Tapasin influences the quantity and quality of MHC/peptide complexes at the cell surface; however, little is understood about the structural features that underlie its effects. Because tapasin, MHC class I, and TAP are transmembrane proteins, the tapasin transmembrane/cytoplasmic region has the potential to affect interactions at the endoplasmic reticulum membrane. In this study, we have assessed the influence of a conserved lysine at position 408, which lies in the tapasin transmembrane/cytoplasmic domain. We found that substitutions at position K408 in tapasin affected the expression of MHC class I molecules at the cell surface, and down-regulated tapasin stabilization of TAP. In addition to affecting TAP interaction with tapasin, the substitution of alanine, but not tryptophan, for the lysine at tapasin position 408 increased the amount of tapasin found in association with the open, peptide-free form of the HLA-B8 H chain. Tapasin K408A was also associated with more folded, β2-microglobulin-assembled HLA-B8 molecules than wild-type tapasin. Consistent with our observation of a large pool of tapasin K408A-associated HLA-B8 molecules, the rate at which HLA-B8 migrated from the endoplasmic reticulum was slower in tapasin K408A-expressing cells than in wild-type tapasin-expressing cells. Thus, the alanine substitution at position 408 in tapasin may interfere with the stable acquisition by MHC class I molecules of peptides that are sufficiently optimal to allow MHC class I release from tapasin.
Several proteins associate concurrently in an assembly complex with the class I/β2-microglobulin (β2m)3 heterodimer in the endoplasmic reticulum (ER), including tapasin, TAP, calreticulin, and ERp57 (1). Tapasin binds directly to the MHC class I H chain in 1:1 stoichiometry (2, 3). In tapasin knockout mice, thermally unstable MHC class I molecules are expressed at the cell surface, indicating that peptides have bound poorly to the MHC molecules (4, 5). In addition, the overall level of MHC class I cell surface expression is reduced in tapasin knockout mice (4, 5). The detrimental effects of tapasin loss on MHC class I assembly have functional consequences, because tapasin knockout mice have low numbers of T cells and a poor antiviral T cell response (4, 5).
In 721.220, a human B lymphoblastoid cell line with a tapasin defect, TAP peptide binding is diminished in comparison with TAP peptide binding in the closely related, but tapasin-expressing, 721.221 cell line (6, 7, 8, 9, 10). TAP binding to the MHC class I H chain is dependent on tapasin (6, 11). The level of TAP is increased by tapasin transfection into 721.220 cells, and the stabilization of TAP by tapasin is associated with an overall increase in the amount of peptide translocated into the ER (12, 13).
Despite some understanding of tapasin’s functions that has been provided by studies of the knockout models, current knowledge of the molecular mechanisms that underlie tapasin’s functions is limited. Structurally, human tapasin is an ER-resident, 428-aa, type I transmembrane protein (3, 14, 15). Tapasin has a region homologous to the Ig constant region, which classifies it within the Ig gene superfamily (3). In a previous study that focused on the role of the tapasin transmembrane/cytoplasmic region, it was found that truncated tapasin (aa 1–393), lacking the transmembrane and cytoplasmic domains, was unable to bridge HLA-B8 to TAP, but did rescue MHC class I surface expression (12). This finding was interpreted as suggesting that tapasin promotion of MHC class I assembly is independent of tapasin interaction with TAP (12). A more recent study analyzed the effects on MHC class I assembly of three soluble human tapasin mutants having different C-terminal truncation sites and a mutant with a single substitution in the transmembrane/cytoplasmic region (tapasin L410F) (16). In this study, the C-terminal portion of tapasin was confirmed to be a TAP interaction site, but it was also found that the MHC class I molecules assembled in the presence of these truncated tapasin mutants were unstable and poorly recognized by peptide-dependent Abs, compared with MHC molecules assembled in the presence of wild-type tapasin (16). These latter results suggest that it is necessary for tapasin to bring MHC class I H chains into association with TAP for optimal peptide loading to occur.
Within the tapasin transmembrane/cytoplasmic region at position 408 is a highly conserved lysine that is present in human, rat, and mouse tapasin (3, 14, 17). This amino acid residue is also conserved in zebrafish tapasin, and the equivalent position is occupied by another positively charged amino acid residue (arginine) in chicken tapasin. The precise location of the tapasin transmembrane domain is not known, but on the basis of hydrophobicity values, it has been predicted to be positions 393–417, or even possibly 393–407 (3). Therefore, the conserved lysine at position 408 likely lies within the transmembrane domain or at its border (Table I). In this study, we have analyzed the impact of this conserved amino acid residue on the function of tapasin. Our data indicate that tapasin position 408 plays a pivotal role, affecting protein interactions within the assembly complex.
Sequence from the human tapasin transmembrane/cytoplasmic region with predicted transmembrane regions bracketed. Ortmann et al. (3 ) suggested the transmembrane region of tapasin consisted of either residues 393–417 or 393–407. The charged lysine at position 408 is indicated in bold.
Materials and Methods
HC10 is a mAb directed against unfolded human class I, especially HLA, -B and -C (18, 19, 20). HC10+ HLA class I is associated with tapasin and other members of the assembly complex (20, 21, 22). W6/32 is a conformation-dependent Ab that detects folded, β2m-associated HLA class I (20, 23, 24). The 64-3-7 and 30-5-7 Abs detect open, peptide-free, and folded, peptide-occupied Ld, respectively (25). Rabbit antisera specific for the N terminus of human tapasin and a hamster anti-mouse tapasin mAb (gracious gifts from Dr. T. Hansen (Washington University, St. Louis, MO)) were used for Western blotting. Abs specific for human TAP, i.e., mAb 148.3 (26) and rabbit anti-human TAP serum (21), were also used. The mAb against ERp57 and rabbit anti-calreticulin serum were purchased from Stressgen.
Human tapasin-deficient cell lines
The 721.220 cell line is a human B lymphoblastoid cell line that was originally thought to express no endogenous tapasin. More recently, this cell line was found to express a mutant tapasin with a truncated leader peptide and without the N-terminal 49 aa that normally follow the tapasin leader sequence (6, 7, 8, 9). This endogenous tapasin mutant is translocated into the ER very inefficiently (and, therefore, is expressed poorly in the ER), and can associate with TAP but not with the MHC class I H chain. The 721.220 and 721.220-B8 cell lines were generous gifts from Drs. T. Spies (Fred Hutchinson Cancer Research Center, Seattle, WA) and P. Cresswell (Yale University, New Haven, CT). The 721.220 cell line expresses HLA-C; however, HLA-C molecules compose a very minimal fraction (only ∼5%) of the total MHC class I molecules present in HLA-transfected 721.220 cells (16, 27, 28).
The human and mouse tapasin mutants were made from wild-type human and mouse tapasin cDNA templates (14, 29) (gracious gifts from Dr. P. Wang (Queen Mary College, University of London, London, U.K.)) using mutagenic oligonucleotides and the QuikChange procedure (Stratagene). The human wild-type and mutant tapasin cDNAs, in the pREP10 vector (Invitrogen Life Technologies), were transfected into the G418-resistant 721.220-B8 cells by electroporation. Clones expressing wild-type or mutant tapasin, as well as HLA-B8, were first selected by incubation of the electroporated cells in medium containing hygromycin and G418, and then Western blotting for both tapasin and B8 was used to select for transfectants that expressed the same amounts of wild-type or mutant tapasin and the same level of B8. Wild-type human tapasin and tapasin K408A, in the pREP10 vector, were also each electroporated into G418-resistant 721.220-Ld cells, and clones matched for both tapasin and Ld expression were selected by Western blotting. Mouse wild-type and mutant tapasin cDNA were cloned into the pMIN retroviral vector (gift from Dr. T. Hansen), packaged using 293E cells, and transduced into a mouse fibroblast cell line derived from a tapasin knockout mouse (a kind gift from Drs. A. Grandea III and L. Van Kaer (Vanderbilt University School of Medicine, Nashville, TN)). These mouse cells were also transduced with Ld in the pLXSH retroviral vector (30). Western blotting for both tapasin and Ld was used to screen transfectants so that clones expressing the same amounts of wild-type or mutant tapasin and of Ld could be selected.
Immunoprecipitations and Western blots
Immunoprecipitations and Western blotting were performed by a method similar to a published protocol (31). For protein immunoprecipitations, the cells were washed in PBS containing 20 mM iodoacetamide (Sigma-Aldrich) three times and lysed in CHAPS lysis buffer. The CHAPS buffer contained 1% CHAPS (Roche Applied Science) in Tris-buffered saline (pH 7.4) with freshly added 0.2 mM PMSF and 20 mM iodoacetamide and a saturating amount of mAb. After 1 h on ice, the lysates were centrifuged to remove cell nuclei and incubated with protein A-Sepharose beads (Amersham Biosciences). The beads were washed in 0.1% CHAPS/20 mM iodoacetamide in TBS (pH 7.4) four times and boiled in 0.125 M Tris (pH 6.8)/2% SDS/12% glycerol/0.02% bromphenol blue to elute the proteins.
The eluted immunoprecipitates were electrophoresed on SDS-PAGE gels (Invitrogen Life Technologies) and transferred to Immobilon-P membranes (Millipore) for Western blots. After overnight blocking in reconstituted dry milk, membranes were incubated in diluted Ab for 2 h, washed three times with 0.05% Tween 20/PBS, and incubated for 1 h in a dilution of biotin-conjugated goat anti-mouse or anti-rabbit IgG (Caltag Laboratories). After three 0.05% Tween 20/PBS washes, the membranes were incubated with streptavidin-conjugated HRP (Zymed) for 1 h, washed with 0.3% Tween 20/PBS three times, and incubated with ECL Western blot developing reagents (Amersham Biosciences). The membranes were exposed to Kodak BioMax film (Eastman Kodak).
When Western blots were performed on cell lysates without an immunoprecipitation step, the samples were processed before electrophoresis as follows. The cells were washed in PBS containing 20 mM iodoacetamide (Sigma-Aldrich) three times and lysed in buffer containing 0.125M Tris (pH 6.8), 2% (w/v) SDS, 12% (v/v) glycerol, 0.02% (w/v) bromophenol blue, and fresh 0.2 mM PMSF without Ab added. The lysates were incubated 1 h on ice, then centrifuged to pellet nuclear material. Samples of the supernatants were boiled before loading onto gels.
Endoglycosidase H (Endo H) assays
The Endo H experiment to analyze whether the mutant tapasin was retained in the ER was performed by a method similar to that described by Tan et al. (16). Cells (1 × 106) were lysed in TBS (pH 7.4) containing 1% Nonidet P-40, and the lysates were held at 4°C for 30 min. The lysates were cleared by centrifugation and the supernatants were mock digested or digested overnight at 37°C with 10 mU of Endo H (ICN Biomedicals) in 50 mM sodium acetate (pH 5.5). The samples were electrophoresed on acrylamide Tris-glycine gels, and Western blots were performed as described above.
For the pulse-chase experiment to monitor the egress of W6/32-positive HLA class I molecules, 721.220-B8+tpsn and 721.220-B8+tpsn K408A cells were grown in methionine- and cysteine-free medium for 30 min and then radiolabeled with [35S]methionine/cysteine for 30 min. Labeled cells were either washed with PBS containing 20 mM iodoacetamide and placed at −80°C or were washed and incubated in medium containing excess nonradioactive methionine and cysteine for up to 4 h.
Following the appropriate chase intervals, the labeled cells were washed with PBS/iodoacetamide, and lysed with 1% CHAPS containing iodoacetamide, PMSF, and W6/32 mAb. Lysates were incubated on ice for 1 h and then centrifuged. The clarified lysates were incubated with protein A-Sepharose beads (Amersham Biosciences) for 45 min. The beads were washed extensively, and then the precipitated protein was eluted from the beads by boiling in 50 μl of 1× SDS-PAGE buffer for 2 min. An 8-μl aliquot of 0.5 M sodium citrate buffer (0.5 M sodium citrate (pH 5.5)/0.38% SDS/0.14 M 2-ME) was added to each eluate, and the volume was adjusted to 80 μl with dH2O.
The eluates were then divided into two aliquots, one of which was treated with 1 mU of Endo H. All samples were incubated overnight at 37°C, boiled for 5 min in 4× protein elution buffer with 8% 2-ME, and loaded onto 10% acrylamide Tris-glycine gels. Following electrophoresis, proteins were transferred to Immobilon-P membranes (Millipore), which were subsequently dried and autoradiographed. Bands were quantitated with a Molecular Dynamics Storm 860 (Amersham Biosciences).
In flow cytometry assays, cells were suspended at 5 × 106/ml in PBS with 0.2% BSA and 0.1% sodium azide. Cell suspension aliquots in volumes of 0.1 ml were distributed to wells in a 96-well plate. The cells were incubated with excess mAb or with BSA/azide/PBS alone (as a control) at 4°C for 30 min, washed twice, and incubated with a PE-conjugated, Fc-specific F(ab′)2 portion of goat anti-mouse IgG (Jackson ImmunoResearch) at 4°C for 30 min. The cells were washed three times, resuspended in BSA/azide/PBS, and analyzed on a FACSCalibur flow cytometer (BD Biosciences). Statistical analyses were done with the CellQuest software (BD Biosciences). For testing cell surface stability of HLA-B8 molecules, 10 μg/ml brefeldin A treatments with subsequent flow cytometry assays were performed, with treatment durations of 0, 10, or 14 h.
Tapasin K408A and K408W mutants were expressed at a level equivalent to wild-type tapasin
The tapasin transmembrane/cytoplasmic region has potential importance to tapasin function via influencing assembly complex protein interactions at the ER membrane. For the purpose of analyzing the specific role of a charged amino acid residue in this region, we created human and murine tapasin mutant cDNAs with amino acid substitutions at position 408. These human and murine mutant cDNAs were transfected into human tapasin-deficient 721.220-B8 cells and mouse tapasin knockout cells (MF-Ld), respectively, in parallel with wild-type human or murine tapasin, as controls. Tapasin K408A and tapasin K408W were expressed in our transfectants in a stable manner, in quantity similar to wild-type tapasin (Fig. 1). Unlike wild-type tapasin, tapasin K408A, or tapasin K408W, two additional human tapasin mutants that we generated for comparison, K408D and A400K/K408A, appeared as multiple bands of slightly increased molecular mass on Western blots probed with anti-tapasin serum, suggesting that these mutants may adopt multiple alternative conformations or have additional modifications (data not shown). Thus, the character of the amino acids in the transmembrane/cytoplasmic region may affect the conformation or posttranslational processing of tapasin.
Tapasin K408 mutants had reduced ability to stabilize TAP
Wild-type tapasin binds to TAP and stabilizes it (6, 11, 12, 13). Compared with wild-type human tapasin, the ability of human tapasin K408A to increase the steady-state level of TAP was impaired, although not completely abrogated (Fig. 2), which indicates that the K408 position in tapasin is important to the capacity of human tapasin to stabilize TAP. A relatively low amount of tapasin K408A and HLA-B8 coprecipitated with TAP from lysates of 721.220-B8+tpsn K408A cells (Fig. 2). We sought to confirm that less tapasin K408A was visible on the tapasin blot because of reduced TAP/tapasin association, and not because of the relatively low level of stabilized TAP available for immunoprecipitation in 721.220-B8+tpsn K408A lysates. To do this, we diluted an aliquot of the 721.220-B8+tpsn anti-TAP immunoprecipitate to the point at which the level of TAP was equivalent to that immunoprecipitated from 721.220-B8+tpsn K408A cell lysates, and then Western blotted both immunoprecipitates for associated tapasin. Per the same amount of TAP, less tapasin K408A was coprecipitated, relative to wild-type tapasin (Fig. 2). A separate human tapasin mutant with a different amino acid substituted for lysine, K408W, also poorly stabilized TAP and interacted with TAP inefficiently (Fig. 2). Thus, human tapasin K408 mutants have poor steady-state association with TAP.
To determine whether tapasin position 408 affects TAP interaction in mouse cells, we compared the ability of murine wild-type and K408A tapasin, transduced into murine tapasin-knockout MF-Ld cells, to associate with TAP and stabilize it. Murine tapasin K408A, compared with wild-type tapasin, was relatively inefficient at binding and stabilizing TAP in mouse cells (Fig. 2). Thus, K408 is influential in TAP association in mouse as well as human cells.
Effect of substitutions at position 408 on the surface expression of MHC class I molecules
The levels of B8 and Ld at the cell surface were assessed by flow cytometry with Ab specific for folded, peptide-occupied, and open, peptide-free conformations. Human tapasin mutants K408A and K408W substantially increased the levels of folded (W6/32+) HLA-B8 molecules at the surface, although somewhat less than did wild-type tapasin (Table II). Cells expressing tapasin mutant K408W were found to have a relatively low level of cell surface HC10+ molecules, and therefore the calculated ratio of cell surface W6/32+ forms to HC10+ forms was higher for 721.220-B8+tpsn K408W (i.e., 2.6) than for 721.220-B8+tpsn (i.e., 1.3) (Table II). This finding suggests that tapasin K408W expression may lead to the expression of more stable surface HLA-B8 molecules. In contrast, transfection of either wild-type human tapasin or human tapasin K408A into 721.220-B8 cells resulted in similar cell surface W6/32+/HC10+ ratios (Table II). Murine tapasin K408A was capable of increasing the amount of folded Ld at the cell surface; however, its expression caused a disproportionate increase in the quantity of open, 64-3-7+ Ld molecules (Table II). Thus, the alanine substitution at position 408 in tapasin resulted in a relatively high prevalence of open Ld molecules at the surface of murine cells.
|Cell Types .||W6/32 .||HC10 .||W6/32/HC10 .||30-5-7 .||64-3-7 .||30-5-7/64-3-7 .|
|MF-Ld + tpsn||274.1||39.4||7.0|
|MF-Ld + tpsn K408A||232.0||65.7||3.5|
|Cell Types .||W6/32 .||HC10 .||W6/32/HC10 .||30-5-7 .||64-3-7 .||30-5-7/64-3-7 .|
|MF-Ld + tpsn||274.1||39.4||7.0|
|MF-Ld + tpsn K408A||232.0||65.7||3.5|
Values are relative log mean fluorescence units. Background values (with secondary Ab but no primary Ab) were assessed for each cell line, and were all ≤10.
To further analyze whether the K408A mutation in tapasin induced significant changes in the MHC class I molecules that reached the cell surface, we examined the cell surface stability of MHC class I molecules assembled in the presence of tapasin K408A vs wild-type tapasin. Cells (721.220-B8+tpsn, 721.220-B8+tpsn K408A, and control 721.220-B8 cells) were treated with brefeldin A for 0, 10, or 14 h, harvested, and analyzed with flow cytometry for surface expression of W6/32+ HLA-B8 molecules. This analysis was repeated four times, and little or no difference was evident in the stability of B8 on 721.220-B8+tpsn vs 721.220-B8+tpsn K408A (data not shown). This observation is consistent with the findings of Zarling et al. (32), who assessed the stability of HLA-B8 assembled in the presence of wild-type tapasin, soluble tapasin, or no tapasin. Soluble tapasin does not bind to TAP (12). Zarling et al. (32) reported that, in five separate analyses, the stability of B8 expressed in the presence of soluble tapasin was either comparable to or slightly lower than that of B8 expressed in the presence of wild-type tapasin. Also, as did Zarling et al. (32), we found that about half of the B8 molecules able to reach the surface in the absence of tapasin (i.e., in 721.220-B8 cells) were lost after a long time period (8 h for Zarling et al. (32) and 10–14 h in our assay (data not shown)). Our results and those of Zarling et al. (32) together suggest that the ability of tapasin to interact with TAP and stabilize it is not a crucial determinant of B8 surface stability.
Tapasin association with open HLA-B8 was increased by the human tapasin K408A mutation
By examining the coprecipitation of the human tapasin K408A and K408W mutants with the open HLA-B8 H chain, we found that interaction between the open B8 H chain and tapasin K408A, but not K408W, was up-regulated in comparison to wild-type tapasin (Fig. 3). Presumably due to the poor binding of tapasin K408A to TAP (Fig. 2), the increased association of tapasin K408A with the open B8 H chain was insufficient to restore B8/TAP association to normal (Figs. 2, bottom panel, and 3). Concurrently with the increased B8 association with tapasin K408A, association with calreticulin was also increased (Fig. 3). Confirming the data shown in Fig. 3, immunoprecipitating calreticulin and probing for associated HLA-B8 also demonstrated that tapasin K408A increased calreticulin association with the H chain (data not shown). Previous studies of MHC class I mutants have indicated that the degree of association of calreticulin with the MHC class I H chain often parallels the association between tapasin and the open H chain (33, 34, 35). Notably, tapasin K408W, which, like K408A, has poor TAP association (Fig. 2), does not associate more strongly with HLA-B8 than does wild-type tapasin (Fig. 3), and thus different amino acids at this position have distinct phenotypic effects.
Tapasin K408A exhibited increased interaction with ERp57
In addition to binding to the MHC class I H chain, tapasin normally binds to ERp57, and MHC class I molecules coexpressed with tapasin mutants that have little or no ability to bind to ERp57 have reduced thermostability and rapid surface turnover (36). ERp57, like TAP, is dependent on the presence of tapasin to be part of the assembly complex (6, 11, 37). We immunoprecipitated ERp57 and probed the immunoprecipitates on a Western blot for ERp57 to confirm equivalent immunoprecipitation of ERp57 from each cell lysate, and then probed the same immunoprecipitates to detect associated tapasin (wild-type or K408A) and coprecipitated TAP. We found that tapasin K408A association with ERp57 was increased relative to wild-type tapasin interaction with ERp57 (Fig. 4). In contrast, and consistent with the inefficient binding of tapasin K408A to TAP (Fig. 2), TAP/ERp57 association was significantly disrupted (Fig. 4).
Tapasin association with open Ld was not increased by the murine tapasin K408A mutation
In contrast to the elevated level of human tapasin K408A bound to open HLA-B8, the amount of mouse tapasin K408A associated with open Ld in murine cells was very similar to the amount of wild-type mouse tapasin that coprecipitated with Ld (Fig. 5). To examine whether the lack of increased interaction between tapasin K408A and Ld was related to Ld itself, rather than to species-specific differences in tapasin or cell-specific differences, we included transfected human 721.220 cells as a control. The association of human tapasin K408A with open Ld in transfected human 721.220-Ld+tpsn K408A cells was also not elevated, suggesting that Ld itself was likely the determining factor (Fig. 5). Ld binds peptides very poorly (38, 39), and therefore such a strong steady-state association may exist between Ld and tapasin that it cannot be up-regulated further by the tapasin K408A mutation. Although the association of tapasin K408A and Ld was found to be normal, the amount of TAP coprecipitated with the open Ld H chain was reduced, consistent with our observation that the mouse tapasin K408A mutation down-regulated tapasin binding and stabilization of TAP (Figs. 2 and 5).
A large pool of folded MHC class I molecules was associated with tapasin K408A
To test whether tapasin K408A exhibited an unusual degree of interaction with folded HLA-B8 molecules, we performed immunoprecipitations with the mAb W6/32 (which recognizes folded, β2m-associated HLA molecules). In the absence of tapasin, there was very little B8 detected in the folded form, but the presence of wild-type tapasin or tapasin K408A allowed folding of HLA-B8 (Fig. 6, left panel). Some wild-type tapasin was detectable in association with W6/32+ HLA class I H chains, but a much larger amount of tapasin K408A coprecipitated with W6/32+ molecules (Fig. 6, left panel). These data suggest that the tapasin K408A mutation may lead to the accumulation of tapasin K408A with folded (W6/32+) MHC class I molecules that are perhaps binding suboptimal peptides and are not competent to dissociate from tapasin.
Substitution of the charged lysine at position 408 might potentially cause release of tapasin from the ER membrane. If released from the ER, tapasin K408A might continue to interact with folded MHC class I molecules during its egress to the cell surface. To determine whether the tapasin K408A mutation allowed the escape of tapasin from the ER, we used Endo H digestion. A Western blot of Endo H-treated supernatant from a 721.220-B8+tpsn K408A lysate, probed with tapasin-specific antiserum, revealed that none of the tapasin K408A mutant molecules were Endo H resistant (Fig. 6, right panel). This finding indicates that the K408A substitution did not result in tapasin release and migration with folded MHC class I molecules.
Tapasin K408A slowed the rate of egress of folded MHC class I molecules
Considering the relatively weak interaction of tapasin K408A with TAP and the large pool of tapasin K408A associated with the MHC class I H chain, we postulated that the rate of migration of HLA-B8 from the ER might be altered when tapasin K408A is expressed. To monitor the egress of folded B8 molecules, we analyzed W6/32+ immunoprecipitates from 721.220-B8+tpsn and 721.220-B8+tpsn K408A cell lysates in a pulse-chase experiment. The immunoprecipitates were divided into two aliquots, one of which was treated with Endo H, which cleaves high mannose, N-linked oligosaccharides. Transition from a high mannose (i.e., Endo H-sensitive) oligosaccharide form to a complex (Endo H-resistant) oligosaccharide form indicates that a protein has passed through the medial Golgi (40). As shown in Fig. 7, relative to the presence of wild-type tapasin, the presence of tapasin K408A resulted in delayed egress of W6/32+ HLA-B8 molecules from the ER.
According to transmembrane region predictions, the highly conserved lysine residue at position 408 could lie either within the transmembrane domain or at its border (Table I). Due to its location, position 408 could conceivably affect interactions with another transmembrane protein, a theory that has been previously proposed (3). As a precedent, charged residues have been shown to interact in the transmembrane domains of the TCR α- and β-chains (41). In our study, we have examined the influence of tapasin position 408 on assembly complex interactions and on the surface expression of the MHC class I molecules HLA-B8 and H-2Ld in human and mouse cells.
Relative to wild-type human tapasin, the tapasin K408A mutant had reduced steady-state association with TAP and stabilized TAP poorly (Fig. 2), but was associated in unusually large quantities with both open and folded forms of HLA-B8 (Figs. 3 and 6). Human tapasin K408W and murine tapasin K408A both interacted poorly with TAP, but were not detectable at higher than normal levels with MHC class I H chains (B8 and Ld, respectively). Considering their similar phenotypes in regard to interactions in the assembly complex, it is perhaps surprising that human tapasin K408W and murine tapasin K408A had different effects on the level of open MHC class I molecules at the cell surface. Specifically, human tapasin K408W decreased the level of cell surface open (HC10+) HLA-B8 molecules, whereas mouse tapasin K408A increased the quantity of surface open (64-3-7+) Ld molecules (Table II). Although the reason for this difference is unknown, it may be related, in part, to the finding that mouse MHC class I molecules are more dependent on the TAP stabilization function of tapasin than are human MHC class I molecules (42). Another consideration is that, due to the structure of its peptide-binding groove, Ld is known to bind peptide ligands weakly relative to certain other mouse MHC class I H chains (38, 39), which may contribute to the impact of the tapasin K408A mutation on the phenotype of the surface Ld molecules.
A tapasin mutant with increased MHC class I association has not been previously reported, although tapasin mutants that do not bind detectably to the MHC class I H chain have been described (13, 43). The increased steady-state association of tapasin K408A with HLA-B8 likely resulted from suboptimal peptide loading that allowed folding and acquisition of the W6/32 epitope, but that was insufficient to permit release of HLA-B8 from tapasin. Such suboptimal loading could be due to effects specific to tapasin K408A (but not tapasin K408W) on human TAP/peptide binding (10), or to altered ability of tapasin K408A to regulate peptide loading of MHC molecules (32, 44). Alternatively, this finding may indicate that the K408A substitution influenced MHC/tapasin binding, perhaps indirectly via conformational changes in the tapasin transmembrane/cytoplasmic region. At the border between their transmembrane and cytoplasmic regions, MHC class I molecules have a conserved, highly basic sequence [RR(K/R)XXXXKGG] (45, 46, 47). Wild-type tapasin may have interacted more weakly than tapasin K408A with MHC class I molecules because of intermolecular repulsion of neighboring basic residues, and the alanine residue in K408A might have reduced repulsions more than the bulky tryptophan in K408W.
As is the case with our tapasin K408 mutants, a previously reported human tapasin mutant with a substitution at position 410 (L410F) in the transmembrane/cytoplasmic domain has weak interaction with human TAP (16). No increase in HLA class I H chain association was reported for the tapasin L410F mutant (16); in contrast, we found human tapasin mutant K408A, although not tapasin K408W, in high amounts in association with HLA-B8 (Fig. 3). In some previous studies that used MHC class I H chain mutants, the association of calreticulin with the H chain was shown to parallel the association of tapasin with the H chain (33, 34, 35). Consistent with the effect of the tapasin K408A mutation on tapasin/MHC class I interaction, association between open HLA-B8 and calreticulin was increased in the presence of tapasin K408A (Fig. 3). However, immunoprecipitation of calreticulin from 721.220-B*4402+tpsn L410F cells coprecipitated similar amounts of B44 and tapasin as immunoprecipitation of calreticulin from wild-type tapasin transfectants (16). Thus, the tapasin K408A mutant, but not the L410F mutant, had elevated steady-state association with calreticulin, as well as with the MHC H chain.
The difference in the degree of MHC class I association between tapasin K408A and L410F might be related to the fact that tapasin K408A does not leave the ER, as shown by its inability to gain Endo H resistance (Fig. 6, right panel). In contrast, a significant proportion of L410F molecules leave the ER and acquire Endo H resistance, which presumably reduces the number of L410F molecules available in the ER for binding to the MHC class I H chain (16). Evidently, the L410F mutation, but not the K408A mutation, affects a signal within the transmembrane domain that contributes to the ER retention of tapasin.
In summary, our goal in this study was to determine how MHC class I molecule assembly complex interactions are regulated by a specific structural feature of tapasin. We found that the substitution of an alanine or a tryptophan for a lysine in the tapasin transmembrane/cytoplasmic region down-regulated association of tapasin with TAP. Two other single amino acid substitutions in the tapasin transmembrane/cytoplasmic region caused tapasin to electrophorese as multiple bands, suggesting that the nature of the amino acid residue at this position may influence tapasin conformation or modification. In addition to its effect on TAP interaction, replacement of the tapasin 408 lysine with an alanine up-regulated the quantity of tapasin and calreticulin associated with HLA-B8, slowed the egress of folded HLA-B8 molecules from the assembly complex, and increased the amount of tapasin interacting with ERp57. Overall, these findings characterize the tapasin transmembrane/cytoplasmic domain as important to the functioning of tapasin in the peptide-loading complex. Studies such as these that extend our understanding of tapasin contribute to our ability to comprehend the full complexity and intricacy of the assembly of MHC class I molecules with antigenic peptides.
We thank Dr. Ted Hansen for gifts of Ab and cell lines, Dr. Ping Wang for gifts of cDNAs, and Drs. Peter Cresswell, Thomas Spies, Andres G. Grandea III, and Luc Van Kaer for cell lines. We also gratefully acknowledge the assistance of the University of Nebraska Medical Center Cell Analysis Facility, Monoclonal Antibody Facility, and Molecular Biology Facility, and the technical support of Adrienne Schwartz, Jennifer Jerrells, and Toni Luke.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health Grant GM57428 (to J.C.S.) and National Institutes of Health Training Grant T32 CA09476 (to J.L.P.).
Abbreviations used in this paper: β2m, β2-microglobulin; ER, endoplasmic reticulum; Endo H, endoglycosidase H; tpsn, tapasin.