WO2005116074A2 - Nucleoproteins displaying native t cell receptor libraries - Google Patents

Abstract

A diverse library of nucleoproteins each displaying on its surface a polypeptide comprising a native TCR α variable domain sequence or native TCR ß variable domain sequence, in which library the diversity resides in the variety of native TCR α variable domain or native TCR ß variable domain sequences displayed. Such a library is useful for identifying a ligand or ligands of a target peptide-MHC or CD1-antigen.

Description

Nucleoproteins displaying native T cell receptor libraries

The invention relates to a diverse library of nucleoproteins each displaying on its surface a polypeptide comprising a native TCR α variable domain sequence or native TCR β variable domain sequence, in which library the diversity resides in the variety of native TCR α variable domain or native TCR β variable domain sequences displayed. Such a library is usefule for identifying a ligand or ligands of a target peptide-MHC (pMHC) complex or CD 1 -antigen.

Background to the Invention

Native TCRs

As is described in, for example, WO 99/60120 TCRs mediate the recognition of specific Major Histocompatibihty Complex (MHC)-peptide complexes by T cells and, as such, are essential to the functioning of the cellular arm of the immune system.

Antibodies and TCRs are the only two types of molecules which recognise antigens in a specific maimer, and thus the TCR is the only receptor for particular peptide antigens presented in MHC, the alien peptide often being the only sign of an abnormality within a cell. T cell recognition occurs when a T-cell and an antigen-presenting cell

(APC) are in direct physical contact, and is initiated by ligation of antigen-specific TCRs with pMHC complexes.

The native TCR is a heterodimeric cell surface protein of the immunoglobulin superfamily which is associated with invariant proteins of the CD3 complex involved in mediating signal transduction. TCRs exist in αβ and γδ forms, which are structurally similar but have quite distinct anatomical locations and probably functions. The MHC class I and class II ligands are also immunoglobulin superfamily proteins but are specialised for antigen presentation, with a highly polymorphic peptide binding site which enables them to present a diverse array of short peptide fragments at the APC cell surface. Two further classes of proteins are known to be capable of functioning as TCR ligands. (1) CD1 antigens are MHC class I-related molecules whose genes are located on a different chromosome from the classical MHC class I and class II antigens. CD1 molecules are capable of presenting peptide and non-peptide (e.g. lipid, glycolipid) moieties to T cells in a manner analogous to conventional class I and class II-MHC- pep complexes. See, for example (Barclay et al, (1997) The Leucocyte Antigen Factsbook 2^nd Edition, Academic Press) and (Bauer (1997) Eur J Immunol 27 (6) 1366-1373)) (2) Bacterial superantigens are soluble toxins which are capable of binding both class II MHC molecules and a subset of TCRs. (Fraser (1989) Nature 339

221-223) Many superantigens exhibit specificity for one or two Vbeta segments, whereas others exhibit more promiscuous binding, i any event, superantigens are capable of eliciting an enhanced immune response by virtue of their ability to stimulate subsets of T cells in a polyclonal fashion.

The extracellular portion of native heterodimeric αβ and γδ TCRs consist of two polypeptides each of which has a membrane-proximal constant domain, and a membrane-distal variable domain. Each of the constant and variable domains includes an intra-chain disulfide bond. The variable domains contain the highly polymorphic loops analogous to the complementarity determining regions (CDRs) of antibodies.

CDR3 of αβ TCRs predominantly interact with the peptide presented by MHC, and CDRs 1 and 2 of αβ TCRs predominantly interact with the peptide and the MHC. The diversity of TCR variable domain sequences is generated via somatic rearrangement of linked variable (V), diversity (D), joining (J), and constant genes

Functional α and γ chain polypeptides are formed by rearranged V-J-C regions, whereas β and δ chains consist of V-D-J-C regions. The extracellular constant domain has a membrane proximal region and an immunoglobulin region. There are single α and δ chain constant domains, known as TRAC and TRDC respectively. The β chain constant domain is composed of one of two different β constant domains, known as TRBCl and TRBC2 (LMGT nomenclature). There are four amino acid changes between these β constant domains, three of which are within the domains used to produce the single-chain TCRs displayed on phage particles of the present invention. These changes are all within exon 1 of TRBCl and TRBC2: N₄K₅->:E N₅ and F₃₇->Y ( GT numbering, differences TRBC1->TRBC2), the final amino acid change between the two TCR β chain constant regions being in exon 3 of TRBCl and TRBC2: Vι->E. The constant γ domain is composed of one of TRGC1, TRGC2(2x) or TRGC2(3x). The two TRGC2 constant domains differ only in the number of copies of the amino acids encoded by exon 2 of this gene that are present.

The extent of each of the TCR extracellular domains is somewhat variable. However, a person skilled in the art can readily determine the position of the domain boundaries using a reference such as The T Cell Receptor Facts Book, Lefranc & Lefranc, Publ. Academic Press 2001.

Recombinant TCRs

The production of recombinant TCRs is beneficial as these provide soluble TCR analogues suitable for the following purposes: • Studying the TCR / ligand interactions (e.g. pMHC for αβ TCRs) • Screening for inhibitors of TCR-associated interactions • Providing the basis for potential therapeutics

A number of constructs have been devised to date for the production of recombinant TCRs. These constructs fall into two broad classes, single-chain TCRs and dimeric

TCRs, the literature relevant to these constructs is summarised below.

Single-chain TCRs (scTCRs) are artificial constructs consisting of a single amino acid strand, which like native heterodimeric TCRs bind to MHC-peptide complexes.

Unfortunately, attempts to produce functional alpha/beta analogue scTCRs by simply linking the alpha and beta chains such that both are expressed in a single open reading frame have been unsuccessful, presumably because of the natural instability of the alpha-beta soluble domain pairing.

Accordingly, special techniques using various truncations of either or both of the alpha and beta chains have been necessary for the production of scTCRs. These formats appear to be applicable only to a very limited range of scTCR sequences. Soo Hoo et al (1992) PNAS. 89 (10): 4759-63 report the expression of a mouse TCR in single chain format from the 2C T cell clone using a truncated beta and alpha chain linked with a 25 amino acid linker and bacterial periplasmic expression (see also Schodin et al (1996) Mol. Immunol. 33 (9): 819-29). This design also forms the basis of the m6 single-chain TCR reported by Holler et al (2000) PNAS. 97 (10): 5387-92 which is derived from the 2C scTCR and binds to the same H2-Ld-restricted alloepitope. Shusta et al (2000) Nature Biotechnology 18: 754-759 and US 6,423,538 report using a murine single-chain 2C TCR constructs in yeast display experiments, which produced mutated TCRs with, enhanced thermal stability and solubility. This report also demonstrated the ability of these displayed 2C TCRs to selectively bind cells expressing their cognate pMHC. Khandekar et al (1997) J. Biol. Chem. 272 (51): 32190-7 report a similar design for the murine D10 TCR, although this scTCR was fused to MBP and expressed in bacterial cytoplasm (see also Hare et al (1999) Nat. Struct. Biol. 6 (6): 574-81). Hilyard et al (1994) PNAS. 91 (19): 9057-61 report a human scTCR specific for influenza matrix protein-HLA-A2, using a Vα-linker-Vβ design and expressed in bacterial periplasm.

Chung et al (1994) PNAS. 91 (26) 12654-8 report the production of a human scTCR using a Vα-linker-Vβ-Cβ design and expression on the surface of a mammalian cell line. This report does not include any reference to peptide-HLA specific binding of the scTCR. Plaksin et al (1997) J. Immunol. 158 (5): 2218-27 report a similar Vα-linker- Vβ-Cβ design for producing a murine scTCR specific for an HIV gpl20-H-2D^d epitope. This scTCR is expressed as bacterial inclusion bodies and refolded in vitro. A number of papers describe the production of TCR heterodimers which include the native disulphide bridge which connects the respective subunits (Garboczi, et al, (1996), Nature 384(6605): 134-41; Garboczi, et al, (1996), J Immunol 157(12): 5403- 10; Chang et al, (1994), PNAS USA 91: 11408-11412; Davodeau et al, (1993), J. Biol. Chem. 268(21): 15455-15460; Golden et al, (1997), J. Imm. Mem. 206: 163-

169; US Patent No. 6080840). However, although such TCRs can be recognised by TCR-specific antibodies, none were shown to recognise its native ligand at anything other than relatively high concentrations and/or were not stable.

In WO 99/60120, a soluble TCR is described which is correctly folded so that it is capable of recognising its native ligand, is stable over a period of time, and can be produced in reasonable quantities. This TCR comprises a TCR α or γ chain extracellular domain dimerised to a TCR β or δ chain extracellular domain respectively, by means of a pair of C-terminal dimerisation peptides, such as leucine zippers. This strategy of producing TCRs is generally applicable to all TCRs.

Reiter et al, Immunity, 1995, 2:281-287, details the construction of a soluble molecule comprising disulphide-stabilised TCR α and β variable domains, one of which is linked to a truncated form of Pseudomonas exotoxin (PE38). One of the stated reasons for producing this molecule was to overcome the inherent instability of single- chain TCRs. The position of the novel disulphide bond in the TCR variable domains was identified via homology with the variable domains of antibodies, into which these have previously been introduced (for example see Brinkmann, et al. (1993), Proc. Natl. Acad. Sci. USA 90: 7538-7542, and Reiter, et al. (1994) Biochemistry 33: 5451- 5459). However, as there is no such homology between antibody and TCR constant domains, such a technique could not be employed to identify appropriate sites for new inter-chain disulphide bonds between TCR constant domains.

As mentioned above Shusta et al (2000) Nature Biotechnology 18: 754-759 report using single-chain 2 C TCR constructs in yeast display experiments. The principle of displaying scTCRs on phage particles has previously been discussed. For example, WO 99/19129 details the production of scTCRs, and summarise a potential method for the production of phage particles displaying scTCRs of the Vα-Linker-Vβ Cβ format. However, this application contains no exemplification demonstrating the production of said phage particles displaying TCR. The application does however refer to a co- pending application:

"The construction of DNA vectors including a DNA segment encoding a sc-TCR molecules fused to a bacteriophage coat protein (gene II or gene VIII) have been described in said pending U.S. application No. 08/813,781."

Furthermore, this application relies on the ability of anti-TCR antibodies or super- antigen MHC complexes to recognise the soluble, non-phage displayed, scTCRs produced to verify their correct conformation. Therefore, true peptide-MHC binding specificity of the scTCRs, in any format, is not conclusively demonstrated. Finally, a further study (Onda et al, (1995) Molecular Immunology 32 (17-18) 1387-

1397) discloses the phage display of two murine TCR α chains in the absence of their respective β chains. This study demonstrated that phage particles displaying one of the TCR α chains (derived from the Al.l murine hybridoma) bound preferentially to the same peptides immobilised in microtitre wells that the complete TCR would normally respond to when there were presented by the murine Class I MHC I-A^d.

Display Methods

It is often desirable to present a given peptide or polypeptide on the surface of a nucleoprotein particle. Such particles may serve as purification aids for the peptide or polypeptide (since the particles carrying the peptide or polypeptide may be separated from unwanted contaminants by sedimentation or other methods). They may also serve as particulate vaccines, the immune response to the surface displayed peptide or polypeptide being stimulated by the particulate presentation. Protein p24 of the yeast retrotransposon, and the hepatitis B surface coat protein are examples of proteins which self assemble into particles. Fusion of the peptide or polypeptide of interest to these particle-forming proteins is a recognised way of presenting the peptide or polypeptide on the surface of the resultant particles.

However, particle display methods have primarily been used to identify proteins with desirable properties such as enhanced expression yields, binding and/or stability characteristics. These methods involve creating a diverse pool or 'library' of proteins or polypeptides expressed on the surface of nucleoprotein particles. These particles have two key features, firstly each particle presents a single variant protein or polypeptide, and secondly the genetic material encoding the expressed protein or polypeptide is associated with that of the particle. This library is then subjected to one or more rounds of selection. For example, this may consist of contacting a ligand with a particle-display library of mutated receptors and identifying which mutated receptors bind the ligand with the highest affinity. Once the selection process has been completed the receptor or receptors with the desired properties can be isolated, and their genetic material can be amplified in order to allow the receptors to be sequenced.

These display methods fall into two broad categories, in-vitro and in-vivo display.

All in-vivo display methods rely on a step in which the library, usually encoded in or with the genetic nucleic acid of a replicable particle such as a plasmid or phage replicon is transformed into cells to allow expression of the proteins or polypeptides.

(Plϋckthun (2001) Adv Protein Chem 55 367-403). There are a number of replicon/host systems that have proved suitable for in-vivo display of protein or polypeptides. These include the following

Phage / bacterial cells plasmid / CHO cells

Vectors based on the yeast 2μm plasmid / yeast cells bacculovirus / insect cells plasmid / bacterial cells In-vivo display methods include cell-surface display methods in which a plasmid is introduced into the host cell encoding a fusion protein consisting of the protein or polypeptide of interest fused to a cell surface protein or polypeptide. The expression of this fusion protein leads to the protein or polypeptide of interest being displayed on the surface of the cell. The cells displaying these proteins or polypeptides of interest can then be subjected to a selection process such as FACS and the plasmids obtained from the selected cell or cells can be isolated and sequenced. Cell surface display systems have been devised for mammalian cells (Higuschi (1997) J Immunol. Methods 202 193-204), yeast cells (Shusta (1999) J Mol Biol 292 949-956) and bacterial cells (Sameulson (2002) J. Biotechnol 96 (2) 129-154).

Numerous reviews of the various in-vivo display techniques have been published. For example, (Hudson (2002) Expert Opin Biol Ther (2001) 1 (5) 845-55) and (Schmitz (2000) 21 (Supp A) S106-S112).

In-vitro display methods are based on the use of ribosomes to translate libraries of mRNA into a diverse array of protein or polypeptide variants. The linkage between the proteins or polypeptides formed and the mRNA encoding these molecules is maintained by one of two methods. Conventional ribosome display utilises mRNA sequences that encode a short (typically 40-100 amino acid ) linker sequence and the protein or polypeptide to be displayed. The linker sequence allow the displayed protein or polypeptide sufficient space to re-fold without being sterically hindered by the ribosome. The mRNA sequence lacks a 'stop' codon, this ensures that the expressed protein or polypeptide and the RNA remain attached to the ribosome particle. The related mRNA display method is based on the preparation of mRNA sequences encoding the protein or polypeptide of interest and DNA linkers carrying a puromycin moiety. As soon as the ribosome reaches the mRNA/DNA junction translation is stalled and the puromycin forms a covalent linkage to the ribosome. For a recent review of these two related in-vitro display methods see (Amstutz (2001) Curr Opin Biotechnol 12 400-405). Particularly preferred is the phage display technique which is based on the ability of bacteriophage particles to express a heterologous peptide or polypeptide fused to their surface proteins. (Smith (1985) Science 217 1315-1317). The procedure is quite general, and well understood in the art for the display of polypeptide monomers. However, in the case of polypeptides that in their native form associate as dimers, only the phage display of antibodies appears to have been thoroughly investigated.

For monomeric polypeptide display there are two main procedures:

Firstly (Method A) by inserting into a vector (phagemid) DNA encoding the heterologous peptide or polypeptide fused to the DNA encoding a bacteriophage coat protein. The expression of phage particles displaying the heterologous peptide or polypeptide is then carried out by transfecting bacterial cells with the phagemid, and then infecting the transformed cells with a 'helper phage'. The helper phage acts as a source of the phage proteins not encoded by the phagemid required to produce a functional phage particle.

Secondly (Method B), by inserting DNA encoding the heterologous peptide or polypeptide into a complete phage genome fused to the DNA encoding a bacteriophage coat protein. The expression of phage particles displaying the heterologous peptide or polypeptide is then carried out by infecting bacterial cells with the phage genome. This method has the advantage of the first method of being a

'single-step' process. However, the size of the heterologous DNA sequence that can be successfully packaged into the resulting phage particles is reduced. Ml 3, T7 and Lambda are examples of suitable phages for this method.

A variation on (Method B) the involves adding a DNA sequence encoding a nucleotide binding domain to the DNA in the phage genome encoding the heterologous peptide be displayed, and further adding the corresponding nucleotide binding site to the phage genome. This causes the heterologous peptide to become directly attached to the phage genome. This peptide/genome complex is then packaged into a phage particle which displays the heterologous peptide. This method is fully described in WO 99/11785. The phage particles can then be recovered and used to study the binding characteristics of the heterologous peptide or polypeptide. Once isolated, phagemid or phage DNA can be recovered from the peptide- or polypeptide-displaying phage particle, and this DNA can be replicated via PCR. The PCR product can be used to sequence the heterologous peptide or polypeptide displayed by a given phage particle.

The phage display of single-chain antibodies and fragments thereof, has become a routine means of studying the binding characteristics of these polypeptides. There are numerous books available that review phage display techniques and the biology of the bacteriophage. (See, for example, Phage Display - A Laboratory Manual, Barbas et al, and (2001) Cold Spring Harbour Laboratory Press).

A third phage display method (Method C) relies on the fact that heterologous polypeptides having a cysteine residue at a desired location can be expressed in a soluble form by a phagemid or phage genome, and caused to associate with a modified phage surface protein also having a cysteine residue at a surface exposed position, via the formation of a disulphide linkage between the two cysteines. WO 01/ 05950 details the use of this alternative linkage method for the expression of single-chain antibody- derived pep tides.

Native TCR Libraries

To date there have been no published accounts of diverse libraries of nucleoproteins displaying polypeptides comprising native TCR variable domains.

Brief Description of the Invention

Native TCR's are heterodimers which have lengthy fransmembrane domains which are essential to maintain their stability as functional dimers. As discussed above, TCRs are useful for research and therapeutic purposes in their soluble forms so display of the insoluble native form has little utility. There have been a number of publications relating to various means of TCR display which are summarised below:

WO 99/18129 contains the statement: "DNA constructs encoding the sc-TCR fusion proteins can be used to make a bacteriophage display library in accordance with methods described in pending U.S. application Serial No. 08/813.781 filed on March 7, 1997, the disclosure of which is incorporated herein by reference.", but no actual description of such display is included in this application. However, The inventors of this application published a paper (Weidanz (1998) J Immunol Methods 221 59-76) that demonstrates the display of two murine scTCRs on phage particles.

WO 01/62908 discloses methods for the phage display of scTCRs and scTCR/ Ig fusion proteins. However, the functionality (specific pMHC binding) of the constructs disclosed was not assessed.

A retrovirus-mediated method for the display of diverse TCR libraries on the surface of immature T cells has been demonstrated for a murine TCR. The library of mutated TCRs displayed of the surface of the immature T cells was screened by flow cytometry using pMHC teframers, and this lead to the identification TCR variants that were either specific for the cognate pMHC, or a variant thereof. (Helmut et al, (2000) PNAS 97 (26) 14578-14583)

The inventors' co-pending application (PCT/GB2003/04636) is appended hereto as Appendix A. This co-pending application is based in part on the finding that functional single chain and dimeric TCRs can be expressed as surface fusions to nucleoprotein particles, and makes available nucleoprotein particles displaying alpha/beta-analogue and gamma delta-analogue scTCR and dTCR constructs. The nucleoprotein particles on which the TCRs are displayed are preferably phage particles but also include self- aggregating particle-forming proteins, virus-derived and ribosome particles. Such nucleoprotein particle-displayed TCRs are useful for purification and screening purposes, particularly as a diverse library of particle displayed TCRs for biopanning to identify TCRs with desirable characteristics such as high affinity for the target MHC- peptide complex. In the latter connection, particle-displayed scTCRs may be useful for identification of the desired TCR, but that information may be better applied to the construction of analogous dimeric TCRs for ultimate use in therapy. The invention also includes high affinity TCRs identifiable by these methods.

In the present invention the displayed TCRs comprise native variable domains attached to the nucleoprotein of choice, which is preferably a ribosome or phage particle, and most preferably a phage particle.

The present invention make available for the first time libraries containing a diverse array of native TCR variable domains displayed on the surface of nucleoproteins. The present invention also provides methods using said libraries for the isolation of a ligand or ligands of a target peptide-MHC (pMHC) complex or CD 1 -antigen complex.

Detailed Description of the Invention

Native TCR Libraries The first step in the production of a diverse library comprising native TCR α and/or native TCR β variable domains is the isolation of DNA encoding a range of native TCR variable domains. As is known to those skilled in the art suitable primers for this isolation step can be prepared by reference to the sequences of the Vα genes and Cα gene of TCR α chains, and the Vβ genes and Cβ genes of TCR β chain respectively. The gene sequence of all known TCR C and V genes can be found in The T Cell

Receptor Factsbook (2^nd Edition), 2001, Lefranc and LeFranc, Academic Press.

In order to provide a source of DNA containing TCRα variable domain Vα and Jα genes, and TCRβ variable domain Vβ, Jβ and Dβ genes peripheral blood mononucleate cells (PBMCs) are isolated from blood samples using techniques known to those skilled in the art. cDNA is then prepared from the PMBCs and used as the template for polymerase chain reaction (PCR)-based amplification of the TCR variable domains. The isolated PCR products comprising the variable domain Open Reading Frames (ORFs) are then cloned into expression systems that allow the gene product of the various variable genes to be expressed in a functional form on the surface of a nucleoprotein. Appendix A hereto describes such expression systems and methods.

In one broad aspect the invention provides a diverse library of nucleoproteins each displaying on its surface a polypeptide comprising a native TCR variable domain sequence or native TCR β variable domain sequence, in which library the diversity resides in the variety of native TCR α variable domain or native TCR β variable domain sequences displayed.

In a preferred embodiment of the invention said nucleoproteins are phage or ribosome particles. Appendix A appended hereto provides a detailed description of the methods required to display single-chain TCRs (scTCRs) on ribosomes.

In a further embodiment of the invention at least some of the displayed polypeptides comprise a native TCR β variable domain sequence located N-terminal to part of a TCR β chain which includes all or part of a TCR β constant domain sequence except the transmembrane domain thereof.

In an alternative embodiment of the invention at least some of the displayed polypeptides comprise a native TCR α variable domain sequence located N-terminal to part of a TCR α chain which includes all or part of a TCR α constant domain sequence except the transmembrane domain thereof.

As stated above, one study (Onda et al, (1995) Molecular Immunology 32 (17-18) 1387-1397) discloses binding of phage particles displaying a murine TCR α chain in the absence of the respective TCR β chain to a peptide immobilised in microtitre wells that the complete TCR would normally respond to when there were presented by the murine Class I MHC I-A^d. However, the present invention provides the first indication that a phage particle displaying a polypeptide comprising a TCR α or β variable domain, absent an associated second TCR variable domain can bind a pMHC.

hi another embodiment of the invention a first displayed polypeptide comprising the native TCR α variable domain sequence or native TCR β variable domain sequence is associated with a second polypeptide comprising a native TCR variable domain sequence located N-terminal to part of a TCR chain which includes all or part of a TCR constant domain sequence except the transmembrane domain thereof, said association being maintained at least in part by a disulfide bond which has no equivalent in native αβ T cell receptors between introduced cysteine residues in the

TCR constant domain sequences of the first and second polypeptides.

In a specific embodiment of the invention the first and second polypeptides are linked by a disulfide bond between cysteine residues substituted for Thr 48 of exon 1 of TRAC*01 and Ser 57 of exon 1 of TRBC1*01 or TRBC2*01 or the non-human equivalent thereof.

A further embodiment of the invention is provided by phage particles or ribosomes on which the first displayed polypeptide is linked by a peptide bond at its C-terminus to a surface exposed amino acid residue of the phage particle or ribosome.

Methods of identifying pMHC or CDl-antigen ligands

The TCR libraries described above are suitable for the identification of pMHC or CDl-antigen ligands. The presence and/or expression level of certain pMHC on the surface of a given cell is related to the disease-state of that cell. For example, certain cancers express the NY-ESO-1 protein and this leads to MHC molecules on the surface of such cancer cells presenting peptides derived from the NY-ESO-1 protein. The SLLMWITQC-HLA-A*0201 complex is an example of such a cancer-related pMHC which is expressed by NY-ESO-l⁺ cancer cells. The isolation of ligands, such as TCRs, which bind to these disease and cancer-related pMHC molecules, provides a basis for targeting moieties capable of delivering therapeutic and/or diagnostic agents agents to the diseased or cancerous cells.

One embodiment of the invention provides a method of identifying a ligand or ligands of a target peptide-MHC (pMHC) complex or CDl-antigen, which method comprises

(a) contacting several members of a library of the present invention in parallel with the target pMHC or CDl-antigen and identifying members which bind to the target pMHC or CDl-antigen, (b) contacting members identified in step (a) in series with the target pMHC or CDl-antigen, and assessing their affinities for the pMHC or CDl-antigen, (c) selecting one or more members having the desired affinity assessed in step (b) and determining the variable domain sequence(s) of the displayed polypeptide(s), (d) creating soluble form TCRs incorporating the thus-determined variable domain sequence(s), and optionally, (e) determining or redetermining as the case may be, the affinities and/or the off-rate for the target pMHC or CDl-antigen of these TCRs, and optionally,

(f) selecting one or more TCRs having the desired affinity and/or off-rate determined in step (b) or optional step (e) as the desired ligand(s).

A specific embodiment of the invention is provided wherein (a) above involves contacting several members of a library of the invention in parallel with an antigen- presenting cell which presents the target pMHC or CDl-antigen and identifying members which bind to the cell-presented target pMHC or CDl-antigen.

A further specific embodiment of the invention is provided a method comprising the following steps prior to step (a) (i) several members of the library are contacted in parallel with an antigen- presenting cell which does not present the target pMHC or CDl-antigen

(ii) members identified in step (i) as not binding to the antigen presenting cell are used as the several members of the library referred to in step (a).

The above steps (i) and (ii) are useful in that they provide an initial check on the specificity of the library members selected by the screening method. This may result in the elimination of some cross-reactive library members that could have been selected as "hits" absent steps (i) and (ii).

A further specific embodiment of the invention is provided wherein the antigen- presenting cell which presents the target pMHC or CDl-antigen is caused to present said target pMHC or CDl-antigen by peptide-pulsing or antigen-pulsing.

Another specific embodiment of the invention is provided wherein the antigen- presenting cell which presents the target pMHC or CDl-antigen is a non-peptide- pulsed antigen presenting cell. Any disease-related or cancerous antigen-presenting cell may be used in this regard.

The above methods and nucleoprotein library of the invention enable identification of αβ dTCRs, TCR α-chains, TCR β-chains, TCR αα homodimers and/or TCR ββ homodimers that will bind to the target pMHC or CDl-antigen

As will be obvious to those skilled in the art the above methods will be equally applicable to the identification of ligands from the library members to any other complexes that are capable of being bound by the library members.

A suitable method for determining the affinity and/or off-rate for the target pMHC is/are determination by Surface Plasmon Resonance. Example 5 herein provides a detailed description of how such measurements are carried out. The Displayed Polypeptides

The following are the preferred designs for the display of polypeptides comprising native TCR α and/or native TCR β variable domains by association with nucleoproteins. It should be noted that these designs are equally suited for use as soluble TCRs absent the associated nucleoprotein.

Displayed dimeric polypeptides

In a further preferred embodiment of the invention the phage-displayed αβ dimeric

TCR polypeptides comprise a first polypeptide wherein a sequence corresponding to a TCR α chain variable domain sequence is fused to the N terminus of a sequence corresponding to a TCR α chain constant domain extracellular sequence, and a second polypeptide wherein a sequence corresponding to a TCR β chain variable domain sequence fused to the N terminus a sequence corresponding to a TCR β chain constant domain extracellular sequence, the first and second polypeptides being linked by a disulfide bond which has no equivalent in native αβ T cell receptors, and wherein one of said first or second polypeptides is linked by a peptide bond at its C-terminus to a surface exposed amino acid residue of the phage particle.

In a specific embodiment of the invention the first and second TCR polypeptides are linked by a disulfide bond between cysteine residues substituted for Thr 48 of exon 1 of TRAC*01 and Ser 57 of exon 1 of TRBC1*01 or TRBC2*01 or the non-human equivalent thereof, and one of said first or second polypeptides are linked by a peptide bond at its C-terminus to a surface exposed amino acid residue of the phage particle.

The residues for mutation to cysteine in order to form the non-native disulfide interchain bind are identified using ImMunoGeneTics (LMGT) nomenclature. (The T cell Receptor Factsbook 2^nd Edition (2001) LeFranc and Lefranc, Academic Press) WO 03/020763 provides a detailed description of the methods required to introduce the specified non-native disulfide interchain bond and alternative residues between which it may be sited. Displayed single chain polypeptides

In a further preferred embodiment of the invention the phage-displayed αβ scTCRpolypeptide may be, for example, one which has a first segment constituted by an amino acid sequence corresponding to a TCR α variable domain sequence fused to the N terminus of an amino acid sequence corresponding to a TCR α chain constant domain extracellular sequence, a second segment constituted by an amino acid sequence corresponding to a TCR β chain variable domain fused to the N terminus of an amino acid sequence corresponding to TCR β chain constant domain extracellular sequence, a linker sequence linking the C terminus of the first segment to the N terminus of the second segment, or vice versa, and a disulfide bond between the first and second chains, said disulfide bond being one which has no equivalent in native αβ T cell receptors.

Alternatively, the displayed scTCR polypeptide may be one which has a first segment constituted by an amino acid sequence corresponding to a TCR α chain variable domain a second segment constituted by an amino acid sequence corresponding to a TCR β chain variable domain sequence fused to the N terminus of an amino acid sequence corresponding to a TCR β chain constant domain extracellular sequence, and a linker sequence linking the C terminus of the first segment to the N terminus of the second segment, or

one which has a first segment constituted by an amino acid sequence corresponding to a TCR β chain variable domain a second segment constituted by an amino acid sequence corresponding to a TCR α chain variable domain sequence fused to the N terminus of an amino acid sequence corresponding to a TCR α chain constant domain extracellular sequence, and a linker sequence linking the C terminus of the first segment to the N terminus of the second segment

dTCR Polypeptide Pair and scTCR Polypeptide

The constant domain extracellular sequences present in the displayed scTCR polypeptides or dTCR polypeptides preferably correspond to those of a human TCR, as do the variable domain sequences. However, the correspondence between such sequences need not be 1 : 1 on an amino acid level. N- or C-truncation, and/or amino acid deletion and/or substitution relative to the corresponding human TCR sequences are acceptable. In particular, because the constant domain extracellular sequences present in the first and second segments are not directly involved in contacts with the ligand to which the scTCR or dTCR binds, they may be shorter than, or may contain substitutions or deletions relative to, extracellular constant domain sequences of native

TCRs.

The constant domain extracellular sequence present in one of the displayed dTCR polypeptide pair, or in the first segment of a displayed scTCR polypeptide may include a sequence corresponding to the extracellular constant Ig domain of a TCR α chain, and/or the constant domain extracellular sequence present in the other member of the pair or second segment may include a sequence corresponding to the extracellular constant Ig domain of a TCR β chain.

In one embodiment of the invention, one member of the displayed dTCR polypeptide pair, or the first segment of the displayed scTCR polypeptide, corresponds to substantially all the variable domain of a TCR α chain fused to the N terminus of substantially all the extracellular domain of the constant domain of an TCR α chain; and/or the other member of the pair or second segment corresponds to substantially all the variable domain of a TCR β chain fused to the N terminus of substantially all the extracellular domain of the constant domain of a TCR β chain.

In another embodiment, the constant domain extracellular sequences present in the displayed dTCR polypeptide pair, or first and second segments of the displayed scTCR polypeptide, correspond to the constant domains of the α and β chains of a native TCR truncated at their C termini such that the cysteine residues which form the native inter-chain disulfide bond of the TCR are excluded. Alternatively those cysteine residues may be substituted by another amino acid residue such as serine or alanine, so that the native disulfide bond is deleted. In addition, the native TCR β chain contains an unpaired cysteine residue and that residue may be deleted from, or replaced by a non-cysteine residue in, the β sequence of the scTCR of the invention.

In one particular embodiment of the invention, the TCR α and β chain variable domain sequences present in the displayed dTCR polypeptide pair, or first and second segments of the displayed scTCR polypeptide, may together correspond to the functional variable domain of a first TCR, and the TCR α and β chain constant domain extracellular sequences present in the dTCR polypeptide pair or first and second . segments of the scTCR polypeptide may correspond to those of a second TCR, the first and second TCRs being from the same species. Thus, the a and β chain variable domain sequences present in dTCR polypeptide pair, or first and second segments of the scTCR polypeptide, may correspond to those of a first human TCR, and the α and β chain constant domain extracellular sequences may correspond to those of a second human TCR. For example, A6 Tax sTCR constant domain extracellular sequences can be used as a framework onto which heterologous α and β variable domains can be fused.

In one particular embodiment of the invention, the TCR α and β chain variable domain sequences present in the displayed dTCR polypeptide pair or first and second segments of the displayed scTCR polypeptide may together correspond to the functional variable domain of a first human TCR, and the TCR α and β chain constant domain extracellular sequences present in the dTCR polypeptide pair or first and second segments of the scTCR polypeptide may correspond to those of a second non- human TCR, Thus the α and β chain variable domain sequences present dTCR polypeptide pair or first and second segments of the scTCR polypeptide may correspond to those of a first human TCR, and the α and β chain constant domain extracellular sequences may correspond to those of a second non-human TCR. For example, murine TCR constant domain extracellular sequences can be used as a framework onto which heterologous human α and β TCR variable domains can be fused.

Linker in the scTCR Polypeptide

For scTCR-displaying nucleoproteins of the present invention, a linker sequence links the first and second TCR segments, to form a single polypeptide strand. The linker sequence may, for example, have the formula -P-AA-P- wherein P is proline and AA represents an amino acid sequence wherein the amino acids are glycine and serine.

For the scTCR displayed by nucleoproteins of the present invention to bind to a ligand, MHC-peptide complex in the case of αβ TCRs, the first and second segments are preferably paired so that the variable domain sequences thereof are orientated for such binding. Hence the linker should have sufficient length to span the distance between the C temiinus of the first segment and the N terminus of the second segment, or vice versa. On the other hand excessive linker length should preferably be avoided, in case the end of the linker at the N-terminal variable domain sequence blocks or reduces bonding of the scTCR to the target ligand.

For example, in the case where the constant domain extracellular sequences present in the first and second segments correspond to the constant domains of the α and β chains of a native TCR truncated at their C termini such that the cysteine residues which form the native interchain disulfide bond of the TCR are excluded, and the linker sequence links the C terminus of the first segment to the N terminus of the second segment.

The linker sequence may consist of, for example, from 26 to 41 amino acids, preferably 29, 30, 31 or 32 amino acids, or 33, 34, 35 or 36 amino acids. Particular linkers have the formula -PGGG-(SGGGG)₅-P- (SEQ LD NO: 1) and -PGGG-

(SGGGG)₆-P-(SEQ ID NO:2) wherein P is proline, G is glycine and S is serine.

Inter-chain Disulfide bond

A principle characterising feature of the preferred dTCR polypeptides and scTCR polypeptides displayed by nucleoproteins of the present invention is a disulfide bond between the constant domain extracellular sequences of the dTCR polypeptide pair or first and second segments of the scTCR polypeptide. That bond may correspond to the native inter-chain disulfide bond present in native dimeric αβ TCRs, or may have no counterpart in native TCRs, being between cysteines specifically incorporated into the constant domain extracellular sequences of dTCR polypeptide pair or first and second segments of the scTCR polypeptide. In some cases, both a native and a non-native disulfide bond may be desirable.

As stated above, WO 03/020763 provides a detailed description of the methods required to introduce the specified non-native disulfide interchain bond and alternative residues between which it may be sited. Required to prepare a diverse library of polypeptides comprising native TCR α and/or β variable domains, are nucleic acids encoding (a) one chain or both chains of a dTCR polypeptide pair fused to a nucleic acid sequence encoding a protein capable of forming part of the surface of a nucleoprotein particle; or (b) nucleic acid encoding a scTCR polypeptide fused to a nucleic acid sequence encoding a protein capable of forming part of the surface of a nucleoprotein particle.

For expression of the nucleoproteins, host cells may be used transformed with an expression vector comprising nucleic acid encoding (a) or (b).

Preferably the expression system comprises phagemid or phage genome vectors expressing nucleic acids (a) or (b). Preferably these phagemid or phage genome vectors is (are) those which encode bacteriophage gill or gVIII coat proteins.

Transformed cells are incubated to allow the expression of nucleoprotein particles displaying the polypeptides comprising native TCR α and/or native TCR β variable domains. These particles can then be used in assays to identify TCR variants which bind to a given pMHC or CDl-antigen complex with the desired affinity and or off- rate characteristics. Any particles that possess the desired characteristics under investigation can then be isolated. The DNA encoding these TCRs can then be amplified by PCR and the sequence determined.

It is known that high expression levels of an exogenous polypeptide may be toxic to the host cell. In such cases, either a host strain which is more tolerant of the exogenous polypeptide must be found, or the expression levels in the host cell must be limited to a level which is tolerated. For example (Beekwilder et al, (1999) Gene 228 (1-2) 23- 31) report that only mutated forms of a potato protease inhibitor (PI2) which contained deletions or amber stop codons would be successfully selected from a phage display library. There are several strategies for limiting the expression levels of an exogenous polypeptide from a given expression system in a host which may be suitable for the limiting the expression levels of a scTCR, or one, or both TCR chains of a dTCR . These strategies are described in Appendix A

Correct pairing of scTCR polypeptide variable domain sequences after expression is preferably assisted by an introduced disulfide bond in the extracellular constant domain of the scTCR. Without wanting to be limited by theory, the novel disulfide bond is believed to provide extra stability to the scTCR during the folding process and thereby facilitating correct pairing of the first and second segments.

Also as mentioned above, for dTCR phage display, one of the dTCR polypeptide pair is expressed as if it were eventually to be displayed as a monomeric polypeptide on the phage, and the other of the dTCR polypeptide pair is co-expressed in the same host cell. As the phage particle self assembles, the two polypeptides self associate for display as a dimer on the phage. Again, in the preferred embodiment of this aspect of the invention, correct folding during association of the polypeptide pair is assisted by a disulfide bond between the constant sequences. Further details of a procedure for phage display of a dTCR having an interchain disulfide bond appear in the Examples contained within Appendix A.

As an alternative, the phage displaying the first chain of the dTCR may be expressed first, and the second chain polypeptide may be contacted with the expressed phage in a subsequent step, for association as a functional dTCR on the phage surface.

One in-vitro TCR display method for biopanning to identify ligands to a given pMHC or CDl-antigen is ribosomal display. Firstly, a DNA library is constructed that encodes a diverse array of native TCR α variable domain sequences or native TCR β variable domain sequences using the techniques discussed above. The DNA library is then contacted with RNA polymerase in order to produce a complementary mRNA library. Optionally, for mRNA display techniques the mRNA sequences can then be ligated to a DNA sequence comprising a puromycin binding site. These genetic constructs are then contacted with ribosomes in-vitro under conditions allowing the translation of the scTCR polypeptide or the first polypeptide of the dTCR polypeptide pair. In the case of the dTCR polypeptide, the second of the polypeptide pairs is separately expressed and contacted with the ribosome-displayed first polypeptide, for association between the two, preferably assisted by the formation of the disulphide bond between constant domains. Alternatively, mRNA encoding both chains of the dTCR polypeptide may be contacted with ribosomes in-vitro under conditions allowing the translation of the TCR chains such that a ribosome displaying a dTCR polypeptide pair is formed. These scTCR- or dTCR-disp laying ribosomes can then used for screening or in assays to identify desired ligands. Any particles that display desired ligands can then be isolated. The mRNA encoding these TCRs can then converted to the complementary DNA sequences using reverse transcriptase. This DNA can then be amplified by PCR and the sequence determined.

scTCR polypeptides or dTCR polypeptide pairs of the present invention may be displayed on nucleoprotein particles, for example phage particles, preferably filamentous phage particles, by, for example, the following two means:

(i) The C-terminus of one member of the dTCR polypeptide pair, or the C-terminus of the scTCR polypeptide, or the C-terminus of a short peptide linker attached to the C- terminus of either, can be directly linked by a peptide bond to a surface exposed residue of the nucleoprotein particle. For example, the said surface exposed residue is preferably at the N-terminus of the gene product of bacteriophage gene III or gene VIII; and

(ii) The C-terminus of one member of the dTCR polypeptide pair, or the C-terminus of the scTCR polypeptide, or the C-terminus of a short peptide linker attached to the C- terminus of either, is linked by a disulfide bond to a surface exposed cysteine residue of the nucleoprotein particle via an introduced cysteine residue. For example, the said surface exposed residue is again preferably at the N-terminus of the gene product of bacteriophage gene III or gene VIII.

Ml 3 and fl are examples of bacteriophages that express gene III and gene VIII gene products.

Method (i) above is preferred. In the case of an scTCR polypeptide, nucleic acid encoding the TCR may be fused to nucleic acid encoding the particle forming protein or a surface protein of the replicable particle such as a phage or cell. Alternatively, nucleic acid representing mRNA but without a stop codon, or fused to puromycin

RNA may be translated by ribosome such that the TCR remains fused to the ribosome particle, hi the case of a dTCR, nucleic acid encoding one chain of the TCR may be fused to nucleic acid encoding the particle forming protein or a cell surface protein of the replicable particle such as a phage or cell, and the second chain of the TCR polypeptide pair may be allowed to associate with the resultant expressed particle displaying the first chain. Proper functional association of the two chains may be assisted by the presence of cysteines in the constant domain of the two chains which are capable of forming an interchain disulfide bond, as more fully discussed below.

Additional Aspects

A ligand (which preferably is constituted by constant and variable sequences corresponding to human sequences) isolated by the method of the invention may be provided in substantially pure form, or as a purified or isolated preparation. For example, it may be provided in a form which is substantially free of other proteins.

The invention also provides a method for obtaining a ligand selected by the method of this invention, which method comprises incubating a host cell comprising nucleic acid encoding that ligand, or of the individual polypeptides of the pair comprising the ligand, under conditions causing expression of the ligand or ligand component, and then purifying said ligand or hgand component. dTCR polypeptide hgands can then be formed by refolding the purified ligand components as described in Example 4. Preferred features of each aspect of the invention are as for each of the other aspects mutatis mutandis. The prior art documents mentioned herein are incorporated to the fullest extent permitted by law.

Examples

The invention is further described in the following examples, which do not limit the scope of the invention in any way.

Figure 1 details the DNA sequence of the pEX922_lG4 plasmid.

Figure 2 is a plasmid map of pEX922_lG4

Example 1- Naϊve Library construction

Overview:

Comprehensive primer sets were designed with reference to the known human sequences described in the "The T Cell Receptor Facts Book, Lefranc & Lefranc, Publ. Academic Press 2001" such that the Vα and Vβ non-immunised native chain repertoire could be amplified by PCR from cDNA template prepared from total RNA isolated from the peripheral blood cells of 20 healthy individuals with serological specificity for HLA-A2. Pooled Vα and pooled Vβ amplification products were subjected to a second round of PCR using common flanking primers in order to introduce sufficiently long constant domain tags. Separate Vα and Vβ fragment pools comprising highly diverse native TCR Vα and Vβ DNA sequences with a portion of the corresponding constant domain DNA fused at the 3 ' ends thereof were thus generated. These pools were randomly assembled into Vα /Vβ pairs using a PCR splicing reaction employing a common intervening fragment derived from the destination phagemid vector. The spliced products were cloned into the destination phagemid vector and electroporated into E. coli strain TGI in order to generate a large library of vector clones comprising randomly paired native TCR chains capable of being displayed on phage.

Purification of Total RNA from Buffy Coat PBMCs:

PBMC samples stored in liquid nitrogen (~10⁶ cells) were gently thawed and cells were pelleted in a microfuge at 2000rpm for 2 min. Supernatants were removed and Total RNA was isolated from cell pellets using QIAshredder spin modules and an RNeasy kit (Qiagen) with adherence to the manufacturers guidelines. The total RNA from each PBMC sample was recovered in 50μl of RNAse free water and it's integrity checked by agarose gel electrophoresis.

cDNA synthesis:

Total RNA from each of 20 samples was pooled (5μl per sample) and cDNA was synthesised as follows: A reaction containing 20μl 0.5μg/μl oligo(dT)_12-18 (Invitrogen), lOOμl pooled RNA, 40μl dNTPs (5mM each dNTP; Abgene) and 80μl sterile water was set up on ice prior to being heated to 65°C for 5 min. The reaction was rapidly chilled on ice and the contents collected by brief centrifugation before adding 80μl 5 x first strand synthesis buffer, 40μl 0.1M DTT, and 20μl sterile water. The solution was mixed by gentle pipetting and incubated at 42°C for 2 min prior to the addition of 20μl Superscript II reverse transcriptase (Invitrogen). The cDNA synthesis reaction was mixed by gentle pipetting and allowed to proceed for lhr at

42°C followed by 15 min at 70°C. Following a brief centrifugation to collect the tube contents the cDNA was cooled on ice and stored at -20°C until use.

PCR1 — Amplification of V and Vβ native sequences from cDNA: In order to amplify individual Vα and Vβ class repertoires from cDNA, discrete PCR reactions were assembled as follows: 39μl water, 5μl Expand polymerase buffer 2 (Roche), 2μl dNTPs (20mM combined stock; Abgene), lμl TRAV_R or TRBV_R reverse primer (lOμM stock), lμl class-specific forward primer (lOμM stock), lμl cDNA, and lμl Expand DNA polymerase (Roche) to give a total volume of 50μl.

PCR cycling was as follows: 94°C for 2 min followed by 35 cycles of 94°C for 15 sec, 55°C for 30 sec, 72°C for 45 sec + 1 sec/cycle. A final extension of 10 min at 72°C was allowed.

The 43 Vα class-specific PCR reactions were pooled as were the 37 Vβ (37) and electrophoresed through a preparative 1.5% agarose gel. Bands of the correct size were excised and gel purified using a Qiagen MinElute kit.

PCR2 — Nested tag modification of Va and Vβ products from PCR1:

Prior to splicing of the Vα and Vβ products from PCR1, the existing terminal tags were modified by limited PCR cycling in order to generate ends that would allow the complementary annealing of a vector-encoded joining fragment. The reactions were assembled as follows: 39.5μl water, 5μl pfu turbo polymerase lOx buffer (Stratagene), 2μl dNTPs (20mM combined stock; Abgene), lμl TRAVFE or TRBVSF forward primer (lOμM stock), lμl TRAVSR or TRBV_R reverse primers (lOμM stock), lμl

Vα or Vβ amplicons from PCR1 , and 0.5 μl pfu Turbo DNA polymerase (Stratagene) to give a total volume of 50μl. PCR cycling was as follows: 94°C for 2 min followed by 20 cycles of 94°C for 15 sec, 55°C for 45 sec, 72°C for 45 sec. A final extension of 10 min at 72°C was allowed. The end modified Vα and Vβ PCR reactions were electrophoresed through a preparative 1.5% agarose gel. Bands of the correct size were excised and gel purified using a Qiagen MinElute kit.

PCR3 — Generation of the vector-encoded joining fragment: The template for the reaction was pEX922_lG4 (FIG 1) containing irrelevant TCR alpha and beta chain DNA sequences. These sequences contain codons encoding the cysteine residues required for eventual formation of an introduced disulfide bond between constant domains of a soluble heterodimeric TCR (as described in, for example WO 03/020763)

A fragment of plasmid pEX922_lG4 (shown in bold in FIG 1) comprising the sequence separating the Vα domain from the start of the Vβ chain, and comprising the introduced TCR α chain cysteine codon referred to above (both introduced cysteine codon are shown highlighted in FIG 1) was amplified by limited PCR. The reaction was assembled as follows: 39.5μl water, 5μl pfu turbo polymerase lOx buffer (Stratagene), 2μl dNTPs (20mM combined stock; Abgene), lμl TRAVSF forward primer (lOμM stock), lμl FABlink_R reverse primer (lOμM stock), lμl pEX922_lG4 vector template, and 0.5μl pfu Turbo DNA polymerase (Stratagene) to give a total volume of 50μl. PCR cycling was as follows: 94°C for 2 min followed by 20 cycles of

94°C for 15 sec, 55°C for 45 sec, 72°C for 45 sec. A final extension of 10 min at 72°C was allowed. The end modified Vα and Vβ PCR reactions were electrophoresed through a preparative 1.5% agarose gel. A Band of the correct size was excised and gel purified using a Qiagen MinElute kit.

PCR4 - Generation of randomly spliced V and Vβ chains:

The modified Vα and Vβ fragments were randomly annealed to either end of the joining fragment prepared in PCR3 and were spliced in an overlap extension reaction by limited PCR cycling. The unit reaction was assembled as follows: 39.5μl water,

5G1 pfu turbo polymerase lOx buffer (Stratagene), 2μl dNTPs (20mM combined stock; Abgene), lμl TRAVFE forward primer (lOμM stock), lμl TRBV_R reverse primer (lOμM stock), lμl each of modified Vα, Vβ, and joining fragment (lOng of each fragment), and 0.5μl pfu Turbo DNA polymerase (Stratagene) to give a total volume of 50μl. Multiple reactions were performed in parallel. PCR cycling was as follows: 94°C for 2 min followed by 20 cycles of 94°C for 30 sec, 59°C for 45 sec, 72°C for 90 sec. A final extension of 10 min at 72°C was allowed. The spliced Vα and Vβ fragments reactions were electrophoresed through a preparative 1.0% agarose gel. A Band of the correct size was excised and gel purified using a Qiagen MinElute kit.

The spliced and purified products of PCR4 were digested with Nco I and BspEI (restriction sites included in the Vα forward primers used for PCRl and the TRB V_R Vβ constant domain primer) and ligated into a derivative of the pEX922_lG4 phage display vector (containing an engineered BspE I site in the Vβ constant domain sequence, and digested with the same) containing an irrelevant TCR Vα and Vβ open reading frame, thus resulting in the substitution of the irrelevant sequence with a large and diverse population of randomly paired native chain sequences. Ligations were carried out at a 3:1 insert to vector ratio using T4 DNA ligase according to standard protocols.

The ligated DNA was electroporated into TGI cells following concentration and desalting on Qiagen MinElute columns.

Electroporation was performed according to the protocols provided by the commercial supplier of the cells (Stratagene) and using ratios of approximately lμg DNA per 50μl electrocompetent cells. A total of ~20μg of ligated DNA was electroporated.

Following electroporation, cells were reclaimed from cuvettes by resuspension in 950μl of prewarmed (37 degrees) SOC medium and directly plated out on 20cm x 20cm 2TY-agar plates containing lOOμg/ml ampicillin and 2% glucose. Plates were incubated overnight at 30°C and the resultant cell lawns were scraped into a small volume of 2TY medium containing 20% glycerol and 2% glucose. Aliquots (250μl) of the library were frozen on dry ice and stored at -80 degrees. The predicted library size was 1 x 10⁹ . PCR Primers

V α Tagged Forward Primer Set (all in the 5 '-3' direction)

AVI (a) aaatactcagccggccatggcccag caaagccttgagcagccctctg

AVl(b) aaatactcagccggccatggcccag caaaacattgaccagcccactg

AV2 aaatactcagccggccatggcccag gaccaagtgtttcagccttccac

AV3 aaatactcagccggccatggcccag cagtcagtggctcagccggaag AV4 aaatactcagccggccatggcccag gctaagaccacccagcccatctc

AV5 aaatactcagccggccatggcccag gaggatgtggagcagagtcttttcc

AV6(a) aaatactcagccggccatggcccag caaaagatagaacagaattccgagg

AV7 aaatactcagccggccatggcccag aaccaggtggagcacagccctc

AV8(a) aaatactcagccggccatggcccag cagtctgtgagccagcataacc AV8(b) aaatactcagccggccatggcccag cagtcagtgacccagcctgacatc

AV8(c) aaatactcagccggccatggcccag cagtcggtgacccagcttgatgg

AV8(d) aaatactcagccggccatggcccag cagtcggtgacccagcttggcag

AV8(e) aaatactcagccggccatggcccag cagtcKgtgacccagcttRRcagc

AV9(a) aaatactcagccggccatggcccag gattcagtggtccagacagaagg AV9(b) aaatactcagccggccatggcccag Rattcagtgacccagatggaagg

AV 10 aaatactcagccggccatggcccag aaccaagtggagcagagtcctcag

AV11 aaatactcagccggccatggcccag catacactggagcagagtccttc

AVI 2 aaatactcagccggccatggcccag aaggaggtggagcagRatYctgg

AV 13 (a) aaatactcagccggccatggcccag gagaatgtggagcagcatccttc AVI 3(b) aaatactcagccggccatggcccag gagagtgtggcgctgcatcttcc

AV14/DV4 aaatactcagccggccatggcccag cagaagataactcaaacccaacc

AV 16 aaatactcagccggccatggcccag cagagagtgactcagcccgagaag

AV 17 aaatactcagccggccatggcccag caacagggagaagaggatcctcag

AV 18 aaatactcagccggccatggcccag gactcggttacccagacagaagg AVI 9 aaatactcagccggccatggcccag cagaaggtaactcaagcgcagac

AV20 aaatactcagccggccatggcccag gaccaggtgacgcagagtcccg AV21 aaatactcagccggccatggcccag caggaggtgacRcagattcctgc

AV22 aaatactcagccggccatggcccag atacaagtggagcagagtcctcc

AV23/DV6 aaatactcagccggccatggcccag cagcaggtgaaacaaagtcctc

AV24 aaatactcagccggccatggcccag ctgaacgtggaacaaRgtcctcag

AV25 aaatactcagccggccatggcccag caacaggtaatgcaaattcctcag

AV26(a) aaatactcagccggccatggcccag gctaagaccacccagccc

AV26(b) aaatactcagccggccatggcccag gctaagaccacacagccaaattc

AV27 aaatactcagccggccatggcccag cagctgctggagcagagccctc

AV29/DV5 aaatactcagccggccatggcccag cagcaagttaagcaaaattcacc

AV30 aaatactcagccggccatggcccag caacaaccagtgcagagtcctc

AV34 aaatactcagccggccatggcccag caagaactggagcagagtcctc

AV35 aaatactcagccggccatggcccag caacagctgaatcagagtcctc

AV36/DV7 aaatactcagccggccatggcccag gacaaggtggtacaaagccctc

AV38 aaatactcagccggccatggcccag cagacagtcactcagtcYcaRcc

AV39 aaatactcagccggccatggcccag ctgaaagtggaacaaaaccctc

AV40 aaatactcagccggccatggcccag aattcagtcaagcagacgggcc

AV41 aaatactcagccggccatggcccag aatgaagtggagcagagtcctcag

Vβ Tagged Forward Primer Set (all in the 5 '-3 ' direction):

BV1 cctttctattctcacagcgcgcag gatactggaattacccagacacc

BV2 cctttctattctcacagcgcgcag gaacctgaagtcacccagactcc

BV3 cctttctattctcacagcgcgcag gacacagcYgtttcccagactcc BV4(a) cctttctattctcacagcgcgcag gacactgaagttacccagacacc

BV4(b) cctttctattctcacagcgcgcag gaaacgggagttacgcagacacc

BV5(a) cctttctattctcacagcgcgcag aRggctggRgtcactcaaactcc

BV5(b) cctttctattctcacagcgcgcag gagRctggagtcacccaaagtcc

BV5(c) cctttctattctcacagcgcgcag gaS gctggagtcacMcaaagtcc BV6(a) cctttctattctcacagcgcgcag aatgctggtgtcactcagaccc

BV6(b) cctttctattctcacagcgcgcag aYtgctgggatcacccaggcacc BV7(a) cctttctattctcacagcgcgcag ggWgctggagtYtcccagtcc

BV7(b) cctttctattctcacagcgcgcag ggtgctggagtctcccagWcYcc

BV7(c) cctttctattctcacagcgcgcag gatactggagtctcccagRacc

BV7(d) cctttctattctcacagcgcgcag atatctggagtctcccacaacc BV9 cctttctattctcacagcgcgcag gattctggagtcacacaaaccc

B V 10 cctttctattctcacagcgcgcag gatgctgRaatcacccagagcc

B V 11 (a) cctttctattctcacagcgcgcag gaagctgRagttgcccagtcYcc

B V 11 (b) cctttctattctcacagcgcgcag gaagctggagtggttcRgtctcc

BV12(a) cctttctattctcacagcgcgcag gatgctggHRttatccagtcacc B VI 2(b) cctttctattctcacagcgcgcag gatgctagagtcacccagacacc

BV13 cctttctattctcacagcgcgcag gctgctggagtcatccagtccc

BV14 cctttctattctcacagcgcgcag gaagctggagttactcagttcc

B V 15 cctttctattctcacagcgcgcag gatgccatggtcatccagaacc

B V 16 cctttctattctcacagcgcgcag ggtgaagaagtcgcccagactcc BV17 cctttctattctcacagcgcgcag gagcctggagtcagccagacc

B V 18 cctttctattctcacagcgcgcag aatgccggcgtcatgcagaacc

BV19 cctttctattctcacagcgcgcag gatggtggaatcactcagtcc

BV20 cctttctattctcacagcgcgcag Rgtgctgtcgtctctcaacatcc

BV21 cctttctattctcacagcgcgcag gacaccaaggtcacccagagacc BV23 cctttctattctcacagcgcgcag catgccaaagtcacacagactcc

BV24 cctttctattctcacagcgcgcag gatgctgatgttacccagaccc

BV25 cctttctattctcacagcgcgcag gaagctgacatctaccagaccc

BV26 cctttctattctcacagcgcgcag gatgctgtagttacacaattccc

BV27 cctttctattctcacagcgcgcag gaagcccaagtgacccagaaccc BV28 cctttctattctcacagcgcgcag gatgtgaaagtaacccagagctcg

BV29 cctttctattctcacagcgcgcag agtgctgtcatctctcaaaagcc

BV30 cctttctattctcacagcgcgcag tctcagactattcatcaatggcc

Alpha Chain Constant Domain Reverse Primer: TRAV_R 5 '-attagtcttgaatttcgaatctctcagctggtacacggcMgggtcagg-3 '

Beta Chain Constant Domain Reverse Primer:

TRB V_R 5 ' -attcgtatagtttgcggccgctccggagtgcacctccttcccattcaccc-3 '

V a Fragment Adaptor Forward Primer:

TRAVFE 5 ' -cttacttgcaaatactcagccggccatggc-3 '

V Fragment Adaptor Reverse Primer:

TRAVSR 5 '-cgaatctctcagctggtacacggc-3 '

Vβ Fragment Adaptor Forward Primer:

TRBVSF 5 '-cctttctattctcacagcgcgcag-3 '

Constant Joining Fragment Forward Primer:

TRAVSF 5 ' -gccgtgtaccagctgagagattcg-3 '

Constant Joining Fragment Reverse Primer:

FABlinkJR. 5 ' -ctgcgcgctgtgagaatagaaagg-3 ' Wherein:

R = A or G Y = C or T K = G or T M = A or C H = A or C or T W = A or T

Example 2 - Isolation of TCRs that bind to a given target peptide-MHC complexe from a diverse phage library displaying native TCR variable domains

The isolation of TCRs that bind to a target peptide-MHC complex from the diverse phage library displaying native TCR variable domains described above was carried out as follows. The initial panning was carried utilising the selection of phage particles (prepared as described above) displaying native TCR variable domains.

Streptavidin-coated paramagnetic beads (Roche) were pre-washed according to manufacturer's protocols. Phage particles, displaying native TCR variable domains at a concentration of 10¹² to 10¹³ cfu in 3% powdered milk- PBS, were pre-mixed with biotinylated SLLMWITQC-HLA-A*0201 at a concentration of 500nM for all three rounds of selection carried out. The mixture of native TCR variable domain-displaying phage particles and SLLMWιTQC-HLA-A*0201 complex was incubated for one hour at room temperature with gentle rotation, and the native TCR variable domain- displaying phage particles bound to SLLMWITQC-HLA-A*0201 Phage biotynlated HLA complexes were rescued for 5 minutes with 100 ml of streptavidin-coated (Roche) magnetic beads which had been blocked with de-biotynalyted 3% Milk powder PBS. After capture of the phage particles, the beads were washed a total of six times (three times in PBS-0.1 % tween20 and three times in PBS) using a Dynal magnetic particle concentrator. After final wash, the beads were re-suspended in lOOμl of freshly prepared PBS and 50μl of the re-suspended beads was used to infect 10 ml of E.coli TGI at OD(600nm)=0.5 freshly prepared for the amplification of the selected phage particles according to established methods.

After the third round of selection, 100 colonies were picked from either the

SLLMWITQC-HLA-A*0201 complex plates and used to inoculate 100 μl of 2TYAG in a 96-well microtiter plate. The culture was incubated at 30°C with shaking overnight. 100 μl of 2TYAG was then sub-inoculated with 2 to 5 μl of the overnight cultures, and incubated at 30°C with shaking for 2 to 3 hours or until the culture became cloudy. To infect the cells with helper phage, the culture was infected with 100 μl of 2TYAG containing 5 x 10⁹pfu helper phages, and incubated at 37°C for 60 minutes. 5 μl of the infected culture was added to 200 μl of 2TYAK ("TYAG + 100 μg/ml Ampicillin and 50 μg/ml Kanomycin) The plates were incubated at 25°C for 20 to 36 hours with shaking at 300 rpm. The cells were precipitated by centrifugation at 3000g for 10 minutes at 4°C. Supernatants were used to screen for high affinity TCR mutants by phage ELISA.

Phage clones which bound to the SLLMWITQC-HLA-A*0201 complex were found during the ELISA screening as determined by their strong ELISA signals (OD⁶⁰⁰ 0.3- 1) relative to control wells (OD⁶⁰⁰ 0.05). This therefore demonstrates that TCRs with the desired specificity had been isolated from said library. Such an approach allows one to isolate TCRs from a phage display library that are capable of binding to any peptide-MHC or CDl-antigen. The general approach involves phagemid DNA encoding the identified TCR being isolated from the relevant E.coli cells using a Mini- Prep kit (Quiagen, UK). PCR amplification can be carried out using the phagemid

DNA as template and a set of primers designed to amplify the soluble TCR a and β chain DNA sequences encoded by the phagemid. The full range of primers required can be deduced by reference to the TCR a and TCR β Vand C sequences. The PCR product is then digested with appropriate restriction enzymes and cloned into an E. coli expression vector with corresponding insertion sites. The amplified TCR a and β chain DNA sequences (which include, as described above, codons encoding the cysteines required to form the introduced constant domain interchain disulfide bond) are then used to produce a soluble TCR as described in WO 03/020763. Briefly, the two chains are expressed as inclusion bodies in separate E.coli cultures. The inclusion bodies are then isolated, de-natured and re-folded together in vitro.

If it is required to characterise the binding characteristics of the soluble TCR, this may be done using the Biacore system:

Biacore surface plasmon resonance characterisation ofsTCR binding to specific pMHC

A surface plasmon resonance biosensor (Biacore 3000™ ) can be used to analyse the binding of an sTCR to its peptide-MHC ligand. This is facilitated by producing single pMHC complexes (described below) which are immobilised to a streptavidin-coated binding surface in a semi-oriented fashion, allowing efficient testing of the binding of a soluble T-cell receptor to up to four different pMHC (immobilised on separate flow cells) simultaneously. Manual injection of HLA complex allows the precise level of immobilised class I molecules to be manipulated easily.

Biotinylated soluble peptide-MHC molecules are refolded in vitro from bacterially- expressed inclusion bodies containing the constituent subunit proteins and synthetic peptide, followed by purification and in vitro enzymatic biotinylation (O'Callaghan et al. (1999) Anal. Biochem. 266: 9-15). Heavy chain was expressed with a C-terminal biotinylation tag which replaces the transmembrane and cytoplasmic domains of the protein in an appropriate construct. Inclusion body expression levels of ~75 mg litre bacterial culture were obtained. The MHC light-chain or β2-microglobulin was also expressed as inclusion bodies in E.coli from an appropriate construct, at a level of ~500 mg/litre bacterial culture. E. coli cells are lysed and inclusion bodies are purified to approximately 80% purity. Protein from inclusion bodies was denatured in 6 M guanidine-HCl, 50 mM Tris pH 8.1, 100 mM NaCl, 10 mM DTT, 10 mM EDTA, and are refolded at a concentration of 30 mg/litre heavy chain, 30 mg/litre β2m into 0.4 M L-Arginine-HCl, 100 mM Tris pH 8.1, 3.7 mM cystamine, 6.6 mM β-cysteamine, 4 mg/ml of the peptide required to be loaded by the MHC molecule, by addition of a single pulse of denatured protein into refold buffer at < 5°C Refolding is allowed to reach completion at 4°C for at least 1 hour.

Buffer is exchanged by dialysis in 10 volumes of 10 mM Tris pH 8.1. Two changes of buffer may be necessary to reduce the ionic strength of the solution sufficiently. The protein solution is then filtered through a 1.5μm cellulose acetate filter and loaded onto a POROS 50HQ anion exchange column (8 ml bed volume). Protein is eluted with a linear 0-500 mM NaCl gradient. Peptide-MHC complex elutes at approximately 250 mM NaCl, and peak fractions are collected, a cocktail of protease inhibitors (Calbiochem) was added and the fractions were chilled on ice.

Biotinylation tagged pMHC molecules are buffer exchanged into 10 mM Tris pH 8.1, 5 mM NaCl using a Pharmacia fast desalting column equilibrated in the same buffer. Immediately upon elution, the protein-containing fractions are chilled on ice and protease inhibitor cocktail (Calbiochem) was added. Biotinylation reagents are then added: 1 mM biotin, 5 mM ATP (buffered to pH 8), 7.5 mM MgC12, and 5 μg/ml BirA enzyme (purified according to O'Callaghan et al. (1999) Anal Biochem. 266: 9- 15). The mixture is then allowed to incubate at room temperature overnight.

The biotinylated peptide-MHC molecules are purified using gel filtration chromatography. A Pharmacia Superdex 75 HR 10/30 column is pre-equilibrated with filtered PBS and 1 ml of the biotinylation reaction mixture is loaded and the column is developed with PBS at 0.5 ml/min. Fractions containing biotinylated peptide-MHC are pooled, chilled on ice, and protease inhibitor cocktail is added. Protein concentration is determined using a Coomassie-binding assay (PerBio) and aliquots of biotinylated pHLA-A*0201 molecules are stored frozen at -20°C. Streptavidin is immobilised by standard amine coupling methods.

Such immobilised complexes are capable of binding both T-cell receptors and the coreceptor CD8αα, both of which may be injected in the soluble phase. Specific binding of TCR is obtained even at low concentrations (at least 40μg/ml), implying the TCR is relatively stable. The pMHC binding properties of sTCR are observed to be qualitatively and quantitatively similar if sTCR is used either in the soluble or immobilised phase. This is a control for partial activity of soluble species and also suggests that biotinylated pMHC complexes are biologically as active as non- biotinylated complexes.

The interactions between the isolated TCRs containing a novel inter-chain bond and its ligand/ MHC complex or an irrelevant HLA-peptide combination, the production of which is described above, can be analysed on a Biacore 3000™ surface plasmon resonance (SPR) biosensor. SPR measures changes in refractive index expressed in response units (RU) near a sensor surface within a small flow cell, a principle that can be used to detect receptor ligand interactions and to analyse their affinity and kinetic parameters. The probe flow cells are prepared by immobilising the individual HLA- peptide complexes in separate flow cells via binding between the biotin cross linked onto β2m and streptavidin which have been chemically cross linked to the activated surface of the flow cells. The assay can be performed by passing sTCR over the surfaces of the different flow cells at a constant flow rate, measuring the SPR response in doing so.

To measure Equilibrium binding constant

Serial dilutions of the TCRs are prepared and injected at constant flow rate of 5 μl min-1 over two different flow cells; one coated with ~1000 RU of the target peptide- MHC complex, the second coated with -1000 RU of non-specific HLA-A2 -peptide complex. Response is normalised for each concentration using the measurement from the control cell. Normalised data response is plotted versus concentration of TCR sample and fitted to a hyperbola in order to calculate the equilibrium binding constant, K_D- (Price & Dwek, Principles and Problems in Physical Chemistry for Biochemists

(2^nd Edition) 1979, Clarendon Press, Oxford).

To measure Kinetic Parameters For high affinity TCRs K_D was determined by experimentally measuring the dissociation rate constant, kd, and the association rate constant, ka. The equilibrium constant K_D is calculated as kd/ka.

TCR is injected over two different cells one coated with -300 RU of the target peptide-MHC complex, the second coated with -300 RU of non-specific HLA-A2 - peptide complex. Flow rate is set at 50 μl/min. Typically 250 μl of TCR at -3 μM concentration was injected. Buffer is then flowed over until the response had returned to baseline. Kinetic parameters are calculated using Biaevaluation software. The dissociation phase is also fitted to a single exponential decay equation enabling calculation of half-life.

Claims

Claims:

1. A diverse library of nucleoproteins each displaying on its surface a polypeptide comprising a native TCR α variable domain sequence or native TCR β variable domain sequence, in which library the diversity resides in the variety of native TCR α variable domain or native TCR β variable domain sequences displayed.

2. A diverse library of nucleoproteins as claimed in claim 1 in which said nucleoproteins are phage or ribosome particles.

3. A diverse library of nucleoproteins as claimed in claim 1 or claim 2 wherein at least some of the displayed polypeptides comprise a native TCR β variable domain sequence located N-terminal to part of a TCR β chain which includes all or part of a TCR β constant domain sequence except the transmembrane domain thereof

4. A diverse library of nucleoproteins as claimed in claim 1 or claim 2 wherein at least some of the displayed polypeptides comprise a native TCR α variable domain sequence located N-terminal to part of a TCR α chain which includes all or part of a TCR α constant domain sequence except the transmembrane domain thereof.

5. A diverse library of nucleoproteins as claimed in claim 3 or claim 4 wherein a first displayed polypeptide comprising the native TCR α variable domain sequence or native TCR β variable domain sequence is associated with a second polypeptide comprising a native TCR variable domain sequence located N-terminal to part of a TCR chain which includes all or part of a TCR constant domain sequence except the transmembrane domain thereof, said association being maintained at least in part by a disulfide bond which has no equivalent in native αβ T cell receptors between introduced cysteine residues in the TCR constant domain sequences of the first and second polypeptides.

6. A diverse library of nucleoproteins as claimed in claim 5 comprising phage particles or ribosomes on which the first displayed polypeptide is linked by a peptide bond at its C-terminus to a surface exposed amino acid residue of the phage particle or ribosome.

7 A diverse library of phage particles as claimed in claim 6 wherein the first and second polypeptides are linked by a disulfide bond between cysteine residues substituted for Thr 48 of exon 1 of TRAC*01 and Ser 57 of exon 1 of TRBC1*01 or TRBC2*01 or the non-human equivalent thereof.

8. A method of identifying a ligand or ligands of a target peptide-MHC (pMHC) complex or CDl-antigen complex, which method comprises

(a) contacting several members of a library as claimed in of the preceding claims in parallel with the target pMHC or CDl-antigen and identifying members which bind to the target pMHC or CDl-antigen, (b) contacting members identified in step (a) in series with the target pMHC or CDl-antigen, and assessing their affinities for the pMHC or CDl-antigen, (c) selecting one or more members having the desired affinity assessed in step (b) and determining the variable domain sequence(s) of the displayed polypeptide(s), (d) creating soluble fonn TCRs incorporating the thus-determined variable domain sequence(s), and optionally, (e) determining or redetermining as the case may be, the affinities and/or the off-rate for the target pMHC or CDl-antigen of these TCRs, and optionally, (f) selecting one or more TCRs having the desired affinity and/or off-rate determined in step (b) or optional step (e) as the desired ligand(s).

9. A method of identifying a ligand or ligands of a target peptide-MHC (pMHC) complex or CDl-antigen complex, which method comprises

(a) contacting several members of a library as claimed in any of claims 1 to 7 in parallel with an antigen-presenting cell which presents the target pMHC or CDl-antigen and identifying members which bind to the cell-presented target pMHC or CDl-antigen

(b) contacting members identified in step (a) as binding to the target pMHC or CDl-antigen in series with the target pMHC or CDl-antigen, and assessing their affinities for the pMHC or CD 1 -antigen,

(c) selecting one or more members having the desired affinity assessed in step (b), and determining the variable domain sequence(s) of the displayed TCRs,

(d) creating soluble form TCRs incorporating the thus-determined variable domain sequence(s), and optionally, (e) determining the affinities and/or the off-rate for the target pMHC or CDl- antigen of these TCRs, and optionally (f) selecting one or more TCRs having the desired affinity and/or off-rate detennined in step (b) or optional step (e) are selected.

10. A method as claimed in claim 9 further comprising the following steps prior to step (a)

(i) several members of the library are contacted in parallel with an antigen- presenting cell which does not present the target pMHC or CDl-antigen (ii) members identified in step (i) as not binding to the antigen presenting cell are used as the several members of the library referred to in step (a).

11. A method as described claim 9 or 10 in which the antigen-presenting cell which presents the target pMHC or CDl-antigen is caused to present said target pMHC or CD 1 -antigen by peptide-pulsing or antigen-pulsing.

12. A method as described claim 9 or 10 in which the antigen-presenting cell which presents the target pMHC or CDl-antigen is a non-peptide-pulsed antigen presenting cell.