Butyrophilin-2A1 Directly Binds Germline-Encoded Regions of the Vγ9Vδ2 TCR and Is Essential for Phosphoantigen Sensing

Abstract

Vγ9Vδ2 T cells respond in a TCR-dependent fashion to both microbial and host-derived pyrophosphate compounds (phosphoantigens, or P-Ag). Butyrophilin-3A1 (BTN3A1), a protein structurally related to the B7 family of costimulatory molecules, is necessary but insufficient for this process. We performed radiation hybrid screens to uncover direct TCR ligands and cofactors that potentiate BTN3A1's P-Ag sensing function. These experiments identified butyrophilin-2A1 (BTN2A1) as essential to Vγ9Vδ2 T cell recognition. BTN2A1 synergised with BTN3A1 in sensitizing P-Ag-exposed cells for Vγ9Vδ2 TCR-mediated responses. Surface plasmon resonance experiments established Vγ9Vδ2 TCRs used germline-encoded Vγ9 regions to directly bind the BTN2A1 CFG-IgV domain surface. Notably, somatically recombined CDR3 loops implicated in P-Ag recognition were uninvolved. Immunoprecipitations demonstrated close cell-surface BTN2A1-BTN3A1 association independent of P-Ag stimulation. Thus, BTN2A1 is a BTN3A1-linked co-factor critical to Vγ9Vδ2 TCR recognition. Furthermore, these results suggest a composite-ligand model of P-Ag sensing wherein the Vγ9Vδ2 TCR directly interacts with both BTN2A1 and an additional ligand recognized in a CDR3-dependent manner.

Keywords: T cell receptor; Vγ9Vδ2; butyrophilin; complementarity-determining region; gamma delta T cell; ligand; phosphoantigen.

Graphical abstract

Figure 1

Identification of BTN2A1 as Factor X (A) Radiation hybrid approach to generate and identify rodent cell-fusion hybrids incorporating portions of human chromosome (Chr) 6 that permit P-Ag sensitization. (B) RNA-seq analysis of prioritized clones generated from fusion with A23 or BW cells. Values for less than three transcripts are merged with the x axis. (C) Arrangement of BTN gene cluster on Chr 6 extracted from genome data viewer GRCh38.p13 (GCF_000001405.39). (D) Production of IFNγ from polyclonal Vγ9Vδ2 T cell lines in response to Zol-treated WT or BTN2^−/− 293T cells. Error bars represent standard deviation for three independent experiments. ^∗∗p < 0.005. (E) Production of IL-2 from TCR-MOP transductants in response to HMBPP-treated WT, BTN2^−/−, BTN2A1^−/−, and BTN2A2^−/− 293T cells. (F) Production of IL-2 from TCR-MOP transductants in response to 20.1 mAb-treated WT, BTN2^−/−, BTN2A1^−/−, and BTN2A2^−/− 293T cells. In (E) and (F), the different colors indicate results from two independent experiments. See also Figure S1.

Figure 2

BTN2A1 and BTN3A1 Synergize to Potentiate P-Ag Sensing in Rodent Cells (A) Expression of BTN2A1, BTN3A1, or both genes in transduced BW cells. (B) Production of IL-2 by TCR-MOP transductants in response to HMBPP-treated CD80⁺ BW cells transduced to express BTN2A1, BTN3A1, both, or untransduced controls. Percentage activation is normalized against the maximum response obtained from CD80⁺ CHO cells expressing both BTN2A1 and BTN3A1 in the presence of 10 μM HMBPP. (C) Production of IL-2 from TCR-MOP transductants in response to HMBPP-treated CD80⁺ CHO cells transduced with either BTN2A1, BTN3A1, both genes, or untransduced controls, with responses normalized as in (B). Error bars in (B) and (C) represent standard deviation for three independent experiments. Differences between untransduced and BTN2-transduced cells were significant (p < 0.05), as were those between the BTN2A1-transductant and BTN2A1+BTN3A1-transductant in the presence of HMBPP. (D) MOP-TCR tetramer staining of transduced BW cells. (E) Staining of transduced 293T cells with Vγ9Vδ2 TCRs. (F) MOP-TCR tetramer staining or anti-BTN2A1 mAb staining of BTN2A1 and BTN3A1-transduced CD80⁺ BW cells versus untransduced controls in the presence and absence of Zol. See also Figure S2.

Figure 3

Direct BTN2A1 Binding to Germline-Encoded Regions of Vγ9 Is Essential for P-Ag Sensing (A) (Top panel) Injection of BTN2A1 IgV (25 μM) over surfaces with immobilized Vγ9Vδ2 TCR (2,457 resonance units (RU)) and control surfaces comprising Vγ4Vδ5 TCR (2,351 RU), Vγ2Vδ1 TCR (1,800 RU), or streptavidin alone. Notably, signals over streptavidin alone and control TCR surfaces are equivalent. (Bottom panel) Injection of BTN2A1 IgV (24 μM) over surfaces with immobilized G115 (Vγ9Vδ2; 3,109 RU), MOP (Vγ9Vδ2; 3,108 RU), and Vγ9Vδ1 (2,774 RU) TCRs and LES TCR control (Vγ4Vδ5; 2,885 RU). (B) Equilibrium affinity measurements and Scatchard analysis (inset) of BTN2A1 IgV binding to the G115 (K_d = 39.5 μM) and MOP (K_d = 48.4 μM) Vγ9Vδ2 TCRs and Vγ9Vδ1TCR (K_d = 47.9 μM). Data in (A) and (B) are representative of eight to nine independent experiments. (C) Model of the BTN2A1-Vγ9 interaction mode based on the proposed BTNL3-Vγ4 interaction, with expanded panel showing potential contacts at the Vγ9-BTN2A1 IgV interface. (D) Effects of seven alanine substitutions in proposed BTN2A1 interface residues on Vγ9Vδ2 TCR interaction, indicating affinity of mutant BTN2A1 relative to WT BTN2A1 calculated in the same experiment. Data shown are representative of two independent experiments. (E) Effects of BTN2A1 R124A and R124E mutations on IL-2 production by TCR-MOP in response to HMBPP-treated BTN3A1 and BTN2A1 expressing CD80⁺ CHO cells. (F) Predicted involvement of Vγ9 HV4 and CDR2 residues in BTN2A1 interaction. (G) Effects of Vγ9-E70 mutation (HV4) on IL-2 production by TCR-MOP in response to HMBPP-treated BTN3A1 and BTN2A1 expressing CD80⁺ CHO cells. (H) Effects of mutations in Vδ2 CDR2 (R51A) or a deletion in CDR3 (ΔCDR3) on IL-2 production by TCR-MOP in response to HMBPP-treated BTN3A1 and BTN2A1 expressing CD80⁺ CHO cells. In (E), (G) and (H), error bars indicate standard deviation for three independent stimulation experiments. Percentage activation is normalized against the maximum response obtained from CHO cells expressing both BTN2A1 and BTN3A1 in the presence of 1 μM HMBPP. Differences between WT and mutants in (E), (G), and (H) were significant, except for TCR-MOP E70A at 1 μM. See also Figure S3.

Figure 4

BTN2A1 Forms Disulphide-Linked Homodimers at the Cell Surface (A) Homology model of BTN2A1 homodimer. (B) C-terminal region of BTN2A1 homology model indicating close proximity of Cys residues. (C) Non-reducing (NR) or reducing (R) SDS-PAGE analysis of CHO-cell expressed BTN2A1 and BTN2A2 protein. (D) Effects of BTN2A1C247W mutation on IL-2 production by TCR-MOP in response to HMBPP-treated CD80⁺ CHO cells expressing BTN2A1 and BTN3A1. Error bars indicate standard deviation for three independent stimulation experiments. Percentage activation is normalized against the maximum response obtained from CHO cells expressing both BTN2A1 and BTN3A1 in the presence of 1 μM HMBPP. (E) Model of Vγ9Vδ2-BTN2A1 interaction incorporating BTN2A1 homodimer formation, and bilateral Vγ9Vδ2 interaction with BTN2A1 IgV domain. See also Figure S4.

Figure 5

Cell-Surface Association of BTN2A1 and BTN3A1 Proteins (A) Anti-BTN2A1-HA immunoprecipitation, combined with anti-BTN3A1-FLAG western blot detection, following cell-surface cross-linking of CHO cells expressing BTN2A1-HA, FLAG-BTN3A1, or both. (B) Anti-BTN3A1 IP (20.1 mAb) of the same lysate combined with anti-BTN3A1-FLAG detection. For (A) and (B), likely monomeric or oligomeric species corresponding to appropriate molecular weight bands are indicated on the right-hand side. Data are representative of four independent experiments. (C) NMR chemical-shift perturbations (CSPs) in selected residues in ¹H-¹⁵N-labeled BTN3A1 IgV (100 μM) following addition of BTN2A1 IgV (100 μM). (D) Graph of chemical shift versus residue number in BTN3A1 IgV. Threshold levels for significant CSPs are indicated by horizontal lines. (E) Mapping of residues whose amide resonances undergo CSPs on the surface of BTN3A1 IgV domain, showing clustering in the CFG face of the domain. Residues are colored in relation to the size of their CSPs, using the thresholds indicated in (D). See also Figure S5.