MHC Class II Transport from Lysosomal Compartments to the Cell Surface Is Determined by Stable Peptide Binding, But Not by the Cytosolic Domains of the α- and β-Chains¹

Presentation to CD4⁺ T lymphocytes of antigenic peptides associated with MHC class II molecules at the surface of APCs is a central step in the initiation of immune responses. MHC class II molecules consist of dimers of transmembrane proteins (the α- and β-chains) that associate with antigenic peptides in compartments of the endocytic pathway (1, 2). Therein, antigenic peptides are generated by partial proteolysis of proteins endocytosed or phagocytosed by the APC (3). MHC class II molecules are targeted to these compartments by a third membrane protein, called the invariant chain (Ii)³ (4). However, sequences in the β-chain lumenal domain also influence MHC class II transport to endosomal compartments (5, 6). Once in endosomes, the Ii chain is degraded, and a small Ii-derived peptide remains associated with the αβ dimer until it is exchanged for another peptide under the control of a nonpolymorphic MHC class II molecule, HLA-DM (7).

In the past few years, progress has been made toward defining the nature of the endocytic compartments where these events take place. In human EBV-transformed B cells (B-EBV) and melanoma cells, lysosomal compartments called MIICs (MHC class II compartment), constitute the main site of MHC class II-peptide association (8, 9, 10, 11). In LPS-stimulated murine B lymphocytes and certain B lymphoma cells, most compartments of the endocytic pathway, including early and late endosomes, as well as lysosomes contain MHC class II-peptide complexes (12, 13). In the mouse B lymphoma cell A20, an endosomal compartment called CIIV (class II vesicle), is the main peptide loading site for MHC class II dimers (14, 15, 16).

Invariant chain mediates MHC class II retention in endosomes (17) and determines their distribution between endosomes and lysosomes. In A20 cells, we have shown that a transient relocalization from endosomal to lysosomal compartments occurs when Ii chain degradation is blocked by the presence of the protease inhibitor, leupeptin (18). In the presence of leupeptin, MHC class II dimers remain associated with a partially degraded form of the invariant chain, Ii-p10, and are redistributed into lysosomes. Upon removal of the protease inhibitor, Ii-p10 is degraded, and peptide-loaded MHC class II molecules are delivered to the cell surface (18).

Interestingly, in mouse immature dendritic cells, MHC class II molecules bound to Ii-p10 accumulate in lysosomal compartments (19). Upon dendritic cell maturation, Ii-p10 is degraded, and MHC class II molecules move to endosomal CIIV, where they load antigenic peptides before reaching the plasma membrane (19). Therefore, depending on the species, the cell type, and even the stage of maturation, MHC class II molecules may traffick through endosomal and/or lysosomal compartments.

Whether from endosomal or lysosomal compartments, the pathway followed by MHC class II-peptide complexes to reach the cell surface is still unclear. Direct fusion of MIICs with the plasma membrane has been reported (20, 21), but the amount of MHC class II molecules thus delivered to the cell surface is likely to be very low (20). Recent results show that the route followed by MHC class II dimers from MIICs to the plasma membrane does not intersect recycling endosomes, and that it probably requires the formation of transport vesicles carrying the MHC class II-peptide complexes (22). Different routes are not mutually exclusive and could correspond to different origins of the compartments (endosomal or lysosomal) where the biogenesis of MHC class II molecules-peptide complexes occurs.

Intracellular trafficking of transmembrane proteins generally involves recognition of signals in their cytosolic domains, by cytosolic factors involved in the formation of transport vesicles. Thus, endosomal targeting (and/or retention) signals in the cytosolic domain of the Ii chain direct αβIi complexes from the trans-Golgi network to endosomal compartments (23, 24, 25), either directly or after a brief appearance at the plasma membrane (26, 27, 28). However, after Ii degradation in endosomes and/or lysosomes, αβ dimers are freed of the transport signals of the Ii chain. This suggests the presence of a signal responsible for transport of αβ complexes from the peptide-loading endocytic compartments to the plasma membrane. Such a signal could lie in the cytosolic tails of the α- or β-chains, which are known to contain signals for cell activation and internalization (29, 30, 31).

Another factor that has been suggested to influence MHC class II transport to the plasma membrane is peptide association. It is known, indeed, that peptide loading correlates with the stabilization of the MHC class II dimers (32), preventing aggregation at acidic pH (33) and degradation (34, 35). An increase in the surface expression of MHC class II complexes after peptide loading has also been shown (36). However, the effect of the conformational modification caused by peptide loading on MHC class II intracellular transport has not been directly examined.

Here we have analyzed the influence of the structure and conformation of MHC class II molecules on their transport from either endosomal or lysosomal compartments to the cell surface. We found that the cytosolic domains of MHC class II molecule α- and β-chains are dispensable for the transport from both endosomal and lysosomal compartments. In contrast, peptide loading determined efficient MHC class II transport from lysosomal compartments to the plasma membrane, whereas it was not required for transport from endosomes to the plasma membrane.

A variant of the A20 murine B lymphoma cell line that has lost Fc receptor expression (IIA1.6) (37) was used to generate the cell lines used here. It was cultured in RPMI medium supplemented with glutamine (2 mM), penicillin (100 U/ml), streptomycin (100 μg/ml), pyruvate (1 mM), β-ME (50 μM), and 5% FCS (Sigma, St. Quentin, France). Stable cell lines expressing transfected I-A^b molecules were grown in the same medium supplemented with 0.5 mg/ml each of G418 (Life Technologies, Cergy, France) and hygromycin (Boehringer Mannheim, Meylan, France). The D1 dendritic cell line has been described previously (38): it was cultured in Iscove’s modified Dulbecco’s medium (Sigma) supplemented with glutamine (2 mM), penicillin (100 U/ml), streptomycin (100 μg/ml), β-ME (50 μM), 10% LPS-free FCS (Life Technologies), and 30% granulocyte-macrophage CSF-expressing NIH-T3 cell-conditioned medium. The Abs used here were mouse anti-mouse I-A^b mAb Y3P and rat anti-mouse lamp-1 mAb (PharMingen, San Diego, CA).

The cDNAs encoding truncated forms of the α and β I-A^b chains were generated by PCR using, respectively, a plasmid encoding the full-length I-A^b α (provided by Dr. R. Germain, National Institutes of Health, Bethesda, MD) and cDNAs made from the I-A^b-transfected T1 cell line (provided by Dr. N. Braunstein, Columbia University, New York, NY) as matrix, and the following primers: 5′-TCTTCTCGAGCAGGATGCCGCGCAGCAGAGC-3′ (I-A^b α 5′), 5′-TCTTGATATCATCTAGAGCCTTGAATGATGAAGATGG-3′ (I-A^b α 3′), 5′-TCTTCTCGAGAGATGGCTCTGCAGATCCCC-3′ (I-A^b β 5′), and 5′-TCTTGATATCATCTAGAGAAAAGGCCAAGCCCGAGG-3′ (I-A^b β 3′).

EcoRV- and XhoI-digested PCR products were cloned downstream of a SRα promoter (39) in expression vectors carrying resistance genes for hygromycin B (NTH2, I-A^b-Δα) and neomycin (NTNeo, I-A^b-Δβ) and were fully sequenced. The full-length I-A^b α- and β-encoding cDNAs (respectively, provided by Dr. R. Germain, National Institutes of Health, and Drs. Avie Barlow and Charles Janeway, Yale University, New Haven, CT) were subcloned in the same NTH2 (I-A^b-α) and NTNeo (I-A^b-β) vectors.

IIA1.6 cells were electroporated (260 V, 975 μF) with 50 μg of each linearized plasmid (NTH2-I-A^b-α or -Δα and NTNeo-I-A^b-β or -Δβ). Three days after electroporation, cells were switched to selection medium (normal medium supplemented with 10% FCS and 1 mg/ml each of G418 and hygromycin) and semisubcloned at 1.5 × 10⁴ cells/16-mm well. Growing cells were analyzed for surface I-A^b expression by FACScan after Y3P Ab staining.

Cells expressing high levels of surface I-A^b were subcloned by limiting dilution in 96-well plates. Cells expressing the ΔαΔβ-transfected molecules at their surface were more difficult to obtain than cells of any of the other combinations (only 30% of the ΔαΔβ clones were positive for I-A^b surface expression, whereas for each of the other three combinations, at least 50% of the clones were positive).

Experiments were performed as previously described (14, 18). Briefly, cells metabolically labeled for 20 min with [³⁵S]methionine/cysteine (1 mCi/ml; Amersham, Les Ulis, France) were chased for various periods of time at 37°C with normal culture medium. When added (in the reversibility experiments), leupeptin (Sigma) was used at 2 mM in 2% FCS-supplemented medium. At the end of the chase, cells were biotinylated for 2 min with NHS-SS-biotin (Pierce, Rockford, IL; 2 mg/ml in PBS at 4°C) and lysed in 0.5% Triton X-100-containing buffer. I-A^b class II molecules were precipitated with Y3P-coated protein A-Sepharose beads (Pharmacia, Saclay, France) for 1 to 2 h at 4°C and then eluted from the beads for 5 min at 95°C in 100 μl of PBS-2% SDS. To analyze SDS-stable MHC class II dimers, this latter step was conducted for 30 to 60 min at room temperature (nonboiled conditions). Ten microliters was kept for subsequent SDS-PAGE analysis (total molecules), and 89 μl was incubated with streptavidin-agarose beads (Pierce) in 1 ml of PBS-2% Triton X-100 for 2 h at 4°C, to precipitate the biotinylated fraction of the I-A^b class II molecules (surface molecules). After washing, precipitated molecules were eluted from the beads in 15 μl of 125 μM DTT-containing SDS-PAGE sample buffer at 95°C for 5 min (or at room temperature for 30–60 min for the SDS-stable dimers experiments) before being analyzed on 12% polyacrylamide SDS gels.

Quantification was conducted either with the ImageQuant software on a Molecular Dynamics PhosphorImager (Sunnyvale, CA) or with Bio-1D software (Vilber-Lourmat, Marne la Vallée, France) after scanning the autoradiographs with a camera (Bio-print system, Vilber-Lourmat). In experiments performed according to the reversibility protocol, the I-A^b molecules newly arrived at the cell surface were estimated by correcting the values at the different chase times with the value of I-A^b molecules expressed at the cell surface at time zero.

Immunofluorescence was performed as previously described (37). Cells were allowed to adhere to a poly-l-lysine-coated glass coverslip for 30 min at room temperature and then fixed in 3% paraformaldehyde in PBS for 20 min, permeabilized in 0.05% saponin, and incubated with primary Abs (Y3P, biotinylated Y3P, or anti-lamp-1) diluted in 0.05% saponin-0.2% BSA-PBS for 30 min at room temperature. For D1 cells, this step was performed after a 30-min preincubation in 2.4G2 (rat anti-mouse Fc receptor) supernatant. After washing, cells were incubated with FITC- or TRITC-conjugated streptavidin or F(ab′)₂ of donkey anti-mouse or donkey anti-rat Abs (Pierce). Cells were mounted in Mowiol and analyzed using a TCS scanning laser confocal microscope (Leica Microscopy and Scientific Instruments, Heerbrugg, Switzerland).

To test the possibility that transport of MHC class II molecules from endocytic compartments to the plasma membrane may require signals present in the α- or β-chain cytosolic domains, we generated cDNAs encoding truncated forms of the murine I-A^b α- and/or β-chains, called, respectively, Δα and Δβ, in which the entire cytosolic domains were deleted (Fig. 1,A). The four possible combinations of the full-length and truncated α- and β-chains were transfected into the IIA1.6 murine B lymphoma cell line, a variant of the A20 cell line (37) expressing endogenous MHC class II I-A^d molecules. Since previous reports have shown that the steady state surface expression of truncated MHC class II molecules is similar to that of full-length molecules (40), we decided to select, for each combination, at least two independent clones expressing similar levels of the transfected I-A^b molecules at their surface (Fig. 1 B).

By immunoprecipitation of mature I-A^b αβ dimers from the four types of metabolically labeled cell lines, we tested whether the transfected molecules are of the sizes expected for full-length and truncated molecules (Fig. 1,C). In all the experiments performed, the signal obtained with the ΔαΔβ clones was weaker than that with the αβ, Δαβ, and αΔβ cells, suggesting that acquisition of the mature conformation recognized by the Y3P Ab is impaired by the double truncation of α and β cytosolic domains. This experiment shows that the transfected I-A^b α- and β-chains do not detectably form heterodimers with the endogenous I-A^d chains; in truncated α (or β)-chain-expressing cells, immunoprecipitation with the I-A^b specific Ab (Y3P) does not coprecipitate the endogenous full-length I-A^d, as shown by the sole presence of a band corresponding to the truncated α (β)-chains after SDS-PAGE analysis (Fig. 1 C).

The cell lines generated were then used to analyze the transport of full-length and truncated αβ dimers from either endosomal or lysosomal compartments to the cell surface. In a previous study, we described in the A20 murine B lymphoma cell line the endosomal compartments in which MHC class II molecules accumulate before associating with antigenic peptides (14, 15). In these cells, MHC class II do not transit through lysosomes, indicating that peptide loading takes place exclusively in endosomal compartments. However, upon leupeptin treatment of I-A^b-expressing A20 cells, newly synthesized MHC class II molecules remained associated with Ii-p10 and relocalized to lysosomal compartments (18). This relocalization was reversible; when leupeptin was washed out, Ii-p10 degradation was completed and MHC class II molecules appeared at the plasma membrane (18). Here, we have used this experimental system to study MHC class II trafficking from either endosomal (in untreated cells) or lysosomal compartments (in leupeptin-exposed cells, reversibility experiments) to the plasma membrane, using metabolically labeled and surface biotinylated cells as previously described (14, 15, 18) (Fig. 2).

If the cytosolic domain of MHC class II molecules is important for their transport from the endocytic pathway to the cell surface, then the kinetics of plasma membrane delivery of full-length and truncated molecules should be different. We therefore conducted experiments according to the first protocol described in Figure 2 on the αβ, Δαβ, and αΔβ clones. After pulse labeling, chase, and surface biotinylation of the cells, mature I-A^b dimers were precipitated with Y3P Ab, and the biotinylated fraction of these molecules was reprecipitated with streptavidin-agarose. In these conditions, the streptavidin-precipitated material represents exclusively cell surface-expressed molecules as evidenced by 1) the absence of precipitation of internal molecules such as Ii, and 2) the absence of precipitation of I-A^b dimers if the biotinylation step is omitted (data not shown) (14, 15, 18).

In all three cell types, the mature dimers began to form about 1 h after the pulse, and they reached the plasma membrane 30 to 60 min later (Fig. 3 A). Interestingly, no delay was observed in the arrival at the surface of either of the truncated dimers. Even in the double mutant ΔαΔβ, the few mature dimers that formed were transported normally to the plasma membrane (not shown). Therefore, MHC class II transport from endosomes to the plasma membrane is independent of the cytosolic domains of the α- and β-chains.

To analyze transport from lysosomes to the cell surface using the second type of protocol described in Figure 2, we first checked that the truncated molecules were relocalized to lysosomal compartments as efficiently as full-length I-A^b in the presence of leupeptin. Figure 3,B shows immunofluorescent staining of I-A^b molecules in a Δαβ-expressing clone in control conditions (Fig. 3,B, upper panels) or after 4 h in the presence of 2 mM leupeptin (Fig. 3,B,lower panels). The transfected I-A^b molecules are mainly localized at the plasma membrane in cells cultured in control conditions, with little intracellular staining and no colocalization with the lysosomal marker, lamp-1 (Fig. 3,B). We have previously shown that the low levels of intracellular class II molecules are due to the rapid kinetics of transport through endosomal CIIV in these cells (14, 18). In contrast, in the presence of leupeptin, most of the I-A^b molecules relocalized to intracellular lamp-1-containing compartments (Fig. 3 B). Similar results were obtained with the αβ-, αΔβ-, and ΔαΔβ-expressing clones (data not shown); they are consistent with our published results obtained using the B4-14 clone (18). Approximately 80% of the newly synthesized class II molecules were found in lysosomes under these conditions (18).

In pulse-chase reversibility experiments, the αβ molecules that had accumulated in lysosomes started to reach the plasma membrane about 2 to 4 h (average timing obtained in two different clones) after leupeptin removal (Fig. 3,C), and the surface signal reached a plateau between 8 and 20 h of chase. Similar results were obtained with the αΔβ-, the ΔαΔβ-, and the Δαβ-expressing cell lines (Fig. 3 C and data not shown). The kinetics obtained were thus similar to those described for the B4-14 cell line. Therefore, transport of I-A^b molecules from either endosomal or lysosomal compartments to the cell surface does not require the cytosolic tail of α- or β-chains.

Besides Ii chain degradation, which could unveil a signal in the α- or β-chain cytosolic domain, the other critical event that could influence MHC class II transport to the cell surface is peptide loading. We next investigated whether peptide binding could influence I-A^b transport from endosomal and lysosomal compartments to the plasma membrane. A conformational modification of MHC class II molecules upon peptide binding has been described previously (32); αβ dimers become stable in the presence of SDS at room temperature and migrate as a 50- to 60-kDa band on SDS gels, instead of the two 32- to 28-kDa bands observed for unstable αβ dimers, or upon boiling of the samples.

We used the protocols described in Figure 2 to compare the surface delivery of newly synthesized SDS-stable and -unstable MHC class II dimers in the I-A^b αβ-expressing B lymphoma cell line. To analyze both SDS-stable and -unstable I-A^b dimers, Y3P-precipitated molecules were eluted from protein A-Sepharose beads in SDS at room temperature, instead of at 95°C. After streptavidin-agarose precipitation of the biotinylated fraction of I-A^b dimers, precipitated molecules were eluted from the beads at room temperature in the presence of DTT (125 μM) to cleave the disulfide bound between biotin and cell surface proteins. Both samples (total and surface MHC class II molecules) were then analyzed by SDS-PAGE.

As shown in Figure 4A, in untreated cells (protocol 1, Fig. 2) both stable (compact, C) and unstable (U) αβ dimers reached the cell surface (Surface) with similar efficiencies; the ratios of compact to unstable molecules were comparable in the total and surface samples. Indeed, quantification of the ratios between surface-expressed and total I-A^b molecules shows that compact dimers are transported to the plasma membrane only slightly more efficiently than the unstable ones. In four independent experiments, the rate of SDS-stable dimer arrival at the surface, defined by the surface-expressed/total molecules ratio as a function of time, is not significantly different from that of the SDS-unstable dimer arrival at the surface (Fig. 4 B). Therefore, the efficiencies of transport of SDS-stable and SDS-unstable complexes are similar under conditions where MHC class II molecules transit through endosomes.

To evaluate the efficiency of transport to the cell surface from lysosomal compartments, we performed reversibility experiments in which transport of SDS-stable and -unstable dimers was measured after leupeptin exposure. Upon leupeptin removal, a fraction of the unstable dimers became compact and was quickly transported to the cell surface (Fig. 4,C), whereas the remaining unstable dimers never reached the cell surface even though they were not associated with Ii-p10 anymore (no Ii-p10 is coprecipitated by the Y3P Ab after 8 h in the absence of leupeptin, see Figs. 4 C and 3C). Therefore, after lysosomal relocalization of newly synthesized MHC class II molecules, in the presence of leupeptin, SDS-stable dimers were very efficiently delivered to the cell surface, whereas unstable dimers were barely detectable there.

Quantification of four different experiments shows that under these conditions, compact dimers are transported to the plasma membrane 10 times more efficiently than unstable dimers (Fig. 4 D). Since compact forms of I-A^b dimers result from a stable MHC class II dimer-peptide association, our results indicate that transport of peptide-loaded I-A^b molecules from lysosomal compartments to the cell surface is more efficient than transport of empty and/or unstable molecules. In contrast, transport from endosomal compartments to the plasma membrane does not significantly discriminate between SDS-stable and -unstable dimers.

As our results were obtained in cells in which MHC class II molecules had been artificially relocalized to lysosomes by leupeptin exposure, we investigated whether the transport of SDS-stable I-A^b complexes in cells in which MHC class II molecules normally accumulate in lysosomal compartments was also selective. We chose the I-A^b-expressing murine D1 dendritic cell line (38), which shows all the characteristics of immature murine dendritic cells. Figure 5 A shows the intracellular localization of I-A^b molecules in the immature D1 cells; MHC class II molecules are mainly concentrated in intracellular compartments containing the lamp-1 lysosomal marker as well as H2-M (not shown).

Pulse-chase experiments were performed on D1 cells according to protocol 1 described for the B lymphoma cell line. The kinetics of mature αβ dimer formation and surface delivery were not different from those observed for the B lymphoma cell line; mature dimers appeared after 1 h of chase and reached the plasma membrane 30 to 60 min later (Fig. 5 B). In this respect, D1 cells behave more like human immature dendritic cells (41) than like freshly explanted mouse immature dendritic cells, in which newly synthesized MHC class II molecules are retained in lysosomal compartments and never reach the cell surface (19).

However, strikingly, SDS-stable complexes were much more efficiently transported to the cell surface than the unstable ones. Quantification (Fig. 5 C) of four independent experiments gave a slope of the surface delivery curve 4 times higher for the stable than for the unstable αβ mature complexes. These results show that in a cell line in which MHC class II molecules traffick through lysosomal compartments, peptide-bound αβ dimers are more efficiently conveyed to the plasma membrane than unstable molecules.

The influence of the origin (endosomal or lysosomal) of the endocytic compartments in which MHC class II molecules are loaded on the peptide repertoire expressed at the APC surface is still unclear. We do know, however, that several factors from the endosomal environment control the nature of the peptides loaded onto MHC class II molecules: the proteolytic enzymes generating the antigenic peptides; the pH, which controls their activity; as well as the nonclassical MHC class II molecules HLA-DM/H2-M and HLA-DO/H2-O (3). Our results show that the nature of the endocytic compartments where peptide loading takes place determines the stability of the peptide-MHC class II complexes that are delivered to the cell surface; both SDS-stable and unstable dimers reach the cell surface from endosomal compartments, whereas only stable dimers efficiently exit lysosomal compartments.

The idea that stable peptide loading (or at least SDS-stability) influences MHC class II surface expression has previously been proposed (33, 36). Germain’s group has shown that incubation of B lymphoma cells with a peptide that stably associates with MHC class II molecules induces an increase in the overall levels of class II expression at the cell surface (33). The same group showed that stable peptide loading on purified class II molecules induced their stabilization, preventing aggregation at acidic pH, and proposed that this stabilization determines their transport out of the endocytic compartments where peptide loading takes place (36). However, no experimental evidence supporting this possibility in terms of cell biology was provided.

We now show that the origin of the endocytic compartments, endosomal or lysosomal, is determinant for transport of the loaded class II molecules to the cell surface. The selectivity in transport to the cell surface of SDS-stable class II molecules most likely results from a selective retention of unstable complexes in lysosomes. Association to the Ii chain fragment Ii-p10 is known to cause MHC class II retention in lysosomal compartments (17, 18). However, selective retention of unstable dimers cannot be imputable exclusively to Ii-p10, since the Ii fragment was completely degraded after 4 to 6 h of reversibility, a time when abundant unstable class II molecules were still retained intracellularly (Fig. 4 C), and most SDS-stable dimers were already at the cell surface.

It is most likely, as previously suggested, that in the acidic environment of lysosomes, unstable class II molecules unfold and aggregate, perhaps with putative lysosomal chaperons, which cause their retention and subsequent degradation. Indeed, we observed a selective degradation of unstable dimers, although the percentage of molecules degraded during the time of our assay did not exceed 20 to 30%. It has also been previously reported that stable peptide binding prevents MHC class II degradation both in vitro and in vivo (34, 35).

However, because the pathways for MHC class II transport to the surface are still poorly understood, the actual mechanisms underlying the retention of SDS-unstable dimers in lysosomes are difficult to understand. Two different mechanisms of MHC class II transport to the plasma membrane have been suggested: direct fusion of the entire MHC class II-containing compartments with the plasma membrane (20, 21), and formation of transport vesicles containing the peptide-MHC class II complexes, targeted, directly or not, to the plasma membrane (22). In their recent work, Pond and Watts (22) show that brefeldin A inhibits the addressing of SDS-stable MHC class II molecules to the plasma membrane, supporting the involvement of vesicle budding in this process.

In contrast, we found that transport to the plasma membrane was not affected by truncation of the α- and β-chain cytosolic domains, suggesting that if vesicle budding is required for transport from the endocytic pathway to the cell surface, sorting of MHC class II into these vesicles is independent of their cytosolic domain. It is known that the cytosolic domain of the transferrin receptor is not required for its recycling from endosomes to the cell surface (42). It is therefore not surprising that transport of MHC class II molecules from endosomal compartments to the cell surface does not depend on their cytosolic domain. However, a requirement of these domains for transport from lysosomes to the cell surface would have been expected, since transport from these compartments to the plasma membrane is in general very inefficient. Our results, therefore, suggest that a constitutive transport pathway of membrane proteins from lysosomes to the cell surface exists, and that lysosomal localization requires retention or internalization and lysosomal targeting signals.

We also confirmed results obtained previously with another MHC class II molecule (40): namely, a negative effect of both α- and β-chain truncation on the efficiency of MHC class II α- and β-chain conversion to mature dimers (in our work, recognized by the Y3P Ab). These results suggest that the truncated α- and β-chains do not fold properly, causing their retention and degradation in the endoplasmic reticulum, as previously suggested (40).

Finally, what could be the biologic relevance of a selective retention of unstable MHC class II dimers in lysosomal, but not endosomal, compartments? The nature of the peptide-loading compartment on MHC class II has been a matter of controversy for the past few years. However, it appears that MHC class II molecules may transit through different endocytic compartments (43), and although this has not been formally demonstrated, it seems likely that peptide loading occurs all along the endocytic pathway.

From early endosomes to lysosomes, the luminal milieu of the compartments changes, becoming increasingly acidic and acquiring more active proteolytic enzymes. This is thought to be correlated with a difference in the antigenic peptides generated in these various compartments, since the most accessible parts of the proteins might give rise to peptides in endosomal compartments, whereas the deeper (i.e., more protected) parts of Ags would need to reach lysosomes to be degraded. A difference in the types of peptides generated in endosomal vs lysosomal compartments has actually been shown recently (44).

The overall expected effect of the preferential retention of unstable peptide-loaded MHC class II molecules in lysosomes is the selection of more stable peptide-MHC class II complexes displayed at the cell surface, as loading occurs in later endocytic compartments. It is also important to consider that in physiologic situations, APC internalize Ags through specific membrane receptors: surface Ig in B lymphocytes and mannose receptors or Fc receptors in monocytes and dendritic cells. It is expected that the higher the affinity of the receptors for the internalized Ags, the longer lasting will be their association during endocytic transport, and the more efficiently will Ags be delivered to lysosomal compartments (as pH drops, low affinity interactions could be disrupted before transport to lysosomes). In vivo, as the Ab B cell response develops, Abs with increasing affinity for the Ag are produced. Therefore, B cells, through their surface Ig, and dendritic cells and monocytes, through their Fc receptors, internalize increasingly stable Ag-Ab complexes. If it were functional in vivo, the selective retention of unstable MHC class II dimers in lysosomal compartments would therefore result in the selection of increasingly stable peptide-MHC class II complexes for expression at the APC surface, as the affinities of the specific Abs produced increase.

We thank R. Germain, and A. Barlow and C. Janeway for providing us with the cDNAs encoding the murine I-A^b α- and β-chains, respectively; N. Braunstein for the I-A^b-expressing T1 cell line; P. Benaroch and I. Mellman for critical reading of the manuscript; and D. Meur for photography.