| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Institute for Genetics, University of Cologne, Cologne, Germany
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Newly synthesized MHC class I molecules associate in the ER transiently with calnexin, calreticulin, and TAP (11, 12, 13, 14, 15). In human cells calnexin binds free unfolded class I heavy chains, whereas calreticulin associates only with assembled but partially folded and empty heavy chain ß2m dimers and remains associated with them after they bind to TAP (14). Another MHC-encoded protein, tapasin (14, 16, 17), is required for the association of heavy chain ß2m dimers with TAP, perhaps because tapasin acts as a bridge (14, 17, 18). The dissociation of TAP-class I complexes is induced by peptide loading onto class I and is therefore dependent on the binding motif of the class I allele involved (11, 12). The release of class I from TAP coincides with the egress of loaded class I molecules from the ER into the secretory pathway (11, 12).
Since free peptides probably have a short life span in the ER (19, 20) and bind to various ER-resident proteins (21, 22, 23), the interaction between empty class I molecules and TAP may enable class I to sample a more abundant and more diverse array of peptides than would be available free in the ER lumen. However, there appears to be no absolute requirement for a physical interaction between class I and TAP for peptide loading. Thus, 1) HLA-A2 loads efficiently with signal peptides in T2 mutant cells lacking TAP (24, 25). 2) In human lymphoblastoid cell lines several alleles appear weakly if at all in immunoprecipitates with TAP, yet display normal loading and are transferred rapidly to the cell surface (26). Finally, 3) cells genetically defective in tapasin were first identified in vitro from defective surface class I expression (14, 27, 28); nevertheless, this defect in intracellular maturation seems to affect different class I alleles differentially, with some showing apparently normal maturation and expression despite no detectable interaction with TAP (27, 28). Thus, physical interaction with TAP is apparently not always prerequisite for class I loading, and at least certain alleles, in still not clearly defined circumstances, can load at a distance, so to speak, with peptides translocated into the ER by TAP. We shall refer to this phenomenon as indirect loading, in contradistinction to direct loading, which occurs within the TAP-class I loading complex.
In contrast to human and mouse, the rat possesses two allelic groups of functionally different transporters (TAP-A and TAP-B), defined by the complex alleles of the TAP2 chain (29, 30). Transporters of the TAP-A group are permissive for essentially all peptides of appropriate length, while transporters of the TAP-B group preferentially transport peptides with large hydrophobic C-termini (31, 32). Different rat MHC class I RT1.A alleles are linked in cis to different TAP alleles (29, 33), and these linkages appear to have functional significance. Thus, RT1.Aa, which has a strong preference for arginine-ended peptides (34), is encoded within the same RT1 haplotype as a permissive TAP-A group transporter (34). Furthermore, expression of functional TAP-A transporter is required for normal peptide loading and fast intracellular maturation of RT1.Aa molecules (30, 35). In contrast, in recombinant animals and transfected cell lines with a homozygous TAP-B background, RT1.Aa shows an unusually long retention within the ER and loads with an anomalous nonideal set of peptides (30, 34, 35). In this discordant situation TAP-B does not deliver peptides suitable for loading into RT1.Aa.
We asked whether, in the genetically discordant situation, RT1.Aa might be retained in the ER in persistent association with TAP-B and, if so, what the consequences might be for an endogenous allele, RT1.Au, that normally loads with peptides delivered by TAP-B. We show that the imbalanced peptide supply in this situation creates competition between the two class I populations for access to TAP. In the absence of functional TAP-A, transfected RT1.Aa accumulates in the ER both in and out of association with TAP-B. Meanwhile, the amount of RT1.Au associated with TAP-B is drastically reduced. However, despite being virtually excluded from TAP association, RT1.Au still matures at an apparently normal rate. In contrast, when the permissive TAP-A transporter required for normal loading by the RT1.Aa allele is transfected into these cells, the pool of immature RT1.Aa molecules partially empties, and normal transient association of both RT1.A alleles with TAP is found. We find additionally that both class I alleles associate indiscriminately with the permissive and restrictive TAP alleles.
We propose that access of class I molecules to TAP is also defined by a competitive situation in the ER and not only by the intrinsic ability of a class I molecule to form a loading complex with TAP. Our data provide further evidence that newly synthesized class I molecules can take alternative routes of loading and assembly, either involving or not involving a physical interaction with TAP. Finally, our data show, perhaps unexpectedly, that TAP itself is probably not the main anchor retaining empty class I molecules in the ER while they await a suitable peptide for loading.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Cells were maintained in RPMI 1640 (Life Technologies, Eggenstein, Germany) supplemented with 5% FCS (Sigma, Deisenhofen, Germany), 1000 U/ml penicillin, and 1 mg/ml streptomycin. Cultures were incubated at 37°C in a humidified atmosphere of 5% CO2. Transfected cell lines were cultured in the presence of 0.1 mM hypoxanthine, 0.4 µM aminopterin, and 16 µM thymidine (Life Technologies) and G418 (1 mg/ml; PAA, Colbe, Germany)
Cell culture medium was made with ultrapure water (0.055 µS/cm)
derived from a combined reverse osmosis/ultrapure water system (ß
75/
UV/UF, USF Seral Reinstwassersysteme, Ransbach-Baumbach,
Germany) equipped with UV (185/254 nm) and ultrafiltration (5000-Da
cutoff).
Cell lines
C58 is a rat T cell lymphoma line carrying the RT1u MHC haplotype (RT1.Au, TAP2u) (36). The C58.RT1.Aa cell line (C58.331) was created by transfection of a cDNA encoding RT1.Aa (37). C58.331-B5 (B5) was created by transfection of C58.RT1.Aa with cDNA encoding TAP-2a (30). C58.331-D7 (D7) was created by transfection of C58.RT1.Aa with cDNA encoding TAP-2u (38).
Antibodies
116/5 is a polyclonal rabbit antiserum recognizing the
C-terminus of rat TAP2 chains (EQDVYAHLVQQRLEA) (39). D90 is a
polyclonal rabbit antiserum recognizing the C terminus of rat TAP1
chains (CYRSMVEALAAPSD) (40). MAC 394 is a monoclonal mouse Ab
against rat TAP2a derived from immunization with
recombinant His-tagged cytoplasmic domain (residues 485–703) of rat
TAP2a. MAC 394 fails to detect TAP2u in Western
blots and immunoprecipitates (see Results and Fig. 3
A) (A. Seelig, manuscript in preparation). F87 and F88 are
both polyclonal rabbit antisera against rat MHC class I RT1.A produced
by immunization with acetic acid-denatured RT1.Aa
molecules. In Western blots and immunoprecipitations F87 recognizes
predominantly the allelic product of RT1.Aa, while F88
reacts with RT1.Aa as well as with RT1.Au
molecules (data not shown). R3/13 is an RT1.Aa-specific
alloantibody (IgG2b) derived from AO anti-DA (RT1u
anti-RT1a) immunization (41), while the
RT1.Aa-specific Ab MAC 30 (IgG2c) was derived from a
PVG-RT1u anti-PVG.R8 immunization (41). NR5/10 is an
RT1.Au-specific alloantibody (IgG2b) derived from
PVG-RT1l anti-AO (RT1l
anti-RT1u) immunization (41). MAC 30 is a partially
conformation-sensitive Ab that in immunoprecipitations favors fully
mature RT1.Aa molecules (42). NR5/10 shows virtually
complete specificity for fully mature RT1.Au molecules (see
Results and Fig. 7, A and C). PA3-900
is a polyclonal rabbit anti-calreticulin antiserum produced by
immunization with recombinant human calreticulin (ABR, Golden, CO).
SPA-600 is a polyclonal rabbit anti-calreticulin antiserum raised
against the C-terminal peptide of human calreticulin (StressGen/Biomol,
Hamburg, Germany). R.gp48N is a polyclonal rabbit anti-peptide Ab
to tapasin corresponding to N-terminal residues 2–20 (14).
|
|
Cells were washed twice in ice-cold PBS before resuspension in lysis buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% digitonin (Fluka, Neu-Ulm, Germany) or 1% Triton X-100 (Sigma), 3 mM MgCl2, 5 mM iodoacetamide (Sigma), and 0.5 mM PMSF (Boehringer Mannheim, Mannheim, Germany)). After 30-min incubation on ice, lysed cells were centrifuged at 200 x g for 15 min. Immunoprecipitations were performed at 4°C with postnuclear supernatants of the cell lysates using anti-TAP2 (116/5), anti-class I (F88 or F87), and anti-calreticulin (SPA-600) Abs in combination with protein A-Sepharose. When MAC 394 was used for immunoisolation, protein A-Sepharose was preloaded with affinity-purified goat anti-mouse IgG (Dianova, Hamburg, Germany). R3/13 and NR5/10 Abs were coupled either directly to cyanogen bromide-activated Sepharose or to protein A-Sepharose (Pharmacia, Freiburg, Germany) preloaded with affinity-purified goat Abs to rat IgG (Dianova). Precipitates were washed five times with PBS containing 0.1% digitonin or 1% Triton X-100. Immune complexes were eluted with 1 ml of 1% acetic acid and 1% Triton X-114 (Sigma) at room temperature (43) followed by phase separation at 37°C. Proteins in the detergent phase were recovered by ethanol/hexane (4/1) precipitation. For isolation of the TAP complex, cell lysates were extracted by stirring with protein A-Sepharose beads conjugated to rabbit anti-TAP antiserum 116/5. After 2 h of incubation and extensive washing in 0.1% digitonin, the Sepharose beads were eluted competitively using 10 µM synthetic peptide (EMBL-732) corresponding to the C terminus of TAP2 (EQDVYAHLVQQRLEA) in 0.1% digitonin in PBS at 4°C. Samples were run on SDS-PAGE (44) under reducing conditions and transferred to nitrocellulose by electroblotting. Before blocking (PBS/5% dry milk powder/0.1% Tween 20) for 16 h at 4°C the membranes were, for some experiments, cut into two parts (between the 46 and 66 kDa molecular mass marker). The upper part was incubated with anti-TAP Abs, and the lower part was incubated with anti-RT1.A Abs. All Abs were diluted in PBS/5% FCS/0.1% Tween-20. Bands were visualized with horseradish peroxidase-conjugated secondary goat anti-rabbit IgG (Dianova) and enhanced chemiluminescence substrate (Amersham, Braunschweig, Germany).
Deglycosylation with endoglycosidase H (endo H) and N-glycosidase F
Endo H (5 mU; Boehringer Mannheim) in 20 µl of incubation buffer (0.1 M 2-ME, 0.01% SDS, and 50 mM sodium citrate, pH 5.5) was added to 10 µl of cell lysate containing 1% Triton X-100 or immunoprecipitated material. Aliquots were incubated for 12 h at 37°C before analysis by SDS-PAGE.
To deglycosylate class I molecules completely, immunoisolated material was preincubated for 5 min at 65°C with 30 µl of 100 mM sodium phosphate buffer, pH 7.5, containing 10 mM EDTA, 1% Triton X-100, 0.3% SDS, and 1% 2-ME. After cooling to room temperature, 4 mU of N-glycosidase F (N-Gase F) (Boehringer Mannheim) was added to each aliquot, and the samples were incubated overnight at 37°C.
Pulse-chase analysis
Cells (1 x 106/ml) were starved for 1.5 h in methionine-free RPMI medium containing 10% dialyzed FCS. Then, [35S]methionine (Amersham) was added for 30 min (100 µCi/ml). The chase was initiated by the addition of excess unlabeled methionine (3 mM). Aliquots were removed at various times of chase, and cells were separated from the supernatant and washed twice in ice-cold PBS before resuspension in lysis buffer (PBS, 1% digitonin, or 1% Triton X-100) containing 5 mM iodoacetamide and 0.5 mM PMSF (Sigma). Immunoprecipitations were performed from equivalent amounts of precleared cell lysates by using anti-TAP Abs (116/5) or allele-specific anti-class I Abs (R3/13, MAC 30, and NR5/10). Precipitates were washed five times with lysis buffer containing 0.1% digitonin (Fluka) or 1% Triton X-100 (Sigma). For identification of coisolated RT1.A alleles, immunoisolated TAP complexes were digested with N-Gase F (see above) and separated by SDS-PAGE under reducing conditions. Gels were stained with Coomassie brilliant blue to control for precipitating Abs. Fluorographs were obtained after different exposure times. For signal quantitation, x-ray films were scanned by microdensitometry using a Joyce-Loebl Chromoscan II (Joyce-Loebl, Gateshead, U.K.). Alternatively, dried gels containing [35S]methionine-labeled proteins were exposed to phosphorimager screens, which were imaged and quantified using the associated hardware and software (Fujix BAS1000, Fuji, Düsseldorf, Germany). Hard copies of digitized fluorographs were produced by using Adobe Photoshop (Adobe Systems, Mountain View, CA) and Canvas (Deneba Software, Miami, FL) software.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As a result of peptide deprivation, the residence of
RT1.Aa in the ER is greatly prolonged if only TAP-B is
available (30, 35, 45). To understand the mechanism of retention of
RT1.Aa in this situation we first compared by Western blot
the steady state distribution of RT1.Aa in the B5 and D7
transfectant cell lines (see Table I). In
contrast to D7, B5 expresses a TAP-A group transporter permissive for
the transport of peptides with the ideal C-terminal arginine.
|
A, tracks 1 and
2) were sensitive to endo H (tracks 3 and
4) and therefore represent immature class I molecules in the
ER/cis-Golgi (30, 35). By densitometry of Fig. 1A
(tracks 3 and 4), comparable amounts of
TAP and total RT1.Aa were present, but the two cell lines
differed strikingly in the allocation of signal between the mature and
immature forms of RT1.Aa (Fig. 1B). The ratio of
ER to post-ER signals of RT1.Aa in D7 is about 3.0, while
in B5 it is about 0.3. From 10 comparable experiments we determined a
mean signal ratio of ER to post-ER RT1.Aa for D7 of
4.5 ± 1.5 and for B5 of 0.27 ± 0.11. Thus in the absence of
TAP-A, RT1.Aa accumulates in the ER.
|
We next asked whether the accumulation of RT1.Aa in
the ER in D7 cells is associated with an increase in RT1.Aa
bound to TAP-B. D7 and B5 cells were lysed in digitonin, and TAP
complexes were immunoprecipitated with anti-TAP2 (116/5) antiserum.
Yields of class I molecules were estimated densitometrically on Western
blots stained with F87 and F88 antiserum (Fig. 2A). A marked relative excess
of RT1.Aa was found associated with TAP in D7 compared with
B5. In the experiment shown, the relative excess is 2.8-fold in favor
of D7 (compare tracks 1 and 2), while total
cellular class I is present in a ratio of 0.8 (compare tracks
4 and 5). In three similar analyses the mean relative
excess of TAP-bound RT1.Aa in D7 was 3.0 relative to B5.
Thus, the failure of TAP-B to deliver suitable peptides for loading
into RT1.Aa induces not only an accumulation of immature
RT1.Aa molecules in the ER, but also accumulation of these
molecules on TAP-B itself.
|
A). A similar result was obtained with cells
biosynthetically labeled for 1 h with
[35S]methionine, from which TAP and associated class I
molecules were immunoprecipitated (Fig. 2B). In C58, which
lacks RT1.Aa, RT1.Au associates strongly with
TAP-B. Similarly, in B5, in which the build-up of RT1.Aa in
the ER is released by the presence of TAP-A, RT1.Au and
RT1.Aa associate efficiently with TAP. In C58.331 and D7,
however, RT1.Au is barely detectable in TAP coprecipitates,
as if accumulation of immature RT1.Aa molecules in the ER
in these cells were competitively inhibiting access of
RT1.Au to TAP. The unidentified, class I-related material
indicated with an asterisk in Fig. 2, A and B,
which may correspond to one or several of the nonclassical class I
molecules coded by the RT1u haplotype, is also apparently
largely competitively inhibited from access to TAP in D7 and
C58.331. TAP-A and TAP-B associate indiscriminately with RT1.Aa and RT1.Au in the cell line B5
We used the TAP2a-specific mAb, MAC 394 (A. Seelig,
manuscript in preparation) to investigate the possibility that
exclusion of RT1.Au from TAP-B in C58.331 and D7 cells is
not related to peptide supply but, rather, is caused by an
intrinsically higher affinity interaction between RT1.Aa
and TAP-B than between RT1.Au and TAP-B. As shown in Fig. 3A, MAC 394 was able to
precipitate TAP only from B5 cells, which contain TAP2a,
and not from C58, which does not. Rabbit anti-TAP2 antiserum 116/5
served as a control.
B5 cells were metabolically labeled for 1 h with
[35S]methionine and were lysed in digitonin buffer. Total
TAP-associated class I was directly immunoprecipitated from one aliquot
with anti-TAP2 antiserum (116/5; Fig. 3B, track
5). Another aliquot was first depleted for TAP-A-associated class
I by three successive rounds of immunoprecipitation with MAC 394.
Immunoprecipitates from all three rounds were kept for analysis (Fig. 3B, tracks 1–3). The residual TAP-B-associated
class I population was then isolated with rabbit anti-TAP2
antiserum (116/5; Fig. 3B, track 4). The
quantitative removal of TAP2-A from the lysate by MAC 394 was confirmed
by a corresponding immunoblot (data not shown). Immunoprecipitates of
RT1.Aa and RT1.Au from further aliquots of the
lysate provided markers for TAP-coisolated RT1.A alleles (Fig. 3B, tracks 6 and 7). Before analysis,
all samples were treated with N-Gase F.
Fig. 3B shows that RT1.Aa and RT1.Au
were coisolated with both TAP-A (track 1) and TAP-B
(track 4). By phosphorimager quantitation, the two
class I alleles bind to the TAP-A transporter with a ratio of about
1:1. Nearly the same binding ratio of class I alleles was obtained for
the TAP-B transporter and for the entire TAP population (Fig. 3B, tracks 4 and 5). A longer exposure
of the gel depicted in Fig. 3B showed that the additional
43-kDa protein from the RT1u haplotype was also coisolated
with the TAP-A transporter (not shown). About fourfold more of both
RT1.A alleles was coisolated with TAP-B than with TAP-A, consistent
with the data shown in Fig. 3A, which indicate TAP2-A to be
clearly <50% of the total TAP2 pool in these cells. In conclusion,
these data from B5 show that the association of RT1.Aa with
TAP-B is not intrinsically favored. Thus, the preferential association
of TAP-B with RT1.Aa in D7 and C58.331 cells must follow
from the RT1.Au-biased peptide delivery available in these
cells. The interaction of an immature class I molecule with TAP is thus
governed also by competitive relations in the ER, not only by intrinsic
differences in the affinity of interaction between individual allelic
products and TAP itself.
In the absence of TAP-A, RT1.Aa accumulates in the free as well as the TAP-B-bound state
We next tested whether the overall accumulation of empty
RT1.Aa molecules in the ER of D7 and C58.331 cells was due
only to the increase in RT1.Aa associated with TAP by
comparing TAP-associated and TAP-unbound immature RT1.Aa
molecules from digitonin lysates of D7 and B5. TAP-associated
RT1.Aa was cleared from lysates by sequential
immunoprecipitation with anti-TAP2 (116/5) antiserum and the
residual immature RT1.Aa quantitated from Western blots.
Fig. 4A shows total
RT1.Aa and TAP2 signals (track a) and
signals after clearance of TAP by immunoprecipitation
(track b) from the two cell lines. In D7 as well as
in B5 TAP was substantially removed from the lysate. From densitometry
(Fig. 4B), approximately 50% of the ER form of
RT1.Aa in D7 cells was not associated with TAP, while in B5
approximately 70% was free. Comparison of the absolute signal
intensities from the two (b) tracks in Fig. 4B
shows a roughly twofold excess of free ER-resident RT1.Aa
molecules in D7 relative to B5. The observed overall increase in
ER-form RT1.Aa in D7 is therefore not caused solely by
increased binding of RT1.Aa to TAP-B. A better description
is that the failure of suitable peptide delivery in these cells results
in a rise in the equilibrium concentration of RT1.Aa in
both TAP-associated and free ER compartments.
|
TAP-B/RT1.Aa complexes recovered from C58.331 and D7
cells also contained calreticulin and tapasin. The TAP complex was
immunoprecipitated from digitonin lysates of D7 cells with
anti-TAP2 antiserum 116/5. Immunoisolated proteins were
eluted competitively with the 116/5-immunizing peptide as described
(see Materials and Methods), electrophoresed, and blotted,
and filter strips were developed with antisera against
RT1.Aa, tapasin, calreticulin, and TAP (Fig. 5A). No attempt was made to
estimate the stoichiometry of components in these complexes. Complexes
of calnexin with TAP and/or RT1.Aa were not identified even
in the presence of cross-linker
dithiobis[succinimidylpropionate] (DSP) (data not shown).
|
) was associated with calreticulin, calreticulin was
quantitatively immunoprecipitated from a digitonin lysate of D7 cells
precleared for TAP with anti-TAP2 antiserum. RT1.Aa
molecules in the lysate before and after precipitation of TAP and
calreticulin were immunoprecipitated with anti-class I antiserum.
Precipitated material was separated by SDS-PAGE, blotted, and probed
with anti-class I antiserum (Fig. 5B, left
panel). Depletion of TAP and calreticulin from the D7 lysate by
immunoisolation was controlled in additional blots probed with
anti-TAP2, anti-TAP1, or anti-calreticulin (Fig. 5B, right panel). Preclearance of TAP was
efficient but had no detectable effect on the amount of calreticulin in
the lysate, indicating that only a small proportion of total
calreticulin is in a TAP complex. Immunoprecipitation with
anti-calreticulin removed this chaperone efficiently from the
lysate.
The results shown in the left panel of Fig. 5B
confirmed our earlier observations (Fig. 4), with about 40% of
immature RT1.Aa in D7 not complexed with TAP-B (compare
track 2 with the sum of tracks 3 and
4). Roughly 15% of ER-localized free RT1.Aa
molecules were coprecipitated with calreticulin (compare track
3 with track 4). Other groups have shown a relative
instability of calreticulin in TAP- and/or tapasin-free class I
complexes (14, 46). High affinity binding of calreticulin to class I
molecules may depend on the presence of TAP and tapasin in the complex.
The small proportion of TAP-free RT1.Aa molecules
coprecipitated with calreticulin may therefore be substantially lower
than the equilibrium value in vivo.
Kinetics of class I/TAP interactions in the presence and the absence of TAP-A
The results presented in Fig. 4 showed an excess of empty
RT1.Aa molecules in both the TAP-bound and free ER pools in
D7 cells. We therefore sought to compare the kinetics of TAP-class I
interaction in D7 and C58.331, in which a large ER pool of
RT1.Aa existed, with that in B5, in which the ER pool is
substantially smaller. Pulse-chase analysis in Fig. 6 shows that the apparent half-life
(defined as the time taken to reach 50% of the highest signal obtained
during the chase) of TAP-class I complexes of RT1.Aa was
longer in C58.331 and D7 (
4 h) than in B5 (
70 min). Thus, in
C58.331 and D7 large amounts of pulse-labeled RT1.Aa
molecules can still be immunoprecipitated with TAP after 4.5 h of
chase. Consistent with the presence of a relatively large preexisting
pool of immature RT1.Aa in these cells, the initial sp.
act. of RT1.Aa coprecipitated with TAP is markedly lower in
D7 and C58.331 than in B5. Coupled with the long apparent half-life of
TAP-precipitable RT1.Aa, this suggests that in these cells
TAP is sampling RT1.Aa from an equilibrium pool that is
relatively stable. In B5, in contrast, the cohort of labeled
RT1.Aa molecules appears to pass rapidly through a
relatively small free pool. This results in high specific activities
associated with TAP and a short apparent half-life in the
TAP-precipitable pool, consistent with the labeled cohort being rapidly
washed out of the system as pulse-labeled molecules successfully load
with TAP-A-derived peptides and leave the ER. In neither D7 nor in
C58.331 cells could detectable amounts of labeled RT1.Au be
coisolated with TAP (compare Fig. 2). In B5, however,
RT1.Au follows kinetics of TAP association and dissociation
virtually identical with those of RT1.Aa (Fig. 6B), suggesting that the two allelic products are not
markedly different in their maturation and loading so long as an
appropriate peptide supply is available for both, as in B5. It further
follows from this result that RT1.Aa molecules that
associate with TAP-B molecules in B5 cells (see Fig. 3) do not
accumulate, in contradistinction to their behavior in D7 and C58.331
cells. This effect is discussed below (see Discussion).
|
We next investigated how the maturation of RT1.Au
molecules is affected by their failure to interact significantly with
TAP in D7 and C58.331 cells. For immunoprecipitation we used two
allele-specific mAbs, MAC 30 (anti-RT1.Aa) and NR5/10
(anti-RT1.Au), which, as the Western blot in Fig. 7A shows, predominantly
recognize mature class I molecules (referred to as the post-ER form).
To analyze maturation, C58.331, D7, and B5 cells were pulse labeled
with [35S]methionine and chased for various times.
RT1.Aa and RT1.Au were immunoprecipitated from
Triton X-100 lysates and analyzed by SDS-PAGE (Fig. 7, B and
C). In B5, RT1.Aa showed rapid maturation as
indicated by the rapid rise in label associated with the mature
(post-ER) heavy chain, in contrast to the slower maturation observed in
C58 and D7 cells (Fig. 7B) and as previously reported (30).
The absolute increase in total precipitated label during the chase
reflects the preference of MAC 30 for mature RT1.Aa
molecules, but late ER forms of RT1.Aa are also apparent.
During the first 90 min of chase we calculated that in C58 and D7,
RT1.Aa maturation is retarded by a factor of 3
relative to B5. In contrast to the situation with RT1.Aa,
the maturation of RT1.Au was essentially identical in
B5, D7, and C58.331 cells (Fig. 7C). The strong
preference of NR5/10 for post-ER forms of RT1.Au is
reflected in the large increase in total precipitated label during the
chase and the very low signal corresponding to an ER form.
In consistency with the maturation data, we found by FACS analysis that RT1.Au is expressed apparently normally at the cell surface of all C58 transfectants (data not shown), while RT1.Aa is significantly underexpressed in C58.331 and D7 relative to B5 (38). In summary, these results demonstrate that despite being excluded from normal TAP association in D7 and C58.331 cells, RT1.Au matures normally and reaches the cell surface with normal kinetics.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Newly synthesized class I molecules face a situation in the ER where access to suitable peptides may well be severely limited. Class I alleles that require relatively rare classes of peptides for loading may have to wait longer than others. Furthermore, free peptides in the ER are rapidly removed (19, 20, 21, 22, 23), no doubt at rates that are at least in part dependent on the sequences of the peptides themselves. Thus, differential TAP association as observed by Neisig et al. (26) in human lymphoblastoid cell lines and attributed to differential strength of the primary class I-TAP interaction, could also result from disproportionate availability of suitable class I-binding peptides. These authors have, indeed, very recently come to this view themselves in the case of the strong TAP binder HLA-Cw4 (51), whose deficiency at the cell surface and persistence in the ER are reminiscent of RT1.Aa expressed in the absence of a permissive TAP, as in our present and earlier experiments (30, 35, 45).
The results of the present study show that if the delivery of class
I-binding peptides is sufficiently unbalanced, competition for access
to TAP may develop between allelic class I variants. Specifically, we
found that the rat class I molecule RT1.Au was excluded
from its normal interaction with TAP-B in the presence of a second
class I allele, RT1.Aa (Figs. 2 and 6), which is poorly
loaded with peptides delivered by TAP-B (30, 35). In this situation,
RT1.Aa accumulated in the ER in both the free and TAP-bound
states (Figs. 1, 2, and 5). This behavior resembled that of class I
molecules following reduction of peptide supply from the cytosol (15).
Therefore, we concluded that the ER accumulation of peptide-starved
RT1.Aa molecules competitively interfered with the access
of RT1.Au to TAP. In support of this, we could demonstrate
that accumulation and long term residence in the ER were eliminated for
RT1.Aa when peptide delivery optimal for both class I
alleles was provided by TAP2-A transfection (Figs. 1, 2, and 6). In
this situation both class I alleles were recovered in transient
association with TAP (Fig. 6), and neither showed significant ER
accumulation. Since the two allelic forms of rat TAP, TAP-A and TAP-B,
apparently associated promiscuously and equally with the two class I
alleles (Fig. 3), the observed effects could not be attributed to a
high affinity interaction of TAP-B with RT1.Aa acting to
exclude a low affinity interaction with RT1.Au.
Our data suggest that access of class I molecules to TAP is also
defined by the particular class I alleles present and by the
availability of suitable class I ligands in the ER, and not only by
variations in the strength of the primary interaction between different
class I alleles and TAP. In the case of RT1.Au, we assume
that suitable peptides for loading are efficiently transported into the
ER by the TAP-B transporter. In the normal case, as in wild-type C58
cells, access of RT1.Au to TAP is not impeded, and most
newly synthesized RT1.Au molecules probably associate with
TAP before they load with peptide in a typical TAP-associated loading
complex. When RT1.Aa is also present, as in D7 or C58.331
cells, its accumulation in the ER and also its extended association
with TAP-B cause a drastic reduction in the access of TAP-B for newly
synthesized RT1.A alleles. Pulse-chase analysis demonstrated for D7 and
C58.331 that newly synthesized RT1.Au molecules are
apparently largely excluded from TAP binding (Fig. 2B and
6A), and the initial specific activity of newly synthesized
RT1.Aa coprecipitated with TAP is markedly lower than that
in B5 (Fig. 6B). The reduction in TAP-RT1.Au
association in D7 and C58.331 compared with C58 and B5 cannot be
explained in terms of a change in the intrinsic affinity of the
TAP-RT1.Au interaction itself. However, our findings are
consistent with the idea that in D7 and C58.331 a normal proportion of
newly synthesized RT1.Au molecules interacts with TAP for a
greatly reduced length of time. Alternatively, a smaller proportion of
newly synthesized RT1.Au molecules may associate with TAP
with normal kinetics. In the former case, loading of RT1.Au
may nevertheless be mediated by physical TAP binding even in the
presence of RT1.Aa, while in the latter case, loading must
be largely distant from TAP. We tend to prefer the second possibility
and hypothesize that in the situation of D7 and C58.331, the average
time required for newly synthesized RT1.Au molecule to
associate with TAP is longer than the average time required to find a
suitable peptide in the ER pool. Without affecting the maturation
process, RT1.Au molecules thus behave as if they are
outcompeted for TAP association by peptide-starved RT1.Aa
molecules. Despite being largely excluded from TAP association,
however, RT1.Au matures and is expressed normally at the
cell surface. Experiments with the proteasome inhibitor
N-acetyl-L-leucyl-L-leucyl-L-norleucinal
(data not shown) have shown that peptide delivery is a prerequisite for
maturation of RT1.Au in D7 cells, and therefore that
RT1.Au molecules transported to the cell surface are loaded
with peptide. Whether a distinct array of peptides is loaded into
RT1.Au when it is competitively excluded from TAP
association has yet to be determined.
Our data taken together speak in favor of a dynamic model in which
TAP-class I association represents sampling by TAP of the pool of free
RT1.A molecules available in the ER. In the case of the D7 and C58.331
cells, the pool of free molecules is disproportionately composed of
RT1.Aa because of their difficulty in acquiring a suitable
peptide. RT1.Au molecules do not accumulate in the ER
because they acquire free peptides and subsequently leave the
compartment (Fig. 7). In contrast, in B5, peptide supply and the pool
of ER-resident RT1.A molecules seem to be more balanced (Fig. 7). The
ratio of RT1.Aa and RT1.Au molecules associated
with TAP thus largely reflects the equilibrium composition of class I
molecules in the ER. In the case of D7 and C58.331, it may be still
further biased in favor of RT1.Aa if RT1.Au
molecules associated with TAP leave the TAP complex earlier, on the
average, than RT1.Aa molecules as a result of efficient
loading with suitable peptides.
The behavior of RT1.Au in D7 and C58.331 cells resembles that of HLA-B56 and HLA-B60 in the human cell line 4778 (26). It is tempting to assume that the proposed competition effect of TAP-class I interaction demonstrated in our experiments may also contribute to the apparent HLA allele specificity of TAP association (26). As noted above, Neisig et al. (51) have recently presented direct evidence that the strong association of HLA-Cw4 with TAP may be due to a relative deficiency in the provision of suitable peptides for loading. It should be noted that our findings in rat do not rule out that class I alleles of other species may have differential affinities for TAP-tapasin complexes. However, possible competitive effects of persistent ER-resident HLA-Cw4 on TAP association of other HLA alleles would be worth investigation.
Our experiments in the cell line B5 showed that the presence of a
functional TAP-A transporter releases RT1.Aa from its
extended confinement in the ER and simultaneously restores normal TAP
interaction to RT1.Au (Fig. 2). In these cells,
RT1.Aa showed the same transient TAP association as
RT1.Au, which is capable of loading with peptides delivered
by the restrictive TAP-B transporter (42) and most likely also by the
permissive TAP-A transporter. Since subsequent experiments showed that
association of TAP and RT1.A alleles is random (Fig. 3),
RT1.Aa molecules in these cells must be making a transient
association with TAP-B as well as with TAP-A even though only TAP-A can
deliver suitable peptides. Evidently, RT1.Aa molecules
associating randomly with TAP-B can nevertheless be loaded by peptides
delivered by TAP-A. It may be that RT1.Aa loads with a
suitable luminal peptide while still associated with TAP-B and is then
released as a result of an indirect loading event (11, 52). Our own
data show that indeed RT1.Aa can be released from TAP-B
association in detergent solution by addition of a suitable exogenous
peptide (M. R. Knittler and J. C. Howard, manuscript in
preparation), but there is no direct evidence even in the presence of
TAP-A for the existence of a free luminal pool of peptides suitable for
loading into RT1.Aa. A second possibility is that in the
absence of indirect peptide loading, the interaction between TAP and
class I is a dynamic and cyclic process of binding and release.
Released but still unloaded class I molecules could reassociate with a
second TAP, whereas class I molecules that bind an appropriate peptide
would be removed from this oscillating process. Thus,
RT1.Aa would successfully load only after release from
TAP-B and subsequent reassociation with TAP-A. If the mean half-life of
TAP-class I complexes in the absence of peptide is relatively short,
such an exchange process would not cause a significant delay in the
mean loading time of RT1.Aa molecules. It should be noted
that if TAP-class I interactions were also labile under in vitro
immunoprecipitation conditions (4°C, detergent lysate, etc.), the
quantitation and interpretation of coisolated components would be
difficult. In fact, our own experience and that of others (11, 12, 15)
is that immunoprecipitated TAP-class I complexes are remarkably stable
in vitro in the presence of mild detergents. We would certainly not
believe that this stability under in vitro conditions speaks against
dynamic behavior in vivo. We therefore suggest that our quantitation of
free and bound RT1.Aa, shown in Figs. 4 and 5, more or less
correctly reflects a snapshot of the in vivo situation.
Our experimental findings provide evidence that ER retention of
RT1.Aa is not directly effected by physical complex
formation with TAP-B, since Western blot analysis of TAP-depleted
digitonin lysates showed that in D7 about half of the ER-accumulated
RT1.Aa molecules were not found in association with TAP-B
at any given moment (Figs. 4 and 5). We cannot exclude the possibility
that a portion of the TAP complex is disrupted upon cell solubilization
and Ab binding. However, the result of the Western blot is in agreement
with the pulse-chase experiments (Fig. 6), which demonstrated that TAP
samples RT1.Aa posttranslationally from an ER-resident free
pool. A small proportion of these ER-resident TAP-free
RT1.Aa molecules could be shown to be associated stably
with calreticulin (Fig. 5), in agreement with Van Leeuwen and Kearse
(53). Class I molecules undergo a complex series of interactions with
the ER-resident chaperones calnexin, calreticulin, and tapasin before
or during assembly with TAP (14, 15, 16, 17, 52), and all three accessory
proteins contain short amino acid sequences characterized as
ER-specific retention signals (16, 17, 54, 55, 56, 57). The prevailing model
asserts that the early chaperone, calnexin, is displaced in favor of
calreticulin by ß2m binding. This chaperone remains
associated when class I, in turn, binds to the TAP-tapasin complex. It
has been noted before that free calreticulin-class I complexes could
exist in equilibrium with TAP-associated class I complexes until
peptide loading occurs (14). Binding to calreticulin may directly
retain free, peptide-receptive, class I molecules in the ER (58). In
support of this, it has been shown that in the absence of TAP and/or
tapasin, calreticulin is associated with ER-resident class I molecules
(14, 18). The interaction between calreticulin and class I seems to be
strengthened by the presence of TAP and tapasin in the complex (14, 46), which may explain why only about 15% of TAP-free
RT1.Aa molecules were coisolated with calreticulin (Fig. 5).
Further work is needed to understand the adaptive significance of this complicated situation. The description we have given still leaves the function of TAP association unclear, and other recently published data further confuse the issue. Most surprisingly, when tapasin is expressed in mutant cells in a soluble form, the block of class I loading and maturation is apparently fully released despite the absence of any detectable association with TAP (59). In these cells, therefore, at least superficially there is no residual defect. Tapasin must, therefore, act as a chaperone aiding loading, whether free or TAP associated, while the advantage of TAP association itself perhaps lies in the more sensitive sampling of rare and possibly labile peptides arising from cytosolic peptidase activity that may have very short ER luminal half-lives and very low ER luminal concentrations. Such peptides would be grossly under-represented in the pool of free luminal peptides loaded by class I molecules unable to associate with TAP, while their opportunity to load on a TAP-associated class I molecule is transiently equal to that of all other peptides. While this makes adaptive sense and is perhaps consistent with early evidence that free cellular pools of certain class I-loadable peptides are essentially absent (19), direct evidence for the existence of such TAP association-dependent peptides is lacking.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Michael R. Knittler, Institute for Genetics, University of Cologne, Zülpicherstr. 47, D-50674 Cologne, Germany. E-mail address:
3 Present address: Biochemiezentrum Heidelberg, University of Heidelberg, Im Neuenheimer Feld 328, D-69120 Heidelberg, Germany.
4 Abbreviations used in this paper: ER, endoplasmic reticulum; endo H, endoglycosidase H; N-Gase F, N-glycosidase F.
Received for publication April 24, 1998. Accepted for publication August 3, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. C. Verweij, D. Koppers-Lalic, S. Loch, F. Klauschies, H. de la Salle, E. Quinten, P. J. Lehner, A. Mulder, M. R. Knittler, R. Tampe, et al. The Varicellovirus UL49.5 Protein Blocks the Transporter Associated with Antigen Processing (TAP) by Inhibiting Essential Conformational Transitions in the 6+6 Transmembrane TAP Core Complex J. Immunol., October 1, 2008; 181(7): 4894 - 4907. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Rufer, R. M. Leonhardt, and M. R. Knittler Molecular Architecture of the TAP-Associated MHC Class I Peptide-Loading Complex J. Immunol., November 1, 2007; 179(9): 5717 - 5727. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Tourdot, M. Nejmeddine, S. J. Powis, and K. G. Gould Different MHC Class I Heavy Chains Compete with Each Other for Folding Independently of {beta}2-Microglobulin and Peptide J. Immunol., January 15, 2005; 174(2): 925 - 933. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Ehses, R. M. Leonhardt, G. Hansen, and M. R. Knittler Functional Role of C-Terminal Sequence Elements in the Transporter Associated with Antigen Processing J. Immunol., January 1, 2005; 174(1): 328 - 339. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Zernich, A. W. Purcell, W. A. Macdonald, L. Kjer-Nielsen, L. K. Ely, N. Laham, T. Crockford, N. A. Mifsud, M. Bharadwaj, L. Chang, et al. Natural HLA Class I Polymorphism Controls the Pathway of Antigen Presentation and Susceptibility to Viral Evasion J. Exp. Med., November 8, 2004; (2004) jem.20031680. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Ford, A. Antoniou, G. W. Butcher, and S. J. Powis Competition for Access to the Rat Major Histocompatibility Complex Class I Peptide-loading Complex Reveals Optimization of Peptide Cargo in the Absence of Transporter Associated with Antigen Processing (TAP) Association J. Biol. Chem., April 16, 2004; 279(16): 16077 - 16082. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Tourdot and K. G. Gould Competition Between MHC Class I Alleles for Cell Surface Expression Alters CTL Responses to Influenza A Virus J. Immunol., November 15, 2002; 169(10): 5615 - 5621. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. C. Smith, A. Gallimore, E. Jones, B. Roberts, J. M. Lord, E. Deeks, V. Cerundolo, and L. M. Roberts Exogenous Peptides Delivered by Ricin Require Processing by Signal Peptidase for Transporter Associated with Antigen Processing-Independent MHC Class I-Restricted Presentation J. Immunol., July 1, 2002; 169(1): 99 - 107. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Tan, H. Kropshofer, O. Mandelboim, N. Bulbuc, G. J. Hammerling, and F. Momburg Recruitment of MHC Class I Molecules by Tapasin into the Transporter Associated with Antigen Processing-Associated Complex Is Essential for Optimal Peptide Loading J. Immunol., February 15, 2002; 168(4): 1950 - 1960. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Velarde, R. C. Ford, M. F. Rosenberg, and S. J. Powis Three-dimensional Structure of Transporter Associated with Antigen Processing (TAP) Obtained by Single Particle Image Analysis J. Biol. Chem., November 30, 2001; 276(49): 46054 - 46063. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Lybarger, Y. Y. L. Yu, T. Chun, C.-R. Wang, A. G. Grandea III, L. Van Kaer, and T. H. Hansen Tapasin Enhances Peptide-Induced Expression of H2-M3 Molecules, but Is Not Required for the Retention of Open Conformers J. Immunol., August 15, 2001; 167(4): 2097 - 2105. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. W. Purcell, J. J. Gorman, M. Garcia-Peydro, A. Paradela, S. R. Burrows, G. H. Talbo, N. Laham, C. A. Peh, E. C. Reynolds, J. A. Lopez de Castro, et al. Quantitative and Qualitative Influences of Tapasin on the Class I Peptide Repertoire J. Immunol., January 15, 2001; 166(2): 1016 - 1027. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Barnden, A. W. Purcell, J. J. Gorman, and J. McCluskey Tapasin-Mediated Retention and Optimization of Peptide Ligands During the Assembly of Class I Molecules J. Immunol., July 1, 2000; 165(1): 322 - 330. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Stevens, K.-H. Wiesmuller, G. W. Butcher, and E. Joly Analysis of peptide length preference of the rat MHC class Ia molecule RT1-Au, by a modified random peptide library approach Int. Immunol., January 1, 2000; 12(1): 83 - 89. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. A. Peh, N. Laham, S. R. Burrows, Y. Zhu, and J. McCluskey Distinct Functions of Tapasin Revealed by Polymorphism in MHC Class I Peptide Loading J. Immunol., January 1, 2000; 164(1): 292 - 299. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Y. L. Yu, N. B. Myers, C. M. Hilbert, M. R. Harris, G. K. Balendiran, and T. H. Hansen Definition and transfer of a serological epitope specific for peptide-empty forms of MHC class I Int. Immunol., December 1, 1999; 11(12): 1897 - 1906. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Y. L. Yu, H. R. Turnquist, N. B. Myers, G. K. Balendiran, T. H. Hansen, and J. C. Solheim An Extensive Region of an MHC Class I {alpha}2 Domain Loop Influences Interaction with the Assembly Complex J. Immunol., October 15, 1999; 163(8): 4427 - 4433. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
TAP2a, tracks 1–3)
followed by precipitation with anti-TAP2 antiserum (116/5; 
