The present invention relates to a method of producing a binding molecule and binding molecules produced by the method.

The existing dogma and the current approach to developing affinity reagents (such as antibodies, antibody mimics, their fragments, aptamers and all other types of affinity reagents) against proteins is that an affinity reagent must be complementary to the surface of the antigen molecule at the interface of the affinity reagent and antigen. The same principle applies to other protein-protein interactions. Therefore, affinity reagents are generally sought and developed against folded proteins, folded fragments, folded protein products of post-translational modification, or any other complete structure of the protein or protein fragment, including peptides or even haptens. In this way, proteins or protein fragments are generally considered to be "complete" unchangeable units against which affinity reagents are developed.

In order to produce effective affinity reagents, the aim is to achieve a high degree of complementarity between the surfaces of the two molecules at their interface. Such an interface is stabilised through a combination of hydrogen bonds (with or without additional water molecules), ionic and van der Waals interactions, and through spatial complementarity between the two molecules. Devising such an interface is the main challenge in the field of antibody research and development.

Currently, two main approaches are used to produce an effective interface between an affinity reagent and an antigen: 1) animal based methods in which an antibody is generated using the immune system of an experimental animal; and 2) by screening large recombinant libraries containing antibodies, antibody mimics, their fragments or aptamers. Such libraries are often randomly generated. A variety of "display" systems exist to ensure that the selected affinity reagent can be identified and amplified, for example cell and phage based display systems, ribosome display system, mRNA molecular display, antibody-ribosome-mRNA complexes (ARM), other molecular display, self-assembly and in vitro compartmentalization systems. Both the traditional animal based and the molecular approaches are time and resource consuming. Further, they constitute a major challenge and a bottleneck in proteomics, antibody research and development, and life sciences in general. All of the above approaches also depend on the availability of the antigen in order to carry out immunisation or in vitro selection. Since the discovery of antibodies and of the monoclonal antibody, no technology has been developed which overcomes these limitations.

The general direction in antibody research and development is to ensure that the antigenic protein (or protein fragment) structure remains intact through the antibody development process. This is to develop antibodies, etc. against the native conformation of a protein.

However, the inventors for this application postulated that affinity reagents, when binding to a particular part of a protein, may compete against the individual structural fragments of the protein itself. For example, where a native protein has two regions of secondary structure which bind together, an affinity reagent could compete with the first region of secondary structure to bind to the second region of secondary structure. This takes advantage of the fact that protein folding is not an irreversible process and that the elements of the protein secondary or supersecondary structure, or protein motifs and small domains already possess the appropriate structure to enable a very good fit with the remaining part of the protein structure.

Protein structure is not permanently fixed but is adaptable and individual elements of the protein's own primary, secondary, tertiary, or quaternary structures are labile, i.e. they can move apart from other parts of the protein and are not fixed in the final native completely folded state (unless covalently cross-linked, for example, with disulphide bonds). Therefore, even after a protein is completely folded into its final native state, portions of the protein may unfold and re-fold to a certain extent. For example, the so called "assisted protein folding" in nature relies on a similar phenomenon to unfold and re-fold imperfectly folded proteins. Protein re-folding is also being widely used in research, e.g. for re-folding misfolded and insoluble protein aggregates (inclusion bodies) following recombinant protein expression (Terashima et al. Process Biochem., 31, 341-345 (1996); Katoh et al Process Biochem., 35, 297-300 (1999); Protein Refolding Kit, Pierce; DUALrefold system, Dualsystems Biotech; iFOLD ^® Protein Refolding Systems, EMD Chemicals Inc; QuickFold™ Protein Refolding Kit, Athena Environmental Sciences, Inc). However, no one has previously thought to take advantage of this phenomenon for producing affinity reagents.

The inventors hypothesised and have experimentally proven that fragments of a protein's own sequence and structure, if presented to the otherwise complete and folded protein, will compete with the corresponding fragments of that protein molecule and will bind to that protein. The inventors have termed this use of fragments of the protein of interest as binding molecules as the artificial antibody approach or "artibody" approach.

The artibody approach is principally different from any existing approaches in that the artibody itself is a fragment of the protein which is used for recognising the complementary elements of that same protein. The artibody approach allows an affinity reagent or binding molecule to be designed based solely on knowledge of the target antigen structure/sequence. This approach is fundamentally different to the affinity reagent generation approaches used previously and has never before been considered as a suitable method for the rational design of new affinity reagents or binding molecules.

Generally, known approaches make use of the binding affinity that two separate proteins may have for each other, i.e. inter-protein interactions. For example, where a protein is a homodimer, the areas in which the two protein subunits bind together may have been investigated. However, this makes use of the binding between two proteins and no conformational change is required to reveal the binding site. The surface of one protein or fragment binds to the surface of the other protein. In contrast, the artibody approach makes use of interactions which are normally present within a single folded protein, i.e. intra-protein interactions. It relies on internal binding forces within a protein which normally hold the protein in its native conformation. This means it can be applied to all proteins and not only those proteins which are capable of binding to other proteins, as in the prior art. International patent application WO 2006/088823 and US Patent Application No. 2005/0026165 relate to probes which can bind to a target protein. However, the probes disclosed in these documents bind to the surface of the target protein, recreating the inter-protein interactions which are normally responsible for forming a multi-protein plaque (as seen in Figure 1 of WO 2006/088823). The probe binds to the surface of the target protein making use of inter-protein interactions that occur between the surface of two target proteins. The probe does not bind in place of a portion of the target protein and does not make use of protein interactions within the target protein.

International patent application WO 99/40435 relates to inhibitors of protein folding. These inhibitors bind to sites which are buried in the native folded protein. They do not bind in place of a part of the target protein located at the surface of the target protein.

J. G. Hall & C. Frieden. P.N.A.S. USA, Vol. 86, pp. 3060-3064 (1989) also relates to inhibitors of protein folding. Again, the inhibitors disclosed in this document bind to internal regions of the protein, not near the surface of the target protein in place of a part of the target protein. The inhibitors are not used on the completely folded protein.

Permanne et al. FASEB Journal, Vol. 16, pp. 860-862 (2002) relates to the use of a 5 residue beta breaker peptide to disrupt amyloid deposition in Alzheimer's disease. The beta breaker peptide simply binds to the surface of the target protein. It does not bind in place of a portion of the target protein.

WO 96/39834 relates to the use of peptides to bind to proteins to stop amyloid formation such as in Alzheimer's disease. These peptides bind to the surface of the target protein rather than binding in place of a portion of the target protein near the surface of the target protein. Dwyer, Life Sciences, Vol. 45, pp. 421-429 (1989) discloses two proteins binding to each other at the surface of the protein. This document does not disclose a peptide binding to the target protein in place of a surface exposed portion of the target protein.

According to a first aspect, the present invention provides a method of producing a binding molecule, the method comprising the steps of: selecting a surface exposed portion of a fully folded target protein; and producing a binding molecule to the target protein, wherein the binding molecule comprises a peptide having a sequence which has substantial identity with at least a substantial part of the sequence of the surface exposed portion of the fully folded target protein thereby allowing the binding molecule to bind to the target protein in place of the surface exposed portion of the target protein.

In the method of the invention, the sequence of the peptide binding molecule is based on the sequence of the portion of the target protein which is displaced when the binding molecule binds to the remaining portion of the target protein. This means that the conformation of the binding molecule in the binding region will be similar or identical to the conformation of the binding region of the surface exposed portion. The binding molecule competes with and displaces the fragment of the target protein to which it corresponds. It displaces the solvent exposed portion of the target protein upon which the peptide of the binding molecule is based.

For the artibody effect to occur, the binding molecule should be close to the target protein and should be able to mimic the structure of the target protein. In order for the peptide of the binding molecule to interact with the target protein by displacing the original fragment (i.e. the solvent exposed portion), the peptide of the binding molecule should assume the spatial position and orientation of the displaced fragment of the target protein.

The term "binding region" when used with respect to the binding molecule means the region defined by the amino acids which are exposed on the surface of the binding molecule and which interact with the target protein in order to cause the binding molecule to bind to the target protein. This is the region of the binding molecule which interacts with the target protein in place of the surface exposed portion.

The term "binding region" when used with respect to the surface exposed portion of the target protein means the region defined by the amino acids which interact with the target protein in order to cause the surface exposed portion to bind to the target protein. This is the region of the surface exposed portion which forms interactions with the target protein and is normally buried when the surface exposed portion is bound to the target protein in its native state.

The target protein that is selected in the method can be any protein to which it is desired to produce a binding molecule. The target protein must have a surface exposed portion upon which the binding molecule can be based. The target protein may be a protein which is formed from a number of subunits, i.e. it is a multimer. In such an embodiment, the protein as a whole is considered to be the target protein. Therefore, the solvent exposed portion should be at the surface of the protein when in its multimeric configuration. An internal portion of one of the subunits, which is on the outside surface of one of the subunits but which is on the inside of the protein in its multimeric state, is not a solvent exposed portion. In an alternative embodiment, the target protein is a protein which is a single peptide or monomer. For example, this may be a single polypeptide chain or multiple polypeptide chains chemically linked together so that a single protein unit is formed. The target protein does not have any quaternary structure.

The term "surface exposed portion" means a portion of the sequence of the target protein which is at the surface of the target protein so that at least one side of the surface exposed portion is on the outside of the protein. The surface exposed portion is at the surface of the target protein when the target protein is in a fully folded state. This may be the native, fully folded conformation of the target protein. Some proteins form dimers where the two proteins bind together at their surfaces. In such a dimer, the areas where the two monomers bind cannot be considered to be a surface exposed portion because, although the area is at the surface of one of the monomers, it is not surface exposed in the native dimer protein. Therefore, the surface exposed portion should be at the surface of the fully folded target protein when it is in its native state. Otherwise, the binding molecule will not be able to bind to the target protein if its binding site is sterically hindered by the presence of, for example, another protein monomer. The surface exposed portion is generally partially embedded in the surface of the protein but must be able to be displaced from the rest of the target protein. The surface exposed portion may form a defined structure at the surface of the target protein. The surface exposed portion may form a stable structure. Preferably, the surface exposed portion is a discreet structural element such as a portion of secondary structure. For example, in a given protein, the sequence at the C-terminal end of the protein may form a helix which lies on the surface of the protein in its native state. Such a helix is one possible surface exposed portion of the protein. This helix may be partially embedded in that it lies in a groove formed by the rest of the target protein. However, at least one side of the helix will be at the surface of the protein and the helix can be displaced from the rest of the target protein. In an alternative embodiment, a protein may have a β-strand which lies at the surface of the protein in its native state and is partially embedded in the protein. Such a β-strand is another possible surface exposed portion of the protein provided that it can be displaced from the rest of the protein to allow a binding molecule, which is based on the β-strand, to bind to the protein. Preferably, the surface exposed portion contains at least one polar amino acid. More preferably, the surface exposed portion contains a mixture of polar and non-polar amino acids.

The surface exposed portion of the target protein may be any suitable structure. The surface exposed portion may be a portion of primary structure, secondary structure, supersecondary structure, motif, domain or tertiary structure element. For example, it may be a helix, a β-strand, a β-turn, a random coil, or a motif such as a helix-turn- helix, β-hairpin motif, or another motif comprising one or more helices and/or β- strands. The surface exposed portion may simply be a portion of secondary structure. The surface exposed portion may be a helix, β-strand or β-hairpin motif.

The sequence of the surface exposed portion can be located anywhere in the sequence of the target protein as long as it forms a structure which is exposed on the surface of the target protein. The surface exposed portion may be located at or near the C- terminal or N-terminal of the target protein. Alternatively, it may be located within the sequence of the target protein, i.e. not at or near the C-terminal or N-terminal of the target protein. Preferably, the surface exposed portion is not located at the N- terminal of the target protein. The surface exposed portion may be located at or near the C-terminal of the target protein. Advantageously, about 90% of proteins have a surface exposed C-terminal domain.

The terms "C-terminal" and "N-terminal" mean that the surface exposed portion is located at the C-terminal end or at the N-terminal end of the target protein. When the surface exposed portion is located near the C-terminal or N-terminal, this means that there may be a stretch of between about 2 and about 10 amino acids between the end of the sequence and the start of the surface exposed portion. Therefore, if the start of a helix was located 8 amino acids from the C-terminal end of a target protein and the helix was selected as the surface exposed portion, this helix would be near the C- terminal of the target protein.

As indicated above, the surface exposed portion of the target protein may be any suitable structure. This may be formed from a continuous stretch of amino acids in the sequence of the target protein, for example, a helix, a β-strand or a β-hairpin motif. Alternatively, it may be formed from distinct stretches of amino acids in the sequence of the target protein which are adjacent in the native conformation; for example, two β-strands which are at a distance from each other in the sequence of the protein but are adjacent in the native protein to form a β-sheet. Preferably, the surface exposed portion is formed by a continuous stretch of amino acids in the sequence of the target protein. This allows the easy design of the binding molecule as it is based on the sequence of the surface exposed portion.

Preferably, the surface exposed portion has at least one flexible end capable of forming a hinge. This allows the surface exposed portion to be displaced from its native position in the target protein relatively easily so that the surface exposed portion allows the binding molecule to bind to the target protein. Having a flexible hinge allows the surface exposed portion to be displaced or to move relatively easily so that the binding molecule can bind in its place. Preferably, the surface exposed portion has two flexible ends.

An end of a surface exposed portion may be flexible as a result of being connected to an amino acid or amino acids which allow a large degree of movement, for example, glycine, serine or alanine, and/or as a result of being connected to another flexible amino acid or acids, for example, lysine, arginine, asparagine, glutamic acid, aspartic acid, glutamine or threonine. It will be readily apparent to a person skilled in the art whether or not the surface exposed portion has a flexible end. For example, the flexibility of amino acid chains is discussed in Karplus and Schulz, Naturwissenschaften, 72, 212 (1985); Huang et al. Angewandte Chemie International Edition, Volume 42, Issue 20, Pages 2269 - 2272 (2003); and Smith et al. Protein Sci. 2003 May; 12(5): 1060-1072. Alternatively, an end of a surface exposed portion may be flexible as a result of being close to an end of the protein (such as the N-terminus or C-terminus) so that the end is not constrained by interactions with other parts of the protein.

An end of the surface exposed portion can be considered to be flexible if its movement is not constrained. The movement of an end of the surface exposed portion might be considered to be constrained if it is prevented from leaving its natively folded position. This may be because of strong non-covalent interactions with surrounding amino acids. Alternatively, it may be because the end is physically restricted from moving, e.g. by being blocked by other elements of the protein structure, because of covalent interactions with other amino acids (for example, through disulphide bonds) or by having insufficient size, such as length, to allow any movement.

The surface exposed portion preferably has a sequence containing between about 3 and about 100 amino acids. More preferably, the surface exposed portion has a sequence which contains between about 5 and about 50 amino acids, more preferably still, between about 8 and about 40 amino acids and, even more preferably, the surface exposed portion has a sequence which contains between about 10 and about 30 amino acids. As will be appreciated by one skilled in the art, the number of amino acids contained in the sequence of the surface exposed portion will depend on the nature of the surface exposed portion itself. For example, if the surface exposed portion is a helix, the sequence may contain between about 8 and about 40 amino acids whereas if the surface exposed portion is an intramolecular loop, the sequence may contain between about 3 and about 15 amino acids.

In order to be able to select a surface exposed portion of the target protein, it is necessary to know enough about the structure of the protein so that it is possible to identify or predict the regions of the protein which are exposed to the surface. With modern techniques, it is well within the grasp of a skilled person to identify or predict a surface exposed portion of a target protein. For example, many proteins have already had their structure resolved and these structures can be found in publicly accessible online databases. For example, protein structures can be found in the Protein Data Bank (http://www.wwpdb.org) or in the Molecular Modeling Database run by the National Centre for Biotechnology Information (NCBI) (http ://www.ncbi.nlm.nih. gov/Structure/MMDB/mmdb . shtml) .

Preferably the target protein is a protein having a known structure.

If the structure of a protein is not known, its 3-D structure can be resolved using techniques such as x-ray crystallography or nuclear magnetic resonance (NMR). Determination of protein structure using x-ray crystallography is well known to those skilled in the art and is described, for example, in the follow references: J. Drenth "Principles of Protein X-Ray Crystallography." Springer, ISBN: 978-0-387-33334-2; D. E. McRee "Practical Protein Crystallography" ISBN: 0-12-486052-4; D. Blow "Outline of Crystallography for Biologists" (Oxford University Press) ISBN: 0-19- 851051-9; and E.E. Lattman "Protein Crystallography: A Concise Guide" ISBN:0801888069.

NMR methods of protein structure determination are described, for example, in the following references: J.Cavanagh et al. "Protein NMR Spectroscopy: Principles and Practice" (Academic Press) ISBN-13: 9780121644918; G.S. Rule et al. "Fundamentals of Protein NMR Spectroscopy (Focus on Structural Biology)" ISBN- 13: 978-1402034992; and A. K. Downing "Protein NMR Techniques (Methods in Molecular Biology)" ISBN-13: 978-1588292469.

If the structure of a protein is not known, regions potentially suitable for the generation of binding molecules can be predicted using a variety of approaches which are well known to those skilled in the art. Many of these can be found at http://www.expasy.org/tools/. For example, some suitable approaches are as follows:

1. Homology (comparative) modelling

This is suitable for predicting 3D protein structures of target proteins homologous to proteins with known 3D structure. See: Arnold K., Bordoli L., Kopp J., and Schwede T. (2006). The SWISS-MODEL Workspace: A web-based environment for protein structure homology modelling. Bioinformatics, 22,195-201; Kopp J. and Schwede T. (2004) The SWISS-MODEL Repository of annotated three-dimensional protein structure homology models Nucleic Acids Research 32, D230-D234; Schwede T, Kopp J, Guex N, and Peitsch MC (2003) SWISS-MODEL: an automated protein homology-modeling server. Nucleic Acids Research 31 : 3381-3385; Guex, N. and Peitsch, M. C. (1997) SWISS-MODEL and the Swiss-PdbViewer: An environment for comparative protein modelling. Electrophoresis 18: 2714-2723; Peitsch, M. C. (1995) Protein modeling by E-mail Bio/Technology 13: 658-660; Bates, P.A., Kelley, L.A., MacCallum, R.M. and Sternberg, M.J.E. (2001) Enhancement of Protein Modelling by Human Intervention in Applying the Automatic Programs 3D-JIGSAW and 3D-PSSM. Proteins: Structure, Function and Genetics, Suppl 5:39-46; Bates, P.A. and Sternberg, M.J.E. (1999) Model Building by Comparison at CASP3: Using Expert Knowledge and Computer Automation. Proteins: Structure, Function and Genetics, Suppl 3:47-54; Contreras-Moreira,B., Bates,P.A. (2002) Domain Fishing: a first step in protein comparative modelling. Bioinformatics 18: 1141-1142; CPHmodels 2.0: X3M a Computer Program to Extract 3D Models. O. Lund, M. Nielsen, C. Lundegaard, P. Worning Abstract at the CASP5 conference Al 02, 2002; Lambert C, Leonard N, De Bolle X, Depiereux E. ESyPred3D: Prediction of proteins 3D structures. Bioinformatics. 2002 Sep;18(9):1250-1256; Combet C, Jambon M, Deleage G & Geourjon C. Geno3D: Automatic comparative molecular modelling of protein. Bioinformatics, 2002, 18, 213-214; and Geourjon C, Combet C, Blanchet C, Deleage G. Identification of related proteins with weak sequence identity using secondary structure information. Protein Sci 2001 Apr;10(4):788-97.

2. Threading

This is suitable for predicting 3D protein structures of target proteins for which no highly homologous proteins with known 3D structure exist. See: Kelley LA and Sternberg MJE. Nature Protocols 4, 363 - 371 (2009); J. Shi, T. L. Blundell, and K. Mizuguchi (2001). FUGUE: sequence-structure homology recognition using environment-specific substitution tables and structure-dependent gap penalties. J. MoI. Biol, 310, 243-257; R. N. Miguel, J. Shi and K. Mizuguchi (2001). Protein Fold Recognition and Comparative Modeling using HOMSTRAD, JOY and FUGUE. In Protein Structure Prediction: Bioinformatic Approach. International University Line publishers, La Jolla, 143-169; Sόding J. (2005) Protein homology detection by HMM-HMM comparison. Bioinformatics 21, 951-960. doi:10.1093^ioinformatics/btil25; Sόding J, Biegert A, and Lupas AN. (2005) The HHpred interactive server for protein homology detection and structure prediction. Nucleic Acids Research 33, W244-W248 (Web Server issue). doi:10.1093/nar/gki40; http://cbsuapps.tc. cornell.edu/loopp.aspx; httpV/compbio.soe.ucsc.edu/SAM TOS/TOS-query.html; Jones, D.T., Taylor, W.R. & Thornton, J.M. (1992) A new approach to protein fold recognition. Nature. 358, 86- 89; Jones, D.T., Miller, R.T. & Thornton, J.M. (1995) Successful protein fold recognition by optimal sequence threading validated by rigorous blind testing. Proteins. 23, 387-397; and Jones, D.T. (1998) THREADER : Protein Sequence Threading by Double Dynamic Programming, (in) Computational Methods in Molecular Biology. Steven Salzberg, David Searls, and Simon Kasif, Eds. Elsevier Science. Chapter 13.

3. Ab initio predictions of 3D protein structures

This is suitable for predicting 3D protein structures of target proteins for which no homologous proteins with known 3D structure exist, or as an independent confirmation of the comparative modelling results. See: Bystroff C & Shao Y. (2002). Fully automated ab initio protein structure prediction using I-SITES, HMMSTR and ROSETTA. Bioinformatics 18 Suppl 1, S54-61; and Bystroff C, Thorsson V & Baker D. (2000). HMMSTR: A hidden markov model for local sequence-structure correlations in proteins. Journal of Molecular Biology 301, 173-90.

4. By using one or a combination of secondary structure prediction methods

Such methods include: SOPM (Geourjon and Deleage, 1994); SOPMA (Geourjon and Deleage, 1995); HNN (Guermeur, 1997); MLRC (Guermeur et al, 1999); DPM (Deleage and Roux, 1987); DSC (King and Sternberg, 1996); GOR I (Gamier et al, 1978); GOR nUGibrat ef a/.. 1987); GOR IV (Gamier et al.. 1996); PHD (Rost and Sander, 1993); PREDATOR (Frishman and Argos. 1996); SIMPA96 (Levin, 1997).

5. By using one or a combination of primary structure analysis tools

5.1 Hydrophilicity, hydrophobicity, polarity, solvent accessibility and exposed/buried residues predictions

See: Kyte J, Doolittle RF. J MoI Biol. 1982 May 5;157(l):105-32; Hopp TP, Woods KR. MoI Immunol. 1983 Apr;20(4):483-9; Grantham R. Science 185:862-864(1974); P.K.Ponnuswamy, M.Prabhakaran, P.Manavalan, Biochim.Biophys.Acta., 623, 301 (1980); Abraham D.J., Leo A.J. Proteins: Structure, Function and Genetics 2:130- 152(1987); Bull H.B., Breese K. Arch. Biochem. Biophys. 161 :665-670(1974); Guy H. R. Biophys J. 47:61-70(1985); Miyazawa S., Jernigen R.L. Macro molecules 18:534-552(1985); Roseman M.A. J. MoI. Biol. 200:513-522(1988); Wolfenden R.V., Andersson L., Cullis P.M., Southgate CCF. Biochemistry 20:849-855(1981); Zimmerman J.M., Eliezer N., Simha R. J. Theor. Biol. 21 :170-201(1968); Eisenberg D., Schwarz E., Komarony M., Wall R. J. MoI. Biol. 179:125-142(1984); Manavalan P., Ponnuswamy P.K. Nature 275:673-674(1978); Black S.D., Mould D.R. Anal. Biochem. 193:72-82(1991); Fauchere J.-L., Pliska V.E. Eur. J. Med. Chem. 18:369- 375(1983); Rao M.J.K., Argos P. Biochim. Biophys. Acta 869:197-214(1986); J.Janin and S.Wodak, J.M.Biol., 125, 357 (1978); Janin J. Nature 277:491- 492(1979); Chothia C J. MoI. Biol. 105:1-14(1976); Rose G.D., Geselowitz A.R., Lesser G.J., Lee R.H., Zehfus M.H. Science 229:834-838(1985); PARKER, JMR; GUO, D; HODGES, RS. BIOCHEMISTRY Volume: 25 Issue: 19 Pages: 5425- 5432 Published: SEP 23 1986; http://www.expasy.org/tools/protscale.html; http://www.isrec.isb-sib.ch/software/SAPS_form.html.

5.2 Transmembrane location prediction

This is to avoid picking a transmembrane region. See Zhao, G., London E. Protein Sci. 15:1987-2001(2006).

5.3 Flexibility predictions

This is to find two flexible regions, which flank a less flexible region of a secondary structure. See P.A.Karplus and G.E.Schulz, Naturwissenschaften, 72, 212 (1985); and Bhaskaran R., Ponnuswamy P.K. Int. J. Pept. Protein. Res. 32:242-255(1988).

5.4 Disordered regions predicitons

This is to avoid disordered regions. No regular structure may indicate flexibility and lack of good intramolecular bonds. See: R. Linding, L.J. Jensen, F. Diella, P. Bork, T.J. Gibson and R.B. Russell. Structure VoI 11, Issue 11, 4 November 2003; and Rune Linding, Robert B. Russell, Victor Neduva and Toby J. Gibson. Nucleic Acids Research, 2003, Vol. 31, No. 13 3701-3708.

5.5 Antigenic predictions

This can be used to identify surface accessible amino acids. See: Kolaskar AS, Tongaonkar PC. FEBS Lett. 1990 Dec 10;276(l-2): 172-4; Welling GW, Weijer WJ, van der Zee R, Welling- Wester S. FEBS Lett. 1985 Sep 2;188(2):215-8; and Jameson, B. A.; Wolf, H. Computer Applications in the Biosciences. Volume: 4, Issue: 1, Pages: 181-186. Published: March 1988.

Alternatively, it is possible to identify surface exposed residues using chemical modification of the protein surface, followed by proteolytic digestion and mass spectrometry or Edman degradation or a combination of other bioanalytical techniques for identifying the modified amino acids.

Once the structure of a protein is known or has been predicted, a skilled person can easily identify surface exposed portions which could be selected to be the basis for the design of an appropriate binding molecule. Similarly, if a skilled person has determined surface exposed portions in a target protein, a skilled person will be able to select a suitable surface exposed portion upon which to base the binding molecule. The characteristics and preferable features of the selected surface exposed portion of the target protein are discussed above.

The need to determine or to predict protein structure may be obviated by making of a set of partially overlapping peptides covering the whole or part of the target protein sequence and experimentally testing each peptide for its binding propensity. Even without the knowledge of the protein structure, such an approach will provide a significantly faster and cheaper way to identify a binding molecule compared to traditional immunisations or in vitro screening techniques.

The sequence of the binding molecule is based on at least a substantial part of the sequence of the surface exposed portion of the target protein. As a result, the binding region of the binding molecule possesses the same or a similar structure as the surface exposed portion or adopts this structure on binding to the target protein by displacing the surface exposed portion. This allows the binding molecule to bind to the rest of the target protein in place of the surface exposed portion with relatively high selectivity and specificity since the binding molecule has the same or a similar sequence compared to the surface exposed portion of the target protein.

The sequence of the binding molecule does not have to be identical to the sequence of the surface exposed portion of the target protein or a part of the surface exposed portion. Not all of the amino acids in the surface exposed portion of the target protein will be involved in interacting or fitting with the rest of the target protein in order to allow the surface exposed portion to bind to the rest of the target protein. Since the surface exposed portion is located at the surface of the target protein, some of the amino acids in the sequence of the surface exposed portion will be on the surface of the protein and, therefore, they will not interact with the rest of the target protein. Therefore, these amino acids do not have to be identical in the binding molecule in order for the binding molecule to have an identical conformation to the surface exposed portion in the binding region. For example, if the surface exposed portion of the target protein is a helix which lies in a groove in the rest of the target protein, only approximately 50% of the amino acids in the helix may interact with and bind to the rest of the protein. These will be the amino acids down one side of the helix which fit into the groove in the rest of the target protein. The amino acids on the other side of the helix may not interact and bind with the groove in the rest of the protein. Therefore, when producing a binding molecule, it is not always necessary for these amino acids in the binding molecule to be identical to the ones in the surface exposed portion because they do not interact with the groove and so may not play a role in binding. So, the amino acids of the binding molecule which are identical to the amino acids of the surface exposed portion are in the binding region of the binding molecule. Amino acids in other areas of the binding molecule can also be identical and this may also be beneficial, for example, it may help to ensure the structure of the binding molecule remains similar to that of the surface exposed portion.

The conformation of the binding molecule in the binding region should be substantially the same as the conformation of the displaced surface exposed portion in the binding region so that the binding molecule can bind to the target protein in place of the surface exposed portion. As will be appreciated by one skilled in the art, the conformation of the binding molecule and the surface exposed portion do not have to be identical in their binding regions for the binding molecule to be able to bind to the target protein. The conformation of the binding molecule in the binding region can differ slightly and the binding molecule will still be able to bind to the target protein. Therefore, not all of the amino acids in the binding region of the binding molecule need to be identical to the amino acids in the binding region of the surface exposed portion of the target protein to still allow the binding region of the binding molecule to bind to the target protein. However, the identity must be sufficient so that the binding region of the binding molecule can bind to the target protein. Preferably, at least about 80% of the amino acids in the binding region of the binding molecule are identical to those in the surface exposed portion, more preferably, at least about 85%, more preferably still, at least about 90%, even more preferably, at least about 95% and, most preferably, 100%. In some instances, it may be preferable to change some of the amino acids in the binding region of the binding molecule so that they are not identical to those in the surface exposed portion. Alternatively, one or more amino acids may be removed from the sequence of the surface exposed portion.

The amino acids in the binding region of the binding molecule which should remain unchanged compared to the surface exposed portion of the target protein are those amino acids which are responsible for the interactions and which provide the complementarity of the shape and surface of the interacting groups, hence causing the surface exposed portion to bind to the rest of the target protein. These amino acids are on the whole the ones with buried side chains and are likely to be, for example, large aliphatic, aromatic or sulphur containing amino acids. They are likely to be mostly hydrophobic amino acids.

Once a skilled person has determined or predicted the structure of a target protein and selected a surface exposed portion, it will be relatively easy for the skilled person to identify which amino acids form part of the binding region as the structure of the protein will give an indication of the orientation of the amino acids and the positioning of their side group with relation to the side groups of other amino acids.

With regard to the overall identity between the binding molecule and the surface exposed portion of the target protein, preferably, the binding molecule has at least about 50% identity with the surface exposed portion or a substantial part thereof. More preferably, the binding molecule has at least about 60% identity with the surface exposed portion or a substantial part thereof, even more preferably, at least about 70% identity, more preferably still, at least about 80% identity, even more preferably, at least about 85% identity, more preferably still, at least about 90% identity, even more preferably, at least about 95% identity, and most preferably, 100% identity. These percentages refer to the identity between the surface exposed portion and the binding molecule based on the sequence of the surface exposed portion. If the binding molecule is complexed or conjugated to another moiety, for example, another protein, the conjugated or complexed moiety is not taken into consideration when assessing the overall identity between the binding molecule and the surface exposed portion of the target protein. In some instances, it may be possible to increase the binding affinity of the binding molecule compared to the surface exposed portion. This can be done by adding one or more amino acids to the binding molecule to increase the number of interactions that take place with the target protein during binding. Preferably, this additional amino acid or acids are added at the end of the binding region of the binding molecule. Alternatively, the additional amino acid or acids may be added in internal regions of the binding region of the binding molecule. Preferably, the additional amino acid or acids provide strong interactions. Suitable amino acids are aliphatic, aromatic, sulphur containing or polar amino acids. This helps to increase the number of interactions which take place between the binding molecule and the target protein thereby helping to improve the binding affinity of the binding molecule.

The binding affinity of the binding molecule may also be improved by increasing the rigidity of the binding molecule. This may be done by introducing cysteine residues in appropriate positions in the sequence to form disulphide bridges which help to stabilise the structure of the binding molecule. Alternatively, the binding molecule can be incorporated into a suitable protein scaffold to help increase the rigidity of the binding molecule. Suitable protein scaffolds can be engineered from any protein having stable tertiary structure and a large number of examples have been reported and are well known to those skilled in the art. For example, suitable protein scaffolds are discussed in Hosse et al. A new generation of protein display scaffolds for molecular recognition. Protein Science (2006), 15:14-27 and Binz et al. Engineering novel binding proteins from nonimmunoglobulin domains. Nature Biotechnology, v.23 n.lO (2005) pp 1257-1268.

The amino acids of the binding molecule may be modified to allow them to interact with the displaced surface exposed portion. The amino acids that can be modified are those which are exposed on the surface of the binding molecule once the binding molecule has bound to the target protein. These amino acids will not be in the binding region of the binding molecule. In effect, the displaced portion of the target protein (the surface exposed portion), which was partially buried, will become exposed upon binding of the binding molecule. This displaced surface exposed portion is likely to be in close proximity to the binding molecule. Therefore, the binding of the binding molecule can be further stabilised if the formally buried amino acids of the surface exposed portion can be made to interact with the 'unused' amino acids on the surface of the binding molecule. This modification process can be done using rational design or affinity selection and saturation. This helps to increase the affinity of the binding molecule for the target protein as the binding molecule binds to the target protein as well as interacting with the displaced surface exposed portion.

The binding molecule has a sequence which is substantially identical with at least a substantial part of the sequence of the surface exposed portion of the target protein. The expression "a substantial part of the sequence of the surface exposed portion" means that the binding molecule may be based on a substantial part of the surface exposed portion. For example, if a helix is selected as the surface exposed portion on which the binding molecule is based, the sequence of the binding molecule does not have to be based on the whole sequence of the helix. The binding molecule may be based on the sequence of the first half of the helix. In this way, the sequence of the binding molecule has identity with a substantial part (i.e. the first half) of the sequence of the surface exposed portion. The binding molecule must have a sequence which is substantially identical to a large enough portion of the surface exposed portion so that it can displace at least some of the surface exposed portion and bind to the target protein in place of the surface exposed portion. Looking at it from an alternative point of view, the part of the sequence of the surface exposed portion on which the binding molecule is based must be large enough so that the binding molecule produced is able to bind to the target protein and displace the surface exposed portion. Preferably, the binding molecule has identity with at least 50%, more preferably at least 70%, even more preferably at least 90%, even more preferably at least 95%, and most preferably 100% of the surface exposed portion of the target protein. It is most preferred that the binding molecule is based on the whole sequence of the surface exposed portion.

The sequence of the binding molecule which allows it to bind to the target protein is preferably between about 3 and about 100 amino acids. More preferably, the sequence of the binding molecule is between about 3 and about 50 amino acids, more preferably still, the sequence of the binding molecule is between about 5 and about 40 amino acids and, even more preferably, between about 10 and about 30 amino acids. Those skilled in the art will appreciate that the number of amino acids contained in, for example, the elements of secondary or supersecondary structure, or short motifs, will vary between different proteins, domains, motifs and secondary structure elements. Therefore, the sequence and the number of amino acids in the sequence of the binding molecule will partly depend on the nature and size of the surface exposed portion that has been selected.

For example, if the surface exposed portion is a helix comprising a sequence of about 50 amino acids, the sequence of the binding molecule may contain anywhere between about 3 and about 50 amino acids. However, if the surface exposed portion is, for example, a helix-turn-helix motif comprising about 100 amino acids, the sequence of the binding molecule may contain anywhere between about 3 and about 100 amino acids. If the surface exposed portion is, for example, a β-hairpin motif comprising about 10 amino acids, the sequence of the binding molecule may be between about 3 and about 10 amino acids.

The binding molecule is preferably specific for the protein target.

Since the binding molecule is a relatively short polypeptide, it can be produced relatively easily using chemical synthesis. Suitable methods are well known to those skilled in the art.

The binding molecule of the invention discussed above may optionally be conjugated to or complexed with an additional moiety. The moiety may be any suitable moiety and the identity of the moiety will depend on the intended function of the conjugate or complex. For example the binding molecule may be conjugated to or complexed with a molecule, such as a polypeptide, which causes homo- or hetero-oligomerisation. For example, the molecule may cause dimerisation, trimerisation, tetramerisation, pentamerisation or polymerisation. The binding molecule may be conjugated to or complexed with a molecule which allows attachment to other molecules or surfaces such as biotin which allows attachment to avidin coated surfaces. The binding molecule may be conjugated to or complexed with functional molecules such as inhibitors, enzymes, toxins, and regulators. Further, the binding molecule may be conjugated to or complexed with peptides, other binding molecules (with the same or different specificity), antibodies, antibody mimics, antibody fragments or antibody mimic fragments which retain their binding functionality, aptamers, functional proteins, structural proteins, nucleic acids, lipids, carbohydrates, natural cell metabolites, chemically synthesised unnatural moieties, aromatic compounds, organic compounds, inorganic compounds, pharmaceutically active ingredients, chemical groups, metals, non-metals, particles, gels, reporter molecules or groups, or labels such as fluorescent, luminescent, quenchers, radioactive, magnetic, non-magnetic, isotopic or spin labels (as in EPR or ESR).

Where the binding molecule is conjugated to an additional moiety, the additional moiety may optionally be conjugated via a linker. Suitable linkers are well known to those skilled in the art.

The binding molecule may comprise other modifications which can be used to improve binding or physical properties of the binding molecule. These include but are not limited to: acylation, acetylation, deacetylation, alkylation, amidation, biotinylation, carboxylation, glutamylation, glycosylation, glycation, glycylation, hydroxylation, iodination, isoprenylation, lipoylation (e.g. prenylation, myristoylation, farnesylation, geranylgeranylation), oxidation, palmitoylation, pegylation, phosphopantetheinylation, phosphorylation, polysialylation, pyroglutamate formation, ubiquitination, citrullination, deamination, deamidation, eliminylation, dehydration, and decarboxylation. These can be similar to the normal types of post-translational modification that can take place which are well known to those skilled in the art.

The binding molecules produced by the invention have a wide range of uses and display a number of advantages over affinity reagents described in the prior art.

The binding molecules of the invention do not rely on any pre-defined protein scaffold, for example, as in the case of antibodies which have a very particular immunoglobulin protein scaffold which holds the complementarity determining regions (CDRs) in place, or in the case of numerous known antibody mimics, for example, as discussed in Hosse et al. A new generation of protein display scaffolds for molecular recognition. Protein Science (2006), 15:14-27 and Binz et al. Engineering novel binding proteins from non-immunoglobulin domains. Nature Biotechnology, v.23 n.lO (2005) pp 1257-1268.

This means that the structure of the binding molecule is much simpler which means that the binding molecule can be produced much more easily. For example, the binding molecule may be produced by chemical synthesis. Further, the binding molecule is capable of folding correctly without any help or scaffold since the protein structure on which the binding molecule is based can fold correctly to give a functional native protein.

Since the binding molecules are based on the sequence of known proteins, it is not necessary to use complicated techniques to produce them such as animal immunisation or in vitro affinity selection using large libraries and a "display system" (e.g. cell display, phage display, ribosome display, mRNA display, ARM, yeast-two- hybrid or similar systems). Further, affinity maturation can be avoided to develop the binding molecules. If affinity maturation is used, it is greatly simplified using the binding molecules of the invention. Further, the binding molecules do not require library screening for the design of the affinity reagent against a known protein target. Instead, the method allows a "rational" approach to complete de novo design of binding molecules based on the target protein structure or even the sequence alone, and does not rely on any other proteins, protein scaffolds or any other scaffolds and does not require the knowledge of other protein sequences or structures.

Although the term "rational design" has been used by others before, it was never used in relation to the approach described above. Instead the term was used in relation to the re-designing of existing proteins for their use as protein scaffolds or antibody mimics to bind other targets. For example, the use and re-engineering of the PDZ domain to recognise and bind new non-natural target sequences as described in Jose Reina et al, Computer-aided design of a PDZ domain to recognize new target sequences. Nature Structural Biology 9, 621 - 627 (2002), S Schneider et al, Mutagenesis and selection of PDZ domains that bind new protein targets. Nat Biotechnol. 1999 Feb; 17(2): 170-5 and M. Ferrer et al., Directed evolution of PDZ variants to generate high-affinity detection reagents. Protein Eng Des SeI. 2005 Apr;18(4):165-73.

Other examples are: 1) the use of polypeptide toxins for their use as scaffolds to display modified sequences or added amino acids, e.g. to engineer a metal binding site, as described in C.Vita et al. Scorpion toxins as natural scaffolds for protein engineering. Proc Natl Acad Sci U S A. 1995 JuI 3;92(14):6404-8; 2) the creation of CD4 protein mimics as ligands to bind to gpl20 protein of the HIV virus as described in L. Martin et al. Rational design of a CD4 mimic that inhibits HIV-I entry and exposes cryptic neutralization epitopes. Nat Biotechnol. 2003 Jan;21(l):71-6; 3) the re-engineering of tetratricopeptide (TPR) protein to use it as a scaffold to create heat shock protein binging molecules, as described in A.L. Cortajarena et al. Protein design to understand peptide ligand recognition by tetratricopeptide repeat proteins. Protein Eng Des SeI. 2004 Apr;17(4):399-409; 4) the re-engineering of the binding site of E.coli periplasmic binding protein (PBP), to make it bind selected ligands in place of the wild-type sugars or amino-acids, as described in L.L.Looger et al. Computational design of receptor and sensor proteins with novel functions. Nature 423, 185-190 (8 May 2003); 5) grafting of the selected epitope, involved in protein-protein interaction with the surface of the target protein, on the solvent-exposed face of the avian pancreatic polypeptide (aPP) to create protein-binding chimera, as described in S.E.Rutledge, et al. Molecular Recognition of Protein Surfaces: High Affinity Ligands for the CBP KIX Domain. J. Am. Chem. Soc, 2003, 125 (47), pp 14336- 14347; 6) the engineering of another protein scaffold from plant homeodomain motif (zinc finger domain), as described in A.H.Kwan et al. Engineering a protein scaffold from a PHD finger. Structure. 2003 JuI; 11(7):803-13; and 7) the re-engineering of the human pancreatic secretory trypsin inhibitor (PSTI) to display a short peptide sequence held between two disulfide bridges, as described in Rόttgen P, Collins J. A human pancreatic secretory trypsin inhibitor presenting a hypervariable highly constrained epitope via monovalent phagemid display. Gene. 1995 Oct 27; 164(2) :243- 50.

Additionally, as the binding molecules are based on the sequence of a known protein and therefore adopt a similar or identical conformation to that known protein, they are naturally relatively highly selective for their target protein because the interactions between the binding molecule and the target protein are the same or similar as those found in the native protein. Therefore, this produces a relatively stable structure. Further, the binding molecules can be denatured without the loss of their binding ability and will subsequently refold to their correct functional state. This makes them very robust so that they are likely to have a relatively long storage life and can survive being subjected to relatively harsh conditions. This means that binding molecules can be regenerated and reused and are therefore advantageous over traditional antibodies in affinity-based assays and especially in such applications where multiple use of the affinity reagent may be required. Such examples include but are not limited to blots, arrays, affinity chromatography, chromatography resins, other immunosorbent substrates including ELISA, beads, particles microparticles, nanoparticles, microfluidic and chip-based applications, Lab-on-a-Chip, Biacore, other SPR based assays, QCM based assays, optical and spectrometric fluorescence and mass spectrometry detection, scanning microscopy and other instrumental measurements. The robust character of the binding molecules makes them especially suitable for therapeutics applications because they are more stable against extreme pH conditions, denaturing media and can be easily made protease-resistant.

A further advantage of the binding molecules of the invention is that the binding molecule may affect the function of the target protein as it may cause its conformation to change to a non-native state. This is as a result of the surface exposed portion of the target protein being displaced from its original position and the associated changes in the target protein conformation. This allows the binding molecules to be used to affect protein function. This can be used to down regulate protein function or disrupt protein-protein interaction networks (especially useful in cancer therapies), thus providing definitive advantage over traditional approaches, e.g. therapeutic antibodies. The ability to affect the function of the target protein is the intrinsic ability of the binding molecule, due to the mechanism of interaction with the target protein. This mechanism is principally different form protein-protein interaction involved in traditional antibody-antigen or antibody mimics-antigen interaction. Since the binding molecules resemble fragments of native proteins which originate from the protein surface (e.g. normally exposed to the immune system), there is a reduced risk of developing an immune response against such a fragment as these portions of protein are likely to have been exposed to the immune system previously.

Other advantages are that the binding molecules are relatively small so higher immobilisation densities can be achieved (e.g. on arrays) and they show increased permeability which is advantageous for use in imaging, cytochemistry, histochemistry, and therapeutic applications. The binding molecules have a defined position within the target protein, hence multiple interaction sites and multiple binding molecules are possible, so improved affinity/specificities are achievable (this is not always possible with large IgG antibodies which, because of their size, will not always be able to bind to more than one binding site on their antigen simultaneously, especially on smaller antigens).

According to a second aspect, the present invention provides a method of producing a binding molecule, the method comprising: producing a binding molecule to a fully folded target protein, wherein the binding molecule comprises a peptide having a sequence which has substantial identity with at least a substantial part of the sequence of a surface exposed portion of the fully folded target protein thereby allowing the binding molecule to bind to the target protein in place of the surface exposed portion of the target protein.

The present invention also provides a binding molecule obtained by the method of the first or second aspects of the invention.

In a third aspect, the present invention provides a binding molecule comprising one of the sequences selected from SEQ ID NOS: 1-39, 41-53 and 55-86, wherein the binding molecule can bind to a target protein displacing a surface exposed portion of the target protein.

Preferably the binding molecule according to the third aspect of the present invention has a maximum length of 100 amino acids, more preferably less than 50 amino acids, more preferably still, less than 30 amino acids and most preferably less than 22 amino acids.

It is further preferred that the binding molecule according to the third aspect of the present invention consists of one of the sequences selected from SEQ ID NOS: 1-39, 41-53 and 55-86.

The above binding molecules can be used in a number of different ways. For example, the binding molecules may be used in pharmaceutical, medical, veterinary, biological and biotechnology research and development applications. The binding molecules can also be used in research, medical diagnostics, therapy and biotechnology.

Accordingly, in a fourth aspect, the invention provides the binding molecule described above for use in therapy or diagnostics. The present invention also provides the use of the binding molecule described above.

In a fifth aspect, the present invention provides a pharmaceutical composition comprising the binding molecule described above in combination with a pharmaceutically acceptable excipient.

In the above description of the invention, The term binding molecule is extensively used. The skilled person will appreciate that the features of the binding molecule are equally applicable to the peptide which forms the binding molecule. In the majority of cases the terms "binding molecule" and "peptide" are exchangeable. In one embodiment, the binding molecule is a peptide which has a sequence that has substantial identity with at least a substantial part of the sequence of the surface exposed portion of the fully folded target protein thereby allowing the binding molecule to bind to the target protein in place of the surface exposed portion of the target protein.

The present invention will now be described in detail, by way of example only, with reference to the figures in which: Figure 1 shows the selected target protein structure - a fragment of human Fibrinogen sequence shown in the form of a double-D dimer from human fibrin (PDB ID: IFZE, panel A) and the devised artibody (named "FBB204", panel B). The dashed line indicates the position of the selected artibody within the target protein structure. Panel C shows the alignment of the artibody sequence with that of the target protein (Fibrinogen). Panel D shows the specific binding of FBB204 artibody as measured using ELISA assay and Peroxidase conjugated FBB204 (2 measurements). Open bars show the binding of FBB204 artibody to Fibrinogen, filled bars show displacement of that binding by the excess of unlabelled FBB204 (filled bars) and therefore prove the specificity of the interaction.

Figure 2 shows the selected target protein structure - a single subunit of human C- reactive protein (PDB ID: 1LJ7, panel A) and the devised artibody (named "CRP206", panel B). The dashed line indicates the position of the selected artibody within the target protein structure. Panel C shows the alignment of the artibody sequence with that of the target protein (CRP). Panel D shows the specific binding of CRP206 artibody as measured using ELISA assay and Peroxidase conjugated CRP206 (2 experiments). Open bars show the binding of CRP206 artibody to CRP, filled bars show displacement of that binding by the excess of unlabelled CRP206 (filled bars) and dashed bars show the binding of CRP206 artibody to an irrelevant target protein (BSA). Both of the negative controls show nearly identical values and either can be used in the calculation of the specific artibody binding.

Figure 3 shows human C-reactive protein (PDB ID: 1LJ7, panel A) and the devised artibodies (named "CRP206" and "CRP205", panel B). The dashed lines indicate the position of the selected artibodies within the target protein structure. Panel C shows the alignment of the CRP206 artibody sequence with that of the target protein (CRP). Panel D shows the alignment of the CRP205 artibody sequence with that of the target protein (CRP). Panel E shows the structure of the face-to-face decamer multisubunit CRP. The position of the artibodies are indicated with arrows. CRP205 is prevented from binding to the multimerised CRP, whilst CRP206 binds (means of two measurements, panel F). Figure 4 shows human myoglobin (PDB ID: 2MM 1) and the devised binding molecules (artibodies named "MG202", "MG201", "MG301", "NlOl" and "MGlOl", panel A). The dashed lines indicate the position of the selected artibodies within the target protein structure. Panels B,C,D and E show the alignment of the "MG202", "MG201", "MG301" and "NlOl" artibody sequences (respectively) with that of the target protein (myoglobin). Artibody "MGlOl" is constructed from two identical "NlOl" fragments. Panel F: specific binding of MGlOl, MG201, MG202 and MG301 artibodies to human myoglobin (all artibodies were conjugated to peroxidase and were used in a standard ELISA type assay). Open bars show normalised values for total binding. Nonspecific binding was determined using the excess of unconjugated artibodies (filled bars) or by measuring binding to an irrelevant protein BSA (dashed bars).

Figure 5 shows that binding affinity of artibodies is comparable with that of monoclonal antibodies. Panel A: binding of a monoclonal antibody clone [40Fl 1] (from Abeam) to human fibrinogen using ELISA assay. Panel B: binding of FBB204 artibody to human fibrinogen using ELISA assay. Both plots show Ln(binding), arbitrary units (the vertical axes) vs. time in seconds (horizontal axes). The slope of the fitted linear curve shows the dissociation rate kd for each of the affinity reagents tested.

Figure 6 shows that binding of an artibody results in structural changes in the target protein, as indicated by the disappearance of binding of a monoclonal antibody [40Fl 1] to the same protein (in the presence of an excess of an artibody FBB204). Both values for measured using identically loaded, assayed and washed wells.

Figure 7 shows green fluorescent protein GFP (PDB ID: IEMA, Panel A) and the devised artibody (named "GFP307", panel B). The dashed line indicate the position of the selected artibody within the target protein structure. Panel C shows the alignment of the "GFP307" artibody sequence with that of the target protein (GFP). Panel D shows fluorescence spectra of the GFP incubated with and without GFP307 artibody. The excitation was at 387 nm in both cases. Thinner dotted line shown the fluorescence emission spectrum of GFP incubated without artibody, and the thicker dotted line shows the emission spectrum of GFP incubated with GFP307 artibody.

Figure 8 shows the selected target protein structure - a single subunit of human C- reactive protein (PDB ID: 1LJ7, panel A) and the devised artibody (named "CRP206", panel B). The dashed line indicates the position of the selected artibody within the target protein structure. Panel C shows design of the CRP306 artibody, which consists of a 16 amino acid long binding region (consisting of two beta strands and a beta turn), followed by an 8 amino acid long linker, followed by 55 amino acid long fragment of a homo-oligomeric pentamerising peptide from Phe-14 protein (which forms a helix, shown as a cylinder on panel C), followed by a three amino acid long C-terminal fragment to allow labelling, cross-linking or immobilisation of the artibody on surfaces. Panel D shows homo-oligopentamerisation mechanism of the CRP306 artibody.

Figure 9 shows the selected target protein structure - a human leptin (PDB ID: 1AX8, panel A) and the devised artibody (named "LP302", panel B). The position of the disulphide bond is indicated on Panel A. The dashed line indicates the position of the selected artibody within the target protein structure. Panel C shows the alignment of the LP302 artibody sequence with that of the target protein (Leptin).

Figure 10 shows the selected target protein structure - a fragment of human Fibrinogen sequence shown in the form of a double-D dimer from human fibrin (PDB ID: IFZE, panel A), the devised binding region (Panel B, this is similar to the FBB204 artibody, as in Figure 1). Panel C shows the artibody "BIA311", which consists of a binding region (underlined) a linker region and a C-terminal region containing four lysine residues (doubly underlined). The dashed line indicates the position of the selected artibody within the target protein structure.

Figure 11 is a surface plasmon resonance curve showing the binding kinetics of Artibody BIA311, immobilised on BIAcore chip CM5, with Fibrinogen (the target antigen). Horizontal axis - time in seconds, vertical axis - specific binding (RU, arbitrary units). Figure 12A is a surface plasmon resonance curve showing the binding kinetics of Artibody BIA311, immobilised on BIAcore chip CM5. The Artibody is contacted with an irrelevant target (4180 - 4480 seconds), washed with Gly/NaCl solution (4623-4923 seconds), and then contacted with Fibrinogen (5100-5400) with 0.1% n- octyl-β-D-glucoside added. Horizontal axis - time in seconds, vertical axis - specific binding (RU, arbitrary units). Figure 12B shows an enlarged portion of the graph of Figure 12 A.

Figure 13A is a surface plasmon resonance curve showing the binding kinetics of Artibody BIA311, immobilised on BIAcore chip CM5, and contacted with progressively higher concentrations of Fibrinogen. A wash cycle is carried out between each Fibrinogen sample. Figure 13B shows association and dissociation curves for each sample of Fibrinogen (as shown in Figure 13A). The curves are overlaid onto each other; washing/regeneration curves are not shown. Figure 13C shows fitting of the dissociation curves using commercially available software (provided with the Biacore2000 instrument). Figure 13D shows fitting of the dissociation curves using commercially available software (provided with the Biacore2000 instrument). Horizontal axis in all figures - time in seconds, vertical axis - specific binding (RU, arbitrary units).

Figure 14A is a surface plasmon resonance curve showing the binding kinetics of Artibody FBB204, immobilised on BIAcore chip CM5 through the SH- group of an internal cysteine using EDC/NHS and PDEA/Na-borate chemistry, when contacted with 0.016% n-octyl-β-D-glucoside (1760 - 1940 seconds), washed with 1% SDS/0.1M NaOH solution (2060 - 2180 seconds), then contacted with 1 mg/ml Fibrinogen in the presence of 0.016% n-octyl-β-D-glucoside (2300 - 2480 seconds), and followed by the next wash/regeneration cycle (2600 - 2720 seconds). Horizontal axis - time in seconds, vertical axis - specific binding (RU, arbitrary units). Figure 14B shows binding of Fibrinogen to Artibody FB204 depending on the n-octyl-β-D- glucoside concentration. Horizontal axis - concentration of n-octyl-β-D-glucoside in %, vertical axis - specific binding (RU, arbitrary units). Figure 15 is a surface plasmon resonance curve showing the immobilisation of TBA Artibodies on a BIAcore chip using EDC/NHS and PDEA/Na-borate chemistry, followed by 3 x 1 min washes with 1%SDS/1M NaOH. First channel was loaded with TBA405, second - with TBA406, third - TBA 407, fourth - TBA 408. Horizontal axis - time in seconds, vertical axis - specific binding (RU, arbitrary units).

Figures 16 A, 16B, 16C and 16D are surface plasmon resonance curves showing real time binding of the recombinant Iy expressed Botulinum toxin A protein to four different TBA Artibodies immobilise in a BIacore chip. In Figures 16A, 16B, 16C and 16D, time 5920 seconds corresponds to the end of the chip regeneration/wash cycle (SDS/NaOH wash); 6040 seconds is the start of application and binding of the Botulinum toxin to each of the TBA Artibodies; 6160 seconds correspond to the end of association, beginning of the dissociation stage; 6160-8060 seconds is the dissociation stage. Timepoint 8060 seconds is the start of the next regeneration stage (SDS/NaOH wash). Artibody TBA405 is shown in Figure 16A. Artibody TBA406 is shown in Figure 16B. Artibody TBA407 is shown in Figure 16C. Artibody TBA408 is shown in Figure 16D. Horizontal axis in all figures - time in seconds, vertical axis - specific binding (RU, arbitrary units).

Examples

In the following examples, the term "artibody" is used which is comparable to the term "binding molecule" used above. However, the meaning of these words should not be considered to be identical. Further, the terms "artibody forming region" and "artibody forming sequence" are used which are comparable to the term "surface exposed portion" used above.

Example 1. Artibody design against human Fibrinogen (Figure 1)

Example 1 shows that:

- an artibody molecule can be designed against a blood protein;

- an artibody molecule can be designed against a glycoprotein;

- an artibody molecule can be designed against a cardiovascular disease marker; - a single secondary structure element can be used to make a binding artibody molecule;

- a beta strand structure can be used to make a binding artibody molecule;

- a C-terminal fragment of a target protein can be used to make a binding artibody molecule;

- a non-identical variant of the target protein can be used to design an artibody against the target protein if required. The structure used to design the FBB204 artibody was of a fibrin dimer - a shorter processed and dimerised form of fibrinogen - but the artibody can bind unprocessed fibrinogen;

- binding part of the artibody sequence can match the target protein sequence 100%;

- artibody can be covalently attached to another protein (e.g. Peroxidase) through one of the artibody ends;

- artibody can be used instead of traditional antibodies or antibody mimics to bind selected protein targets in an affinity assay;

- that artibody can be used to bind selected protein targets in an enzyme linked affinity assays, such as ELISA;

- that artibody can be designed in a way that quaternary structure of the target protein will not affect artibody binding, (i.e. D-dimer vs. fibriogen)

Detailed description of Example 1 :

Selected protein target: Fibrinogen;

Artibodv forming sequence chosen: MSMKIRPFFPQQ (SEQ ID NO: 1);

Artibody name: FBB204;

Artibody sequence: MSMKIRPFFPQQGSC (SEQ ID NO: 2);

Modification - none, the fragment corresponds to human Fibrinogen sequence;

Labelling - two amino acid- long linker and one other amino acid added (C-term):

MSMKIRPFFPOOGSC (SEQ ID NO: 2) to introduce SH- group for labelling.

Peroxidase was attached covalently to the FBB204 through C-terminal Cysteine;

Assay: direct binding ELISA assay, read at 450 nm;

Result: FBB204 artibody binds Fibrinogen, the binding is specifically displaced by the excess of unlabelled FBB204, thus proving the specificity of the interaction.

Example 2. Artibodv design against human C-reactive protein (Figure 2) Example 2 further shows that:

- an artibody molecule can be designed against an acute-phase protein;

- an artibody molecule can be designed against an inflammatory marker protein;

- an artibody molecule can be designed against a protein linked to diabetes, hypertension, myocardial and cerebral infarcts;

- an artibody can be designed against a protein which is differentially expressed in some forms of cancer, for example colon cancer or in response to the diet, smoking, physical exercise, and behavioural traits;

- an artibody molecule can be designed against a protein which has more than one subunit;

- more than one secondary structure element can be used to make a binding artibody molecule;

- an artibody molecule may contain a beta turn;

- artibody forming region does not have to be at the C-terminus and can be in the middle of the target protein sequence;

- binding part of the artibody sequence may differ slightly from the target protein sequence and does not have to match the target protein sequence 100%;

- artibody can be attached to another protein (e.g. Peroxidase) through one of the internal (not terminal) amino acids;

- the value of the specific artibody binding can be obtained either by comparing total binding and the binding to the same target but measured in the presence of an excess of unlabelled artibody, or alternatively, by comparing the total binding and the binding to an irrelevant protein target.

Detailed description of the Example 2:

Selected protein target: C-reactive protein (CRP);

Artibodv forming sequence chosen: KDIGYSFTVGGSEILFE (SEQ ID NO: 3);

Artibody name: CRP206;

Artibody sequence: KDIGYSFTVGGCEILFE (SEQ ID NO: 4);

Modification - one amino acid (Glycine) substituted for Cysteine (underlined above) - this makes 6% amino acid sequence difference compared to the target protein;

Labelling - Peroxidase was attached covalently to the CRP206 through the Cysteine side chain; Assay: direct binding ELISA assay, read at 450 nm;

Result: CRP206 artibody binds CRP, the binding is specifically displaced by the excess of unlabelled CRP206. Alternatively, artibody binding to BSA protein can be used as the negative control. The above two negative controls prove the specificity of the interaction.

Example 3. Multiple artibodies against human C-reactive protein (Figure 3)

Example 3 further shows that:

- more than one artibody can be designed against a selected protein target;

- that artibody binding can be used to discriminate between different folding states of the target protein, or ligand binding by the target protein, such as for example the structural changes (face-to-face decamer multisubunit formation by CRP from two independent pentamers) associated with Ca ²⁺ binding by CRP;

- that artibody ELISA assays are not limited of a single wavelength and can be read at different wavelength, for example 660 nm;

Detailed description of the Example 3:

Selected protein target: C-reactive protein (CRP);

Artibodv forming sequences chosen: KDIGYSFTVGGSEILFE (SEQ ID NO: 3) and

LSPDEINTIYLGGPFS (SEQ ID NO: 5);

Artibody names: CRP206 and CRP205 respectively;

Artibody sequences: KDIGYSFT VGGCEILFE (SEQ ID NO: 4) and

LSPDEINTIYLGGPFSGSC (SEQ ID NO: 6);

Modifications - one amino acid (Glycine) substituted for Cysteine (CRP206), two amino acid-long linker and one other amino acid added to introduce SH- group for labelling (CRP205). Peroxidase was attached covalently to the cysteines in both artibodies;

Assay: direct binding ELISA assay, read at 660 nm;

Result: CRP206 artibody binds decamer multisubunit CRP, whilst CRP205 artibody is prevented from binding due to the shielding of the binding site by the complex quaternary structure of the decamer multisubunit CRP, showing that artibodies can be used to study structural traits of target proteins. Example 4. Artibodies against human myoglobin (Figure 4)

Example 4 further shows that:

- an artibody molecule can be designed against a protein which is not normally present in the bloodstream;

- an artibody molecule can be designed against a protein which is normally expressed and present in solid tissues;

- an artibody molecule can be designed against a tissue marker;

- an artibody molecule can be designed against a protein which normally has only one subunit;

- an alpha helix structure can be used to make a binding artibody molecule;

- only part of an alpha helix structure can be used to make a binding artibody molecule;

- N-terminal fragment of a target protein can be used to make an artibody molecule;

- artibodies can be designed against discontinuous epitopes ;

- binding part of the artibody sequence does not have to match the target protein sequence 100% and may differ significantly from the artibody forming fragment of the target protein sequence;

- artibody may contain more than one copy of the binding fragment in one artibody molecule;

- a multimeric artibody can be designed by joining multiple artibodies in a single molecule, i.e. covalently.

Detailed description of the Example 4:

Selected protein target: Myglobin;

Artibodv forming sequences chosen: GLSDGEWQL VLNVW (SEQ ID NO: 7) (N- terminal artibody forming fragment) and KDMASNYKELGFQG (SEQ ID NO: 9) (C- terminal artibody forming fragment);

Artibody sequences: CGLSDGEWQL VLNVW (SEQ ID NO: 8) (MG202),

KDMASNYKELGFQGC (SEQ ID NO: 10) (MG201), ASNYKELGFQGSSSC (SEQ

ID NO: 11) (MG301), GNASNYKLGSGSGSC (SEQ ID NO: 12) (NlOl) and

GNASNYKLGSCSGNASNYKL (SEQ ID NO: 13) (MGlOl).

Modifications - N-terminal Cysteine added to introduce SH- group for labelling

(MG202); C-terminal Cysteine added to introduce SH- group for labelling (MG201); three amino acid-long linker and one Cysteine added to introduce SH- group for labelling (MG301); two and five amino acid-long fragments added (N-terminal and C- terminal to the artibody binding site, one amino acid removed from the middle of the artibody forming region and one Cysteine added to introduce SH- group for labelling (NlOl); a dimer of two NlOl artibodies, to which two and five amino acid-long fragments added (N-terminal and C-terminal to the first binding region, one amino acid removed from the middle of each of the two binding regions, a Cysteine is used to introduce SH- group for labelling (MGlOl). NlOl and MGlOl are based on discontinuous sequence epitopes of the target protein. Peroxidase was attached covalently to the cysteines in all the artibodies. Assay: direct binding ELISA assay, read at 450 nm;

Result shown on Figure 4 (panel F) include the MG301 and MGlOl binding to myoglobin. MGlOl artibody which have only 7 amino acid long binding region(s) and -13% sequence difference with the target protein artibody forming region (one amino acid omitted form the 8 amino acid-long fragment) still bind myoglobin protein.

Example 5. Artibody affinity exemplified with FBB204 anti-Fibrinogen artibody (Figure 5)

Example 5 shows that:

- an artibody molecule possesses binding affinity comparable to that of a monoclonal antibody;

Detailed description of the Example 5:

Selected protein target: Fibrinogen;

Artibodv forming sequence chosen: MSMKIRPFFPQQ (SEQ ID NO: 14);

Artibody name: FBB204;

Artibody sequence: MSMKIRPFFPQQGSC (SEQ ID NO: 15);

Modification - none, the fragment corresponds to human Fibrinogen sequence

Labelling - two amino acid- long linker and one other amino acid added (C -term):

MSMKIRPFFPOOGSC (SEQ ID NO: 15) to introduce SH- group for labelling.

Peroxidase was attached covalently to the FBB204 through C-terminal Cysteine;

Assay: direct binding ELISA assay, various incubation times, read at 450 nm; Monoclonal antibody used for comparison - clone [40Fl 1] (from Abeam), peroxidase conjugated antibody;

Assay: identical to that of FBB204 - direct binding ELISA assay, various incubation times, read at 450 nm;

Result: [40FI lI monoclonal antibody binding yields k _d = 5x10 ⁵ s ^"1 , FBB204 artibody binding yielded kd = 9x10 ⁵ s ^"1 (kd is the dissociation rate, not the dissociation constant

(KD))-

Example 6. Artibody binding cause structural changes to the target protein (Figure 6)

Example 6 shows that:

- binding of an artibody molecule may cause structural changes to the target protein;

- an artibody molecule can be designed to disrupt or change the structure of the target protein.

Detailed description of the Example 6:

Selected protein target: Fibrinogen;

Artibodv forming sequence chosen: MSMKIRPFFPQQ (SEQ ID NO: 14);

Artibody name: FBB204;

Artibody sequence: MSMKIRPFFPQQGSC (SEQ ID NO: 15);

Modification - none, the fragment corresponds to human Fibrinogen sequence;

Labelling - two amino acid- long linker and one other amino acid added (C -term):

MSMKIRPFFPOOGSC (SEQ ID NO: 15) to introduce SH- group for labelling.

Peroxidase was attached covalently to the FBB204 through C-terminal Cysteine;

Monoclonal antibody used - clone [40Fl 1] (from Abeam), peroxidase conjugated antibody.

Assay: the binding of monoclonal antibodies in absence and in the presence of the

FBB204 artibody.

Result: FBB204 artibody abolishes binding of the monoclonal antibody [40Fl 1] to human fibrinogen.

Comment: The binding of monoclonal antibodies is abolished either because the binding epitope is shielded or because it disappears. FBB2204 does not bind to the

"outside" of the protein as traditional affinity reagents do, so it can not itself shield the monoclonal antibody binding epitope. FBB204 is designed to displace the corresponding artibody-forming region form the target protein. The only shielding effect therefore can come from the displaced fragment of the native fibrinogen (which proves the change on the target protein structure). Another alternative could be the disappearance of the epitope due to changes in the epitope structure. This would also directly prove the change in the target protein structure in response to the binding of FBB204 artibody.

Example 7. Artibody binding cause functional changes to the target protein (Figure 7)

Example 7 shows that:

- N-terminal fragment of a target protein can be used to make an artibody molecule;

- Artibody can be designed against a protein of animal origin;

- Artibody can be designed against a jellyfish protein;

- Artibody can be designed against a protein which is naturally fluorescent;

- Artibody can be designed against a protein which acquires ability to fluoresce following its recombinant production;

- binding of an artibody molecule may cause functional changes to the target protein;

- Artibody can be used to affect function of recombinantly produced protein;

- an artibody molecule can be designed to disrupt or change the function of the target protein, such as for example, but not limited to R&D and therapeutic applications.

Detailed description of the Example 7:

Selected protein target: Green fluorescent protein (GFP);

Artibodv forming sequence chosen: GEELFTGVVPIL VELDGDV (SEQ ID NO: 16);

Artibody name: GFP307;

Artibody sequence: C GEELFTG V VPIL VELD GDV (SEQ ID NO: 17);

Modification - none, the fragment corresponds to GFP sequence;

Labelling - one Cysteine amino acid added (N-teπn): CGEELFTGVVPIL VELDGDV

(SEQ ID NO: 17) to introduce a SH- group for optional labelling.

GFP: fluorescent GFP protein was produced using in vitro transcription and translation system RTS 100 E. coli HY Kit from Roche. Assay: GFP was incubated with or without GFP307 artibody and with an irrelevant artibody; following an incubation fluorescent spectra were taken.

Result: GFP307 artibody causes changes in the fluorescent properties of GFP, this confirms that artibodies can be used to modulate functional properties of the target proteins.

Example 8. Multimeric Artibody design against human C-reactive protein (Figure 8)

Example 8 further shows that:

- a multimeric artibody can be designed by joining multiple artibodies non-covalently, for example by self-assembly or polymerisation; this can be achieved using molecules, such as but not limited to the elements of protein tertiary and quaternary structure, self-assembling proteins or other molecules capable of self-assembly such as nucleic acids, small molecules, and their combinations;

- protein helixes can be used to produce oligomeric artibodies.

Detailed description of the Example 8:

Selected protein target: C-reactive protein (CRP);

Artibodv forming sequence chosen: KDIGYSFTVGGSEILFE (SEQ ID NO: 3);

Artibody name: CRP306;

Artibody sequence:

KDIGYSFTVGGSEILFsgggsgggssnakfdqfssdfqtfnakfdqfsndfnafrsd fqafkddfarfnqrfdn fatkyKKC (SEQ ID NO: 18);

This includes a 16 amino acid long binding region (shown in capital letters, this is similar to CRP206 artibody), followed by an 8 amino acid long linker (underlined), followed by 55 amino acid long fragment of a homo-oligomeric pentamerising peptide from Phe-14 protein, followed by a three amino acid long C-terminal fragment to allow labelling, cross-linking or immobilisation of the artibody on surfaces (shown in capital letters and underlined).

Example 9. Artibodv design against Leptin protein (Figure 9)

Example 9 further shows that:

- an artibody molecule can be designed against a hormone; - an artibody molecule can be designed against a protein which is processed and has the signal peptide removed by signal peptidases;

- an artibody molecule can be designed against a protein involved in metabolism;

- an artibody molecule can be designed against a protein involved in modulating immune response, angiogenesis, and reproduction;

- an artibody molecule can be designed against a protein produced by adipose tissue, placenta ovaries, skeletal muscle, stomach, mammary epithelial cells, bone marrow, pituitary and liver;

- an artibody molecule can be designed against a protein structure which contains intramolecular disulphide bonds.

Detailed description of the Example 9:

Selected protein target: Leptin;

Artibodv forming sequence chosen: KVQDDTKTLIKTIVTRIN (SEQ ID NO: 19);

Artibody name: LP302;

Artibody sequence: CGKVQDDTKTLIKTIVTRIN (SEQ ID NO: 20);

Modification - none, the fragment corresponds to human Leptin sequence;

Labelling - one amino acid-long linker and one additional Cysteine added at the N- terminus to introduce SH- group for labelling or for the attachment of peroxidase.

Example 10. Correcting the orientation of artibodv immobilisation. Real time affinity assays. (Figure 10)

Example 10 further shows that:

- an artibody molecule can be devised and modified to change the distribution of polar residues along the artibody molecule;

- an artibody molecule can be devised and modified to change the distribution of charged residues along the artibody molecule;

- an artibody molecule can be devised and modified to change the distribution of charge along the artibody molecule;

- an artibody molecule can be devised and modified to achieve directional immobilisation on surfaces;

- an artibody molecule can be immobilised covalently on a solid surface;

- an artibody molecule can be immobilised on a Biacore chip surface; - an artibody molecule can be immobilised using NHS / EDC chemistry;

- an artibody molecule can be covalently linked to carboxylic group;

- an artibody molecule can be used in real time affinity assays;

- an artibody molecule can be used in SPR based affinity assays.

Detailed description of the Example 10:

Selected protein target: Fibrinogen;

Artibodv forming sequence chosen: MSMKIRPFFPQQ (SEQ ID NO: 1);

Artibody name: BIA311 ;

Artibody sequence: MSMKIRPFFPQQGSGSKKKK (SEQ ID NO: 21);

Modification - none, the fragment corresponds to human Fibrinogen sequence;

Immobilisation - four amino acid-long linker and for C-terminal terminal lysines to facilitate the immobilisation of the artibody on a derivatised gold surface of the CM5 chip (Biacore).

Example 11. Typical approach to ARTIBODY® design

The method described below is a typical approach to designing an artibody.

Step l :

If Protein structure is known:

- find structure in the data bank e.g. NCBI structure, PDB.

If Protein structure is not known:

- predict the structure;

- experimentally resolve the structure; or

- use overlapping peptides.

Step 2:

Select suitable C-terminal (or N-terminal or intra-molecular) peptide, preferably corresponding to a discrete element of existing structure, flanked by flexible hinges (e.g. GIy,) or small amino acids (Ala, Ser) - these will be the likeliest places where the native structure will "let go" and bend away. Preserve the native protein structure core by not interfering with the majority of the buried amino acid side chains (mostly hydrophobic), such as large aliphatic, aromatic or sulphur-containing amino acids.

Step 3 (optional):

Accommodate the existing displaced element of the secondary structure by rational design or affinity saturation of the side chains of the devised ARTIB OD Y® facing the solvent and the displaced native sequence.

Step 4 (optional):

Extend the ARTIBODY fragment to include one or a few strong interacting amino acids, such as large aliphatic, aromatic or sulphur- containing amino acids.

Add required modification to improve physical properties of the artibody, or to mimic native post-translational modifications.

Step 5 (optional):

Enhance the structure by introducing disulphide bridges or otherwise, by using known protein scaffolds suitable to accommodate the chosen structure.

Step 6 (optional):

Devise linkers, adapters, labels, other proteins or biologically relevant molecules.

Example 12

Artibody BIA311 was immobilised on BIAcore chip CM5 using NHS/EDC coupling chemistry. This was contacted with Fibrinogen (the target antigen). The polypeptide sequence of the BIA311 Artibody is MSMKIRPFFPQQGSGSKKKK (SEQ ID NO: 21) (the same as that of Example 10).

Figure 11 shows that fibrinogen binds to the BIA311 Artibody with fast kinetics.

This example shows that: - Artibody molecules can be used instead of traditional polyclonal or monoclonal antibodies, or other affinity reagents;

- Artibody molecules can be immobilised on planar surfaces without the loss of binding activity;

- Artibody molecules can be immobilised by covalent binding;

- Artibody molecules can be immobilised by covalent binding through reactive groups such as amino acid side chains;

- Artibody molecules can be immobilised by covalent binding through an amino acid or amino acids located at or close to the termini of Artibody sequence (not situated at the position away from the termini of the Artibody molecule);

- Artibody molecules can be immobilised by covalent binding through lysine amino acid side chains;

- Artibody molecules can be immobilised by covalent binding through lysine amino acids added to the main sequence of the Artibody polypeptide with the aim of covalent cross-linking;

- Artibody molecules can be directed towards the chip surface through the engineering of the desired physico-chemical properties, such as but not limited to their pi;

- Artibody molecules can be used with surface plasmon resonance (SPR) detection equipment;

- Artibody molecules bind their target when immobilised on BIAcore / GE Healthcare sensors;

- Artibody molecules are capable of fast binding rate kinetics;

- Binding of Artibody molecules can reach saturation in time, when dissociation rate will be equal to association rate.

Example 13

Artibody BIA311 was immobilised on BIAcore chip CM5 and contacted with an irrelevant target (4180 - 4480 seconds), washed with Gly/NaCl solution (4623-4923 seconds) and then with Fibrinogen (5100-5400) to which 0.1% n-octyl-β-D-glucoside was added. The polypeptide sequence of the BIA311 Artibody is given in Example 10. Figure 12A shows no specific binding to the irrelevant target (FBB204), that the chip withstands a wash cycle, and that the Artibody binds Fibrinogen. The same sensor surface was used, with the same number of immobilised BIA311 molecules, as in Example 12. FBB204 is an Artibody devised against Fibrinogen and both BIA311 and FBB204 have the same active sequence based on the target Fibrinogen protein. FBB204 does not bind to the sensor surface, indicating that any interaction seen with Fibrinogen is not because the protein possesses the same sequence fragment, but because the Artibody binds to the complimentary protein structure, i.e. the rest of the Fibrinogen, by displacing the targeted region.

Figure 12B shows a zoomed in section of the binding graph (Figure 12A). This figure shows the increased binding of Fibrinogen to the BIA311 Artibody in the presence of n-octyl-β-D-glucoside (compared to binding without the additive as shown in Figure 11).

This Example shows that:

- Artibody molecules can bind their targets specifically;

- Artibody molecules do not bind in absence of their target proteins;

- Artibody molecules do not bind themselves;

- Artibody molecules require the target protein structure to be present to show binding;

- Artibody molecules immobilised on sensors and surfaces can withstand washing and regeneration cycles;

- Artibody molecules could be regenerated by standard washing buffers suitable for washing antigens bound to standard antibodies, such buffers as acidic glycine and/or high salt;

- Artibody molecules can be used to bind their targets in the presence of binding modulators;

- Artibody molecules can be used to bind their targets in the presence of detergents;

- Binding of Artibody molecules to their targets is not negatively affected by detergents;

- Detergents can be used to improve binding of Artibody molecules. Example 14

Artibody BIA311 was immobilised on BIAcore chip CM5 and contacted with a number of Fibrinogen samples (the target antigen) having different concentrations.

In Figure 13 A, the large peaks just below 21,500 correspond to wash cycles (Gly/NaCl solution). Figure 13A shows that BIA311 is regenerated following each binding cycle. Progressively higher concentrations of Fibrinogen result in stronger binding.

Figure 13B shows association and dissociation curves overlaid onto each other; washing/regeneration curves are not shown.

Figure 13C shows fitting of the dissociation curves using commercially available software (provided with the Biacore2000 instrument).

Figure 13D shows fitting of the dissociation curves using commercially available software (provided with the Biacore2000 instrument).

Fitting of the binding curves results in the following binding rates for the BI A311 Artibody: kd ~ 7 x 10 ^~3 ; ka ~ 2 x 10 ³ ; KD ~ 3.5 x 10 ^"6 .

This Example shows that:

- Artibody molecules can bind their protein target and be regenerated/washed in a repetitive manner;

- Artibodies are suitable for use with commercial instruments designed for real time binding analysis;

- Artibody binding kinetics can be analysed using numerical tools and software designed for traditional affinity reagents and protein-protein interaction analysis;

- Artibody binding kinetics parameters are similar to a typical TCR+MHC/peptide binding which is of the order of 10 ^"6 M. Example 15

Artibody FBB204 was immobilised on BIAcore chip CM5 through the SH- group of an internal cysteine using EDC/NHS and PDEA/Na-borate chemistry and contacted with 0.016% n-octyl-β-D-glucoside (1760 - 1940 seconds), washed with 1% SDS/0.1M NaOH solution (2060 - 2180 seconds), then contacted with 1 mg/ml Fibrinogen in the presence of 0.016% n-octyl-β-D-glucoside (2300 - 2480 seconds), and followed by the next wash/regeneration cycle (2600 - 2720 seconds). The polypeptide sequence of the FBB204 Artibody is given in Example 1. The binding kinetics are shown in Figure 14A.

Figure 14B shows binding of Fibrinogen to Artibody FB204 depending on the n-octyl- β-D-glucoside concentration.

This Example shows that:

- Artibody molecules can be immobilised by covalent binding through Cysteine amino acid side chains;

- Artibody molecules can be immobilised using EDC/NHS and PDEA/Na-borate chemistry;

- Artibody molecules can be immobilised by covalent binding through an amino acid internal to the Artibody sequence (not situated at the position close to the C- or N- termini of the Artibody molecule);

- Artibody molecules can be stable and do not loose their binding ability after treatment with very harsh regeneration/washing buffers, such as SDS and NaOH;

- Binding modulators can be used to exert a different effect on Artibody binding depending on their concentration;

- Binding modulators can increase or decrease Artibody binding and the range of modulator concentrations could be used to find the optimum concentration;

- Detergents affect binding of Artibody molecule to their targets and can be used to improve binding of Artibody molecules;

- Detergents can increase or decrease Artibody binding and the range of modulator concentrations could be used to find the optimum concentration. Example 16

Table 1 below shows the fragments of the Botulinum toxin A protein which were used to design Artibodies against Botulinum toxin A (TBA405-408). Numbers in the left hand column refer to the amino acid position in the sequence of the toxin and the original Botulinum sequence (database Ace. No. A5HZZ9 or P 10845) is shown in parenthesis. Artibody sequences are shown on the right. These were synthesised chemically and used for binding Botulinum toxin A protein in ELISA and on BIAcore CN5 chip.

Table 1

(1025-NNSKIYINGRLIDQKPI-1041) TRA405 NNSKIYINGCLIDQKPI (SEQ ID NO: 22) (SEQ ID NO: 23)

(1121-KYVDVNNVGIRGYMYLKGPR-1140) TBA406 KYVDVNNVGICGYMYLKGPR (SEQ ID NO: 24) (SEQ ID NO: 25)

(1140-RGSVMTTNIYLNSSLYR-1156) TBA407 RGSVMTTCIYLNSSLYR (SEQ ID NO: 26) (SEQ ID NO: 27)

(1246-DIGFIGFHQFNNIAKLVASN-1265) TBA408 DIGFIGFHQFNCIAKLVASN (SEQ ID NO: 28) (SEQ ID NO: 29)

The TBA Artibodies were immobilised on a BIAcore chip using EDC/NHS and PDEA/Na-borate chemistry, followed by 3 x 1 min washes with 1%SDS/1M NaOH. This is shown in Figure 15. (First channel was loaded with TBA405, second - with TBA406, third - TBA 407, fourth - TBA 408).

This Example shows that:

- an Artibody molecule can be designed against a bacterial protein;

- an Artibody molecule can be designed against a protein target which is produced by expression in standard expression systems, such as a bacterial expression system;

- an Artibody molecule can be designed against a toxin;

- multiple different Artibody molecules can be designed against the same target protein;

- different Artibody molecules can be immobilised on a single sensor;

- different Artibody molecules are loaded using automated liquid handling systems, for example, the system employed in BIAcore2000 instruments;

- different Artibody molecules can be treated in sequential manner, for example, being loaded one by one; - different Artibody molecules can be treated in parallel, for example, being washed in a single washing step, i.e. in parallel;

- a designed Artibody molecule can differ in its sequence from the respective sequence of the target region of the target protein;

- Artibodies can be washed with stringent washing buffers such as 1%SDS 0.1M NaOH.

Example 17

Four different TBA Artibodies were immobilise in the same BIacore chip (as described in Example 16) and contacted with the recombinantly expressed Botulinum toxin A protein.

Figure 16 shows real time binding of the recombinantly expressed Botulinum toxin A protein to the four different TBA Artibodies immobilise in the BIacore chip. In Figure 16A, 16B, 16C and 16D, time 5920 seconds corresponds to the end of the chip regeneration/wash cycle (SDS/NaOH wash); 6040 seconds is the start of application and binding of the Botulinum toxin to each of the TBA Artibodies; 6160 seconds correspond to the end of association, beginning of the dissociation stage; 6160-8060 seconds is the dissociation stage. No significant dissociation is detectable during that time indicating high affinity of interaction. Timepoint 8060 seconds is the start of the next regeneration stage (SDS/NaOH wash). Artibody TBA405 is shown in Figure 16 A. Artibody TBA406 is shown in Figure 16B. Artibody TBA407 is shown in Figure 16C. Artibody TBA408 is shown in Figure 16D.

This Example shows that:

- multiple different Artibody molecules can be used in a single binding experiment;

- multiple different Artibody molecules can be used on a single solid surface or support;

- multiple Artibody molecules can be washed and regenerated in a single washing step, where the same regeneration or washing buffer is passed over each of the different sensor surfaces; - Artibodies can bind their target and do not wash away easily, indicating good binding affinity.

Example 18

Example 18 description:

Table 2 shows the designed Artibodies against the ionotropic Glutamate [NMDA] receptor protein, subunits NRl and NR2. Receptor sequences used to design the Artibodies are shown on the left. The location of the target regions is near the agonist and antagonist or co-agonist binding sites of this channel- forming receptor. PDB database entries "2A5T" and "3BTA" were used to model the Artibodies. Numbers refer to the amino acid position in the sequence of the receptor and the receptor protein sequence (database numbers are shown on the right of the species names). Artibody names and the sequences of the chemically synthesised Artibodies are shown on the right. Underlined amino acids indicate spacers, linkers and amino acids with chemically reactive groups for subsequent chemical modifications. Control sequences (negative controls for relevant Artibodies also shown).

Example 18 shows that:

- Artibodies can be designed using existing protein structure, for example, available from protein structure databases;

- Artibodies can be designed against human proteins;

- Multiple Artibodies can be designed against human proteins;

- Artibodies can be designed against animal proteins;

- Multiple Artibodies can be designed against animal proteins;

- Artibodies can be designed against animal proteins but can be used on human proteins;

- Artibodies can be designed against human proteins but can be used on animal proteins;

- Artibodies can be designed against one species and can be used for a different species;

- Multiple Artibodies can be designed against matching human and non-human proteins or against matching proteins from different species; - The same Artibody can be designed to target multiple species within the same Genus;

- The same Artibody can be designed to target proteins from different Genus;

- The same Artibody can be designed to have different spacers, linkers and groups suitable for downstream uses, for example, chemical covalent labelling through an - SH group of an introduced cysteine or amino groups of introduced Lysines;

- An Artibody may have one or more chemically reactive site introduced specifically for labelling or cross-linking and immobilisation (e.g. R1N401, R1N403);

- An Artibody can be designed based on the regions of high sequence similarity between different proteins;

- Individual amino acids within Artibody sequences can be changed. For example, individual amino acids exposed to solvent only could be substituted with different amino acids containing different side chains;

- A negative "control" Artibody can be designed based on the same target sequence;

- A negative "control" Artibody can be designed to have an amino acid sequence differing from that of the specific Artibody;

- A negative "control" Artibody can be designed to have the same amino acid composition as that of the specific Artibody;

- A negative "control" Artibody can be designed by reversing the amino acid sequence of the specific artibody;

- An Artibody can be designed to target a membrane protein;

- An Artibody can be designed to target a receptor;

- Artibodies can be designed to target multiple subunits of the same protein target;

- An Artibody can be designed to target an agonist binding site of a receptor;

- An Artibody can be designed to target a co-agonist binding site of a receptor;

- Multiple Artibodies can be designed to target multiple sites of the same protein target;

- An Artibody can be designed to have one or multiple sites suitable for subsequent covalent labelling of a molecular load with the aim of targeting or visualising the protein; these could be for example a fluorescent or radioactive label or a drug;

- An Artibody can be synthesised chemically. Table 2

Glutamate [NMDA] receptor subunit zeta-1 (NR1 or R1N401 QSSVDIYFRRQVELSGC

NMDA-R1): (SEQ ID NO: 30)

686-QSSVDIYFRRQVELS-700,

[Homo sapiens] Q05586; (SEQ ID NO: 31 )

686-QSSVDIYFRRQVELS-700,

[Rattus norvegicus] P35439; (SEQ ID NO: 32)

686-QSSVDIYFRRQVELS-700,

[Mus musculus] P35438; (SEQ ID NO: 33)

707-QSSVDIYFRRQVELS-721 ,

[Canis lupus familiaris] Q5R1 P0; (SEQ ID NO: 34)

702-SSVDMYFRRQVELS-715,

[Drosophila yakuba] B4PVB0; (SEQ ID NO: 35)

699-SSVDMYFRRQVELS-712,

[Drosophila virilis] B4LZB5; (SEQ ID NO: 36)

702-SSVDMYFRRQVELS-715,

[Drosophila sechellia] B4I414; (SEQ ID NO: 37)

702-SSVDMYFRRQVELS-715,

[Drosophila melanogaster] Q24418; (SEQ ID NO: 38) same as above R1N601 QSSVDIYFRRQVELSGSKKK

(SEQ ID NO: 39) negative control for the Artibodies devised against the R1X602 GSLEVQRRFYIDVSSQSKKK sequences above (SEQ ID NO: 40)

Glutamate [NMDA] receptor subunit epsilon-1 (2A R2N403 RFNQRGVEDALVSLKGSC or NMDAR2A): (SEQ ID NO: 41 )

707-RFNQRGVEDALVSLK-721 , [Rattus norvegicus] Q00959; (SEQ ID NO: 42) 708-FNQRGVEDALVSLK-721 , [Mus musculus] P35436; (SEQ ID NO: 43) 708-FNQKGVEDALVSLK-721 , [Pan troglodytes] Q5IS45; (SEQ ID NO: 44) 708-FNQKGVEDALVSLK-721 , [Homo sapiens] Q12879; (SEQ ID NO: 45) 709-FNQRGVDDALLSLK-722, [Canis lupus familiaris] Q5R1 P3; (SEQ ID NO: 46) Glutamate [NMDA] receptor subunit epsilon-2 (2B or NMDAR2B): 709-FNQRGVDDALLSLK-722, [Homo sapiens] Q13224; (SEQ ID NO: 47) 709-FNQRGVDDALLSLK-722, [Mus musculus] Q01097; (SEQ ID NO: 48) 709-FNQRGVDDALLSLK-722, [Rattus norvegicus] Q00960; (SEQ ID NO: 49) Glutamate [NMDA] receptor subunit epsilon-3 (2C or NMDAR2C)

706-FNQRSVEDALTSLK-719, [Homo sapiens] Q14957; (SEQ ID NO: 50) 706-FNQRSVEDALTSLK-719, [Rattus norvegicus] Q00961 ; (SEQ ID NO: 51 ) 706-FNQRSVEDALTSLK-719, [Mus musculus] Q01098; (SEQ ID NO: 52) same as above R2N603 RFNQRGVEDALVSLKSGSKK

(SEQ ID NO: 53) negative control for the Artibodies devised against the R2X604 KLSVLADEVGRQNFRSGSKK sequences above (SEQ ID NO: 54) Example 19

Example 19 description:

Design of Artibodies against Transient receptor potential channel 3 (TrpC3 or TRP-3), for example, see Becker EB, Oliver PL, Glitsch MD, Banks GT, Achilli F, Hardy A,

Nolan PM, Fisher EM, Davies KE. Proc Natl Acad Sci USA. 2009 106(16):6706-l l.

Artibodies were designed against a protein which at the moment of Artibody design did not have any quaternary structure (3D structure) available. The exact transmembrane topology of the protein was also not known. The Artibodies were designed after the structure of this protein was modelled using bioinformatics tools available on-line. At various stages of the Artibody design, the following tools were used. These are listed as examples and the Artibody design is not limited to the use of listed tools. Those with ordinary skill in the art will understand that to practice this invention alternative tools of protein structure analysis and modelling tools could be used. These tool were:

•BLAST sequence comparison and similarity analysis from NCBI;

•Fasta3 - sequence comparison and similarity analysis from EBI;

•PropSearch - Structural homo log search using a 'properties' approach at Montpellier;

•ScanProsite - Scans a sequence against PROSITE or a pattern against the UniProt

Knowledgebase (Swiss-Prot and TrEMBL);

•MotifScan - Scans a sequence against protein profile databases (including

PROSITE);

•ProDom - Compares sequences with ProDom search utility;

•ProtParam - Physico-chemical parameters of a protein sequence (amino-acid and atomic compositions, isoelectric point, extinction coefficient, etc.);

•SignalP - Prediction of signal peptide cleavage sites;

•NetCGlyc - C-mannosylation sites in mammalian proteins;

•NetOGlyc - Prediction of O-GalNAc (mucin type) glycosylation sites in mammalian proteins;

•NetGlycate - Glycation of epsilon amino groups of lysines in mammalian proteins;

•NetNGlyc - Prediction of N-glycosylation sites in human proteins;

•OGPET - Prediction of O-GalNAc (mucin-type) glycosylation sites in eukaryotic

(non-protozoan) proteins;

•YinOYang - O-beta-GlcNAc attachment sites in eukaryotic protein sequences; •PSORT - Prediction of protein subcellular localization;

•TargetP - Prediction of subcellular location;

•TMpred - Prediction of transmembrane regions and protein orientation (EMBnet-

CH);

•ProtParam - Physico-chemical parameters of a protein sequence (amino-acid and atomic compositions, isoelectric point, extinction coefficient, etc.);

•Compute pI/Mw - Compute the theoretical isoelectric point (pi) and molecular weight (Mw) from a UniProt Knowledgebase entry or for a user sequence;

•Protein GRAVY - the GRAVY (grand average of hydropathy) value analysis of protein sequences from the Sequence Manipulation Suite running of JavaScript programs;

•SAPS - Statistical analysis of protein sequences at EMBnet-CH;

•Coils - Prediction of coiled coil regions in proteins (Lupas's method) at EMBnet-CH;

•2ZIP - Prediction of Leucine Zippers;

•ePESTfind - Identification of PEST regions;

•PredictProtein - PHDsec, PHDacc, PHDhtm, PHDtopology, PHDthreader,

MaxHom, EvalSec from Columbia University;

•SWISS-MODEL - An automated knowledge-based protein modelling server;

•CPHmodels - Automated neural-network based protein modelling server;

•DisEMBL - Protein disorder prediction;

•Rasmol and Cn3D protein structure visualization tools;

•PubMed biomedical literature databases and search tools, including protein sequence and structure databases, sequence analysis tools, Conserved Domains database and

UniGene databases available on-line from U.S. National Library of Medicine National

Institutes of Health;

•Sequences and structures were also visualised by printing complete sequences, sequence alignments and Artibody sequences and sequence comparisons on paper for easier visualisation and document archiving.

Artibodies TR3-701 to 706 are designed to bind to the region containing the third Ankyrin repeat and the labile fragment of TRCP3 structure between Ank3 and Ank4 domains. These Artibodies (701, 702, 703, 704, 705, 796, 707) are designed to disrupt the interaction of TRPC3 with its binding partners and affect protein Io causation/targeting .

These Artibodies are designed specifically to target the region shown as situated between "NH2" and transmembrane domain "1" shown on Figure 3 from Becker EB, Oliver PL, Glitsch MD, Banks GT, Achilli F, Hardy A, Nolan PM, Fisher EM, Davies KE. Proc Natl Acad Sci U S A. 2009 106(16):6706-l l.

Artibody TR3-707 is designed to target the region situated after the fourth Ank repeat which is predicted to match TRP 2 superfamily domain, which partially overlaps with the N-term of the predicted coiled-coil domain. This Artibody (TR3-707) is designed to bind to TRPC3 as well as disrupt its interactions with binding partners and affect protein localisation/targeting.

These Artibodies are designed specifically to target the region shown as situated between "NH2" and transmembrane domain "1" shown on Figure 3 from Becker EB, Oliver PL, Glitsch MD, Banks GT, Achilli F, Hardy A, Nolan PM, Fisher EM, Davies KE. Proc Natl Acad Sci U S A. 2009 106(16):6706-l 1 and bind further away form the N-terminus of the protein compared to Artibodies TR3-701 to 706

Artibody TR3-708 is designed to target the region situated between two of the eight predicted transmembrane or membrane-associated domains. At the time of TR3-708 design no accurate knowledge of TRPC protein topology existed. This artibody (TR3- 708) is also designed to help to reveal protein topology and affect TRCP channel properties.

These Artibodies are designed specifically to target the region shown as situated between the assumed transmembrane domains "3" and "4" shown on Figure 3 from Becker EB, Oliver PL, Glitsch MD, Banks GT, Achilli F, Hardy A, Nolan PM, Fisher EM, Davies KE. Proc Natl Acad Sci U S A. 2009 106(16):6706-l 1. Artibody TR3-709 is designed to target the region situated immediately after the proline-rich region and extending nearly to the predicted coiled-coil / IP3R3 binding region.

Artibody PR-709 is designed to affect binding of interacting proteins in this region and to affect function of either of the proteins.

All Artibodies have been made synthetically, all Artibodies were synthesised to contain one fluorescent group (TAMRA) and have protected C-termini. All Artibodies have d-Serine at their C-termini and all but one artibody are amidated at their C- temini. All Artibodies should therefore be suitable for both in vitro and in vivo applications. Most of the Artibodies are strongly acidic, TR3-704 is mildly acidic, TR3-706/707 are basic. TR3-709 is prone to oxidation.

The following section shows the final Artibody sequences and their alignment with human Short transient receptor potential channel 3 (Short=TrpC3; AltName: Full=Transient receptor protein 3; Short=TRP-3). Database number Q13507. The alignments were correct at the data of alignment.

Artibody TR3-701: TAMRA_GKRLTLSPCEQQLQDDDFYAYDEDG (SEQ ID NO: 55)

Query 2 KRLTLSPCEQQLQDDDFYAYDEDG 25 (SEQ ID NO: 56)

KRLTLSPCEQ+LQDDDFYAYDEDG Sbjct 131 KRLTLSPCEQELQDDDFYAYDEDG 154 (SEQ ID NO: 57)

Artibody TR3-702: TAMRA_GQDDDFYAYDEDG (SEQ ID NO: 58)

Query 2 QDDDFYAYDEDG 13 (SEQ ID NO: 59)

QDDDFYAYDEDG Sbjct 143 QDDDFYAYDEDG 154 (SEQ ID NO: 59)

Artibody TR3-703: TAMRA-GESKTLNVKENLARGGSKRKTKSPCEQQLQDDDFYAYDEDG (SEQ ID NO: 60)

Query 15 GGSKRKTKSPCEQQLQDDDFYAYDEDG 41 (SEQ ID NO: 61)

SKR T SPCEQ+LQDDDFYAYDEDG Sbjct 128 AASKRLTLSPCEQELQDDDFYAYDEDG 154 (SEQ ID NO: 62)

Artibody TR3-704:

TAMRA SNHPGFAASKRLTLSPCEQEG (SEQ ID NO: 63) Query 2 NHPGFAASKRLTLSPCEQE 20 (SEQ ID NO: 64)

NHPGFAASKRLTLSPCEQE Sbjct 123 NHPGFAASKRLTLSPCEQE 141 (SEQ ID NO: 64)

Artibody TR3-705 :

GQDDDFYAYDEKGTRFSPDG (SEQ ID NO: 65)

Note that this Artibody contains fluorescent label (TAMRA) attached to the side chain of the only "Lys" in the middle of the peptide (underlined).

Query 2 QDDDFYAYDEDGTRFSPD 19 (SEQ ID NO: 66)

QDDDFYAYDEDGTRFSPD Sbjct 143 QDDDFYAYDEDGTRFSPD 160 (SEQ ID NO: 66)

Artibody TR3-706: TAMRA_SKGYVRIVEAILNHPKG (SEQ ID NO: 67)

Query 1 SKGYVRIVEAILNHP 15 (SEQ ID NO: 68)

SKGYVRIVEAILNHP Sbjct 111 SKGYVRIVEAILNHP 125 (SEQ ID NO: 68)

Artibody TR3-707: TAMRA_GCRDSFSHSRSRINAYKG (SEQ ID NO: 69)

Query 4 DSFSHSRSRINAYK 17 (SEQ ID NO: 70)

DSFSHSRSRINAYK Sbjct 206 DSFSHSRSRINAYK 219 (SEQ ID NO: 70)

Artibody TR3-708: TAMRA_SYVQESDLSEVTLPPEIQYFTYARDKWLP (SEQ ID NO: 71)

Query 1 SYVQESDLSEVTLPPEIQYFTYARDKWLPS 30 (SEQ ID NO: 72)

SYVQESDLSEVTLPPEIQYFTYARDKWLPS Sbjct 503 SYVQESDLSEVTLPPEIQYFTYARDKWLPS 532 (SEQ ID NO: 72)

Artibody TR3-709:

TAMRA_GGCRRRRLQKDIEMGMGNSK (SEQ ID NO: 73)

Query 1 CRRRRLQKDIEMGMGNSK 18 (SEQ ID NO: 74)

CRRRRLQKDIEMGMGNSK Sbjct 729 CRRRRLQKDIEMGMGNSK 746 (SEQ ID NO: 74)

Example 19 further shows that Artibodies can be designed using existing on-line tools and databases, JavaScript programs and published scientific papers.

Example 20

Design of Artibodies against Prostate-specific antigen (PSA). The Artibodies were designed using the same approach as described in previous examples. PSA Artibodies are designed to target more than one splicing variant of PSA. Example 20 shows that:

- An Artibody can be designed to target PSA;

- An Artibody can be designed to target a human protein which is a forensic marker;

- An Artibody can be designed to target a human protein which is a disease marker;

- An Artibody can be designed to target a human protein which is a marker of cancer;

- An Artibody can be designed to target a human protein which is a known marker of prostate cancer;

- An Artibody can be designed to target a protein produced by the cells in a gland;

- An Artibody can be designed to target a protein produced by the cells in an exocrine gland such as the prostate gland;

- An Artibody can be designed to target a protein produced by the cells in female breast;

- An Artibody can be designed to target a protein which is present in small quantities in the serum under normal conditions;

- An Artibody can be designed to target a protein which is present in small quantities in the urine under normal conditions;

- An Artibody can be designed to target a protein which is in larger quantities in semen;

- An Artibody can be designed to target a protein which is elevated in response to healthy physiological activity, such as sexual intercourse, followed by ejaculation, which usually accompanies male orgasm.

- An Artibody can be used in the existing tests available for the early detection of prostate cancer;

- An Artibody can be designed to target a glycoprotein;

- An Artibody can be designed to target an enzyme;

- An Artibody can be designed to target a protease;

- An Artibody can be designed to target a serine protease;

The section below shows the final Artibody sequences for targeting PSA protein and the introduced modifications:

PSA-801-C:

GEVVHYRKWIKDTICANP (SEQ ID NO: 75) PSA-802-C:

CSGEP AQLTD AVGG- Amid (SEQ ID NO: 76) note, Amidated C-term

PSA-803-C:

LPDQEP ALSEC-Amid (SEQ ID NO: 77) note, Amidated C-term

PSA-804-C:

Acet-S VILLGRHSLCHPEDTGQ VFQG- Amid (SEQ ID NO: 78) note, Acetylated N-term; Amidated C-term

PSA-805-C:

KEPCEFLTPG-amid (SEQ ID NO: 79) note, Amidated C-term

PSA-806-C:

(Acet)-ECTGQ VFQVSHSFPHG (SEQ ID NO: 80) note, Acetylated N-term;

PSA-807-C:

(Acet)-YDMSLLCNRFL-(Amid) (SEQ ID NO: 81) note, Acetylated N-term; Amidated C-term

PSA-808-C:

(Acet)-YDMSLLKNRFLCPGDD-(Amid) (SEQ ID NO: 82) note, Acetylated N-term; Amidated C-term

PSA-809-C:

(Acet)-YDMSLLKNRYLCGGSKGISSALS (D-Serine) (SEQ ID NO: 83) note, Acetylated N-term; D-Serine at C-term

APSA-810-C:

(Acet)-GYTFTTYYRCEYGFNSGYSSFDGDS-(Amid) (D serine) (SEQ ID NO: 84) note, Acetylated N-term; Amidated C-term; D-Serine at C-term