Absence of mucosal-associated invariant T cells in a person with a homozygous point mutation in MR1

Whole exome sequencing of a patient revealed a homozygous mutation MR1

A 31-year-old male patient was identified with unexplained primary immunodeficiency, characterised by tattoo-associated persistent human papilloma virus (HPV)⁺ warts (Fig. S1). The patient also had a history of immune-related complications. Namely, treatment-refractory Campylobacter gastroenteritis when he was 25 years old, and development of pneumonia secondary to varicella-zoster virus infection that caused lung scarring when he was an infant (see supplementary text for full case report). The patient was selected for whole exome sequencing when the cause for this immunodeficiency could not be clinically determined. Whole exome sequencing revealed that the patient was homozygous for two rare single nucleotide variants (SNV) with immune-related functions (Fig. 1A, see Table S1 and S2 for full list of SNV).

The first SNV rs35337543, resulting in a splice variant of IFIH1, has been previously characterised as loss-of-function mutation of the MDA5 protein (15), and further examination of this gene mutation in the patient can be found in the rs35337543 SNV report (see Supplementary text and Fig. S2). Therefore, we focused on the patient’s second homozygous immune-related SNV that is yet to be characterised, rs41268456 which is a G>A substitution at position 92 of the MR1 transcript (c.92G>A). It occurs in an exon region that results in an arginine to histidine substitution at position 31 of the amino acid sequence, corresponding to position 9 of the mature MR1 protein after the signal peptide is cleaved (R9H) (Fig S3A). R9H has a combined annotation dependent depletion (CADD) score of 9.12, a polyphen score of 0.92 with a prediction of probably damaging. This mutation resides within the ligand-binding pocket of the MR1 amino acid sequence (Fig. 1B), with R9 being a conserved residue across mammalian species (Fig. S3B). Structural studies have shown that R9 lies within the A’-pocket of MR1 and interacts directly with the ligand, but this interaction differs depending on the ligand bound. For example, R9 forms hydrogen bonds with the ribityl 3’-hydroxyl group and the uracil C-2 carbonyl oxygen of 5-OP-RU (Fig. 1C), while making hydrogen bonds to the ring carbonyl oxygen of 6-FP (4, 5, 16). The patient was born to nonconsanguineous parents who were both heterozygous for the MR1^R9H mutation, confirming an autosomal recessive inheritance pattern (Fig. 1B and D). The homozygous population frequency for rs41268456 is 0.000106 (15 individuals from a total cohort of 141 456), with a slightly higher frequency of 0.000203 in the European population, based on the Genome Aggregation Database (gnomAD) cohort (17) (Fig. 1E). However, there is no clinical information available for the 15 homozygous individuals in the gnomAD database and there are yet to be any scientific studies that report the biological implications of this mutation.

The MR1^R9H/R9H patient lacked circulating MAIT cells

To determine the functional implications of the MR1^R9H mutation, the frequency of MAIT cells in the peripheral blood of the MR1^R9H/R9H patient was examined and compared to healthy age and gender matched donors with wild-type (WT) MR1 (Fig. S3C). The MR1^R9H/R9H patient had no MAIT cells present based on MR1–5-OP-RU antigen-loaded tetramer staining or by staining with the MAIT cell surrogate markers: T cell receptor (TCR) Vα7.2 and CD161 (Fig. 2A). To confirm that the MR1^R9H/R9H patient lacked functional MR1–5-OP-RU-reactive MAIT cells, we stimulated peripheral blood mononuclear cells (PBMC) with the MAIT cell antigen 5-OP-RU (6). In contrast to the response of PBMC from healthy controls, we failed to detect any circulating T cells producing TNF from the patient (Fig. 2B). We also confirmed that the Vα7.2⁺ CD161⁻ population did not display any difference compared to healthy donors in their reactivity to stimulation (Fig. S5). We next examined the circulating αβ T cell populations from the MR1^R9H/R9H patient. He had a normal frequency of NKT cells but was mildly lymphopenic with reduced CD4 and CD8 T cell total cell counts and frequencies (Fig. S6A and Table S3). The MR1^R9H/R9H T cells were also mostly comprised of memory T cells and showed a robust effector response to stimulation (Fig. S6B–C). Thus, the MR1^R9H/R9H patient has a selective loss of MAIT cells, while all other T cell subsets remained intact.

The MR1^R9H/R9H patient’s antigen presenting cells were defective in antigen presentation

We next wanted to understand whether the lack of MAIT cells in the MR1^R9H/R9H patient was a result of his antigen presenting cells being unable to present MAIT cell activating ligands. To test this, we pulsed B cells with 5-OP-RU and then cultured them with T cells from an allogenic MR1^WT/WT donor. We found that MR1^R9H/R9H B cells had a diminished capacity to activate the allogenic MAIT cells compared to B cells from a MR1^WT/WT donor (Fig. 2C). Thus, the MR1^R9H mutation has a direct impact on MR1-restricted antigen presentation.

To directly examine the impact of the MR1^R9H mutation on ligand binding, we tested the capacity of MR1^R9H/R9H B cells to upregulate MR1 in response to 5-OP-RU and the non-stimulatory MR1 ligand Ac-6-FP, both of which induce WT MR1 upregulation (18). We observed that, although the MR1^R9H/R9H patient’s B cells could upregulate MR1 upon addition of Ac-6-FP, 5-OP-RU did not induce MR1 upregulation (Fig. 3A). This observation was confirmed by examining MR1 upregulation on C1R cells overexpressing either MR1^R9H or WT MR1 (Fig. 3B and C). Thus, the MR1^R9H mutation appears to selectively disrupt 5-OP-RU binding to MR1.

The MR1^R9H mutation caused structural changes to the A’-pocket of MR1

To understand how the MR1^R9H mutation impacts ligand binding, we expressed and refolded MR1^R9H in the presence of Ac-6-FP and 5-OP-RU, then crystallised their co-complexes with a MAIT cell TCR and subsequently solved the structures to high resolution (see Table S4 for all crystal structure supporting data). Surprisingly, MR1^R9H was refolded in the presence of 5-OP-RU but the A’-pocket was empty (Fig. 4A). WT MR1 requires binding of a ligand for stabilisation and correct folding (5). However, the MR1^R9H structure appears to not require the presence of a ligand to refold, similar to how the MR1^K43A mutant protein can be refolded without ligand (19). Indeed, MR1^R9H does refold in the absence of any exogenously added ligand (Fig. S7). Within the empty MR1^R9H A’-pocket, the imidazole side chain of the H9 residue formed new van der Waals interactions with the indole group of W69, causing W69 to flip and reorientate toward the base of the A’-pocket (Fig. 4A and Fig. S8B). In addition, water molecules found within the A’-pocket established a network of H-bonds with K43, H58, H9 and S24 residues. Together, these hydrophobic and polar interactions stabilised MR1^R9H enabling it to be refolded without ligand. Moreover, this reorientation of W69 and the presence of H9 partially fills the A’-pocket and sterically precludes 5-OP-RU binding (Fig. 4B). Interestingly however, MR1^R9H did refold and complex with Ac-6-FP, where the ligand formed a Schiff base with K43 of MR1 (Fig. 4C and S7D). Here, the reorientation of W69 does not interfere with Ac-6-FP binding, but forms hydrophobic interactions with the ligand in an analogous manner to how Ac-6-FP binds to WT MR1 (Fig. 4D). Consequently, while the non-stimulatory ligand can still bind MR1^R9H, the stimulatory ligand cannot.

The MR1^R9H mutation caused structural instability of MR1 impacting its cellular trafficking

Thermal stability assays showed that the empty MR1^R9H molecule was less stable compared to MR1^R9H-Ac-6-FP, and much less stable than WT MR1 when it is bound to either 5-OP-RU or Ac-6-FP (Fig. 5A). To examine whether the unstable nature of MR1^R9H impacted on its cellular trafficking, we examined its maturation through the secretory pathway by testing endoglycosidase H (endo-H) sensitivity. It has been previously shown that WT MR1 is largely retained in the endoplasmic reticulum (ER) (endo-H sensitive), and the presence of ligand causes it to refold and traffic through the Golgi apparatus to the cell surface (endo-H resistant) (18). We observed that MR1^R9H resides mostly in the ER, similar to WT MR1, and in the presence of Ac-6-FP it could traffic to the cell surface (Fig. 5B). Whereas upon addition of 5-OP-RU the MR1^R9H remained in the ER. Furthermore, when we examined the stability of MR1 molecules at the cell surface by measuring its rate of internalisation, MR1^R9H was internalised faster than WT MR1, both without ligand and in the presence of Ac-6-FP (Fig. 5C). These results confirm the MR1^R9H ligand discrimination that we have previously described above and also demonstrates the unstable nature of MR1^R9H at the cell surface.

The MR1^R9H/R9H patient did not have any detectable MR1-restricted T cell populations

To examine the MR1^R9H/R9H patient’s MR1-restricted T cell populations, we stained PBMC with WT MR1 and MR1^R9H tetramers. The patient did not have any WT MR1-restricted T cell populations present (Fig. S9). We also did not observe any difference between MR1^WT/WT controls’ and MR1^R9H/R9H patient’s MR1^R9H-Ac-6-FP or -empty tetramer⁺ T cell frequencies, as no distinct populations were detected. This lack of TCR reactivity towards MR1^R9H was confirmed by surface plasmon resonance (SPR) experiments (Fig. S10). Thus, the MR1^R9H/R9H patient does not have any detectable MR1-restricted T cell populations.

Immune phenotyping of the MR1^R9H/R9H patient’s PBMC revealed that all other immune cell populations were present

We next wanted to determine whether the MR1^R9H/R9H patient’s immune system was altered or compensating for the absence of MAIT cells. Immune cell phenotyping revealed that the MR1^R9H/R9H patient’s circulating natural killer (NK) cell and monocyte frequencies were similar or slightly lower (for NK cells) compared to MR1^WT/WT controls, and there were no major differences to the NK or monocyte subsets (Fig. S11). However, when examining the B cell subsets, we observed a reduction in memory B cells in the MR1^R9H/R9H patient (Fig. S12). This lack of memory B cells could be an indication of deficient humoral immunity and is a typical observation made in a range of primary immunodeficiencies, such as: specific antibody deficiency and common variable immunodeficiency (20). However, the MR1^R9H/R9H patient does still have detectable plasma cells (Fig. S12B), normal serum Ig levels (Table S5), and a robust vaccine response (see supplementary text) indicating that their low frequency of memory B cells does not have a considerable impact on the ability to mount humoral immune responses. These results confirm that the MR1^R9H/R9H patient’s immune cell subsets are otherwise intact, with only a possible impairment in their memory B cell compartment.

MR1^R9H/R9H patient’s Vγ9/Vδ2⁺ T cells were over-expanded and fully functional in response to various stimuli

We also examined the MR1^R9H/R9H patient’s γδ T cells and observed a 7-fold higher circulating frequency compared to healthy donors (Fig. 6A). This expanded γδ T cell population was almost exclusively comprised of Vγ9/Vδ2⁺ T cells (Fig. 6B) that had a distinct CD27⁻ CD28⁻ phenotype (Fig. 6C). This is not a hallmark of HPV infection associated with primary immunodeficiency, as we also analysed samples from patients with confirmed CXCR4 gain-of-function mutations resulting in WHIM (warts, hypogammaglobulinemia, immunodeficiency, and myelokathexis) syndrome in which patients are highly susceptible to HPV infections (21, 22). The WHIM syndrome patients that tested HPV⁺ were examined and found to have a normal frequency of γδ T cells (Fig. 6A–B). We tested the functionality of the MR1^R9H/R9H patient’s Vγ9/Vδ2⁺ T cells and found that they exhibited intact effector functions in response to a range of different stimuli similar to MR1^WT/WT controls (Fig. 6D). Paired TCR repertoire analysis revealed that the MR1^R9H/R9H patient’s Vγ9/Vδ2⁺ T cells are clonally expanded, with one dominant clone comprising ~30% of the γδ T cell population (Fig. 6E), which is not typically seen in Vγ9/Vδ2⁺ T cells from healthy donors (23). An unusual feature of the MR1^R9H/R9H patient’s Vγ9/Vδ2⁺ T cell population was that 50% of the TCR repertoire used the TRDJ3 gene segment, compared to healthy adults which predominantly use TRDJ1 (Fig. 6F).

Next we examined the total TNF/IFNγ production by PBMC in response to stimuli. The MR1^R9H/R9H PBMC showed a robust TNF/IFNγ response, for which 89% of the cells producing cytokine in response to E. coli were the Vγ9/Vδ2⁺ T cells (Fig. 7A). To confirm this robust antibacterial response, we challenged PBMC ex vivo to a variety of bacteria and examined the total cytokine production and cytotoxic response. We found that the patient’s PBMC could produce proinflammatory cytokines and cytotoxic granules to all bacteria strains to the same extent as healthy donors (Fig. 7B and Fig. S13). These results suggest that the MR1^R9H/R9H patient is still able to produce a robust effector response when challenged with bacteria.

Study Design

The MR1^R9H/R9H patient was recruited to the Genetic Immunology Flagship of Australian Genomics Health Alliance, which was approved by human research ethics committees at ACT Health Human Research Ethics Committee and Melbourne Health (protocols ETH.8.17.171E and HREC/16/MH/251) for whole exome sequencing to determine the underlying cause of his primary immunodeficiency. Immune phenotyping of the patient’s blood samples was done with ethical approval for additional samples collected on three separate occasions at Monash Medical Centre from the Monash Health Human Research Ethics Committee (project ID: 52046). The MR1^R9H/R9H patient’s parents’ blood was also collected for immune phenotyping. For comparison to the patient, age- and gender-matched healthy donors were recruited and blood samples taken with ethical approval from the Monash University Human Research Ethics Committee (project ID: 19488). To provide a HPV⁺ control group for immune phenotyping comparison, WHIM patient samples were collected as part of a clinical research protocol approved by the Institutional Review Board of the National Institute of Allergy and Infectious Diseases, NIH. Written informed consent was obtained for all study volunteers in accordance with the Declaration of Helsinki. The study was unblinded and sample sizes were limited by the availability of donors but adequate to detect differences between sample groups.

Whole exome sequencing

Genomic DNA was isolated from blood using the QIAamp DNA Blood Kit (Qiagen) according to the manufacturer’s instructions. For whole exome sequencing, Input DNA was analysed for integrity then fragmented by nebulization. DNA library preparation was performed using Agilent High Sensitivity whole exome capture and libraries were sequenced as paired end reads (100 bp runs) on an Illumina HiSeq 2500. Bioinformatics analysis was performed as previously described (35).

Human samples

Patients were recruited to the study at the Monash Medical Centre and a blood sample taken during a routine appointment. Healthy adult blood donors were from laboratory volunteers, from buffy packs obtained from Australian Red Cross Lifeblood or were obtained from the National Institutes of Health (NIH) Clinical Center Department of Transfusion Medicine. Peripheral blood mononuclear cells (PBMC) were isolated from all blood samples and buffy packs by gradient centrifugation using Lymphoprep™ (STEMCELL Technologies) or Ficoll-Paque PLUS density gradient media (GE Healthcare). Cells were then either used fresh or cryogenically frozen for later analysis.

Polymerase chain reaction (PCR) confirmation of IFIH1 splice variation

RNA was extracted from PBMC samples using the PureLink™ RNA Mini Kit (Ambion) and cDNA produced using SuperScript™ IV First-Strand Synthesis System (Invitrogen™) following the manufacturer’s instructions. The cDNA was then used in a standard PCR reaction with GoTaq® Green Master Mix (Promega) following manufacturer’s instructions using the following primers designed to amplify the regions flanking exon 8 of the IFIH1 transcript: IFIH1.F, 5’-ACGAAGCAAGCCAAAGCTGA and IFIH1.R 5’-GGCCCATTGTTCATAGGGTTGA. Samples were then run on a 2% agarose gel with a 100 bp DNA ladder (New England Biolabs). Band size of the PCR product indicated whether the exon 8 was present (258 bp) or absent (139 bp).

TaqMan® qPCR assay for IFNA1 and IFNB1 induction

PBMC were cultured with or without 10 μg/mL Poly(I:C) high molecular weight (HMW)/LyoVec™ (InvivoGen) for 24 hr. RNA was extracted and cDNA synthesised as described in the previous section. The cDNA was then used in a TaqMan® qPCR assay using TaqMan™ Fast Advanced Master Mix (Applied Biosystems) with TaqMan® Gene Expression Assays for IFNA1 and IFNB1 and the reference gene HPRT1 (Thermo Fisher Scientific). Samples were then run on a CFX Connect Real-Time PCR Detection System (Bio-Rad) and quantified using the Comparative CT Method (2^-ΔΔCT method).

Sanger sequencing of MR1

RNA was extracted and cDNA synthesised as described in the previous section. The cDNA was then used in a standard PCR reaction with GoTaq® Green Master Mix following manufacturer’s instructions using the following primers designed to amplify position R9 on the MR1 transcript: MR1.F, 5’-GGAACTGATGGCGTTCCTGT and MR1.R, 5’-GTGGTGATAGGGTGCGAGTC. Samples were then sequenced with BigDye Terminator v3.1 (Applied Biosystems) following manufacturer’s instructions and run on an Applied Biosystems 3730S capillary sequencer (Micromon Genomics, Monash University). Results were visualised using SnapGene® Viewer.

Antibodies and staining reagents

APC anti-human CD56 (NCAM16.2), BB700 anti-human CD19 (SJ25C1), BUV395 anti-human CD3 (NCAM16.2), BUV395 anti-human CD8 (RPA-T8), BV786 anti-human IFN-γ (4S.B3), FITC anti-human IgM (G20–127), PE Streptavidin, PE-CF594 anti-human IgD (IA6–2) and PerCP-Cy™5.5 anti-human TNF (MAb11) were purchased from BD Biosciences. Alexa Fluor® 700 anti-human CD16 (3G8), Alexa Fluor® 700 anti-human CD4 (SK3), APC or Alexa Fluor® 700 anti-human CD161 (HP-3G10), Brilliant Violet 605™ anti-human CD38 (HIT2), Brilliant Violet 605™ anti-human TCR Vα7.2 (3C10), Brilliant Violet 650™ anti-human CD28 (CD28.2), Brilliant Violet 711™ anti-human CD45RA (HI100), Brilliant Violet 785™ anti-human CD45RO (UCHL1), PE anti-human CD14 (M5E2), PE anti-human IL-2 (MQ1–17H12), PE/Cy7 anti-human CD27 (M-T271), PE/Cy7 anti-human CD3 (UCHT-1) and PE/Dazzle 594 anti-human CD27 (M-T271) were purchased from Biolegend. APC anti-human TCR Vδ2 (123R3), APC-Vio770 anti-human TCRγ/δ (REA591), FITC anti-human TCR Vδ1 (REA173) and VioBlue anti-human TCRα/β (REA652) were purchased from Miltenyi Biotec. PC5 anti-human TCR Vγ9 (IMMU 360) was purchased from Beckman Coulter. A hybridoma cell line was used to produce the MR1 8F2.F9 and 26.5 antibodies (36). PBS-57-loaded CD1d tetramer was obtained from the NIH tetramer core facility. Dead cells were excluded using the viability dye LIVE/DEAD® Fixable Aqua Dead Cell Stain Kit (Life Technologies). For immunoprecipitation experiments, a polyclonal rabbit antiserum was generated by the Walter and Eliza Hall Institute Antibody Facility against a peptide corresponding to the final 15 residues of the cytosolic tail of human MR1 (PREQNGAIYLPTPDR) (anti-MR1-CT).

Flow cytometry and cell sorting

For standard flow cytometry cell staining, cells were first stained with viability dye for 10 min at room temperature, followed by antibodies diluted in phosphate buffered saline with 10% foetal bovine serum (Sigma) for 20 min on ice. Flow cytometry data was collected on a BD LSRFortessa™ X20. Cell sorting was performed on BD FACSAria™ Fusion Cell Sorter. Data was analysed using FlowJo™ cell analysis software (FlowJo, LLC).

Ligands and compounds

Production of 5-OP-RU has been described previously (6). Ac-6-FP was purchased from Schircks Laboratories.

Mucosal-associated invariant T (MAIT) cell activation assay

PBMC were cultured with 1 nM 5-OP-RU for 4 hr, then Brefeldin A solution (eBioscience) was added and cells cultured for 18 hr. Cells were collected and stained with viability dye, prior to fixation and permeabilisation using the eBioscience™ Intracellular Fixation & Permeabilisation Buffer Set (Invitrogen). Antibodies diluted in permeabilisation buffer were then incubated with cells for 45 min at room temperature.

Cell lines

Production of C1R cells overexpressing MR1 under the control of the p-MSCV-IRES-eGFP (pMIG) vector have been previously described (37). Using this protocol we produced C1R cells expressing a single amino acid substitution R9H in MR1 by designing gBlock gene fragments (IDT) and cloning it into the pMIG vector and transfecting it into 293T cells with the retroviral packaging vectors (pPAM-E and pVSVG) using Fugene 6. The supernatant containing viral particles was then filtered through a 0.45 μm filter and added to WT C1R cells with polybrene to produce a stable transduction of cells overexpressing MR1^R9H. The transduced C1R.MR1^R9H cell line was then sorted for high expression of GFP and MR1.

Bacterial cultures

The bacteria: Escherichia coli, Pseudomonas aeruginosa, Staphylococcus aureus, Staphylococcus epidermidis and Enterococcus faecalis were grown in Lysogeny broth (LB) overnight, then diluted 1:100 and grown to log phase for 3 hr. OD₆₀₀ was measured and used to determine concentration.

MR1 upregulation assay

Cells were cultured in the presence of 5-OP-RU (1 μM) or Ac-6-FP (10 μM) for 4 hr (unless specified otherwise) prior to staining with viability dye. Cells were then stained with biotinylated MR1 8F2.F9 for 30 min on ice, followed by streptavidin-PE (and any directly conjugated primary antibodies, if being used) for 20 min on ice. Cells were then analysed on a flow cytometer.

Expression, Refold and Purification of MR1-β2m and MAIT TCR proteins

Human WT MR1-β2m and MR1^R9H-β2m were refolded in the presence of either Ac-6-FP or 5-OP-RU, as described previously (4). Human MR1^R9H-β2m-empty was refolded in the absence of any ligand, producing refolded protein, confirmed by native gel and size exclusion chromatography (Fig, S6). Soluble A-F7 (TRAV1–2/TRBV6–1) and TRAV1–2/TRBV6–4 MAIT TCR proteins were refolded from inclusion bodies, as described previously (38, 39). Refolded MR1-ligand and TCR proteins were purified by three consecutive purification steps: crude anion exchange, S200 15/60 size exclusion chromatography and HiTrap-Q HP anion exchange. The protein purity was then assessed using sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and quantified using a NanoDrop™ spectrophotometer.

Generation of MR1 tetramers

C-terminal cysteine-tagged human-MR1-β2m (complexed with ligand where indicated) were refolded and purified, as described in the previous section. The purified proteins were reduced by dithiothreitol (DTT), buffer exchanged into HEPES buffered saline (10 mM HEPES pH 7.3, 150 mM NaCl, 1 mM EDTA), and then biotinylated by incubation with biotin-maleimide overnight at 4℃. The biotinylated samples were purified by gel filtration and biotinylation efficiency was assessed by native gel. Biotinylated proteins were then tetramerised with PE-conjugated streptavidin and used to stain cells at a 1:50 dilution.

Thermal stability assay

A thermal shift assay of the WT MR1 and MR1^R9H proteins was performed in a real-time detection system (Corbett RotorGene 3000) using the fluorescent dye SYPRO® Orange (Sigma) to monitor the protein unfolding upon heating. Both WT and MR1^R9H (complexed with ligand where indicated) were purified by gel filtration just before the experiment and eluted in buffer of 10 mM Tris-HCl (pH 8), 150 mM NaCl. The samples were then heated from 25–95℃ with a heating rate of 1℃/min. The fluorescence intensity was measured (excitation 530 nm, emission 610 nm) and used to infer protein unfolding. The half maximum melting point (Tm50) represents the temperature for which 50% of the protein is unfolded.

Protein Crystallization, Structure Determination and Refinement

Purified TRAV1–2-TRBV6–1 TCR was mixed with either WT or MR1^R9H proteins (complexed with ligand where indicated) in 1:1 molar ratio at concentration of 2–4 mg/mL and incubated on ice for 4 hr. Hanging-drop method was employed to produce crystals with a precipitant consisting of 100 mM BTP (pH 6.1–6.7), 8–14% PEG3350 and 200 mM sodium acetate, as reported previously (38). Hexagonal complex crystals formed within a week at 20℃ and were flash frozen in liquid nitrogen after a short soak in reservoir solution with 14% glycerol and 10× molar amount of the loaded ligand (either 5-OP-RU or Ac-6-FP) for cryo-protection. X-ray diffraction data was collected at 100 K on the Australian Synchrotron at either MX1 or MX2 beamlines. Diffraction images were processed using XDS (40) and programs from the CCP4 suite (41) and Phenix package (42). Both ternary structures were determined by molecular replacement using PHASER (43), where modified TCR-MR1 ternary complex (PDB: 4NQC) was used as a search model. Model building was performed in COOT (44), accompanied with iterative rounds of refinement using Phenix.refine (42). The Grade Webserver and Phenix tools were used to build and to generate ligand restraints. The models were validated using MolProbity (45) and graphical representations were generated using PyMOL Molecular Graphics System, Version 1.8 (Schrödinger, LLC, New York, NY).

Immunoprecipitation

For immunoblotting and detection of MR1, cells were lysed in 0.5% IGEPAL CA-630 (Sigma-Aldrich), 50 mM Tris-HCl (pH 7.4), 5 mM MgCl₂ with Complete Protease Inhibitor Cocktail (Roche), and nuclei separated by centrifugation at 13,000 × g for 10 min. Lysates were precleared with normal rabbit serum (Sigma-Aldrich) and protein G-Sepharose then with protein G-Sepharose alone. MR1 was immunoprecipitated using anti-MR1-CT and protein G-Sepharose, and then washed three times with NET buffer (0.5% IGEPAL CA-630, 50 mM Tris-Cl pH 7.4, 150 mM NaCl, 5 mM EDTA) and treated with Endoglycosidase Hf (New England Biolabs) according to the manufacturer’s instructions. Protein samples were denatured with reducing SDS-PAGE sample buffer, separated on NuPAGE 4–12% Bis-Tris precast gel (Life Technologies) and immunoblotted onto nitrocellulose membrane using the iBlot system (Life Technologies). Band density was quantified using FIJI analysis software.

Internalisation assay

To measure MR1 internalisation, a fluorescence internalisation probe (FIP) using 8F2.F9 MR1 antibody was generated as previously described (46). Briefly, the FIP conjugated to MR1 contains a DNA sequence conjugated to the fluorescent dye Cy5, which can be quenched by using a quencher conjugated to the complementary DNA sequence. Any Cy5 signal not quenched indicates protection due to internalisation. Cells were labelled with MR1-FIP on ice, then washed and incubated at 37°C for the indicated times prior to being placed on ice to halt internalisation. Cells were washed and resuspended with or without the quencher, and Cy5 fluorescence analysed by flow cytometry. The Cy5 signal remaining after quenching indicates internalised MR1 and was calculated as a percentage of the initial signal before internalisation.

γδ T cell activation assay

PBMC were cultured and pulsed with a range of stimulants. For cytokine stimulation: 100 U/mL human interleukin (IL)-2 (Roche), 200 ng/mL human IL-12 (Miltenyi Biotec) and 200 ng/mL human IL-18 (Biolegend), two doses 72 hr apart, with brefeldin A (BFA) (Biolegend) added after second dose and cultured for 18 hr. For bacterial stimulation: E. coli at multiplicity of infection (MOI) 50 was added to PBMC, left for 1 hr prior to adding 50 μg/mL gentamicin (Sigma) and BFA and cultured for 18 hr. For (E)-4-Hydroxy-3-methyl-but-2-enyl pyrophosphate (HMBPP) stimulation: 10 ng/mL of HMBPP (Cayman Chemicals) was added and cultured for 4 hr before adding BFA and cultured for 18 hr. For a positive control: 50 ng/mL phorbol 12-myristate 13-acetate (PMA) (Sigma) and 1 µg/mL ionomycin (Sigma) were added with BFA and cultured for 18 hr. Following stimulation, cells were collected and stained with viability dye and stained with surface antibodies for 20 min on ice, prior to fixation and permeabilisation using the eBioscience™ Intracellular Fixation & Permeabilisation Buffer Set (Invitrogen). Intracellular antibodies were then incubated for 45 min diluted in permeabilisation buffer.

Single cell γδ TCR sequencing

PBMC were surface stained with antibodies and γδ T cells were single cell sorted directly into individual wells in a 96-well PCR plate containing 2 µL of SensiFAST cDNA Synthesis Kit reaction mix (Bioline) with 0.1% Triton X-100. Samples were incubated according to manufacturer’s instructions. TCRγ and TCRδ sequences were amplified using two rounds of nested PCR using GoTaq mastermix (Promega) and primers for Vδ2: TCTGGGCAGGAGTCATGT (external) and GAAAGGAGAAGCGATCGGTAAC (internal); Cδ: GCAGGATCAAACTCTGTTATCTTC (external) and TCCTTCACCAGACAAGCGAC (internal); Vγ1–8: CTGGTACCTACACCAGGAGGGGAAGG (external) and TGTGTTGGAATCAGGAVTCAG (internal); Vγ9: AGAGAGACCTGGTGAAGTCATACA (external) and GGTGGATAGGATACCTGAAACG (internal); and Cγ: CTGACGATACATCTGTGTTCTTTG (external) and AATCGTGTTGCTCTTCTTTTCTT (internal). PCR products were visualized on a 2% agarose gel, and products of successful reactions were incubated ExoSAP-IT™ PCR Product Cleanup Reagent (Applied Biosystems) before sequencing with BigDye Terminator v3.1 (Applied Biosystems) following the manufacturer’s instructions and running on an Applied Biosystems 3730S capillary sequencer (Micromon Genomics, Monash University). Results were visualised using SnapGene® Viewer.

Bulk γδ TCR repertoire sequencing

RNA was purified from 10 000 sorted Vγ9⁺Vδ2⁺ T cells using an RNAmicro kit (Qiagen) according to the manufacturer’s instructions. Amplicon rescued multiplex-PCR and next generation sequencing (NGS) methods (47) were used to analyse the sorted γδ T cell population from three healthy donors. Following initial first-round reverse transcription PCR using high concentrations of gene-specific primers, universal primers were used for the exponential phase of amplification (48). (Patent: WO2009137255A2). All cDNA synthesis, amplification, NGS library preparation were performed using the TRD/TRG iR-profile kits (HTGI-M or HTDI-M; iRepertoire, Inc. Huntsville, USA) and subsequent libraries were pooled and deep sequenced using an illumina MiSeq NGS platform (Micromon Genomics, Monash University).

Allogenic T cell/B cell coculture assay

B cells were isolated from PBMC using B Cell Isolation Kit II, human (Miltenyi Biotec) and T cells from an allogenic donor were isolated using CD2 MicroBeads, human (Miltenyi Biotec) and cell isolation was performed using the Miltenyi Biotec MidiMACS™ separator following the manufacturer’s instructions. B cells were cultured with 5-OP-RU at indicated concentrations for 4 hr. The B cells were then washed twice prior to addition of BFA and allogenic (MHC miss-matched, as MR1 is monomorphic) T cells at 1:1 effector:target ratio and cultured for 18 hr. Cells were collected and stained with viability dye, prior to fixation and permeabilisation using the eBioscience™ Intracellular Fixation & Permeabilisation Buffer Set (Invitrogen). Antibodies diluted in permeabilisation buffer were then incubated for 45 min at room temperature.

Whole PBMC ex vivo bacterial challenge

PBMC were cultured with the indicated bacteria at MOI 50 for 1 hr, then 50 μg/mL gentamicin was added and cultured for 24 hr. Supernatant was then collected and the cytokine and cytotoxic profile examined using the LEGENDplex™ CD8/Natural Killer (NK) panel following the manufacturer’s instructions. Samples were then run on a BD LSRFortessa™ X20 and data analysed using the LEGENDplex™ Data Analysis Software.

Statistical Analysis

Statistical analysis was performed using GraphPad Prism version 8 for macOS, GraphPad Software, San Diego, California USA. For multiple comparisons with two independent variables two-way ANOVA with Geisser-Greenhouse correction and Sidak’s multiple comparison test was used. Significant result is indicated where P < 0.05.