Article

Immunity,

Vol. 3, 197-205,

August,

1995, Copyright

0 1995 by Cell Press

DM Enhances Peptide Binding to Class II MHC by Release of Invariant Chain-Derived Peptide Melanie A. Sherman, Dominique A. Weber, and Peter E. Jensen Department of Pathology Emory University School of Medicine Atlanta, Georgia 30322

Summary Major histocompatibllity complex (MHC) class II molecules bind antigenic peptides rapidly after biosynthesis in antigen-presenting cells (APCs). By contrast, the rate of peptide binding to purified class II molecules is remarkably slow. We find that purified HLA-DR molecules bind peptides rapidly in the presence but not the absence of HLA-DM, a recently identified heterodlmer required for efficient antigen processing. The same effect is seen with immunoprecipitated DM, suggesting that DM interacts directly with DR. Class IIassociated Invariant chain peptides (CLIP) are selectively and rapidly released from DR during incubation with DM at pH 5. We conclude that DM is a cofactor that enhances peptide binding to DR molecules through a mechanism involving peptide exchange. Introduction Major histocompatibility complex (MHC) class II ab dimers assemble in the endoplasmic reticulum (ER) as a nonameric complex with invariant chain (Ii) in a stoichiometric fashion (Marks et al., 1990; Roche et al., 1991). Ii is a nonpolymorphic, non-MHC-encoded protein that is essential for the proper assembly (Peterson and Miller, 1990; Anderson and Miller, 1992) and transport of class II dimers (Lamb et al., 1991; Lotteau et al., 1990; Guagliardi et al., 1990; Neefjes et al., 1990; Bakke and Dobberstein, 1990). In addition, the invariant chain may block the association of peptides with class II in the ER (Teyton et al., 1990), preventing the binding sites from being saturated with endogenous peptides. Binding of invariant chain and peptide to class II is mutually exclusive (Roche et al., 1992; Teyton et al., 1990; Rocheand Cresswell, 1990a, 1991; Newcomb and Cresswell, 1993). The class II-ii complex is targeted to the endosomal pathway, where the Ii protein is degraded and class II is released. Recent reports have impiicated a specialized endocytic compartment, MIIC, where empty class II may encounter and load peptides derived from exogenous antigen (Qiu et al., 1994; Amigorena et al., 1994; Tulp et al., 1994; West et al., 1994). CLIP is the region of invariant chain (residues 81-105) that is responsible for both association with class II and inhibition of peptide binding (Romagnoli and Germain, 1994; Bijlmakers et al., 1994). The CLIP peptide has been eluted from both mouse and human class Ii molecules (Hunt et al., 1992; Rudensky et al., 1991; Riberdy et al., 1992; Chicz et al., 1992) and is probably a product of Ii proteolysis in the endocytic pathway. Class II-CLIP com-

plexes can be generated in vitro by digestion of DR aj3li with the endosomal proteases cathepsins 6 and D (Awa and Cresswell, 1994). The rate of association and dissociation of the peptideclass II complex are unusual compared with other receptor-iigand interactions. In experiments using purified class II, the rate of association is very slow, requiring 24-48 hr to reach saturation (Buus et al., 1986; Roche and Cresswell, 1990b). Once formed, the complex is stable with an even slower dissociation rate (Buus et al., 1986). The dissociation kinetics are consistent with the physiological role of class II molecules, which present a variety of peptides by displaying them on the antigen-presenting ceil (APC) surface long enough to be scrutinized byT cells. However, the observed association rate using purified class II does not explain the rapid physiological peptide loading that occurs within the APC in vivo. Many laboratories report that antigen-derived peptide associates with class II in cells after only 20-60 min (Roosneck et al., 1988; Davidson et al., 1991; Marsh et al., 1992; Harding and Geuze, 1992). In addition, a large fraction of class II molecules become stable as dimers in SDS (which is indicative of stable association with peptide; Germain and Hendrix, 1991) within a few hoursof biosynthesis (Amigorena et al., 1994; Castellino and Germain, 1995). These observations imply the presence of one or more cofactors expressed in APC that are responsible for the enhanced rate of peptide binding to class II. A number of antigen-processing mutant ceil lines have been described that can present peptide but not protein antigens (Mellins et al., 1990, 1991; Riberdy and Cresswell, 1992; Ceman et al., 1992). Although these cells express wild-type ii and class II molecules, the class II isolated from these cells exhibit certain biochemical differences when compared with class II in normal cells. First, although the aj3 dimers assemble normally, they lack certain conformationally specific antibody epitopes (Riberdy and Cresswell, 1992). Second, the class II molecules do not acquire stability in SDS (Riberdy and Cresswell, 1992). Finally, class II heterodimers from mutant cells are predominantly associated with the CLIP peptide (Riberdy et al., 1992). Studies with these mutant cell lines have mapped the defect to the MHC-encoded nonclassical class II protein termed DM (Morriset al., 1994; Fling et al., 1994). Expression of DM in the mutant ceils restores conformational epitopes and SDS stability to class II, and association with CLIP is reduced (Denzin et al., 1994). DM localizes to the class II-containing endocytic vesicles but is not found at the cell surface (Karlsson et al., 1994; Denzin et al., 1994; Sanderson et al., 1994). A recent study found that DM, class II, and Ii are the minimal factors required for formation of mature SDS-stable class II in non-APCs (Karlsson et al., 1994). Several mechanisms have been suggested for DM function. It may act as a molecular chaperone, to target either class ii or antigenic peptides to an endocytic compartment where peptide loading occurs. Alternatively, DM may be

Immunity 199

a CLIP “sink”, which absorbs invariant chain peptides created during Ii degradation, preventing class II from being saturated with CLIP instead of antigenic peptides. Finally, DM may be a catalyst for the peptide loading step, either helping to release CLIP from class II or directly enhancing the binding of peptides toclass II molecules. Todistinguish between these possibilities, we reconstituted purified mature DRl into DM-containing membranes and measured the peptide loading of these class II molecules. We found that DM directly enhances peptide loading of class II by facilitating CLIP release. These results identify DM as the factor responsible for rapid peptide loading within APC and explain the phenotype of the mutant APCs.

confirmed by Western blot analysis using specific rabbit antisera (Figure 2A). Liposomes of D&expressing L cell membranes were reconstituted with purified DRl. Again, DRl showed greater ability to bind peptide during a 3 hr incubation in the presence of DM compared with DRl in control liposomes (Figure 28). The enhanced binding in the presence of DM is most striking during short incubations, as shown in Figure 2C. After 32 hr, the quantity of DRl-peptide complexes in L cell liposomes approached that in LDM liposomes. Peptide loading was optimal at low pH (Figure 2D), consistent with the hydrogen ion concentration present in endosomal compartments. These results point to DM as a cofactor that may catalyze peptide binding to available class II molecules.

Results MHC-Linked Genes Control Peptide Loading by DRl Molecules Purified class II molecules bind peptide antigens extremely slowly, requiring 24-48 hr to reach saturation (Buus et al., 1986; Roche and Cresswell, 1990b; Jensen, 1991). However, class II molecules load with peptides within a few hours after synthesis in vivo (Davidson et al., 1991; Amigorena et al., 1994; Castellino and Germain, 1995; Roosneck et al., 1988), raising the possibility that cofactors may facilitate peptide loading under physiological conditions (Jensen, 1991). This possibility was supported by preliminary studies demonstrating that the rate of peptide binding is enhanced after reconstitution of purified class II proteins into liposomes containing membrane components from B lymphoblastoid cells (data not shown). Membranes from wild-type Tl and mutant T2 lymphoblastoid cells were reconstituted into liposomes with purified DRl by detergent dialysis to determine whether a putative cofactor(s) is encoded in the MHC. T2 cells have a large homozygous deletion in the MHC, which includes all class II structural genes and the genes that encode DM (DeMarset al., 1984; Erlich et al., 1986; Salteret al., 1985). Each liposome preparation contained similar amounts of DRl as determined by immunoassay (data not shown). Biotinylated peptides were incubated with the liposomes at acidic pH, and biotin-peptide/class II complexes were measured. DRl bound significantly more peptide during a 3 hr incubation when reconstituted into Tl membranes compared with T2 membranes. This enhanced binding was observed using three peptides with binding specificities for DRl, but not with a control peptide having a different binding motif (Figure 1). Tl expresses endogenous DRi’(Salter et al., 1985), which did not bind to the peptides used in this experiment. These results indicate that acomponent(s) in Tl but not T2 membranes enhances peptide binding to DRl molecules. Enhanced Peptide Binding in the Presence of HLA-DM Tl cells express a number of MHC class II-encoded proteins that are not present in T2 cells, including DM. To identify DM specifically as the cofactor involved in efficient peptide loading, we transfected L cells with both DMA and DMB genes. Expression of DMa and DMp proteins was

A Direct Effect of DY on Peptide Binding by Detergent-Solublized DR1 A number of experiments demonstrated that the observed activity depends on the concentration of DM. Liposomes were generated by mixing different proportions of solubilized membranes from L cells and LDM transfectants before addition of DRl and dialysis. The enhancement of peptide binding was proportional to the relative quantity of LDM membranes (Figure 3A). In addition, the activity of membranes from various sublines of LDM transfectants correlated with the level of DM expression measured in Western blots (Figure 38). We found that membrane reconstitution was not required for DM to function. Detergent-solubilized LDM membranes but not L cell membranes enhanced peptide binding by purified DRl (Figure 3C). The above results did not exclude the possibility that DM may act indirectly by influencing the concentration or activity of another factorthat actsdirectlyon DRl. Toaddress this possibility, we

Y

Figure branes

I

1. Peptide

I

Binding

by DRl

in Reconstituted

I

Tl and T2 Mem-

Liposomes containing 50 ug TI or T2 membrane protein and 20 pmol DRl (if indicated) were incubated with 1 uM biotinylated peptide for 3 hr at 37%. After solubilization with NP40 buffer, biotin-peptide/ class II complexes were detected by europium fluorescence assay as described in Experimental Procedures.

prg Enhances

Peptide

Binding

--c-

2

hours

3

4

5

LDM

6

7

PH

Figure 2. Effect of DM on Peptide Binding to DRl in L Cell Transfectant Membrane Liposomes (A) Western blot analysis of DM protein expression in cell lines. Membranes (15 ug) from Tl , T2, L cells (L), and L cell sublines transfected with DMA and DMB genes (LDM) were resolved on SDSPAGE and probed with antisera to DMa or DMf3. The approximate apparent molecular masses of DMa and DM5 were 34 and 30 kDa, respectively. (B) Enhanced peptide binding to DRl reconstituted into liposomes containing LDM compared with L cell membranes. Liposomes (50 sg protein) from LDM and untransfected L cells containing 20 pmol DRI were incubated with biitinylated peptides (1 pM) for 3 hr at 37% (pli 5) in the presence of protease inhibitors. Peptide binding was measured as described in Experimental Procedures. (C) Kinetics of peptide binding in the presence of DM. DRi-containing liposomes were incubated with 1 uM biotin-MAT (17-31) at 37% (pH 5) for the indicated length of time and bound peptide was measured. (D) Effect of pH on MAT (17-31) binding to DRI in the presence of DM. Liposomes containing 20 pmol DRI were incubated with 1 uM biotinMAT (17-31) in 0.1 M citrate/phosphate buffer at the indicated pH for 3 hr at 37% and bound peptide was measured.

studied the effectof immunoprecipitated DM moleculeson peptide binding. DM was captured from detergentsolubilized Tl membranes with antiserum specific for the cytoplasmic domain of DMp. Captured DM directly enhanced peptide binding by purified DRl (Figure 3D). No enhancement was seen in control groups with preimmune serum or T2 cell membranes. The same results were obtained with DM captured from solubilized LDM membranes (Figure 3E). DM Induces the Release of Class II-Associated Invariant Chain Peptldes Similar amounts of peptide-class II complexes are generated after long incubations whether DM is present or not.

The population of class II that is available to bind peptide could be empty or bound with easily removed peptides, such as CLIP. To determine directly whether DM can induce the dissociation of peptide, preloaded biotin-peptide/DRl complexes were incubated with LDM membranes. DM induced the dissociation of CLIP but not the high affinity antigenic peptides HA (306-316) (Figure 4A) or MAT (17-31) (data not shown). Addition of unlabeled MAT (17-31) did not further enhance DM-induced CLIP dissociation. A role for DM in facilitating the replacement of CLIP with antigenic peptide was further supported by the results of experiments with DR3 isolated from DM-deficient cells. A major fraction of DR3 molecules isolated from T2 transfec-

Immunity 200

LDM I.

25

0

50 % LDM

B 1000

Reconstituted

75

loo

125

membranes

Liposomes

’

T

Lb T

DMa -

membranes antibody

t C

DMB

n DW

control

E

DRa -

membranes, antibody~DM~

L

LDM control

LDM DMf3

Tl control

Tl DMfj

Figure 3. Direct Effect of DM on Peptide Binding by DRI (A) Bolubilized L and LDM membranes were mixed in various relative amounts, keeping the final membrane concentration in each sample equal. DRI (0.4 pM) was added before detergent dialysis. The liposomes (containing 20 pmol DRl) were incubated with 1 pM biotin-MAT (17-31) for 3 hr and biotin-peptide/DRl complexes were measured as described in Experimental Procedures. (B) LDM sublines expressing different amounts of DMa show corresponding degrees of enhanced peptide binding. Membranes from Tl, T2, L, and several LDM sublines were reconstituted into liposomes with DRI as described above. Each liposome preparation contained exactly the same amount of membrane protein as measured by BCA assay. Borne liposomes were incubated with 1 pM biiin-MAT (17-31) for 3 hr and biitinpeptide/class II complexes were quantifisd with the europium assay. The remaining liposomes were solubilixad in SDS sample buffer and run in reducing SDS-PACE. Western blots were used to measure relative amounts of DMs and DR as described in Experimental Procedures. We also found that the presence of DM5 as determined by Western analysis is absolutely required for function (data not shown). (C) Detergent-solubilized membranes containing DM enhance peptide binding to purified DRl. L and LDM membranes were solubilized in 0.5% NP-40 and serial dilutions were incubated with 10 pmol DRl, 1 pM biotin-MAT (17-31). and protease inhibitors at pli 5 for 3 hr at 37% followed by measurement of bound peptide.

DM Enhances 201

A *Cm

Peptide

Binding

biotin-HA(306-318)DRl

,

bidin-CLIP/DRl

I200

1

L L

LDM

LDM

L

LDM

LIiM

LDM

MA&-PI)

M.&Jl)

B LLDM peptide

-

+ I -

+

lane stable w

Figure

4. DM Induces

the Dissociation

1 4.2)

- 2

P 3

4.8(46)

7.0(85)

9.2)

5.5(54)

1.2(15)

1.3 (12)

4

of CLIP from DR Molecules

(A) Dissociation of synthetic CLIP from DRI. Preformed DRl complexes containing biotin-CLIP or biotin-HA (309-319) (2 pmol) were incubated with 50 ug solubilized membranes in the presence or absence of 5 uhf unlabeled MAT (17-31) for 5 hr at 37OC and remaining biotin-peptide/ DRl complexes were measured. (8) Formation of SDS-stable DR3 dimers in the presence of DM. DR3 (2 pmol) was incubated with 25 ug solubilized L or LDM membranes for 5 hr at pH 5. Some samples contained 5 pM HSP(3-13). After pH neutralization and addition of 2% SDS, unheated samples were analyzed by SDS-PAGE and Western blotting with DR-specific antisera. Relative quantities of unstable and stable dimers were determined by densitometry. Percent of each sample in the stable versus unstable form is indicated in parentheses. (C) Enhanced peptide binding to DR3 in the presence of DM. DR3 (0.5 pmol), isolated from transfected mutant T2 cells (Riberdy et al., 1992) was incubated for 5 hr with 1 uM biotin-HSP(3-13) in the presence of 50 ug solubilized L or LDM membranes.

tants is associated with CLIP and these complexes are unstable in SDS (Riberdy et al., 1992). We isolated DR3 from T2 transfectants and found that approximately 50% was unstable in SDS. The fraction of unstable molecules was greatly reduced after a 5 hr incubation in the presence of DM (Figure 48). The shift from unstable to stable dimers in the presence of DM is not dependent on additional pep-

tide. It is likely that peptides or unfolded proteins (Jensen, 1993) contributed by the solubilized membranes bind to DR3 to form stable dimers in the presence of DM. The binding of the peptide antigen, biotin-HSP(3-13) to DR3 was markedly enhanced in the presence of DM (Figure 4C). Thus, our observations support the conclusion that DM facilitates the replacement of physiologically loaded

(D) Activity of DM captured from Tl membranes. Solubilized membranes from Tl or T2 cells were incubated been preloaded with preimmune serum (control) or antiserum specific for the cytoplasmic domain of DMf3. incubated for 3 hr with 2 pmol purified DRl and 1 pM biotin-MAT (17-31) at 37OC and bound peptide was (E) Activity of DM captured from LDM membranes. DM was captured as in (D) and incubated for 3 hr with (17-31).

with protein A-Sepharose that had After washing, the Sepharose was measured. 20 pmol DRl and 1 uM biotin-MAT

Immunity 202

CLIP by antigenic peptide. DM may also enhance peptide binding to mature DRl by inducing the release of other peptides that share physical characteristics with CLIP. Discussion This study demonstrates that DM enhances peptide binding by DR molecules through a mechanism involving the release of previously bound peptides. This is primarily a kinetic effect and only a subpopulation of molecules is susceptible to peptide exchange. Our results suggest that DM acts through a direct interaction with class II molecules because immunoprecipitated DM enhances peptide binding to purified DRl . The presence of DM is clearly required for this activity, because the degree of enhancement is proportional to the concentration of DM, and no enhancement is seen with control immunoprecipitates from DMnegative cells or with preimmune serum. We cannot rule out the presence of another factor, coprecipitating with DM. No proteins have yet been identified that coprecipitate with DMalDM8 isolated from metabolically labeled cells (Denzin et al., 1994; Sanderson et al., 1994). However, the data presented here do not formally exclude the possibility that a coprecipitating factor may be involved. Our results demonstrate that CLIP but not the high affinity peptides, HA (306318) and MAT (17-31) are rapidly released from DRl molecules in the presence of DM. CLIP dissociation is nearly complete after a 5 hr incubation with DM. The weight of evidence suggests that CLIP is associated with class II molecules through the peptide-binding site (Romagnoli and Germain, 1994; Bijlmakers et al., 1994; Malcherek et al., 1995; Sette et al., 1995; Sinigaglia and Hammer, 1995). The rapid peptide loading of DR3 molecules isolated from T2 transfectants can readily be accounted for by a mechanism involving the release of CLIP to generate empty peptide binding sites. Supporting this conclusion, we observed that 65%-800/o of SDSunstable DR3 molecules are converted to stable dimers during a 5 hr incubation with DM. Since CLIP-DR complexes(Mellinset al., 1990; Riberdyet al., 1992)andempty DR molecules (Stern and Wiley, 1992) are unstable in SDS, whereas many other peptide-class II complexes are stable (Germain and Hendrix, 1991) this conversion to stable molecules reflects exchange of CLIP for other polypeptides. The release of CLIP may also account for the effect of DM on peptide binding to DRl molecules isolated from DM-expressing LG2 B lymphoblastoid cells. Despite the expression of DM, a significant fraction of mature DRl molecules in these cells are also associated with CLIP (Urban et al., 1994). We speculate that DM binds to class II molecules and either induces a conformational change or shifts an equilibrium towards a conformation associated with rapid release of CLIP. It is possible that DM interacts only with a subset of class II molecules that assume a specific conformation, which is dependent on the nature of bound peptide. CLIP-bearing DR molecules have a different conformation from DR molecules bearing most other peptides acquired under physiological conditions. Unlikeother peptide complexes, CLIP-DR complexes are unstable in SDS

(Mellins et al., 1990; Riberdy et al., 1992) and have reduced affinity for certain monoclonal antibodies (Mellins et al., 1990, 1991; Morris et al., 1994; Fling et al., 1994), reflecting the absence of specific conformational epitopes. Alternatively, DM may interact with all class II molecules, inducing a conformation that favors the release of selected peptides that have nonoptimal anchor residues. CLIP differs from most peptides in that it is able to bind to many different class II alleles. This promiscuous binding specificity is possible because the peptide has amino acid side chains at key anchor positions that can be accommodated by pockets formed by polymorphic residues in different class II alleles (Malcherek et al., 1995; Sette et al., 1995; Sinigaglia and Hammer, 1995). The lack of allele-specific anchors in the peptide may compromise stability of binding. Several reports demonstrate that CLIP-DR complexes are particularly sensitive to low pH (Riberdy et al., 1992; Urban et al., 1994; Awa and Cresswell, 1994). CLIP, unlike most peptides, is selectively released from DRl molecules during a 24 hr incubation at pH 4 (Urban et al., 1994). This sensitivity to acid-induced dissociation may reflect nonoptimal binding interactions with DR. Peptide binding by purified (Jensen, 1991; Sadegh-Nasseri and Germain, 1991; Reay et al., 1992) and cell surface (Jensen, 1990; Sherman et al., 1994) class II molecules is generally optimal at low pH. Studies with conformation-sensitive probes indicate that class II molecules undergo a definite structural transition at mildly acidic pH, approximating the pH found in acidic endosomal vesicles (H. A. Runnels, J. C. Moore, and P. E. J., submitted). The enhanced peptide binding observed at low pH may, at least in part, be a result of the release of CLIP or other peptides from a subpopulation of class II molecules. The marked pH dependence of peptide binding is preserved when binding is facilitated by DM (Figure 2D). It is possible that the interaction of DM with DR may be influenced by pH-induced changes in the structure of DM or the charge of critical contact residues in DM or DR. Alternatively, the low pH conformer of DR may be a preferred substrate for DM. The latter possibility seems more likely given the documented effect of pH on the structure and function of purified class II molecules. The ability to catalyze the binding of antigenic peptides to CLIP-bearing class II molecules is sufficient to explain the phenotype of mutant cells without postulating other roles previously considered for DM. One need not propose a chaperone function for DM, since a trafficking requirement for class II or antigen has been eliminated in our experimental system. Although acidic pH has been shown to enhance the dissociation of CLIP from DR molecules, the large fraction of molecules associated with CLIP in DM-deficient cells (Riberdy et al., 1992) indicates that the pH encountered in endosomal compartments is not low enough to release CLIP from more than a small fraction of the molecules during the transport of newly synthesized DR to the cell surface. As previously demonstrated, occupancy with CLIP as opposed to other peptides accounts for the large fraction of SDS-unstable molecules in mutant cells (Mellins et al., 1990; Riberdy et al., 1992) and CLIP

DM Enhances 203

Peptide

Binding

blocks the binding of antigenic peptides (Riberdy et al., 1992; Romagnoli and Germain, 1994; Bijlmakers et al., 1994), accounting for the observed deficiency in antigen processing function (Mellins et al., 1990; Riberdy and Cresswell, 1992; Ceman et al., 1992; Morris et al., 1994). Our results help to resolve an apparent contradiction in the kinetics of peptide-class II formation. Purified class II binds peptide extremely slowly, requiring 24-48 hr to reach saturation (Buus et al., 1988; Roche and Cresswell, 1990b). By contrast, peptide loading occurs rapidly under physiological conditions in viable APC (Roosneck et al., 1988; Ceppellini et al., 1989; Davidson et al., 1991; Amigorena et al., 1994; Castellino and Germain, 1995). Although the peptide association rate can be strikingly enhanced at the low pH characteristic of endosomal compartments (Jensen, 1991; Reay et al., 1992) this factor alone cannot account for the enhanced kinetic seen in vivo. In earlier studies using T cell activation as an indirect measure of peptide binding to cell surface class II molecules, we found that binding approached maximal levels within 2-4 hr, raising the possibility that a cofactor present in the APC membrane may facilitate peptide binding to class II molecules (Jensen, 1990, 1991). However, more recent biochemical peptide binding experiments demonstrate that this cofactor is present in intracellular but not cell surface membranes (data not shown). The results of the present study support the conclusion that DM is the cofactor responsible for efficient peptide loading in APC. In mutant cells lacking functional DM, most class II molecules reach the cell surface without binding antigenic peptide because of the slow intrinsic rate of CLIP dissociation and peptide exchange.

AGTAGGTATGGGT-3’; Notl-DMAIR (minus strand, 3’ end of DMA) 5’-CAGCGCGGCCGCTGGCATCAAACTCTGGTCTGGA-3’; KpnlDMB/F (plus strand, 5’ end of DMB) 5%AGTGGTACCATCTTTACAGAGCAGAGCATGATC-3’; Notl-DMBIR (minus strand, d’end of DMB) 5’-CAGCGCGGCCGCTCCTTCTCACTTGGAGTGGAAGTTG-3’. Amplified products of DMA (643 bp) were subcloned into eukaryotic expression vector pRC/RSV (neomycin+) and DMB (657 bp) into pCEP4 (hygromycin B+) (both from lnvitrogen Corporation). DAP-3 cells (2.5 x Iq (a f/r subclone of murine L cells) were transfected with 2.5 ug each DMA and DMB constructs resuspended with 20 ug lipofectin reagent (GIBCO BRL). After 24 hr, double selection was initiated with 500 pg/ml each G416 (GIBCO BRL) and hygromycin B (Sigma). Stable DM transfectants were cloned by limiting dilution and selected for high expression of both DMa and DM6 chains (analyzed by Western blots) and maximal effect on DRl peptide binding. Antlbodlea L243 anti-DR a6 (Lampson and Levy, 1960) and LB3.1 (Stern and Wiley, 1992) were purified from hybridoma culture supernatant using protein A-Sepharose beads (Pharmacia). Rabbi antisera to DMa and DM6 were prepared as follows: Antiserum to the soluble portion of DMa (sDMA) was made by immunizing a rabbit with a recombinant sDMa-GST (glutathione-s-ribosyl transferase) fusion protein. A 701 bp fragment of DMA (from the 5’ end of DMA to the transmembrane region) was amplified from DMA cDNA construct using as forward primer: BamHI, 5’-CCTAGGATCCGTGTGGCAAGAAGGTATGGGT-3’ and reverse primer EcoRI, 5”CGCGGAATTCTTACTCCAGCAGATCTGAGGGCAG-3’, and subcloned into pGEX-2T (Pharmacia) for expression in Escherichia Coli BL2l(D2). sDMa-GST fusion protein was purified from bacterial extracts and used as immunogen. Rabbit sera were extensively absorbed with GST purified from extracts of bacterial cultures transformed with pGEX-PT alone, and used at l/l000 dilution in Western blot analyses. Rabbit anti-DM6 antiserum was made by immunizing with the DM6 C-terminal synthetic peptide, GHSSMPLPGSNYSEGWHIS, conjugated to keyhole limpet hemocyanin. Affinity-purified antibodies were eluted from Sepharose beads conjugated with the same peptide.

Peptides Peptides were synthesized as previously described (Jensen, 1991). MAT (17-31). sequence: SGPLKAEIAQRLEDV; HIV gp41(577-595) sequence: GIKQLQARILAVERYLKDQ; NP7TPQ(535-551), sequence: LDPLIRGLLARPAKLQV; Rabies NS(lOl-120) sequence: GVQIVRQIRSGERFLKIWSQ; HSP(3-13) sequence:KTIAYDEEARR; and CLIP, sequence:LPKPPKPVSKMRMATPLLMGALPM. Peptides were labeled with biotin at the o amino group by reaction with biotin amidocaproate N-hydroxysuccinimide (Jensen, 1991).

Reconetltutlon of DR in Membranes and Llposomes DRl was purified from the Epstein-Barr virus-transformed homozygous LG2 B cell line (DRBl’OlOl) and DR3 was purified from T2 transfectants(Riberdyetal., 1992). usingamonoclonal antibody LB3.1 (Stern and Wiley, 1992) immunoaffinity column as previously described (Jensen, 1991). Purified DR was stored in buffer containing 1% Noctylglucoside at 4OC and analyzed by SDS-PAGE and Coomassie bluestaining. Protein was measured by bicinchoninic acid (BCA) assay (Pierce, Rockford, Illinois). Membranes were prepared from Tl and T2 cells or trypsinized L cells by hypotonic lysis (10 mM Tris [pH 6.01) in the presence of protease inhibitors: 5 mM phenanthroline, 4 mM PMSF, 15 mM pepstatin, 566 mM TPCK, 1 mM benzamidine, 1 mM iodoacetamide, 6 mM NEM, 10 mM leupeptin, 270 mM TLCK, and 10 mM EDTA as previously described (all available from Sigma Chemical Company, St. Louis, Missouri). Suspensions were cleared of debris and nuclei by centrifugation at 3500 x g for IO min at 4OC, and membranes pelleted from supernatants by centrifugation at 22,000 x g for 30 min at 4OC. Membranes were solubilized in 0.5% s-octyi glycopyranoside in phosphatebuffered saline (PBS) with protease inhibitors in the presence or absence of 0.4 uM purified DRl. The solubilized membranes were extensively dialyzed against PBS to form liposomes. They were used directly in binding assays. Protein content was measured by BCA assay.

Constructlon and Expression of HLA-DMA and HLA-DMB cDNA Expression Vectors Total RNA was prepared from LG2 lymphoblastoid cells using TRlzol reagent (GIBCO BRL). cDNA was synthesized by reverse transcriptase-polymerase chain reaction (RT-PCR) using GIBCO BRL reagent kit. Full-length HLA-DMA and HLA-DMB were amplified from cDNA by PCR using Taq polymerase (Boehringer-Mannheim Corporation, Indianapolis, Indiana) and standard PCR conditions. The primers used to amplify cDNA clones were designed with a restriction site for cloning: Spel-DMA/F (plus strand, 5’ end of DMA) 5’-CTGTGTGACT-

Peptida Bindlng Assay Peptide-class II binding assay has been previously described (Sherman et al., 1994). In brief, assays were performed in triplicate and reproduced at least three times. Each sample was comprised of liposomes containing 50 ug membrane protein and 20 pmol DRl (if indicated). The liposomes were incubated with 1 mM biotin-peptide in 0.1 M citrate/phosphate buffer (pH 5) and protease inhibitors for 3 hr at 37OC. An equal volume of solubilization buffer (1% NP-40, 0.1% Tween 20. 0.1 M NaCI, 400 mM Tris [pH 7.51, 1 mglml bovine serum albumin) was added to each sample before incubation on ice for 30

Experimental

Procedures

call Culture Tl and T2 human (.174 x CEM. Tl and ,174 x CEM. T2) hybrids (Salter et al., 1965) were cultured in DMEM containing 10% fetal calf serum, 2 mM L-glutamine, 2 mM sodium pyruvate, and the antibiotics penicillin-streptomycin and gentamicin. L cells (DAP-3, t/r subclone of murine L cells) were cultured in RPM1 containing the same additives and 50 mM 2-ME (RIO). Cells were maintained at 37OC in 7% CO*. T2-DR3 cells (Riberdy et al., 1992) were a gift of P. Cresswell (Yale University, New Haven, Connecticut). Media and additives were purchased from GIBCO BRL (Grand Island, New York).

Immunity 204

min. The samples were added to assay plates coated with the anti-DR antibody L-243. After washing, excess europium-avidin was added and fluorescence was read at 612 nm as described (Sherman et al., 1994). Results, expressed as fluorescence counts per second (cps x lO-3), are comparable within each assay but not between assays.

lmmunoblot

Analysis

Western blots were performed as described previously (Sherman et al., 1994) using donkey anti-rabbit horseradish peroxidase and chemiluminescent substrate (Amersham). Quantification of bands in immunoblots was accomplished by optical densitometry using a high resolution CCD camera and the Macintosh-based software, Image (W. Rasband). Using Image, optical density values were determined for each band. The black level was adjusted to ensure linearity of the densitometry analysis.

Peptlde

Blnding

Assay

In the Presence

of Captured

DY

Membranes were solubilized in PBS containing 0.5% NP-40 and protease inhibitors. Solubilized membranes were added to tubes containing purified DRl and peptide in buffer with a final concentration of 0.5% NP-40 and 0.1 M citrate/phosphate (pH 5). Protein A-Sepharose (25 @ample) was incubated with 25 pl serum for 1 hr at 4OC and washed twice with PBS followed by incubation with 100 pg membrane protein for 1 hr at 4OC. After further washes, the Sepharose was incubated with purified DRI and peptide in 50 pl volumescontaining 0.2% NP-40, 0.1 M citrate/phosphate buffer (pH 5) and protease inhibitors.

Releass

of CLIP from

Preformed

DRl-CLIP

Complexes

Preformed biotin-peptide/DRl complexes were made by incubating 6.6 uM DRl with 60 uM biotin-CLIP (LPKPPKPVSKMRMATPLLMGALPM) or biotin-HA (306-318) (Stern and Wiley, 1992) for 72 hr at 37OC. Unbound peptide was removed by extensive dialysis. The complexes were used in peptide binding assays with L or LDM membranes as described above.

Acknowledgments We thank C. Gutekunst, R. advice and P. Cresswell for Spotts and J. Moore provided was supported by research Health. M. A. S. is a Howard Received

June

Hendrix, and J. W. Nichols for helpful generously providing T2-DR3 cells. E. excellent technical assistance. This work grants from the National Institutes of Hughes Predoctoral Fellow.

16, 1995; revised

June

J. R., Webster, P., and Mellman, I. (1994). of new class II MHC molecules in a novel in B lymphocytes. Nature 369, 113-120.

Anderson, M. S., and Miller, J. (1992). Invariant chain can function as a chaperone protein for class II major histocompatibility complex molecules. Proc. Natl. Acad. Sci. USA 89, 2262-2266. Awa, R. R., and Cresswell, P. (1994). and dissociation of HLA-DR complexes peptides. Immunity 1, 763-774. Bakke, O., and Dobberstein, ant chain contains a sorting 83, 707-716.

In vivo and in vitro formation with invariant chain-derived

8. (1990). MHC class k-associated signal for endosomal compartments.

invariCell

Bijlmakers, M. J., Benaroch, P., and Ploegh, H. L. (1994). Mapping functional regions in the lumenal domain of the class II-associated invariant chain. J. Exp. Med. 180, 623-629. Buus, S., Sette, A., Colon, S. M., Jenis, D. M., and Grey, H. M. (1966). Isolation and characterization of antigen-la complexes involved in T cell recognition. Cell 47, 1071-1077.

conformation and impaired 149, 754-761.

anti-

Ceppellini, R., Frumento, G., Ferrara, G. B., Tosi, R.. Chersi. A., and Pernis, 8. (1969). Binding of labelled influenza matrix peptide to HLA DR in living B lymphoid cells. Nature 339, 392394. Chicz, R. M., Urban, R. G., Lane, W. S., Gorga, J. C., Stern, L. J., Vignali, D. A., and Strominger, J. L. (1992). Predominant naturally processed peptides bound to HLA-DRl are derived from MHC-related molecules and are heterogeneous in size. Nature 358, 764-766. Davidson, H. W., Reid, P. A., Lanzavecchia. A., and Watts, C. (1991). Processed antigen binds to newly synthesized MHC class II molecules in antigen-specific B lymphocytes. Cell 67, 105-l 16. DeMars, R., Chang, C. C., Shaw, S., Reitnauer, P. J., and Sondel, P. M. (1964). Homozygous deletions that simultaneously eliminate expressionsof class I and class II antigensof EBV-transformed B-lymphoblastoid cells. I. Reduced proliferative responses of autologous and allogeneic T cells to mutant cells that have decreased expression of class II antigens. Hum. Immunol. II, 77. Denzin, L. K., Robbins, N. F., Carboy-Newcomb, C., and Cresswell, P. (1994). Assembly and intracellular transport of HLA-DM and correction of the class II antigen-processing defect in T2 cells. Immunity 7, 5x95-606. Erlich, H.. Lee, J. S., Petersen, J. W., Bugawan, T., and DeMars, R. (1966). Molecular analysis of HLA class I and class II antigen loss mutants reveals a homozygous deletion of the DR. W, and part of the DP region: implications for class II gene order. Hum. Immunol. 16, 205-219. Fling, S. P., Arp, B., and Pious, D. (1994). HtA-DMA are both required for MHC class Il/peptide complex gen-presenting cells. Nature 368, 554-556. Germain, occupancy reticulum

and -DMB genes formation in anti-

R. N., and Hendrix, L. R. (1991). MHC class II structure, and surface expression determined by post-endoplasmic antigen binding. Nature 353, 134-139.

Guagliardi, L. E., Koppelman, B., Blum, J. S., Marks, M. S., Cresswell, P., and Brodsky, F. M. (1990). Co-localization of molecules involved in antigen processing and presentation in an early endocytic compartment. Nature 343, 133-139. Harding, C. V., and Geuze, U. J. (1992). Class II moleculesare present in macrophage lysosomes and phagolysosomes that function in the phagocytic processing of listeria monocytogenes for presentation to T cells. J. Cell. Biol. 179, 531-542.

Jensen, P. E. (1990). Regulation J. Exp. Med. 171, 1779-1769.

of antigen

presentation

by acidic pH.

Jensen, P. E. (1991). Enhanced class II major histocompatibility Med. 774, 1111-1120.

binding of peptide antigen to purified glycoproteins at acidic pH. J. Exp.

Jensen, P. E. (1993). Acidification and disulfide reduction can be sufficient to allow intact proteins to bind class II MHC. J. Immunol. 150, 3347-3356. Karlsson, L., Peleraux, A., Lindstedt, R., Liljedahl, M., and Peterson, P. A. (1994). Reconstitution of an operational MHC class II compartment in nonantigen-presenting cells. Science 266, 1569-1573. Lamb, C. A., Yewdell, J. W.. Bennink, J. R., and Cresswell, P. (1991). Invariant chain targets HLA class II molecules to acidic endosomes containing internalized influenza virus. Proc. Natl. Acad. Sci. USA 88, 5996-6002. Lampson, L. A., and Levy, R. (1980). Two populations of la-like molecules on a human cell line. J. Immunol. 125, 293-299.

of MHC and apImmu-

Lotteau, V., Teyton, L., Peleraux, A., Nilsson, T.. Karlsson, L., Schmid, S. L., Quaranta, V., and Peterson, P. A. (1990). Intracellular transport of class II MHC molecules directed by invariant chain. Nature 348. 600-605.

R.. Long, E. O., and DeMars, R. (1992). MHC expresses normal levels of transgene encoded

Malcherek, G., Gnau, V., Jung, G., Rammensee, H.-G., and Melms, A. (1995). Supermotifs enable natural invariant chain-derived peptides to interact with many major histocompatibility complex+lass II mole-

Castellino, F., and Germain, R. N. (1995). Extensive trafficking class II-invariant chain complexes in the endocytic pathway pearance of peptide-loaded class II in multiple compartments. nity 2. 73-66. Ceman, S., Rudersdorf, class II deletion mutant

that have abnormal ability. J. Immunol.

Hunt, D. F., Michel, H., Dickinson, T. A., Shabanowitz. J., Cox, A. L., Sakaguchi, K., Appella, E., Grey, H. M., and Sette, A. (1992). Peptides presented to the immune system by the murine class II major histocompatibility complex molecule I-Ad. Science 258, 1617-I 620.

23, 1995.

References Amigorena, S., Drake, Transient accumulation endocytic compartment

class II molecules gen presentation

DM Enhances 205

cutes.

Peptide

J. Exp. Med.

Binding

787, 527-536.

determining

Marks, M. S., Blum, J. S., and Cresswell, P. (1990). Invariant chain trimers are sequestered in the rough endoplasmic reticulum in the absence of association with HLA class II antigens. J. Cell Biol. 777, 639-855. Marsh, E. W., Dalke, D. P., and Pierce, S. K. (1992). Biochemical evidence for the rapid assembly and disassembly of processed antigen-major histocompatibility complex class II complexes in acidic vesicles of B cells. J. Exp. Med. 775, 425-436.

MHC class

II structure.

Nature

353, 167-170.

Salter, R. D., Howell, D. N., andcresswell, P. (1965). Genes regulating HLA Class I antigen expression in T-B lymphoblast hybrids. Immunogenetics 27, 235-246. Sanderson, F., Kleijmeer, M. J.. Kelly, A.. Verwoerd, D., Tulp, A., Neefjes, J. J., Geuze, H. J., and Trowsdale. J. (1994). Accumulation of HLA-DM, a regulator of antigen presentation, in MHC class II compartments Science 268, 1566-I 569.

Mellins, E., Smith, L., Arp, 6.. Cotner, T., Celis, E., and Pious, D. (1990). Defective processing and presentation of exogenous antigens in mutants with normal HLA class II genes. Nature 343, 71-74.

Sette, A., Southwood, S., Miller, J., and Appella. E. (1995). of major histocompatibility complex class II to the invariant derived peptide, CLIP, is regulated by allelic polymorphism II. J. Exp. Med. 787, 677-683.

Mellins, E., Kempin, S., Smith, L., Monji, T., and Pious, D. (1991). A gene required for class II-restricted antigen presentation maps to the major histocompatibility complex. J. Exp. Med. 174, 1607-1615.

Sherman, M. A., Runnels, H. A., Moore, J. C., Stern, L. J., and Jensen, P.E. (1994). Membrane interactions influence the peptide binding behavior of DRI. J. Exp. Med. 779, 229.

Morris, P., Shaman, J., Attaya, M., Amaya, M., Goodman, S., Bergman, C., Monaco, J. J., and Mellins, E. (1994). An essential role for HLA-DM in antigen presentation by class II major histocompatibility molecules. Nature 368, 551-554.

Sinigaglia, F., and Hammer, J. (1995). Motifs and supermotifs class II binding peptides. J. Exp. Med. 787, 449-451.

Neefjes, J. J., Stollorz, V., Peters, P. J., Geuze, H. J., and Ploegh, H.L. (1990). The biosynthetic pathway of MHC class II but not class I molecules intersects the endocytic route. Cell 67, 171-163. Newcomb, J. Ft., and Cresswell, nous peptides bound to purified from invariant chain-associated 499-507. Peterson, nological 174.

P. (1993). Characterization of endogeHLA-DR molecules and their absence alpha beta dimers. J. Immunol. 750,

M.. and Miller, J. (1990). Invariantchain influencestheimmurecognition of MHC class II molecules. Nature 345, 172-

Qiu, Y., Xu, X., Wandinger Ness, A., Dalke, D. P., and Pierce, S.K. (1994). Separation of subcellular compartments containing distinct functional forms of MHC class II. J. Cell Biol., 725, 595605. Reay, P. A., Wettstein, D.A., and Davis, M. M. (1992). pH dependence and exchange of high and low responder peptides binding to a class II MHC molecule. EMBO J. 77, 2629-2639. Riberdy, J. M., and Cresswell, P. (1992). The antigen-processing mutant T2 suggests a role for MHC-linked genes in class II antigen presentation. J. Immunol. 748, 2586-2590. Riberdy, J. M., Newcomb. J. R., Surman, M. J., Barbosa, J. A., and Cresswell, P. (1992). HLA-DR molecules from an antigen-processing mutant cell line are associated with invariant chain peptides. Nature 360, 474-477. Roche, P. A., and Cresswell, P. (199Oa). Invariant chain association with HLA-DR molecules inhibits immunogenic peptide binding. Nature 345, 615-616. Roche, P. A., and Cresswell, influenza hemagglutininderived ncr. 744, 1649-1666.

P. (1990b). High-affinity binding of an peptide to purified HlA-DR. J. Immu-

Roche, P. A., and Cresswell, P. (1991). Proteolysis of the class IIassociated invariant chain generates a peptide binding site in intracellular HlA-DR molecules. Proc. Natl. Acad. Sci. USA 88, 3150-3154. Roche, P. A., Marks, M. S., and Cresswell, P. (1991). Formation of a nine-subunit complex by HLA class II glycoproteins and the invariant chain. Nature 354, 392394. Roche, P. A., Teletski, C. L., Karp, D. R., Pinet, V., Bakke, O., and Long, E. 0. (1992). Stable surface expression of invariant chain prevents peptide presentation by HLA-DR. EMBCI J. 7 7, 2641-2647. Romagnoli, P., andGermain, R. N. (1994). TheCLIP regionof invariant chain plays a critical role in regulating major histocompatibility complex class II folding, transport, and peptide occupancy. J. Exp. Med. 780, 1107-1113. Roosneck, E., Demotz. S., Corradin, Kinetics of MHC-antigen complex cells. J. Immunol. 740, 4079-4070.

G., and Lanzavecchia. A. (1986). formation on antigen-presenting

Rudensky. A., Preston Hurlburt, P.. Hong, S. C., Barlow, A., and Janeway, C. A., Jr. (1991). Sequence analysis of peptides bound to MHC class II molecules. Nature 353, 622-627. Sadegh-Nasseri,

S., and Germain,

R. N. (1991).

A role for peptide

in

Stern, L. J., and Wiley, D. C. (1992). The Human HLA-DRl assembles as empty a6 heterodimers antigenic peptide. Cell 88, 465-477.

Binding chainin class

for MHC

class II MHC protein in the absence of

Teyton, L., O’Sullivan, D., Dickson, P. W., Lotteau, V.. Sette, A., Fink, P., and Peterson, P. A. (1990). Invariant chain distinguishes between the exogenous and endogenous antigen presentation pathways. Nature 348, 39-44. Tulp, A., Verwoerd, D., Dobberstein, 8.. Ploegh, H. L., and Pieters, J. (1994). Isolation and characterization of the intracellular MHC class II compartment. Nature 389, 120-126. Urban, R. G., Chicz, R. M., and Strominger, J. L. (1994). Selective release of some invariant chain-derived peptides from HLA-DRl molecules at endosomal pH. J. Exp. Med. 780, 751-755. West, M. A., Lucocq, J. M., and Watts, C. (1994). Antigen processing and class II MHC peptidbloading compartments in human B-lymphoblastoid cells. Nature 369, 147-I 51.