1. Biochemistry and Chemical Biology
  2. Immunology and Inflammation
Download icon

A loop structure allows TAPBPR to exert its dual function as MHC I chaperone and peptide editor

  1. Lina Sagert
  2. Felix Hennig
  3. Christoph Thomas  Is a corresponding author
  4. Robert Tampé  Is a corresponding author
  1. Institute of Biochemistry, Biocenter, Goethe University Frankfurt, Germany
Research Article
  • Cited 2
  • Views 847
  • Annotations
Cite this article as: eLife 2020;9:e55326 doi: 10.7554/eLife.55326

Abstract

Adaptive immunity vitally depends on major histocompatibility complex class I (MHC I) molecules loaded with peptides. Selective loading of peptides onto MHC I, referred to as peptide editing, is catalyzed by tapasin and the tapasin-related TAPBPR. An important catalytic role has been ascribed to a structural feature in TAPBPR called the scoop loop, but the exact function of the scoop loop remains elusive. Here, using a reconstituted system of defined peptide-exchange components including human TAPBPR variants, we uncover a substantial contribution of the scoop loop to the stability of the MHC I-chaperone complex and to peptide editing. We reveal that the scoop loop of TAPBPR functions as an internal peptide surrogate in peptide-depleted environments stabilizing empty MHC I and impeding peptide rebinding. The scoop loop thereby acts as an additional selectivity filter in shaping the repertoire of presented peptide epitopes and the formation of a hierarchical immune response.

eLife digest

Cells in the body keep the immune system informed about their health by showing it fragments of the proteins they have been making. They display these fragments, called peptides, on MHC molecules for passing immune cells to inspect. That way, if a cell becomes infected and starts to make virus proteins, or if it becomes damaged and starts to make abnormal proteins, the immune system can ‘see’ what is happening inside and trigger a response.

MHC molecules each have a groove that can hold one peptide for inspection. For the surveillance system to work, the cell needs to load a peptide into each groove before the MHC molecules reach the cell surface. Once the MHC molecules are on the cell surface, the peptides need to stay put; if they fall out, the immune system will not be able to detect them. The problem for the cell is that not all peptides fit tightly into the groove, so the cell needs to check each one before it goes out. It does this using a protein called TAPBPR.

TAPBPR has a finger-like structural feature called the "scoop loop", which fits into the end of the MHC groove while the molecule waits for a peptide. It was not clear, however, what this loop actually does. To investigate, Sagert et al. mutated the scoop loop of TAPBPR to see what happened to MHC loading in test tubes.

The experiments revealed that the scoop loop plays two important roles. The first is to keep the MHC molecule stable when it is empty, and the second is to hinder unsuitable peptides from binding. The scoop loop sticks into one side of the groove like a tiny hairpin, so that pushed-out, poorly fitting peptides cannot reattach. At the same time, it holds the MHC molecule steady until a better peptide comes along and only releases when the new peptide has slotted tightly into the groove.

Understanding how cells choose which peptides to show to the immune system is important for many diseases. If cells are unable to find a suitable peptide for a particular illness, it can stop the immune system from mounting a strong response. Further research into this quality control process could aid the design of new therapies for infectious diseases, autoimmune disorders and cancer.

Introduction

Nucleated cells of higher vertebrates provide information about their health status by presenting a selection of endogenous peptides on MHC I molecules at the cell surface. By sampling these peptide-MHC I (pMHC I) complexes, CD8+ T lymphocytes are able to detect and eliminate infected or cancerous cells (Blum et al., 2013; Rock et al., 2016). In a process called peptide editing or proofreading, peptides derived from the cellular proteome are selected for their ability to form stable pMHC I complexes. This peptide editing is known to be catalyzed by the two homologous MHC I-specific chaperones tapasin (Tsn) and TAP-binding protein-related (TAPBPR) (Fleischmann et al., 2015; Hermann et al., 2015; Morozov et al., 2016; Neerincx and Boyle, 2017; Tan et al., 2002; Thomas and Tampé, 2019; Wearsch and Cresswell, 2007; Wearsch et al., 2011). The selection of high-affinity MHC I-associated peptide epitopes is of pivotal importance not only for immunosurveillance by effector T lymphocytes, but also for priming of naïve T cells and T cell differentiation. As an integral constituent of the peptide-loading complex (PLC) in the endoplasmic reticulum (ER) membrane, the ER-restricted Tsn functions in a ‘nanocompartment’ characterized by a high concentration of diverse, optimal peptides. The peptides are shuttled into the ER by the heterodimeric ABC (ATP-binding cassette) transporter associated with antigen processing TAP1/2, the central component of the PLC (Abele and Tampé, 2018). In the ER, most peptides are further trimmed by the aminopeptidases ERAP1 and ERAP2 to an optimal length for binding in the MHC I groove (Evnouchidou and van Endert, 2019; Hammer et al., 2007). In contrast to Tsn, TAPBPR operates independently of the PLC and is also found in the peptide-depleted cis-Golgi network (Boyle et al., 2013). Fundamental insights into the architecture and dynamic nature of the Tsn-containing PLC have come from a recent cryo-EM study of the fully-assembled human PLC (Blees et al., 2017), while the basic principles underlying catalyzed peptide editing have been elucidated by crystal structures of the TAPBPR-MHC I complex (Jiang et al., 2017; Thomas and Tampé, 2017a): TAPBPR stabilizes the peptide-binding groove in a widened conformation primarily through the MHC I α2–1 helix, distorts the floor of the binding groove, and shifts the position of β2-microglobulin (β2m). Furthermore, one of the two TAPBPR-MHC I complex structures revealed a remarkable structural feature in TAPBPR named the scoop loop (Thomas and Tampé, 2017a). In TAPBPR, this loop is significantly longer than the corresponding region in Tsn, which was not resolved in the X-ray structure of Tsn (Dong et al., 2009). Notably, the scoop loop of TAPBPR is located in the F-pocket region of the empty MHC I binding groove (Figure 1A,B). By anchoring the C-terminal part of the peptide, the F pocket region is crucially involved in defining pMHC I stability (Abualrous et al., 2015; Hein et al., 2014). The scoop loop occupies a position that is incompatible with peptide binding and displaces or coordinates several key MHC I residues responsible for binding the C terminus of the peptide. We therefore proposed that the scoop loop can be regarded as a surrogate for the C terminus of the displaced peptide, stabilizing the inherently labile empty MHC I molecule (Thomas and Tampé, 2017a). At the same time, by occupying a region critical to peptide binding, the scoop loop might allow only high-affinity peptides to re-enter the MHC I binding groove after displacement of sub-optimal peptide. The proposed importance of the scoop loop for TAPBPR function has recently been scrutinized in a study by Ilca et al. investigating TAPBPR scoop-loop variants using immunopeptidomics and cell-based assays (Ilca et al., 2018). Ilca et al. found that a specific leucine residue in the scoop loop facilitates peptide displacement on MHC I allomorphs favoring hydrophobic peptide side chains in their F pocket. Here, we aimed to clarify the role of the scoop loop during TAPBPR-catalyzed peptide editing using in vitro interaction and peptide-exchange studies with defined, purified components. We demonstrate that the scoop loop is of critical importance for TAPBPR-mediated stabilization of empty MHC I clients in peptide-depleted environments and contributes to peptide quality control during editing by impeding released peptide to rebind in the MHC I groove. Collectively, our data support a crucial role for the TAPBPR scoop loop in establishing a hierarchical immune response.

Expression and purification of different TAPBPR scoop-loop variants and MHC I chaperone clients.

(A) X-ray structure of the TAPBPR-MHC I complex in cartoon representation (PDB ID: 5OPI). The zoom-in shows how the TAPBPR scoop loop (purple) is inserted into the F-pocket region of the MHC I peptide-binding groove that is occupied by the C terminus of the peptide before peptide displacement. (B) 2Fo-Fc electron density of the X-ray structure in the region of the scoop loop, contoured at 0.8σ. The width of the helix cartoons has been reduced to facilitate visualization of the electron density. The viewing direction is indicated by the black arrow in panel (A). (C) Sequence alignment of the scoop-loop region in the TAPBPR constructs used in this study. (D, E) Purified proteins used in the current study were analyzed by non-reducing SDS-PAGE. The MHC I allomorphs H2-Db (mouse) and HLA-A*02:01 (human) were refolded in the presence of β2m and peptide. (F) The TAPBPR proteins, injected at different concentrations to facilitate comparison, eluted as monodisperse samples during size-exclusion chromatography (SEC). Abbreviations: MHC I hc: MHC I heavy chain; wt: wildtype; Tsn: tapasin; SL: scoop loop; M: protein marker; kDa: kilodalton; A280: absorption at 280 nm; V0: void volume; Vt: total volume.

Results

Design of TAPBPR scoop-loop variants

To investigate the function of the scoop loop, we prepared two human TAPBPR variants: TAPBPRTsn-SL, in which the TAPBPR scoop loop was replaced with the corresponding shorter loop of Tsn, and TAPBPRΔSL, in which the original scoop loop was essentially deleted by replacing it with three glycine residues to preserve proper folding of the MHC I chaperone (Figure 1C). The ER-lumenal domains of wildtype (wt) TAPBPR and the variants, each harboring a C-terminal histidine tag, were expressed in insect cells and purified from the cell culture supernatant via immobilized-metal affinity chromatography (IMAC) and size-exclusion chromatography (SEC). As MHC I chaperone clients, we chose mouse H2-Db and human HLA-A*02:01, which are known to interact with TAPBPR (Hermann et al., 2013; Ilca et al., 2019; Morozov et al., 2016). HLA-A*02:01, the major MHC I allomorph in the Caucasian population and found in more than 50% of the global population, presents a diverse spectrum of immunodominant autoimmune, viral, and tumor epitopes and is therefore medically highly relevant (Boucherma et al., 2013). The MHC I allomorphs were expressed in E. coli as inclusion bodies and refolded in the presence of β2m and fluorescently-labeled or photo-cleavable peptide (Rodenko et al., 2006). The highly pure TAPBPR variants and pMHC I complexes eluted as monodisperse samples at expected size during SEC (Figure 1D–F).

Scoop-loop variants have reduced chaperone activity towards peptide-free MHC I

During peptide exchange, MHC I molecules go through a peptide-free high-energy intermediate state after peptide release and before re-entry of a new peptide. A hallmark of peptide editors like TAPBPR is their ability to recognize and chaperone this intermediate until it is located in a peptide-rich environment where a high-affinity peptide ligand can enter the MHC I binding groove (Thomas and Tampé, 2019; Thomas and Tampé, 2017b). To scrutinize the role of the scoop loop in chaperoning empty MHC I, we tested the ability of our TAPBPR variants to stabilize peptide-free H2-Db. Hence, H2-Db (10 µM) loaded with a photo-cleavable peptide was incubated with TAPBPR (3 µM) under UV exposure. Subsequent SEC analysis revealed that both TAPBPRTsn-SL and TAPBPRΔSL are, in principle, competent to form complexes with MHC I (Figure 2A). However, in comparison to TAPBPRwt (Figure 2A,B), the amount of H2-Db complex detected for TAPBPRTsn-SL and TAPBPRΔSL during SEC was reduced by around 40% and 90%, respectively (Figure 2C). After reanalysis of the MHC I chaperone complexes by SEC, the mutant complexes were mostly dissociated, indicating kinetic instability (Figure 2—figure supplement 1A). In contrast, isolation and reinjection of the wt complex showed that it remained stable for the duration of the experiment (Figure 2—figure supplement 1A,B). Yet, in the presence of a high-affinity peptide, even the TAPBPRwt-MHC I complex dissociated, in accordance with the role of TAPBPR as a chaperone that stabilizes the MHC I as long as no optimal peptide is present (Figure 2—figure supplement 1B). Taken together, these findings demonstrate that the scoop loop is crucial to an extended lifetime of the chaperone-client complex, enabling the escorting of empty MHC I by TAPBPR in a peptide-deficient environment.

Figure 2 with 1 supplement see all
Complex formation between MHC I and TAPBPR variants.

(A) H2-Db (10 µM) loaded with a photo-cleavable peptide (RGPGRAFJ*TI, J* denotes photocleavable amino acid) was irradiated with UV light in the presence of TAPBPRwt (3 µM, red), TAPBPRTsn-SL (blue), or TAPBPRΔSL (yellow) and subsequently analyzed by SEC. The different elution volumes of the first main peak, marked by dashed lines, already hint at different complex stabilities. (B) Deconvolution of size-exclusion chromatogram from TAPBPRwt complex formation (experiment independent of the sample shown in (A)). The experimental chromatogram (red) was deconvoluted using three Gaussian functions (gray) that can be ascribed to the TAPBPR-H2-Db complex (1.06 mL), free TAPBPR (1.12 mL), and free H2-Db (1.20 mL). The sum of the three Gaussians is shown as dotted curve. The residual plot depicted beneath the main panel shows the difference between the experimental data and the sum. (C) Stability of complexes formed by TAPBPRwt, TAPBPRTsn-SL, and TAPBPRΔSL, respectively, as judged by the area of the complex peak obtained by deconvolution. Data represent mean ± SD (n = 2).

Scoop-loop variants retain their function in catalyzing peptide dissociation from MHC I

After investigating the chaperone activity of the TAPBPR scoop-loop mutants, we tested their ability to displace MHC I-bound peptide. To this end, we employed an in-vitro peptide exchange assay similar to the one previously described for measuring the activity of Tsn (Fleischmann et al., 2015; Chen and Bouvier, 2007). Dissociation of medium-affinity fluorescent peptide from refolded and purified p*MHC I (p* denotes fluorescently-labeled peptide) was monitored by fluorescence polarization after addition of a 1000-fold molar excess of unlabeled high-affinity competitor peptide in the absence or presence of TAPBPR (Figure 3A). The large molar excess of unlabeled competitor peptide ensures that once a fluorescent peptide dissociates, it does not rebind, but is replaced by an unlabeled competitor-peptide molecule. The observed rate constant is thus solely determined by the dissociation rate constant of the fluorescent peptide. The condition of this assay mimics the environment of the PLC, where optimal, high-affinity peptides abound. For the mouse MHC I allomorph H2-Db, TAPBPRwt and the scoop-loop variants accelerated the uncatalyzed peptide release (2.53 ± 0.37 × 10−3 s−1) to a similar extent. The TAPBPRΔSL mutant lacking the entire scoop loop exhibited slightly reduced activity (7.68 ± 1.17 × 10−3 s−1) compared to the wt protein (10.41 ± 0.54 × 10−3 s−1), whereas TAPBPRTsn-SL was slightly more active (12.64 ± 1.03 × 10−3 s−1) (Figure 3B,C). When we performed the experiment at a much lower TAPBPR concentration (75 nM), the TAPBPRs retained their activity, and the gradual activity differences between the variants remained (Figure 3—figure supplement 1). This suggests that TAPBPRwt and the scoop-loop mutants have similar affinities for H2-Db. TAPBPRwt was even able to catalyze displacement of a high-affinity peptide from H2-Db, although the catalytic effect was considerably smaller (1.8-fold acceleration) than for H2-Db loaded with the medium-affinity peptide (4.1-fold acceleration) (Figure 3—figure supplement 2A,B). In a second set of experiments, we analyzed peptide dissociation from the human MHC I allomorph HLA-A*02:01. Similar to the experiments with H2-Db, in a peptide-rich environment (1000-fold molar excess of peptide), the highest catalytic activity towards HLA-A*02:01 was observed for TAPBPRTsn-SL, followed by TAPBPRwt and TAPBPRΔSL; yet, the differences in activity between the three TAPBPRs were more pronounced, and the acceleration of the uncatalyzed peptide dissociation from HLA-A*02:01 (1.90 ± 0.04 × 10−3 s−1) by TAPBPRTsn-SL (26.31 ± 2.59 × 10−3 s−1) and TAPBPRwt (15.79 ± 0.71 × 10−3 s−1) was significantly higher than for H2-Db, while the activity of TAPBPRΔSL (8.52 ± 1.18 × 10−3 s−1) remained almost the same (Figure 3D,E).

Figure 3 with 2 supplements see all
Peptide release from H2-Db and HLA-A*02:01 in peptide-rich environment.

(A) Schematic of peptide displacement assay. (B) Peptide dissociation kinetics from H2-Db (300 nM) loaded with fluorescently-labeled peptide (TQSC*NTQSI) was monitored in real time by fluorescence polarization. The arrow indicates the addition of a 1000-fold molar excess of unlabeled high-affinity competitor peptide (ASNENMETM) without TAPBPR (black trace) or in combination with 1 µM TAPBPR (red, blue, and yellow traces). (C) Average dissociation rate constants of uncatalyzed and catalyzed peptide dissociation from H2-Db, using the same conditions as in (B). Data represent mean ± SD (n = 2–6). (D) Representative fluorescence polarization traces of uncatalyzed and catalyzed peptide (FLPSDC*FPSF) dissociation from HLA-A*02:01 (300 nM). The arrow indicates the addition of a 1000-fold molar excess of unlabeled competitor peptide (FLPSDEEPYV, 300 µM) with and without TAPBPR (1 µM). (E) Average dissociation rate constants of uncatalyzed and catalyzed peptide dissociation from HLA-A*02:01, using the same experimental conditions as in (D). Data represent mean ± SD (n = 3). (F) Peptide dissociation from H2-Db (300 nM) after addition (arrow) of unlabeled competitor peptide (300 µM) without TAPBPR or in combination with the interface mutants TN6-TAPBPR and TN3-Ala-TAPBPR (1 µM each), respectively. A TAPBPRwt-catalyzed peptide release reaction is shown as reference. The average dissociation rate constants in the presence of TN6 (koff = 2.53 ± 0.30×10−3 s−1) and TN3-Ala (koff = 4.23 ± 0.45×10−3 s−1) are shown in panel (C). Abbreviations: β2m: β2-microglobulin; MHC I hc: MHC I heavy chain; pMHC I: peptide-MHC I; mP: milli-polarization units; wt: wildtype; Tsn: tapasin; SL: scoop loop.

The validity of our peptide exchange assay was confirmed by two interface mutants of TAPBPRwt, TN3-Ala and TN6. The TN3 (E72K) and TN6 (E185K, R187E, Q189S, Q261S) mutants were initially described for Tsn to significantly reduce or abolish MHC I binding (Dong et al., 2009). The impact of the TN6 mutations on MHC I interaction was later confirmed for TAPBPR (Morozov et al., 2016). According to the TAPBPR-MHC I crystal structures (Jiang et al., 2017; Thomas and Tampé, 2017a), the residue in TAPBPR (E105) corresponding to the mutated residue in Tsn-TN3 forms a hydrogen bond with the swung-out Y84 of the MHC heavy chain, which is involved in coordinating the C-terminus of the peptide in liganded MHC. We reasoned that a mutation to Ala instead of Lys might increase the mutational effect and therefore generated the TN3-Ala mutant. Two of the mutated residues in TN6 (R210 and Q212) are part of the jack hairpin of TAPBPR and form several interactions with MHC I heavy-chain residues, while Q275 lies in the interface with the α2–1 helix and the β8 sheet in the floor of the MHC I binding groove. Consequently, TN3-Ala and TN6 displayed drastically reduced activity towards H2-Db in our peptide-exchange experiment, with peptide dissociation rate constants close to the value of the uncatalyzed reaction (Figure 3C,F). In summary, the results of our exchange assays demonstrate that under peptide-rich condition, the tested TAPBPR variants differ gradually in their displacement activity in an allomorph-dependent manner. But even the TAPBPRΔSL mutant lacking the scoop loop is still able to substantially accelerate peptide dissociation from MHC I.

The scoop loop acts as an internal peptide competitor

In the TAPBPR-MHC I crystal structure, the scoop loop binds in the F pocket region of the MHC binding groove and appears to act as a surrogate for the peptide C terminus (Thomas and Tampé, 2017a). This notion is corroborated by our SEC analyses, which show that the scoop loop stabilizes peptide-free MHC I. We therefore wondered if the scoop loop impedes rebinding of displaced peptide and functions ‘in cis’ as a tethered, internal peptide competitor in the F pocket with extremely high effective concentration. To test this hypothesis, we modified the peptide exchange assay for H2-Db and HLA-A*02:01 by adding in a first step only TAPBPR without competitor peptide, which allowed us to monitor the change in free and bound fluorescent peptide under the influence of peptide rebinding in the presence of TAPBPR (Figure 4A). This condition mimics the physiological environment TAPBPR is operating in, where optimal replacement peptides are scarce. Strikingly, after addition of the different TAPBPRs to H2-Db loaded with fluorescent peptide, the polarization changes, which correspond to the changes in the ratio of free to bound peptide, diverged dramatically (Figure 4B). Peptide dissociation was most pronounced for TAPBPRwt with the native scoop loop, reaching ~ 60% peptide release, whereas only ~ 12% of the peptide population was released from H2-Db by TAPBPRTsn-SL, and almost no decrease in polarization was caused by TAPBPRΔSL. Similar to our original peptide exchange assay (Figure 3), differences between the two MHC I allomorphs were observed: In comparison to H2-Db, TAPBPRTsn-SL-induced peptide dissociation from HLA-A*02:01 was significantly stronger, approaching the level of peptide release induced by TAPBPRwt (Figure 4—figure supplement 1A). Peptide release was also peptide-dependent, as H2-Db loaded with a high-affinity peptide led to a significantly smaller decline in bound peptide (Figure 3—figure supplement 2C). After addition of competitor peptide (2nd step), the observed dissociation rate constants were in the same range as the values determined for the one-step experiment. Moreover, the level of released peptide after TAPBPR addition was titratable and reached saturation at 3 µM TAPBPR (Figure 4C–E, Figure 4—figure supplement 1B). Under the given conditions, TAPBPRwt was able to dissociate 70% (H2-Db) and 80% (HLA-A*02:01) of total MHC I-associated peptide, respectively (Figure 4C, Figure 4—figure supplement 1B). These results suggest that the scoop loop interferes with re-binding of displaced peptide. It can only be completely dislodged from the MHC I binding pocket by a high-affinity peptide. The scoop loop thus acts as a crucial selectivity filter during peptide editing on MHC I.

Figure 4 with 1 supplement see all
The scoop loop acts as a selectivity filter during peptide editing.

(A) Schematic of two-step peptide exchange assay. (B) Peptide displacement from H2-Db (300 nM) loaded with fluorescently-labeled peptide (TQSC*NTQSI) was monitored by fluorescence polarization after addition of TAPBPR (1 µM, first arrow) and after subsequent addition of a 1000-fold molar excess of unlabeled high-affinity competitor peptide (ASNENMETM, 300 µM, second arrow). (C) Titration of peptide-loaded H2-Db (300 nM) with varying concentrations of TAPBPRwt (first arrow) and final addition of a 1000-fold molar excess of unlabeled high-affinity competitor peptide (300 µM, second arrow). (D) Peptide displacement from H2-Db (300 nM) loaded with fluorescently-labeled peptide monitored by fluorescence polarization after addition of 3 µM and 10 µM TAPBPRTsn-SL, respectively (first arrow), and after subsequent addition of a 1000-fold molar excess of unlabeled high-affinity competitor peptide (300 µM, second arrow). (E) Peptide displacement from H2-Db (300 nM) loaded with fluorescently-labeled peptide monitored by fluorescence polarization after addition of TAPBPRΔSL (3 µM, first arrow) and after subsequent addition of a 1000-fold molar excess of unlabeled high-affinity competitor peptide (300 µM, second arrow). Data shown in (B)-(E) are representative of three independent measurements.

Discussion

Tsn and TAPBPR are MHC I-dedicated chaperones, which facilitate loading and selective exchange of antigenic peptides and thereby generate stable pMHC I complexes that shape a hierarchical immune response. The molecular underpinnings of their chaperone and peptide proofreading activities have only recently been uncovered by crystal structures of the TAPBPR-MHC I complex (Jiang et al., 2017; Thomas and Tampé, 2017a). Notably, one of the X-ray structures resolved a loop structure, termed the scoop loop, that is wedged into the F-pocket region of the empty MHC I binding groove and has been postulated to play an important role during peptide exchange (Thomas and Tampé, 2017a). Here, we show that the TAPBPR scoop loop is indeed critically important in chaperoning intrinsically unstable empty MHC I clients in a peptide-depleted environment. This is illustrated by the reduced chaperone activity of TAPBPRTsn-SL, which harbors the shorter Tsn scoop loop, and by the dramatically reduced lifetime of the TAPBPRΔSL complex. In a peptide-rich, PLC-like environment, emulated by our one-step displacement experiments, the TAPBPRTsn-SL mutant displays the highest activity, while TAPBPRΔSL retains the ability to displace peptide. The latter observation appears to be in contrast to the study by Ilca et al. which found that TAPBPR with a mutated, but full-length scoop loop loses its ability to effectively mediate peptide dissociation (Ilca et al., 2018). In addition to stabilizing the chaperone-MHC I complex, we demonstrate that the TAPBPR scoop loop acts as an internal peptide competitor, and thus, as a selectivity filter in the discrimination between low- and high-affinity peptides. Although a direct competition appears to be the most obvious explanation for the effect on peptide rebinding, we cannot exclude that the scoop loop exerts its influence on peptide rebinding by an allosteric mechanism. The peptide-filtering activity seems to be allomorph-dependent for TAPBPRTsn-SL. Our current interpretation of this allomorph specificity is that the Tsn scoop loop interacts more strongly with the F-pocket region of HLA-A*02:01 and is therefore able to impede peptide rebinding more efficiently than in the case of H2-Db. In contrast, TAPBPRwt shows a strong peptide release activity towards both MHC I allomorphs.

Based on the new insights, we propose the following model of TAPBPR-catalyzed peptide optimization on MHC I (Figure 5): The large concave surface formed by the N-terminal domain of TAPBPR mediates its initial encounter with a suboptimally-loaded MHC I, assisted by the C-terminal domain of TAPBPR, which contacts the α3 domain of the MHC I heavy chain and β2m. TAPBPR facilitates the release of low- to medium-affinity peptides primarily by widening the peptide-binding groove through the MHC I α2–1-helix, fastening the peptide-coordinating Tyr84, distorting the floor of the binding groove, and shifting the position of β2m (Jiang et al., 2017; Thomas and Tampé, 2017a). This remodeling is made possible by the intrinsic plasticity of MHC I molecules (Bailey et al., 2015; Garstka et al., 2011; McShan et al., 2019; Natarajan et al., 2018; Thomas and Tampé, 2017b; van Hateren et al., 2017; van Hateren et al., 2015; Wieczorek et al., 2017), and it appears to be induced primarily by structural elements of TAPBPR that lie outside the scoop loop. As a result, the TAPBPRΔSL mutant lacking the scoop loop is still able to catalyze peptide displacement. Once the suboptimal peptide has been released, the scoop loop occupies the position of the peptide C terminus in the F-pocket region. The scoop loop thereby contributes to the stabilization of the peptide-deficient binding groove. Our two-step peptide exchange — mimicking a peptide-depleted environment — demonstrates that the scoop loop functions at the same time as a peptide selectivity filter by impeding re-binding of the replaced peptide, either through direct competition with the C terminus of the incoming replacement peptide or through an allosteric mechanism. Hence, the scoop loop contributes to the significant affinity decrease of incoming peptides for the MHC I groove in the presence of TAPBPR (McShan et al., 2018). Assuming a mode of direct competition, the replacement peptide would dock in the MHC I groove first with its N terminus, before it competes with the TAPBPR scoop loop over the F pocket region (Hafstrand et al., 2019; Thomas and Tampé, 2017a). Negative allosteric coupling between different parts of the MHC I molecule might play a role in the final release of TAPBPR (McShan et al., 2018). The shorter scoop loop in Tsn suggests that its selective pressure on the replacement peptide is weaker than in TAPBPR. Indeed, our fluorescence polarization and SEC analyses show that the tapasin scoop loop in TAPBPRTsn-SL is less efficient in preventing re-binding of dissociated peptide.

Proposed mechanistic functions of the scoop loop in catalyzed peptide proofreading.

MHC I molecules bound to low-affinity peptide are recognized by the peptide editor (TAPBPR) (step 1). The editor lowers the peptide affinity of the suboptimally-loaded MHC I and induces dissociation of the low- to medium-affinity peptide (step 2). The scoop loop, which inserts into the F-pocket region of the peptide-binding groove, crucially contributes to the stabilization of the empty MHC I. In the absence of suitable peptides, empty MHC I clients are thereby held in a stable state until they can be loaded with an optimal epitope, for example in the PLC. Re-binding of the low-affinity peptide (step 3) is impeded by the scoop loop, through direct competition and/or via allosteric means. Only high-affinity peptides are able to compete with the editor over key regions of the peptide-binding groove (step 4) to eventually displace the scoop loop and the editor from the MHC I (step 5). The displaced editor is now ready for a new round of peptide selection, and the stable pMHC I complex is licensed to travel via the Golgi apparatus to the cell surface.

Physiologically, these observations might be explained by the fact that Tsn functions within the PLC, a ‘nanocompartment’ characterized by an abundant and diverse supply of optimal peptides, reaching a bulk concentration of up to 16 µM before the TAP transporter is arrested by trans-inhibition (Grossmann et al., 2014). Moreover, Tsn is supported by other PLC chaperones in stabilizing empty MHC I clients. In contrast, TAPBPR operates as a single MHC I-dedicated chaperone outside the PLC in environments where the concentration of high-affinity peptides is drastically lower and MHC I clients have to be stabilized in a peptide-receptive state for extended periods of time. Long-term stabilization of suboptimally-loaded or empty MHC I by TAPBPR also allows the major ER/cis-Golgi glycoprotein folding sensor UGGT1 (UDP-glucose:glycoprotein glucosyltransferase 1) to re-glucosylate the MHC I molecule in order to feed it back into the calnexin/calreticulin cycle and/or allow recruitment of the MHC I to the PLC (Neerincx et al., 2017; Thomas and Tampé, 2019). In conclusion, the evidence provided by our study indicates that the scoop loop is evolutionarily fine-tuned to enable Tsn and TAPBPR to accomplish their dual function as chaperone and proofreader in the specific subcellular location they operate in. By serving both as a stabilizing element and as selectivity filter in TAPBPR, the scoop loop influences peptide editing and impacts the repertoire of MHC I-associated epitopes presented on the cell surface.

Materials and methods

Key resources table
Reagent type
(species) or resource
DesignationSource or referenceIdentifiersAdditional information
Gene (human)TAPBPRwtPMID:29025996lumenal domain
Gene (human)TAPBPRΔSLThis study
(Figure 1C, Materials and methods section)
lumenal domain
Gene (human)TAPBPRTsn-SLThis study
(Figure 1C, Materials and methods section)
lumenal domain
Gene (human)TAPBPRTN3-AlaPMID:19119025lumenal domain
Gene (human)TAPBPRTN6PMID:19119025lumenal domain
Gene (human)HLA-A*02:01This study
(Materials and methods section)
ectodomain
Gene (human)β2-microglobulinPMID:29025996
Gene (mouse)H2-DbPMID:29025996ectodomain
Strain, strain background (Escherichia coli)DH10BacThermo Fisher Scientific10361012chemically competent
Strain, strain background (Escherichia coli)BL21(DE3)Sigma-AldrichCMC0014chemically competent
Recombinant DNA reagentpET-22Novagen/
Merck
Millipore
69744vector for protein expression in E. coli
Recombinant DNA reagentpET-28Novagen/
Merck
Millipore
69864vector for protein expression in E. coli
Recombinant
DNA reagent
pFastBacI-gp67PMID:29025996transfer vector for Bac-to-Bac system
Cell line (Spodoptera frugiperda)Sf9Thermo Fisher Scientific11496015
Cell line (Spodoptera frugiperda)Sf21Thermo Fisher Scientific11497013
Peptide, recombinant proteinRGPGRAFJ*TI
(photo-P18-I10)
PMID:26869717J* denotes photo-cleavable amino acid
Peptide, recombinant proteinASNENMETMIEDB: epitope ID 4602competitor peptide for H2-Db
Peptide, recombinant proteinFLPSDEEPYVThis study
(Materials and methods section)
competitor peptide for HLA-A*02:01
Peptide, recombinant proteinTQSC*NTQSIThis study
(Materials and methods section)
C* denotes
TAMRA-labeled Cys
Peptide, recombinant proteinFLPSDC*FPSFThis study
(Materials and methods section)
C* denotes
TAMRA-labeled Cys
Peptide, recombinant proteinASNC*NMETMThis study
(Materials and methods section)
C* denotes
TAMRA-labeled Cys
Chemical compound, drugTAMRA-5 maleimideThermo Fisher ScientificT6027
Chemical compound, drugTAMRA-6 C2 maleimideThermo Fisher Scientific48180
Chemical compound, drugFmoc-3-amino-3-(2-nitro)phenyl-propionic acidPeptechCAS #:
517905-93-0
Software, algorithmPrism 6GraphPad Software
Software, algorithmFityk 1.3.1DOI:
10.1107/S0021889810030499
OtherSuperdex 200 Increase 10/300GE Healthcare28990944SEC column
OtherSuperdex 200 Increase 3.2/300GE
Healthcare
28990946SEC column
OtherSuperdex 75 Increase 3.2/300GE Healthcare29148723SEC column
OtherHiLoad
Superdex 75 16/60
GE Healthcare28989333SEC column
OtherFluorolog-3Horiba Jobin Yvonspectro-fluorometer
OtherÄkta PurifierGE Healthcareprotein purification
OtherAgilent 1200Agilentanalytical SEC
OtherLiberty BlueCEM Corporationpeptide synthesizer
OtherX-tremeGENE HPSigma-Aldrich6366236001transfection reagent

DNA constructs

Request a detailed protocol

The DNA constructs of human β2m, the ectodomain of mouse H2-Db, and TAPBPRwt were identical to the ones previously described (Thomas and Tampé, 2017a), except for position 97 in TAPBPRwt, which contained the native cysteine. The TAPBPR scoop loop mutants TAPBPRTsn-SL and TAPBPRΔSL were generated by overlap extension PCR, the TN3-Ala and TN6 mutants were generated by site-directed mutagenesis. The TN3-Ala and TN6 mutants harbored the same mutations that were described for the corresponding mutants of Tsn (Dong et al., 2009), except that in TN3-Ala E105 was mutated to alanine. TAPBPRTsn-SL, TAPBPRΔSL, TN3-Ala, and TN6 all contained the C97A mutation. Human HLA-A*02:01 (amino acids 1–278) was cloned into pET-28 (Novagen, Merck Millipore, Darmstadt, Germany) and ended in a C-terminal His6-tag preceded by a linker (sequence: HE). The amino acid numbering of TAPBPR is based on the mature protein as defined by N-terminal sequencing (Zhang and Henzel, 2004).

Protein expression

Request a detailed protocol

Human β2m and the ectodomains of mouse H2-Db and human HLA-A*02:01 were expressed as inclusion bodies in Escherichia coli BL21(DE3) as described before (Rodenko et al., 2006; Thomas and Tampé, 2017a). TAPBPR proteins were expressed in Spodoptera frugiperda (Sf21 or Sf9) insect cells according to standard protocols for the Bac-to-Bac system (Thermo Fisher Scientific, Waltham, MA). A high-titer recombinant baculovirus stock was used to infect the insect cells at a density of 1.5–2.0 × 106 cells/mL, which were cultivated in Sf-900 III SFM medium (Thermo Fisher Scientific) at 28°C. The cell culture medium containing secreted TAPBPR was harvested 72 hr after infection.

Refolding and purification of β2m

Request a detailed protocol

β2m was refolded by dialysis essentially as described previously (Rodenko et al., 2006) and purified by SEC on a Superdex 75 column (GE Healthcare, Piscataway, NJ) in HEPES-buffered saline (1xHBS: 10 mM HEPES pH 7.2, 150 mM NaCl). Purified protein was concentrated by ultrafiltration (Amicon Ultra 3 kDa MWCO, Merck Millipore).

Peptide synthesis and labeling

Request a detailed protocol

The following peptides were used: the photo-cleavable peptide photo-P18-I10 (RGPGRAFJ*TI) (H2-Db) [J*=3-amino-3-(2-nitro)phenyl-propionic acid], the non-fluorescent competitor peptides ASNENMETM (H2-Db) and FLPSDEEPYV (HLA-A*02:01), as well as the fluorescently labeled peptides TQSC*NTQSI (H2-Db), FLPSDC*FPSF (HLA-A*02:01), ASNC*NMETM (H2-Db) (C* denotes TAMRA-labeled cysteine). Non-natural peptide epitopes were designed based on their theoretical affinities according to the NetMHCpan server (Jurtz et al., 2017). While TQSC*NTQSI and FLPSDC*FPSF were constructed to have medium affinity (500–600 nM), ASNC*NMETM and the competitor peptides were designed to be high-affinity (8–10 nM) ligands. Peptides were synthesized using standard Fmoc solid-phase chemistry and purified by C18 reversed-phase HPLC. The identity of peptides was verified either by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) or by electrospray ionization-mass spectrometry (ESI-MS). In order to site-specifically label peptides with fluorophores, 10.5 µM peptide were incubated with 26 µM TAMRA-5-maleimide (single isomer, Thermo Fisher Scientific) or TAMRA-6 C2 maleimide (Lumiprobe, Hannover, Germany) (used for labeling of FLPSDC*FPSF) overnight at 4°C. Labeled peptides were purified by C18 reversed-phase HPLC, and their identity was confirmed by ESI-MS.

Refolding and purification of MHC I allomorphs

Request a detailed protocol

H2-Db and HLA-A*02:01 were refolded from inclusion bodies by rapid dilution in the presence of purified β2m and peptide according to established protocols (Rodenko et al., 2006). Refolded MHC I complexes were purified by SEC (Superdex 200 Increase 10/300, GE Healthcare) in 1xHBS and concentrated by ultrafiltration (Amicon Ultra, Merck Millipore).

Purification of TAPBPR proteins

Request a detailed protocol

TAPBPR proteins were purified from the insect cell culture medium by IMAC according to a protocol published earlier (Thomas and Tampé, 2017a), polished by SEC (Superdex 200 Increase 10/300, GE Healthcare) in 1xHBS, and concentrated by ultrafiltration (Amicon Ultra, Merck Millipore).

Peptide exchange

Request a detailed protocol

Dissociation of fluorescently labeled peptide from MHC I was monitored at 23°C in 1xHBS by fluorescence polarization (Fluorolog-3 spectrofluorometer, Horiba Jobin Yvon, Bensheim, Germany) with λex/em of 530/560 nm. One-step and two-step dissociation assays were carried out with 300 nM MHC I loaded with TAMRA-labeled peptide, 1 µM TAPBPR, and 300 µM competitor peptide. Dissociation rate constants were determined in GraphPad Prism using a one-phase exponential decay regression.

MHC I-chaperone complex formation

Request a detailed protocol

In the presence of purified TAPBPR (3 µM), photo-P18-I10-loaded H2-Db (10 µM) was irradiated with UV light (36 nm, 185 mW/cm2, 120 s) on ice and afterwards incubated for 10 min at room temperature. Samples were subsequently centrifuged at 10,000xg for 10 min and analyzed by analytical SEC on a Superdex 75 (3.2/300) column (GE Healthcare). SEC runs were conducted in 1xHBS and monitored by absorbance at 280 nm. Chromatograms were deconvoluted into three Gaussian functions using the program Fityk 1.3.1 (Wojdyr, 2010). The amount of complex was assessed by the area of the complex peak.

TAPBPR-MHC I complex stability

Request a detailed protocol

Purified peptide-deficient TAPBPRwt-H2-Db, TAPBPRTsn-SL-H2-Db, and TAPBPRΔSL-H2-Db complexes were analyzed via analytical SEC either on a Superdex 75 (3.2/300) or a Superdex 200 (3.2/300) column (GE Healthcare) at a flow rate of 0.075 mL/min. A separate sample of purified TAPBPRwt-H2-Db complex was incubated with a 100-fold molar excess of high-affinity peptide prior to re-analysis by SEC.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41

Decision letter

  1. Pamela J Bjorkman
    Reviewing Editor; California Institute of Technology, United States
  2. Tadatsugu Taniguchi
    Senior Editor; Institute of Industrial Science, The University of Tokyo, Japan
  3. Malini Raghavan
    Reviewer; University of Michigan, United States
  4. Scheherazade Sadegh-Nasseri
    Reviewer; Johns Hopkins University School of Medicine, United States
  5. Efstratios Stratikos
    Reviewer

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

These careful experiments to analyze the role of the scoop loop in TAPBPR's chaperone activity during peptide loading onto the MHC I proteins will add to the general understanding of antigen presentation.

Decision letter after peer review:

Thank you for submitting your article "A loop structure allows TAPBPR to exert its dual function as MHC I chaperone and peptide editor" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Tadatsugu Taniguchi as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Malini Raghavan (Reviewer #1); Scheherazade Sadegh-Nasseri (Reviewer #2); Efstratios Stratikos (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

TAPBPR functions as a chaperone and peptide editor during the assembly of MHC class I molecules. Two different crystal structures of TAPBPR-MHC class I complexes have been solved (Jiang et al., 2017; Thomas and Tampe, 2017). While the two structures show overall similarities in the interaction modes, the location of the scoop loop (approximately residues 22-25 of TAPBPR) in the F-pocket region of peptide-free MHC class I have been in debate (Natarajan et al., Crit Rev Biochem Mol Biol., 2019). More recent studies support a model in which the scoop loop plays an important role in TAPBPR-mediated peptide dissociation from MHC class I (Ilca et al., 2018). Using purified human and mouse MHC class I molecules, and purified TAPBPR and its scoop loop mutant, the new study by Sagert et al. further analyzes the role of the scoop loop in TAPBPR function. Overall, this is a very well written manuscript that presents well-thought experiments and interesting results. The experiments are well-performed and controlled, and the results are generally clear-cut, under the tested conditions. Concerns with the manuscript relate to some of the experimental interpretations and conclusions leading to the model.

Essential revisions:

1) Figure 2 nicely presents the formation of TAPBPR/H2-Db and how mutating the scoop loop can affect the relative amount of complex formed. While the experiment is highly informative, I am not persuaded that it accurately reports kinetic instability as the authors claim. Rather it is equally possible that it reports thermodynamic destabilisation. Given the reported interactions of the scoop loop with the MHC I peptide-binding groove, lack of the scoop loop or sequence substitution would be expected to weaken the interaction, something that could explain the observations in Figure 2. Can the authors more directly measure the affinity of TAPBPR for MHC I? Even if this is not easily achievable, I think the authors should reconsider their claims for kinetic instability, since deconvolution of kinetics vs. thermodynamics is probably impossible by this particular experiment.

2) Given the result presented in Figure 2, how do the authors explain the finding that scoop loop variants largely retain their ability to catalyse peptide dissociation (Figure 3)? Is it because the scoop loop interactions with MHC are secondary to the overall TAPBPR-MHC interactions? If the scoop loop variants have a reduced affinity for MHC, would an effect be more obvious if experiments in Figure 3 are repeated at lower concentrations?

3) Relative to a TAPBPR chimeric construct containing the shorter scoop loop of tapasin, the wild type TAPBPR is shown to form more stable (higher affinity complexes) with MHC I (Figure 2), although the peptide dissociation function of the wild type TAPBPR is less efficient than the chimera with the tapasin scoop loop (Figure 3). The stronger peptide dissociation effect of a TAPBPR with the tapasin scoop loop compared with TAPBPRwt should be interpreted, explicitly discussed, and taken into account in the overall model of Figure 5.

4) Notably, mutations of other key TABBPR residues at the TAPBPR-MHC class I interface (TN3 and TN6) are shown in the representative figure to have more significant effects on peptide dissociation than a TAPBPR mutant in which the scoop loop was mutated to a 3xGly sequence, suggesting a dominant role for TN3 and TN6-mediated interactions in inducing peptide dissociation. The dominant role of TN3 and TN6-mediated interactions in inducing peptide dissociation should be acknowledged (including within the model), and their quantifications shown in the same way as Figure 3C and 3E.

5) In a variant of the peptide dissociation experiment, peptide dissociation was measured in the presence of the three TAPBPR variants, but in the absence of excess unlabeled peptide (Figure 4 and Figure 4—figure supplement 1). Under this set of conditions, wild type TAPBPR was more efficient at inducing peptide dissociation from H2-Db than the two scoop loop mutants, whereas wild type TAPBPR and the tapasin scoop loop mutant appear to have similar efficiencies towards peptide dissociation from HLA-A2. These data are used to argue for a model in which the scoop loop directly interferes with rebinding of a displaced peptide (internal competitor). However, concerns with these interpretations are:

a) Other models (such as different allosteric effects induced by the three scoop loop variants of TAPBPR) could explain the varying degrees of inhibition of rebinding of dissociated peptides by the TAPBPR variants, which are also predicted to have different affinities for MHC class I. Model validation will need data that can more clearly sort out the difference between an internal competitor model and alternative allosteric effects-based models. If the differences cannot be easily sorted out, at minimum, the model of Figure 5 should be changed to indicate that both the internal competitor model and allosteric model are likely to be relevant.

b) The effects of the three TAPBPR constructs on inhibition of peptide rebinding vary between HLA-A2 and H2-Db (comparing Figure 4 with Figure 4—figure supplement 1). The variable effects of tapasin scoop loop on Db vs. HLA-A2 should be expanded, interpreted, and explicitly discussed.

6) Does the experiment in Figure 4 report a reduced "functional" ability of the scoop loop mutants to remove the peptide from MHC or a reduced affinity of the TAPBPR variants for MHC? This difference is relevant to our understanding of the distinct functionality of the scoop loop (versus being a hinge that makes additional interactions with the MHC). Perhaps the authors could test this by titrating the less active TAPBPRTsn-SL to higher concentrations to see if they can get similar results to the TAPBPRwt-SL.

7) The importance of the scoop loop has recently been demonstrated in the study by Ilca et al. (2018), but the latter paper is only referenced to highlight the differences in conclusions, rather than many other overall similarities in findings. The Ilca et al. paper should be referenced within the Introduction for better acknowledgement of prior contributions to the same questions, and similarities with the present study pointed out in addition to the differences-within the Discussion.

8) Finally, for readers and for the field, this section in the Introduction is very confusing "Furthermore, one of the two TAPBPR-MHC I complex structures revealed a remarkable structural feature in TAPBPR named the scoop loop (Thomas and Tampe, 2017). […] The scoop loop occupies a position that is incompatible with high-affinity peptide binding and displaces or coordinates several key MHC I residues, including Y84, T143, K146, and W147, which are responsible for binding the C terminus of the peptide." The authors should rephrase to clarify what has been discovered regarding the scoop loop previously, what is in debate, and specify the open questions that this study will address.

9) As noted above, model validation will need data that can more clearly sort out the difference between an internal competitor model and alternative allosteric effects-based models. Additional higher resolution structural data, if possible to acquire within two months, may be helpful because the loop was modeled as inserting into the F-pocket in one of the structures (Thomas and Tampe, 2017) but not in the second (Jiang et al., 2017).

Overall, this is a well-executed study, but the data do not integrate well with the proposed model. In line with the above points, the data are more consistent with a relatively minor (and possibly MHC-context dependent) contribution of the scoop loop to peptide editing, rather than the current emphasis of a critical role. This message should be evident within the Abstract and throughout the manuscript.

https://doi.org/10.7554/eLife.55326.sa1

Author response

Essential revisions:

1) Figure 2 nicely presents the formation of TAPBPR/H2-Db and how mutating the scoop loop can affect the relative amount of complex formed. While the experiment is highly informative, I am not persuaded that it accurately reports kinetic instability as the authors claim. Rather it is equally possible that it reports thermodynamic destabilisation. Given the reported interactions of the scoop loop with the MHC I peptide-binding groove, lack of the scoop loop or sequence substitution would be expected to weaken the interaction, something that could explain the observations in Figure 2. Can the authors more directly measure the affinity of TAPBPR for MHC I? Even if this is not easily achievable, I think the authors should reconsider their claims for kinetic instability, since deconvolution of kinetics vs. thermodynamics is probably impossible by this particular experiment.

This is indeed an important point. Directly determining the affinity of TAPBPR for peptide-deficient MHC I, as a measure of chaperone activity, is very difficult, as peptide-deficient MHCs are intrinsically unstable and will precipitate. We tried to determine Kd values via microscale thermophoresis using MHC I loaded with a photo-cleavable peptide; however, neither T-jump nor thermophoresis delivered a useful signal upon complex formation. In addition, we also aimed to determine kofffor the complexes using fluorescently-labeled TAPBPRs, but the scoop loop mutants turned out to be unstable upon fluorophore attachment. This is why we resorted to SEC analysis using unlabeled proteins. We cannot exclude different equilibrium dissociation constants for the complexes. However, our re-injection experiments (Figure 2—figure supplement 1; SEC traces for the two TAPBPR scoop-loop mutants have now been added) clearly show that the three complexes have different kinetic stabilities. Furthermore, we have repeated the peptide displacement experiments at a much lower TAPBPR concentration (75 nM) (new Figure 3—figure supplement 1) and observed gradual activity differences between the three TAPBPR constructs that were similar to the previous measurements at 1 µM, suggesting that the TAPBPRs have affinities for (suboptimally-loaded, and most likely empty) MHC I that are at least in the same range.

2) Given the result presented in Figure 2, how do the authors explain the finding that scoop loop variants largely retain their ability to catalyse peptide dissociation (Figure 3)? Is it because the scoop loop interactions with MHC are secondary to the overall TAPBPR-MHC interactions? If the scoop loop variants have a reduced affinity for MHC, would an effect be more obvious if experiments in Figure 3 are repeated at lower concentrations?

Initially, we chose the high TAPBPR concentrations on purpose to test if the mutants are still active in catalyzing peptide displacement. We have now repeated the experiment at a much lower TAPBPR concentration (75 nM) (new Figure 3—figure supplement 1) and observed the same relative catalytic activities between the three TAPBPR constructs as before (at 1 µM), suggesting that the TAPBPRs have similar affinities for (suboptimally-loaded, and most likely empty) MHC I. Then, why are the different complex stabilities not reflected in the one-step peptide-displacement assay? We think there might be two reasons: First, the chaperone activity, i.e. the ability to stabilize peptide-free MHC I for a prolonged period of time, does not come into effect in our displacement assay, as the assay is carried out in the presence of a 1000-fold molar excess of high-affinity competitor peptide. Secondly, the scoop loop does not appear to be of primary importance for the displacement step in peptide exchange. Other TAPBPR-mediated, scoop loop-independent changes might be more important for induced peptide dissociation (widening of MHC binding groove, salt bridge to Y84, distortion of groove floor, shift of β2m).

3) Relative to a TAPBPR chimeric construct containing the shorter scoop loop of tapasin, the wild type TAPBPR is shown to form more stable (higher affinity complexes) with MHC I (Figure 2), although the peptide dissociation function of the wild type TAPBPR is less efficient than the chimera with the tapasin scoop loop (Figure 3). The stronger peptide dissociation effect of a TAPBPR with the tapasin scoop loop compared with TAPBPRwt should be interpreted, explicitly discussed, and taken into account in the overall model of Figure 5.

As explained in our answer to point 2, we do not think that the chaperone activity (Figure 2) can be directly correlated with the activity in our peptide-displacement assay (Figure 3). Currently, we do not have an explanation for the slightly higher activity of the TAPBPRTsn-SL variant. However, our aim of the displacement measurements was to demonstrate that the two scoop-loop variants are still generally able to catalyze peptide dissociation. Notably, the conditions of the assay (high concentration of competitor peptide) resemble the physiological environment tapasin is operating in, i.e. the peptide-rich environment of the PLC. The tapasin scoop loop might thus be expected to operate optimally under these experimental conditions.

4) Notably, mutations of other key TABBPR residues at the TAPBPR-MHC class I interface (TN3 and TN6) are shown in the representative figure to have more significant effects on peptide dissociation than a TAPBPR mutant in which the scoop loop was mutated to a 3xGly sequence, suggesting a dominant role for TN3 and TN6-mediated interactions in inducing peptide dissociation. The dominant role of TN3 and TN6-mediated interactions in inducing peptide dissociation should be acknowledged (including within the model), and their quantifications shown in the same way as Figure 3C and 3E.

It is beyond debate that the Y84-coordinating glutamate in TAPBPR (mutated in the TN3 mutant) and residues mutated in TN6 are crucial for TAPBPR-catalyzed peptide displacement. Catalysis of peptide displacement by TAPBPR involves several induced changes in the MHC I, including a stabilized open conformation of the peptide binding groove, distortion of the groove floor, and displacement of β2m. These changes are mediated by parts of the TAPBPR molecule that are outside the scoop loop and comprise the greater part of the TAPBPR-MHC I interface. Especially, the TN6 mutations affect the core of the TAPBPR-MHC I interface and are therefore expected to drastically reduce displacement activity. We definitely do not claim that the scoop loop is more important for peptide displacement than the aforementioned aspects of catalysis. In the current manuscript we focus solely on the role of the scoop loop, and the main purpose of the experiments shown in Figure 3 was to demonstrate that our two scoop-loop variants are still able to catalyze peptide dissociation.

We have now included bar diagrams for the TN3 and TN6 measurements in Figure 3C.

5) In a variant of the peptide dissociation experiment, peptide dissociation was measured in the presence of the three TAPBPR variants, but in the absence of excess unlabeled peptide (Figure 4 and Figure 4—figure supplement 1). Under this set of conditions, wild type TAPBPR was more efficient at inducing peptide dissociation from H2-Db than the two scoop loop mutants, whereas wild type TAPBPR and the tapasin scoop loop mutant appear to have similar efficiencies towards peptide dissociation from HLA-A2. These data are used to argue for a model in which the scoop loop directly interferes with rebinding of a displaced peptide (internal competitor). However, concerns with these interpretations are:

a) Other models (such as different allosteric effects induced by the three scoop loop variants of TAPBPR) could explain the varying degrees of inhibition of rebinding of dissociated peptides by the TAPBPR variants, which are also predicted to have different affinities for MHC class I. Model validation will need data that can more clearly sort out the difference between an internal competitor model and alternative allosteric effects-based models. If the differences cannot be easily sorted out, at minimum, the model of Figure 5 should be changed to indicate that both the internal competitor model and allosteric model are likely to be relevant.

We have modified the manuscript and the text of Figure 5 to include the possibility of allosteric effects.

b) The effects of the three TAPBPR constructs on inhibition of peptide rebinding vary between HLA-A2 and H2-Db (comparing Figure 4 with Figure 4—figure supplement 1). The variable effects of tapasin scoop loop on Db vs. HLA-A2 should be expanded, interpreted, and explicitly discussed.

We now explicitly describe the observed allomorph-specific activity and included a possible interpretation of this phenomenon.

6) Does the experiment in Figure 4 report a reduced "functional" ability of the scoop loop mutants to remove the peptide from MHC or a reduced affinity of the TAPBPR variants for MHC? This difference is relevant to our understanding of the distinct functionality of the scoop loop (versus being a hinge that makes additional interactions with the MHC). Perhaps the authors could test this by titrating the less active TAPBRPTsn-SL to higher concentrations to see if they can get similar results to the TAPBPRwt-SL.

We clearly show that the differences between wildtype TAPBPR and the scoop loop mutants persist after saturation (see titrations with increasing TAPBPR concentrations in Figure 4C and D, and the saturating concentration in Figure 4E). This demonstrates that the differences are not due to reduced affinities of the mutants.

7) The importance of the scoop loop has recently been demonstrated in the study by Ilca et al. (2018), but the latter paper is only referenced to highlight the differences in conclusions, rather than many other overall similarities in findings. The Ilca et al. paper should be referenced within the Introduction for better acknowledgement of prior contributions to the same questions, and similarities with the present study pointed out in addition to the differences-within the Discussion.

We now reference the Ilca et al. paper in the Introduction and contrast its cell-based experimental approach to our in vitro approach using defined, purified protein components.

8) Finally, for readers and for the field, this section in the Introduction is very confusing "Furthermore, one of the two TAPBPR-MHC I complex structures revealed a remarkable structural feature in TAPBPR named the scoop loop (Thomas and Tampe, 2017). […] The scoop loop occupies a position that is incompatible with high-affinity peptide binding and displaces or coordinates several key MHC I residues, including Y84, T143, K146, and W147, which are responsible for binding the C terminus of the peptide." The authors should rephrase to clarify what has been discovered regarding the scoop loop previously, what is in debate, and specify the open questions that this study will address.

We have simplified and rephrased the paragraph to describe the current state of knowledge and open questions.

9) As noted above, model validation will need data that can more clearly sort out the difference between an internal competitor model and alternative allosteric effects-based models. Additional higher resolution structural data, if possible to acquire within two months, may be helpful because the loop was modeled as inserting into the F-pocket in one of the structures (Thomas and Tampe, 2017) but not in the second (Jiang et al., 2017).

In contrast to our TAPBPR-H2-Db structure (new Figure 1B showing electron density of scoop loop, contoured at 0.8 σ), the TAPBPR-H2-Dd complex (Jiang et al., 2017) does indeed lack density for the scoop loop. This is most likely due to the presence of a disulfide-linked 5-mer peptide in the binding groove of the H2-Dd construct used by Jiang et al., which interferes with stable scoop-loop binding. (Electron density for this covalently bound 5-mer peptide is also lacking, probably due to mobility). Higher-resolution crystallographic data might provide more detailed insights into the binding mode of the scoop loop, but the given time frame for revision (2 months) is definitely too short to obtain these data. It has taken us years of very focused efforts and numerous crystals to obtain the TAPBPR-H2-Db structure at the current resolution.

Overall, this is a well-executed study, but the data do not integrate well with the proposed model. In line with the above points, the data are more consistent with a relatively minor (and possibly MHC-context dependent) contribution of the scoop loop to peptide editing, rather than the current emphasis of a critical role. This message should be evident within the Abstract and throughout the manuscript.

We have slightly toned down our statements about the role of the scoop loop in peptide editing. However, we still think that the observed effects on chaperone activity and peptide rebinding are striking.

https://doi.org/10.7554/eLife.55326.sa2

Article and author information

Author details

  1. Lina Sagert

    Institute of Biochemistry, Biocenter, Goethe University Frankfurt, Frankfurt, Germany
    Contribution
    Data curation, Formal analysis, Visualization, Methodology, Writing - original draft
    Competing interests
    No competing interests declared
  2. Felix Hennig

    Institute of Biochemistry, Biocenter, Goethe University Frankfurt, Frankfurt, Germany
    Contribution
    Data curation, Formal analysis
    Competing interests
    No competing interests declared
  3. Christoph Thomas

    Institute of Biochemistry, Biocenter, Goethe University Frankfurt, Frankfurt, Germany
    Contribution
    Conceptualization, Data curation, Formal analysis, Supervision, Validation, Visualization, Writing - original draft, Writing - review and editing
    For correspondence
    c.thomas@em.uni-frankfurt.de
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-7441-1089
  4. Robert Tampé

    Institute of Biochemistry, Biocenter, Goethe University Frankfurt, Frankfurt, Germany
    Contribution
    Conceptualization, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Writing - original draft, Project administration, Writing - review and editing
    For correspondence
    tampe@em.uni-frankfurt.de
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-0403-2160

Funding

European Research Council (ERC_AdG 789121)

  • Robert Tampé

Deutsche Forschungsgemeinschaft (TA 157/12-1)

  • Robert Tampé

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Christian Winter for help with peptide synthesis and analytical SEC. We thank Dr. Rupert Abele, Dr. Simon Trowitzsch, Andrea Pott, Inga Nold, and all members of the Institute for Biochemistry for discussion and comments. The support by the European Research Council (ERC Advanced Grant 789121 to RT) and the German Research Foundation (Reinhart Koselleck Project TA 157/12–1 to RT) is gratefully acknowledged.

Senior Editor

  1. Tadatsugu Taniguchi, Institute of Industrial Science, The University of Tokyo, Japan

Reviewing Editor

  1. Pamela J Bjorkman, California Institute of Technology, United States

Reviewers

  1. Malini Raghavan, University of Michigan, United States
  2. Scheherazade Sadegh-Nasseri, Johns Hopkins University School of Medicine, United States
  3. Efstratios Stratikos

Publication history

  1. Received: January 20, 2020
  2. Accepted: March 12, 2020
  3. Accepted Manuscript published: March 13, 2020 (version 1)
  4. Version of Record published: April 2, 2020 (version 2)

Copyright

© 2020, Sagert et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 847
    Page views
  • 156
    Downloads
  • 2
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Biochemistry and Chemical Biology
    2. Structural Biology and Molecular Biophysics
    Roman O Fedoryshchak et al.
    Research Article

    PPP-family phosphatases such as PP1 have little intrinsic specificity. Cofactors can target PP1 to substrates or subcellular locations, but it remains unclear how they might confer sequence-specificity on PP1. The cytoskeletal regulator Phactr1 is a neuronally-enriched PP1 cofactor that is controlled by G-actin. Structural analysis showed that Phactr1 binding remodels PP1's hydrophobic groove, creating a new composite surface adjacent to the catalytic site. Using phosphoproteomics, we identified mouse fibroblast and neuronal Phactr1/PP1 substrates, which include cytoskeletal components and regulators. We determined high-resolution structures of Phactr1/PP1 bound to the dephosphorylated forms of its substrates IRSp53 and spectrin aII. Inversion of the phosphate in these holoenzyme-product complexes supports the proposed PPP-family catalytic mechanism. Substrate sequences C-terminal to the dephosphorylation site make intimate contacts with the composite Phactr1/PP1 surface, which are required for efficient dephosphorylation. Sequence specificity explains why Phactr1/PP1 exhibits orders-of-magnitude enhanced reactivity towards its substrates, compared to apo-PP1 or other PP1 holoenzymes.

    1. Biochemistry and Chemical Biology
    2. Microbiology and Infectious Disease
    Kuan-Yi Lu et al.
    Research Article

    Phosphatidylinositol 3-phosphate (PI(3)P) levels in Plasmodium falciparum correlate with tolerance to cellular stresses caused by artemisinin and environmental factors. However, PI(3)P function during the Plasmodium stress response was unknown. Here, we used PI3K inhibitors and antimalarial agents to examine the importance of PI(3)P under thermal conditions recapitulating malarial fever. Live cell microscopy using chemical and genetic reporters revealed that PI(3)P stabilizes the digestive vacuole (DV) under heat stress. We demonstrate that heat-induced DV destabilization in PI(3)P-deficient P. falciparum precedes cell death and is reversible after withdrawal of the stress condition and the PI3K inhibitor. A chemoproteomic approach identified PfHsp70-1 as a PI(3)P-binding protein. An Hsp70 inhibitor and knockdown of PfHsp70-1 phenocopy PI(3)P-deficient parasites under heat shock. Furthermore, PfHsp70-1 downregulation hypersensitizes parasites to heat shock and PI3K inhibitors. Our findings underscore a mechanistic link between PI(3)P and PfHsp70-1 and present a novel PI(3)P function in DV stabilization during heat stress.