Direct evidence that the N-terminal extensions of the TAP complex act as autonomous interaction scaffolds for the assembly of the MHC I peptide-loading complex
The loading of antigenic peptides onto major histocompatibility complex class I (MHC I) molecules is an essential step in the adaptive immune response against virally or malignantly transformed cells. The ER-resident peptide-loading complex (PLC) consists of the transporter associated with antigen processing (TAP1 and TAP2), assembled with the auxiliary factors tapasin and MHC I. Here, we demonstrated that the N-terminal extension of each TAP subunit represents an autonomous domain, named TMD0, which is correctly targeted to and inserted into the ER membrane. In the absence of coreTAP, each TMD0 recruits tapasin in a 1:1 stoichiometry. Although the TMD0s lack known ER retention/retrieval signals, they are localized to the ER membrane even in tapasin-deficient cells. We conclude that the TMD0s of TAP form autonomous interaction hubs linking antigen translocation into the ER with peptide loading onto MHC I, hence ensuring a major function in the integrity of the antigen-processing machinery.
KeywordsABC transporter Antigen processing Membrane protein interaction Macromolecular membrane complex Tapasin
ER-Golgi intermediate compartment
- MHC I
Major histocompatibility complex class I
Transporter associated with antigen processing
The adaptive immune system of jawed vertebrates is responsible for detection and elimination of virus-infected or malignantly transformed cells, thus playing an essential role in survival. Information about the cellular proteome is presented to cytotoxic T cells on the cell surface in the form of complexes of MHC I molecules with antigenic peptides derived from intracellular proteins [1, 2, 3]. A large portion of the cellular proteome is degraded by the proteasome. A fraction of these antigenic peptides is transported into the lumen of the endoplasmic reticulum (ER) by the transporter associated with antigen processing (TAP), a heterodimeric ABC complex composed of TAP1 (ABCB2) and TAP2 (ABCB3). The loading of peptides onto MHC I takes place within the peptide-loading complex (PLC), a multisubunit machinery consisting of TAP1/2, MHC I heavy chain/β2-microglobulin, the chaperone calreticulin, the oxidoreductase ERp57, and tapasin (Tsn). The latter is an essential adapter molecule within the PLC, as it bridges the peptide donor (TAP) with the peptide acceptor (MHC I) [4, 5]. For a large number of MHC I alleles, tapasin is required for loading of high-affinity peptides onto MHC I, a process known as peptide editing [6, 7]. After loading their peptide cargo, MHC I complexes travel via the secretory pathway to the cell surface. Large efforts have been made to understand the assembly, organization, and function of the PLC. For most ER-lumenal parts, structural information is available [8, 9, 10, 11, 12]. However, TAP cannot yet be crystallized, and only a homology model of the coreTAP complex is available based on the X-ray structures of bacterial ABC transporter Sav1866 and the NBD1 of TAP1 [13, 14, 15].
Each coreTAP subunit consists of six transmembrane helices plus the nucleotide-binding domain. CoreTAP shares significant homology with other ABC transporters and is necessary and sufficient for peptide binding and transport . In contrast, the unique N-terminal domain with four putative transmembrane-spanning segments, called TMD0, shares no homology to any other known protein. Deletion of the first transmembrane-spanning segment of TAP destroys its ability to interact with tapasin . However, there is no structural information available for the TMD0 of TAP1 and TAP2. Despite their critical role in PLC assembly, the TMD0s constitute the least understood part of this macromolecular machinery. TMD0 of TAP2 (hereafter named TMD0TAP2) is about 30 amino acids shorter than that of TAP1 (TMD0TAP1). Although their sequences markedly differ, both TMD0s seem to fulfill similar functions in tapasin binding , in spite of data indicating a functional asymmetry of greater significance of the rat TMD0TAP2 in PLC function . In this study, the functionality and structural integrity of the TMD0s of both TAP subunits were addressed with regard to their subcellular localization, binding, and stoichiometry of tapasin.
Materials and methods
Cloning and constructs
TMD0TAP1 and TMD0TAP2 were cloned into pcDNA3.1(+) (Invitrogen, Darmstadt, Germany) via XhoI/EcoRI and KpnI/NotI, respectively. To generate TMD0TAP1 (aa 1–164 of TAP1) with a C-terminal myc-tag and TMD0TAP2 (aa 1–127 of TAP2) with a C-terminal HA-tag, the following primer pairs were used: 5′-GTCGACGAATTCATGGCTAGCTCTAGGTG-3′ and 5′-GTCGACCTCGAGTCACAGATCCTCTTCTGAGATGAGTTTTTGTTCGGATCCGCCGGGCACCCAG-3′ for TMD0TAP1, 5′-GTCGACGGTACCAGATCTACCATGCGGCTCCCTGACCTG-3′ and 5′-GTCGACGCGGCCGCTCAAGCGTAGTCTGGGACGTCGTATGGGTAGGATCCCTTCTCCTGGGCTCC-3′ for TMD0TAP2. CoreTAP1 (amino acids 165–748, containing an N-terminal methionine) was amplified with primer pairs 5′-CGATTACTCGAGATGGGTCAGGGCGGCTC-3′and 5′-CGATTAGAATTCCCTTCTGGAGCATCTGC-3′ and cloned into pEGFP-N3 (BD Biosciences, Franklin Lakes, NJ, USA) via XhoI and EcoRI sites. CoreTAP2-mCerulean (amino acids 125-716 plus N-terminal methionine) containing a C-terminal StrepII-tag was amplified with the primer pair 5′-CATGCTTAAGATGGCCCAGGAGAAGGAGCAGGACC-3′and 5′-CCGCTCGAGTCACTTCTCGAATTGTGGGTGAGACCAAGC-3′, then cloned into pcDNA3.1(+) via AflII and XhoI. Tsn-TMD0 constructs were amplified via PCR with the following primers: 5′-AGATCTATGAAGTCCCTGTCTCTGCTCCTCG -3′ as forward primer for both constructs, 5′-ATCGCGGCCGCTCACAGATCCTCTTCTGAGATGAGTTTTTGTTCACCTCCAGGCACCCAAAGACTACC-3′ for Tsn-TMD0TAP1 containing a C-terminal myc-tag and 5′-GTCGACGCGGCCGCTCAAGCGTAGTCTGGGACGTCGTATGGGTAGGATCCTTTTTCTTGGGCACCTGGTGGAC-3′ for Tsn-TMD0TAP2 with a C-terminal HA-tag. A 34 amino acid long flexible glycine-serine linker was inserted between the C-terminus of tapasin and the N-terminus of the TMD0. Both constructs were cloned into pcDNA3.1(+) via BamHI/NotI. For the TAP1/TAP2 coexpression plasmid, TAP1 was cloned into the MCS2 of pVitro2-neo-mcs (Invivogen, San Diego, CA, USA) via BglII and NheI, TAP2 was cloned into MCS1 via BamHI and SalI.
Cell lines and transfection
HeLa cells were cultured in DMEM (PAA Laboratories, Pasching, Austria) supplemented with 10 % fetal calf serum (FCS; Biochrom, Berlin, Germany). M553, a human tapasin-deficient melanoma cell line [19, 20], was maintained in RPMI 1640 with 10 % FCS. Transfection of M553 was performed using XtremeGene HP (Roche, Grenzach-Wyhlen, Germany) according to the manufacturer’s instructions. HeLa cells were transfected with 18 mM branched polyethyleneimine (PEI) with a DNA-to-PEI ratio of one to three. After 24–48 h, cells were harvested and used for the indicated experiments.
For immunofluorescence experiments, the following antibodies were used: mouse anti-myc 4A6 (Millipore, Billerica, MA, USA), mouse anti-HA HA-7 (Abcam, Cambridge, UK), mouse anti-TAP1 148.3 , and mouse anti-StrepII (IBA BioTAGnology, Göttingen, Germany). As organelle makers, rabbit anti-Calreticulin (polyclonal IgG fraction; Sigma-Aldrich, Steinheim, Germany), anti-ERGIC-53 (Sigma-Aldrich), and anti-GM130 (Golgi; EP892Y; Abcam) were used. Detection of the primary antibodies was done with donkey anti-mouse-Alexa488 (Invitrogen), goat anti-rabbit-Cy3 (Dianova, Hamburg, Germany), and, for simultaneous detection of GFP and StrepII together with organelle marker, goat anti-mouse-Alexa633 (Invitrogen). For immunoblotting, anti-myc 4A6, anti-HA HA-7, anti-SRP54 (BD Bioscience), anti-MHC I hc HC10  and mAb 7F6, raised against amino acids 21GPAVIECWFVEDASGKG35 of human tapasin, were used.
Membrane preparation and carbonate extraction
Membranes were prepared from 5 × 106 transiently transfected HeLa cells. The cell pellet was mixed with 50 mM ice-cold Tris buffer (pH 7.3) containing 250 mM sucrose and protease inhibitor mix (Serva, Heidelberg, Germany). Cells were pulped with a glass homogenizer. Post-nuclear supernatants were collected by centrifugation for 10 min at 700g and membranes were sedimented by centrifugation at 100,000g for 30 min at 4 °C. Membranes were resuspended in 0.1 M Na2CO3 buffer pH 11.5 and incubated on ice for 15 min, followed by centrifugation at 100,000g for 20 min at 4 °C. The supernatant was collected. The remaining pellet was resuspended in carbonate buffer pH 11.5, adjusted to 1.6 M sucrose and overlaid with 1.25 M and 0.25 M sucrose in the same buffer. Centrifugation was performed for 90 min at 100,000g and the floating membranes were collected. Proteins were precipitated by chloroform–methanol and subsequently analyzed by SDS-PAGE (15 %).
Immunofluorescence and image processing
Transiently transfected HeLa cells grown on cover slips were fixed 30 min with 4 % formaldehyde in PBS at room temperature, quenched with 50 mM glycine for 10 min, and permeabilized with 0.1 % Triton X-100 for 20 min. After blocking with 5 % bovine serum albumin for 30 min, cells were stained with the primary antibodies followed by secondary antibodies for 1 h at room temperature. Nuclei were visualized by DAPI staining (Dianova). Preparations were mounted in 10 % (w/v) Mowiol (Calbiochem/Merck, Nottingham, UK). Samples were analyzed with a confocal laser-scanning microscope (LSM 510; Zeiss, Germany) equipped with a Plan-Apochromat 63x/1.4 Oil DIC objective. Deconvolution of the images was performed using non-blind 2D deconvolution of AutoQuantX2 (MediaCybernetics, Bethesda, MD, USA). Non-transfected cells were removed from the picture prior to determination of Pearson’s coefficients, and a threshold was manually set to remove background. For colocalization analysis, the JACoP plug-in  for ImageJ was used to calculate Pearson’s coefficients out of 8–10 individual cells per construct and organelle marker. For coreTAP1 and coreTAP2 co-expression, Pearson’s coefficients were calculated for coreTAP1-GFP and the organelle marker.
For coimmunoprecipitation, sheep anti-mouse Dynabeads (Invitrogen) were loaded with mouse anti-myc 4A6 (TMD0TAP1 and Tsn-TMD0TAP1) or mouse anti-HA HA-7 (TMD0TAP2 and Tsn-TMD0TAP2) for 2 h at 4 °C. Control precipitations were performed using a non-specific isotype-matched mouse antibody. After coating, the beads were washed three times with 1 ml IP buffer containing 20 mM Tris/HCl, pH 7.4, 0.1 % BSA, 150 mM NaCl, and 5 mM MgCl2. For each sample, 2.5 × 106 HeLa cells were plated in a 14.5-cm dish the day before the transfection. Then, 24 h (Tsn-TMD0s) or 48 h (TMD0s) after transfection, the cells were harvested and solubilized in 1 ml buffer containing 20 mM Tris/HCl, pH 7.4, 150 mM NaCl, 5 mM MgCl2, 1 % digitonin, and 1 % protease inhibitor mix (Serva) for 1 h at 4 °C. Solubilized cells were centrifuged for 30 min at 100,000g at 4 °C. After pre-clearing, the supernatant was incubated with pre-coated magnetic beads for 1 h at 4 °C. After washing, beads were isolated by a magnet and bound protein was eluted with 30 μl SDS sample buffer for 10 min at 65 °C. Eluates were analyzed by SDS-PAGE and subsequent immunoblotting.
Protease K digestion
Membranes were prepared as described above. After centrifugation at 100,000g, membranes were resuspended in ice-cold PBS containing 5 mM CaCl2. Proteinase K (0.6 units, Sigma Aldrich) were added in the absence or presence of 0.1 % Triton-X 100, and incubated on ice for 30 min. Digestion was stopped by the addition of PMSF (5 mM final), followed by incubation at 95 °C for 10 min in SDS sample buffer. Samples were separated by SDS-PAGE (10 %) and subsequent immunoblotting with the indicated antibodies.
Expression and membrane insertion of the isolated TMD0s
Subcellular localization of the TAP domains
TMD0s of TAP1 and TAP2 form autonomous platforms for tapasin recruitment
TMD0s of TAP1 and TAP2 localize to the ER independently of tapasin
Each TMD0 provides one single binding site for tapasin
In this study, we demonstrated that the isolated TMD0s of TAP1 and TAP2 are independently targeted to the ER membrane and form autonomous interaction hubs for tapasin recruitment. They are not required for the ER localization and transport function of the coreTAP complex . A similar functional partition into a core ABC transport complex and extra N-terminal domains was shown for other ABC transporters. Some of these N-terminal extensions have an impact on the subcellular trafficking or function of the ABC transporter, while others seem to serve different purposes, extending the fascinating properties of ABC transporters towards receptor and channel functions. The lysosomal polypeptide transporter ABCB9 (TAP-like), which forms a homodimeric complex, possesses a TMD0 that can be expressed separately. Although this TMD0 is dispensable for mere transport activity, it is required for the lysosomal localization of the transporter. The core transport complex lacking the TMD0s is targeted to the plasma membrane but redirected into lysosomes upon separate expression of TMD0 . However, an intrinsic affinity of TMD0 for the core translocation complex, as seen for ABCB9, was not observed for TAP (data not shown).
Mitochondrial ABC transporters, also members of the ABCB family, have N-terminal extensions containing very long mitochondrial pre-sequences [30, 31, 32]. In contrast to non-mitochondrial ABC transporters, these N-terminal extensions are not membrane spanning per se. Nevertheless, they are essential for correct targeting, since deletion of the leader sequence in ABCB10 and its yeast homologue MDL1 resulted in mistargeting of the transporter to the ER membrane [31, 32]. Vice versa, fusion of the N-terminal extension of ABCB7 (residue 1–135) to the dihydrofolate reductase leads to mitochondrial targeting of the fusion protein .
Most full-length ABC transporters of the subfamily C also display an extra N-terminal domain (named MSD or TMD0), which typically comprises five transmembrane-spanning segments (reviewed in ). MRP1 and MRP2, for example, harbor TMD0s with strategic functions in targeting of the transporters to the plasma membrane [34, 35]. As for TAP, the TMD0 of MRP1 is dispensable for transport activity per se . The sulfonylurea receptor SUR1 (ABCC8) is the regulatory subunit of the KATP channel KIR6.2. Mutations in KIR6.2 are associated with congenital hyperinsulinism [37, 38] and neonatal diabetes [39, 40]. TMD0SUR1 is the interaction and regulation domain to KIR6.2 [41, 42]. The TMD0SUR1, expressed in the absence of the remaining core ABC transporter, retains its ability to associate with and gate KIR6.2 . In conclusion, some ABC transporters present an interesting division of work using several independent membrane-spanning domains: one forming the gated translocation pathway and one with functions in regulation, trafficking, or binding of interaction partners. However, the TMD0s of TAP1 and TAP2 constitute prime examples of ER-resident ABC proteins.
A number of mechanisms are known that retain or retrieve proteins in or to the ER. Soluble proteins are retrieved from post-ER compartments by the KDEL motif that binds to the KDEL receptor. For type I membrane proteins, C-terminal di-lysine motifs (KKxx or KxKxx) interact with COPI vesicles for retrieval to the ER . Type II membrane proteins are retained by a double-arginine motif at the N-terminus . However, neither coreTAP nor any of the TMD0s contain these characteristic signals. Nevertheless, direct evidence for ER localization of both TMD0s is delivered in this work by immunofluorescence analysis. TMD0TAP1 and TMD0TAP2 were predominantly found in the ER, with a minor fraction entering the post-ER compartment, ERGIC. This distribution resembles that of wt TAP and is in agreement with a previous study that found a minor fraction of TAP active in the ERGIC . Of note, the escape of TMD0TAP2 into the ERGIC is enhanced compared to TMD0TAP1. These data imply that the TMD0s of both TAP subunits are retained in the ER, but the unidentified retention or retrieval factor acts more strongly in TMD0TAP1 than in TMD0TAP2. This retention/retrieval signal must be directly located within each TMD0, because ER localization is not mediated by the interaction with the KKxx motif-containing partner tapasin. The localization of wt TAP was formerly suggested to be independent of tapasin . Here, we demonstrate that this is also the case for the isolated membrane interaction hubs TMD0 of TAP1 and TAP2. In addition, we provide direct evidence that coreTAP1 and coreTAP2 localize to the ER membrane. Escape to post-ER compartments was weak and comparable to that of full-length TAP, demonstrating that the mechanism for TAP ER retention is preserved in coreTAP.
The observation that the independently expressed TMD0s of both TAP subunits are able to recruit tapasin gives direct evidence that the TMD0s are folded and functional independent of coreTAP. Furthermore, coimmunoprecipitation experiments with Tsn-TMD0 fusions demonstrate that each TMD0 only bears one single binding site for tapasin. This suggests a tapasin-to-TAP subunit ratio of 1:1, which is in perfect agreement with the latest study on this subject . However, we cannot formally rule out the possibility that a second tapasin-binding site is formed by combination of the very C-terminal parts of TMD0TAP1 and TMD0TAP2 as well as the N-terminus of coreTAP1 and coreTAP2. Tsn-TMD0TAP1 was reproducibly found to recruit less MHC I than Tsn-TMD0TAP2 (Fig. 6). This might reflect a greater importance for TMD0TAP2 in MHC I loading. This is a reasonable speculation, since for chicken and all other avian MHC I loci sequenced so far, only TAP2 harbors a TMD0, while TAP1 has no TMD0 and is therefore equivalent to coreTAP1 [46, 47]. The exact interaction sites of TMD0TAP1 and TMD0TAP2 for tapasin remain unknown. For tapasin, a number of candidate residues have been proposed, including residues F397, F401, G405, K408, and W412 of mouse tapasin , and K408 of human tapasin . For TAP, it was previously shown that removal of the first transmembrane helix destroys the tapasin-TAP interaction . Whether this is due to an altered overall folding of the truncated N-terminal domain or missing essential residues will require future approaches. The structural analysis of the TMD is one approach to resolve this important interaction hub.
In conclusion, we show for the first time that the TMD0s of both TAP subunits form autonomous interaction scaffolds for the assembly of the MHC I peptide-loading complex in the ER membrane. Each TMD0 connects ERp57, calreticulin, and peptide-receptive MHC I via a single tapasin molecule to the peptide supplier TAP. According to this, the PLC can be subdivided into three functional modules: (1) peptide binding and transport by the coreTAP complex, (2) peptide loading and editing by the Tsn-ERp57/MHC I subcomplex, and (3) the conjunction of these functions, accomplished by the interaction hubs TMD0TAP1 and TMD0TAP2.
We thank Drs. David Parcej and Andreas Hinz for providing the coreTAP2 constructs. The German Research Foundation (SFB 807 Transport and Communication across Biological Membranes and TA157/7 to R.T.), the European Drug Initiative on Channels and Transporters (EDICT to R.T.) funded by the European Commission Seventh Framework Program, and the Japan Society for the Promotion of Science (JSPS grant 20228001 to K.U.) supported this work.
This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.
- 19.Cormier JN, Panelli MC, Hackett JA, Bettinotti MP, Mixon A, Wunderlich J, Parker LL, Restifo NP, Ferrone S, Marincola FM (1999) Natural variation of the expression of HLA and endogenous antigen modulates CTL recognition in an in vitro melanoma model. Int J Cancer 80:781–790PubMedCrossRefGoogle Scholar
- 22.Stam NJ, Vroom TM, Peters PJ, Pastoors EB, Ploegh HL (1990) HLA-A- and HLA-B-specific monoclonal antibodies reactive with free heavy chains in western blots, in formalin-fixed, paraffin-embedded tissue sections and in cryo-immuno-electron microscopy. Int Immunol 2:113–125PubMedCrossRefGoogle Scholar
- 42.Gloyn AL, Pearson ER, Antcliff JF, Proks P, Bruining GJ, Slingerland AS, Howard N, Srinivasan S, Silva JMCL, Molnes J, Edghill EL, Frayling TM, Temple IK, Mackay D, Shield JPH, Sumnik Z, van Rhijn A, Wales JKH, Clark P, Gorman S, Aisenberg J, Ellard S, Njølstad PR, Ashcroft FM, Hattersley AT (2004) Activating mutations in the gene encoding the ATP-sensitive potassium-channel subunit Kir6.2 and permanent neonatal diabetes. New Engl J Med 350:1838–1849PubMedCrossRefGoogle Scholar