HLA-DO is a negative modulator of HLA-DM-mediated MHC class II peptide loading

950

Research Paper

HLA-DO is a negative modulator of HLA-DM-mediated MHC class II peptide loading S.M. van Ham*, E.P.M. Tjin*, B.F. Lillemeier†, U. Grüneberg†, K.E. van Meijgaarden‡, L. Pastoors*, D. Verwoerd*, A. Tulp*, B. Canas§, D. Rahman§, T.H.M. Ottenhoff‡, D.J.C. Pappin§, J. Trowsdale¶ and J. Neefjes* Background: Class II molecules of the major histocompatibility complex become loaded with antigenic peptides after dissociation of invariant chainderived peptides (CLIP) from the peptide-binding groove. The human leukocyte antigen (HLA)-DM is a prerequisite for this process, which takes place in specialised intracellular compartments. HLA-DM catalyses the peptideexchange process, simultaneously functioning as a peptide ‘editor’, favouring the presentation of stably binding peptides. Recently, HLA-DO, an unconventional class II molecule, has been found associated with HLA-DM in B cells, yet its function has remained elusive. Results: The function of the HLA-DO complex was investigated by expression of both chains of the HLA-DO heterodimer (either alone or fused to green fluorescent protein) in human Mel JuSo cells. Expression of HLA-DO resulted in greatly enhanced surface expression of CLIP via HLA-DR3, the conversion of class II complexes to the SDS-unstable phenotype and reduced antigen presentation to T-cell clones. Analysis of peptides eluted from HLA-DR3 demonstrated that CLIP was the major peptide bound to class II in the HLADO transfectants. Peptide exchange assays in vitro revealed that HLA-DO functions directly at the level of class II peptide loading by inhibiting the catalytic action of HLA-DM. Conclusions: HLA-DO is a negative modulator of HLA-DM. By stably associating with HLA-DM, the catalytic action of HLA-DM on class II peptide loading is inhibited. HLA-DO thus affects the peptide repertoire that is eventually presented to the immune system by MHC class II molecules.

Background Antigen presentation, a key event in the functioning of the immune system, is a well-controlled process, regulated by a number of independent factors. Antigens derived from exogenous sources, such as pathogenic bacteria, are presented to the immune system after intracellular degradation and association with class II molecules of the major histocompatibility complex (MHC) [1]. The association step occurs in specialised intracellular compartments, termed MIIC (MHC class II compartments) [2], to which the class II heterodimers are directed from the endoplasmic reticulum (ER) after association with a third glycoprotein, the invariant chain (Ii) [3]. During transport, Ii is degraded by proteases like cathepsin S [4], until only a nested set of peptides, termed CLIP for class II-associated invariant chain peptides, remains bound to the peptide-binding groove of the class II molecules [5,6]. Although spontaneous exchange of CLIP for antigenic peptides might occur in the MIICs and is favoured by the

Addresses: *Department of Cellular Biochemistry, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands. †Human Immunogenetics Laboratory, Imperial Cancer Research Fund, 44 Lincoln’s Inn Fields, London WC2A 3PX, UK. ‡Department of Immunohaematology and Bloodbank, Leiden University Medical Center, Leiden, the Netherlands. §Protein Sequencing Laboratory, Imperial Cancer Research Fund, 44 Lincoln’s Inn Fields, London WC2A 3PX, UK. ¶Department of Pathology, University of Cambridge, Tennis Court Road, Cambridge CB2 1QP, UK. Correspondence: S.M. van Ham E-mail: [email protected] Received: 8 October 1997 Revised: 23 October 1997 Accepted: 23 October 1997 Published: 7 November 1997 Current Biology 1997, 7:950–957 http://biomednet.com/elecref/0960982200700950 © Current Biology Ltd ISSN 0960-9822

lysosomal-like pH of these compartments [7,8], an additional factor is essential for this process. Mice or cell lines carrying a mutation in the genes encoding the heterodimer human leukocyte antigen (HLA)-DM are defective in antigen presentation and express predominantly class II molecules that still contain CLIP [9–12]. HLADM is localised in the MIICs, where it catalyses peptide exchange by enhancing the dissociation of CLIP and stabilising the peptide-free state of the class II molecules. Simultaneously, HLA-DM functions as a peptide ‘editor’ in that its catalytic action promotes dissociation of peptides that do not bind stably to the class II backbone, thus skewing the peptide repertoire that is eventually expressed at the cell surface in the direction of stably binding antigenic peptides [13–21]. Recently, another unconventional MHC molecule, HLADO, has been found tightly associated with HLA-DM in B cells [22]. HLA-DO, a heterodimeric complex of the HLA-DOα (formerly termed HLA-DNα or HLA-DZα)

Research Paper HLA-DO is negative modulator of HLA-DM van Ham et al.

[23] and HLA-DOβ chains [24], localises to the MIICs upon association with HLA-DM [22]. In mice lacking H2-M, the mouse equivalent of HLA-DM, or in human cell lines bearing a mutation in HLA-DM, HLA-DO expression is lowered and confined to the ER, demonstrating that HLA-DO requires HLA-DM association for efficient egress from the ER ([22]; C. Thomas, C. Roucard, J.T. and S.M.v.H., unpublished observations). Although the association between HLA-DM and HLADO might indicate that the functions of both complexes are linked, the action of HLA-DO has remained as yet unknown. The function of HLA-DO was investigated by introducing HLA-DOαβ or HLA-DOαβ coupled to green fluorescent protein (HLA-DOαβ–GFP) [25] into cell lines that express functional class II and HLA-DM molecules but lack the expression of HLA-DO molecules. Expression of either form of HLA-DO markedly increased the amount of class II molecules at the cell surface that still contained CLIP and greatly reduced the amount of class II molecules associated with antigenic peptides. Consequently, HLA-DO expression hampered antigen presentation to T cells. Peptide-exchange assays in vitro showed that HLA-DO binding to HLA-DM interfered with class II peptide exchange by directly inhibiting the catalytic action of HLA-DM.

951

Results and discussion The function of HLA-DO was explored by transfection of HLA-DOα [23] and HLA-DOβ [24] cDNA expression constructs into Mel JuSo cells. These cells stably express the components needed for antigen presentation by class II molecules, including Ii, HLA-DM and HLA-DR3 [21], but no HLA-DO could be detected by immunoblotting analysis (see later). (a) DOβ–GFP

DRα

10µm

(b) DOαβ–GFP

DRα

DOαβ–GFP

DMα

DOαβ–GFP

CD63

Figure 1

10µm

28 kDa

Lysate 2

Anti-GFP

Total lysate

(c) Lysate 2

Confocal and western-blot analysis of HLA-DOβ–GFP (DOβ–GFP) single and HLA-DOαβ–GFP double transfectants. (a) Confocal analysis demonstrates that the HLA-DOβ–GFP single chain is retained in the ER. HLA-DOβ–GFP transfectants were fixed with methanol and stained with an antibody against class II HLA-DRα chain. Fluorescence from GFP is shown in the left panel, whereas HLA-DRα staining (DRα) is shown in the right panel. The middle panel is a overlay of the left and right images. HLA-DOβ–GFP stains the nuclear envelope and a fine reticular network, whereas class II is seen in perinuclear vesicles that do not co-localise with HLA-DOβ–GFP. (b) HLA-DOαβ–GFP is observed in lysosomal vesicles containing HLA-DM and class II upon confocal analysis. Double transfectants were fixed and stained with antibodies against HLA-DRα, HLA-DM or CD63, as indicated (DRα, DMα and CD63, respectively). The left panel is the GFP fluorescence; the right panel shows the staining for the other proteins, and the middle panel represents the overlay, with co-localisation shown as yellow. (c) Cell lysates of HLA-DOαβ–GFP transfectants were analysed by SDS–PAGE before (total lysate) or after various rounds of immunoprecipitation with anti-GFP antibodies (lysate 2). The first immuno-isolate was loaded as well (lane Anti-GFP). Proteins were then transferred to nitrocellulose and the filter was treated with the antiHLA-DMα antibody 5C1. Two closely migrating bands were visualised in the total lysate, the lower band corresponding to the high-mannose carbohydrate containing HLA-DMα chain (the ER form) and the higher band corresponding to the most abundant, complex-carbohydratecontaining HLA-DMα chain (transported through the Golgi to the MIIC). The upper band was quantitatively removed by the anti-GFP isolations, whereas the lower band mostly remained in the lysate fractions, probably because it was not fully associated with HLADOαβ–GFP in the ER.

HLA–DMα

Current Biology

952

Current Biology, Vol 7 No 12

Transfection of an HLA-DOβ–GFP cDNA expression construct into Mel JuSo cells resulted only in a fine reticular staining that resembled staining for the ER (Figure 1a): double-labeling for class II expression revealed that class II molecules were transported to perinuclear vesicles as normal and did not co-localise with the free HLADOβ–GFP chain. We next transfected the cells with cDNA expression vectors for HLA-DOα and HLADOβ–GFP [25], to verify that HLA-DO was expressed in all cells. Confocal analysis of clones stably expressing HLA-DOαβ–GFP showed that the HLA-DOαβ–GFP heterodimer co-localised with HLA-DR3 and HLA-DM in acidic, CD63-positive MIIC vesicles (Figure 1b). Apparently, both α and β chains of HLA-DO need to be expressed for correct intracellular transport. Two-dimensional electrophoretic analysis of immunoprecipitations with an anti-GFP serum demonstrated the formation of a complex between HLA-DOβ–GFP and HLA-DOα, and association of HLA-DO with the HLA-DM complex, consistent with previous findings [22]. In contrast to the findings in B cells (data not shown), depletion experiments indicated that HLA-DM was quantitatively associated with HLA-DOαβ–GFP in the Mel JuSo transfectants (Figure 1c). To investigate the differential effects of HLA-DO expression on peptide presentation by class II molecules, flow cytometric analyses were performed. Clones expressing HLA-DOαβ showed a similar expression of HLA-DR3 as control transfectants, but with greatly enhanced surface presentation of CLIP (Figure 2a). Fluorescence-activated cell sorter (FACS) analyses of Mel JuSO cells transfected with both empty vectors showed low levels of surface expression of class II-associated CLIP (Figure 2b). In contrast, HLA-DOαβ–GFP transfectants showed highly enhanced CLIP presentation. As the cloned transfectants lost HLA-DOαβ–GFP expression, they concomitantly lost CLIP surface expression, which dropped to a level similar to that found in control transfectants; however, class II expression levels remained unaffected (Figure 2b). These findings point to an antagonistic effect of HLA-DO on release of CLIP from the HLA-DR3 molecule, resulting in a similar phenotype as in HLA-DM-negative cells [26,27]. Stability of complexes in SDS can be used as to measure the level of binding of antigenic peptides, but not of CLIP, to HLA-DR3 molecules [27]. When compared with control cells, a FACS-sorted, HLA-DOαβ–GFP-positive population showed a large reduction in the amount of SDS-stable HLA-DR3 complexes by western-blot analyses, with the total level of HLA-DR3 expression remaining unchanged (Figure 3a). This reduction did not result from a decreased level of HLA-DM expression in the HLA-DOαβ–GFP transfectants (Figure 3a). The amount of SDS-stable class II complexes was also greatly reduced in the HLA-DOαβ transfectant (Figure 3b), indicating

Figure 2

Expression of the HLA-DO complex in Mel JuSo cells enhances cellsurface expression of HLA-DR3-associated CLIP. (a) FACS analysis of 5,000 cloned Mel JuSo cells transfected with vectors only (DOαβ–) or HLA-DOαβ (DOαβ+) showing background staining using secondary PE-conjugated antibodies only (dotted lines), class II-specific staining using monoclonal antibody L243 [40] and CLIP-specific staining using CerCLIP.1 [41]. Histogram staining profiles of live gated cells are shown. (b) FACS analysis of cloned transfectants with vectors only (DOαβ–GFP–) or HLA-DOαβ–GFP (DOαβ–GFP+) showing background staining only (–), class II-specific or CLIP-specific staining. The vertical axis represents GFP fluorescence and the horizontal axis represents phycoerythrin (PE) fluorescence, each in arbitrary units on a logarithmic scale.

that the GFP tag did not adversely affect the function of HLA-DO, as also seen in Figure 2. The effect of HLA-DO expression was examined further by comparison of peptides eluted from HLA-DR3 molecules that had been affinity-purified from FACS-sorted, HLA-DOαβ–GFP transfectants with peptides obtained in a similar way from HLA-DO-negative cells. The HPLC (high pressure liquid chromatography) profile of the eluted

Research Paper HLA-DO is negative modulator of HLA-DM van Ham et al.

class II peptides was clearly altered upon HLADOαβ–GFP expression, with the appearance of a set of peaks eluting with 30–35% acetonitrile (Figure 4a). This set of peaks strongly resembled the HPLC pattern observed for HLA-DM-negative cells [26,27]. Analysis of the peptide content of these peaks by matrix-assisted laser desorption (MALD) time-of-flight mass spectrometry (MS) [28] identified peptides of 21–24 residues in length, that were present in relatively high amounts (Figure 4b). The peptide masses corresponded to three CLIP variants [26,27]. This finding was confirmed by digesting the peptides with trypsin, derivatisation with SPA (n-succinimidyl2 (3-pyridyl) acetate) and sequencing of the carboxy-terminal tryptic fragments by low-energy collisioninduced dissociation (CID) tandem MS (Figure 4b) [29,30]. In the corresponding fractions in the HLA-DOnegative control, only minute amounts of these CLIP peptides could be detected (Figure 4b), with the concomitant appearance of small amounts of numerous other peptides. In the HLA-DOαβ–GFP transfectant, other peptides besides CLIP peptides were also detected, but in much lower amounts and mainly of different masses than the peptides detected in the control transfectant, consistent with the observation that HLA-DO expression reduces, but does not abolish formation of SDS-stable class II complexes. Together, these data suggest that the majority of HLA-DR3 molecules are complexed to CLIP upon quantitative association of HLA-DOαβ–GFP and HLA-DM. We next addressed the question of the mode of action by which HLA-DO influences class II peptide presentation. Expression of HLA-DO did not alter the intracellular distribution of HLA-DM and class II molecules (as determined by subcellular fractionation experiments), nor did it change the maturation rate of SDS-stable class II complexes (data not shown). Peptide association and dissociation studies in vitro were performed on HLA-DR3–CLIP molecules isolated from T2.DR3 cells [31] using cell lysates from FACS-sorted HLA-DOαβ–GFP-expressing cells or control transfectants. Addition of lysates containing HLA-DM without HLA-DO catalysed the dissociation of the biotinylated full-length CLIP(81–104) peptide [32] from HLA-DR3 molecules, but not of the stably bound ApoB(2877–2894) peptide [32] (Figure 5a), in agreement with previous findings [16]. Addition of lysates from transfectants expressing HLA-DO were catalytically defective for CLIP(81–104) dissociation, despite containing similar quantities of HLA-DM as control transfectants. Moreover, whereas HLA-DM lysates greatly enhanced exchange of free biotinylated CLIP(81–104) or ApoB(2877–2894) peptides on purified HLA-DR3–CLIP complexes, this catalytic effect was greatly reduced by addition of HLA-DO–HLA-DM lysates (Figure 5b). These experiments were repeated with lysates from purified MIIC compartments containing equal amounts of HLA-DM and HLA-DM–HLA-DO complexes (Figure 5c, inset). These

953

Figure 3 (a) DOαβ– GFP

– nb

b

nb

DOαβ– – GFP

–

DOαβ– GFP

b

68 αβ 43 α 28

DOβ–GFP

Anti-DR3α (1B5)

Anti-DMα (5C1)

Anti-DOβ–GFP (αGFP)

(b) – b

DOαβ– nb

b

–

DO αβ

–

DOαβ

nb

68 αβ 43 α 28

DOβ Anti-DR3α (1B5)

Anti-DMα (5C1)

Anti-DOβ

Current Biology

Reduction of amount of SDS-stable HLA-DR3 complexes upon HLADO expression. (a,b) Immunoblotting of whole cell lysates of 1 × 106 Mel JuSo cells transfected with vectors only (–) or with HLADOαβ–GFP (DOαβ–GFP) or HLA-DOαβ (DOαβ), followed by detection of the total amount of HLA-DMα (DMα), HLA-DOβ–GFP (DOβ–GFP) and SDS-stable HLA-DR3α (DR3α) complexes. The HLA-DR3α and HLA-DR3β chains of SDS-stable HLA-DR3 dimers remain associated in SDS-containing sample buffer under non-boiling (nb) conditions, but dissociate upon boiling (b) conditions. Arrowheads indicate the position of the HLA-DR3αβ dimers (αβ) or DR3α monomer (α). HLA-DMα and HLA-DOβ–GFP were detected upon boiling of the samples. The position of the HLA-DOβ–GFP protein is indicated, with additional bands representing HLA-DO-unrelated, cross-reactive species. Molecular sizes are indicated on the left (in kDa). For detection of HLA-DMα, the HLA-DMα-specific monoclonal antibody 5C1 [34] was used and HLA-DR3 complexes were detected using the HLA-DR3α-specific monoclonal antibody 1B5 [42].

results demonstrated that HLA-DM–HLA-DO complexes from MIICs were 70–80% less efficient in the formation of HLA-DR3–ApoB(2877–2894) complexes than were the HLA-DM complexes alone (Figure 5c). We tentatively conclude that HLA-DO affects class II peptide presentation by acting directly as a negative modulator of the chaperone HLA-DM, resulting in an increase in CLIP bound to HLA-DR3 and reduced stability of class II molecules. If HLA-DO association to HLA-DM is negatively modulating HLA-DM action, presentation of antigen to proliferative T cells should be reduced as well. The antigen-presenting capacities of control and HLADOαβ–GFP-expressing Mel JuSo cells were assessed by pulsing the cells with recombinant hsp65 protein of

954

Current Biology, Vol 7 No 12

Figure 4 DO–

DO αβ–GFP (a)

0.3

Absorbance 214 nm

Absorbance 214 nm

F3

F2

F1

0.3

F6 F4

0.1 32 (b)

F5

0.1 34

Time (min)

32

Relative intensity 2334.8

34

Time (min)

Relative intensity F1

1300

F4

2000 1000

1100 2000 3000 Mass (Da) KPPKPVSKMR MATPLLMQALP Relative intensity 2545.2

2000

3000

Relative intensity 1700

F2

1000

Mass (Da)

Expression of HLA-DOαβ–GFP changes the peptide repertoire that is bound to HLA-DR3. (a) HPLC profiles of acid-eluted peptides from HLA-DOαβ–GFP-positive cells (DOαβ–GFP) and cells transfected with vectors only (DO–). The absorbance of specific eluate fractions at 214 nm is shown against their retention time after start of the gradient. The peaks labelled F1–3 for HLADOαβ–GFP correspond to peaks labelled F4–6 respectively for HLA-DO–. (b) Analysis of fractions F1–6 by MALD time-of-flight MS, together with identified masses and amino acid sequences of major HLA-DOαβ–GFP peptides. The relative intensities of CLIPs may be assessed by comparison to the 100 fmol peak of [Glu] fibrinopeptide B (indicated by the asterisk), added as an internal mass calibrant. The underlined regions of sequence shown for F1–3 were fully sequenced by lowenergy tandem MS. The arrows indicate the position of the almost absent corresponding peptide in HLA-DO– samples.

F5

750 1300 2000 3000 Mass (Da) LPKPPKPVSKMR MATPLLMQALP Relative intensity 1000

2000

3000

Mass (Da)

Relative intensity

2676.3 F3

500

1250

F6

750 2000 3000 Mass (Da) LPKPPKPVSKMR MATPLLMQALP

2000

3000

Mass (Da)

Current Biology

Mycobacterium tuberculosis and subsequently using them to stimulate T cells in a proliferation assay using an HLADR3-restricted T-cell clone recognising hsp65 epitope p3–13 [33]. Although both control cells and HLADOαβ–GFP-expressing cells were able to present the p3–13 epitope of intact hsp65 specifically, HLA-DOαβ–GFP expression clearly reduced the antigen-presenting capacity of the cells (Figure 6). No significant proliferative response was detected in response to unpulsed cells. To our knowledge, HLA-DO is the first protein that has been identified to function as an inhibitor of a chaperone (HLA-DM). To fully appreciate the effect of HLA-DO on HLA-DM-mediated class II peptide loading, we have generated cells in which HLA-DM shows quantitative association with HLA-DO. Even under these conditions, class II peptide loading is not completely abolished; part of the class II molecules still succeed in binding antigenic

peptides, as demonstrated by the residual amount of SDS-stable class II complexes and the diminished, but still significant, antigen-presenting capacity of HLADOαβ–GFP-expressing transfectants. Thus, instead of blocking HLA-DM action, the association of HLA-DO with HLA-DM is down-modulating HLA-DM-mediated class II peptide presentation. The observation that HLADO expression seems to be restricted to certain specialised antigen-presenting cells [22,34] as well as to distinct regions of the thymus [34] suggests that HLA-DO may affect the peptide repertoire presented by class II molecules only in particular situations. In B cells, part of HLA-DM is associated with HLA-DO which would result only in partial down-modulation of HLA-DM-mediated class II peptide loading. Primary B cells may require optimal peptide loading onto class II molecules only upon uptake of antigen via the B-cell receptor. Although there is no evidence as yet for free HLA-DO molecules outside

Research Paper HLA-DO is negative modulator of HLA-DM van Ham et al.

955

Figure 5 CLIP(81–104)

(a)

(c)

ApoB(2877–2894)

75 50 25 0 0.0

0.5

1.0

1.5

100 75 50 0 0.0

2.0

Incubation time (h)

0.2 0.1 1.0

1.5

Incubation time (h)

2.0

Absorbance 405 nm (arbitrary units)

Absorbance 405 nm (arbitrary units)

0.3

0.5

1.0

1.5

2.0

DM

DO/DM

0.4

0.2

ApoB(2877–2894)

0.4

0 0.0

0.5

Incubation time (h)

CLIP(81–104)

(b)

0.6

25 Relative absorbance (arbitrary units)

DR3–peptide complexes (%)

DR3–peptide complexes (%)

0.8 100

0.5 0.4 0.0

0.3 0.2 0.1 0 0.0

DM

0.5

1.0

1.5

DO/DM

Buffer

2.0

Incubation time (h)

Current Biology

Association of HLA-DO with HLA-DM abolishes HLA-DM-mediated catalysis of class II peptide loading. (a) Time course of dissociation of biotinylated CLIP(81–104) and ApoB(2877–2894) from purified HLADR3 molecules in the presence of lysis buffer (dots), cell lysates containing equal amounts of HLA-DM (squares) or HLA-DO–HLA-DM (triangles). Dissociation is shown as relative values compared to the initial amount of HLA-DR3–peptide complexes. (b) Time course of association of biotinylated CLIP(81–104) and ApoB(2877–2894) with HLA-DR3–CLIP in the presence of lysis buffer (dots), cell lysates containing equal amounts of HLA-DM (squares) or HLA-DO–HLA-DM (triangles). Association is shown as increase in absorbance (405 nm)

in arbitrary units. (c) Association of biotinylated ApoB(2877–2894) to HLA-DR3–CLIP after 50 min incubation with lysosomal lysates containing HLA-DM or HLA-DO–HLA-DM or lysis buffer only as indicated. Association is depicted as absolute values in absorbance at 405 nm (in arbitrary units). Inset: western blot using the HLA-DMαspecific monoclonal antibody 5C1 [34], demonstrating the presence of equal amounts of HLA-DM in the applied HLA-DM and HLA-DM–HLADO lysates. The data shown are a representative set of values from 3–4 individual experiments. Titration of added lysates to the experiments showed that the inhibition of HLA-DM activity correlated with the amount of lysate added (data not shown).

the ER, it may be possible that HLA-DO is released from HLA-DM upon triggering of the B-cell receptor or down-regulated to shift the peptide-presentation repertoire in particular circumstances, an issue that is currently under study.

which antigens will be expressed for the generation of a specific immune response.

Conclusions HLA-DO is a negative modulator of HLA-DM, the catalyst of antigen presentation via MHC class II molecules. Mel JuSo cells expressing the HLA-DOαβ heterodimer show inhibition of HLA-DM-mediated CLIP release from the class II molecules. This inhibition results in a profoundly diminished formation of class II–antigenic peptide complexes and consequently a reduced capacity to present antigens to T-cell clones. Thus, antigen presentation by MHC class II molecules is a more tightly controlled process than had been anticipated; HLA-DM alone skews the expressed peptide repertoire in the direction of stably binding antigens, whereas association of HLA-DO with HLA-DM counteracts this ushering mechanism. The balance of HLA-DO and HLA-DM expression in particular circumstances may therefore determine

Materials and methods Construction, characterisation and culture of HLA-DO–GFP transfectants The cDNAs encoding HLA-DOβ [24] and HLA-DOα [23] were cloned into pcDNA3 (Invitrogen) and a variant of pCEP4 (Invitrogen) that is disabled for episomal replication, respectively. A fusion construct of a FACS-optimised mutant of GFP to the carboxyl terminus of HLA-DOβ was generated by PCR; the HLA-DOβ cDNA was amplified using a primer overlapping the HLA-DOβ start codon and a primer specific for the last eight HLA-DOβ codons, omitting the stop codon but including a 3′ extension specific to the first four codons of GFP. The GFP cDNA was amplified using a primer specific to the first eight codons of GFP, with a 5′ extension specific to the last four codons of HLA-DOβ, and a primer specific to the last seven GFP codons. The resulting products were used as templates for a subsequent PCR using the primers specific to the amino terminus of HLA-DOβ and the carboxyl terminus of GFP. The final construct was checked by sequence analysis and cloned into pcDNA3. The melanoma cell line Mel JuSO (HLA-A1, HLA-B8, HLA-Cw7, HLA-DR3 and HLA-DQ2 as determined by DNA typing) was transfected using the calcium-phosphate precipitation procedure. Transfectants were selected in Iscove’s medium with 10% FCS, 2000 µg/ml G418 and 600 µg/ml hygromycin (Gibco BRL). HLADOαβ–GFP positive clones were identified by confocal analysis [35]

956

Current Biology, Vol 7 No 12

Figure 6

Biochemical analyses

[3H]-thymidine incorporation (cpm)

15,000

10,000

5,000

0 1,000

5,000 APCs per well

Mel JuSo + hsp65 Mel JuSo/DOαβ–GFP + hsp65 Mel JuSo Mel JuSo/DOαβ–GFP Current Biology

Antigen-presenting capacity of control and HLA-DOαβ–GFPexpressing cells for epitope p3–13 from Mycobacterium tuberculosis hsp65 protein. T-cell proliferation of the hsp65 p3–13 reactive, HLADR3-restricted clone Rp15 1-1 (expressed as the amount of [3H]thymidine incorporation in cpm) as a function of the number of antigenpresenting cells (APC) per well. Mel JuSo cells expressing HLADOαβ–GFP and Mel JuSo cells lacking its expression were cultured in the presence or absence of purified hsp65 protein as indicated. Data shown are from a representative experiment performed twice in quadruple; s.e.m. < 10%. T-cell proliferation in response to Mel JuSo cells pulsed with HLA-DR1-restricted peptide p411–425 was not significant. Titration experiments of the number of antigen-presenting cells or the dose of protein antigen tested showed dose-dependent effects on T-cell proliferation as expected (data not shown). and two-dimensional electrophoretic analysis of immunoprecipitations using an anti-GFP antiserum. HLA-DOαβ positive clones were identified by a rabbit antiserum raised against the carboxy-terminal peptide of HLADOβ (SGNEVSRAVLLPQSC) and by western-blot analysis for expression of HLA-DOβ with complex type N-linked glycans, demonstrating egress from the ER, which is HLA-DOα-dependent (data not shown).

FACS and confocal analysis For FACS analysis, cells were stained with saturating amounts of unlabelled primary antibody and phycoerythrin-conjugated F(ab′)2 rabbit anti-mouse IgG (H + L; Zymed) and analysed on a FACScan flow cytometer (Becton Dickinson). For immunofluorescence labelling, cells were fixed in ice-cold methanol and incubated with various antibodies followed by Texas red-conjugated secondary antibodies (Molecular Probes) in saturating amounts, as described [35]. Confocal analysis was performed using a 600MRC equipped with an argon/krypton laser (BioRad). Green fluorescence was detected at λ 520–560 nm after excitation at λ = 488 nm. Texas red fluorescence was detected at λ > 585 nm after excitation at λ = 568 nm.

To isolate antigen-presenting vesicles containing HLA-DM or HLADM–HLA-DO complexes, FACS-sorted GFP-positive cells were homogenised and postnuclear supernatant was separated by a linear 0.6–1.4 M sucrose gradient [36]. The vesicular fractions containing lysosomes were identified by measurement of β-hexosaminidase activity [36], and concentrated by dilution of sucrose in the solution and ultracentrifugation [36]. Lysates of purified MIICs or transfected cells were prepared in 50 mM Tris-HCl buffer containing 0.5% NP-40, 5 mM EDTA and protease inhibitors pH 8.0, and nuclei and debris were removed by centrifugation. Western-blotting analysis was used to determine that there were similar quantities of HLADM in the lysates. The association and dissociation rates of biotinylated peptides from affinity-purified DR3 molecules were measured essentially as described [16,37]. In brief, for dissociation analyses of bound biotinylated peptides, DR3 complexes were preloaded with 2 µM biotinylated peptides for 3 days (37°C, pH 4.5) in binding buffer containing 0.2% NP-40. Excess peptides were removed by 10-K ultrafiltration (Amicon). The time course of peptide dissociation was determined during 2 h at 37°C in binding buffer containing 0.2% NP-40 using 30 nM HLA-DR3–peptide complexes in the presence or absence of cell lysates (equivalent to 5 × 106 cells) or purified MIIC lysates containing HLA-DO–HLA-DM complexes and an excess of 50 µM unlabeled ApoB(2877–2894) peptide. Association of biotinylated peptides with HLA-DR3 was determined by adding 2 µM biotinylated peptides to 30 nM HLADR3–CLIP complexes in binding buffer containing 0.2% NP-40, pH 4.5, at 37°C. MHC–peptide complexes were immunoprecipitated with immobilised L234 antibody and peptides were detected via biotinylated streptavidin. The absorbance at 405 nm was measured by an enzyme-linked immunosorbent assay reader (Multiskan Plus, Titertek) and non-specific signals (quadruplicates, typically 15% of maximal absorbance) were subtracted from the data. Immunoprecipitation procedures and western-blot analyses were performed as previously described [27,38].

MHC class II peptide isolation, RP–HPLC, mass spectrometry and peptide sequencing Peptides were eluted from purified HLA-DR3 complexes as described [16,39]. RP–HPLC analysis was performed on a SMART system equipped with a µRPC C2/C18 SC 2.1/10 column (Pharmacia Biotech): eluent A, 0.1% trifluoroacetic acid (TFA); eluent B, 70% acetonitrile containing 0.1% TFA; gradient, 0–60% in 60 min; flow rate 50µl/min. Collected fractions were analysed by MALD time-of-flight MS [16] with 100 fmol [Glu] fibrinopeptide B (mass 1570.6 Da) added as an internal mass calibrant. For amino-acid sequence analysis, peptides were digested with 50 ng trypsin and reacted with n-succinimidyl-2 (3-pyridyl) acetate (SPA) to enhance β-ion abundance and facilitate sequence analysis by tandem MS [29]. Derivatised peptides of the CLIPs were then fully sequenced by low-energy collision-induced dissociation (CID) using a Finnigan MAT TSQ7000 tandem triple quadrupole MS [30].

T-cell proliferation assays For antigen presentation experiments, Mel JuSo cells were suspended in Iscove’s modified Dulbecco’s medium supplemented with 10% pooled human serum, irradiated (8000 rad) and seeded in 96-well flatbottomed microtiter plates at various cell concentrations, previously determined to trigger optimal T-cell proliferation (781, 1563 and 3125 cells/well) [33]. As antigens were added, purified hsp65 protein from Mycobacterium tuberculosis, the HLA-DR3-restricted epitope hsp65 peptide p3–13 or the control HLA-DR1 restricted peptide p411–425, and 104 T cells from the HLA-DR3-restricted, p3–13 specific T-cell clone Rp15 1-1 were then added. After 66 h in culture, 1 µCi [3H]thymidine was added to each well, and 18 h later cells were collected on glass-fibre filter strips and the radioactivity incorporated in the DNA was assessed by liquid scintillation counting [33]. No effect of HLA-DOαβ–GFP expression on T-cell proliferation was observed when Mel JuSo cells were exogeneously labelled with hsp65 p3–13 peptide (data not shown).

Research Paper HLA-DO is negative modulator of HLA-DM van Ham et al.

Acknowledgements E.P.M. Tjin and B.F. Lillemeier contributed equally to this work. We gratefully acknowledge I. Correa for the pCEP4 variant, B. Cormack for the GFP cDNA, D. Shima for the anti-GFP serum and P. Cresswell for the T2.DR3 cells and the CerCLIP.1 antibody. We thank E. Noteboom, P. Spee, E. Reits and R. Wubbolts for helpful assistance and J. Lardy for DNA typing of the Mel JuSo cells. This work was supported by fellowships to S.M.v.H. (Pioneer grant), to B.F.L. (Carl-Duisberg Stiftung), U.G. (Boehringer Ingelheim Fonds) and J.T. (Wellcome).

References 1. Neefjes JJ, Stollorz V, Peters PJ, Geuze HJ, Ploegh HL: The biosynthetic pathway of MHC class II but not class I molecules intersects the endocytic route. Cell 1990, 61:171-183. 2. Peters PJ, Neefjes JJ, Oorschot V, Ploegh HL, Geuze HJ: Segregation of MHC class II molecules from MHC class I molecules in the Golgi complex for transport to lysosomal compartments. Nature 1991, 349:669-676. 3. Avva RR, Cresswell P: In vivo and in vitro formation and dissociation of HLA-DR complexes with invariant chain-derived peptides. Immunity 1994, 1:763-774 . 4. Riese RJ, Wolf PR, Brömme D, Natkin LR, Villadangos JA, Ploegh HL, Chapman HA: Essential role for cathepsin S in MHC class IIassociated invariant chain processing and peptide loading. Immunity 1996, 4:357-366. 5. Ghosh P, Amaya M, Mellins E, Wiley DC: The structure of an intermediate in class II MHC maturation: CLIP bound to HLA-DR3. Nature 1995, 378:457-462. 6. Roche PA, Cresswell P: Invariant chain association with DR molecules inhibits immunogenic peptide binding. Nature 1990, 345:615-618. 7. Riberdy JM, Newcomb JR, Srunam MJ, Barbosa JA, Cresswell P: HLA-DR molecules from an antigen-processing mutant cell line are associated with invariant chain peptides. Nature 1992, 360:474-477. 8. Jensen PE: Enhanced binding of peptide antigen to purified class II major histocompatibility glycoproteins at acidic pH. J Exp Med 1991, 174:1111-1120. 9. Fling SP, Arp B, Pious D: HLA-DMA and -DMB genes are both required for MHC class II–peptide complex formation in antigenpresenting cells. Nature 1994, 368:554-558. 10. Morris P, Shaman J, Attaya M, Amaya M, Goodman S, Bergman C, et al.: An essential role for HLA-DM in antigen presentation by class II major histocompatibility molecules. Nature 1994, 368:551-554. 11. Miyazaki T, Wolf P, Tourne S. Waltzinger C, Dierich A, Barois N, et al.: Mice lacking H2-M complexes, enigmatic elements of the MHC class II peptide-loading pathway. Cell 1996, 84:531-541. 12. Martin WD, Hicks GG, Mendiratta SK, Leva HI, Ruley HE, Van Kaer L: H2-M mutant mice are defective in the peptide loading of class II molecules, antigen presentation, and T cell repertoire selection. Cell 1996, 84:543-550. 13. Sloan VS, Cameron P, Porter G, Gammon M, Amaya M, Mellins E, et al.: Mediation by HLA-DM of dissociation of peptides from HLADR. Nature 1995, 375:802-806. 14. Sherman MA, Weber DA, Jensen PE: DM enhances peptide binding to class II MHC by release of invariant chain-derived peptide. Immunity 1995, 3:197-205. 15. Denzin LK, Cresswell P: HLA-DM induces CLIP dissociation from MHC class II alpha beta dimers and facilitates peptide loading. Cell 1995, 82:155-165. 16. van Ham SM, Grüneberg U, Malcherek G, Bröker I, Melms A, Trowsdale J: Human histocompatibility leukocyte antigen (HLA)DM edits peptides presented by HLA-DR according to their ligand-binding motifs. J Exp Med 1996, 184:2019-2024. 17. Weber DA, Evavold BD, Jensen PE: Enhanced dissociation of HLADR-bound peptides in the presence of HLA-DM. Science 1996, 274:618-620. 18. Kropshofer H, Vogt AB, Moldenhauer GJH, Blum JS, Hämmerling GJ: Editing of the HLA-DR-peptide repertoire by HLA-DM. EMBO J 1996, 15:6144-6154. 19. Denzin LK, Hammond C, Cresswell P: HLA-DM interactions with intermediates in HLA-DR maturation and a role for HLA-DM in stabilizing empty HLA-DR molecules. J Exp Med 1996, 184:21532165. 20. Kropshofer H, Arndt SO, Moldenhauer, G, Hämmerling GJ, Vogt AB: HLA-DM acts as a molecular chaperone and rescues empty HLADR molecules at lysosomal pH. Immunity 1997, 6:293-302.

957

21. Sanderson F, Kleijmeer, MJ, Kelly AP, Verwoerd D, Tulp A, Neefjes JJ, et al.: Accumulation of HLA-DM, a regulator of antigen presentation, in MHC class II compartments. Science 1994, 266:1566-1569. 22. Liljedahl M, Kuwana T, Fung-Leung W-P, Jackson MR, Peterson PA, Karlsson L: HLA-DO is a lysosomal resident protein which requires association with HLA-DM for efficient intracellular transport. EMBO J 1996, 15:4817-4824. 23. Young JA, Trowsdale J: The HLA-DNA (DZA) gene is correctly expressed as a 1.1 kb mature mRNA transcript. Immunogenetics 1990, 31:386-388. 24. Tonnelle C, DeMars R, Long EO: HLA-DOb: a new b chain gene in HLA-D with a distinct regulation of expression. EMBO J 1985, 4:2839-2847. 25. Heim R, Cubbit AB, Tsien RY: Improved green fluorescence. Nature 1995, 373:663-664. 26. Sette A, Ceman S, Kubo RT, Sakaguchi K, Appella E, Hunt DF, et al.: Invariant chain peptides in most HLA-DR molecules of an antigenprocessing mutant. Science 1992, 258:1801-1804. 27. Riberdy JM, Newcomb JR, Surnam MJ, Barbosa JA, Cresswell P: HLA-DR molecules from an antigen-processing mutant cell line are associated with invariant chain peptides. Nature 1992, 360:474-477. 28. Mock KK, Sutton CW, Cottrell JS: Sample immobilisation protocol for matrix-assisted laser desorption mass spectrometry. Rapid Commun Mass Spectrom 1992, 6:233-238. 29. Sherman NE, Yates NA, Shabanowitz J, Hunt DF, Jeffery WA, BartletJones M, et al.: A novel N-terminal derivative designed to simplify peptide fragmentation. In Proceedings of the 43rd ASMS conference of Mass Spectrometry and Allied Topics , May 21–26. Atlanta, Georgia; 1995:626-627. 30. Hunt DF, Yates JR, Shabanowitz J, Winston S, Hauer CR: Protein sequencing by tandem mass spectrometry. Proc Natl Acad Sci USA 1986, 84:6223-6237. 31. Riberdy JM, Cresswell P: The antigen-processing mutant T2 suggests a role for MHC-linked genes in class II antigen presentation. J Immunol 1992, 148:2586-2590. 32. Malcherek G, Gnau V, Jung G, Rammensee H-G, Melms A: Supermotifs enable natural invariant chain-derived peptides to interact with many major histocompatibility complex-class II molecules. J Exp Med 1995, 181:527-536. 33. Geluk A, van Meijgaarden KE, Janson AAM, Drijfhout J, Meloen RH, De Vries RRP, et al.: Functional analysis of DR17(DR3)-restricted mycobacterial T cell epitopes reveals DR17 binding motif and enables the design of allele-specific competitor peptides. J Immunol 1992, 149:2864-2871. 34. Douek DC, Altmann DM: HLA-DO is an intracellular class II molecule with distinctive thymic expression. Int Immunol 1997, 9:355-364. 35. Wubbolts R, Fernandez-Borja M, Oomen L, Verwoerd D, Janssen H, Calafat J, et al.: Direct vesicular transport of MHC class II molecules from lysosomal structures to the cell surface. J Cell Biol 1996, 135:611-622. 36. Fernandez-Borja M, Verwoerd D. Sanderson F, Aerts H, Trowsdale J, Tulp A, Neefjes J: HLA-DM and MHC class II molecules codistribute with peptidase-containing lysosomal subcompartments. Int Immunol 1996, 8:625-640. 37. Malcherek G, Falk K, Rötzschke O, Rammensee H-G, Stevanovic S, Gnau V, et al.: Natural peptide ligand motifs of two HLA molecules associated with myasthenia gravis. J Immunol 1994, 153:11411149. 38. Sanderson F, Thomas C, Neefjes J, Trowsdale J: Association between HLA-DM and HLA-DR in vivo. Immunity 1996, 4:1-20. 39. Falk K, Rotzschke O, Stevanovic S, Jung G, Rammensee H: Allelespecific motifs revealed by sequencing of self-peptides eluted from MHC molecules. Nature 1991, 368:290-296. 40. Lampson LA, Levy R: Two populations of Ia-like molecules on a human cell line. J Immunol 1980, 125:293-299. 41. Denzin LK, Robbins NF, Carboy-Newcomb C, Cresswell P: Assembly and intracellular transport of HLA-DM and correction of the class II antigen-processing defect in T2 cells. Immunity 1994, 1:595-606. 42. Adams TE, Bodmer JG, Bodmer WF: Production and characterization of monoclonal antibodies recognizing the achain subunits of human Ia alloantigens. Immunology 1983, 50:613-624.