, and were completely undetectable after a 60-min chase. Similarly, newly synthesized [3H]leucine-labeled Ia molecules associated with the CSPG were detectable immediately after a 20-min pulse, and after a 75-min chase, but could not be detected in association with the CSPG after a 300-min chase. Virtually no CSPG that was similar in size to that associated with Ia molecules was found free in the cells or was secreted into the media. The results in this report are compatible with the hypothesis that some or all of the Ia molecules associate transiently with the CSPG, or that a small fraction of Ia molecules associate permanently with the CSPG in a short-lived complex. These studies tend to favor a role for the CSPG in the biosynthesis of Ia rather than in intracellular trafficking or in intercellular communication.

Cellular distribution of the Ia-associated chondroitin sulfate proteoglycan.

The Ia-associated chondroitin sulfate proteoglycan (CSPG) found in anti-Ia and anti-invariant chain immunoprecipitates was originally detected in [35S] sulfate-labeled extracts derived from unseparated populations of splenocytes. To determine whether the CSPG was produced only by a subpopulation of spleen cells, we examined various cell populations for their ability to produce the CSPG. We found that B lymphocytes were the predominant source of CSPG in the spleen. The synthesis of the Ia-associated CSPG in spleen cell cultures was not diminished by the depletion of T cells or adherent cells. Moreover, the CSPG was readily detected in lysates derived from the Lyb-5- B cell subsets of xid mice, splenocytes from athymic (nude) mice, and in vitro B cell hybridomas. Peritoneal exudate macrophages from indomethacin-treated mice were also found to be capable of producing the CSPG. In all of the studies performed to date, no dissociation of the synthesis of the CSPG from the synthesis of Ia was observed in any cell type. We therefore tentatively conclude that all cells that synthesize conventional Ia molecules also synthesize the CSPG. Finally, we have been able to use anion exchange chromatography to prepare proteoglycan-enriched fractions to isolate the CSPG. This purification step has allowed us to convincingly demonstrate that the CSPG can be labeled with amino acids, and is a necessary step for detecting amino acid-labeled CSPG. This purification step method was used in the accompanying report to begin a quantitative examination of the Ia/CSPG complex, to monitor the kinetics of CSPG synthesis and association with Ia, and to determine its subcellular localization.

Structural characterization of murine Ia antigen N-linked oligosaccharides and localization of specific structures to two unique alpha-chain glycosylation sites.

The sequence of N-linked oligosaccharides of differentially glycosylated murine I-Ak alpha-(alpha 2- and alpha 3-) and beta-chains was determined. I-Ak beta-chains predominantly bear a biantennary complex oligosaccharide with a core fucose, and with the peripheral sequence SA----Gal----GlcNAc----Man. The I-Ak alpha-chain has two N-linked glycosylation sites at Asn-82 and Asn-122. When Lubrol-insoluble alpha 3-chains are examined they are found to bear high-mannose oligosaccharides of either the Man9GlcNAc2 or Man8GlcNAc2 type at both sites. When Lubrol-soluble alpha 2-chains are examined, in about 85% of the molecules the Asn-82 site bears a biantennary complex oligosaccharide with core fucose, and with the peripheral sequence SA----Gal----GlcNAc----Man. Interestingly, the Asn-122 site bears a variety of structures. In about 50% of the molecules, the structure at Asn-122 is a biantennary complex oligosaccharide without core fucose and with the peripheral sequence SA----Gal----GlcNAc----Man. In addition, it can bear other complex structures which we did not define further. The apparently restricted addition of fucose to the oligosaccharide at the alpha-Asn-82 site, even when both alpha-sites bear biantennary complex structures with the same peripheral sequence, is a feature unique to this system. The unusual variety of structures present at the alpha-Asn-122 site may indicate differential processing in different cell types.

Antigen presentation by the BCL1 murine B cell line: in vitro stimulation by LPS.

We examined the antigen-presenting capacity of BCL1 tumor cells, which are capable of differentiating in vitro with respect to immunoglobulin synthesis/secretion under the influence of LPS. In vivo passaged BCL1 cells depleted of host cell contamination either by positive selection employing panning with anti-lambda reagents, or by elimination of latex-ingesting adherent cells, are capable of MHC-restricted antigen presentation to a GAT-immune T cell line. The BCL1 cells act as antigen-presenting cells when freshly explanted, but gradual loss of this function occurs, and cells cultured for 3.5 days cannot present antigen unless LPS is included during the culture period. BCL1 cells are equivalently Ia+ after the culture period with or without LPS stimulation. Other B cell lines capable of antigen presentation appear to express this trait constitutively, and the in vivo passaged BCL1 line is therefore unique among B cell lines in having antigen-presenting cell function that can be modulated. The data suggest that freshly explanted or LPS-cultured BCL1 cells are heterogeneous with respect to antigen-presenting capacity, and the basis for this heterogeneity is being sought. BCL1 offers an opportunity to study requirements for antigen presentation by B cells.

Identification of a sulfate-bearing molecule associated with HLA class II antigens.

The human Ia antigens (DR, DS, and SB), determined by genes contained within the HLA complex on chromosome 6, are glycoprotein heterodimers consisting of a Mr approximately equal to 34,000 alpha chain and a Mr approximately equal to 28,000 beta chain. As a result of studies exploring the possibility that alpha or beta (or both) might be sulfated, a unique component of the oligomeric Ia antigen complex was discovered. When anti-Ia immunoprecipitates from Nonidet P-40 lysates of [35S]sulfate-labeled lymphoid cells were analyzed by NaDodSO4/PAGE, a molecule of considerable size heterogeneity (Mr 40,000-70,000) was observed. This component was present in both anti-DR and anti-DS immunoprecipitates prepared from both human tonsil cells and lymphoblastoid B-cell lines but was not observed in control precipitates or in association with immunoglobulin or class I HLA molecules. Preliminary biochemical studies indicate that this Mr 40,000-70,000 molecule is polyanionic, disperse in molecular weight, and sensitive to protease digestion. The sulfate-bearing moiety of this component was resistant to Pronase but sensitive to chondroitinase ABC, indicating that this molecule belongs to the chondroitin sulfate class of proteoglycans.

Identification of a new component in the murine Ia molecular complex.

In this report, we describe a previously unidentified component in the murine Ia antigen complex. SDS-PAGE analysis of anti-Ia immunoprecipitates prepared from spleen cells biosynthetically labeled with 35S-sulfate showed no detectable incorporation of 35SO4 into alpha, beta, or Ii chains but did not reveal the presence of a novel sulfate-bearing molecule of considerable molecular weight heterogeneity (46-69-kdaltons). The 46-69-kdalton molecule could be precipitated with monoclonal antibodies specific for I-A, I-E, and Ii glycoproteins but was not seen in control precipitates, nor in association with IgG or class I MHC molecules. Preliminary biochemical characterization indicated that the 46-69-kdalton product is extremely polydisperse, both in charge and apparent molecular weight, is sensitive to proteases, and bears the sulfate moiety on a large pronase-resistant structure. These results suggested this component might be a proteoglycan. Definitive identification of this component as a proteoglycan was accomplished by selective enzymatic degradation experiments which showed that the sulfate-bearing component of the 46-69-kdalton molecule is chondroitin 6-sulfate.