Mice
B10.A (B10. A/SgSnSlc H2a), B10.D2 (B10. D2/nSnSlc H2d), B10.S (B10. S/SgSlc H2s), C57BL/10 (C57BL/10SnSlc H2b) congenic, Balb/c (BALB/cCrSlc) and C57BL/6 J (C57BL/6JJmsSlc) mice were purchased from Japan SLC. SJL/J mice (Jackson strain no: 000686) were purchased from Charles River Laboratories Japan. NOD (NOD/SHiJcl) mice were purchased from CLEA JAPAN. To induce diabetes, cyclophosphamide (C2236, TCI) was administered to seven-week-old female NOD mice on day0 and day1425. MD4 BCR transgenic mice (Jackson strain no: 002595) were kindly provided by Prof. Tomonori Kurosaki (Osaka University). MD4 mice were backcrossed for over 15 generations to the B10.A-H2 h4 /(4 R) SgDvEgJ strain (Jackson strain no: 001150) purchased from the Jackson laboratory. Mice were kept in specific pathogen-free cages in a controlled environment, and experimental/control animals were co-housed. Mice were euthanized using carbon dioxide inhalation. All experiments were approved by the Animal Research Committee of the Research Institute for Microbial Diseases, Osaka University and conducted in accordance with the guidelines of the committee (R05-01-0).
Cells
The human embryonic kidney 293 T cells (RCB2202) were purchased from the RIKEN Cell Bank, and LK35.2 cells (ATCC HB-98) expressing I-Ak and NK92 cells (ATCC CRL-2407) were obtained from the American Type Culture Collection. The invariant chain (Ii) or H-2Mα deleted LK35.2 cells were generated by the CRISPR-Cas9 system using the pX330 vector (Addgene: ID42230) inserted into the primers (Ii:5’-caccgtacaccggtgtctctgtcc-3’ and 5’-aaacggacagagacaccggtgtac-3’, H-2Mα: 5’-caccgattcccaacatagggctct-3’ and 5’-aaacagagccctatgttgggaatc-3’). The Expi293F cells were purchased from Thermo Fisher Scientific. Each cell line was tested regularly for Mycoplasma contamination using PCR.
Plasmids and transfection
Plasmids for MHC class II molecules and 3A9 single-chain TCR-Fc fusion protein were constructed as previously described12. The cDNAs were cloned into the pME18S expression vector. MHC-II mutants were generated by site-directed mutagenesis. 293 T cells were transiently co-transfected with MHC-II and pMxs-GFP using PEI Max (24765, Polysciences), the peptide was added on day 2, and GFP-positive cells were analyzed 3 days post-transfection. The gating strategy of Ab titers and interaction analysis of monoclonal iTab and MHC-II mutants was shown in Supplementary Fig. 1b. The plasmids for the 3A9 single-chain TCR-Fc fusion protein were generated from cDNA derived from the 3A9 T cell hybridoma with 3 × (GGGS) linkers between the extracellular domains of the α and β chains. The construct was cloned into the pME18S expression vector, which contains the human IgG1 constant region. Plasmids for mouse CD4 (accession No.: NM_013488) and TCR α and β genes of 3A9, 2E5, 4E3, 5B6, 7A5 and SPL1.1 were synthesized (Integrated Device Technology) according to the published TCR gene sequence and cloned into pMxs retroviral vector22,26. The generation of these CD4 and TCRαβ stable transfectants was achieved with PLAT-E retroviral packaging cells with an amphotropic envelope. The sequence encoding 11–72 heavy and light chains from hybridoma cDNA was cloned into a pcDNA3.4 expression vector containing the SLAM signal sequence. Also, 11–72 heavy or light chain mutations were introduced using a QuikChange multi-mutagenesis kit (200514, Agilent). Recombinant Human MOG (accession No.: HSU64567) fused with GS-linker and 6×His tag was also cloned into pcDNA3.4, then transfected to Expi293F cells. The DNA sequences of these constructs were confirmed by sequencing (ABI3130xl).
Peptides
The HEL peptide library, overlapping by 10 residues, was purchased from SCRUM Inc. The 25-amino-acid OVA peptide library, overlapping by 8 residues, was purchased from GenScript. Other peptides were also obtained from GenScript, with a purity of > 90% as determined by high-performance liquid chromatography. The epitopes encoded in the different monomeric constructs used here include; OVA323–339(ISQAVHAAHAEINEAGR), HEL41–70(QATNRNTDGSTDYGILQINSRWWCNDGRTP), HEL41–61(QATNRNTDGSTDYGILQINSR), HEL48–70(DGSTDYGILQINSRWWCNDGRTP), HEL48–61(DGSTDYGILQINSR), HEL48–62(DGSTDYGILQINSRW), HEL48–63(DGSTDYGILQINSRWW), HEL48–64(DGSTDYGILQINSRWWA), HEL48–64(62A)(DGSTDYGILQINSRAWA), HEL48–64(63A)(DGSTDYGILQINSRWAA), HEL48–64(53,54W)(DGSTDWWILQINSRWAA), HEL48–64(59,60K)(DGSTDYGILQIKKRWAA), HEL48–64(59,60W)(DGSTDYGILQIWWRWAA), biotinylated HEL48–64(BioGSGSDGSTDYGILQINSRWWA), Sm-P40234–246(PKSDNQIKAVPAS), Sm-P40237–246(DNQIKAVPAS), Sm-P40237–249(DNQIKAVPASQAL), Sm-P40234–246-WWA(PKSDNQIKAVPASWWA), PKS-HEL52–61(PKSDYGILQINSR), PKS-HEL52–64(PKSDYGILQINSRWWA), HEL48–61(53,56A) (DGSTDAGIAQINSR), HEL48–64(53,56A) (DGSTDAGIAQINSRWWA), PLP136–151(RVSHSLGKWLGHPDKF), biotinylated PLP136–151(BioGSGSRVSHSLGKWLGHPDKF), PLP139–151(HSLGKWLGHPDKF), PLP136–151(H147K) (RVSHSLGKWLGKPDKF), PLP139–151(H147K) (HSLGKWLGKPDKF), MOG35–55(MEVGWYRSPFSRVVHLYRNGK), InsB9–25(SHLVEALYLVCGERGFF). We employed the HEL48–64 peptide, in which the C-terminal cysteine was substituted with alanine, in order to inhibit the formation of S-S bonds in this paper.
Immunization
Mice were immunized subcutaneously (s.c.) with 100 μg of HEL (L6876, Sigma-Aldrich), OVA (A5503, Sigma-Aldrich), recombinant MOG or 30nmol of peptides in Complete Freund’s adjuvant (F5881, Sigma-Aldrich). Pre-immunization with mutated peptides was conducted using Incomplete Freund’s adjuvant (263910, BD Difco). Serum samples were collected from immunized six to eight-week-old female mice after a period of 2–4 weeks, and other immunization experiments also used the same mouse strain. B cell hybridomas were established according to standard protocols. Briefly, whole lymphocytes from lymph nodes were fused to P3U1 cells to generate hybridomas, which were then screened using MHC-II-transfectants pulsed with peptide by flow cytometry. The iTab clones were generated at a frequency of less than 1% of the total screened hybridomas. Antibodies from serum or mAbs were purified using Protein A and G sepharose (17127903 and 17061805, GE Healthcare) on a Profinia protein purification system (Bio-Rad) and gel filtration on an AKTA pure 25 (GE Healthcare)
Flow cytometry
APCs were prepared by stimulating them with peptides or proteins for either an overnight period or 36 h. The cells or beads were stained with immunized mouse serum at a dilution of 1:100–1000 (cell) or 1:200–4000 (beads), and then reacted with allophycocyanin (APC)-conjugated anti-mouse IgG or Alexa Fluor 647-conjugated anti-mouse IgG1 for LK35.2 cell staining as a secondary antibody. Finally, the dead cells were stained with propidium iodide (PI). The Aw3.18 antibody was purified from the supernatant of a hybridoma culture (CRL-2826, ATCC). Antibody titers were determined by serial dilution of positive serum. Anti-HEL or HEL peptide antibodies in serum were removed by Streptavidin (SA) conjugated Sepharose beads (17511301, GE Healthcare) fused with biotinylated antigen. For the detection of anti-peptide antibodies or anti-HEL antibodies, SA (Z02043, Genscript) was labeled with aldehyde/ sulfate latex beads (A37304, Invitrogen), and then biotinylated peptide or protein was mixed. The binding of the human receptor-Fc fusion 3A9 TCR-Fc was examined by premixed with an APC-conjugated anti-human IgG Fc antibody. The TCR-Fc competitive assay was conducted by reacting the diluted serum before the premixed-TCR-Fc, followed by staining. APC-conjugated anti-mouse CD4 was employed in the reporter assay. For the depletion of CD4 T cells, anti-CD4 antibody (BE0003-1, BioXcell) (0.2 mg) was administered intraperitoneally (i.p.) on day -2. For intracellular staining, cells were fixed and permeabilized using the BD Cytofix/Cytoperm Fixation/Permeabilization Solution Kit (554714, BD Biosciences), then stained with anti-mouse H2-DM, Ii (151002, Biolegend) and APC-conjugated anti-rat IgG. Flow cytometric analyses were conducted using either a FACSCalibur or a FACSVerse flow cytometer (BD Biosciences). The data were subsequently analyzed using FlowJo v10 software (FlowJo, LLC). For the details of antibodies used for flow cytometry, see Supplementary Table 5.
Immunoprecipitation and mass spectrometry
The sample preparation for mass spectrometry was performed based on previous reports27. Briefly, HEL protein pulsed LK35.2 cells were lysed by a cell lysis buffer containing 20 mM Tris-HCl pH, 8.0, 1 % 3-[(3-cholamidopropyl) dimethylammonio]−1-propanesulfonate (CHAPS) (347-04723, DOJINDO), 5 mM ethylenediaminetetraacetic acid (EDTA), Protease inhibitor cocktail (P8340, Sigma-Aldrich), 0.1 mM iodoacetamide, and 1 mM phenylmethanesulfonylfluoride (PMSF) (27327-81, Nacalai, Japan) for 60 min at 4 °C on a rotator and cleared for 20 min of centrifugation at 21,000 × g. For immunoprecipitation, the cell lysates were incubated with biotinylated anti-mouse MHC-II (205303, Biolegend)-coupled SA Sepharose overnight at 4 °C on a rotator. The MHC-II antibody-coupled SA Sepharose capturing MHC-II-peptide complexes was washed six times with washing buffer 1 containing 250 mM NaCl, 50 mM Tris-HCl, pH 8.0 and six times with washing buffer 2 containing 50 mM Tris-HCl, pH 8.0. For the detection of peptides, the immunoprecipitated products were eluted using 10 % acetic acid (01-0280-5, Sigma-Aldrich). For analysis of MHC-II binding peptides, the eluted immunoprecipitated products were filtered through VIVASPIN 6, MWCO 10,000, PES (VS0601, Sartorius). The lyophilized peptides were reconstituted in 0.1% formic acid. After cleanup using a C18 tip column, the samples were subjected to LC-MS analysis. Mass spectrometric analysis was performed using a quadrupole time-of-flight mass spectrometer coupled with trapped ion mobility spectrometry (timsTOF Pro 2, Bruker Daltonics) equipped with a nanoHPLC system (nanoElute 2, Bruker Daltonics) and a nanocapillary C18 column (NTCC-360/75-3-105, Nikkyo Technos). Acquired raw data were analyzed using PEAKS Xpro (Bioinformatics Solutions Inc.), searching against HEL supplemented with a common contaminant database. Peptide identification was filtered using a score threshold of − 10lgP ≥ 15.
Reporter assay
The TCR reporter cells were mouse T cell hybridomas that had been stably transfected with NFAT-GFP and FLAG-tagged DAP12, as previously described28,29. The original hybridoma TCR αβ chains were depleted using the CRISPR-Cas9 system with the pX330 vector inserted into the primers (α chain; 5’-CACCGTGCCGAAAACCATGGAATC-3’ and 5’-AAACGATTCCATGGTTTTCGGCAC-3’, βchain; 5’-CACCGAGAAATGTGACTCCACCCA-3’ and 5’-AAACTGGGTGGAGTCACATTTCTC-3’). Following cell depletion, mouse CD4 was stably transfected into the cells, and each TCR was introduced into the CD4-positive cells. The cells were sorted using a Cell Sorter (SH800Z, Sony). The TCR reporter cells (96-well plate at 2 × 104 cells per well) were co-cultured with HEL protein-pulsed LK35.2 cells or PLP136–151 peptide-loaded 293 T cells transfected with I-As (4 × 104 cells per well) and supplemented with or without antibodies for 16 h. The expression of GFP was analyzed using flow cytometry (gating strategy of reporter assay was shown in Supplementary Fig. 4d).
In vitro proliferation and IL-2 secretion assays
CD4+ T cells (2 × 105 cells per well) were isolated from peripheral lymph nodes of peptide-immunized mice using BioMag anti-mouse IgG (310007, QIAGEN) and a mouse CD4 T cell isolation kit (130-104-454, Miltenyi Biotec). Splenocytes as APC (6 × 105 cells per well) from wild-type mice were treated with Ack buffer and subsequently treated with mitomycin C. The cells were incubated with HEL48–64 peptides for 72 h in Advanced RPMI 1640 (49140-15, Nacalai tesque) supplemented with 1% FBS (Hyclone), Glutamax (35050061, Invitrogen) and penicillin/streptomycin (26253-84, Nacalai tesque), and pulsed with 1 μCi (3H) thymidine for 18 h before harvesting and reading on a beta counter reader (NET027, Perkin Elmer). Supernatants were collected 24 h later for the measurement of IL-2 by ELISA ready-set-go reagent set (88–7024, eBioscience).
Delayed-type hypersensitivity assays
Ten days post-immunization of HEL48–64 peptide, mice were challenged with 30 nmol of the same peptide, which was injected intradermally into the footpad. The right footpad was injected with the peptide, while the left footpad was injected with vehicle (PBS) as a baseline control. The degree of swelling was quantified 24 h post-challenge using a Thickness Gauge (SM-112, TECLOCK). Monoclonal antibodies (1 mg per injection) were administered intraperitoneally (i.p.) on days 3 and 6.
Antibody-dependent cellular cytotoxicity assays
NK9230,31 cells stably transfected with chimeric FcR (mouse FcγRIV extracellular domain and human FcγR III intracellular domain) (96-well V-bottom plate at 2 × 105 cells per well) were co-cultured with or without HEL48–64 peptide-loaded 293 T cells transfected with I-Ak (2 × 104 cells per well) and supplemented with antibodies for 6 h. The ratio of PI-positive dead cells was analyzed using flow cytometry. In vivo assay, B cells were isolated from the spleen of eight to ten-week-old female B10.A(4 R) and B10.A(4 R)-MD4 mice using the mouse B cell isolation kit (130-090-862, Miltenyi Biotec). These B cells were co-cultured with 10ng/ml IL-4 (404-ML-010, R&D Systems), 5 μg/ml anti-CD40 antibody (102812, Biolegend), 5 μg/ml anti-IgM antibody (115-006-075, Jackson Immuno Research) with or without HEL protein for 18h. These cells were mixed at a 1:1 ratio, labeled using the Cell trace violet proliferation kit (C34571, Invitrogen), and then transferred into eight-week-old female wild-type B10.A(4 R) mice. After 30 min, 0.5 mg of antibody was administered intravenously. Three days later, labeled+ cells were analyzed by flow cytometry. Anti-mouse IgMa antibody was used to distinguish wild-type B cells (labeled+, IgMa−) and MD4 B cells (labeled+, IgMa+). The gating strategy was shown in Supplementary Fig. 5h. The ratio of MD4 B cells to wild-type B cells was calculated. This ratio was further adjusted by setting the average ratio in mice without co-culture with HEL, treated with control antibody as 1.0.
Sample preparation for cryoEM analysis
The construction of soluble MHC class II molecules has been previously described32. The I-Akα plasmids were fused with a GSSGSSG-linker, TEV protease cleavage site, leucine zipper and His-tag at the C-terminus, and HEL48–64 peptide was bound to the I-Akβ plasmids with a GGGGSLVPRGSGGGGSGS-linker, which was attached via a GSSGSSG-linker, TEV protease cleavage site, leucine zipper and a Flag-tag at the C-terminus. The plasmids were subsequently inserted into the pcDNA3.4 expression vector. The recombinant protein was obtained via the Expi293 expression system. Monoclonal antibodies were obtained from hybridoma cells by digestion to the Fab fragment using immobilized papain (20341, Sigma-Aldrich). The purified HEL peptide-MHC-II protein by Ni Sepharose excel (17371201, cytiva) was mixed with the Fab fragment, and incubated on ice for 30 min. The mixture was loaded onto a Superdex 200 10/300 column (cytiva) equilibrated with 20 mM Tris-HCl buffer (pH 8.0) containing 150 mM NaCl and eluted with the same buffer. Fractions containing the complex were collected and concentrated to approximately 4 mg/ml by ultrafiltration.
Cryo-EM data collection
Holey carbon film-coated copper grids (Quantifoil R1.2/1.3 Cu 200 mesh; Microtools GmbH) pretreated by Au sputtering were glow-discharged for 10 s using an ion coater (JEC-3000FC; JEOL). A solution of 4.0 µL (1.0 mg/mL) of the MHC-II–HEL peptide–Fab complex, suspended in a buffer containing 20 mM Tris-HCl (pH 8.0) and 100 mM NaCl, was applied to the grid and blotted with filter paper for 6 s. The grid was then immediately plunged into a bath of cooled ethane using a FEI Vitrobot Mark IV (Thermo Fisher Scientific) under 100% humidity at 4 °C. Subsequently, the grids were examined using a CRYO ARM 300 electron microscope (JEOL)33, equipped with a cold-field emission gun and an in-column energy filter with a slit width of 20 eV. Dose-fractionated images were recorded using a K3 Summit direct electron detector (AMETEK) in counting mode, with a dose rate of ~ 99 e− Å−2 across 100 frames. All images were collected using SerialEM34 with AI-assisted hole detection via yoneoLocr35, at a nominal magnification of 100,000 ×, corresponding to a pixel size of 0.495 Å. The defocus range was estimated to − 0.8 µm to − 5.0 µm. A total of 17,175 movies were acquired.
Cryo-EM data processing
Data processing was performed using CryoSPARC 4.2.036 and Relion 3.1.437. The collected movies were aligned, dose-weighted, and averaged through motion correction implemented in CryoSPARC. The contrast transfer function (CTF) was estimated using Ctffind-4.1.1338. Subsequently, micrographs with an estimated resolution below 10 Å were excluded from the data processing. A total of 1,327,931 particles in 12,394 micrographs were selected through Blob picker and subjected to the first two-dimensional (2D) classification, template picker, and second 2D classification. Then, an initial three-dimensional (3D) map (C1 symmetry) was reconstructed through ab-initio reconstruction in CryoSPARC. Following homogeneous and non-uniform refinement yielded a 3D map at 3.7 Å resolution. The micrographs and 3D map obtained were exported to RELION 3.1.4, where particles were picked up again using Topaz and 3D reference picking due to a tendency for preferred orientations of the particles. Subsequent 2D and 3D classification in RELION yielded a total of 786,526 particles. Several rounds of 3D refinement, CTF refinement, and Bayesian polishing resulted in a 3D map at 2.91 Å resolution. However, the obtained 3D map was highly disordered, making model building difficult. Therefore, the selected particles and the 3D map were imported back into CryoSPARC and subjected to 3D classification (20 classes) again, resulting in the selection of 50,032 high-quality particles. Additionally, 3D flexible refinement (3DFlex)37 improved the disorder of the 3D map, resulting in a 3D map at 3.09 Å resolution based on the gold-standard Fourier Shell Correlation cutoff (0.143) criterion, through post-processing in RELION. Furthermore, an enhanced map was generated using EMready39, a program for improving the interpretability of the 3D map by leveraging similarity and correlation-guided deep learning, which was used as a reference for model building.
Atomic model building
An initial model was generated using the structure of murine MHC-II class II I-Ak with a peptide from hen egg lysozyme (PDB code: 1IAK21) and H-2 class II histocompatibility antigen (PDB code: 2PXY40) through the SWISS-MODEL server (https://swissmodel.expasy.org/). These models were fitted into the map using UCSF Chimera41. The fitted initial model was manually adjusted with COOT42 to better fit the 3D map, followed by refinement using Phenix43 and REFMAC5 in CCP-EM44,45. The refinement statistics of the final model were obtained using the comprehensive validation program in Phenix.
Cytokine secretion assay and adoptive transfer of EAE
Female SJL/J mice, aged 10–14 weeks, were immunized s.c. with 100 μl of a CFA emulsion containing 200 μg of Mycobacterium tuberculosis H37Ra (231141, Difco Laboratories) and 15nmol of PLP136–151 peptide. Monoclonal anti-PLP iTab (1 mg per injection) was administrated i.p. on days 0,4 and 8. Cytokine release was evaluated by culturing lymphocytes from lymph nodes on day 10 in vitro with 0.1 μM of PLP136–151 peptide or an irrelevant peptide (OVA323–339) for 48 h. Supernatants were collected and analyzed using a mouse Th1/Th2/Th17 cytokine kit (560486, BD Bioscience) by flow cytometry. Lymphocytes were cultured for an additional 48 h, after which 1 × 107 cells were transferred to naïve SJL/J mice. The animals were evaluated daily for clinical signs of disease as follows: (0) No disease; (0.5) Loss of tail tonus; (1) Complete tail paralysis; (2) Tail paralysis with hind limb weakness; (3) Complete hind limb paralysis; (4) Hind limbs paralyzed, weakness in forelimbs; (5) Tetraplegia; (6) Death.
iTab or peptide therapy in EAE
EAE was induced by immunization with 100 μl of a CFA emulsion containing 250 μg of Mycobacterium tuberculosis H37Ra and 50 nmol of PLP136–151 peptide. In iTab therapy, Monoclonal anti-PLP iTab (1 mg per injection) was administered i.p. on days 0,4 and 8. Two weeks before the onset of EAE, a pre-immunization procedure was conducted in which 50nmol of PLP peptide, which mutated for the pathogenic TCR recognition site at H147K, was administered s.c. with IFA.
Statistical analysis
Data were compared by Student’s t- test, one-way or two-way ANOVA tests. p-values < 0.05 were considered statistically significant.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.