High throughput pMHC-I tetramer library production using chaperone-mediated peptide exchange

Abstract

Peptide exchange technologies are essential for the generation of pMHC-multimer libraries used to probe diverse, polyclonal TCR repertoires in various settings. Here, using the molecular chaperone TAPBPR, we develop a robust method for the capture of stable, empty MHC-I molecules comprising murine H2 and human HLA alleles, which can be readily tetramerized and loaded with peptides of choice in a high-throughput manner. Alternatively, catalytic amounts of TAPBPR can be used to exchange placeholder peptides with high affinity peptides of interest. Using the same system, we describe high throughput assays to validate binding of multiple candidate peptides on empty MHC-I/TAPBPR complexes. Combined with tetramer-barcoding via a multi-modal cellular indexing technology, ECCITE-seq, our approach allows a combined analysis of TCR repertoires and other T cell transcription profiles together with their cognate antigen specificities in a single experiment. The new approach allows TCR/pMHC interactions to be interrogated easily at large scale.

Fig. 1. Linking peptide specificities with T cell transcriptomes.

Fluorophore-labeled, empty MHC-I/TAPBPR multimers are loaded with peptides of interest on a 96-well plate format and individually barcoded with DNA oligos designed for 10x Genomics and Illumina compatibility. TAPBPR and excess peptide, along with free oligo, are removed by centrifugation and the multimers are pooled together. A single patient sample can be stained with the pooled multimer library, and collected by fluorescence-activated cell sorting. Tetramer associated oligos and cellular mRNA from individual cells are then barcoded using 10x Genomics gel beads, followed by cDNA synthesis, library preparation and library sequencing. This workflow enables the transcriptome and paired αβ TCR sequences to be linked with pMHC specificities in a single experiment.

Fig. 2. Structure-guided design of placeholder peptides for murine MHC-I molecules.

Crystal structures showing stabilizing contacts in the A-pocket of the MHC-I peptide-binding groove for: a P18-I10/H-2D^d (PDB 3ECB¹) with the bound peptide shown in light red. b Structure of p29/H-2L^d (PDB 1LD9²), and QL9/H-2L^d (PDB 3TF7³), with bound peptides shown in yellow. Insets focus on the A-pocket of the MHC-I peptide binding groove, with polar contacts between the N-terminal residue of the peptide and the indicated MHC-I residues shown as dotted yellow lines. Full peptide sequences are provided, with the first residue indicated in green.

Fig. 3. Capturing empty HLA-A*02:01/TAPBPR complexes for peptide exchange.

a Structure-based design of goldilocks peptides: comparison of polar contacts between HLA-A*02:01 and the N-terminal region of LLFGYPVYV (TAX) peptide (upper left). _LFGYPVPYV (gTAX) (upper right), Ac-LLFGYPVYV (bottom right) and lLFGYPVYV where l = D-Leucine (bottom left). Structures were modeled using PDB ID IDUZ. b Peptide complex thermal stabilities of HLA-A*02:01 bound to TAX, lLFGYPVYV, Ac-LLFGYPVYV and gTAX. c SEC TAPBPR binding assays of TAX/HLA-A02:01 (left), gTAX/HLA-A02:01 (right). d Native gel electrophoresis of HLA-A*02:01/TAPBPR complex incubated with a 10-fold molar excess of a non-specific peptide (p29, YPNVNIHNF) or varying molar excess of a specific, high-affinity peptide (TAX). 12% polyacrylamide native gels were run at 90 V for 5 h at 4 °C before visualization with InstantBlue (Expedeon). Data shown are representative of triplicate gel assays. e Competitive binding of TAMRA-TAX to purified HLA-A*02:01/TAPBPR complexes from (c) as a function of increasing peptide concentration, measured by fluorescence polarization. f Conceptual diagram of TAPBPR-mediated capture and peptide loading on empty MHC-I molecules. g Bio-Layer Interferometry analysis of TAPBPR dissociation from HLA-A*02:01 in the presence of peptides of different affinities. Data shown in e and g are representative of triplicate assays (n = 3) and error-bars are standard deviations from the mean. Source data are provided as a Source Data file.

Fig. 4. TAPBPR-mediated peptide exchange on HLA-A*24:02 and HLA-A*68:02.

a Structure-based design of goldilocks peptides: analysis of A-pocket hydrogen bonds observed in the X-ray structure of HLA-A*24:02/Nef_134-10 (RYPLTFGWCF) and homology-based model of HLA-A*28:02/TAX (LLFGYPVYV). Peptide residues are labelled in red, A-pocket MHC residues in black. The structure of HLA-A*24:02/Nef_134-10 was obtained from PDB ID 3QZW. The structure of HLA-A*68:02/TAX was modeled using a related structure (PDB ID 4HX1) as input. b Native gel electrophoresis analysis showing MHC-I/TAPBPR complex dissociation in the presence of 10-fold molar excess of relevant, high-affinity peptides PHOX2B (QYNPIRTTF), Nef_134-10 (lanes 5 and 6) and TAX, MART-1 (lanes 10 and 11) for complexes prepared using refolded HLA-A*24:02 and HLA-A*68:02 with gNEF (lane 2) and gTAX (lane 7) goldilocks peptides, respectively (protein yields of approximately 6 and 8 mg from a 1 L refolding reaction). The exchanged pMHC molecular species on lanes 5,6,10 and 11 are indicated with arrows, showing electrophoretic mobilities that depend on the charge of the bound peptide. Both complexes remain bound in the presence of 10-fold molar excess of the irrelevant peptide p29 (YPNVNIHNF) (lanes 4 and 9). 12% polyacrylamide native gels were run at 90 V for 5 h at 4 °C before visualization with InstantBlue (Expedeon). Data shown are representative of triplicate gel assays. All protein samples used in a and b were derived from the same peptide exchange experiment, and the gels were processed in parallel. c Overlaid Differential Scanning Fluorimetry temperature profiles of: goldilocks/MHC-I (red) and high-affinity peptide/MHC-I (blue) prepared using chaperone-mediated exchange of the goldilocks for high-affinity peptides, followed by purification of the pMHC peak by SEC. Thermal stabilities are shown: HLA-A*24:02 refolded with gNEF (YPLTFGWCF) or exchanged with NEF (RYPLTFGWCF) (left), and HLA-A*68:02 refolded with gTAX (_LFGYPVYV) or exchanged with TAX (LLFGYPVYV) (right). The peaks at approximately 63 °C correspond to the thermal melt of the β₂m light chain. Data shown are representative of triplicate assays and error-bars are standard deviation from the mean. Source data are provided as a Source Data file.

Fig. 5. Flow cytometry using TAPBPR-exchanged pMHC-I tetramers.

a Representative flow cytometric analysis of top row; murine cells expressing the B4.3.2 TCR, middle row; DMF5 human T cells expressing the MART-1 TCR, bottom row; NY-ESO-1 human T cells expressing the NY-ESO-1 TCR. Columns depict staining with a fixed excess (1 µg/mL) of P18-I10/H2-D^d, MART-1/HLA-A*02:01 and NY-ESO-1/HLA-A*02:01 tetramers prepared either by conventional refolding, exchange of peptide on stoichiometric MHC-I/TAPBPR complexes, TAPBPR-mediated peptide exchange using a catalytic (1:100) TAPBPR to MHC-I molar ratio or by stoichiometric exchange of a mismatched peptide, not recognized by the respective TCRs. b Histogram plots of tetramer staining. c Tetramer titration of P18-I10/H-2D^d (top), MART-1/HLA-A*02:01(middle) and NY-ESO-1/HLA-A*02:01 (bottom), prepared by exchange of each peptide on the corresponding stoichiometric MHC-I/TAPBPR complexes. Percentage of cells staining positive with tetramer over a serial two-fold dilution series were plotted and EC₅₀ values calculated by curve fitting to a sigmoidal line (with R² values in the 0.97–0.99 range), using Graph Pad Prism version 8 (GraphPad Software, La Jolla California USA). Data shown are representative of triplicate assays (n = 3) and error-bars are standard deviations from the mean. Gating strategies used for sorting tetramer-positive cells are outlined in Supplementary Fig. 4. Source data are provided as a Source Data file.

Fig. 6. Fine-tuning MHC-I/TAPBPR interactions through α3 domain mutants.

a Alignment of the α₃ domain sequences from murine H-2D^d, H-2L^d and human HLA-A*02:01. Conserved residues are highlighted in yellow. The M228T mutation site is highlighted in blue. * indicates residues directly participating in TAPBPR interactions, as shown in published mutagenesis studies and crystal structures. b TAPBPR/H-2D^d α₃ domain interface from PDB ID 5WER (left panel) and with the M228T mutation modeled (right panel). c Size exclusion chromatograms of H-2D^d M228T refolded with either high-affinity P18-I10 or goldilocks gP18-I10 peptides, with and without TAPBPR. d LC/MS analysis of the peptide region from SEC-purified gP18/H-2D^dM228T and H-2D^dM228T/TAPBPR peaks from (c). Data shown are representative of triplicate SEC and LC/MS experiments.

Fig. 7. Identification of paired αβ TCR sequences with their antigen specificities.

a Recovery of MART-1 tetramer barcodes on DMF5 cells from PBMC-DC co-cultures spiked with 1% DMF5 cells. Number of MART-1 tetramer reads among DMF5 positive (DMF5-TCR) and negative (NB-TCR) cells. Cells are classified as positive/negative according to sequencing reads of the DMF5 TCR, where >10 DMF5 reads was used as a cutoff to classify DMF5+ cells. Box graphs display the distribution of MART-1 tetramer reads from DMF5+ (>10 DMF5 reads n = 76) and DMF5- (≤10 DMF5 reads, n = 927) cells. Upper/lower bars and box boundaries indicate the 95th/5th and 75th/25th percentiles, respectively, horizontal box lines indicate medians. Statistical significance was assessed using a two-tailed Mann-Whitney test (p-value < 0.0001). b Distribution of antigen specificities identified from tetramer + /CD8+ T cells from human splenocytes and the number of tetramer-barcode read per cell. Each dot represents a single cell, n = 102 in total, error bars correspond to one standard deviation from the mean. c, d V(D)J usage of cells identified as specific for the EBV-BRLF1 (YVLDHLIVV) and NY-ESO-1 (SLLMWITQA) antigens. All TRAVJ (n = 11/35) and TRBVJ (n = 14/35) chains identified are represented. e TCR CDR3 sequences identified for antigen-specific T cells. Consensus BRLF-1 TRA and NY-ESO-1 TRB chains are highlighted in orange and red, respectively. Gating strategies used for sorting tetramer+ cells are outlined in Supplementary Fig. 11. Source data are provided as a Source Data file.