γ9δ2 T cells detect mevalonate diphosphate via BTN3A3

Butyrophilins (BTNs) act as critical sensors of phosphoantigens (pAgs), mediating the immune response to these small molecules. The natural ligands for BTN3A1 include the exogenous (E)-4-hydroxy-3-methyl-but-2-enyl diphosphate (HMBPP) and endogenous pAgs dimethyl allyl diphosphate (DMAPP) and isopentenyl diphosphate (IPP), all of which can bind to BTN3A1 followed by activation of γ9δ2 T cells. BTN3A1 shares similar structural domains and exhibits high amino acid sequence homology with BTN3A3. However, the potential functions of BTN3A3 and whether it possesses ligands of its own remain unknown. Here, we show that a single point mutation to BTN3A3 allows it to detect HMBPP, suggesting that the protein may have its own ligand. Mevalonate diphosphate (MPP), a naturally occurring endogenous compound, exhibits high docking scores with both BTN3A1 and BTN3A3. Functional testing revealed that MPP activates γ9δ2 T cells, albeit with a higher EC₅₀ (300 µM) compared to DMAPP (7.1 µM) and IPP (30 µM). ³¹P NMR experiments confirmed MPP binding to both BTN3A1 and BTN3A3 internal domains while exposing the ability of BTN2A1 to form a complex with BTN3A3-MPP. 6-Fluoromevalonate (6-FM), an inhibitor targeting MPP decarboxylase, also activated γ9δ2 T cells. LC-MS showed that 6-FM-treated CHO-K1 cells accumulated substantial MPP while risedronate, an inhibitor of farnesyl diphosphate synthase, primarily increased DMAPP/IPP. Interestingly, in coculture experiments, transfection of CHO-K1 cells with either BTN2A1 + BTN3A1 or BTN2A1 + BTN3A3 enabled their response to 6-FM. Both BTN2A1 + BTN3A3-transfected CHO-K1 cells and BTN3A1 knockout K562 cells could detect 6-FM, despite their absence of BTN3A1. Our findings provide evidence that MPP functions as a natural ligand of BTN3A3 and establish a foundation for elucidating biological functions of BTN3A3 complementary to the established BTN3A1 pathway.