Direct recognition of an intact foreign protein by an αβ T cell receptor

Abstract

αβ T cell receptors (αβTCRs) co-recognise antigens when bound to Major Histocompatibility Complex (MHC) or MHC class I-like molecules. Additionally, some αβTCRs can bind non-MHC molecules, but how much intact antigen reactivities are achieved remains unknown. Here, we identify an αβ T cell clone that directly recognises the intact foreign protein, R-phycoerythrin (PE), a multimeric (αβ)₆γ protein complex. This direct αβTCR-PE interaction occurs in an MHC-independent manner, yet triggers T cell activation and bound PE with an affinity comparable to αβTCR-peptide-MHC interactions. The crystal structure reveals how six αβTCR molecules simultaneously engage the PE hexamer, mediated by the complementarity-determining regions (CDRs) of the αβTCR. Here, the αβTCR mainly binds to two α-helices of the globin fold in the PE α-subunit, which is analogous to the antigen-binding platform of the MHC molecule. Using retrogenic mice expressing this TCR, we show that it supports intrathymic T cell development, maturation, and exit into the periphery as mature CD4/CD8 double negative (DN) T cells with TCR-mediated functional capacity. Accordingly, we show how an αβTCR can recognise an intact foreign protein in an antibody-like manner.

Fig. 1. PE binds to 4D4 TCR⁺ cells.

A BW58 cells were transduced with the 4D4 TCR sequence which was identified amongst single cell BALB/c Jα18^−/− thymocytes sorted as CD1d-α-GlcADAG-tetramer-(SAV-PE)⁺ CD1d-α-GalCer-tetramer (SAV-BV421)^- (as per Almeida et al, 2019) and assessed for binding to α-GlcADAG-, α-GalCer-, sulfatide-loaded or unloaded CD1d tetramers (SAV-PE conjugated). Flow cytometry plots show TCRβ versus GFP expression on top and mean fluorescence intensity (MFI) of SAV-PE conjugated tetramers on TCRβ⁺GFP⁺ gated cells. CD1d-α-GlcADAG-reactive (A11B8.2 and A10B8.2, CD1d-sulfatide-reactive (XV19) and CD1d-α-GalCer-reactive VB8-STD TCR-expressing clones were included as controls. B Graded amounts of plate-bound CD1d loaded with α-GlcADAG, α-GalCer or vehicle (Veh) were assessed for their ability to activate the 4D4 clone compared to A11B8.2, A10B8.2 or VB8-STD control lines. Plots show flow cytometric detection of CD69 upregulation (mean of 2 experiments) or IL-2 secretion using a capture bead assay after 16 h. C α-GalCer-loaded CD1d was tetramerised using SAV-PE (top plot) or SAV-BV421 (bottom plot) and assessed for staining of the BW58 lines expressing the 4D4 TCR (red) or control A10B8.2 TCR (blue) by flow cytometry. D SAV-PE, SAV-PE-CY7, SAV-APC, SAV-APC-CY7, SAV-BV421, and SAV-PB (all from Becton Dickinson) were assessed for their ability to stain the 4D4 (red) or the control V1168 NKT (blue) TCR+ lines. E Two distinct isoforms of non-conjugated PE (from Prozyme and from Thermo Fisher Scientific) were assessed for their ability to stain 4D4 (red) or control A10B8.2 (blue) TCR+ cell lines, and analysed in a 10% SDS-PAGE gel (top). The ability of different PE conjugates (from the indicated suppliers) to stain the same cell lines was investigated by flow cytometry (bottom). F Anti-PE mAb (50µg/ml) or G soluble 4D4 TCR, and the control A11B8.2 NKT TCR or CD1a-restricted BK6 TCR were pre-incubated with SAV-PE or PE alone and assessed for impact on staining by FACS of the 4D4 cell line. Data in Fig. 1A–G are representative of 2 independent experiments, except for the gel in E, which was one of two runs (reducing and non-reducing) from one experiment, and the BK6 control in G which was used in one experiment. Source data are provided as a Source Data file.

Fig. 2. PE activates cells expressing the 4D4 TCR.

A Plate bound or soluble versions of non-conjugated PE (from Prozyme) and various forms of SAV-conjugated PE (from different suppliers) at 50 or 100 μg/mL—respectively—were assessed for their ability to activate a BW58 cell line expressing the 4D4 TCR or a control NKT TCR (VII68). Anti-CD3 (10 μg/mL) mAb was included as a positive control for stimulation. After 16 h the supernatants were assayed for the presence of IL-2 by cytometric bead array. B Titrating amounts of plate bound SAV-PE, SAV-APC, SAV-FITC and non-conjugated PE were assessed for their ability to elicit IL-2 production by the 4D4 or the control A11B8.2 NKT TCR line. In order to investigate if PE induced activation could be prevented by a neutralising antibody or protein denaturation, the top concentration of each sample was pre-incubated with anti-PE mAb (20µg/ml), or incubated at 90 °C for 2 min. Data in A and B is representative of 2 independent experiments except for the heat-inactivated samples and the SAV-FITC samples in B which were one experiment. C After 16 h of stimulation with plate bound molecules as indicated, 4D4 (grey bar) or VII68 (white bar) cells were stained for CD44, CD69 CD25 or TCRβ surface expression. Mean Fluorescence Intensity fold variation relative to no activation, error bars represent ± SEM for each marker across 3 independent experiments, each represented by the individual dots. Single-molecule imaging of TCRβ (D), or pCD3ζ (E), clustering in BW58 thymoma cells transduced with the 4D4, 2C12 or VII68 TCRs following stimulation on a supported lipid bilayer decorated with either ICAM-1 only, ICAM-1 and SAV-PE or ICAM-1 and CD1d-α-GalCer. Inset bright-field images show thymoma cells used for single-molecule imaging. Scale bar: 5 µm. F Single-molecule data were analysed using DBSCAN across three experiments with a total number of cells (shown as dots) TCRβ /pCD3ζ for ICAM-1; 4D4 (40/42), 2C12 (40/42), VII68 (40/42), for PE; 4D4 (40/44), 2C12 (42/42), VII68 (42/40) and for CD1d-α-GalCer; 4D4 (40/45), 2C12 (39/44), VII68 (40/40). Data were expressed as mean ± SEM. One-way ANOVA was used for comparing stimulation against an ICAM-1 control. ns = not significant. *P  ≤  0.05; **P  ≤  0.01; ***P  ≤  0.001; ****P  ≤  0.0001 (exact P values shown in Supplementary Table 1). Source data provided as a Source Data File.

Fig. 3. Biophysical analysis of the 4D4 interaction with PE.

A Size distribution plot derived from sedimentation velocity analysis in an analytical ultracentrifuge (SV-AUC). Data is shown for soluble 4D4 TCR alone, PE alone or a mixture of both at a 10:1 (TCR:PE) molar ratio. Sedimentation coefficient (S) was determined and used to calculate the molecular weight (MW) of each component in solution. The estimated MW of the complex was compared to the MW of PE alone and used to estimate the molar binding ratio of TCR:PE (n), as shown in the table below. B On a separate SV-AUC experiment 4D4 TCR was incubated at different ratios with a fixed amount of PE (3:1-6:1 TCR:PE molar ratio), results are presented as in A. In the zoomed region of the graph, numbers on top of peaks represent S for each sample. A, B are representative of 2 independent SV-AUC experiments. C Electrophoresis mobility shift assay: 4D4 TCR was incubated at different ratios (1:1–7:1) with a fixed amount of PE and complex formation investigated in a 7.5% native PAGE gel. On separate wells PE or 4D4 were ran alone. Data is representative of 2 independent assays. D Representative Isothermal titration calorimetry trace (upper) and binding isotherm (lower) following serial injections of soluble 4D4 TCR (17.7 μM) into a PE solution (Prozyme; 660 μM). The addition of buffer alone (PBS pH 8) to PE and of PE to buffer alone were subtracted from the values presented to exclude buffer induced heat fluctuations. Data is representative of two independent experiments. The dissociation constant (K_D), stoichiometry of binding (n), enthalpy (ΔH) and entropy (ΔS) for the binding reaction were estimated for each experiment and averaged, as shown in the table below. E Surface plasmon resonance of the 4D4 TCR and 2C12 type I NKT TCR binding to immobilised PE, mouse (m)CD1d-α-GalCer, I-A^b-CLIP, H2-D^b-NP₃₆₆ and mMR1-5-OP-RU. Data are shown for two independent SPR experiments, each colour coded as orange and blue, with the equilibrium binding curves plotting the mean derived from two SPR runs per experiment, used to estimate the dissociation constant (K_D). N.D. = not determined. Source data are provided as a Source Data file.

Fig. 4. Overall architecture of the 4D4 TCR-PE complex and recognition mode.

A Cartoon representation of the symmetry-expanded 4D4 TCR–PE complex, showing the 4D4 TCR (grey) interface recognising a symmetry-related epitope upon the hetero-hexameric PE assembly (red). B The asymmetric unit of the 4D4 TCR–PE complex (4D4 TCR and PE α- and β-chains shown, light grey, dark grey, dark red, and pink, respectively).

Fig. 5. Molecular interactions at the 4D4 TCR-PE interface.

A Complementarity determining region (CDR)-mediated contacts from the 4D4 TCR to PE and the molecular contacts therein (B). The CDR loops are coloured as follows: CDR1α yellow, CDR2α orange, CDR3α wheat, CDR1β light blue, CDR2β light green, CDR3β dark blue on the PE α-chain show as grey cartoon, with the Vα and Vβ centres of mass (COM) shown as grey and black spheres, respectively. The molecular surface is coloured according to the respective CDR-mediated contacts. The 4D4 TCR α-chain interactions from the CDR1α and CDR3α are shown in (C, D). The 4D4 TCR β-chain interactions from the CDR3β are shown in (E). Hydrogen and Van der Waals interactions are shown in black with salt bridge contacts shown as yellow dashed lines.

Fig. 6. Structural homology of the PE paratope recognised by the 4D4 TCR.

Cartoon representation and alignment of the PE α-chain paratope (in red) recognised by the 4D4 TCR to; A MHC-I, B MR1, C CD1a, D MICA, E T10, F T22, G M157, H Fc neonatal receptor, I HFE, J Haemoglobin, K Neuroglobin, and L Hell’s gate globin.

Fig. 7. 4D4 TCR⁺ T cells develop in vivo in a retrogenic mouse model.

A Flow cytometry analysis of bone marrow (BM), thymus (Thy), spleen (Sp), lymph node (LN), liver, lung, and blood of 4D4 retrogenic mice (Rg) or wild type (WT) C57BL/6. Representative plots of GFP and TCRβ expression on 7AAD⁻ CD19⁻ CD11b⁻ cells amongst CD45.2⁺ (from donor transduced BM) or CD45.1⁺ (recipient) in Rg mice or CD45.2⁺ from WT mice. B CD4 and CD8 co-receptor expression and C SAV-PE binding among CD45.2⁺ TCRβ ⁺ GFP⁺ Rg cells in comparison to CD45.1⁺ TCRβ⁺ GFP⁻ cells from the same Rg mice, or CD45.2⁺ TCRβ⁺ GFP⁻ cells from WT mice. Data in (B) are from one Rg mouse, representative of n = 12 Rg mice (acquired over 5 independent experiments for thymus and spleen, or 3 experiments for blood), n = 10 Rg mice for liver and BM (4 experiments), n = 7 Rg mice for lung and LN (3 experiments); and from one WT mouse representative of n = 7 WT mice (4 experiments) for spleen and thymus, n = 6 WT mice for lung, LN and BM (3 experiments), n = 4 WT mice for liver (3 experiments). Data in (C) are from one Rg and from one WT mice representative of 2 experiments with n = 2 or n = 3 Rg and WT organs, with the exception of the liver which is from 1 experiment with n = 3 Rg or n = 3 WT mice. Graphs showing assay variation for (B) and (C) are in Supplementary Fig. 10A and 10B. D Paired TCR analysis of single cells sorted as TCRβ⁺ GFP⁺ or TCRβ⁺ GFP⁻ from the spleen of a retrogenic mouse. Nomenclature as per the IMGT database. Source data are provided as a Source Data file.