Background Hitherto the only small molecules known to bind to ligand binding and receptor proteins were those natural molecules of biological origin which are known to bind to the usually highly specific functional sites of such proteins, or synthetic analogues of those natural molecules. Common examples of these small molecules are the haptens, polypeptides and epitopes binding to the variable or complimentarity determining region (CDR) of antibodies and other immunoglobulins, biotin binding to avidin or streptavidin, glucose binding to concanavalin A and the tachykinins binding to tachykinin receptors. Whilst other protein molecules are known to bind to other sites on such ligand binding and receptor proteins, such as Protein A or G binding to immunoglobulins, or antibodies raised to bind to epitopes on the surface of many proteins, these binding proteins are of a particularly large size which limits their usefulness in separation, detection and treatment.

The use of antibodies, another major tool in the life sciences, relies heavily on affinity chromatography as a purification tool. In particular, antibody purification by immobilised Protein A is a frequently-used separation technique both in the laboratory and for pilot scale manufacture.

Protein A, which is isolated from the cell walls of the pathogenic bacterium <BR> <BR> <BR> <BR> <BR> Staphylococcus aureus, and Protein G, which is isolated from the cell wall of a ß haemolytic Streptococcus G strain, are extensively used as ligands for the affinity purification of polyclonal and monoclonal antibodies. A number of affinity supports to which these proteins are immobilised are available. Protein A and Protein G offer some differences in selectivity for the source and subtype of the antibody to be purifie. The use of proteins such as Protein A and Protein G in affinity chromatography is summarised in an article by S R Narayaman (Journal of Chromatography 658, 1994, pages 237-258).

Large proteins such as Protein A and Protein G are fragile molecules and their biological activity is prone to changes in their protein structure. Thus the storage and use of such protein matrixes requires expert handling and careful attention. Retaining the biological activity after each purification cycle is essential for the re-use of affinity chromatography columns. However, each purification cycle involves the use of harsh conditions (such as low pH) which can denture protein structure. Consequently, affinity chromatography columns uslng Protein A or Protein G have a limited life, which is progressively reduced in normal use and accelerated by improper maintenance or use in difficult conditions.

Commercially available Protein A and (in particular) Protein G are relatively expensive. W R Trumble et al (Protein Engineering, 7, No. 5 1994, pages 705-713) have investigated whether shortening the protein used on an affinity column would reduce the likelihood of the protein structure of the protein being changed in use or storage, and report an attempt to identify the minimum portion of a monodomain IgG Fc binding protein that retained Fc binding ability. This paper indicates that the smallest Fc binding protein which could be produced would be of the order of 45-55 amino acids long. Another paper has recently reduced this sequence to 33 residues (Proc. Natl. Acad. Sci. 93,5688-5692;1996, Biophysical Journal 1992,62,87-91.). Frick et al (Proc. Natl. Acad. Sci. USA 1992,89, 8532-8536 derived a 11 mer peptide from Protein G and found it to bind specifically to IgG Fc site. <BR> <BR> <BR> <BR> <BR> <BR> <P>A further paper (J. Neuroimmunol. 1993+ 48: 2,199-203) has shown that the covalently membrane bound 12-28 amino acid domain, derived from the natural amyloid polypeptide, binds Immunoglobulin G at its hinge region. This 17 residue polypeptide however binds with very high affinity resisting dissociation by denaturants. An analogue of the 17 mer polypeptide with almost identical hydropathic profile also showed binding to Ig G while a control polypeptide having a scrambled amino acid sequence and thus different hydropathic profile, shows minimal binding.

G Fassina et al (J. Molecular Recognition 1996,9,561-569) have identifie a synthetic tetrameric tripeptide which mimics Protein A in its ability to recognise the Fc portion of immunoglobulin G.

However, this tetrameric polypeptide is not a linear animo acid sequence. It comprise 4 copies of a 3-mer peptide (YTR) on a common glycine core which branches with lysyl residues and in addition the latter tetrameric polypeptide is shown to bind specifically to the Fc fragment of IgG. In this context it is also noteworthy that Sloostra et al (Molecular Diversity, 1 (1995) 87-96) made all possible trimers (8000 peptides) and carried out screens against three different antibodies. These studies reveal that the only linear sequence trimer peptides which bound the antibodies were those which corresponded to mimic the linear or non linear part of the native epitopes. The YTR linear peptide was not identifie as an antibody binding sequence.

Other groups have produced paralogs, which are short polypeptides that simulate the binding site for the antigen on a molecule antibody. Such polypeptides have high specificity for the antigen, and are reviewed in the article by Narayaman (supra), but require knowledge of an antibody amino acid sequence and are therefore not appropriate for general use. Peptides which bind specifically to a protein have been identifie using Phage or chemical libraries screened against known specificities or binding receptors (Eur.

J. Biochem. 1974,43 71-375. Anal Biochem. 1979,97,302-308. Biotech. Bioeng 1995,47, 288-297.). Again such peptides are specific ligands of those proteins.

The inventors have surprisingly found short polypeptide sequences which have the ability to bind both antibodies and several other proteins, typically those with ligand binding and receptor functions, but not to the majority of enzymes tested.

The protein binding polypeptides can be distinguished according to this invention in three important and previously unknown and unexpected ways: their particularly small size (between 2 and 50 and typically 4 to 30 amino acids); their binding at a different site to those sites known to define the function of the proteins bound; and their binding substantially to two or more unrelated proteins which may or may not have ligand binding functions. Typical proteins bound by the polypeptides are immunoglobulins (eg antibodies) and related receptors (eg antibody receptors on B cells, and T cell receptors on T cells of the cellular immune system); immunoglobulin binding proteins (eg Protein A, Protein G), lectins (eg concanavalin A), vitamin binding proteins (eg avidin, streptavidin). The polypeptides of the invention are advantageous in the fields of separation, detection and treatment, particularly in applications involving the binding of the polypeptides to proteins with protein binding and receptor functions.

Disclosure of the Invention A first aspect of the invention provides a protein binding polypeptide not directly derived jal protein-bindin protein, the protein binding polypeptide comprising 2-50 amino acids preferably 4 to 30. The polypeptide preferably comprises less than 17 amino acids. In a preferred embodiment, the polypeptide comprises 13 amino acids.

The precise nature of the polypeptide protein interaction in the present invention may be elucidated by detailed structural analysis. It is possible that the polypeptides of the present invention may interact with themselves via aggregation or with other proteins. The binding and dissociation may in many, but not all cases, be governed by aggregation phenomena. Indeed the polypeptides in the invention may influence aggregation properties of other proteins including ß amyloid polypeptides.

The term"not directly derived"from a protein binding protein means that the sequence identity between polypeptide and any stretch, especially a contagious stretch, of sequence taken from said protein is under 90%. The term"polypeptide"is used herein in a broad sense to indicate that a particular molecule comprises a plurality of amino acids joined together by peptide bonds. It therefore inclues within its scope substances, which may sometimes be referred to in the literature as peptides, polypeptides or proteins.

The polypeptide may be further truncated below 13 amino acids.

Preferably, the amino acid sequence of the protein binding polypeptide inclues at least two of non polar amino acids which are separated by n polar amino acids where n is 0 or 1 and the said sequence has the ability to bind to at least two or more unrelated other proteins and where the full length of the sequence bears less than 90% identity with any stretch of sequence present in the said protein Preferably, the protein binding polypeptides of the present invention include at least one Gln residue adjacent to at least one non-polar residue.

The protein binding polypeptides of the present invention may be non-branching and are capable of binding to a greater range of polypeptides guch as lectins, vitamin binding proteins and immunoglobulin binding proteins (including Protein A itself).

The polypeptide may bind to sites on a protein which do not compete with natural protein active sites. For example, the polypeptide preferably binds an antibody at a site or sites not competing-with the antigen binding site of the antibody.

The polypeptide of the invention has several avantages over the prior art protein-binding protein-molecules such as Proteins A and G.

The polypeptide of the present invention is distinguished from the polypeptides described in J. Neuroimmunol (1993) supra. Unlike the peptides of the present invention, these peptides were membrane bound polypeptides binding specifically to IgG at hinge region only and failed to show binding to Fc or Fab fragments. There are further distinctions that the amyloid peptides required specific hydropathic profile, were longer in nature and showed reluctance to reversing the binding in the presence of strong denaturing conditions.

Due to its smaller size, the polypeptide of the invention has a higher capacity than conventional protein binding proteins that is to say it binds more protein per unit weight of polypeptide. The polypeptide of the present invention has been found to bind strongly to enzyme labelle rabbit and goat IgG antibody, and native antibodies from a range of animal species (goat, human, dog, cat, horse), but not to the common enzymes used in labels in such immunological procedures.

The use of immobilised proteins to purify other proteins by affinity chromatography is well known in the art. Affinity chromatography, which is based upon the ability of molecules in solution to bind specifically to immobilised ligands or receptors on solid phase is simple in concept. It is performed in a column containing a ligand derivatised matrix, molecules to be separated from crude preparations binding specifically and tightly to the matrix, whereas most of the contaminants, which lack specific binding sites, are washed away. The specifically absorbe molecules are then eluted with desorbing agents and collecte.

Affinity chromatography has become an important method for the purification of molecules for use as research probes, diagnostic tools and therapeutic agents.

In the case of relatively non-specific ligands (eg dyes and other short polypeptides), the binding of proteins is typically not very efficient, serving only to resolve some of the desired proteins into"bands"or separated fractions eluting from separation media. In contrast, the polypeptide ligands according to this invention typically allow the contaminants to be washed away whilst retaining the desired proteins on or within the separation media.

The short length of the polypeptide sequence means that the polypeptide is more stable in the harsh conditions used in affinity chromatography columns and also has improved storage properties. It is an important aspect of this invention that such harsh conditions can be avoided which is of particular avantage to the proteins being separated. Typically bound proteins can be eluted from the polypeptides of the invention using reagents such as dilute acid solutions or by using buffers or electrolytes of higher ionic strength than used to bind the polypeptides. In one example, antibodies bound onto one polypeptide in accordance with the invention in weak buffer (eg 10 mM Tris HCI pH 7) can be eluted in a stronger buffer. It will be apparent to those skilled in the art that loading and elution buffers of this type can be designed for particular proteins and polypeptide combinations according to this invention.

The polypeptides of this invention may be made by recombinant DNA methods.

Alternatively, the polypeptides of the invention may be made synthetically. This reduces the risk of pyrogenic substances, typically from the cell envelope of bacteria, contaminating a particular product when the polypeptides of the invention are used with affinity columns.

Conventional protein binding proteins Protein A and Protein G are derived from the cell envelope of pathogenic bacteria, and may comprise such pyrogenic substances as contaminants. The pyrogenic substances produce fever in animals, which often makes traditional protein preparations unsuitable for medical use or requires careful and extensive production and quality analysis procedures.

The production of the short polypeptides of the invention by synthetic means has the additional avantage that the cost of producing them is significantly reduced in comparison to the conventional isolation of known protein binding polypeptides. This results in a significant reduction in the cost of affinity chromatography columns including the polypeptides of the invention compare with conventional columns.

Production of the polypeptides of the invention synthetically has the further avantage that the structure of the polypeptide can be altered by known techniques to improve its stability.

For example, some or all labile bonds of the polypeptide can be chemically modifie to prevent atAack by proteolytic enzymes, thus making the polypeptide non-biodegradable and resistant to microbiological attack. This is very difficult to achieve with native Protein A or Protein G even through recombinant DNA technology.

The structure of the polypeptide may be modifie to enhance its binding properties. For example, residues such as cystine may be introduced into the sequence via which residues the polypeptide can be attache to an affinity chromatography matrix.

In a preferred embodiment, the polypeptide is capable of binding an antibody or a protein enzyme conjugate. Preferred protein-enzyme conjugates inclues: Protein A-horseradish peroxidase (HRP), concanavalin A-HRP, and avidin-HRP. Such conjugates find application in immunoassays and related immunological procedures.

A second aspect of the invention provides a polypeptide including the following amino acid sequence: TRNGQVLQGAIKG and functional equivalents thereof.

A third aspect of this invention provides a polypeptide including the following amino acid sequence: GQVLQGAIKG and functional equivalents thereof.

A fourth aspect of this invention provides a polypeptide including the following amino acid sequence: DMHDFFVGLM and functional equivalents thereof.

A fifth aspect of this invention provides a polypeptide including the following amino acid sequence: APVGTDKELSDLLDF and functional equivalents thereof.

A sixth aspect of the invention provides a polypeptide including the following amino acid sequence: SRAQILQAAG and functional equivalents thereof.

A seventh aspect of the invention provides a polypeptide including the following amino acid sequence: KIGQFLIQFAGAFLSILQGLTLRAAEKQAG and functions equivalents thereof.

Functional equivalents include polypeptides comprising additions, deletions, and/or substitutions to the above sequence having the same or similar protein binding abilities as the above polypeptide. The determination of functional equivalents of the above sequence is within the scope of the skilled worker. For example a polypeptide having the sequence AIKG derived from the polypeptides of the second and third aspects of the invention binds protein.

Functional equivalents of the polypeptide of the invention may have improved protein binding abilities. One or more amino acids of the polypeptide may be replace by an amino acid having similar properties.

Amino acids having similar properties include Amino acids having aliphatic side chains: gly, ala, val, leu, ile, pro.

Aliphatic hydroxyl amino acids: ser, thr.

Aromatic amino acids: phe, tyr, trp Basic amino acids: lys, arg, his, asn, gln.

Acidic amino acids: asp, glu.

Sulphur containing amino acids: cys, met.

Non-polar amino acids are amino acids not including acidic and basic amino acids.

The polypeptide may also contain synthetic amino acids such as aminoadipic acid, aminobutyric acid, desmosine, sarcosine, norvaline, norleucine and ornithine, ß -alanine, homocysteine, citrulline, cyclohexylalanine, chlorophenylalanine, cystine, dehydrproli ne, homocitrulline, homoserine, hydroxyproline, fl hydroxyvaline, penicillamine, statine.

Preferably the polypeptide binds to a site or sites on a protein which do not compete with the natural protein active sites, that is to say the normal function of the active site is not substantially affecte. Where the protein is an antibody, the polypeptide may bind to the antibody at a site or sites which do not compete with a normal antigen binding site of the antibody.

In a preferred embodiment, polymers of a polypeptide according to the invention may be prepared. These have the avantage that several protein binding sites may be provided on the same molecule at the same time. The polymer may be a homopolymer or a eopolymer of a polypeptide in accordance with the invention together with another suitable polypeptide.

In a preferred embodiment, a polypeptide of the invention is used to modify another protein, by incorporation of the sequence of the polypeptide into the amino acid sequence of the protein, so that that protein can bind to other proteins. Proteins which may have the polypeptide of the invention added or substituted to their amino acid sequence include Protein A and Protein G. This improves the protein's ability to bind antibodies and other proteins, and increases its binding affinity to other molecules. Alternatively, the polypeptide sequence of a polypeptide in accordance with the invention may be added to other proteins to enable novel protein conjugates to be made.

The polypeptides of the invention may be used as affinity chromatography agents.

According the polypeptides of the invention may be conjugated to one or more solid materials suitable for use in affinity chromatography such as, acrylic polymers cross-linked dextran, silica, glass, agarose, methacrylamide-methylbisacrylamide, cellulose, vinyl polymers, polyacrylamide or combinations thereof. The polypeptides may be covalently or non-covalently attache to such substrats by any means known in the art.

For instance the sequence GQVLQGAIKG can be assemble on solid phase removing a portion for testing after each amino acid. In this example a Lys residue is incorporated with temporary side chain protection such as Fmoc which can be removed (prior to testing) with 20% piperidine in DMF without peptide cleavage from the resin In this way the inventors were able to scan the whole of GQVLQGAIKG polypeptide of the invention and found that relative to control expriment even the short sequences for instance AIKG are able to bind proteins.

Preferably the polypeptides of the invention are used to purify antibodies, or similar immunoglobulins, such as T cell receptor, lectins, streptavidin, avidin, or fragments or derivatives thereof, their ligands, ligands, substrates, antigens or other analytes.

Antibody binding proteins, such as protein A and G, are also widely used in the implementation of diagnostics tests or assays, including those assays, tests and monitoring methods using biosensor devices (eg surface plasmon resonance, surface acoustic wave, or other such microelectronic, optoelectronic devices). It will be apparent to those skilled in the art that the polypeptides of the invention may similarly be used, where their unique ability to bind a broad range of commonly used diagnostic molecules very close to the active surface of such sensor devices is particularly advantageous. The small size of the polypeptide of the invention enables further improvements to be made in such assays, tests and monitoring methods, particularly those where the distance between reacting components needs to be short, which methods are typically called proximity assays.

Another typical example would be the attachment of a fluorescent dye (e. g. dansyl) to the polypeptide such that the binding of the peptide-dye complex to the protein changes the fluorescence measured. Alternatively, the natural (e. g. tryptophan) fluorescence of the protein may be used or fluorescence may be introduced into the polypeptide. In both cases, it will be apparent to those skilled in the art that the fluorescence of the polypeptide or protein may be coupled in such a way as to change the nature of the fluorescence (its intensity, wavelength of emission or excitation, or the time scale or polarisation of the fluorescence). In the particular case of coupling, the fluorescence between two or more molecules, it is of particular avantage that the polypeptide enables the distance between the two molecules to be much shorter than would be possible using a protein such as Protein A or G. While the polypeptide may not bind to the functional site of proteins, its binding to the protein can present a molecule, such as a fluorophore, sufficiently close to the functional site that the fluorophore responds to binding processes occurring at the functional site. For example, the well known processes of fluorescence quenching or resonance energy transfer may be used. It will be similarly apparent to those skilled in the art that molecules or other materials may be attache to the polypeptide so as to interfere with the normal operation of the functional site of the protein.

In some cases, it can be advantageous to bind the polypeptide covalently to its site of binding on the protein molecule by procedures well established in the art, which typically involve the use of heterobifunctional cross-linking agents, whose rection to couple the polypeptide to the protein may include photochemical methods.

Similar methods used in separation and diagnostic tests and assays may be used in the context of microbiological, animal cell or viral diagnostics tests, assay and monitoring procedures. Some of these cells can possess the proteins to which the polypeptides of the invention bind, or the proteins can be introduced to bind to the micro-organisms, cells or viruses, by methods well known to those skilled in the fjeld.

In a similar fashion, the polypeptides of the invention can be used to treat micro-organisms, cells or viruses by attaching a bioactive agent, drug or their carriers to the polypeptide by well established methods.

The polypeptides of the invention can also be introduced onto the surface of larger drug molecules, such as those produced by biotechnology processes, commonly known as biopharmaceuticals, cells, micro-organisms, viruses, macromolecules, polymers or other particles or materials, such as medical implants, that are introduced into biological samples or the body of an animal. A particular problem in these procedures is that the molecule, cell, particle or material so introduced is often treated as foreign by the animal body, such that various processes (e. g. immune responses) result in unfavourable rections in the body.

One of the early phases of these unfavourable responses is the attachment of proteins, such as antibodies, which label or opsonise the foreign matter introduced, which provokes the unfavourable response. The binding of proteins, such as immunoglobulins, by the polypeptides of the invention attache to such foreign matter by the processes described in this invention, notably not involving the functional site of the oposonising protein (e. g. antibody), may prevent or minimise the unfavourable biological response. The foreign matter bearing the polypeptides bind proteins present in the host which are not recognised as foreign by the host, and furthermore may bind them in such a way that they do not present their normal labelling or opsonisation function.

As the polypeptide of the invention is able to bind to more than one protein at once it is possible to target one protein to another via interaction with the polypeptide. Similarly an antibody bound to the polypeptide of the invention may be targeted to specific site and another protein could then be targeted to the same site, and vice versa, by interaction with the polypeptide.

The binding of the polypeptide to proteins may influence the specific functional properties of those proteins and this can be exploite to control the function of protein. In one embodiment the aggregation of Alzheimer polypeptides may be controlled by binding to polypeptides of the present invention leading to treatment by minimising fibril formation.

It is known that some small synthetic molecules (e. g. dyes) do bind to protein molecules.

These are distinguished from the polypeptides of the invention by their binding to a much wider or different range of proteins (e. g. including enzymes) and their binding at both functional and other sites on the protein. In principle, short polypeptides may also be designed to bind to proteins in the manner exemplified by the above dyes, for example by presenting the basic, acidic or hydrophobic properties of the amino acid side chains, which would be expected to bind to many such sites on many protein molecules. Similarly polypeptides can be designed to bind to particular motifs on a protein. However, it will be clear to those skilled in the art that the polypeptides according to this invention demonstrate markedly different and unexpected properties. Indeed, the polypeptides according to this invention may indicate the presence on ligand binding and receptor proteins of a previously unknown common or similar site or structural motif, which is substantially absent at least from one other major class of proteins with catalytic functions- the enzymes.

Brief Description of the Drawings The polypeptides in accordance with the invention and their production will now be described by way of example only, with reference to the accompanying drawings Figures 1 to 11 in which: Fig. 1 is an HPLC trace of purifie peptide TRNGQVLQGAIKG; Fig. 2 shows antibody binding to 10mer peptide GQVLQGAIKG and its Ala scan derivatives; Fig. 3 is binding and elution profile from peptide affi-prep-10 column.

Fig. 4 is a fluorescence spectrum recorde by excitation at 490nm showing quenching of FITC fluorescence by anti-FITC in the presence and absence of the 10 mer peptide; Fig. 5 shows binding of HRP, GARP, Fc fragment and Fab fragment to 10mer peptide GQVLQGAIKG. Control experiments in which no peptide was present are marked by C) Fig. 6 Shows binding of goat anti rabbit peroxidase to 11 mer peptide (NDNGVDGETWY) derived from natural antibody binding protein (Proc. Natl. Acad. Sci. USA 1992,89, 8532-8536) compare to peptide GQVLQGAIKG. Control experiments in which no peptide was present are marked by 0.

Fig. 7 shows binding of peroxidase conjugated antibody and fragments Fab and Fc to peptide GQVLQGAIKG which has been immobilised on affinity matrix affi-prep 10.

Control experiments in which no peptide was present are marked by 0.

Fig 8. Shows association and dissociation of different concentrations of Goat IgG with BSA-peptide conjugate immobilised on CM 5 chip. The curves from top to bottom are for IgG conce rations 1,0.8,0.6,0.4,0.2,0.1,0.05 and 0.025 RM respectively.

Fig 9. Shows association and dissociation of different concentrations of Goat IgG with multimeric 10 mer peptide immobilised on CM 5 chip. The curves from top to bottom are for IgG concentrations 1,0.8,0.6,0.4,0.2,0.1,0.05,0.025,0.0125 and 0.00625, uM respectively.

Fig. 10. Shows binding of goat antirabbit peroxidase (GARP) to a 30 mer peptide (KIGQFLIQFAGAFLSILQGLTLRAAEKQAG) and also a conjugate of 10 mer peptide (GWVLQGAIKG) with BSA measured by ELISA.

Fig. 11. Shows binding of GARP to a peptide (SRAQILQQAG) sequence taken from Flagella protein (J. Mol. Biol 1991 219: 471-480) and the same figure also shows that the protein itself does not bind.

Fig. 12 is a chromatographic profile indicating binding and elution of Goat IgG from multimeric peptide (GWVLQGAIKG) column.

The preparation of polypeptides in accordance with this invention is now described by way of example only.

Example 1: Preparation, purification and characterisation of polypeptides.

Many methods are known for synthesising polypeptides by solution phase and solid phase chemistries. The polypeptide can be readily prepared by solid-phase synthesis as follows using well established protocols. For instance we used Boc chemistry developed by Merrifield to synthesise a polypeptide having the sequence TRNGQVLQGAIKG.

MBHA resin (0. SmMoles) was used. The side chain protecting group for Lys was 2-ClZ.

Each synthetic cycle consiste of (i) a 2min and 25min deprotection with 50% TFA/DCM (ii) neutralisation with 5% DIPEA/DCM and (iii) coupling with l. 5mMoles amino acid, l. 5mMoles BOP and 4. immoles DIPEA in DMF for 40 mins. A second coupling was used when necessary to drive the rection to almost completion (>99.8% yield). At the end of synthesis the polypeptide was cleaved with HF by known procedure. Typically the polypeptide resin was treated with 20mol HF, 0. 5g thiocresol and 0.75g p-cresol and after evaporation of HF, extraction was carried out with 50% acetic acid/water. The polypeptide was purifie on C-8 reverse phase Vydac semi-prep column using linear gradient of 20% acetonitrile/0.1 % TFA to 80% acetonitrile/0.1 % TFA over 45 mins. The product peak was lyophilised and analyse by HPLC.

Other sequences accrding to the invention could be similarly produced. The side chain protected amino acids used in other sequences were Boc-Arg (Tos)-OH, Boc-Asp (OcHx)-OH, Boc-Glu (OBzl)-OH, Boc-Lys (2-CL-Z)-OH, Boc-Lys (Fmoc)-OH, Boc-Ser (BzL)-OH and Boc-Thr (Bzl)-OH.

Biotin could be coupled in identical manner to amino acids using BOP activation as described above. For multimeric peptide the first residue to couple was Fmoc-Lysine (Fmoc)-OH. The Fmoc groups were removed using 20% Piperidine in DMF. Repeating this procedure again yielded the lysine core for extending four peptide chains in the usual manner.

Figure 1 shows an HPLC trace of the purifie polypeptide: TRNGQVLQGAIKG 25, ag polypeptide applied to a C-18 Vydac column running gradient of 0.1% TFA to 80% acetonitrile/0.1 % TFA in 30 mins. Detection wavelength was 218nm.

Techniques for synthesising polypeptides with different sequences and similar properties are well known. For instance new sequences may be discovered by the common method of constructing polypeptide libraries. Similarly existing sequences can be chemically modifie by removing, adding or replacing or substituting amino acids or analogues which are not required for activity. Sections of sequences may be combine from different polypeptides to make a new polypeptide. In one typical example of the method, the sequence TRNGQVLQGAIKG was reduced to 10 residues to give the sequence GQVLQGAIKG and this was further modifie by substituting each amino acid at a time, with another residue (Ala) generating several sequences (termed"Ala scan"polypeptides) as below: (1) AQVLQGAIKG (2) GAVLQGAIKG (3) GQAI » GAIKG (4) GQVAQGAIKG (5) GQVLAGAIKG (6) GQVLQAAIKG (7) GQVLQGAIKG (8) GQVLQGAAKG (9) GQVLQGAIAG (10) GQVLQGAIKA In Fig. 2, polypeptides (1)- (10) with Ala residue at positions 1-10 were screened for their binding to GARP (Goat anti-rabbit peroxidase) as described in Example 3. The absorbance reflets the binding of each analogue. The polypeptides produced by this substitution technique of Fig. 2 show that the Ala scan polypeptide sequences 1 to 8 each bound the protein to a different level.

Using amino acids other than Ala, thousands of analogues can be made. In this way information can be gained regarding the significance of each residues leading to discovery of new polypeptides.

Example 2. Screening of proteins binding to a polypeptide There are many techniques known for measuring the binding of proteins to molecules including polypeptides either in solution or by attaching to solid surfaces. We typically employed a 96 microwell ELISA plate which was coated, in replicates of eight, overnight with 100 Hl of 20 Hg/ml of polypeptide dissolve in buffer such as 50mM sodium carbonate pH 9.6. The coated plate then went through series of steps: (1) Wash--3 times) with Phosphate buffered saline pH 7.4 containing 0. 1% Tween (PBS-T), (2) Block the plate by incubating for 90 mins with 100, ul per well of PBS-T and wash three times with the same.

(3) Incubation for 60 mins with 100, ul of binding protein solution. The protein may be a labelle protein such as antibody or avidin or any other labelle with HRP or other enzymes or reporte groups. The concentration of protein solution will depend on the amount of label.

(4) Wash three times with PBS-T and monitor response by a technique depending on the label attache. For instance when Horse Radish Peroxides (HRP) is the label the wells could be incubated with a substrate such as 5-amino salicylic acid dissolve in 50mm Sodium phosphate pH 6 buffer containing 0.0 1% (vu) of fresh hydrogen peroxide. The response may then be measured colourimetrically after short incubation (e. g. 30 mins) by a reader or visually. Fig 5 shows binding of immunoglobulins and two fragments where the enzyme peroxidase is used as label.

In an alternative format, the label could be attache or bound secondary to the protein (step 2 above). For instance when rabbit IgG is used, the procedure following coating of polypeptide would then comprise: (1) Wash (3 times) with Phosphate buffered saline pH 7.4 containing 0.1% Tween (PBS-T) (2) Block the plate by incubating for 90 mins with 100, ul per well of PBS-T and wash three times with the same.

(3) Incubation for 60 mins with 100, ul of binding protein solution prepared in PBS-T.

(4) Wash three times with PBS-Tween and incubate with GARP (5) Wash three times with PBS-T and monitor response with a substrate such as 5-aminosalicylic acid dissolve in 50mM Sodium phosphate pH 6 buffer containing 0. 0 1 %- (VN) of fresh hydrogen peroxide.

There are numerus variations which are well known in thé art of solid phase assays and which can easily be made to these protocols.

For instance for indirect Peptide-Antibody Binding Assay the following procedure was adopte. Plates were coated by incubation for 1 hr at 37°C with 100 pl/well of IgG solution at 20, ug/ml in carbonate buffer, pH 9.6 (15 mM Na2C03; 35 mM NaHC03); plates were washed three times with tris-buffered saline (25 mM, pH 7.4) containing 0.1 % Tween-20 (TBS-T); 100 pl/well of TBS-T was used to block the uncoated well surface by incubation for 3 hr at 37°C; plates were washed as before; 100 pl of biotinylated peptide at 20tg/ml in TBS-T containing 0.2 % DMSO was placed into each well followed by incubation at 37°C for 1 hr; excess and unbound biotinylated peptide was washed; IgG-bound biotinylated-peptide in each well was detected by incubation with 1001 of ExtrAvidin-peroxidase conjugate in TBS-T (1: 1000 dilution as supplie and recommended by manufacturer) for 1 hr at 37°C; plates were washed as before to remove excess and unbound conjugate; bound conjugate was then detected by incubation with 5-amino salicylic acid and optical readings at 450 nm were determined as described previously.

Biotinylated GQVLQGAIKG showed significant binding to polyclonal IgG from various sources and some proteins relative to the appropriate controls (Table 1 and 2). In this assay avidin exhibited some non-specific binding to the proteins. Nevertheless, signal from peptide-protein interaction was apparent.

Accordingly biotinylated GQVLQGAIKG is an example of a polypeptide in accordance with the invention which binds more than one protein.

Example 3. Screening of polypeptides binding to a typical protein In this example Goat antirabbit peroxidase is used as typical protein.

Different polypeptides were coated in microwells as above and following steps performed.

(1) Wash (3 times) with Phosphate buffered saline pH 7.4 containing 0.1% Tween (PBS-T) (2) Block the plate by incubating for 90 mins with 100, ul per well of PBS-T and wash three times with the same.

(3) Incubation for 60 mins with 1001il of GARP (4) Wash three times with PBS-T and monitor response by incubating with a substrate such as 5-aminosalicylic acid dissolve in 50mM Sodium phosphate pH 6 buffer containing 0.01% (VN) of fresh hydrogen peroxide. The response may then be measured colourimetrically after short incubation (e. g 30 mins) by a reader or visually.

In one typical example the sequence GQVLQGAIKG was compare to 11 mer sequence derived from Protein G (Proc. Natl. Acad. Sci. USA 1992,89,8532-8536). Figure 6 shows that the peptide derived from natural protein is unable to bind IgG in the same manner as our sequence.

Fig. 10 shows the binding of GARP to a 30mer peptide of sequence KIGQFLIQFAGAFLSILQGLTLRAAEKQAG. In these cases an improved response is apparent due to the longer peptide than a shorter peptide as adsorption and or binding may be improved. A 10 mer peptide of sequence GQVLQGAIKG immobilised on BSA protein as in Example 9 also shows comparable binding to the 30 mer peptide (Fig. 10).

Using a similar screening method we coated the plates with Flagellin protein (sequence published J. Mol.. Biol 1991,219,471-480) and found that it was unable to bind to GARP.

Based on the possible binding motifs in our sequence we were able to synthesise the peptide SRAQILQQAG and show that this binds to GARP. It is thus possible to derive useful protein binding polypeptide sequences which show different binding properties to their natural full protein sequence (Fig 11).

Example 4: Screening of proteins binding to polypeptides directly on resin Short polypeptides do not usually adsorb efficiently to microwells. There are several ways such polypeptides can be screened for their binding ability. One of the ways is to immobilise the polypeptide covalently to solid surface. In this regard, ELISA plates with derivatiseSsurfaces are commercially available for linking molecules to surfaces. Such methods can easily be applied to bind shorter sequences and then screen in the usual manner described above. In an alternative way, the sequences can be synthesised on the solid phase by well known techniques and portions of the resin removed at various stages of the synthesis. The resin can then be used instead of the microwell as a support for the protein binding polypeptide of the invention. The washing steps analogous to the ELISA method reporte above can be carried out by mixing the resin with desired solution and separation effected by bench top microcentrifuge. In this way the inventors were able to scan the whole of TRNGQVLQGAIKG polypeptide of the invention. and found that even short sequences, for instance AIKG, are able to bind proteins. In a typical example the polypeptide (GQVLQGAIKG) was assemble on acid resistant resin but using Fmoc amino acids instead of the Boc used in example 1. A small amount of resin (lOmg) was removed after assembly of each amino acid and treated with 95% TFA/5% water mixture to cleave side chains. The resin was washed with dichloromethane and methanol and dried.

Next the resin was incubated in lml solution of PBS-T to block non specific sites. The ELISA steps as described in examples 2,3 and 4 could be performed on the resin using lml solution volumes followed by centrifugation to recover the resin after each wash step. At the last stage after adding the substrate and allowing rection to take place the response was measured by recording absorbance reading of the supernatant.

Using this method the binding of sequences as short as 4 residues, for instance AIKG linked to resin, to proteins (e. g. GARP) can be shown.

Example 5: Preparation of polypeptide affinity column There are several ways and chemistries available to prepare affinity columns with a wide range of matrices. (c. f Immobilised affinity ligand techniques (Academic Press 1992) or Bioaffinity Chromatography (Elsevier Science Publications 1993)). In a typical example the inventors used a commercially available preactivated Affi-prep column and immobilised the polypeptide of the invention in a high performance stainless steel column.

A 13mer polypeptide was synthesised with a 6 carbon spacer, aminohexanoic acid, and was immobilised in dilute buffer at pH7.8 at polypeptide concentration of 10mg/ml. The coupling was allowed to proceed overnight by recirculating the polypeptide solution through the column. The next day any remaining activated groups were treated with 0. 1M ethanolamlue solution.

In one example Proteins-HRP conjugates binding to Affiprep-10 immobilised peptide were measured by incubation of small amount of matrix (Preblocked to minimise non specific binding) in PBS-T buffer and washing off unbound material by centrifugation.

The bound conjugates were then estimated by incubation with 5-aminosalicylic acid as described for ELISA measurement in example 2 except that the rection mixture was centrifuged at low speed and supernatant used for recording the optical density. Affi-prep 10 matrix blocked with ethanolamine was used as a control matrix. Figure 7 shows that the immobilised sequence is able to bind proteins.

Immobilisation of multimeric GQVLQGAIKG peptide on amino sepharose matrix (AH-Sepharose 4B Pharmacia) was done as follows. lg of AH-Sepharose in 4ml of PBS was treated with 0. 5ml of 8% Glutaraldehyde solution. for 30 mins. Excess reagent was removed by washing the resin with distille water on a sintered funnel. A 2 fold molar excess of multimeric peptide in 50mM bicarobnate buffer pH 9.6 containing 10% DMSO was coupled to this activated matrix for 3hrs. Unbound peptide was washed by filtration with buffer followed by 10% acetic acid followed by water and ethanol. The matrix was resuspended in 25mol buffer and few crystals of sodium borohydride added. The washings used to remove unbound peptides were repeated. The coupling of peptide was qualitatively measured using a ninhydrin test. The matrix was packed into a short column (0.8cm X 10cm) and equilibrated with binding buffer.

Example 6: Separation andlor screening of proteins using polypeptide afrmity colunm The column produced in Example 6 was attache to the HPLC system and proteins were detected by using UV detector fixed at 280nm wavelength. Typically protein (0.2 to 0.5 to mg) was loaded onto the polypeptide column, and equilibrated with suitable buffer such as 10mM Tris-HCL pH 8, by rheodyne injector at a flow rate of 0.2m1/min. The effluent was continuously monitored at wavelength of 280nm and an elution profile obtained. The bound protein was then eluted by applying elution buffer such as 3M Guanidine hydrochloride or 0.1% TFA. The binding and elution profile can be seen in Fig. 3 which represents the binding and elution profile from affinity column when HRP enzyme (trace A) and goat anti-rabbit IgG antibody (trace B) were applied. The binding was achieved in 10mM tris-HCL pH7.4 buffer also used to equilibrate the column while the elution was made possible with the 0.1% TFA solution. In this example, the HRP is not appreciably bound and is thus eluted almost in the void. In contrast, the IgG is bound and eluted when 0.1 % TFA is applied.

The multimeric peptide column prepared in example 5 was equilibrated with 6ml of Phosphate buffered saline (PBS) at pH 7.4. A 0.2mg amount of Goat IgG (Sigma) in 0. 5ml PBS was applied to the column at a flow rate of 0.4m1/min. After loading the sample the binding buffer was applied to wash off unbound protein and absorbance measured continually at 280nm. We then applied 3ml solution of 3M Guanidine hydrochloride to elute off the bound protein. Figure 12 shows the chromatographic profile indicating binding and elution of typical protein from this column.

Example 7: Screening different compound for binding to polypeptides.

Using the methodologies of ELISA and affinity chromatography the inventors screened several antibodies from different sources and the data is tabulated below to indicate binding to typical polypeptide sequence. The proteins marked showed either no binding or insignificant levels of binding.

Table 1. Antibody Binding Determination by Goat IgG Yes ELISA & Column Human IgG Yes ELISA & Column Rat IgG Yes ELISA & Column Mouse IgG weak ELISA Human IgA Yes ELISA Donkey IgG Yes ELISA Guinea pig weak ELISA Dog Yes Column Sheep Yes Column Horse Yes Column Pig Yes Column Cow Yes Column The results depicted in Table 1 indicate that almost all sources of antibody showed binding and elution to varying degrees. Similarly other compound can be screened and some of the ones screened are tabulated below.

Table 2. Protein Binding Determination by Alkaline phosphate No ELISA & Column HRP No ELISA & Column BSA No Column Galactosidase No Column Human transferrin Yes ELEIS Strepatvidin/avidin Yes ELISA Protein A Yes ELISA Protein G Yes ELISA Concanavalin Yes ELISA Chymotrypsin Weak Column Example 8. Binding site on IgG.

In an expriment to demonstrate that the polypeptide of the invention does not bind directly at the normal binding site of the protein we used anti-FITC antibody. This antibody is known to quench >95% fluorescence of fluorescein upon specific binding at the antigen binding site. This quenching assay was performed in 2ml of l0mM Tris-HCL pH 7.4 buffer containing 3: 1 of 10, ug/ml fluorescein solution in the presence and absence of antibody. As shown in Fig. 4 5: 1 of antibody was able to quench most of the fluorescence upon binding.

The same level of binding was observe in the presence of 10: g of polypeptide (GQVLQGAIKG). The data indicates that the binding of the polypeptide does not influence the normal functioning of protein to appreciable extent. However it can not be ruled out whether the polypeptide of the invention binds close to the functional site or whether the non immobilised polypeptide behaves differently than that free in solution.

Example 9 Measurement of binding by optical biosensor The peptide sequence GQVLQGAIKG was immobilised on BSA using the glutaraldehyde method as follows. Bovine Serum albumin (4mg) was dissolve in 0.75mol of 10mM Sodium Phosphate pH 7.4 buffer. Glutaraldehyde (0. 25mol of 8% solution) was added and mixture stirred for 30mins at room temperature. The Excess Glutaradehyde was removed by Gel filtration on PD-10 column. The peptide (lOmg dissolve in minimum volume of DMSO) was added to the activated BSA and conjugation allowed to proceed for 3 hrs.

Unconjugated Peptide was removed by dialysis, centrifugation and further Gel filtration.

The conjugate was immobilised on CM5 chip, using EDC coupling, by flowing across the sensor chip according to the manufacturers description (BIAcore). BSA was used in the control flow cell. The peptide protein interaction was studies using different concentrations of proteins in order to obtain optimum conditions for measuring binding constants. In a typical example the antibody was bound in Phosphate buffered saline and regeneration effected with 3M Guanidine hydrochloride solution. The binding affinities (kD) were determined. Figure 8 and 9 shows the association and dissociation progress curves. The multimeric peptide could be immobilised in identical manner and binding evaluated. The kD values obtained were estimated to be 4 x 10-7M and 1 x 10-7 M respectively for the multimeric and BSA conjugated 10 mer peptide.