The MHC I loading complex: a multitasking machinery in adaptive immunity

Transcription

1 Review The MHC I loading complex: a multitasking machinery in adaptive immunity Sabine Hulpke 1 and Robert Tampé 1,2 1 Institute of Biochemistry, Biocenter, Goethe-University Frankfurt, Max-von-Laue-Strasse 9, Frankfurt am Main, Germany 2 Cluster of Excellence Macromolecular Complexes, Goethe-University Frankfurt, Max-von-Laue-Strasse 9, Frankfurt am Main, Germany Recognition and elimination of virally or malignantly transformed cells are pivotal tasks of the adaptive immune system. For efficient immune detection, snapshots of the cellular proteome are presented as epitopes on major histocompatibility complex class I (MHC I) molecules for recognition by cytotoxic T cells. Knowledge about the track from the equivocal protein to the presentation of antigenic peptides has greatly expanded, leading to an astonishingly elaborate understanding of the MHC I peptide loading pathway. Here, we summarize the current view on this complex process, which involves ABC transporters, proteases, chaperones, and endoplasmic reticulum (ER) quality control. The contribution of individual proteins and subcomplexes is discussed, with a focus on the architecture and dynamics of the key player in the pathway, the peptide-loading complex (PLC). Immune detection leads to pathogen and tumor clearance Humans are threatened by numberless pathogens every second of their lives. Bacteria, fungi, and viruses all cause human diseases, and in some cases even pandemics. In addition, malignantly transformed cells are an enemy from within. To cope with these threats, evolution has built the sophisticated system of adaptive immunity. To identify and eliminate infected or transformed cells, information on the target cell must be transmitted to the effector cells of the immune system. The adaptive immune system is based on a constant turnover of cellular proteins, which are converted into peptides. These are loaded onto MHC I molecules in the context of the multisubunit PLC (Figure 1). This macromolecular machinery is centered on the transporter associated with antigen processing (TAP), an ABC transporter that translocates antigenic peptides into the ER lumen [1]. The PLC further includes, besides MHC I heavy chain and b 2 -microglobulin (b 2 m), the adapter chaperone tapasin (Tsn), the ER-resident Corresponding author: Tampé, R. (tampe@em.uni-frankfurt.de) Keywords: ABC transporters; proteasome; chaperones; ER quality control; macromolecular complexes; MHC class I molecules; tapasin; transporter associated with antigen processing /$ see front matter ß 2013 Elsevier Ltd. All rights reserved. oxidoreductase ERp57, and lectin-like chaperone calreticulin (Crt). CD8 + cytotoxic T cells scan MHC I complexes for antigenic cargo derived from virus- or tumor-associated proteins and eventually eliminate the target cell by inducing apoptosis to prevent systemic spread of the disease. In this review, we discuss the MHC I peptide-loading pathway, which spans multiple compartments and involves a set of specialized and generic receptors, chaperones, and transporters. Our main focus is on the PLC and its function, composition, and dynamics. The components and interplay of the PLC are covered, with special attention paid to the molecular architecture, stoichiometry, and dynamics of this macromolecular complex. To begin, the origin of antigenic peptides is discussed. The generation of antigenic peptides One source of peptides for MHC I presentation are defective ribosomal products (DRiPs), which are polypeptides that fail to acquire their native conformation and are rapidly delivered to the 26S/20S proteasome complex [2]. Hence, antigenic peptides can emerge very soon after protein translation and help in the very early detection of viral infection. In addition, unwanted or redundant proteins are, after ubiquitylation, subject to turnover by the proteasome, and thus provide an alternative source for MHC I ligands [3]. The relative contribution of DRiPs and proteasomal breakdown products to the pool of MHC I binding peptides, however, remains to be elucidated. Proteasomes come in different types: the standard housekeeping proteasome and the immunoproteasome, which can be induced by interferon-g and interferon-b, along with the TAP and MHC I [4]. The immunoproteasome is perfectly suited for production of peptides for the MHC I antigen presentation pathway because it preferentially cleaves after hydrophobic and basic amino acids, matching the preferences of TAP and MHC I for the C terminus of peptides. The N termini produced by the proteasome differ from those favored by TAP and MHC I. In addition, the peptides generated are approximately 3 22 residues long, so most of them are either too short or too long to be transported and loaded onto MHC I, which binds 8 11mers. After proteasomal degradation, the resulting peptides may be trimmed by cytosolic aminopeptidases such as leucine aminopeptidase, puromycin-sensitive 412 Trends in Biochemical Sciences August 2013, Vol. 38, No. 8

2 Cytotoxic T cell T cell receptor An gen presenta on via MHC I CD8 Plasma membrane Cellular proteome Cytosol Proteasome Golgi Pep des Ribosome Translocon TAP1 TAP2 BiP MHC I heavy chain β 2 m Calnexin Calre culin ERp57 Tsn ERAAP PLC ER membrane ER lumen TiBS Figure 1. Antigen processing via MHC I molecules. After translation, the newly synthesized MHC I heavy chain associates with BiP and calnexin. Subsequently, assembly with soluble b 2 m and exchange of calnexin with Crt take place. This MHC I/Crt subcomplex now associates with TAP/Tsn ERp57 to form a fully assembled PLC, through which peptide transport and loading onto MHC I are spatiotemporally coordinated. The transmembrane and nucleotide-binding domains of TAP are colored in blue and red, respectively. The ER-associated aminopeptidase ERAAP might transiently interact with the PLC. After peptide loading, the peptide MHC I complex dissociates from TAP/ Tsn ERp57 and traffics to the cell surface via the secretory pathway, where the peptide MHC I complexes are monitored by cytotoxic CD8 + T cells. BiP, binding immunoglobulin protein; MHC I, major histocompatibility complex class I; b 2 m, b 2 -microglobulin; PDI, protein disulfide isomerase; PLC, peptide-loading complex; Crt, calreticulin; TAP, transporter associated with antigen processing; Tsn, tapasin. aminopeptidase, bleomycin hydrolase, and tripeptidyl peptidase-ii (TPP-II). Besides these general trimming peptidases, nardilysin, thimetoligopeptidase, and the insulin-degrading enzyme are responsible for the generation of a few clearly defined epitopes. However, several studies have questioned the necessity of these peptidases; their real impact on the formation of antigenic peptides therefore remains unknown [5,6]. For example, a critical role in the generation of many MHC I ligands was proposed for TPP-II [7], but other studies using, for instance, TPP-II knockout mice, found no major defects in antigen presentation [8,9]. On the whole, although peptide trimming by cytosolic peptidases seems essential in some cases, the net effect might be a negative one, meaning that the peptidases destroy more MHC I ligands than they create [5]. Considering the short half-life of proteasome-generated peptides of less than 10 s [10], efficient MHC I antigen processing would benefit from a spatial link between the proteasome and TAP. However, such a direct connection has never been demonstrated convincingly. After transport into the ER lumen by TAP [1], peptides are further N- terminally processed by the ER aminopeptidase associated with antigen processing ERAAP (ERAP1/2 in humans) until they fit the length requirements for the MHC I binding pocket [11 13]. Taken together, the data suggest that antigenic peptides are derived from either DRiPs or native proteins and are shaped by the actions of the proteasome and additional peptidases. Before MHC I loading, peptides have to traverse the ER membrane, a task performed by the translocation complex TAP. The core TAP complex: the gateway for peptide delivery into the ER The heterodimeric TAP complex, composed of TAP1 (ABCB2) and TAP2 (ABCB3), is a member of the ABC transporter superfamily. The common blueprint for all ABC transporters is their organization into two transmembrane domains (TMDs) and two nucleotide-binding domains (NBDs). A cycle of engagement and disengagement of the two NBDs on ATP binding and hydrolysis is thought to be coupled to conformational changes in the TMDs to move proteasomal degradation products across the ER membrane [14]. TAP can be functionally dissected into the core TAP complex and N-terminal extensions at each subunit, called TMD 0 (Figure 2). As the central functional unit of the antigen translocation complex, core TAP comprises inner 2 6 transmembrane helices as well as the NBDs and is essential and sufficient for peptide binding and transport [15]. Peptide binding is ATP-independent [16,17], whereas peptide transport requires ATP hydrolysis [18]. The peptide-binding site was mapped to the cytosolic loop between transmembrane helices 4 and 5 and the cytosolic stretch following transmembrane helix 6 of core TAP1 and TAP2 [19]. TAP represents a prime example of how substrate specificity, selectivity, and diversity are combined. The 413

3 Author's personal copy Review Trends in Biochemical Sciences August 2013, Vol. 38, No. 8 ERp57 Calre culin Pep deloaded MHCI MHCI heavy chain Tsn β2m TMD0 ER membrane TMD0 Proteasome Pep de CoreTAP1/2 Ti BS Figure 2. Dynamic assembly and disassembly of the PLC. The model is assembled based on the X-ray structures of HLA-B*4402 (PDB 3KPL), Tsn-ERp57 (PDB 3F8U), Crt (PDB 3RG0), and the homology model of coretap1/2 [14]. The interaction sites are summarized in Table 1. The exact docking and positioning of the MHC I subcomplexes are notional. Together with the Tsn ERp57 conjugate (salmon and magenta), the TAP1/2 heterodimer forms the core of the PLC. Crt (grey) escorts the empty MHC I heavy chain b2m dimer (lime green and yellow) to Tsn, leading to assembly of the active, loading-competent PLC. A fully assembled PLC can contain up to two Tsn ERp57/MHC I/Crt subcomplexes. Peptides transported by TAP are now loaded onto MHC I and are affinity-proofread by the Tsn ERp57 conjugate. The newly formed, stable peptide MHC I complex dissociates from the PLC, making way for a new, peptide-receptive MHC I to be incorporated into the PLC. MHC I, major histocompatibility complex class I; Crt, calreticulin; TMD, transmembrane domain; Tsn, tapasin; PLC, peptide-loading complex; TAP, transporter associated with antigen processing. substrate specificity of human TAP has been most intensively studied. The three N-terminal residues and the last C-terminal residue of the peptide are critical for recognition by TAP. A clear bias towards hydrophobic and basic residues at the C terminus of the peptide was found. Proline is strongly disfavored at position 2, whereas positively charged residues at positions 1 3 have a positive effect on TAP binding [14]. Despite these anchor positions, TAP is promiscuous concerning peptide length and binds peptides from eight up to 40 residues. Binding and transport are optimal for peptides of 8 16 and 8 12 amino acids, respectively [16,20]. The discrepancy observed between binding and transport length optimum could be due to the experimental set-ups, which involved direct binding or indirect transport assays. Overall, these findings raise an important question: how does the binding site accommodate peptides of such different size and sequence? The answer comes from a recent electron paramagnetic resonance (EPR) study that suggests that both ends of the peptide are fixed in the binding pocket, whereas the central part is mobile and unlikely to interact with TAP [21]. It is therefore thought that the middle of the peptide protrudes from the binding pocket, adopting an extended kink conformation, which enables TAP to accommodate peptides of different length and sequence. On the basis of these 414 experimental data, two binding pockets with different electrostatic potentials have been proposed in a TAP homology model for binding the peptide N and C termini [14,22,23]. It was revealed that TAP is stereospecific, because introduction of D-amino acids drastically reduced peptide binding and translocation [24,25]. Peptides containing extremely large side chains bound to TAP, but translocation was inhibited, indicating steric constraints within the translocation pathway [24,26]. In summary, TAP recognizes its substrate peptides by N- and C-terminal anchor positions, thereby combining specificity with variability because the sequence in the middle of the peptide can be diverse. Once the peptide has crossed the ER membrane, the PLC components need to be assembled at the TAP transport machinery to guarantee optimal peptide loading. Membrane interaction hubs: bridging antigen transport and MHC I loading Each TAP subunit possesses an N-terminal extension consisting of a predicted bundle of four transmembrane helices [15,27]. These TMD0 domains function as autonomous interaction platforms that are targeted to the ER membrane independently of the core TAP complex (Figure 2) [28]. Although TMD0 of TAP1 and TAP2 shares

4 Table 1. Interactions within the PLC Interaction partner Interaction sites Refs 1 2 Partner 1 Partner 2 TAP1 TAP2 NBD1, intracellular loops, TM domains NBD2, intracellular loops, TM domains [14,23] TAP Tsn TMD 0 TM helix; K408, L410 (human); F397, F401, G405, K408, W412 (mouse) MHC I Tsn a2 domain (aa ) a3 domain (aa 222) MHC I Crt N-core glycosylation N86 Lectin domain ERp57 Crt b domain (ERp57) ERp57 Tsn C57 a domain TAP [15,28,37] Tsn [29 31] N- and C-terminal ER-lumenal IgG-like domain MHC I [52 54,92] Tsn [38,49 51] P domain (Crt) (aa ) C95 ER-lumenal domain [73,74] [49,63,64] NBD, nucleotide-binding domain; TM, transmembrane; TAP, transporter associated with antigen processing; Tsn, tapasin; MHC I, major histocompatibility complex class I; Crt, calreticulin. less than 17% sequence identity, each provides a single binding site for the adapter chaperone Tsn, even across different species (human and rodent) [28]. Substitution of the L410 residue in human Tsn destroys the interaction with TAP (Table 1); it was therefore speculated that the interaction is mediated by a leucine zipper-like motif [29]. Contrary to this notion, substitution of K408 in the middle of the transmembrane helix of human Tsn [30] and combined substitution of F397, F401, G405, K408, and W412 in mouse Tsn abolishes binding to TAP [31]. Notably, Tsn interaction sites in TMD 0 of TAP1 and TAP2 have not been identified so far. By recruiting Tsn, TMD 0 guarantees close proximity between TAP and MHC I, which is crucial for efficient MHC I loading [32]. By enhancing the local peptide concentration, peptide-receptive MHC I molecules could sample a greater number of different peptides and thus increase the chance of binding peptides that display the lowest dissociation rate. In short, the TMD 0 domains form interaction platforms for PLC assembly. To obtain peptide MHC complexes of high stability, not only a connection to TAP, but also a puzzling phenomenon called peptide editing are necessary. Peptide editing and affinity proofreading Tsn, a type I membrane glycoprotein of 48 kda, is a key component of the PLC [33]. By binding simultaneously to TAP and MHC I, it links the two major tasks of the PLC: peptide transport and MHC I loading. This dual liaison is important for overall PLC function, because a soluble Tsn lacking the transmembrane domain is unable to support peptide loading to its full extent [29,34 36]. Furthermore, Tsn binding stabilizes the TAP complex, because TAP levels are significantly decreased in the absence of Tsn [37 39]. Along the same lines, the thermal stability of TAP is increased in the presence of Tsn [40]. The most prevalent function of Tsn is peptide editing/optimization, which is responsible for preferential loading of high-affinity peptides onto MHC I [41 45]. This function is of great importance for some MHC I alleles (these are called Tsndependent) because they are unable to associate with optimal peptides in the absence of Tsn. MHC I complexes are meant to present the antigenic cargo for a relatively long time at the cell surface (3 7 days), so their stability is crucial for successful immune detection, making peptide editing a key task of the PLC. MHC I molecules loaded with suboptimal peptides are retained in the PLC. The current view is that after de novo synthesis, MHC I heavy chain (hc; 44 kda) first encounters the chaperones BiP (binding immunoglobulin protein) and calnexin (Cnx). On b 2 m association, Crt replaces Cnx, both of which have overlapping functions as ER lectin-like chaperones [46]. Subsequently, the remaining PLC components, including Tsn, ERp57, and TAP, form a macromolecular complex that catalyzes MHC I recognition, translocation, and loading of high-affinity peptides. Only monoglucosylated MHC I hc can be incorporated into the PLC, so the actions of glucosidases I and II are necessary to create monoglucosylated MHC I. After removal of the last glucose residue, the interaction with the PLC is weakened and MHC I may traffic via the secretory pathway to the plasma membrane. Prematurely dissociated, unstable MHC I may be rescued by re-glucosylation by UDP-glucose-glycoprotein glucosyltransferase 1 [47,48]. MHC I hc consists of three domains, a1 a3. The a3 domain harbors the binding site for the CD8 T cell co-receptor, whereas a1 and a2 form the peptide-binding cleft, which accommodates peptides of 8 11 residues (Figure 3A). The lumenal domains of MHC I hc and Tsn also provide an extended contact surface. Concerning the exact interaction site(s) of MHC I with Tsn, studies identified the involvement of a number of partially contradictory regions or residues. Multiple regions in the N-terminal and the C-terminal ER lumenal IgG-like domain of Tsn contribute to MHC I binding [49 51]. Deletion of the first 50 aa results in diminished MHC I binding [38]. The substitution T134K in MHC I leads to complete disruption of the MHC Tsn interaction, and it has been suggested that several surrounding residues positively or negatively influence Tsn-mediated MHC I/TAP association [52 54]. The dependence of peptide loading onto MHC I on Tsn varies greatly between different alleles. B*44:02 is highly Tsn-dependent, whereas B*44:05 is largely Tsn-independent [55]. This is intriguing, because the two alleles differ only in a single residue at position 116, aspartate in B*44:02 and tyrosine in B*44:05. It is thought that this residue influences the tendency of the MHC I peptidebinding site to open up in the absence of peptide [56]. This hints at a model in which a wide open, peptide-receptive binding pocket is a precondition for peptide binding, and opening of the pocket can be supported by residues in the peptide-binding cleft or by Tsn. The present view is that Tsn acts as chaperone on the MHC I peptide binding cleft 415

5 Author's personal copy Review Trends in Biochemical Sciences August 2013, Vol. 38, No. 8 (A) Lec n domain 3 ERp57-Crt 2 TMD0-Tsn Calre culin (b P-domain) (TM helices) Tsn MHC I TMD0 TMD0 CoreTAP1/2 ERp57 P-domain 1 MHC I (pep de (binding pocket) 5 nm (B) b domain a domain Calre culin 6 Tsn-ERp57 (C95 C57) ERp57 90 P-domain Lec n domain α2 b domain a domain MHC hc 5 MHC I-Tsn (α2 and α3) β2m Tsn 4 MHC I-Crt (N86 lec n d.) α3 Ti BS Figure 3. Structural organization of the PLC. (A) Top view of the PLC and (B) side views of ER-lumenal domains of the PLC. After translocation by core TAP, peptides are shuttled into the binding pocket of MHC I (1, indicated by the black arrow). The MHC I loading subcomplex assembles independently at TMD0 of TAP1 and TAP2 (2, cyan and green circles). Crt contacts the b domain of ERp57 (3) and the N-glycan chain on MHC I (4), thereby stabilizing the PLC by multivalent interactions. The MHC I heavy chain (hc) contacts Tsn via its a2 and a3 domain (5), whereas Tsn forms a mixed disulfide with ERp57 (6). PLC, peptide-loading complex; HLA, human leukocyte antigen; Tsn, tapasin; Crt, calreticulin; MHC I, major histocompatibility class I; b2m, b2-microglobulin; TAP, transporter associated with antigen processing. and induces a more open conformation, thereby accelerating the dissociation of low-affinity peptides and promoting the binding of high-affinity peptides [43 45]. MHC I complexes loaded with an optimal peptide overcome the energy barrier necessary to convert to a closed conformation, which causes MHC I to dissociate from Tsn and leave the PLC [43]. Other studies, however, did not find differences in the affinity of the bound peptide to MHC I in the 416 absence and presence of Tsn [57]. Moreover, the stability of the peptide MHC I complex is not entirely determined by the affinity of the peptide cargo [57,58]. In addition, allelespecific differences in the interaction of MHC I with Tsn/ TAP have been found [59]. Future experiments to determine the relative contribution of the different functions of Tsn to MHC I peptide loading might solve these discrepancies. Since the interaction between Tsn and MHC I has

6 not been quantified directly so far, the mechanistic basis of peptide editing is still not known. Tsn, which contains an ER retrieval signal (KKxx) at its C terminus, is involved in the retrograde transport of peptide-receptive MHC I from the Golgi [60,61]. A similar function has been attributed to Crt [62]. Thus, Crt and Tsn may have redundant functions in this context. This assumption fits well with the observation that an engineered PLC with covalently linked Tsn TAP retains its ability to load peptides onto MHC I [32]. Thus, Tsn that cannot dissociate from TAP and thus cannot leave the ER is fully active in antigen processing. In summary, Tsn is a key component of the PLC, because it ensures the loading of peptides onto MHC I that results in kinetically stable peptide MHC I complexes. However, for optimal PLC interplay and functionality, the chaperones ERp57 and Crt are needed. Accessory chaperones for MHC I loading: ERp57 and Crt In the PLC, Tsn is covalently linked to the oxidoreductase ERp57 by a mixed, intermolecular disulfide bridge between Cys95 of Tsn and Cys57 of ERp57 (Table 1) [63,64]. Evidence has emerged that Tsn and ERp57 form a functional unit [45]. The X-ray structure of the Tsn ERp57 conjugate demonstrates that the ER-lumenal part of Tsn adopts an L-like conformation and interacts with both catalytic domains, a and a 0, of ERp57 (Figure 3) [49]. ERp57 is necessary for PLC structural integrity [65], whereas its redox activity is dispensable for PLC function [66,67]. Undoubtedly, however, ERp57 is relevant for antigen processing, because its deletion leads to reduced MHC I surface expression in a mouse model [68] and suboptimally loaded MHC I in human cells [63]. As an ER lectin-like chaperone, Crt binds to the monoglucose residue of N-glycosylated polypeptides during quality control of protein folding in the secretory pathway. In addition, Crt fulfills specific roles in the MHC I peptide loading pathway by promoting MHC I surface presentation [65,69 71]. In contrast to Crt-proficient cells, Crtdeficient cells required pulsing with higher concentrations of ovalbumin for efficient T cell stimulation [72]. Within the PLC, Crt is engaged in multiple interactions (Table 1). Tsn binding to MHC I depends on N-core monoglycosylation at N86 of the heavy chain (Figure 3). In addition, Crt interacts with the b domain of ERp57 via the tip of its P domain, a b-strand hairpin domain (Figure 3) [73,74]. The glycan of Tsn is also involved in recruitment of Crt into the PLC [75]. Nevertheless, the individual contribution of each of these interactions remains controversial [48,67,71,75,76]. Although the redox activity of ERp57 is not necessary for basic PLC function, ERp57 and Crt seem to stabilize and support the function of the PLC. The list of key components is now complete and we focus on the stoichiometric assembly and disassembly of the PLC. Stoichiometry and dynamics within the PLC It is thought that the PLC assembles on a single TAP heterodimer. The number of associated factors is less clear. The X-ray structure of the Tsn ERp57 heterodimer allows us to conclude that Tsn and ERp57 exist in a well-defined 1:1 stoichiometry. Nonetheless, the stoichiometry of TAP, Tsn, and MHC I remained controversial. In two initial studies, one or four Tsn molecules were found in the PLC [33,77]. In rat human hybrid PLCs, two human Tsn molecules per rat TAP were observed; the two Tsn molecules alternated recruitment of MHC I so that a PLC contained either one or no MHC I molecules at any given moment [78]. The 2:1 ratio of Tsn to TAP is supported by the existence of Tsn-binding sites on TAP1 and TAP2 [15,79], both of which can be occupied simultaneously [28,32]. Recent findings that the TMD 0 of each TAP subunit harbors a single Tsn binding site exclude the possibility of a stoichiometry exceeding 2:1 [28], especially because Tsn does not form homo-oligomers [80]. The Tsn-to-TAP ratio of 2:1 was recently proven by single molecule analysis, demonstrating that PLCs exist as complexes with one or two associated MHC I, whose ratio depends on the peptide supply provided by TAP and the MHC I alleles involved (Figure 2) [81]. Notably, one Tsn molecule (and hence one MHC I) per PLC is sufficient to promote MHC I surface presentation to wild type levels, whereas PLCs with two bound Tsn molecules support MHC I surface presentation even exceeding wild type levels [32]. The finding that PLCs with one or two MHC I molecules exist supports the idea that the interaction TMD 0 platforms of TAP1 and TAP2 act independently of one another (Figure 3). Interestingly, the PLC including TAP undergoes changes in membrane dynamics during peptide translocation/loading. Lateral diffusion of TAP1-GFP in the ER membrane decreases when TAP translocates peptides. By contrast, TAP mobility increases under conditions of ATP or peptide depletion and after inhibition by viral inhibitors [82]. However, the nature of these structural rearrangements is largely unknown. Relatively little is known about PLC assembly and dissociation. According to the observation that Tsn covalently linked to TAP gives fully functional PLCs [32], TAP and the Tsn ERp57 conjugate presumably form the rather static PLC core (Figure 2). Crt associates with MHC I before assembly with the remaining PLC components, and therefore Crt seems to escort MHC I hc/b 2 m dimers to the core PLC (Figure 2) [48]. Tsn ERp57 mediates loading of high-affinity peptides onto MHC I and retains them in the ER until an optimal peptide is bound [45,83,84]. After loading, peptide MHC complexes leave the PLC together with Crt (Figure 2) [48], and possibly Tsn in some cases [60,61], en route to the Golgi. It was recently found that the Tsn-related protein (TAPBPR) escorts MHC I through the Golgi [85]. TAPBPR is structurally similar to Tsn, but does not bind ERp57 and Crt, and associates preferentially with human leukocyte antigen (HLA)-A alleles. In contrast to Tsn, however, it is not incorporated into the PLC and is therefore thought to fulfill post-er functions [85]. In the Golgi, suboptimally loaded MHC I can be retrieved to the ER [60 62] or otherwise proceed to the cell surface. The quality control of loaded MHC I molecules may not be restricted to the ER, and some steps may also occur in post-er compartments [86,87]. Taken together, data indicate that fully assembled PLC comprises one TAP1/2 heterodimer, two Tsn ERp57 conjugates, and up to two MHC I and Crt molecules. Notwithstanding, the exact stoichiometry is subject to change 417

7 according to the circumstances. We can now combine all our current knowledge into a functional model of PLC action. A dynamic PLC model On the basis of the data outlined here, we propose the following model for PLC function. As the central translocation complex, TAP assembles a maximum of two Tsn ERp57 conjugates via its two TMD 0 interaction hubs. Newly synthesized MHC I molecules interact first with Cnx, which is replaced by its soluble homolog Crt on b 2 m binding. This preassembled MHC I subcomplex docks to the TAP/Tsn ERp57 complex, forming the loading-competent active PLC, which consists of one TAP heterodimer, two Tsn ERp57 conjugates, and one or two MHC I complexes decorated with Crt. Within this multitasking assembly, core TAP acts as the peptide supplier, whereas the TMD 0 of each TAP subunit is essential in recruiting Tsn, and thus provides proximity between the peptide translocation pore core TAP and the peptide acceptor MHC I, which seems to be important for optimal peptide loading (Figure 3) [32]. Tsn affinity proofreads peptide MHC I complexes, and thereby catalyzes peptide exchange in favor of peptides with very slow dissociation rates (Tampé lab, unpublished data). Kinetically stable peptide MHC I complexes are released from the PLC. Finally, MHC I molecules travel via the secretory pathway to the cell surface and the PLC can re-form for the next round of peptide loading. Notably, multiple MHC I molecules that are loaded with the same epitopes start to cluster in the Golgi, resulting in patches of similar peptide MHC I complexes on the cell surface [88]. This increases the avidity of TCR MHC I interaction, leading to an enhanced T cell sensitivity. For efficient antigen processing, a sequence of key events must be fulfilled. (i) A peptide supply suitable for MHC I loading by TAP is required. (ii) There must be spatial proximity between the peptide donor TAP and the peptide acceptor MHC I, granted by the membrane interaction of TMD 0 and Tsn. (iii) Peptide-receptive MHC I must be retained in the PLC, assured by Tsn ERp57. (iv) Before dissociation of the peptide MHC complex from the PLC, MHC I must bind high-affinity peptides to ensure that a stable complex is formed; this peptide editing function is also attributed to the Tsn ERp57 conjugate. For Box 1. Outstanding questions What is the structure of TAP and how will it add to our understanding of peptide translocation and the conformational changes associated with this process? How does the TMD 0 of TAP provide a generic and essential interaction platform for PLC assembly? What is the mechanistic basis for peptide editing by Tsn ERp57? What are the affinities between different components of the PLC? Where are the exact binding sites between the different PLC components located and how redundant are they? What is the overall architecture of the PLC? How can the structures of different subcomponents be integrated? What are the major structural rearrangements during PLC assembly and disassembly? What are the role and mechanism of TAP in cross-presentation and has the PLC a physiological impact beyond classical MHC I loading? efficient loading, at least one Tsn-ERp57/MHC I complex must be present in the PLC. (v) The peptide MHC I complex must dissociate from the PLC; this exact mechanism is not understood yet. Although knowledge about the MHC I peptide loading pathway is quite advanced, culminating in the model explained above, a number of questions remains unsolved (Box 1). This opens the way for further research directions to complete our understanding of immune recognition and adaptive immunity. Concluding remarks The physiological importance of the MHC I antigen processing machinery is underlined by the manifold mechanisms that viruses and tumors have developed to prevent immune recognition. The abundant cowpox and herpes viruses are to be found among the ones that block the PLC [14]. A profound understanding of the exact mechanisms leading to immune evasion may support the development of anti-viral (and possibly anti-tumor) drugs and this topic will therefore be the subject of further research. Finally, it remains to be mentioned that antigen presentation by MHC I is not entirely dependent on the proteasome and the PLC, although the vast majority of peptides are loaded via this classical pathway. However, there are several exceptions to this [89,90]. Examples are epitopes derived from signal sequences and certain viral epitopes, both of which are not generated in the cytosol but in the ER lumen and therefore do not rely on TAP for ER entry. For latent membrane protein 2 of Epstein Barr virus, a TAP-independent but proteasome-dependent pathway has been reported [91]. Owing to recent advances, structural information about the ER-lumenal parts of the PLC is available. For TAP, the core complex can only be modeled on the structure of related ABC exporters, and therefore the eagerly awaited X-ray structure of the TAP complex constitutes the missing link. Sophisticated integrative approaches that combine information from X-ray crystallography, cryo-electron microscopy, cross-linking, mass spectrometry, magnetic resonance spectroscopy, molecular dynamics, and coarsegrain simulation will be instrumental in reconstruction of the architecture of this macromolecular complex in adaptive immunity. Acknowledgments We thank all members of our laboratory for helpful comments on the manuscript. The German Research Foundation (SFB 807, Transport and Communication across Biological Membranes) and the European Drug Initiative on Channels and Transporters (EDICT to R.T.) funded by the European Commission Seventh Framework Program supported this work. References 1 Abele, R. and Tampé, R. (2009) Peptide trafficking and translocation across membranes in cellular signaling and self-defense strategies. Curr. Opin. Cell Biol. 21, Yewdell, J.W. (2011) DRiPs solidify: progress in understanding endogenous MHC class I antigen processing. Trends Immunol. 32, Farfan-Arribas, D.J. et al. (2012) Using intein catalysis to probe the origin of major histocompatibility complex class I-presented peptides. Proc. Natl. Acad. Sci. U.S.A. 109, Huber, E.M. et al. (2012) Immuno- and constitutive proteasome crystal structures reveal differences in substrate and inhibitor specificity. Cell 148,

8 5 van Endert, P. (2011) Post-proteasomal and proteasome-independent generation of MHC class I ligands. Cell. Mol. Life Sci. 68, Rock, K.L. et al. (2010) Proteases in MHC class I presentation and cross-presentation. J. Immunol. 184, Reits, E. et al. (2004) A major role for TPPII in trimming proteasomal degradation products for MHC class I antigen presentation. Immunity 20, Firat, E. et al. (2007) Analysis of direct and cross-presentation of antigens in TPPII knockout mice. J. Immunol. 179, Kawahara, M. et al. (2009) Analysis of the role of tripeptidyl peptidase II in MHC class I antigen presentation in vivo. J. Immunol. 183, Reits, E. et al. (2003) Peptide diffusion, protection, and degradation in nuclear and cytoplasmic compartments before antigen presentation by MHC class I. Immunity 18, Saric, T. et al. (2002) An IFN-g-induced aminopeptidase in the ER, ERAP1, trims precursors to MHC class I-presented peptides. Nat. Immunol. 3, Saveanu, L. et al. (2005) Concerted peptide trimming by human ERAP1 and ERAP2 aminopeptidase complexes in the endoplasmic reticulum. Nat. Immunol. 6, Serwold, T. et al. (2002) ERAAP customizes peptides for MHC class I molecules in the endoplasmic reticulum. Nature 419, Parcej, D. and Tampé, R. (2010) ABC proteins in antigen translocation and viral inhibition. Nat. Chem. Biol. 6, Koch, J. et al. (2004) Functional dissection of the transmembrane domains of the transporter associated with antigen processing (TAP). J. Biol. Chem. 279, van Endert, P.M. et al. (1994) A sequential model for peptide binding and transport by the transporters associated with antigen processing. Immunity 1, Uebel, S. et al. (1995) Requirements for peptide binding to the human transporter associated with antigen processing revealed by peptide scans and complex peptide libraries. J. Biol. Chem. 270, Neefjes, J.J. et al. (1993) Selective and ATP-dependent translocation of peptides by the MHC-encoded transporter. Science 261, Nijenhuis, M. and Hämmerling, G.J. (1996) Multiple regions of the transporter associated with antigen processing (TAP) contribute to its peptide binding site. J. Immunol. 157, Koopmann, J.O. et al. (1996) Translocation of long peptides by transporters associated with antigen processing (TAP). Eur. J. Immunol. 26, Herget, M. et al. (2011) Conformation of peptides bound to the transporter associated with antigen processing (TAP). Proc. Natl. Acad. Sci. U.S.A. 108, Corradi, V. et al. (2012) The human transporter associated with antigen processing: molecular models to describe peptide binding competent states. J. Biol. Chem. 287, Oancea, G. et al. (2009) Structural arrangement of the transmission interface in the antigen ABC transport complex TAP. Proc. Natl. Acad. Sci. U.S.A. 106, Gromme, M. et al. (1997) The rational design of TAP inhibitors using peptide substrate modifications and peptidomimetics. Eur. J. Immunol. 27, Uebel, S. et al. (1997) Recognition principle of the TAP transporter disclosed by combinatorial peptide libraries. Proc. Natl. Acad. Sci. U.S.A. 94, Gorbulev, S. et al. (2001) Allosteric crosstalk between peptide-binding, transport, and ATP hydrolysis of the ABC transporter TAP. Proc. Natl. Acad. Sci. U.S.A. 98, Schrodt, S. et al. (2006) Membrane topology of the transporter associated with antigen processing (TAP1) within an assembled functional peptide-loading complex. J. Biol. Chem. 281, Hulpke, S. et al. (2012) 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. Cell. Mol. Life Sci. 69, Tan, P. et al. (2002) 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. 168, Petersen, J.L. et al. (2005) A charged amino acid residue in the transmembrane/cytoplasmic region of tapasin influences MHC class I assembly and maturation. J. Immunol. 174, Papadopoulos, M. and Momburg, F. (2007) Multiple residues in the transmembrane helix and connecting peptide of mouse tapasin stabilize the transporter associated with the antigen-processing TAP2 subunit. J. Biol. Chem. 282, Hulpke, S. et al. (2012) Molecular architecture of the MHC I peptideloading complex: one tapasin molecule is essential and sufficient for antigen processing. FASEB J. 26, Ortmann, B. et al. (1997) A critical role for tapasin in the assembly and function of multimeric MHC class I TAP complexes. Science 277, Everett, M.W. and Edidin, M. (2007) Tapasin increases efficiency of MHC I assembly in the endoplasmic reticulum but does not affect MHC I stability at the cell surface. J. Immunol. 179, Vigneron, N. et al. (2009) Functional significance of tapasin membrane association and disulfide linkage to ERp57 in MHC class I presentation. Eur. J. Immunol. 39, Rizvi, S.M. and Raghavan, M. (2010) Mechanisms of function of tapasin, a critical major histocompatibility complex class I assembly factor. Traffic 11, Lehner, P.J. et al. (1998) Soluble tapasin restores MHC class I expression and function in the tapasin-negative cell line.220. Immunity 8, Bangia, N. et al. (1999) The N-terminal region of tapasin is required to stabilize the MHC class I loading complex. Eur. J. Immunol. 29, Garbi, N. et al. (2003) A major role for tapasin as a stabilizer of the TAP peptide transporter and consequences for MHC class I expression. Eur. J. Immunol. 33, Raghuraman, G. et al. (2002) Tapasin interacts with the membranespanning domains of both TAP subunits and enhances the structural stability of TAP1TAP2 complexes. J. Biol. Chem. 277, Williams, A.P. et al. (2002) Optimization of the MHC class I peptide cargo is dependent on tapasin. Immunity 16, Howarth, M. et al. (2004) Tapasin enhances MHC class I peptide presentation according to peptide half-life. Proc. Natl. Acad. Sci. U.S.A. 101, Chen, M. and Bouvier, M. (2007) Analysis of interactions in a tapasin/ class I complex provides a mechanism for peptide selection. EMBO J. 26, Praveen, P.V. et al. (2010) Tapasin edits peptides on MHC class I molecules by accelerating peptide exchange. Eur. J. Immunol. 40, Wearsch, P.A. and Cresswell, P. (2007) Selective loading of high-affinity peptides onto major histocompatibility complex class I molecules by the tapasin ERp57 heterodimer. Nat. Immunol. 8, Chapman, D.C. and Williams, D.B. (2010) ER quality control in the biogenesis of MHC class I molecules. Semin. Cell Dev. Biol. 21, Zhang, W. et al. (2011) A role for UDP-glucose glycoprotein glucosyltransferase in expression and quality control of MHC class I molecules. Proc. Natl. Acad. Sci. U.S.A. 108, Wearsch, P.A. et al. (2011) Essential glycan-dependent interactions optimize MHC class I peptide loading. Proc. Natl. Acad. Sci. U.S.A. 108, Dong, G. et al. (2009) Insights into MHC class I peptide loading from the structure of the tapasin-erp57 thiol oxidoreductase heterodimer. Immunity 30, Turnquist, H.R. et al. (2001) A region of tapasin that affects L d binding and assembly. J. Immunol. 167, Simone, L.C. et al. (2012) Productive association between MHC class I and tapasin requires the tapasin transmembrane/cytosolic region and the tapasin C-terminal Ig-like domain. Mol. Immunol. 49, Lewis, J.W. et al. (1996) Point mutations in the alpha 2 domain of HLA- A2.1 define a functionally relevant interaction with TAP. Curr. Biol. 6, Peace-Brewer, A.L. et al. (1996) A point mutation in HLA-A*0201 results in failure to bind the TAP complex and to present virusderived peptides to CTL. Immunity 4, Yu, Y.Y. et al. (1999) An extensive region of an MHC class I alpha 2 domain loop influences interaction with the assembly complex. J. Immunol. 163, Peh, C.A. et al. (1998) HLA-B27-restricted antigen presentation in the absence of tapasin reveals polymorphism in mechanisms of HLA class I peptide loading. Immunity 8,

9 56 Sieker, F. et al. (2008) Differential tapasin dependence of MHC class I molecules correlates with conformational changes upon peptide dissociation: a molecular dynamics simulation study. Mol. Immunol. 45, Zarling, A.L. et al. (2003) Tapasin is a facilitator, not an editor, of class I MHC peptide binding. J. Immunol. 171, Roder, G. et al. (2011) Tapasin discriminates peptide human leukocyte antigen-a*02:01 complexes formed with natural ligands. J. Biol. Chem. 286, Neisig, A. et al. (1996) Allele-specific differences in the interaction of MHC class I molecules with transporters associated with antigen processing. J. Immunol. 156, Paulsson, K.M. et al. (2002) Association of tapasin and COPI provides a mechanism for the retrograde transport of major histocompatibility complex (MHC) class I molecules from the Golgi complex to the endoplasmic reticulum. J. Biol. Chem. 277, Paulsson, K.M. et al. (2006) The double lysine motif of tapasin is a retrieval signal for retention of unstable MHC class I molecules in the endoplasmic reticulum. J. Immunol. 176, Howe, C. et al. (2009) Calreticulin-dependent recycling in the early secretory pathway mediates optimal peptide loading of MHC class I molecules. EMBO J. 28, Dick, T.P. et al. (2002) Disulfide bond isomerization and the assembly of MHC class I-peptide complexes. Immunity 16, Peaper, D.R. et al. (2005) Tapasin and ERp57 form a stable disulfidelinked dimer within the MHC class I peptide-loading complex. EMBO J. 24, Stepensky, D. et al. (2007) Aggregate formation by ERp57-deficient MHC class I peptide-loading complexes. Traffic 8, Peaper, D.R. and Cresswell, P. (2008) The redox activity of ERp57 is not essential for its functions in MHC class I peptide loading. Proc. Natl. Acad. Sci. U.S.A. 105, Zhang, Y. et al. (2009) ERp57 does not require interactions with calnexin and calreticulin to promote assembly of class I histocompatibility molecules, and it enhances peptide loading independently of its redox activity. J. Biol. Chem. 284, Garbi, N. et al. (2006) Impaired assembly of the major histocompatibility complex class I peptide-loading complex in mice deficient in the oxidoreductase ERp57. Nat. Immunol. 7, Kienast, A. et al. (2007) Redox regulation of peptide receptivity of major histocompatibility complex class I molecules by ERp57 and tapasin. Nat. Immunol. 8, Gao, B. et al. (2002) Assembly and antigen-presenting function of MHC class I molecules in cells lacking the ER chaperone calreticulin. Immunity 16, Ireland, B.S. et al. (2008) Lectin-deficient calreticulin retains full functionality as a chaperone for class I histocompatibility molecules. Mol. Biol. Cell 19, Fu, H. et al. (2009) Calreticulin maintains the low threshold of peptide required for efficient antigen presentation. Mol. Immunol. 46, Oliver, J.D. et al. (1999) ERp57 functions as a subunit of specific complexes formed with the ER lectins calreticulin and calnexin. Mol. Biol. Cell 10, Frickel, E.M. et al. (2002) TROSY-NMR reveals interaction between ERp57 and the tip of the calreticulin P-domain. Proc. Natl. Acad. Sci. U.S.A. 99, Rizvi, S.M. et al. (2011) Distinct functions for the glycans of tapasin and heavy chains in the assembly of MHC class I molecules. J. Immunol. 186, Del Cid, N. et al. (2010) Modes of calreticulin recruitment to the major histocompatibility complex class I assembly pathway. J. Biol. Chem. 285, Li, S. et al. (1997) Cloning and functional characterization of a subunit of the transporter associated with antigen processing. Proc. Natl. Acad. Sci. U.S.A. 94, Rufer, E. et al. (2007) Molecular architecture of the TAP-associated MHC class I peptide-loading complex. J. Immunol. 179, Leonhardt, R.M. et al. (2005) Critical role for the tapasin-docking site of TAP2 in the functional integrity of the MHC class I-peptide-loading complex. J. Immunol. 175, Bangia, N. and Cresswell, P. (2005) Stoichiometric tapasin interactions in the catalysis of major histocompatibility complex class I molecule assembly. Immunology 114, Panter, M.S. et al. (2012) Dynamics of major histocompatibility complex class I association with the human peptide-loading complex. J. Biol. Chem. 287, Reits, E.A. et al. (2000) The major substrates for TAP in vivo are derived from newly synthesized proteins. Nature 404, Schoenhals, G.J. et al. (1999) Retention of empty MHC class I molecules by tapasin is essential to reconstitute antigen presentation in invertebrate cells. EMBO J. 18, Grandea, A.G., 3rd et al. (2000) Impaired assembly yet normal trafficking of MHC class I molecules in tapasin mutant mice. Immunity 13, Boyle, L.H. et al. (2013) Tapasin-related protein TAPBPR is an additional component of the MHC class I presentation pathway. Proc. Natl. Acad. Sci. U.S.A. 110, Garstka, M. et al. (2007) Peptide-receptive major histocompatibility complex class I molecules cycle between endoplasmic reticulum and cis-golgi in wild-type lymphocytes. J. Biol. Chem. 282, Ghanem, E. et al. (2010) The transporter associated with antigen processing (TAP) is active in a post-er compartment. J. Cell Sci. 123, Lu, X. et al. (2012) Endogenous viral antigen processing generates peptide-specific MHC class I cell-surface clusters. Proc. Natl. Acad. Sci. U.S.A. 109, Lautscham, G. et al. (2003) TAP-independent antigen presentation on MHC class I molecules: lessons from Epstein Barr virus. Microbes Infect. 5, Del Val, M. et al. (2013) Are membrane proteins favored over cytosolic proteins in TAP-independent processing pathways? Mol. Immunol. 55, Lautscham, G. et al. (2003) Identification of a TAP-independent, immunoproteasome-dependent CD8 + T-cell epitope in Epstein Barr virus latent membrane protein 2. J. Virol. 77, Suh, W.K. et al. (1999) Interaction of murine MHC class I molecules with tapasin and TAP enhances peptide loading and involves the heavy chain alpha3 domain. J. Immunol. 162,