Ligand discovery for T cell receptors

1. A library of polypeptides, or of nucleic acids encoding the library of polypeptides, wherein each polypeptide of the library of polypeptides has a structure of the formula P-L1-β-L2-α-L3-T, wherein

(a) each of L1, L2 and L3 are flexible linkers,

(b) α is a soluble form of a binding domain from an α chain of a human class I MHC protein or a human class II MHC protein,

(c) β is a soluble form of a binding domain from a β chain of a human class II MHC protein or a β2 microglobulin of a human class I MHC protein,

(d) T is a cell tether or transmembrane sequence, and

(e) P is a peptide ligand;

wherein the binding domains of (b) and (c) are:

(i) an allele of HLA-DRA and an allele of HLA-DRB15, wherein the allele of HLA-DRA comprises amino acid changes {F12S, M23K} and the allele of HLA-DRB15 comprises amino acid change {P11S}, or

(ii) H2-IE k a and H2-IE k b, wherein H2-IE k a comprises amino acid changes {I8T, F12S, L14T, A56V} and H2-IE k b comprises amino acid changes {W6S, L8T, L34S}.

2. The library of claim 1 wherein the library comprises at least 10 6 different polypeptides.

3. The library of claim 1 , wherein the peptide ligand is from 8 to 20 amino acids in length.

4. The library of claim 1 , wherein the library of polypeptides has limited diversity at an MHC anchor position.

5. The library of claim 4 , wherein the limited diversity at the MHC anchor position is limited to a group of known anchor residues at a known MHC anchor position, or a conservative substitution for a known anchor residue at a known MHC anchor position.

6. The library of claim 1 , wherein the MHC binding domains comprise an alpha 1 domain and an alpha 2 domain of a human class I MHC protein and a β2 microglobulin.

7. A method of identifying a set of peptides that bind to a T cell receptor (TCR) of interest, in a specific MHC protein context, the method comprising:

(a) contacting the TCR of interest with a population of host cells which express the library of polypeptides of claim 1 on the surface of the host cells.

8. The method of claim 7 , further comprising,

(b) selecting for a host cell that expresses a single chain polypeptide that binds to the TCR of interest;

(c) iterating the selecting step for at least three rounds to identify a final selected population of polynucleotides;

(d) sequencing the polynucleotides of the final selected population to determine a dataset of possible amino acids for each amino acid position of the polypeptide;

(e) inputting the dataset from d, into computer readable medium to generate a search algorithm comprising instructions operable to cause one or more data processing apparatuses to perform operations comprising: generating a n×20 matrix from one or more positional frequencies of the dataset from (d), wherein n is the number of amino acid positions in the polypeptide library

(f) searching a sequence database with the search algorithm using an n-position sliding window alignment with scoring a product of positional amino acid frequencies from a substitution matrix and an aligned segment containing at least one amino acid having a frequency that is below a cutoff is excluded as a match to identify a set of peptides that bind to the TCR of interest.

9. A cell line which expresses the library of polypeptides of claim 1 on the cell surface.

10. The cell line of claim 9 , wherein the cell line comprises a yeast cell line.