| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Disrupts Antigen Presentation Without Altering Endosomal Localization1
*
School of Medicine, University of Pennsylvania, Philadelphia, PA 19104;
Department of Pediatrics, Stanford University Medical Center, Stanford, CA 94305;
Department of Medicine, North Shore University Hospital-New York University School of Medicine, Manhasset, NY 11030; and
§
Virginia Mason Research Center, Seattle, WA 98101
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Ser) that creates a novel
glycosylation site at Asn94. The mutant DR molecules are
primarily associated with nested fragments of invariant chain (class
II-associated invariant chain peptides), and their interaction with
HLA-DM is impaired. Here we further analyzed the defect in 10.24.6
cells. Expressing Ser96 mutant DRA cDNA in DRA-null cells
recapitulated the 10.24.6 phenotype, indicating that the mutation
causes the Ag presentation defect. A mutation to Ala96
,
which does not introduce an extra glycan, generated a normal phenotype;
the critical role of the glycan was further supported by experiments in
which N-glycosylation was blocked by tunicamycin. We also
evaluated whether the 10.24.6 mutation affected DR3 maturation or
trafficking. Metabolic labeling and subcellular fractionation showed
that assembly, endosomal transport, and invariant chain proteolysis of
mutant DR3 molecules were similar to wild-type. A slight delay in
export from the endoplasmic reticulum to the Golgi apparatus in 10.24.6
cells probably did not contribute significantly to the Ag presentation
defect, because the abundance of DM and mutant DR in peptide-loading
compartments was normal at steady state. Our results indicate that
proper localization of these molecules does not depend on their
interaction. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
or -DMß
accumulate class II-CLIP complexes (11, 12), implicating HLA-DM in CLIP
release and normal peptide loading. In vitro, HLA-DM accelerates
dissociation of CLIP and other peptides from MHC class II molecules
(13, 14, 15, 16, 17) and stabilizes empty class II molecules against
denaturation (18, 19).
We have previously described a mutant B-LCL, 10.24.6, in which aberrant
glycosylation of HLA-DR molecules is associated with an Ag presentation
defect; 10.24.6 cells present exogenously supplied peptides, but not
endocytosed protein Ags, to DR-restricted T cells (20). Even though DM
expression is normal in 10.24.6 cells, the HLA-DR molecules in 10.24.6
resemble those found in DM-deficient cells: they lack some
conformational mAb epitopes, have reduced stability in SDS (an
indication of abnormal peptide loading (11, 12, 20, 21, 22, 23)), and
accumulate CLIP. The genetic basis for aberrant DR glycosylation in
10.24.6 is a single point mutation in the HLA-DRA gene
(Pro96Ser) that generates a novel
N-glycosylation site at position Asn94. That
this mutation is also responsible for the presentation defect was
suggested by partial restoration of Ag presentation when 10.24.6 cells
were transfected with wild-type DRA cDNA (20). The mutant DR from
10.24.6 cells has reduced affinity for HLA-DM (18, 24) and is resistant
to HLA-DM-catalyzed CLIP release in vitro (13), implying that the
expression of CLIP-loaded DR in 10.24.6 cells reflects a lack of DM-DR
interaction in vivo.
Here we have further analyzed the mechanism by which Ag presentation is disrupted in 10.24.6 cells. We provide evidence that the DRA point mutation causes the peptide-loading defective phenotype, dissect the contribution of the additional glycan vs substitution of the conserved Pro96 residue to the functional defect, and characterize the effect of the 10.24.6 mutation upon intracellular trafficking of HLA-DR and DM.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The cell line 8.1.6 is a human B-LCL hemizygous for
HLA-DR/DQ/DMB that expresses the HLA class II specificities DR3,
DRw52a, DQ2, and two copies of DP4.1 (25). Mutant 10.24.6 was isolated
after ethyl methane sulfonate mutagenesis of 8.1.6 and immunoselection
with the mAb 16.23 plus complement (20). The cell line 9.22.3 is
DR-null due to deletion of the remaining DRA gene from 8.1.6.
Although 9.22.3 expresses DR3 and DRw52a ß-chains, it lacks surface
expression of HLA-DR molecules (25). Cells were maintained in RPMI 1640
with 25 mM HEPES, 2 mM L-glutamine, and 15%
iron-supplemented bovine calf serum (Life Technologies,
Gaithersburg, MD).
Introduction of wild-type and mutant DRA chains into DRA-null cells
The Pro96Ser mutation of DRA from 10.24.6
was introduced into DRA-null 9.22.3 cells by retroviral transduction.
Ser96 mutant DRA cDNA was PCR-amplified using the primers
DRA-U1 (5' CGA GAA GGA TCC ACT CCC AAA AGA GCG CGC CCA A 3') and DRA-L2
(5' CAG TGA TCT GAA TTC TAA GAA ACA CCA TCA CCT CC 3') and Pfu DNA
polymerase (Stratagene, La Jolla, CA). Primer-encoded BamHI
and EcoRI sites were used for directional cloning into the
retroviral vector, pBMN (kind gift of Dr. G. Nolan, Stanford
University, Stanford, CA). Using
Ca3(PO4)2 transfection in the
presence of 25 µM chloroquine, the vector (13.5 µg) was transiently
transfected into the retroviral packaging cell line, 
NXA, to
generate a cellfree recombinant retroviral supernatant, as described
(26). For infection, 2.5 x 106 9.22.3 cells were
exposed to 5 ml of retroviral supernatant containing 4 µg/ml of
polybrene (Sigma, St. Louis, MO) for 24 h. Subpopulations
expressing large amounts of virally encoded DRA-Ser96
chains (5–10% of cells, as determined by flow cytometry) were sorted
by FACS, using the anti-DR mAb L243.
The transfectant Ala96 was generated by electroporation of 9.22.3 cells with a pRC/CMV vector (Invitrogen, San Diego, CA) containing a mutated DRA with a proline to alanine substitution at position 96. The point mutation was introduced by site-directed mutagenesis using overlap extension PCR (27). Two days after transfection, cells were selected in 1 mg/ml of G418 (Life Technologies) and cloned by limiting dilution. Clones with comparable HLA-DR expression were identified by flow cytometry.
As controls, empty, LacZ, and wild-type DRA (Pro96) constructs were introduced into 9.22.3 cells. All constructs were checked by sequencing the inserts (Stanford University Protein and Nucleic Acid Facility, Stanford, CA).
HLA-DR surface immunophenotyping
mAb 16.23 recognizes a polymorphic determinant on mature HLA-DR3
molecules (28). mAb ISCR3 recognizes HLA-DR dimers (4, 29). mAb L243
reacts with DR dimers and recognizes a monomorphic determinant on
HLA-DR (30, 31). The mAb CerCLIP.1, a gift from P. Cresswell, (Yale
University, New Haven, CT) recognizes the N terminus of CLIP (32). The
mAb HB10.a recognizes HLA-DR and -DPß-chains (33). Cells were
incubated with varying amounts of primary Ab, followed by excess goat
anti-mouse IgG-FITC (Life Technologies) and analyzed on a FACScan
flow cytometer (Becton Dickinson, Lincoln, NJ) using log amplification.
Results were converted to mean linear fluorescence intensities (in
arbitrary units) using Lysys II software (Becton Dickinson). Binding of
DM-dependent Abs was corrected for total DR expression (as measured by
L243 binding) for each cell line using the following formula: %
antibody binding ratio = 100 x [MFI(16.23 or CerCLIP.1) -
MFI(secondary Ab alone)]/[MFI(L243) - MFI(secondary Ab alone)],
where MFI is the mean fluorescence intensity at saturating mAb
concentrations.
Metabolic labeling and immunoprecipitation
Cells were pulse labeled for 20 min using 0.14 mCi/ml [35S]protein labeling mix (Dupont NEN, Wilmington, DE) and chased in medium containing excess cold methionine and cysteine for various times. To block N-glycosylation, cells were preincubated with 8 µg/ml of tunicamycin for 2 h and maintained in the presence of tunicamycin during pulse labeling and for the first 90 min of overnight chase. Cells were lysed in buffer containing 1% Nonidet P-40, and DR molecules were immunoprecipitated from precleared extracts using ISCR3 or L243 mAbs and protein A-Sepharose (Pharmacia, Piscataway, NJ). For analysis of SDS-stable complexes, ISCR3 immunoprecipitates were eluted by incubation at room temperature in SDS-PAGE sample buffer containing 1% SDS and 3% 2-ME for 75 min, and half of the eluate was boiled. Alternatively, L243 immunoprecipitates were boiled in 0.6% SDS, 1% 2-ME; eluates were split in half and digested with or without 2 mU of endoglycosidase H (Endo H) at pH 5.5, 37°C overnight. Samples were analyzed by 12% SDS-PAGE and fluorography. For quantitation, band intensities were measured using an Epson ES-1200C flatbed scanner and the public domain NIH Image program (version 1.60; developed by the National Institutes of Health and available on the internet at http://rsb.info.nih.gov/nih-image/). For each lane, the percentage of SDS-stable complexes was calculated as: 100 x (intensity of SDS-stable DR dimer band)/[(total intensity of monomeric DR bands) - (intensity of SDS-stable DR dimer band)].
T cells and T cell proliferation assays
The alloreactive, DR3-specific T cell clone 4.26B was kindly provided by Dr. A. Johnson (Georgetown University, Washington, DC) and propagated as described (34). The tetanus toxoid-specific, DR-restricted T cell clone DN-TT was maintained as described (35). In proliferation assays, T cells (2 x 104) were cultured for 60 h either with 2 x 105 mitomycin C-treated B-LCL in the presence of varying amounts of tetanus toxoid (for DN-TT) or with varying amounts of B-LCL alone (for 4.26B) in 0.2 ml of complete RPMI 1640 medium. T cell stimulation was measured by incorporation of [3H]thymidine (1 µCi/well) during the last 10 to 16 h of culture. Each experiment was performed at least twice with similar results, and results were expressed as median cpm of triplicate cultures.
Subcellular fractionation and analysis for organelle markers
Homogenization and fractionation were performed using an adaptation of a previously described protocol (7). All steps were conducted at 4°C. Approximately 2.5 x 108 washed cells were suspended in 1.2 ml of homogenization buffer (0.25 M sucrose, 1 mM EDTA, 10 mM HEPES, pH 7.3, and 1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride) and subjected to 100 pounds per square inch of nitrogen gas for 10 min in a cell disruption bomb (Parr Instrument, Moline, IL). The cells were homogenized using two strokes in a Dounce homogenizer fitted with a loose pestle. The homogenate was spun at 500 x g (2500 rpm in an Eppendorf microfuge) for 10 min to remove nuclei. The 1.5 ml of pooled supernatant and nuclear pellet wash was layered on top of 26.5 ml of 17% Percoll (Pharmacia) in homogenization buffer (1.06 g/ml) on a 2-ml cushion of 2.5 M sucrose. Percoll gradients were spun at 40,000 x g in an SS-34 rotor (Sorvall, Newtown, CT) for 40 min. Fifteen 2-ml fractions were collected from the bottom of the gradient.
Plasma membrane was identified by the activity of 5'-nucleotidase, assayed as described (36). Lysosomes were identified by ß-hexosaminidase activity, assayed as described (37). Other proteins were detected by immunoblotting of solubilized fractions, essentially as described (25). Briefly, fractions were solubilized in 1% Nonidet P-40 and SDS-sample buffer with 3% 2-ME, boiled, resolved by SDS-PAGE, and transferred to polyvinylidene difluoride membranes. Blocked membranes were incubated with primary Abs, followed by goat anti-mouse IgG or donkey anti-rabbit IgG horseradish peroxidase conjugate (Life Technologies). Bound Ab was detected using an enhanced chemiluminescence substrate (Renaissance substrate kit; Dupont NEN). Calnexin, a resident ER protein, was detected using the murine anti-calnexin Ab AF8, kindly provided by Dr. Michael Brenner (Harvard Medical School, Boston, MA) (38). Rab 5 and rab 7, markers for early and late endosomes, respectively (39), were detected with 4F11, a murine anti-rab 5 mAb, and affinity purified anti-rab 7 polycolonal rabbit antiserum, both kindly provided by Dr. Angela Wandinger-Ness (Northwestern University, Evanston, IL) (40, 41). DRß was detected with the murine Ab HB10.a (33), described above. K455, a rabbit anti-class I antiserum, was kindly provided by Dr. Lars Karlsson (R. W. Johnson Pharmaceutical Research Institute, La Jolla, CA) (42). 47G.S4, a murine anti-DMß mAb, was a gift from Dr. Susan Pierce (Northwestern University, Evanston, IL) (43).
For immunoprecipitation from subcellular fractions, cells (9 x 107 cells per time point) were labeled as described above with L-[35S]methionine/cysteine labeling mix for 20 min. At each chase point, an aliquot of cells was homogenized and fractionated. Nonidet P-40 was added to each fraction at a final concentration of 1%. DR dimers from the fractions were immunoprecipitated using L243, and samples were processed for SDS-PAGE and fluorography as described above.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
substitution are
defective for CLIP exchange and peptide loading in vivo
In earlier studies, we identified the DRA mutation in 10.24.6
cells and showed that transfection of the cells with a wild-type DRA
cDNA complemented their Ag presentation defect (20). Complementation
was incomplete in assays such as expression of particular DR3 epitopes,
due, we suspected, to continued expression of the mutant molecules in
the transfectants. To establish whether the phenotype of 10.24.6 cells
was derived solely from the known mutation, wild-type and
Ser96 DRA cDNAs were introduced into the DRA-null,
DM+ B-LCL, 9.22.3, by retroviral transduction. Cells
transduced with wild-type DRA display surface DR molecules with a
phenotype indicating normal peptide loading (Fig. 1
A). They express
relatively high levels of the DM-dependent epitope recognized by mAb
16.23 (24% of L243 staining) and low levels of CLIP, as measured by
the mAb CerCLIP.1 (10% of L243; the absolute levels are not
significantly above those seen for LacZ-transduced controls). For
comparison, the mean 16.23 fluorescence obtained for 8.1.6 wild-type
progenitors typically ranges from 20 to 45% of the fluorescence
obtained with L243 (32% in the experiment shown; Figure 1B
and data not shown). mAb 16.23 binds about 10-fold less well to 9.5.3
mutants lacking DM (3% of L243 staining), with 10.24.6 cells
displaying intermediate levels of binding (11% of L243). The
CLIP-reactive CerCLIP.1 mAb reacts well with both DM-null 9.5.3 cells
(71% of L243 staining) and the 10.24.6 DRA mutant (52%), but poorly
with 8.1.6 cells (7%). The 9.22.3 cells transduced with retrovirus
encoding Ser96 DRA cDNA have a phenotype very similar to
that of 10.24.6 cells (Fig. 1A): the mutant DR
molecules accumulate CLIP (CerCLIP.1 staining is 40% of L243) and
react less well with 16.23 (10% of L243 staining) than the wild-type.
These results confirm that the DRA mutation, rather than adventitious
defects of 10.24.6 introduced by random mutagenesis or clonal
variation, causes the presentation-defective HLA-DR phenotype in
10.24.6 cells.
|
The Pro96Ser mutation in DRA of 10.24.6 cells
introduces two structural changes in DR. First, it replaces a proline
residue that is conserved among murine and human class II molecules
(44) and interacts with residue 118 of the DR ß-chain (10). Second,
the substitution generates a target site for N-linked
glycosylation at Asn94 of the DR-chain. Either change
might affect DR conformation or impair interactions with DM. To dissect
the effects of the two structural changes, we mutated Pro96
to Ala, which does not result in glycosylation at Asn94.
The mutant construct was transfected into the DRA-null cell line,
9.22.3, generating Ala96 cells. We also produced
transfectants with wild-type DR (Pro96) and empty vector
as controls.
When measured by flow cytometry with a monomorphic anti-DR
mAb, L243, or two DRß-specific Abs, HB10.a (Fig. 1
C) and CD6b.1 (20, data not shown), HLA-DR
expression is clearly observed for the Pro96 wild-type and
Ala96 mutant transfectants, with the Ala96
mutant DR being expressed at somewhat higher levels. For both
transfectants, HLA-DR expression is lower than for 8.1.6 progenitor
cells, the DMß-null mutant 9.5.3, and the DR glycosylation mutant
10.24.6 (Fig. 1B). Empty-vector transfected 9.22.3
cells are not stained by the anti-DR mAbs, except for HB10.a, which
is known to cross-react with HLA-DP expressed on 9.22.3 (Fig. 1C).
Staining with DM-dependent mAbs reveals a normal peptide-loading
phenotype for both the Pro96 and Ala96
transfectants. At saturation, 16.23 staining is about one-fifth of L243
staining (18 and 23%, respectively, in the experiment shown), which is
at the lower end of the range seen for 8.1.6 progenitor cells and
twofold higher than for mutant 10.24.6 (Fig. 1B). In
contrast to cells harboring the Ser96 DRA mutation, both
the Pro96 and Ala96 transfectants stain weakly
with CerCLIP.1 (5 and 10% of L243, respectively; Fig. 1C) as does the empty vector transfectant. Together,
these results argue that CLIP release and normal peptide loading are
not greatly affected by replacing the conserved Pro96
residue of DR by alanine. Consistent with this conclusion, the
majority of HLA-DR molecules are stable to SDS-induced dimer
dissociation in both Pro96 and Ala96
transfectants (data not shown).
We also compared the ability of the Pro96 and
Ala96 transfectants to stimulate proliferation of
allospecific and Ag-specific, DR3-restricted T cells (Fig. 2), which have been shown previously to
be DM dependent (11). As expected, the alloreactive clone is stimulated
well by 8.1.6 progenitor cells, but not by 10.24.6 or the empty
vector-transfected DRA-null cell line, 9.22.3 (Fig. 2A). Stimulation by Ala96 is intermediate
between 8.1.6 and the wild-type Pro96 transfectant,
correlating with levels of HLA-DR expression (cf. Fig. 1, B
and C). Comparable results are obtained with a
tetanus toxoid-specific clone (Fig. 2B). At limiting
Ag concentrations, 9.22.3-empty vector transfectants and 10.24.6 cells
fail to induce proliferation and Pro96, Ala96,
and 8.1.6 cells stimulate increasing levels of T cell proliferation. At
higher Ag concentrations, a low level of proliferation is seen with
mutant 10.24.6 cells, while Pro96, Ala96, and
progenitor 8.1.6 cells all stimulate similar high levels of
proliferation. Thus, the Pro96Ala substitution in DRA is
tolerated well for presentation of different DM-dependent antigenic
determinants to T cells.
|
To test the contribution of the extra glycosylation at
Asn94 to the DR phenotype of 10.24.6, we treated cells
with tunicamycin, an inhibitor of N-linked glycosylation,
and measured the acquisition of an SDS-stable conformation as a marker
for normal peptide loading ((11, 20); Fig. 3). In tunicamycin-treated cells, the
monomeric DR chains are of reduced m.w. compared with untreated cells,
indicating efficient blocking of N-glycan addition. In
addition, the treated cells have diminished amounts of precipitated DR
molecules. In both 8.1.6 and 10.24.6, a novel band at the expected m.w.
for the unglycosylated SDS-stable DR dimer appears in the unboiled,
tunicamycin-treated samples. In contrast, DMB-null 9.5.3 cells have no
detectable unglycosylated SDS-stable DR dimers, indicating that their
formation requires DM. As expected, untreated 10.24.6 mutants contain
fewer SDS-stable complexes than untreated 8.1.6 progenitors.
Interestingly, tunicamycin treatment decreases the percentage
(quantitated by densitometry) of SDS-stable wild-type DR3 complexes
from progenitor 8.1.6 by about half (from 42 to 19% in the experiment
shown), whereas the SDS stability of mutant DR3 molecules is unchanged
or slightly increased (from 13 to 17%). As a result, after treatment,
the percentage of stable complexes in 10.24.6 cells is comparable to
that in 8.1.6 cells. These data imply that tunicamycin facilitates
peptide loading of the mutant DR3 molecules, although it also
interferes with acquisition of SDS stability. The results argue that
the reduced ability of the mutant DR3 molecules in 10.24.6 cells to
load stabilizing peptides is in large part due to the novel
glycosylation at position 94 of the mutant DR-chain.
|
The extra glycan on the mutant DR molecules of 10.24.6 cells
could interfere with their intracellular trafficking either by reducing
interaction with targeting molecules or by causing a DR conformation
that lacks transport competence. To determine whether altered
intracellular transport contributed to the phenotype of 10.24.6 cells,
we analyzed DR molecule trafficking by pulse-chase analysis and
gradient fractionation.
To measure rates of ER-to-Golgi transport, metabolically labeled DR was
immunoprecipitated and treated with Endo H, which cleaves high mannose,
but not complex N-glycans after Golgi processing. DRß has
one N-glycan and wild-type DR has two, only one of which
is converted to the complex form (45). The mutant DR in 10.24.6 has
a third N-linked glycan. Immediately after labeling, the DR
dimers from both cell lines are Endo H sensitive (Fig. 4). For both wild-type and mutant DR
molecules, processed subpopulations with an Endo H-resistant ß-chain
and a partially resistant -chain are first detected by 45 min of
chase and are the dominant species after long chase times (>5 h).
However, between 1 and 3 h, more of the DR protein remains Endo H
sensitive in 10.24.6 than in 8.1.6, implying that in 10.24.6, the DR
molecules reach the medial Golgi apparatus more slowly than in the
wild-type progenitor cells. Note that for mutant DR, only one of the
three glycans appears to be processed, as the Endo H-digested wild-type
and mutant DR migrate identically.
|
, at 3 h of
chase, bands corresponding to endosomally generated Ii degradation
intermediates (leupeptin-induced peptide, LIP, and small LIP, or SLIP)
(7, 46, 47) and CLIP coprecipitate with the DR dimers in both 8.1.6 and
10.24.6. As expected from surface Ab binding, CLIP persists throughout
the chase in 10.24.6 cells, whereas in 8.1.6, it remains at low levels
reflecting CLIP exchange. Normal steady-state localization of mutant HLA-DR and HLA-DM in 10.24.6 cells
Subcellular fractionation was used to examine whether the
steady-state distribution of HLA-DR molecules was altered in 10.24.6
cells. In 8.1.6 cells, HLA-DR molecules accumulate intracellularly in a
high density late endosomal/prelysosomal compartment that can be
separated from plasma membrane, ER, and early and most late endosomes
by fractionation on 17% Percoll density gradients
(Fig. 5).4 In these gradients,
a single peak of MHC class I molecules is found to cofractionate with
the plasma membrane marker, 5'-nucleotidase. In contrast, HLA-DR shows
a bimodal distribution with a large amount of DR in the light, plasma
membrane-containing fractions and a second peak, comprising about 30%
of total cellular DR, in the dense fractions. The distribution of
HLA-DM molecules is also bimodal, but the great majority of DM is in
the high density peak, which contains the lysosomal enzyme
ß-hexosaminidase and a small fraction of the late endosomal protein
rab 7, but lacks the transferrin receptor and markers for ER and plasma
membrane. The distribution of marker proteins and MHC class I molecules
in mutant 10.24.6 is indistinguishable from that seen for progenitor
8.1.6 cells, indicating that their organelle distribution is similar.
Of importance, the distribution of HLA-DR and HLA-DM is also comparable
in the mutant and progenitor cells (Fig. 5). Although reduced levels of
mutant DR molecules in MIIC were reported in a previous study (24), we
have seen normal levels of mutant DR in high density compartments of
10.24.6 cells in five of six independent gradients. Whatever the basis
for the observed differences, our results imply that the poor endosomal
DR accumulation reported by Sanderson et al. (24) likely was not due to
the novel glycan. Thus, despite the small delays in maturation of
mutant DR molecules in 10.24.6, their abundance in high density
(peptide-loading) compartments at steady state is not detectably
affected by the mutation.
|
Gradient fractionation of biosynthetically labeled cells was used
to directly compare the rates at which wild-type and mutant DR
molecules reach high density compartments (Fig. 6). At each chase time, gradient
fractions were immunoprecipitated to determine the location and Ii
association of DR molecules. In progenitor 8.1.6 cells, at 45 min of
chase, DR-Ii complexes are enriched both in fractions 3 to 5 and
fractions 12 to 14 (Fig. 6A). At this time, the DR
molecules are partially Endo H sensitive but have not yet undergone Ii
proteolysis (cf. Fig. 4). These results suggest that the DR molecules
in the light fractions are ER derived and those in the dense fractions
are newly arrived in high density endosomal compartments. By 3 h
of chase, more DR dimers, now associated with LIP and SLIP, are found
in the dense fractions 11 to 15. At the same time, DR molecules are
found in fractions 3 and 4 and are no longer associated with Ii,
consistent with their having reached the plasma membrane. By 8 h
of chase, the bulk of DR from progenitor 8.1.6 has mature glycans and
is found in fractions 3 and 4, with a subset in fractions 13 to 15, a
small proportion of which are associated with LIP, SLIP, and
CLIP.
|
B) reveals
that at 45 min of chase, mutant DR molecules are complexed with Ii and
are primarily detected in fractions 3 to 5 with a peak at fraction 4.
Thus, fewer mutant than wild-type DR molecules have reached loading
compartments at this time. Mutant DR/Ii complexes continue to be
detected in the light fractions at 3 h of chase, and dense
fractions (fractions 12 to 15) start to accumulate mutant DR molecules
primarily associated with intact Ii. By 8 h of chase, however, the
mutant molecules are associated with CLIP in the light fractions and
with low levels of LIP and SLIP as well as CLIP in dense fractions.
These results show that, except for a slight delay in endosomal
transport and the failure to release CLIP, the mutant molecules undergo
a similar sequence of trafficking and maturation steps as wild-type DR
undergoes. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The two approaches used to evaluate the importance of aberrant DR
glycosylation to the 10.24.6 phenotype give mutually consistent
results. However, each is subject to some limitations. Tunicamycin
prevents all N-glycan addition, not only addition of the
extra glycan on DR. Thus, it is perhaps not surprising that fewer DR
dimers can be precipitated from drug-treated cells and that
unglycosylated wild-type DR molecules have reduced SDS stability.
Normal N-glycosylation of newly synthesized DR molecules may
be important for their assembly, recognition by mAbs, and SDS stability
once loaded with peptide. In addition, glycosylation of other proteins
may be important for the function of cellular machinery involved in DR
assembly and maturation. Because only selected aspects of the 10.24.6
phenotype could be evaluated in tunicamycin-treated cells, we also used
site-directed mutagenesis. The ProAla mutation is considered less
conservative than the ProSer change (48); thus, it should have been,
if anything, more disruptive to Ag presentation if mutation of residue
96 were the critical factor. However, the results of this experiment do
not formally rule out the possibility that Ser96 makes a
contribution to the mutant phenotype.
Taken together with published results, our data support the view that the weakened interaction of 10.24.6-derived DR with DM is a consequence of direct steric interference by the extra glycan. In the crystal structure of DR3 (10), Asn94 is exposed to solvent and thus available for glycosylation without unduly disturbing DR conformation. Mutant DR maturation, Ab reactivity, and reactivity with T cells after peptide binding at the cell surface appear normal, as does DM-independent peptide binding capacity in vitro (13, 20). Therefore, any conformational defects in the mutant DR3 molecules must be quite subtle. The slightly reduced rates of export of from the ER in 10.24.6 cells might indicate increased recognition by ER chaperons, but this does not necessarily reflect malfolding. Alternative explanations include recognition of the additional glycan by ER chaperons such as calnexin (49, 50, 51) and subtle differences in ER quality control between 10.24.6 and 8.1.6 cells. The view that in endosomes, the extra glycan directly interferes with DM interactions is consistent with other studies attempting to map DM interaction sites on HLA-DR. The additional glycan is on the same lateral face of the HLA-DR structure as the CLIP N terminus, the epitope of the CerCLIP.1 mAb that is known to block DM-dependent CLIP release (14). We propose that DM interacts with MHC class II molecules via this face.
Our results add to the growing body of evidence that DM-DR interactions
do not regulate endosomal trafficking of either molecule. Even though
DM-DR interactions are greatly diminished by the DR mutation in
10.24.6, steady-state analysis of gradient fractions indicates that
both molecules accumulate in putative MIIC, which are defined here by
their high density and content of late endosomal and lysosymal marker
proteins. This conclusion is extensively corroborated by immunoelectron
microscopic analysis of 10.24.6 and related cells, which shows normal
localization of mutant DR3 and DM in multivesicular
compartments.4 Furthermore, except for poor CLIP release,
mutant DR molecules are processed normally in MIIC and reside there for
some time. These findings indicate that normal DM-DR interaction is not
required for DR retention and Ii proteolysis in MIIC. Our findings are
consistent with observations that DR molecules travel to MIIC in EBV-B
cell mutants lacking DM (52, 53), and that the cytoplasmic tails of
both human and murine DM contain lysosomal targeting sequences that
function in class II-negative nonlymphoid cells (54, 55). Nevertheless,
our data do not rule out the possibility that persistence of a subset
of DR molecules (e.g., empty molecules) in MIIC may be facilitated by
normal interactions with DM.
The modestly reduced rate of ER export in 10.24.6 cells may subtly
delay subsequent steps of DR maturation and contribute to slightly
reduced cell surface expression (note the different levels of L243 and
HB10.a staining in Fig. 1B). However, the delay does
not prevent transient accumulation of mutant DR in MIIC or endosomal Ii
processing. Therefore, it probably is not responsible for the inability
of mutant DR to be loaded with a normal array of endosomal peptides.
Normal DM-DR interactions do not appear to be important for efficient
ER export, because ER-to-Golgi transport and endosomal targeting of
wild-type DR3 are normal in 8.1.6-derived mutants that lack expression
of HLA-DMß4 (56). In this regard, HLA-DR3 differs from
the nonclassical MHC class II molecule, HLA-DO, whose export from the
ER requires coexpression of, and association with, DM in the ER
(57).
Despite their similar phenotype, 10.24.6 cells probably differ from
DM-null cells in that they have normal levels of endosomal DM, and
hence presumably of associated DO. Thus, the observation that peptide
exchange is impaired both in DM-null cells and in the DR mutant
suggests that in the intact cell, the peptide-loading defects are not
primarily related to a need for endosomal targeting of DO in Ag
presentation.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Elizabeth Mellins, Department of Pediatrics, Stanford University Medical Center, Room H306B, Stanford, CA 94305-5208. E-mail address:
3 Abbreviations used in this paper: B-LCL, B lymphoblastoid cell line; CLIP, class II-associated invariant chain peptides; Endo H, endoglycosidase H; ER, endoplasmic reticulum; Ii, invariant chain; MIIC, MHC class II compartment; LIP, leupeptin-induced peptide; SLIP, small leupeptin-induced peptide.
4 E. Stang, C. Guerra, M. Amaya, Y. Paterson, O. Bakke, and E. D. Mellins. DR/CLIP and DR/peptide complexes colocalize in prelysosomes in human B lymphoblastoid cells. J. Immunol., in press.
Received for publication July 14, 1997. Accepted for publication January 6, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
ß dimers and facilitates peptide loading. Cell 82:155.[Medline]
chain residues located on the outer loops are involved in nonpolymorphic and polymorphic antibody-binding epitopes. Hum. Immunol. 39:253.[Medline]
This article has been cited by other articles:
|
|
 |
|
  L.-E. Fallang, S. Roh, A. Holm, E. Bergseng, T. Yoon, B. Fleckenstein, A. Bandyopadhyay, E. D. Mellins, and L. M. Sollid Complexes of Two Cohorts of CLIP Peptides and HLA-DQ2 of the Autoimmune DR3-DQ2 Haplotype Are Poor Substrates for HLA-DM J. Immunol., October 15, 2008; 181(8): 5451 - 5461. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. H. Rinderknecht, M. P. Belmares, T. L. W. Catanzarite, A. J. Bankovich, T. H. Holmes, K. C. Garcia, N. K. Nanda, R. Busch, S. Kovats, and E. D. Mellins Posttranslational Regulation of I-Ed by Affinity for CLIP J. Immunol., November 1, 2007; 179(9): 5907 - 5915. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. van Lith and A. M. Benham The DM{alpha} and DMbeta Chain Cooperate in the Oxidation and Folding of HLA-DM J. Immunol., October 15, 2006; 177(8): 5430 - 5439. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. C. Doebele, A. Pashine, W. Liu, D. M. Zaller, M. Belmares, R. Busch, and E. D. Mellins Point Mutations in or Near the Antigen-Binding Groove of HLA-DR3 Implicate Class II-Associated Invariant Chain Peptide Affinity as a Constraint on MHC Class II Polymorphism J. Immunol., May 1, 2003; 170(9): 4683 - 4692. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. R. Hegde, R. A. Tomazin, T. W. Wisner, C. Dunn, J. M. Boname, D. M. Lewinsohn, and D. C. Johnson Inhibition of HLA-DR Assembly, Transport, and Loading by Human Cytomegalovirus Glycoprotein US3: a Novel Mechanism for Evading Major Histocompatibility Complex Class II Antigen Presentation J. Virol., October 2, 2002; 76(21): 10929 - 10941. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Stratikos, L. Mosyak, D. M. Zaller, and D. C. Wiley Identification of the Lateral Interaction Surfaces of Human Histocompatibility Leukocyte Antigen (HLA)-DM with HLA-DR1 by Formation of Tethered Complexes That Present Enhanced HLA-DM Catalysis J. Exp. Med., July 15, 2002; 196(2): 173 - 183. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. S. Arneson, J. F. Katz, M. Liu, and A. J. Sant Hydrogen Bond Integrity Between MHC Class II Molecules and Bound Peptide Determines the Intracellular Fate of MHC Class II Molecules J. Immunol., December 15, 2001; 167(12): 6939 - 6946. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. S. Patil, A. Pashine, M. P. Belmares, W. Liu, B. Kaneshiro, J. Rabinowitz, H. McConnell, and E. D. Mellins Rheumatoid Arthritis (RA)-Associated HLA-DR Alleles Form Less Stable Complexes with Class II-Associated Invariant Chain Peptide Than Non-RA-Associated HLA-DR Alleles J. Immunol., December 15, 2001; 167(12): 7157 - 7168. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. S. Patil, F. C. Hall, S. Drover, D. R. Spurrell, E. Bos, A. P. Cope, G. Sonderstrup, and E. D. Mellins Autoantigenic HCgp39 Epitopes Are Presented by the HLA-DM-Dependent Presentation Pathway in Human B Cells J. Immunol., January 1, 2001; 166(1): 33 - 41. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C.-L. Chou and S. Sadegh-Nasseri Hla-Dm Recognizes the Flexible Conformation of Major Histocompatibility Complex Class II J. Exp. Med., December 18, 2000; 192(12): 1697 - 1706. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Busch, Z. Reich, D. M. Zaller, V. Sloan, and E. D. Mellins Secondary Structure Composition and pH-dependent Conformational Changes of Soluble Recombinant HLA-DM J. Biol. Chem., October 16, 1998; 273(42): 27557 - 27564. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Faubert, A. Samaan, and J. Thibodeau Functional Analysis of Tryptophans alpha 62 and beta 120 on HLA-DM J. Biol. Chem., January 18, 2002; 277(4): 2750 - 2755. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. J. McFarland, J. F. Katz, C. Beeson, and A. J. Sant Energetic asymmetry among hydrogen bonds in MHC class II{middle dot}peptide complexes PNAS, July 31, 2001; 98(16): 9231 - 9236. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
