NZ623518B2 - Modulation of structured polypeptide specificity - Google Patents

Abstract

Discloses a peptide ligand specific for human Kallikrein comprising a polypeptide comprising at least three reactive groups, separated by at least two loop sequences, and a molecular scaffold which forms covalent bonds with the reactive groups of the polypeptide such that at least two polypeptide loops are formed on the molecular scaffold, wherein the loops of the peptide ligand comprise three, four or five, but less than six, amino acids. ops are formed on the molecular scaffold, wherein the loops of the peptide ligand comprise three, four or five, but less than six, amino acids.

Description

Modulation of ured polypeptide specificity The present invention generally relates to polypeptides which are covalently bound to lar scaffolds such that two or more peptide loops are subtended between attachment points to the scaffold. In particular, the invention describes peptides which are specific for the human protease plasma rein.

Cyclic peptides are able to bind with high affinity and target specificity to protein targets and hence are an attractive molecule class for the development of therapeutics. In fact, several cyclic es are already successfully used in the clinic, as for example the cterial peptide vancomycin, the immunosuppressant drug cyclosporine or the anticancer drug ocreotide (Driggers, et al., Nat Rev Drug Discov 2008, 7 (7), 608-24). Good binding properties result from a relatively large interaction surface formed between the peptide and the target as well as the reduced conformational flexibility of the cyclic structures. Typically, macrocycles bind to surfaces of several hundred square angstrom, as for example the cyclic peptide CXCR4 antagonist CVX15 (400 Å2; Wu, B., et al., Science 330 (6007), 1066-71), a cyclic peptide with the Arg-Gly-Asp motif binding to integrin aVb3 (355 Å2) , J. P., et al., Science 2002, 296 (5565), 151-5) or the cyclic peptide inhibitor upain-1 binding to urokinase-type plasminogen activator (603 Å2; Zhao, G., et al., J Struct Biol 2007, 160 (1), 1-10).

Due to their cyclic configuration, peptide ycles are less flexible than linear peptides, leading to a smaller loss of entropy upon binding to targets and ing in a higher g affinity. The reduced flexibility also leads to locking target-specific conformations, increasing binding icity compared to linear peptides. This effect has been exemplified by a potent and selective inhibitor of matrix metalloproteinase 8, MMP- 8) which lost its selectivity over other MMPs when its ring was opened (Cherney, R. J., et al., J Med Chem 1998, 41 (11), 1749-51). The favorable g properties achieved through macrocyclization are even more pronounced in multicyclic peptides having more than one peptide ring as for example in vancomycin, nisin or actinomycin.

Different research teams have previously tethered polypeptides with cysteine residues to a synthetic lar structure (Kemp, D. S. and McNamara, P. E., J. Org. Chem, 1985; Timmerman, P. et al., ChemBioChem, 2005). Meloen and co-workers had used tris(bromomethyl)benzene and related molecules for rapid and quantitative cyclisation of multiple e loops onto synthetic scaffolds for structural y of protein surfaces rman, P. et al., ChemBioChem, 2005). Methods for the generation of candidate drug compounds wherein said compounds are generated by linking cysteine containing polypeptides to a molecular scaffold as for example tris(bromomethyl)benzene are disclosed in and . /077062 discloses a method of selecting a candidate drug compound. In particular, this nt discloses various scaffold molecules comprising first and second reactive groups, and contacting said scaffold with a further le to form at least two linkages between the scaffold and the further molecule in a coupling reaction.

WO2006/078161 discloses binding compounds, immunogenic compounds and peptidomimetics. This document discloses the artificial synthesis of various collections of es taken from existing proteins. These peptides are then combined with a constant tic peptide having some amino acid changes introduced in order to produce combinatorial libraries. By ucing this diversity via the chemical linkage to separate peptides featuring various amino acid changes, an increased opportunity to find the desired binding activity is provided. Figure 1 of this document shows a schematic entation of the synthesis of various loop peptide constructs. The ucts sed in this document rely on –SH functionalised peptides, typically comprising cysteine residues, and heteroaromatic groups on the scaffold, typically comprising benzylic halogen tuents such as bis- or tris-bromophenylbenzene. Such groups react to form a thioether linkage between the peptide and the ld.

We recently developed a phage display-based combinatorial approach to generate and screen large libraries of bicyclic peptides to targets of interest (Heinis, et al., Nat Chem Biol 2009, 5 (7), 502-7; see also international patent application WO2009/098450). y, combinatorial libraries of linear peptides containing three cysteine residues and two regions of six random amino acids (Cys-(Xaa)6-Cys-(Xaa)6-Cys) were displayed on phage and cyclised by covalently linking the cysteine side chains to a small molecule (tris-(bromomethyl)benzene). Bicyclic peptides isolated in affinity selections to the human proteases cathepsin G and plasma Kallikrein (PK) had nanomolar inhibitory constants.

The best inhibitor, PK15, inhibits human PK (hPK) with a Ki of 3 nM. Similarities in the amino acid sequences of several ed bicyclic peptides suggested that both peptide loops contribute to the binding. PK15 did not inhibit rat PK (81% sequence identity) nor the homologous human serine proteases factor XIa (hfXIa; 69% sequence identity) or thrombin (36% sequence identity) at the highest tration tested (10 mM) (Heinis, et al., Nat Chem Biol 2009, 5 (7), 502-7). This g suggested that the bicyclic inhibitor is highly specific and that other human trypsin-like serine proteases will not be inhibited. A synthetic, small ic inhibitor such as PK15 having the above described potency and target ivity has ial application as a therapeutic to control PK activity in hereditary angioedema, a life-threatening disease which is characterized by recurrent episodes of edema or to t contact activation in cardiopulmonary bypass surgery.

The peptide PK15 was isolated from a y based on the peptide PK2, HACSDRFRNCPLWSGTCG-NH2 , in which the second 6-amino acid loop was randomised.

The sequence of PK15 was H-ACSDRFRNCPADEALCG-NH2, and the IC50 binding constant for human Kallikrein was 1.7 nM. y of the Invention We have analysed the specificity of structured polypeptides selected against human Kallikrein from a number of libraries with different loop lengths. As a , we have succeeded in isolating structured peptides capable of binding Kallikrein with improved binding specificities; and/or which at least provide the public with a useful choice.

In a first aspect, there is ed a peptide ligand specific for human Kallikrein comprising a polypeptide comprising at least three reactive groups, separated by at least two loop sequences, and a molecular scaffold which forms covalent bonds with the reactive groups of the polypeptide such that at least two polypeptide loops are formed on the molecular scaffold, wherein the loops of the peptide ligand comprise three, four or five, but less than six, amino acids.

Surprisingly, we have found that peptides comprising less than 6 amino acids in each loop can have a much higher binding affinity for Kallikrein. For example, the 5x5 peptides described herein achieve binding constants of 0.08nM or less.

In one embodiment, the loops of the peptide ligand comprise three amino acids and the polypeptide has the consensus sequence rRVxGr, wherein Gr is a reactive group.

For example, the polypeptide may be one of the polypeptides set forth in Table 3.

In another embodiment, the loops of the peptide ligand comprise five amino acids and a first loop comprises the consensus sequence GrGGxxNGr, wherein Gr is a ve group.

For example, two adjacent loops of the polypeptide may comprise the consensus sequence GrGGxxNGrRxxxxGr.

For example, the polypeptide may be one of the peptides set forth in Table 4.

In one embodiment, the loops of the peptide ligand se five amino acids and a first loop comprises the motif GrxW/FPxK/RGr, wherein Gr is a reactive group. In the present context, the reference to a “first” loop does not necessarily denote a particular position of the loop in a sequence. In some embodiments, however, the first loop may be proximal loop in an amino us to carboxy terminus peptide ce. For example, the polypeptide further comprises a second, distal loop which comprises the motif GrT/LHQ/TxLGr. Examples of sequences of the first loop include G Gr, GrxWPSRGr, GrxFPFRGr and RGr. In these examples, x may be any amino acid, but is for example S or R.

In one embodiment, the loops of the peptide ligand comprise five amino acids and a first loop comprises the motif GrxHxDLGr, n Gr is a reactive group.

In one embodiment, the loops of the peptide ligand se five amino acids and a first loop comprises the motif GrTHxxLGr, wherein Gr is a reactive group.

In one ment, the polypeptide comprises two adjacent loops which comprise the motif GrxW/FPxK/RGrT/LHQ/TDLGr.

We have shown that the nature of certain positions can influence other positions in the sequence. In particular, experiments conducted with peptides 06-34 and 0603 demonstrate that positions 1 and 6 influence position 4. Preferably, position 4 is A only if positions 1 and 6 are S and T respectively.

In the examples herein, numbering refers to the positions in the loops, and ignores the reactive groups. Thus, in GrxW/FPxK/RGrT/LHQ/TDLGr, x is in position 1 and T/ L in position 6.

For example, the polypeptide may be one of the polypeptides set forth in Table 4, Table 5 or Table 6.

For example, the ptide ligand may comprise one of the polypeptides set forth in one of Tables 4 to 6.

In the ing embodiments, the reactive group is preferably a reactive amino acid.

Preferably, the reactive amino acid is cysteine.

Variants of the polypeptides according to this aspect of the invention can be prepared as bed above, by fying those residues which are available for mutation and preparing libraries which include mutations at those positions. For example, the polypeptide 06-56 in Table 4 can be mutated without loss of activity at positions Q4 and T10 (see Examples below). Polypeptide ligands comprising mutations at these positions can be selected which have improved binding activity in comparison with 06-56.

In a r aspect, there is provided a polypeptide ligand according to the ing aspect of the invention, which comprises one or more non-natural amino acid substituents and is resistant to protease degradation.

We have found that certain non-natural amino acids permit binding to plasma Kallikrein with nM Ki, whilst increasing residence time in plasma icantly.

In one embodiment, the tural amino acid is selected from N-methyl Arginine, homo-arginine and hydroxyproline. Preferably, N-methyl and homo-derivatives of Arginine are used to replace ne, and proline 3 can be preferably replaced by hydroxyproline, azetidine carboxylic acid, or an alpha-substituted amino acid, such as aminoisobutyric acid. In another embodiment, ne may be replaced with guanidylphenylalanine.

In one embodiment, the polypeptide comprises a first loop which comprises the motif GrxWPARGr, wherein P is replaced with azetidine carboxylic acid; and/or R is replaced with N-methyl arginine; and/or R is replaced with homoarginine; and/or R is replaced with guanidyl-phenylalanine.

In one embodiment, the polypeptide comprises a first loop which comprises the motif GrxFPYRGr, wherein R is ed with yl arginine; and/or R is replaced with homoarginine, and wherein proline is replaced by azetidine carboxylic acid; and/or R is replaced with guanidyl-phenylalanine.

Certain statements that appear below are broader than what appears in the statements of the invention above. These statements are ed in the sts of providing the reader with a better understanding of the invention and its practice. The reader is directed to the accompanying claim set which defines the scope of the invention.

Also described is a method for producing a mutant polypeptide ligand to produce an improved level of binding activity for a target over that of a parent polypeptide ligand, wherein the parent polypeptide ligand comprises a polypeptide comprising at least three ve , separated by at least two loop sequences, and a molecular scaffold which forms covalent bonds with the reactive groups of the polypeptide such that at least two polypeptide loops are formed on the molecular scaffold, comprising the steps of: (a) for each of two or more amino acid positions in each of the loop sequences, producing n different ies of mutants, each library consisting of parent polypeptides in which one of said amino acid positions in the loop sequence has been mutated by replacement with one of n different non-parental amino acids; (b) ing each library for binding to the parental target, and scoring each mutation; (c) identifying the amino acid positions at which mutations are tolerated; (d) producing one or more mutant polypeptides comprising one or more ons located at the amino acid positions identified in step (c).

In one embodiment, step (d) comprises preparing a library comprising polypeptides which incorporate mutations at two or more of the amino acid positions identified in step (c), and screening the y for polypeptides with an improved level of binding activity for the target.

The value of n can be selected according to the number of ent mutants it is ed to create in each library. For example, if mutants comprising all possible natural amino acids are desired, n can be 20. If non-natural amino acids are included, such as N- methylated amino acids, n can be greater than 20, such as 22 or 23. For example, n can be 2 or more; 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20.

Also described is a library of ptide s, wherein the ptide ligands comprise a polypeptide comprising at least three reactive groups, separated by at least two loop sequences, and a molecular scaffold which forms covalent bonds with the reactive groups of the polypeptide such that at least two polypeptide loops are formed on the lar scaffold, said library consisting of m different mutants of a polypeptide ligand in which a defined amino acid position in the loop sequences has been mutated by replacement with one of m different amino acids, wherein m is at least 2.

Also described is a set of libraries of ptide ligands, wherein the polypeptide ligands se a polypeptide comprising at least three reactive groups, separated by at least two loop sequences, and a molecular scaffold which forms covalent bonds with the reactive groups of the polypeptide such that at least two polypeptide loops are formed on the molecular scaffold, which set comprises two or more ies of polypeptide ligands, each of said libraries of polypeptide ligands consisting of m different mutants of a polypeptide ligand in which a defined amino acid position in the loop sequences has been mutated by replacement with one of m different amino acids.

Preferably, m is between 2 and 20; in embodiments, m is at least 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20, or more, as set out in respect of n above.

Also described is a peptide ligand comprising a polypeptide comprising at least three reactive groups, separated by at least two loop sequences, and a molecular scaffold which forms covalent bonds with the reactive groups of the polypeptide such that at least two polypeptide loops are formed on the lar scaffold, wherein the peptide is modified by the incorporation of at least one non-natural amino acid.

Preferably, the peptide ligand as described is protease-resistant. The non-natural amino acid substitution(s) increase the level of protease resistance of the polypeptide.

In one embodiment, the non-natural amino acid is selected from N-methyl Arginine, homo-arginine and ine carboxylic acid and guanidylphenylalanine. Preferably, N- methyl and homo-derivatives of Arginine are used to replace Arginine, and azetidine carboxylic acid replaces proline. In another embodiment, Arginine may be replaced with guanidyl-phenylalanine.

Brief description of the figures Figure 1 Phage selection of ic peptides. (a) Bicyclic peptide phage libraries.

Random amino acids are indicated as 'X', alanine as 'A' and the constant three cysteine residues as 'C'. (b) Format of chemically synthesized bicyclic peptide structures having loops of 3, 5 or 6 amino acids. The structures are generated by linking linear peptides via three cysteine side chains to tris-(bromomethly)benzene . Amino acids that vary in the ic es are indicated with 'Xaa'. (c-e) Sequences of bicyclic peptides isolated from library 5x5 (c), library 3x3 A (d) and library 3x3 B (e). Similarities in amino acids are highlighted by shading.

Figure 2 Comparison of the surface amino acids of hPK and homologous serine proteases. (a) Structure of hPK (PDB entry 2ANW) with surface representation. Atoms of amino acids being exposed to the surface and closer than 4, 8 and 12 Å to idine (in grey) bound to the S1 pocket are stained more darkly. (b) Structure of hPK. The side chains of amino acids that are different in hfXIa are highlighted. (c) Structure of hPK. The side chains of amino acids that are different in rPK are highlighted.

Figure 3 Pictorial representation of the method used for determination of red residues for on in ptide ligands.

Figure 4 Analysis of amino acid substitutions in peptide 06-34 (Table 4) on the binding of the e to plasma Kallikrein at 2nM. For each position, the effect of various mutations at that on is shown, in comparison to the parent sequence.

Figure 5 Analysis of amino acid substitutions in e 06-56 (Table 4) on the binding of the peptide to plasma Kallikrein at 2nM. For each position, the effect of various mutations at that position is shown, in comparison to the parent sequence.

Figure 6 Analysis of amino acid tutions in peptide 06-56 (Table 4) on the binding of the peptide to plasma Kallikrein at 10nM. For each position, the effect of various mutations at that position is shown, in comparison to the parent sequence.

Figure 7 Mass spec output g the mass spectra of Ac34-18(TMB)-NH2 after exposure to 35% rat plasma, at t0, 1 day, 2 days and 3 days (method 1). Mass accuracies vary somewhat due to interfering ions and low concentrations of fragments; however identification of discrete proteolytic fragments is possible.

Figure 8 Chemical structures of metabolites M1, M2, M3 of Ac34-18(TMB)-NH2 identified after exposure to rat plasma.

Figure 9 Chemical structure of the Ac34-18(TMB)-NH2 lead Figure 10 Enzyme inhibition assay of Kallikrein by the 34-18(TMB)-NH2 lead and its 1st loop scrambled derivatives. A dramatic reduction in ty is observed, underlining the importance of the integrity of the WPAR pharmacophore.

Figure 11 al structures of arginine and its analogues.

Figure 12 Chemical structures of Trp and potential hydrophobic analogues Figure 13 Chemical structures of Pro and potential constrained analogues Figure 14 ative Kallikrein inhibition by Aze3, NMeArg5 and doubly modified Ac- 0618(TMB)-NH2.

Figure 15 Chemical structures of Alanine and derivatives thereof Figure 16 Comparative Kallikrein tion by F2Y4, F2Y2 HR5 and doubly modified Ac- 0618(TMB)-NH2.

Figure 17 Human plasma Kallikrein binding of particular motifs at positions 2, 3, 4 & 5 (with positions 1, 6, 7, 8, 9 &10 fixed to those of 0603).

Detailed Description of the Invention Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the art, such as in the arts of peptide chemistry, cell culture and phage display, nucleic acid chemistry and biochemistry. Standard techniques are used for molecular biology, genetic and biochemical methods (see Sambrook et al., Molecular Cloning: A Laboratory Manual, 3rd ed., 2001, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY; Ausubel et al., Short Protocols in Molecular Biology (1999) 4th ed., John Wiley & Sons, Inc.), which are incorporated herein by reference.

A peptide ligand, as referred to , refers to a peptide covalently bound to a molecular scaffold. Typically, such peptides comprise two or more ve groups which are capable of g covalent bonds to the scaffold, and a sequence subtended between said reactive groups which is referred to as the loop sequence, since it forms a loop when the peptide is bound to the scaffold. In the present case, the peptides comprise at least three reactive groups, and form at least two loops on the scaffold.

The reactive groups are groups capable of forming a covalent bond with the molecular scaffold. lly, the reactive groups are present on amino acid side chains on the e. Examples are amino-containing groups such as cysteine, lysine and selenocysteine.

Specificity, in the context herein, refers to the ability of a ligand to bind or otherwise interact with its cognate target to the exclusion of entities which are similar to the target.

For example, specificity can refer to the ability of a ligand to inhibit the interaction of a human enzyme, but not a homologous enzyme from a different species. Using the approach described , specificity can be ted, that is sed or sed, so as to make the ligands more or less able to ct with homologues or gues of the ed target. Specificity is not intended to be synonymous with ty, affinity or avidity, and the potency of the action of a ligand on its target (such as, for example, binding affinity or level of inhibition) are not necessarily related to its specificity.

Binding activity, as used herein, refers to quantitative binding measurements taken from g assays, for e as described herein. Therefore, binding ty refers to the amount of peptide ligand which is bound at a given target concentration.

Multispecificity is the ability to bind to two or more targets. Typically, g peptides are capable of binding to a single target, such as an epitope in the case of an antibody, due to their conformational properties. However, peptides can be developed which can bind to two or more targets; dual specific antibodies, for example, as known in the art as referred to above. In the present invention, the peptide ligands can be capable of binding to two or more targets and are therefore be multispecific. Preferably, they bind to two targets, and are dual specific. The binding may be independent, which would mean that the binding sites for the targets on the peptide are not structurally hindered by the binding of one or other of the targets. In this case both targets can be bound independently.

More generally it is expected that the binding of one target will at least partially impede the binding of the other.

There is a fundamental difference between a dual specific ligand and a ligand with specificity which encompasses two related targets. In the first case, the ligand is specific for both targets individually, and interacts with each in a specific manner. For example, a first loop in the ligand may bind to a first target, and a second loop to a second . In the second case, the ligand is ecific because it does not differentiate between the two targets, for example by interacting with an epitope of the targets which is common to both.

In the context of the present invention, it is possible that a ligand which has ty in respect of, for example, a target and an orthologue, could be a ific ligand.

However, in one embodiment the ligand is not bispecific, but has a less precise specificity such that it binds both the target and one or more orthologues. In general, a ligand which has not been selected t both a target and its orthologue is less likely to be bispecific as a result of modulation of loop .

If the ligands are truly ific, in one embodiment at least one of the target specificities of the ligands will be common amongst the ligands selected, and the level of that specificity can be modulated by the methods disclosed herein. Second or further specificities need not be shared, and need not be the t of the procedures set forth herein.

A target is a molecule or part thereof to which the peptide ligands bind or otherwise interact with. Although binding is seen as a prerequisite to activity of most kinds, and may be an activity in itself, other activities are envisaged. Thus, the present ion does not require the measurement of binding directly or indirectly.

The molecular scaffold is any molecule which is able to connect the peptide at le points to impart one or more structural features to the e. It is not a cross-linker, in that it does not merely replace a disulphide bond; instead, it provides two or more attachment points for the peptide. Preferably, the molecular scaffold comprises at least three ment points for the peptide, referred to as scaffold reactive groups. These groups are capable of reacting to the ve groups on the peptide to form a covalent bond. Preferred structures for molecular scaffolds are described below.

Screening for g activity (or any other desired activity) is conducted according to methods well known in the art, for instance from phage display technology. For example, targets immobilised to a solid phase can be used to identify and isolate binding members of a repertoire. Screening allows selection of members of a repertoire according to d characteristics.

The term library refers to a e of heterogeneous polypeptides or nucleic acids. The library is composed of members, which are not identical. To this extent, library is synonymous with oire. ce differences between library members are responsible for the diversity present in the y. The library may take the form of a simple mixture of polypeptides or nucleic acids, or may be in the form of organisms or cells, for example ia, viruses, animal or plant cells and the like, transformed with a y of nucleic acids. Preferably, each individual organism or cell contains only one or a limited number of y members.

In one embodiment, the nucleic acids are incorporated into expression vectors, in order to allow expression of the polypeptides encoded by the nucleic acids. In a preferred embodiment, therefore, a y may take the form of a population of host organisms, each organism containing one or more copies of an expression vector containing a single member of the library in nucleic acid form which can be expressed to produce its corresponding polypeptide member. Thus, the population of host organisms has the potential to encode a large repertoire of genetically diverse polypeptide variants.

In one embodiment, a library of nucleic acids encodes a repertoire of polypeptides. Each nucleic acid member of the library preferably has a ce related to one or more other members of the y. By related sequence is meant an amino acid sequence having at least 50% identity, for example at least 60% identity, for example at least 70% identity, for example at least 80% identity, for example at least 90% identity, for example at least 95% identity, for example at least 98% ty, for example at least 99% identity to at least one other member of the y. Identity can be judged across a contiguous segment of at least 3 amino acids, for example at least 4, 5, 6, 7, 8, 9 or 10 amino acids, for e least 12 amino acids, for example least 14 amino acids, for example least 16 amino acids, for example least 17 amino acids or the full length of the reference sequence.

A repertoire is a collection of ts, in this case polypeptide variants, which differ in their sequence. Typically, the location and nature of the reactive groups will not vary, but the sequences forming the loops between them can be randomised. Repertoires differ in size, but should be considered to comprise at least 102 members. Repertoires of 1011 or more members can be constructed.

A set of polypeptide ligands, as used herein, refers to a plurality of polypeptide ligands which can be subjected to selection in the methods described. Potentially, a set can be a repertoire, but it may also be a small collection of polypeptides, from at least 2 up to 10, , 50, 100 or more.

A group of polypeptide ligands, as used herein, refers to two or more ligands. In one embodiment, a group of ligands comprises only ligands which share at least one target specificity. lly, a group will consist of from at least 2, 3, 4, 5, 6, 7, 8, 9 or 10, 20, 50, 100 or more ligands. In one embodiment, a group consists of 2 ligands.

(A) Construction of Peptide Ligands (i) lar scaffold Molecular scaffolds are described in, for example, WO2009098450 and nces cited therein, particularly WO2004077062 and WO2006078161.

As noted in the foregoing documents, the molecular scaffold may be a small molecule, such as a small organic molecule.

In one ment the molecular scaffold may be, or may be based on, natural monomers such as nucleosides, sugars, or steroids. For example the molecular scaffold may comprise a short polymer of such es, such as a dimer or a trimer.

In one embodiment the molecular scaffold is a compound of known toxicity, for e of low toxicity. Examples of suitable compounds e cholesterols, nucleotides, steroids, or existing drugs such as tamazepam.

In one embodiment the molecular scaffold may be a macromolecule. In one embodiment the molecular scaffold is a macromolecule composed of amino acids, nucleotides or carbohydrates.

In one embodiment the molecular scaffold comprises reactive groups that are capable of ng with functional group(s) of the polypeptide to form covalent bonds.

The molecular scaffold may comprise chemical groups as amines, thiols, alcohols, ketones, aldehydes, nitriles, carboxylic acids, esters, s, alkynes, azides, anhydrides, succinimides, maleimides, alkyl halides and acyl halides.

In one embodiment, the molecular scaffold may comprise or may consist of tris(bromomethyl)benzene, especially 1,3,5-Tris(bromomethyl)benzene ’), or a derivative thereof.

In one embodiment, the molecular scaffold is 2,4,6-Tris(bromomethyl)mesitylene. It is similar to 1,3,5-Tris(bromomethyl)benzene but contains additionally three methyl groups ed to the e ring. This has the advantage that the additional methyl groups may form further ts with the polypeptide and hence add additional structural constraint.

The molecular scaffold useful in the invention contains chemical groups that allow onal groups of the polypeptide of the encoded y described herein to form covalent links with the molecular scaffold. Said chemical groups are selected from a wide range of functionalities including amines, thiols, alcohols, s, aldehydes, nitriles, carboxylic acids, esters, alkenes, alkynes, anhydrides, succinimides, maleimides, azides, alkyl halides and acyl s. (ii) Polypeptide The reactive groups of the polypeptides can be provided by side chains of natural or nonnatural amino acids. The reactive groups of the ptides can be selected from thiol groups, amino groups, carboxyl groups, guanidinium groups, phenolic groups or hydroxyl groups. The reactive groups of the polypeptides can be selected from azide, keto- carbonyl, , vinyl, or aryl halide groups. The reactive groups of the polypeptides for linking to a molecular scaffold can be the amino or carboxy termini of the polypeptide.

In some embodiments each of the reactive groups of the polypeptide for linking to a molecular scaffold are of the same type. For example, each reactive group may be a cysteine e. r s are provided in WO2009098450.

In some embodiments the reactive groups for g to a molecular scaffold may comprise two or more different types, or may comprise three or more different types. For example, the reactive groups may comprise two cysteine residues and one lysine residue, or may comprise one cysteine residue, one lysine residue and one N-terminal amine.

Cysteine can be employed because it has the advantage that its reactivity is most different from all other amino acids. Scaffold reactive groups that could be used on the molecular scaffold to react with thiol groups of cysteines are alkyl halides (or also named halogenoalkanes or haloalkanes). Examples are bromomethylbenzene (the ld reactive group exemplified by TBMB) or iodoacetamide. Other scaffold reactive goups that are used to couple selectively compounds to nes in ns are maleimides.

Examples of maleimides which may be used as molecular scaffolds include: tris-(2- maleimidoethyl)amine, tris-(2-maleimidoethyl)benzene, tris-(maleimido)benzene.

Selenocysteine is also a natural amino acid which has a similar reactivity to cysteine and can be used for the same reactions. Thus, wherever cysteine is mentioned, it is typically acceptable to substitute cysteine unless the context suggests otherwise.

Lysines (and primary amines of the N-terminus of peptides) are also suited as reactive groups to modify peptides on phage by linking to a molecular scaffold. However, they are more abundant in phage ns than cysteines and there is a higher risk that phage particles might become cross-linked or that they might lose their ivity. Nevertheless, it has been found that s are especially useful in intramolecular reactions (e.g. when a molecular scaffold is already linked to the phage peptide) to form a second or consecutive linkage with the molecular scaffold. In this case the lar scaffold reacts preferentially with lysines of the displayed peptide (in particular lysines that are in close proximity). ld reactive groups that react selectively with primary amines are succinimides, des or alkyl halides. In the bromomethyl group that is used in a number of the accompanying examples, the electrons of the benzene ring can stabilize the cationic transition state. This particular aryl halide is therefore 100-1000 times more reactive than alkyl halides. Examples of succinimides for use as molecular scaffold include tris-(succinimidyl riacetate), 1,3,5-Benzenetriacetic acid. Examples of aldehydes for use as molecular scaffold include Triformylmethane. Examples of alkyl halides for use as molecular scaffold include 1,3,5-Tris(bromomethyl)-2,4,6- trimethylbenzene, 1,3,5-Tris(bromomethyl) e, 1,3,5-Tris(bromomethyl)-2,4,6- triethylbenzene.

The amino acids with reactive groups for linking to a molecular scaffold may be d at any suitable positions within the polypeptide. In order to influence the particular structures or loops created, the positions of the amino acids having the reactive groups may be varied by the skilled operator, e.g. by manipulation of the nucleic acid encoding the polypeptide in order to mutate the polypeptide ed. By such means, loop length can be manipulated in accordance with the t teaching.

For example, the polypeptide can comprise the sequence AC(X)nC(X)mCG, wherein X stands for a random natural amino acid, A for alanine, C for cysteine and G for glycine and n and m, which may be the same or different, are s between 3 and 6. (iii) Reactive groups of the polypeptide The molecular scaffold useful in the invention may be bonded to the polypeptide via functional or reactive groups on the polypeptide. These are typically formed from the side chains of particular amino acids found in the polypeptide r. Such ve groups may be a cysteine side chain, a lysine side chain, or an N-terminal amine group or any other suitable reactive group. Again, details may be found in WO2009098450.

Examples of reactive groups of natural amino acids are the thiol group of cysteine, the amino group of lysine, the carboxyl group of aspartate or glutamate, the guanidinium group of arginine, the phenolic group of ne or the hydroxyl group of . Nonnatural amino acids can provide a wide range of reactive groups including an azide, a keto-carbonyl, an alkyne, a vinyl, or an aryl halide group. The amino and carboxyl group of the termini of the polypeptide can also serve as reactive groups to form covalent bonds to a molecular scaffold/molecular core.

The ptides of the invention contain at least three reactive groups. Said polypeptides can also contain four or more reactive groups. The more reactive groups are used, the more loops can be formed in the molecular scaffold.

In a preferred embodiment, polypeptides with three reactive groups are generated.

Reaction of said polypeptides with a molecular scaffold/molecular core having a three- fold rotational symmetry generates a single product isomer. The tion of a single product isomer is favourable for several s. The nucleic acids of the compound ies encode only the primary sequences of the polypeptide but not the isomeric state of the molecules that are formed upon reaction of the polypeptide with the lar core. If only one product isomer can be formed, the assignment of the nucleic acid to the product isomer is clearly defined. If multiple t s are formed, the nucleic acid can not give information about the nature of the product isomer that was isolated in a screening or selection process. The formation of a single product isomer is also advantageous if a specific member of a library is synthesized. In this case, the chemical reaction of the polypeptide with the molecular scaffold yields a single t isomer rather than a mixture of isomers.

In another embodiment of the invention, polypeptides with four reactive groups are generated. Reaction of said polypeptides with a molecular scaffold/molecular core having a tetrahedral symmetry generates two product isomers. Even though the two different product isomers are encoded by one and the same nucleic acid, the isomeric nature of the isolated isomer can be determined by chemically synthesizing both s, separating the two isomers and testing both isomers for binding to a target ligand.

In one embodiment of the invention, at least one of the reactive groups of the polypeptides is orthogonal to the remaining reactive groups. The use of orthogonal reactive groups allows the directing of said orthogonal reactive groups to specific sites of the molecular core. Linking strategies ing orthogonal reactive groups may be used to limit the number of t isomers formed. In other words, by choosing distinct or different reactive groups for one or more of the at least three bonds to those chosen for the remainder of the at least three bonds, a particular order of bonding or directing of specific reactive groups of the polypeptide to specific positions on the molecular scaffold may be usefully achieved.

In r embodiment, the ve groups of the polypeptide of the invention are reacted with molecular linkers wherein said linkers are capable to react with a lar ld so that the linker will intervene between the molecular scaffold and the polypeptide in the final bonded state.

In some embodiments, amino acids of the members of the libraries or sets of polypeptides can be replaced by any natural or non-natural amino acid. Excluded from these exchangeable amino acids are the ones harbouring functional groups for crosslinking the polypeptides to a molecular core, such that the loop sequences alone are exchangeable. The exchangeable polypeptide sequences have either random sequences, constant sequences or sequences with random and constant amino acids.

The amino acids with reactive groups are either located in defined positions within the polypeptide, since the on of these amino acids determines loop size.

In one embodiment, an polypeptide with three reactive groups has the sequence (X)lY(X)mY(X)nY(X)o, wherein Y represents an amino acid with a reactive group, X represents a random amino acid, m and n are numbers between 3 and 6 defining the length of intervening polypeptide segments, which may be the same or different, and l and o are numbers between 0 and 20 defining the length of flanking polypeptide segments.

Alternatives to thiol-mediated conjugations can be used to attach the lar scaffold to the peptide via covalent interactions. Alternatively these techniques may be used in modification or attachment of further moieties (such as small molecules of interest which are distinct from the molecular scaffold) to the polypeptide after they have been selected or isolated as bed herein – in this embodiment then clearly the attachment need not be covalent and may embrace non-covalent ment. These methods may be used instead of (or in combination with) the thiol mediated methods by producing phage that y proteins and peptides bearing unnatural amino acids with the requisite chemical ve groups, in combination small molecules that bear the complementary reactive group, or by incorporating the unnatural amino acids into a chemically or recombinantly synthesised polypeptide when the molecule is being made after the selection/isolation phase. Further details can be found in WO2009098450 or , et al., Nat Chem Biol 2009, 5 (7), 502-7. (iv) Combination of loops to form multispecific molecules Loops from peptide ligands, or repertoires of peptide ligands, are advantageously combined by cing and de novo synthesis of a polypeptide incorporating the combined loops. atively, nucleic acids encoding such polypeptides can be synthesised.

Where repertoires are to be combined, particularly single loop repertoires, the nucleic acids encoding the oires are advantageously digested and re-ligated, to form a novel repertoire having ent combinations of loops from the constituent repertoires.

Phage vectors can include nkers and other sites for ction enzymes which can provide unique points for cutting and relegation the vectors, to create the desired multispecific peptide ligands. Methods for manipulating phage libraries are well known in respect of dies, and can be applied in the present case also. (v) Attachment of Effector Groups and Functional Groups Effector and/or onal groups can be attached, for example, to the N or C termini of the polypeptide, or to the molecular scaffold. riate effector groups include antibodies and parts or fragments thereof. For instance, an effector group can include an antibody light chain constant region (CL), an antibody CH1 heavy chain , an antibody CH2 heavy chain domain, an antibody CH3 heavy chain domain, or any combination thereof, in addition to the one or more constant region domains. An effector group may also comprise a hinge region of an antibody (such a region normally being found between the CH1 and CH2 domains of an IgG molecule).

In a further preferred embodiment, an effector group ing is an Fc region of an IgG molecule. Advantageously, a peptide ligand-effector group as described herein comprises or consists of a peptide ligand Fc fusion having a tb ife of a day or more, two days or more, 3 days or more, 4 days or more, 5 days or more, 6 days or more or 7 days or more. Most advantageously, the peptide ligand according to the present invention comprises or consists of a e ligand Fc fusion having a tb half-life of a day or more.

Functional groups include, in general, binding , drugs, reactive groups for the attachment of other entities, functional groups which aid uptake of the macrocyclic peptides into cells, and the like.

The ability of peptides to penetrate into cells will allow peptides against intracellular targets to be effective. Targets that can be accessed by es with the ability to penetrate into cells include transcription factors, intracellular signalling les such as tyrosine kinases and molecules involved in the apoptotic pathway. Functional groups which enable the penetration of cells include peptides or chemical groups which have been added either to the peptide or the molecular scaffold. Peptides such as those derived from such as VP22, HIV-Tat, a homeobox protein of Drosophila (Antennapedia), e.g. as described in Chen and Harrison, mical Society ctions (2007) Volume 35, part 4, p821 "Cell-penetrating peptides in drug development: enabling intracellular s" and “Intracellular delivery of large molecules and small peptides by cell penetrating peptides” by Gupta et al. in ed Drug Discovery Reviews (2004) Volume 57 9637. Examples of short peptides which have been shown to be efficient at translocation through plasma membranes include the 16 amino acid penetratin peptide from Drosophila Antennapedia n (Derossi et al (1994) J Biol. Chem. Volume 269 p10444 ‘’The third helix of the Antennapedia homeodomain ocates through ical membranes’’), the 18 amino acid ‘model amphipathic peptide’ (Oehlke et al (1998) Biochim Biophys Acts Volume 1414 p127 “Cellular uptake of an alpha-helical amphipathic model peptide with the potential to deliver polar compounds into the cell interior non-endocytically’’) and arginine rich regions of the HIV TAT protein. Non peptidic approaches include the use of small molecule mimics or SMOCs that can be easily ed to biomolecules (Okuyama et al (2007) Nature Methods Volume 4 p153 ‘Smallmolecule mimics of an a-helix for efficient transport of proteins into cells’. Other chemical strategies to add inium groups to molecules also enhance cell penetration (Elson- Scwab et al (2007) J Biol Chem Volume 282 p13585 dinylated Neomcyin Delivers Large ive Cargo into cells through a heparin Sulphate Dependent Pathway”). Small molecular weight molecules such as steroids may be added to the molecular scaffold to enhance uptake into cells.

One class of functional groups which may be attached to peptide ligands includes antibodies and binding fragments thereof, such as Fab, Fv or single domain fragments.

In particular, antibodies which bind to proteins capable of increasing the half life of the peptide ligand in vivo may be used.

RGD peptides, which bind to integrins which are present on many cells, may also be incorporated.

In one embodiment, a peptide ligand-effector group as described has a tb half-life selected from the group consisting of: 12 hours or more, 24 hours or more, 2 days or more, 3 days or more, 4 days or more, 5 days or more, 6 days or more, 7 days or more, 8 days or more, 9 days or more, 10 days or more, 11 days or more, 12 days or more, 13 days or more, 14 days or more, 15 days or more or 20 days or more. Advantageously a peptide ligand-effector group or composition as described will have a tb half life in the range 12 to 60 hours. In a further embodiment, it will have a t ife of a day or more. In a further embodiment still, it will be in the range 12 to 26 hours.

Functional groups include drugs, such as cytotoxic agents for cancer therapy. These include ting agents such as tin and carboplatin, as well as oxaliplatin, mechlorethamine, cyclophosphamide, mbucil, ifosfamide; Anti-metabolites including purine s azathioprine and mercaptopurine)) or pyrimidine analogs; plant alkaloids and terpenoids including vinca alkaloids such as stine, Vinblastine, lbine and Vindesine; Podophyllotoxin and its tives etoposide and teniposide; Taxanes, including paclitaxel, originally known as Taxol; omerase inhibitors including thecins: irinotecan and topotecan, and type II inhibitors including amsacrine, etoposide, etoposide phosphate, and side. Further agents can include mour antibiotics which include the immunosuppressant dactinomycin (which is used in kidney transplantations), doxorubicin, epirubicin, bleomycin and others.

Possible effector groups also include enzymes, for instance such as carboxypeptidase G2 for use in enzyme/prodrug therapy, where the peptide ligand replaces antibodies in ADEPT. (vi) Synthesis It should be noted that once a polypeptide of interest is isolated or identified as described herein, then its subsequent synthesis may be simplified wherever possible. Thus, groups or sets of polypeptides need not be produced by recombinant DNA techniques. For example, the ce of polypeptides of interest may be determined, and they may be manufactured synthetically by standard techniques followed by reaction with a molecular scaffold in vitro. When this is performed, standard chemistry may be used since there is no longer any need to preserve the functionality or integrity of the genetically encoded carrier particle, such as phage. This enables the rapid large scale preparation of e material for r downstream experiments or validation. In this , large scale preparation of the candidates or leads identified by the methods as described could be accomplished using conventional chemistry such as that disclosed in Timmerman et al.

Also described is the manufacture of polypeptides or conjugates selected as set out herein, wherein the manufacture comprises optional further steps as explained below. In one embodiment, these steps are carried out on the end product polypeptide/conjugate made by chemical synthesis, rather than on the phage.

Optionally amino acid residues in the polypeptide of interest may be substituted when manufacturing a conjugate or x e.g. after the initial isolation/identification step.

Peptides can also be extended, to incorporate for example another loop and therefore introduce multiple specificities.

To extend the peptide, it may simply be extended chemically at its inus or C- us or within the loops using orthogonally protected lysines (and analogues) using standard solid phase or solution phase try. Standard protein chemistry may be used to introduce an activatable N- or C-terminus. atively additions may be made by fragment condensation or native al ligation e.g. as described in (Dawson PE, Muir TW, Clark-Lewis I, Kent, SBH. 1994. Synthesis of Proteins by Native Chemical Ligation. Science 266:776-779), or by enzymes, for example using subtiligase as described in (Subtiligase: a tool for semisynthesis of proteins Chang TK, Jackson DY, Burnier JP, Wells JA Proc Natl Acad Sci U S A. 1994 Dec 20;91(26):12544-8 or in Bioorganic & nal Chemistry Letters Tags for labelling protein N-termini with subtiligase for mics Volume 18, Issue 22, 15 November 2008, Pages 6000- 6003 Tags for labeling protein N-termini with subtiligase for proteomics; Hikari A.I.

Yoshihara, Sami Mahrus and James A. Wells).

Alternatively, the peptides may be extended or modified by further conjugation h disulphide bonds. This has the onal advantage of allowing the first and second peptide to dissociate from each other once within the reducing environment of the cell. In this case, the molecular scaffold (eg. TBMB) could be added during the chemical synthesis of the first peptide so as to react with the three cysteine groups; a r cysteine could then be appended to the N-terminus of the first peptide, so that this cysteine only reacted with a free cysteine of the second peptide. r techniques apply equally to the synthesis/coupling of two bicyclic and bispecific macrocycles, potentially creating a tetraspecific molecule.

Furthermore, addition of other functional groups or effector groups may be accomplished in the same manner, using appropriate chemistry, coupling at the N- or C-termini or via side . In one embodiment, the coupling is conducted in such a manner that it does not block the activity of either entity. (vii) Peptide modification To develop the bicyclic peptides (Bicycles; peptides conjugated to molecular scaffolds) into a suitable drug-like molecule, r that be for ion, inhalation, nasal, ocular, oral or topical administration, a number of ties need considered. The following at least need to be ed into a given lead Bicycle: • protease stability, whether this concerns Bicycle ity to plasma proteases, epithelial ("membrane-anchored") proteases, gastric and intestinal proteases, lung surface proteases, intracellular proteases and the like. Protease stability should be maintained between different species such that a e lead candidate can be developed in animal models as well as administered with ence to • replacement of oxidation-sensitive residues, such as tryptophan and methionine with oxidation-resistant analogues in order to improve the ceutical stability profile of the molecule • a ble solubility profile, which is a function of the proportion of charged and hydrophilic versus hydrophobic residues, which is important for formulation and absorption purposes • correct balance of charged versus hydrophobic residues, as hydrophobic residues nce the degree of plasma protein binding and thus the concentration of the free available fraction in plasma, while charged residues (in particular arginines) may influence the interaction of the peptide with the phospholipid membranes on cell es. The two in combination may influence half-life, volume of distribution and exposure of the peptide drug, and can be tailored according to the al endpoint. In addition, the correct combination and number of charged versus hobic residues may reduce irritation at the injection site (were the peptide drug administered aneously). • a tailored half-life, depending on the clinical indication and treatment regimen. It may be prudent to develop an unmodified molecule for short exposure in an acute illness management setting, or develop a bicyclic peptide with chemical modifications that enhance the plasma half-life, and hence be optimal for the ment of more chronic disease states. ches to stabilise therapeutic peptide candidates against proteolytic degradation are numerous, and p with the peptidomimetics field (for reviews see Gentilucci et al, Curr. Pharmaceutical Design, (2010), 16, 3185-203, and Nestor et al, Curr. Medicinal Chem (2009), 16, 4399-418).

These include • Cyclisation of peptide • N- and C-terminal capping, usually N-terminal acetylation and C-terminal amidation.

• Alanine scans, to reveal and potentially remove the proteolytic attack site(s).

• D-amino acid replacement, to probe the steric requirements of the amino acid side chain, to increase proteolytic stability by steric hindrance and by a sity of D- amino acids to stabilise b-turn conformations (Tugyi et al (2005) PNAS, 102(2), 413–418).

• N-methyl/N-alkyl amino acid ement, to impart proteolytic protection by direct modification of the scissile amide bond (Fiacco et al, Chembiochem. (2008), 9(14), 2200–3). N-methylation also has strong effect on the nal angles of the peptide bond, and is believed to aid in cell penetration & oral availability (Biron et al (2008), Angew. Chem. Int. Ed., 47, 2595 –99) • Incorporation of non-natural amino acids, i.e. by employing - ric/isoelectronic side chains that are not recognised by proteases, yet have no effect on target potency - Constrained amino acid side chains, such that proteolytic hydrolysis of the nearby peptide bond is conformationally and sterically d. In particular, these concern proline analogues, bulky ains, Cadisubstituted derivatives (where the simplest tive is Aib, H2NC (CH3)2-COOH), and cyclo amino acids, a simple derivative being amino-cyclopropylcarboxylic acid).

• Peptide bond surrogates, and examples include - N-alkylation (see above, i.e. CO-NR) - Reduced peptide bonds (CH2-NH-) - Peptoids (N-alkyl amino acids, NR-CH2-CO) - Thio-amides (CS-NH) - Azapeptides (CO-NH-NR) - Trans-alkene (RHC=C-) - Retro-inverso (NH-CO) - Urea surrogates (NH-CO-NHR) • e backbone length modulation - i.e. b2/3- amino acids, (NH-CR-CH2-CO, NH-CH2-CHR-CO), • Substitutions on the alpha-carbon on amino acids, which constrains ne conformations, the simplest tive being Aminoisobutyric acid (Aib).

It should be explicitly noted that some of these modifications may also serve to deliberately improve the y of the e against the target, or, for example to identify potent substitutes for the oxidation-sensitive amino acids (Trp and Met). It should also be noted that the Bicycle lead Ac34-18(TMB)-NH2 already harbours two modifications that impart resistance to proteolytic degradation, these being N/C-terminal g, and (bi)cyclisation.

(B) Repertoires, sets and groups of polypeptide ligands (i) Construction of Libraries Libraries intended for selection may be constructed using techniques known in the art, for example as set forth in /077062, or biological systems, including phage vector systems as described herein. Other vector systems are known in the art, and include other phage (for instance, phage lambda), bacterial plasmid expression vectors, eukaryotic cell-based expression s, including yeast vectors, and the like. For example, see WO2009098450 or Heinis, et al., Nat Chem Biol 2009, 5 (7), 502-7.

Non-biological systems such as those set forth in WO2004/077062 are based on conventional chemical screening ches. They are simple, but lack the power of biological systems since it is impossible, or at least impracticably onerous, to screen large libraries of peptide ligands. However, they are useful where, for instance, only a small number of peptide ligands needs to be screened. Screening by such individual assays, however, may be time-consuming and the number of unique molecules that can be tested for binding to a specific target generally does not exceed 106 chemical entities.

In contrast, ical screening or selection methods generally allow the sampling of a much larger number of different molecules. Thus ical s can be used in application of the invention. In biological procedures, molecules are assayed in a single reaction vessel and the ones with favourable properties (i.e. binding) are physically ted from ve molecules. Selection strategies are ble that allow to generate and assay simultaneously more than 1013 individual compounds. Examples for powerful ty selection techniques are phage display, ribosome display, mRNA display, yeast display, bacterial display or RNA/DNA aptamer methods. These biological in vitro ion methods have in common that ligand repertoires are encoded by DNA or RNA. They allow the propagation and the identification of selected ligands by sequencing. Phage display technology has for e been used for the isolation of antibodies with very high binding affinities to virtually any target.

When using a biological , once a vector system is chosen and one or more nucleic acid sequences encoding polypeptides of interest are cloned into the library vector, one may generate diversity within the cloned molecules by undertaking mutagenesis prior to expression; alternatively, the d proteins may be sed and selected before mutagenesis and additional rounds of selection are performed.

Mutagenesis of nucleic acid sequences encoding structurally optimised polypeptides is carried out by standard molecular methods. Of ular use is the polymerase chain reaction, or PCR, (Mullis and a (1987) Methods Enzymol., 155: 335, herein incorporated by reference). PCR, which uses multiple cycles of DNA replication catalysed by a thermostable, pendent DNA polymerase to amplify the target sequence of interest, is well known in the art. The uction of various antibody libraries has been discussed in Winter et al. (1994) Ann. Rev. Immunology 12, 433-55, and references cited therein.

Alternatively, given the short chain lengths of the polypeptides according to the invention, the variants are preferably synthesised de novo and inserted into suitable expression vectors. Peptide synthesis can be carried out by rd techniques known in the art, as described above. Automated peptide synthesisers are widely available, such as the Applied Biosystems ABI 433 (Applied tems, Foster City, CA, USA) (ii) Genetically encoded diversity In one embodiment, the polypeptides of interest are genetically encoded. This offers the advantage of enhanced diversity together with ease of handling. An example of a genetically polypeptide library is a mRNA display library. Another example is a replicable genetic display package (rgdp) library such as a phage display library. In one embodiment, the polypeptides of interest are genetically d as a phage display library.

Thus, in one embodiment the complex comprises a replicable genetic display package (rgdp) such as a phage particle. In these ments, the nucleic acid can be comprised by the phage genome. In these embodiments, the polypeptide can be comprised by the phage coat.

In some embodiments, the invention may be used to produce a genetically encoded combinatorial library of polypeptides which are generated by translating a number of c acids into corresponding polypeptides and linking molecules of said molecular scaffold to said polypeptides.

The genetically encoded combinatorial library of polypeptides may be generated by phage display, yeast display, ribosome display, bacterial display or mRNA display.

Techniques and ology for ming phage display can be found in WO2009098450.

In one embodiment, screening may be med by contacting a library, set or group of polypeptide ligands with a target and isolating one or more member(s) that bind to said In another embodiment, individual members of said library, set or group are ted with a target in a screen and members of said library that bind to said target are identified.

In another embodiment, members of said library, set or group are simultaneously contacted with a target and members that bind to said target are selected.

The target(s) may be a peptide, a protein, a polysaccharide, a lipid, a DNA or a RNA.

The target may be a receptor, a receptor ligand, an enzyme, a e or a cytokine.

The target may be a prokaryotic protein, a eukaryotic protein, or an archeal n. More ically the target ligand may be a mammalian protein or an insect protein or a bacterial protein or a fungal protein or a viral protein.

The target ligand may be an enzyme, such as a protease.

It should be noted that the invention also embraces polypeptide ligands isolated from a screen as described herein. In one embodiment the ing method(s) further comprise the step of: manufacturing a quantity of the polypeptide ed as capable of binding to said targets.

The invention also relates to peptide ligands having more than two loops. For example, tricyclic polypeptides joined to a molecular ld can be created by joining the N- and C- termini of a bicyclic polypeptide joined to a molecular scaffold according to the present invention. In this manner, the joined N and C termini create a third loop, making a tricyclic polypeptide. This embodiment need not be carried out on phage, but can be carried out on a polypeptide–molecular scaffold conjugate as bed herein. g the N- and C- termini is a matter of routine peptide chemistry. In case any ce is needed, the C-terminus may be activated and/or the N- and C- termini may be extended for example to add a ne to each end and then join them by disulphide bonding.

Alternatively the joining may be accomplished by use of a linker region incorporated into the N/C termini. Alternatively the N and C termini may be joined by a conventional peptide bond. Alternatively any other suitable means for joining the N and C termini may be employed, for example N-C-cyclization could be done by standard techniques, for example as disclosed in Linde et al. Peptide Science 90, 671-682 (2008) "Structureactivity relationship and metabolic stability s of backbone cyclization and N- ation of melanocortin peptides", or as in Hess et al. J. Med. Chem. 51, 1026-1034 (2008) "backbone cyclic peptidomimetic melanocortin-4 receptor agonist as a novel orally administered drug lead for treating obesity". One advantage of such tricyclic molecules is the avoidance of proteolytic ation of the free ends, in particular by exoprotease action. r advantage of a tricyclic polypeptide of this nature is that the third loop may be utilised for generally applicable functions such as BSA binding, cell entry or transportation effects, tagging or any other such use. It will be noted that this third loop will not typically be available for selection (because it is not produced on the phage but only on the polypeptide-molecular scaffold conjugate) and so its use for other such biological functions still advantageously leaves both loops 1 and 2 for selection/creation of specificity. (iii) Phage purification Any suitable means for cation of the phage may be used. Standard techniques may be applied. For example, phage may be purified by filtration or by precipitation such as PEG itation; phage particles may be produced and purified by polyethylene-glycol (PEG) precipitation as described previously. Details can be found in WO2009098450.

In case r guidance is needed, reference is made to s et al (Protein Engineering Design and Selection 2004 17(10):709-713. Selection of optical biosensors from chemisynthetic antibody libraries.) In one embodiment phage may be purified as taught therein. The text of this publication is specifically orated herein by reference for the method of phage purification; in particular reference is made to the materials and methods n starting part way down the right-column at page 709 of Jespers et al.

Moreover, the phage may be purified as published by Marks et al J.Mol.Biol vol 222 pp581-597, which is specifically incorporated herein by reference for the particular description of how the phage production/purification is carried out. (iv) Reaction try The present disclosure makes use of chemical conditions for the modification of polypeptides which ageously retain the function and integrity of the genetically encoded element of the product. Specifically, when the genetically encoded element is a polypeptide displayed on the surface of a phage encoding it, the chemistry advantageously does not compromise the ical integrity of the phage. In general, conditions are set out in WO2009098450.

(C) Use of polypeptide ligands according to the invention Polypeptide ligands selected according to the method described may be employed in in vivo therapeutic and prophylactic applications, in vitro and in vivo diagnostic applications, in vitro assay and t applications, and the like. Ligands having selected levels of specificity are useful in applications which involve testing in non-human animals, where reactivity is desirable, or in diagnostic applications, where cross-reactivity with homologues or paralogues needs to be carefully lled. In some applications, such as vaccine applications, the ability to elicit an immune response to predetermined ranges of antigens can be exploited to tailor a vaccine to specific diseases and pathogens.

Substantially pure peptide ligands of at least 90 to 95% homogeneity are preferred for administration to a mammal, and 98 to 99% or more neity is most preferred for pharmaceutical uses, especially when the mammal is a human. Once purified, partially or to homogeneity as d, the selected polypeptides may be used diagnostically or therapeutically (including extracorporeally) or in developing and performing assay procedures, immunofluorescent stainings and the like vite and Pernis, (1979 and 1981) Immunological Methods, Volumes I and II, Academic Press, NY).

The peptide ligands of the t invention will typically find use in preventing, suppressing or treating inflammatory states, allergic hypersensitivity, cancer, bacterial or viral infection, and autoimmune disorders (which include, but are not d to, Type I es, multiple sclerosis, rheumatoid arthritis, systemic lupus erythematosus, Crohn's disease and myasthenia gravis).

In the instant ation, the term "prevention" es administration of the protective composition prior to the induction of the disease. "Suppression" refers to administration of the composition after an inductive event, but prior to the al appearance of the disease. "Treatment" involves administration of the protective composition after disease symptoms become manifest.

Animal model s which can be used to screen the iveness of the peptide ligands in protecting against or treating the disease are available. The use of animal model systems is facilitated by the present invention, which allows the development of polypeptide ligands which can cross react with human and animal targets, to allow the use of animal models.

Methods for the testing of systemic lupus matosus (SLE) in susceptible mice are known in the art (Knight et al. (1978) J Exp. Med., 147: 1653; Reinersten et al. (1978) New Eng. J : Med., 299: 515). Myasthenia Gravis (MG) is tested in SJL/J female mice by inducing the disease with soluble AchR protein from another species (Lindstrom et al. (1988) Adv. Inzn7unol., 42: 233). Arthritis is induced in a susceptible strain of mice by injection of Type II collagen t et al. (1984) Ann. Rev. l., 42: 233). A model by which adjuvant arthritis is induced in susceptible rats by injection of cterial heat shock protein has been described (Van Eden et al. (1988) Nature, 331: 171). Thyroiditis is induced in mice by administration of thyroglobulin as described (Maron et al. (1980) J.

Exp. Med., 152: 1115). Insulin dependent diabetes mellitus (IDDM) occurs lly or can be induced in certain strains of mice such as those described by Kanasawa et al. (1984) Diabetologia, 27: 113. EAE in mouse and rat serves as a model for MS in human.

In this model, the demyelinating disease is induced by administration of myelin basic protein (see Paterson (1986) Textbook of Immunopathology, Mischer et al., eds., Grune and Stratton, New York, pp. 3; McFarlin et al. (1973) Science, 179: 478: and Satoh et al. (1987) J ; Immunol., 138: 179). lly, the present peptide ligands will be utilised in purified form together with pharmacologically appropriate carriers. Typically, these carriers include aqueous or alcoholic/aqueous solutions, emulsions or suspensions, any including saline and/or buffered media. Parenteral vehicles include sodium chloride solution, Ringer's dextrose, dextrose and sodium chloride and lactated Ringer's. le physiologically-acceptable adjuvants, if necessary to keep a polypeptide complex in suspension, may be chosen from thickeners such as carboxymethylcellulose, nylpyrrolidone, n and alginates.

Intravenous vehicles include fluid and nutrient replenishers and electrolyte replenishers, such as those based on Ringer's dextrose. Preservatives and other additives, such as antimicrobials, antioxidants, ing agents and inert gases, may also be present (Mack (1982) Remington's ceutical Sciences, 16th Edition).

The peptide ligands of the present invention may be used as tely administered compositions or in ction with other agents. These can include antibodies, dy fragments and various immunotherapeutic drugs, such as cylcosporine, methotrexate, adriamycin or cisplatinum, and toxins. Pharmaceutical compositions can include "cocktails" of various cytotoxic or other agents in conjunction with the selected antibodies, receptors or binding proteins f of the present invention, or even combinations of selected polypeptides according to the present invention having different specificities, such as polypeptides selected using different target s, whether or not they are pooled prior to administration.

The route of administration of pharmaceutical compositions as described herein may be any of those commonly known to those of ordinary skill in the art. For therapy, including without limitation immunotherapy, the selected antibodies, receptors or binding proteins thereof of the invention can be administered to any patient in accordance with standard techniques. The administration can be by any appropriate mode, including parenterally, intravenously, intramuscularly, intraperitoneally, transdermally, via the pulmonary route, or also, appropriately, by direct on with a catheter. The dosage and ncy of administration will depend on the age, sex and condition of the patient, concurrent administration of other drugs, counterindications and other ters to be taken into account by the clinician.

The peptide ligands of this invention can be lyophilised for storage and reconstituted in a suitable carrier prior to use. This technique has been shown to be effective and art-known lyophilisation and reconstitution techniques can be employed. It will be appreciated by those skilled in the art that lyophilisation and reconstitution can lead to varying degrees of activity loss and that use levels may have to be adjusted upward to compensate.

The compositions containing the present peptide ligands or a cocktail thereof can be administered for prophylactic and/or therapeutic treatments. In n eutic applications, an adequate amount to lish at least partial inhibition, suppression, modulation, killing, or some other measurable ter, of a population of selected cells is defined as a "therapeutically-effective dose". Amounts needed to achieve this dosage will depend upon the severity of the disease and the general state of the patient's own immune system, but generally range from 0.005 to 5.0 mg of selected peptide ligand per kilogram of body weight, with doses of 0.05 to 2.0 mg/kg/dose being more commonly used. For prophylactic applications, itions containing the present peptide ligands or cocktails thereof may also be administered in similar or slightly lower dosages.

A composition containing a peptide ligand according to the present invention may be utilised in prophylactic and therapeutic settings to aid in the alteration, inactivation, g or removal of a select target cell population in a mammal. In addition, the selected repertoires of polypeptides described herein may be used extracorporeally or in vitro selectively to kill, deplete or otherwise ively remove a target cell tion from a heterogeneous collection of cells. Blood from a mammal may be combined extracorporeally with the selected peptide ligands y the red cells are killed or otherwise removed from the blood for return to the mammal in accordance with standard techniques.

(D) on of Polypeptides The desired diversity is typically generated by varying the selected molecule at one or more positions. The positions to be changed are selected, such that libraries are constructed for each individual position in the loop sequences. Where appropriate, one or more positions may be omitted from the ion procedure, for instance if it s nt that those positions are not available for mutation without loss of activity.

The variation can then be achieved either by randomisation, during which the resident amino acid is replaced by any amino acid or analogue thereof, natural or synthetic, producing a very large number of variants or by replacing the resident amino acid with one or more of a defined subset of amino acids, producing a more limited number of variants.

Various methods have been reported for introducing such diversity. Methods for mutating selected positions are also well known in the art and e the use of mismatched oligonucleotides or degenerate oligonucleotides, with or without the use of PCR. For example, several tic antibody libraries have been d by targeting mutations to the antigen binding loops. The same techniques could be used in the context of the present invention. For example, the H3 region of a human tetanus toxoid-binding Fab has been randomised to create a range of new binding specificities (Barbas et al. (1992) Proc. Natl. Acad. Sci. USA, 89: 4457). Random or semi-random H3 and L3 s have been appended to germline V gene segments to e large libraries with mutated framework regions nboom- & Winter (1992) R Mol. Biol., 227: 381; Barbas et al. (1992) Proc. Natl. Acad. Sci. USA, 89: 4457; Nissim et al. (1994) EMBO J, 13: 692; Griffiths et al. (1994) EMBO J, 13: 3245; De Kruif et al. (1995) J. Mol. Biol., 248: 97).

Such diversification has been extended to include some or all of the other antigen binding loops (Crameri et al. (1996) Nature Med., 2: 100; Riechmann et al. (1995) BiolTechnology, 13: 475; Morphosys, W097/08320, supra).

However, since the polypeptides used in the present invention are much smaller than antibodies, the preferred method is to synthesise mutant ptides de novo.

Mutagenesis of structured polypeptides is described above, in connection with library construction.

The invention is further described below with reference to the following examples.

Materials and Methods Cloning of phage libraries Phage libraries were generated according to Heinis et al., Nat Chem Biol 2009, 5 (7), 502-7). In Heinis et al, the genes encoding a semi-random peptide with the sequence Xaa-Cys-(Xaa)3-Cys-(Xaa)3-, the linker Gly-Gly-Ser-Gly and the two disulfide-free s D1 and D2 (Kather, et al., J Mol Biol 2005, 354 (3), 666-78) were cloned in the correct orientation into the phage vector fd0D12 to obtain ‘library 3x3’. The genes ng the peptide repertoire and the two gene 3 domains were step-wise created in two consecutive PCR reactions. First, the genes of D1 and D2 were PCR amplified with the two primer prepcr (5’-GGCGGTTCTGGCGCTGAAACTGTTGAAAGTAG-3’) and sfi2fo (5’- GAAGCCATGGCCCCCGAGGCCCCGGACGGAGCATTGACAGG-3’; restriction site is underlined) using the vector fdg3p0ss21 (Kather, et al., J Mol Biol 2005, 354 (3), 666-78) as a template. Second, the DNA ng the random peptides was ed in a PCR reaction using the primer sficx3ba: 5’- TATGCGGCCCAGCCGGCCATGGCANNKTGTNNKNNKNNKTGCNNKNNKNNKNNKTGTNNKG GGCGGTTCTGGCGCTG-3’ (restriction site is underlined), and sfi2fo. The ligation of 55 and 11 mg of SfiI-digested fd0D12 plasmid and PCR product yielded 5.6x108 colonies on 20x20 cm chloramphenicol (30 mg/ml) 2YT plates. Colonies were scraped off the plates with 2YT media, supplemented with 15% glycerol and stored at -80°C.

Construction of the libraries described herein employed the same que to generate the andom peptide Pro-Ala-Met-Ala-Cys-(Xaa)3-Cys-(Xaa)3-Cys for a 3x3 library for example, and therefore replaced the sficx3ba primer sequence with: 5’- TATGCGGCCCAGCCGGCCATGGCATGTNNKNNKNNKTGCNNKNNKNNKTGTGGCGGTTCTG GCGCTG-3’. Libraries with other loop lengths were generated following the same methodology.

Phage selections Glycerol stocks of phage ies were diluted to OD600=0.1 in 500 ml loramphenicol (30 mg/ml) cultures and phage were produced at 30°C over night (15-16 hrs). Phage were purified and chemically modified as described in Heinis, et al., Nat Chem Biol 2009, 5 (7), 502-7 Biotinylated hPK (3 mg) (IHPKA, from human plasma, Innovative Research, Novi, MI, USA) was incubated with 50 ml pre-washed magnetic streptavidin beads (Dynal, M-280 from ogen, Paisley, UK) for 10 minutes at RT.

Beads were washed 3 times prior to blocking with 0.5 ml washing buffer (10 mM Tris-Cl, pH 7.4, 150 mM NaCl, 10 mM MgCl2, 1mM CaCl2) containing 1% BSA and 0.1% Tween for 30 minutes at RT with rotation. Chemically modified phage (typically 1010-1011 t.u. dissolved in 2 ml washing buffer) were concomitantly blocked by addition of 1 ml g buffer containing 3% BSA and 0.3% Tween 20. Blocked beads were then mixed with the blocked chemically modified phage and incubated for 30 minutes on a ng wheel at RT. Beads were washed 8 times with washing buffer containing 0.1% Tween 20 and twice with washing buffer before incubation with 100 ml of 50 mM glycine, pH 2.2 for 5 minutes. Eluted phage were erred to 50 ml of 1 M Tris-Cl, pH 8 for neutralization, incubated with 30 ml TG1 cells at OD600=0.4 for 90 minutes at 37 C and the cells were plated on large 2YT/chloramphenicol plates. One or two additional rounds of panning were performed using the same procedures. In the second round of selection, neutravidin-coated magnetic beads were used to prevent the enrichment of streptavidinspecific peptides. The neutravidin beads were prepared by reacting 0.8 mg neutravidin (Pierce, Rockford, IL, USA) with 0.5 ml tosyl-activated magnetic beads , M-280 from Invitrogen, Paisley, UK) ing to the supplier’s instructions.

Cloning and expression of human, monkey and rat PK The catalytic domain of human, monkey and rat PK was expressed in mammalian cells as an inactive precursor having a pro-peptide connected N-terminally via a proTEV ge site to the catalytic domain. The expression vector was cloned and the protein expressed, activated and purified as described as follows. Synthetic genes coding for a PK signal sequence, a polyhistidine tag, a proTEV cleavage site, mature catalytic domain of PK and a stop codon were purchased from Geneart sburg, Germany) (Supplementary materials). d DNA containing the synthetic genes for human, monkey (Macaca mulatta) and rat PK was prepared and the gene transferred into the pEXPR-IBA42 mammalian expression vector (IBA Biotechnology, Göttingen, y) using the restriction enzyme pair XhoI and HindIII (Fermentas, s, Latvia) and T4 DNA ligase (Fermentas). The ligated plasmids were transformed into XL-1 blue electrocompetent cells (Stratagene, Santa Clara, USA) and plated onto 2YT agar plates containing ampicillin (10 μg/ml). DNA from the three sion vectors (termed mPK, rPK and hPK) was produced and the correct sequences confirmed by DNA sequencing (Macrogen, Seoul, South Korea).

The three orthologous plasma Kallikreins were expressed in mammalian cells as follows. 50 ml of suspension-adapted HEK-293 cells were grown in serum-free ExCell 293 medium (SAFC ences, St. Louis, MO) in the presence of 4 mM glutamine and the histone deacetylase inhibitor valproic acid (3.75 mM) in an orbitally shaken 100 ml flask at 180 rpm in an ISFW incubator (Kühner AG, Birsfelden, Switzerland) at 37 °C in the presence of 5% CO2. The nic kidney 93) cells at high cell density (20 x 106 cells/ml) (Backliwal, et al/. Biotechnol Bioeng 2008, 99 (3), 721-7) were transfected with the three plasmids (300mg/ml) using linear polyethylenimine (PEI, Polysciences, Eppenheim, Germany). At the end of the 7-day production phase, cells were harvested by centrifugation at 2'500 rpm for 15 min at 4°C. Any additional cell debris was removed from the medium by filtration through 0.45 μm PES nes (Filter-top 250 ml low protein binding TPP). The polyhistidine-tagged protein was purified by Ni-affinity chromatography using Ni-NTA resin, washing buffer (500mM NaCl, 25mM Na2HPO4, pH7,4) and n buffer (500 mM NaCl, 25 mM Na2HPO4, pH 7,4, 500 mM imidazole).

The n was partially activated with (50 units) proTEV (Promega, Madison, Wisconsin, USA) and additionally purified by Ni-affinity tography and gel filtration (PD10 column, 150 mM NaCl, 0,5 mM EDTA, 50 mM HEPES, pH 7).

Development of polypeptides with improved g activity isation of individual positions Library construction: In order to map the amino-acids in the Kallikrein binding bicyclic peptides a set of small ies was constructed. For a bicycle comprised of 2 loops of 5 residues, 10 separate libraries were generated each with randomisation at a particular codon in the peptide sequence. Oligonucleotides were designed for each library in order to mutate the phage genome DNA by site-directed mutagenesis. The mutagenesis incorporated randomisation of the codon of interest (change to NNS), and removal of a unique ApaL1 restriction site from the template genome ce. The mutagenesis product was purified using QIAgen QIAquick PCR purification kit with elution into ultrapure water. Each library was used to separately transform TG1 E coli by electroporation with a BioRad Micropulser machine (Ec1 program) and 1mm BioRad cuvette. After 1 hour ry at 37C in 1ml SOC media, the library transformants were grown overnight in 25ml 2TY broth containing antibiotic to selectively grow library transformants only. The bacteria were harvested by centrifugation and the library phage DNA was purified from the E coli using a QIAgen d Plus Midi kit and eluted in distilled water. The purified DNA was digested with ApaL1 for 2 hours in New England Biolabs buffer 4 to remove the parent al. After digestion, the DNA was repurified using QIAgen PCR purification kit (as above) and used to transform TG1 (electroporation; as described above). ing the 1 hour recovery in SOC, transformants were plated on LB-agar plates containing selective antibiotic and colonies allowed to grow overnight at 37C.

Assay of binding of individual clones: Library ormant colonies were picked at random and grown as individual cultures in 2TY broth containing selective antibiotic. The picked colonies were DNA-sequenced using a QIAgen PyroMark Q96 DNA sequencer to reveal the amino-acid substitution present in each clone. Where isolated, a clone of each unique substitution was assayed for human plasma Kallikrein binding as follows. The phage-containing supernatant was harvested from the culture and phage were cyclised with tris bromomethyl benzene(TBMB) based on the methods of Heinis et al (Nature Chemical y vol. 5 pp 502-507 (2009)). The purified phage from this process were assayed for binding to biotinylated human plasma Kallikrein using a homogeneous platebased binding assay; assay read-out measured on a BMG h Pherastar FS plate reader. The quantitative binding data from cate assay s was averaged (mean) and expressed as signal:background (where ound was a sample assayed with no target material). The signal:background was expressed as a % of the parallel parent . Error bars denote standard deviation of the mean. Assays shown are representative of at least 2 ndent experiments. The assay data was correlated with the peptide sequences. Substitutions marked in grey were not tested (a clone was not isolated from the random library sampling). A sample of a non-binding (arbitrary) bicycle was assayed in parallel to illustrate the assay baseline.

Randomisation of peptide domains Library construction: Small phage libraries were generated according to the methods of Heinis et al as described in ng of phage ies’ above. The sficx3ba primer was modified such that the bicycle-encoding portion was based on a parent 5x5 bicycle (5x5: two 5-residue loops) DNA sequence with only 4-6 codons randomized to NNS. The randomized codons were those encoding the peptide domain/motif of interest.

Assay of binding of individual clones: Library transformant colonies, or selection output colonies, were picked and grown as individual cultures in 2TY broth containing selective otic. The picked colonies were DNA-sequenced using a QIAgen PyroMark Q96 DNA cer to reveal the amino-acid substitution present in each clone, and were d for human plasma Kallikrein binding as follows. The phage-containing supernatant was ted from the culture and phage were cyclised with tris bromomethyl benzene(TBMB) based on the methods of Heinis et al (Nature Chemical y vol. 5 pp 502-507 (2009)). The purified phage from this process were assayed for binding to biotinylated human plasma Kallikrein using a homogeneous plate-based binding assay; assay read-out measured on a BMG Labtech Pherastar FS plate reader. The quantitative g data from duplicate assay samples was averaged (mean) and expressed as :background. Assay data shown is representative of at least 2 independent experiments. The assay data was correlated with the peptide sequences.

Synthesis and purification of bicyclic es Peptide sequences are shown in Tables 1 and 2. Peptide synthesis was based on Fmoc chemistry, using a Symphony peptide siser manufactured by Peptide Instruments.

Standard Fmoc-amino acids were ed (Sigma, Merck), with the following side chain protecting groups: Arg(Pbf); Asn(Trt); Asp(OtBu); Cys(Trt); GIu(OtBu); Gln(Trt); His(Trt); Lys(Boc); Ser(tBu); u); c), Tyr(tBu) ). The coupling reagent was HCTU (Pepceuticals), diisopropylethylamine (DIPEA, Sigma) was employed as a base, and deprotection was achieved with 20% piperidine in DMF (AGTC). Syntheses were performed at 100 umole scale using 0.37 mmole/gr Fmoc-Rink amide AM resin (AGTC), Fmoc-amino acids were utilised at a four-fold excess, and base was at a four-fold excess with respect to the amino acids. Amino acids were dissolved at 0.2 M in DMF, HCTU at 0.4 M in DMF, and DIPEA at 1.6 M in N-methylpyrrolidone (Alfa Aesar). Coupling times were lly 30 s, and deprotection times 2 x 2.5 minutes. Fmoc-N- methylglycine (Fmoc-Sar-OH, Merck) was coupled for 1 hr, and deprotection and ng times for the following residue were 20 min and 1 hr, respectively. After synthesis, the resin was washed with dichloromethane, and dried. Cleavage of side-chain protecting groups and from the support was ed using 10 mL of 95:2.5:2.5:2.5 v/v/v/w TFA/H2O/iPr3SiH/dithiothreitol for 3 hours. Following cleavage, the spent resin was removed by filtration, and the filtrate was added to 35 mL of diethylether that had been cooled at -80 deg C. Peptide pellet was centrifuged, the etheric supernatant discarded, and the peptide pellet washed with cold ether two more times. Peptides were then resolubilised in 5-10 mL acetonitrile-water and lyophilised. A small sample was d for analysis of purity of the crude product by mass spectrometry (MALDI-TOF, Voyager DE from Applied Biosystems). Following lyophilisation, peptide powders were taken up in 10 mL 6 M guanidinium hydrochloride in H2O, supplemented with 0.5 mL of 1 M dithiothreitrol, and loaded onto a C8 Luna preparative HPLC column (Phenomenex).

Solvents (H2O, acetonitrile) were ied with 0.1 % heptafluorobutyric acid. The gradient ranged from 30-70 % acetonitrile in 15 minutes, at a flowrate of 15/20 mL /min, using a Gilson preparative HPLC system. Fractions containing pure linear peptide material (as identified by MALDI) were combined, and modified with trisbromomethylbenzene (TBMB, . For this, linear e was diluted with H2O up to ~35 mL, ~500 uL of 100 mM TBMB in acetonitrile was added, and the reaction was initiated with 5 mL of 1 M NH4HCO3 in H2O (pH 8). The reaction was d to proceed for ~30 -60 min at RT, and lyophilised once the reaction had completed (judged by MALDI). Following lyophilisation, the modified peptide was purified as above, while replacing the Luna C8 with a Gemini C18 column (Phenomenex), and changing the acid to 0.1% trifluoroacetic acid. Pure fractions containing the correct TMB-modified material were pooled, lyophilised and kept at -20 deg C for storage.

Non-natural amino acids were ed from the sources set forth in Table 7.

Bulky or hindered amino acids (NMe-Ser, NMe-Trp, NorHar, 4PhenylPro, Agb, Agp, NMe- Arg, Pen, Tic, Aib, Hyp, NMe-Ala, NMe-Cys, Al, A, Dpg, 1NAl, 2NAl, Aze, 4BenzylPro, Ind) were usually coupled for 1 hours (20 min deprotection), and 6 hrs for the residue that followed (20 min ection). HCTU was used as a coupling reagent as before. Scale was usually at 50 umole.

Enzyme assays Functional enzyme assays were conducted in 10mM Tris HCl, 150mM NaCl, 10mM MgCl2, 1mM CaCl2 and 1mg/mL BSA (all Sigma UK) pH7.4 at 25°C in solid black 96 well . Briefly 26.5pM human plasma rein (purchased from Stratech, UK) or 500pM rat plasma Kallikrein (expressed and purified in house) were incubated in the absence or presence of increasing concentrations of test peptide for 15 minutes before addition of the fluorogenic ate rg-AMC (Enzo Lifesciences UK) to a final assay concentration of 100µM in 4% DMSO. Release of AMC was measured using a Pherastar FS (BMG Labtech), excitation 360nm, emission 460nm. The rate of the linear phase of the reaction, typically 5 to 45 minutes, was calculated in MARS data analysis software (BMG labtech). The rate was then used to calculate the IC50 and Ki in Prism (GraphPad).

A four parameter inhibition non-linear regression equation was used to calculate the IC50. The One site – fit Ki on used to calculate the Ki, constraining the Ki to the Km for the substrate which is 150µM. All Ki/IC50 values are the mean of at least two independent experiments, and at least three for peptides with Ki values lower than 1 nM.

Peptides were dissolved as the TFA-salts in their powder form, and stock solutions were usually prepared in water. All solutions were centrifuged and filtered (20 mm syringe filters) prior absorption measurement at 280 nm. Extinction coefficients were calculated based on the Trp/Tyr content of the peptide, and that of TMB (the TMB core, when contained in a peptide, has an e of ~300 M-1cm-1). For peptides containing non-natural amino acids with suspected chromophoric properties (i.e. NorHar, 4PhenylPro, 3Pal, 4Pal, Tic, he, 4,4-BPAl, 3,3-DPA, 1NAl, 2NAl, 4BenzylPro, Ind) concentrations were ined by weighing the powder and dissolving the peptide in a defined quantity of water. These were ed independently, twice, for peptides with a Ki to rein at 1 nM or less.

Plasma ity profiling Three methods were employed to assess the stability of es (peptides conjugated to molecular scaffolds) in plasma.

Method 1: A rapid plasma stability profiling assay was developed that employed mass spectrometric detection (MALDI-TOF, Voyager DE, Applied Biosystems) of the parent mass, until the time when the parent peptide mass was no longer observable.

Specifically, 200 uM of peptide was incubated in the ce of 35% rat or human plasma (Sera labs, using citrate as agulant) at 37 deg C, which was supplemented with 1 x PBS (derived from a 10 xPBS Stock, Sigma). At various time points (i.e. t = 0, 3, 24 hrs, henceafter daily up to 10 days), 2 uL of sample was added to 18 uL of 30 mM ammonium bicarbonate in a 1:1 mixture of acetonitrile:H2O. Samples were frozen at -80 deg C until the time of analysis. For mass spectrometric is that determines the approximate detection window of the peptide, the acetonitrile:H2O-diluted sample of a given time point was spotted directly (0.7 uL) onto the MALDI plate. Matrix (alphacyanocinnamic acid, Sigma, ed as a ted solution in 1:1 itrile:water containing 0.1% trifluoroacetic acid) was layered over the sample (1 uL). At a similar laser intensity setting on the MALDI TOF, the time could then be determined until parent peptide was no longer detectable. It should be noted that this is a qualitative assay serves to detect relative changes in plasma stability.

Method 2: To obtain ity data more rapidly, peptides were also ed in 95% plasma.

Here, PBS was omitted, and a 1 mM peptide stock (in DMSO) was directly diluted into plasma (i.e. 2.5 uL stock into 47.5 uL plasma), giving a final concentration of 50 uM. 5 uL samples were taken at appropriate time points and frozen at -80 deg C. For analysis, the samples were defrosted, mixed with 15 uL of 1:1 acetonitrile:methanol, and centrifuged at 13k for 5 min. 5 uL of the peptide-containing supernatant was aspirated and mixed with mM ammonium bicarbonate in a 1:1 mixture of acetonitrile:H2O. 1 uL of this was then spotted on the MALDI plate and analysed as described above. As above, it should be noted that this is a qualitative assay serves to detect relative changes in plasma ity.

Method 3: To obtain plasma stability quantitatively, peptide stock solutions (1 mM in DMSO) were shipped to Biofocus, UK, who performed the analysis. Peptides were diluted to 100 uM with water, and diluted 1:20 in plasma (5 uM final concentration, with the plasma at 95%), sampled as appropriate, precipitated as above, and quantified using a Waters Xevo TQ-MS.

Example 1: Identification of preferred residues for binding activity From the examples of 5x5 peptides shown in Table 4 it is possible to identify amino acids that are conserved between peptides with binding activity. To determine which residues were preferred for binding activity, representatives from two of the identified families of es were studied further. These were peptides 06-34, which comprises a CXWPARC motif in the first loop of the bicycle, and peptide 06-56, which comprises a CGGxxNCR motif across both loops of the bicycle. For each peptide sequence, a set of phage libraries was created in which 9 of the loop es were kept constant and the other residue was randomised so that any acid could be expressed in the library at that position. (See ‘Randomisation of individual positions – Library construction’ in Methods above.) For each library a set of 20 randomly selected phage clones were screened for binding to human Kallikrein in a phage binding assay to identify the critical residues for target binding. (See ‘Randomisation of individual positions – Assay of binding of individual clones’ in Methods above.) The data from this experiment are shown in Figures 4 - 6.

For peptide 06-34 (Figure 4), it is clear that Arg1 of the e can be replaced with a variety of different acids and binding to human plasma rein is retained or enhanced. By contrast, replacement of residues 2, 3, 4, 5 (Trp2, Pro3, Ala4, Arg5) by most amino acids greatly reduced the signal seen in an assay that was set up with a stringent cut-off for high affinity s. Val6 can be replaced by many different cids and binding activity is retained or enhanced. Replacement of other residues in the second loop indicated that only Leu10 could be replaced by a variety of different aminoacids whilst retaining ty. Positions 7, 8, and 9 have limited capacity for substitution and no substitutions were identified that enhanced binding.

For peptide 06-56 (Figures 5 and 6) it is clear that glycines at position 1 and 2 are the greatly preferred residues for g to plasma Kallikrein as are arginine, tryptophan and threonine at positions 6, 8, and 9. Glutamine at position 4 and ine at position 10 can be replaced by a variety of residues whilst retaining good g activity. The three ing residues – proline at position 1, asparagine at position 5 and threonine at on 7 have limited capacity for substitution.

Analysis of amino-acid replacements From the preceding analysis it is apparent that for 06-34, position one and position six can be replaced by a variety of amino-acids and still retain binding ty equal or greater than that of the parent peptide. To evaluate whether these observations would hold with isolated synthetic peptides, a set of peptides was designed according to the findings in Figure 4, where Arg1 was replaced by a serine, and where Val6 was tuted by either threonine, methionine or leucine. Peptides employing the various combinations of these tutions were also synthesised. These substitutions produced a greater binding signal in the assay (Table 5).

All of the variant synthetic peptides had approximately equivalent or ed activity against human plasma Kallikrein in enzyme inhibition assays compared to the 06-34 parent peptide, indicating that this type of analysis could be used to fine-tune target binding affinities, and suggesting a route to identifying lead peptide candidates of very high potencies.

The peptides were also tested against rat plasma Kallikrein in isolated enzyme assays.

Substitution of Arg1 to Ser1 had a al impact on activity against rat Kallikrein, whereas substitutions of Val6 to threonine, methionine or leucine generated peptides with markedly increased potency against rat plasma Kallikrein. Activity to human Kallikrein was fully retained. Thus, by ining positions amenable to tutions, peptides with ble properties, such as target orthologue cross-reactivity, can be identified.

To demonstrate the possibility of replacing these two positions with non-natural amino acids so as to have the capacity to introduce functionalities or properties that are not present in the parent peptide, Arg1 and Val6 in 0603 were replaced with either alanine or N-methylglycine (sarcosine), or with N-methyl serine on position 1, and evaluated for binding. Remarkably, as shown in Table 6, positions 1/6 are amenable to removal of the side chain altogether, as the R1A/V6A (0603 ) peptides retained full potency compared to the parent. Replacement of residues 1,6 with N-methylglycine (0603 NMeGly1,6) caused a reduction in y, however the binding affinity remained in the low nanomolar range. uction of an N-methylserine at position 1 causes a ten-fold loss in y, but binding remains in the picomolar range. Thus, certain positions in the bicycle can be identified that allow changes in the peptide backbone structure or side chains, which could allow for rate enhancement of protease stability, ed solubility, reduced aggregation potential, and introduction of orthologous functional .

Example 2: Detailed analysis of WPAR domain.

The WPAR motif identified from Example 1 was analysed in the context of the 0603 peptide, in order to identify alternatives or improvements to the WPAR motif. A library was ucted where positions 1, 6, 7, 8, 9 &10 of 0603 were fixed and positions 2, 3, 4 & 5 were ised (see ‘Randomisation of peptide domains – Library construction’ in Methods above). Selections against human plasma Kallikrein were performed at a y of encies (see Phage selections’ in Methods above). All output sequences were fied and ed for target binding (see ‘Randomisation of peptide domains – Assay of binding of individual ’ in Methods above). Table 17 lists each unique sequence, its relative abundance in the selection output (frequency), and a rank number according to target binding strength.

Table 17 shows that WPAR motif confers the best binding to human plasma Kallikrein, although other Kallikrein binding sequences are retrieved from selections in high abundance. These include, but are not cted to: WPSR, WPAR, WSAR, WPFR, WPYR, FPFR, & FPFR. The most effective and abundant motifs at positions 2, 3, 4 & 5 can be summarised as: W/ K/ F P x R Table 18(A) shows that WPAR & WPSR were most abundant in the more stringent selection outputs; FPFR & FPYR were abundant in the lower stringency selection outputs. This would te that WPAR-like ces are stronger binders than FPFR- like sequences. Analysis of each motif (within the 0603 context) in the target binding assay (Table 18(B)), reveals that WPAR at positions 2, 3, 4 & 5 of the 0603 sequence is the optimal sequence for Kallikrein binding.

Example 3: Optimisation of sequence outside WPAR The WPAR motif and its variants have been studied within the context of peptide 06 03. Figure 1 demonstrates that some positions outside of the WPAR motif can maintain Kallikrein binding when substituted for other residues. In order to study the non-WPAR determinants of Kallikrein binding, a phage library was generated with a fixed-WPAR sequence and all other positions randomised (CxWPARCxxxxxC) as described in misation of peptide domains – Library construction’ in Methods above. 80 random library members were isolated directly from the library pool (no selection) and assayed for binding to Kallikrein at both high and low stringency (see ‘Randomisation of peptide domains – Assay of binding of individual ’ in Methods above). These library s, which contain random sequences outside the WPAR, showed little or no binding to human plasma Kallikrein (data not shown), indicating that the presence of a WPAR motif alone is not sufficient to retain measureable rein binding: the rest of the bicycle sequence must also contribute or influence the interaction.

Selections against human plasma Kallikrein were performed with this library in order to study the non-WPAR determinants of Kallikrein binding, and to isolate the optimal WPAR- ning peptide sequence. Over 150 selection output sequences were isolated and screened for binding to human plasma Kallikrein (as described in Methods above). The sequences were ranked in order of rein binding and the top 50 sequences were aligned in Table 19. Table 19 shows that the residue at position 1 does not affect Kallikrein binding, but a strong consensus for Histidine is seen at position 7 (which ts findings in Example 1 above). The peptide 03 – derived from the work in Example 1 – is one of the best ces. The composition of the second loop shows clear trends which confer strong Kallikrein binding when with a WPAR motif.

The best ontaining binders to human plasma Kallikrein have the trend: C X W P A R C T/ Q/ L H T D L C H7, D9 and L10 are heavily conserved in WPAR-containing Kallikrein binding ces.

Two motifs in within the second bicycle loop (positions 6-10) were identified: 1. C X W P A R C T H Q/ T D L C (positions 6, 7 &10: “THxxL”) 2. C X W P A R C T/ Q/ L H T D L C (positions 7, 8, & 10: “xHxDL”) Over 120 identified human plasma Kallikrein s (selection output sequences) were grouped 2 different ways, according to their derivation from motifs “THxxL” or “xHxDL”.

For all groups, the average Kallikrein binding assay signal for output sequences was noted as a measure of Kallikrein binding for a given group (Table 20).

The rein binding assay data shown in Table 20 demonstrates that ’ and ‘xHxDL’ motifs result in the best Kallikrein binding when in a bicyclic peptide with a WPAR motif. The combination of the 2 motifs, ‘THxDL’, gives the highest binding to human plasma Kallikrein and includes the ‘THQDL’ second loop sequence of the 0603 peptide.

Example 4: Systematic analysis of plasma stability.

For a Kallikrein-inhibiting bicycle, it is pertinent to obtain an adequate se stability profile, such that it has a low protease-driven clearance in plasma or other relevant environments. In a rapid comparative plasma stability assay (methods section, method1) that observed the progressive disappearance of parent peptide in rat plasma, it was found that the N-terminal alanine (which is present at the time of selections and was originally included in tic peptides of lead sequences) is rapidly removed across all bicycle sequences tested by both rat and human plasma. This degradation was avoided by synthesising a lead candidate lacking both N- and C-terminal alanines. To remove potential recognition points for amino- and carboxypeptidases, the free amino-terminus that now resides on Cys 1 of the lead candidate is capped with acetic anhydride during peptide synthesis, leading to a molecule that is N-terminally acetylated. In an equal measure, the C-terminal cysteine is synthesised as the amide so as to remove a potential recognition point for carboxypeptidasese. Thus, ic lead candidates have the following generic sequence: Ac-C1AA1AA2AAnC2AAn+1AAn+2AAn+3C3(TMB)-NH2, where "Ac" refers to N-terminal ation, "-NH2" refers to C-terminal amidation, where "C1, C2, C3" refers to the first, second and third cysteine in the ce, where "AA1" to "AAn" refers to the position of the amino acid (whose nature “AA” is defined by the selections described above), and where "(TMB)" indicates that the e sequence has been cyclised with TBMB or any other suitable reactive scaffold.

Due to the high affinity of Ac34-18(TMB)-NH2 to both human (Ki = 0.17 nM) and rat Kallikrein (IC50 = 1.7 nM), we chose this Bicycle for lead development. Using the same rapid plasma stability profiling assay described above, Ac34-18(TMB)-NH2 had an observability window of about 2 days ds section, method 1), which s to a rat plasma halflife of ~ 2 hrs (as determined quantitatively by LC/MS, see below, table 23 , method 3).

In an effort to identify the lytic recognition site(s) in Ac34-18(TMB)-NH2, the peptide was sampled in 35% rat plasma over time (method 1), and each sample was analysed for the progressive appearance of peptide fragments using MALDI-TOF mass spectrometry. The parent mass of Ac34-18(TMB)-NH2 is 1687 Da. Over time (Figure 7), fragments appear of the masses 1548.6 (M1), 1194.5 (M2), and 1107.2 (M3). From the sequence of 34-18(TMB)-NH2 (Ac-C1S1W2P3A4R5C2L6H7Q8D9L10C3-NH2), it can be calculated that the peak of M1 ponds to Ac34-18(TMB)-NH2 lacking Arg5 (-R5). This appears to be the initial proteolytic event, which is followed by removal of the o acid segment WPAR in Ac34-18(TMB)-NH2 (M2, -WPAR), and finally the entire first loop of Ac34-18(TMB)-NH2 is d (M3, -SWPAR) (Figure 8). From this data, it is evident that Arg5 of Ac34-18(TMB)-NH2 is the main rat plasma protease recognition site that is sible the degradation of the Bicycle. e substitutions and scrambling of first loop: Having identified Arg5 in constituting the recognition site for rat plasma proteases, a campaign of chemical synthesis of Ac34-18(TMB)-NH2 derivatives was undertaken with the aim of identifying candidates with higher plasma proteolytic stability. Crucially, such modifications should not affect the potency against human or rat Kallikrein. An initial exploration regarding the role of the WPAR sequence/pharmacophore (Figure 9, 10) was performed by replacing W2P3 with A2A3 or A2Q3 and by scrambling parts or the entire first loop of the bicycle. Table 8 below shows the ces and the respective ties against Kallikrein.

From these data it is clear that concomitant removal of W2P3 ically reduces g to Kallikrein by a factor of 0, effectively rendering the molecule cologically inert. The importance of the correct sequence of the amino acids is underlined by the four scrambled peptides (Scram2-4), as all of them display a substantial reduction in ty towards Kallikrein (Figure 10). Curiously, all peptides have a roughly identical rat plasma ity profile en 1 to 2 days, method 1), ting that plasma protease recognition relies on the presence of the arginine (Figure 7), and not on its position within the sequence.

Next, five derivatives of Ac34-18(TMB)-NH2 were generated where W2, P3, A4, R5, and C2 were replaced with their respective D-enantiomeric counterparts (Table 9).

From the data it is clear that D-amino acid replacement of A4, R5, and C2 increase peptide stability towards plasma proteases. As Arg5 excision by rat plasma proteases appears to be the first event in peptide degradation, the initial hydrolysis of peptide bonds will occur on the N- and/or C-terminal side of Arg5. It is plausible that replacing the amino acids to either side of Arg5 with their D-enantiomers blocks adjacent peptide bond hydrolysis through steric hindrance. Indeed, this is an effect that has been observed previously (Tugyi et al (2005) PNAS, 102(2), 413–418).

The detrimental effect of D-amino acid substitution on affinities to rein is striking in all cases; losses in potencies range from 300- (D-Arg5) to 45000-fold (D-Trp2). This underlines the importance of the correct three-dimensional y of these sidechains to the Kallikrein bicycle binding pocket. Equally striking is the effect of D-Ala4: here, changing the orientation of a single methyl group (being the Ala side chain) reduces the affinity 7000-fold.

N-methylations: Next, residues in the first loop were systematically replaced with their yl counterparts. N-methylation serves as a straightforward protection of the peptide bond itself; however, due to the absence of the amide hydrogen, addition of steric bulk (the methyl group) and changes in preferred torsional angles, losses in potencies are expected.

Table 10 summarises the data.

N-methylation of amino acids in loop 1 displays an altogether less drastic detrimental effect on potency. In ular, N-methylation of Arg5 still yields a single digit nanomolar binder (20-fold reduction in affinity compared to wildtype peptide), and its rat plasma stability exceeds the assay time (fragmentation of the peptide in the MS was not observable), making this an tive ed lead candidate. As with the o acid substitutions, N-methylation of residues adjacent to Arg5 imparts enhanced stability to the peptide, presumably through steric interference affecting protease-catalysed ysis of peptide bonds N and/or C-terminal to Arg5. Of note, Ser1 can be N-methylated without a icant loss in potency, ting that the integrity of the peptide backbone in this position is not essential for binding.

Arginine substitutions: Given the importance of Arg5 in recognition by rat plasma proteases, a set of arginine analogues were tested in the 34-18(TMB)-NH2 lead. The chemical structures are shown in Figure 11, and the potency versus stability data is shown in Table 11. ngly, all arginine analogues increase the stability of the peptide beyond the assay window time, confirming the importance of the integrity of Arg5 in plasma protease recognition. Increasing (HomoArg) or decreasing the length of the side chain (Agb, Agp) both decrease affinity, however the HomoArg analogue still yields a very good binder (Ki = 2.1 nM), with enhanced stability. Lengthening the amino acid backbone by one methylene group in Arg5 (a so-called beta-amino acid) while retaining the same side chain (b-homoArg5) also yields a binder with ed stability, r at the price of a more significant reduction in affinity (Ki = 8.2 nM). Replacing the aliphatic part of the Arg side chain with a phenyl ring yields a nce stabilised, bulkier and rigidified guanidyl-containing side chain (4GuanPhe). Of all the Arg analogues tested, 4GuanPhe had the greatest affinity (2-fold reduction compared to wildtype), at an enhanced plasma stability. Interestingly, the guanidylphenyl group is urally close to the known small molecule Kallikrein inhibitor benzamidine (Stürzebecher et al (1994), Novel plasma Kallikrein inhibitors of the benzamidine type. Braz J Med Biol Res. 27(8):1929-34; Tang et al (2005), Expression, crystallization, and three-dimensional structure of the tic domain of human plasma Kallikrein. J.Biol.Chem. 280: 41077-89). Furthermore, derivatised Phenylguanidines have been ed as selective inhibitors of another serine protease, uPA (Sperl et al, (4-aminomethyl)phenylguanidine derivatives as nonpeptidic highly selective inhibitors of human urokinase (2000) Proc Natl Acad Sci U S A. 97(10):5113-8.). Thus, Ac34-18(TMB)-NH2 containing 4GuanPhe5 can be viewed as a small molecule inhibitor, whose selectivity is imparted by the nding Bicyclic peptide. This can comprise a principle for other e-based inhibitors, where a known small molecule tor of low selectivity is “grafted” onto a Bicycle in the correct position, leading to a molecule of superior potency and selectivity.

Modification of the Arg guanidyl-group itself, either by methylation (SDMA, NDMA), removal of the positive charge (Cit, where the guanidyl group is replaced by the isosteric but uncharged urea group) or deletion of the Arg altogether (D Arg) has strongly detrimental s on Kallikrein binding potency. Thus, the integrity and presence of the guanidyl group is crucial, while the nature of the sidechain connecting to the guanidyl group or backbone at Arg5 is not. Of note, Arg5 may also be replaced by lysine, however again at reduced affinities (see WPAK peptide).

In summary, data this far indicates that Ac34-18(TMB)-NH2 employing either HomoArg, NMeArg or 4GuanPhe as ne replacements could constitute plasma stability enhanced candidates with high affinities.

Example 5: Improving the potency of a lead candidate through tural modifications and combination with plasma-stability enhancing modifications.

Improving the potency of a given bicyclic ate can be feasibly ed through several isms. These have been partially addressed in Example 4, and can be ten as follows: 1. Incorporating hydrophobic moieties that exploit the hydrophobic effect and lead to lower off rates, such that higher affinities are achieved. 2. Incorporating charged groups that t long-range ionic interactions, leading to faster on rates and to higher affinities (see for example Schreiber et al, Rapid, electrostatically assisted association of proteins (1996), Nature Struct. Biol. 3, 427-31) 3. Incorporating additional constraint into the peptide, by i.e. - aining side chains of amino acids correctly such that loss in entropy is minimal upon target g - Constraining the torsional angles of the backbone such that loss in entropy is l upon target binding - Introducing additional cyclisations in the molecule for identical reasons. (for reviews see ucci et al, Curr. Pharmaceutical Design, (2010), 16, 3185-203, and Nestor et al, Curr. Medicinal Chem (2009), 16, 4399-418).

Tryptophan and hydrophobic analogue substitutions: Initially, a range of hydrophobic amino acids were substituted into the Trp2 site to fy candidates that could replace the oxidation sensitive tryptophan, and to identify candidates that could increase potencies (addressing the first point . The side chains of these amino acids are shown in Figure 12, and affinity data is summarised in Table 12 below.

As expected, none of the modifications increase plasma stability. 2-Naphtylalanine is most closely related to Trp2 and displays a potency slightly weaker than wildtype, making this a good, oxidation-resistant replacement for Trp2. stingly, 3,3-DPA2 has a structure that is very dissimilar to Trp, yet the corresponding peptide retains high potency. This may indicate that the Trp contacting pocket on Kallikrein could be exploited for higher affinity binding by identifying a correctly ed hydrophobic entity.

Proline Analogues: Next, we were interested in determining the role of Pro3 in the WPAR cophore in Ac34-18(TMB)-NH2. 4-hydroxy- or 4-fluoro-trans (L)-proline (HyP3, 4FluoPro3) were chosen for their known property in inducing additional rigidity and helicity on the peptide backbone (Figure 15, Table 13). Additionally, the presence of the yl on HyP probes the solvent accessibility of the proline side chain. Ki's of the respective derivatives were almost identical to that of wildtype, indicating that any effects on the peptide backbone are negligible, but also demonstrating that the side chain is accessible. To elaborate this r, two additional derivatives of 34-18(TMB)-NH2 were tested, which contained bulky extension on the g-carbon of the Pro3 ain (4Phenyl-Pro, 4Benzyl- Pro). The former displayed a striking preservation of potency, while the latter was severely impacted, demonstrating that the Pro side chain is accessible, but limited to ct modifications only. Despite the steric bulk in these modifications, plasma stability was identical to that of wildtype. Thus, these modifications do not improve ivity against other proteases.

To probe the effect of proline ring size on binding, the highly constrained 4-membered Pro analogue ine carboxylic acid (Aze), and the more flexible 6-membered ring (pipecolic acid, Pip) were substituted for Pro3. Ac34-18(TMB)-NH2 Aze3 binds Kallikrein with the highest affinity of all derivatives so far, surpassing that of wildtype by a factor of 3 (Figure 14). There appears to be an inverse relationship between ring size and Ki, which would t that conformational constraint at position 3 of Ac34- 18(TMB)-NH2 is key to a tightly binding molecule.

The flexibility of the proline side chain in tolerating large bulky groups is underlined by the cyclic proline analogues Tic, NorHar and Ind (Figure 13). Particularly for the latter two cases, affinities are still well in the one digit nanomolar range.

Finally, we sought to probe the requirement for the ring structure at Pro3 altogether. To this end, we chose aminoisobutyric acid (Aib, Figure 13, Table 13), which, due to its double methyl substitution at the alpha , has a strong structural effect on the neighbouring amino acids in inducing a or 310 helicity (Toniolo et al , Biopolymers 33, 1061-72; Karle et al (1990), Biochemistry 29, 6747-56). Remarkably, this nonnatural non-cyclic amino acid is well tolerated in place of Pro3, at a Ki of 1.2 nM. Thus, the role of Pro3 in the WPAR pharmacophore is to introduce a constraint onto the peptide backbone. This constraint can be ed by employing a proline analogue with reduced ring size (see Aze3). sely, the proline ring can be replaced vely efficiently with non-cyclic but structure-inducing amino acids, such as Aib.

Miscellaneous Analogues: In table 10, it was shown that Ser1 in loop 1 of Ac34-18(TMB)-NH2 could be N- methylated with very minor impact on potency (0.5 versus 0.17 nM Ki in WT). We sought to determine whether this location tolerated a large double substitution on Ca at position 1. To this end, Ser1 was replaced with Dpg pylglycine) (Figure 15). The affinity of this peptide to Kallikrein is at 1.1 nM, indicating that on 1 is very flexible in accommodating virtually any bulky residue. Thus, this position in loop 1 could be exploited for deliberate inclusion of desirable chemical functionalities or groups, including lising amino acids, radio labels, dye labels, linkers, conjugation sites et cetera.

Several alanine analogues were also tested at position 4. As already seen with the N- methyl and D-alanines (Table 2, 3), Ala4 is highly sensitive to the steric orientation at Ca, or to cation on the backbone itself. Two more derivatives of this class underline this, as elongation of the peptide backbone at Ala4 4) ically reduces affinity (~20 mM). As expected from D-Ala4, Aib4 s ty to almost the same extent (289 nM, Figure 15 and Table 14). Remarkably, extension of the Ala sidechain by one ene (Aba4) appears to enhance the ty to Kallikrein.

Finally, the central cysteine (Cys2) was replaced with a bulkier and more constrained analogue, penicillamine (Pen, Figure 15) in the hope of increasing proteolytic stability due to reduced spatial access to the ourging Arg5 protease recognition point. Indeed, rat plasma ity was slightly enhanced, however potency d significantly, underlining the importance of the full integrity of this structural scaffold-connecting residue.

Combination of plasma stability ing and potency enhancing tural amino acids into a single Bicycle lead Non-natural substitutions in Ac34-18(TMB)-NH2 that retained appreciable potency and maximal rat plasma stability (as determined by method 1) were the Arg5 variants homoarginine (HomoArg5), 4-guanidylphenylalanine (4GuanPhe5) and N-methyl arginine (NMeArg5). Non-natural substitutions in Ac34-18(TMB)-NH2 that increased potency compared to the wildtype peptide was the Pro3 analogue azetidine carboxylic acid (Aze3) and the Ala4 analogue 2-aminobutyric acid (Aba4). Thus, Aze3, Aba4 were combined with the protease stability enhancing HomoArg5, 4GuanPhe5 and NMeArg5 to ine whether this would yield peptide candidates with high plasma stability and increased potency.

Table 15 and 16 present the affinities of the various constructs, together with the plasma stabilities.

Firstly, quantitative determination of rat plasma halflives (4th column, Table 15) of ne analogue containing peptides revealed that Arg5 N-methylation was most potent in protecting the peptide (t1/2 >20 hrs) followed by HomoArg5 and GuanPhe5. The strong protective effect of Arg N-methylation is perhaps not surprising as it directly prevents hydrolysis of the e bond. Upon inclusion of Aze3 in these compounds, the affinity of these peptides could be enhanced in all cases, making Ac-(0618) Aze3 HomoArg5 and Ac-(0618) Aze3 NMeArg5 attractive candidates for further development (Table , Figure 16).

The ty enhancing effect of Aba4 could not be reproduced in the context of Aze3 and any of the ne analogues, as Ki values were higher than those observed without Aba4. Thus, the potency enhancing effects of Aze3 are independent of the type of ne substitution, while those of Aba4 are likely not.

Finally, the activity s rat Kallikrein of these peptides is reduced significantly (Table ). However, these values are relative and not quantitative at this stage as the n preparation of rat Kallikrein is not trivial and contained impurities.

Example 6: Plasma stability enhancement of the Trp-free FPYR Kallikrein e lead and affinity enhancement by Aze3 From the selection output in Examples 1-4 we discovered several sequences resembling Ac34-18(TMB)-NH2 that had a high abundance, but contained altered WPAR motifs.

These were WPSR and FPYR. The latter in particular is interesting as it lacks the ion-sensitive tryptophan.

Bicycles containing WPSR, FPYR, WPYR and FPAR were synthesised and compared against the WPAR parent e (Table 21) As expected, none of the peptides displayed a significantly different plasma ity. The replacement of Trp2 with Phe2 incurs a 40-fold reduction in Ki, underlining the requirement of the bulkier Trp2 side chain. However, this reduction can be compensated by replacing Ala4 with Tyr4 (giving the FPYR motif), so that the affinity increases again to almost that of the wildtype WPAR sequence (Ki = 0.46 nM). Thus, there is a cooperative interplay between the residues at position 2 and position 4 of the Ac34-18(TMB)-NH2 e. Given the high target binding affinity and lack of Trp2 in Ac34-18(TMB)-NH2 Phe2Tyr4, this candidate was investigated for increasing rat plasma half life employing the approach as described in the example above. r, we investigated the interplay between Phe2 and Tyr4 by substituting these residues with non-natural amino acid analogues. tural tutions of Phe2/Tyr4 in Ac34-18(TMB)-NH2 Phe2Tyr4 We performed a non-exhaustive set of syntheses incorporating replacements on Phe2 or Tyr4 in the Ac34-18(TMB)-NH2 Phe2Tyr4 lead. Non-natural amino acids were chosen from the same set as in Figure F6, and affinity data is summarised in Table 22.

Here, substitution with any of the amino acids tested is generally well tolerated, regardless whether the sidechain is a heteroaromatic (3Pal, 4Pal), aromatic and bulky (1Nal, 2Nal, 4,4-BPal) or a cycloaliphatic (Cha) entity. 3Pal is well tolerated at position 2 (Ki = 0.91), which is interesting as Pal contains an ionisable group (which could i.e. be exploited for formulation). It appears, however, that the original Phe2/Tyr4 combination remains most potent.

Stabilisation of Ac34-18(TMB)-NH2 Phe2Tyr4 in rat plasma and effect of azetidine3 substitution 34-18(TMB)-NH2 Phe2Tyr4 was prepared with the homo-arginine, 4- guanidylphenylalanine and N-methylarginine substitutions, in absence and presence of Aze3. HomoArg/4Guanphe are well tolerated, with Ki values almost identical to the parent Phe2Tyr2 peptide (Table 23,), and rat plasma stability was enhanced by a factor of 13 (t1/2 = 12.2 hrs, Table 23). Moreover, IC50 values for rat rein are similar to that of parent, indicating this to be an attractive candidate for in vivo s.

Pro3 to Aze3 substitution in the FPYR context again yielded peptide ates with enhanced affinity, indeed a peptide with a Ki less than 1 nM was generated that would likely have a greater half-life than 20 hrs in rat (Ac-(0618) Phe2 Aze3Tyr4 NMeArg5).

Unless otherwise stated, any methods and materials similar or equivalent to those described herein can be used in the ce or testing of the present invention. Methods, devices, and materials suitable for such uses are described above. All publications cited herein are incorporated herein by reference in their entirety for the purpose of describing and disclosing the ologies, ts, and tools ed in the publications that might be used in connection with the invention.

The term ‘comprising’ as used in this specification and claims means ‘consisting at least in part of’. When interpreting statements in this specification and claims which includes the ‘comprising’, other es besides the features prefaced by this term in each statement can also be present. Related terms such as ‘comprise’ and ‘comprised’ are to be interpreted in similar manner.

In this specification where reference has been made to patent specifications, other external documents, or other s of information, this is generally for the purpose of providing a context for discussing the es of the invention. Unless specifically stated otherwise, reference to such external documents is not to be construed as an admission that such documents, or such sources of information, in any jurisdiction, are prior art, or form part of the common general knowledge in the art.

Tables Bicyclic Number Ki (nM) peptide of amino Orthologous proteases Paralogous proteases acids in loops Human Monkey Rat Human Human Human Human plasma plasma plasma factor Xia thrombin plasmin factor XIIa kalikrein Kallikrein Kallikrein (hfXIa) (hPK) (mPK) (rPK) PK15 6x6 3 4 941 75'000 > 75'000 > 75'000 > 75'000 2A2 5x5 5 6 7 > 75'000 > 75'000 > 75'000 > 75'000 2A10 5x5 18 23 66 75'000 > 75'000 30'000 > 75'000 3B3 3x3 8 11 24 57 > 75'000 > 75'000 > 75'000 3B8 3x3 7 9 12 52 > 75'000 > 75'000 > 75'000 P16 3x3 40 37 53 699 > 75'000 > 75'000 > 75'000 Table 1 Target specificity of bicyclic peptides with different loop lengths. Indicated are Ki values for hPK and different paralogous and orthologous proteases. Ki values are means of at least two measurements. ed region Number Sequence identity with human plasma Kallikrein (hPK) of amino Orthologous proteases gous proteases acids Monkey Rat Human Human Human Human plasma plasma factor XIa thrombin plasmin factor XIIa Kallikrein Kallikrein (hfXIa) (mPK) (rPK) All amino acids 95% 81% 69% 36% 34% 35% Surface amino 4 Å 14 100% 100% 100% 71% 86% 79% acids within a 8 Å 19 100% 100% 83% 79% 74% 63% specific distance of the active site* 12 Å 41 100% 93% 84% 61% 54% 56% Table 2 Sequence homologies around the active site of paralogous and orthologous serine proteases of hPK. (*) Based on the crystal structure of hPK (PDB entry 2ANW) n the bound benzamidine ligand in the S1 pocket was chosen as center.

Table 3: 3x3 peptides Kallikrei Peptid Thrombi Sequence n Av Ki e n Ic50 (nM) ACFKHCRVAC 3B8 0.95 A A C F K H C R V A C A >10000 RVAC 3A3 43.1 A A C F P K C R V A C A 3B9 ACFDPCRVICA A C F D P C R V I C A 90.8 ACFKNCRVNC 3B2 A A C F K N C R V N C A ACFNKCRVNC 06-64 A A C F N K C R V N C A 4.8 RVNC 06-94 A A C F K Q C R V N C A 0.7 >10000 ACFYKCRVNC 06-71 A A C F Y K C R V N C A 15.2 ACFKACRVNC 3B3 0.59 A A C F K A C R V N C A >10000 Table 4: 5x5 peptides Kalli krein Thro Fac Sequence Av mbin tor Ki Ic50 XIIa (nM) ACAWPARC - A C A W P A R C L T V D L C A <0.1* >100 >10 LTVDLCA 01 00 000 - ACRWPARC <0.3* >100 >10 34 VHQDLCA A C R W P A R C V H Q D L C A 00 000 - ACSWPARC 0.4 >100 >10 57 NHQDLCA A C S W P A R C N H Q D L C A 00 000 - ACRWPARC 0.5 >100 >10 59 LTTSLCA A C R W P A R C L T T S L C A 00 000 0.49 A2 ACRWPARC >100 >10 ) T THQNYCA A C R W P A R C T H Q N Y C A 00 000 06 ACTWPARC >100 >10 - THQNWCA A C T W P A R C T H Q N W C A 00 000 - ACFPSHDC A C F P S H D C D G R R M C A 1.27 14 A >100 >10 3 00 000 - ACGGPQNC 2.1 >100 >10 56 RTWTTCA A C G G P Q N C R T W T T C A 00 000 ACNWPYRC >100 >10 7 LHTDLCA A C N W P Y R C L H T D L C A 00 000 - ACSWPYRC 5.8 >100 >10 61 LHQDYCA A C S W P Y R C L H Q D Y C A 00 000 64 ACGVPYRC >100 >10 T THQEMCA A C G V P Y R C T H Q E M C A 00 000 A2 ACTWPARC >10 * TMQNWCA A C T W P A R C T M Q N W C A 000 1277 63 ACADPWAC >100 >10 T LFRRPCA A C A D P W A C L F R R P C A 00 000 1E ACAWPARC >100 >10 6 LTTSLCG A C A W P A R C L T T S L C G 0.16 00 000 2A ACTYPYKCL HQNLCA A C T Y P Y K C L H Q N L C A 4.98 1B ACAWPAKC 1 LTRELCA A C A W P A K C L T R E L C A 8.1 1F ACGGYNNC 7 RAFSYCA A C G G Y N N C R A F S Y C A 2.2 Table 5 - 06-34 – substitutions based on identification of non-critical residues with natural amino acids Peptide Sequence IC50 human PK IC50 rat PK (nM) (nM) 06-34 ACRWPARCVHQDLCA* 0.19 7.38 01 ACSWPARCVHQDLCA 0.15 6.12 0602 ACRWPARCTHQDLCA 0.16 1.09 0603 (01+02) ACSWPARCTHQDLCA 0.082 0.87 0604 ACRWPARCMHQDLCA 0.075 0.56 0605 RCLHQDLCA 0.076 0.62 0617 (01+04) ACSWPARCMHQDLCA 0.073 0.44 0618 (01+05) ACSWPARCLHQDLCA 0.070 0.56 0619 KCLHQDLCA 0.19 4.67 (01+05+R/K) *: Residue numbering is from left to right, where residues 1-5 are in loop 1, and residues 6-10 are in loop 2.

Table 6- 06-34 – substitutions based on identification of non critical residues with N- ated amino acids Ki human PK Peptide Sequence (nM) 06-34 ACRWPARCVHQDLCA 0.128 0603 – Ala1,6 ACAWPARCAHQDLCA 0.147 0603 - N-MeGly1,6 AC N-MeGWPARCN-MeG HQDLCA 24.8 0618 ACSWPARCLHQDLCA 0.040 0618 –N-MeSer1 ACN-MeSWPARCLHQDLCA 0.560 Table 7 Short Supplier name Full chemical name AGTC D-Asp Fmoc-D-Asp(tBu)-OH c NDM-Arg Fmoc-Nwωdimethyl-L-arginine c r Fmoc - Nα - methyl - O - t - butyl - L - serine Anaspec NMe-Trp Fmoc - Nα - methyl - L - tryptophan Anaspec NorHar Fmoc-L-1;2;3;4-tetrahydro-norharmancarboxylic acid Anaspec lPro Fmoc-(2S;4S)phenyl-pyrrolidinecarboxylic acid Iris Biotech Agb Fmoc-L-Agb(Boc)2-OH Iris Biotech Agp Fmoc-L-Agp(Boc)2-OH Iris h β-Ala Fmoc-beta-Ala-OH Iris Biotech Cit Fmoc-Cit-OH Iris Biotech D-Cys Fmoc-D-Cys-OH Iris Biotech β-HArg Fmoc-L-beta-HArg(Pbf)-OH Iris Biotech NMe-Arg Fmoc-L-MeArg(Mtr)-OH Iris Biotech 3Pal Fmoc-L-3Pal-OH Iris Biotech 4Pal Fmoc-L-4Pal-OH Iris h Pen Fmoc-Pen(Trt)-OH Iris Biotech D-Pro Fmoc-D-Pro-OH Iris Biotech Tic Fmoc-L-Tic-OH Iris Biotech D-Trp Fmoc-D-Trp-OH Merck Novabiochem Aib Fmoc-Aib-OH Merck Novabiochem D-Ala Fmoc-D-Ala-OH Merck Novabiochem D-Arg Fmoc-D-Arg(Pbf)-OH Merck Novabiochem 4GuanPhe Fmoc-Phe(bis-Bocguanidino)-OH Merck Novabiochem D-Gln Fmoc-D-Gln(Trt)-OH Merck Novabiochem D-His Fmoc-D-His(Trt)-OH Merck Novabiochem Hyp yp(tBu)-OH Merck Novabiochem D-Leu Fmoc-D-Leu-OH Merck Novabiochem a Fmoc-L-MeAla-OH Merck Novabiochem NMe-Cys Fmoc-N-Me-Cys(Trt)-OH Merck Novabiochem SDMA Fmoc-SDMA(Boc)2-ONa Merck Novabiochem HArg Fmoc-L-HArg(Boc)2-OH h Corporation 4,4-BPAl Fmoc-L-4, 4'-Biphenylalanine Peptech Corporation 3,3-DPA Fmoc-L-3,3-Diphenylalanine Peptech Corporation Dpg Fmoc-Dipropylglycine Peptech Corporation 1NAl Fmoc-LNaphthylalanine Peptech Corporation 2NAl Fmoc-LNaphthylalanine Peptech Corporation Pip Fmoc-L-Pipecolic acid Polypeptide Group Aba Fmoc-Laminobutyric acid Polypeptide Group Aze Fmoc-L-azetidinecarboxylic acid Polypeptide Group 4BenzylPro )-Fmocbenzyl-pyrrolidinecarboxylic acid Polypeptide Group Cha Fmoc-beta-cyclohexyl-L-alanine Polypeptide Group 4FluoPro (2S,4R)-Fmocfluoro-pyrrolidinecarboxylic acid Polypeptide Group Ind Fmoc-L-Indolinecarboxylic acid Table 8 Ki (nM) Observable in Peptide Sequence (human rat plasma, for kallikrein) days Ac-(0618) wildtype Ac-CSWPARCLHQDLC 0.17 2 Ac-(0618) A2A3 Ac-CSAAARCLHQDLC 18545 1 Ac-(0618) A2Q3 Ac-CSAQARCLHQDLC 15840 1 Ac-(0618) Scram1 Ac-CPSAWRCLHQDLC 1091 2 Ac-(0618) Scram2 SPRCLHQDLC 11355 2 Ac-(0618) Scram3 Ac-CAPWSRCLHQDLC 1892 1 Ac-(0618) Scram4 Ac-CWARSPCLHQDLC 67500 1 Table 9 Comparative effects of o acid substitution on potency and rat plasma stability.

Ki (nM) (human Observable Peptide in rat plasma, kallikrein) for days Ac-(0618) wildtype 0.17 2 18) D-Trp2 7558 2 Ac-(0618) D-Pro3 680 3 Ac-(0618) D-Ala4 1203 >10 Ac-(0618) D-Arg5 52 >10 Ac-(0618) D-Cys2 234 >10 Table 10 Comparative s of N-methylation of loop 1 residues and Cys2 on potency and rat plasma stability.

Ki (nM) (human Observable Peptide in rat plasma, kallikrein) for days Ac-(0618) wildtype 0.17 2 Ac-(0618) NMeSer1 0.5 3 Ac-(0618) NMeSer1, NMeAla4 444 >10 Ac-(0618) NMeTrp2 228 5 Ac-(0618) NMeAla4 343 >10 Ac-(0618) NMeArg5 3.5 >10 Ac-(0618) NMeCys2 418 10 Table 11 Comparative effects of arginine analogues in Ac34-18(TMB)-NH2 on potency and stability. Note that the D Arg cation did not y any inhibition up to 100 mM peptide.

Ki (nM) (human Observable Peptide in rat plasma, kallikrein) for days Ac-(0618) wildtype 0.17 2 Ac-(0618) HomoArg5 2.1 >10 Ac-(0618) Agb5 83 >10 18) Agp5 1770 >10 Ac-(0618) bhomoArg5 8.2 >10 Ac-(0618) 4GuanPhe5 0.3 >10 Ac-(0618) SDMA5 1415 >10 Ac-(0618) NDMA5 510 >10 Ac-(0618) Cit5 7860 >10 Ac-(0618) Δ Arg5 >100000 >10 Table 12 ative affinity effects of hydrophobic amino acids substituting Trp2 in Ac34- 18(TMB)-NH2 able Peptide Ki (nM) (human in rat plasma, kallikrein) for days Ac-(0618) wildtype 0.17 2 Ac-(0618) 1NAL2 10.7 2 Ac-(0618) 2NAL2 0.50 2 18) 3Pal2 59 2 Ac-(0618) 4Pal2 72 2 Ac-(0618) Cha2 4.7 2 Ac-(0618) 4,4,BPal2 464 2 Ac-(0618) 3,3-DPA2 1.5 2 Ac-(0618) NorHar2 24 2 Table 13 Comparative affinities obtained for proline derivatives with gamma-carbon substituents, analogues of varying ring sizes, bi/tricyclic derivatives, and constrained amino acids such as Aib Observable Peptide Ki (nM) (human in rat , kallikrein) for days Ac-(0618) wildtype 0.17 2 Ac-(0618) HyP3 0.41 2 Ac-(0618) 4FluoPro3 0.24 2 Ac-(0618) 4Phenyl Pro3 0.58 2 Ac-(0618) 4Benzyl Pro3 191 2 Ac-(0618) Aze3 0.06 2 Ac-(0618) Pip3 0.26 2 Ac-(0618) Tic3 13.51 2 Ac-(0618) NorHar3 2.99 2 Ac-(0618) Ind3 1.35 2 Ac-(0618) Aib3 1.20 2 Table 14 Comparative effects of miscellaneous substitutions of Ser1, Ala4, and Cys2 Observable Peptide Ki (nM) (human in rat plasma, kallikrein) for days Ac-(0618) wildtype 0.17 2 Ac-(0618) Dpg1 1.09 2 Ac-(0618) Aba4 0.07 2 Ac-(0618) β-Ala4 17450 10 Ac-(0618) Aib4 289 7 Ac-(0618) Cys2ToPen2 2162 5 Table 15 Comparative enhancement in potency d by oration of Aze3 in plasmastabilised candidates. 1 ative stabilities estimated according to method 1. 2 The true halflife of peptide stabilities in rat plasma was determined according to method 3. 3 IC50 values are relative, not absolute.

Ki (nM) Observable in t1/2 (hrs) Peptide IC50 (nM) (human rat plasma, in rat rein) for days1 (rat kallikrein) 3 plasma2 Ac-(0618) wildtype 0.17 2.0 2.3 1.7 Ac-(0618) HomoArg5 2.1 >10 10.7 64 Ac-(0618) 4GuanPhe5 0.34 >10 2.8 21 Ac-(0618) NMeArg5 3.5 >10 >20 98 Ac-(0618) Aze3 HomoArg5 0.14 >10 nd nd 18) Aze3 4GuanPhe5 0.17 >10 nd nd Ac-(0618) Aze3 NMeArg5 1.30 >10 nd nd Table 16 Effect on potency upon inclusion of Aba4 in peptides ning Aze3 and the plasmastabilising modifications NMeArg5, HomoArg5, and 4GuanPhe5.

Peptide Ki (nM) (human kallikrein) Ac-(0618) wildtype 0.17 Ac-(0618) Aze3 Aba4 NMeArg5 2.8 Ac-(0618) Aze3 Aba4 HomoArg5 0.9 Ac-(0618) Aze3 Aba4 4GuanPhe5 0.2 Table 17 42 unique Kallikrein binders were identified from selections using a randomised WPAR motif at positions 2, 3, 4 & 5 within the 0603 sequence. The sequences were ranked according to Kallikrein binding and the ve abundance in the total selection outputs was noted.

Sequence Rank Frequency C S W P A R C T H Q D L C 1 24 C S W P S R C T H Q D L C 2 51 C S F P F R C T H Q D L C 3 17 C S W L A R C T H Q D L C 4 8 C S F P Y R C T H Q D L C 5 12 C S F P F K C T H Q D L C 6 4 C S W A A R C T H Q D L C 7 1 C S H P Y R C T H Q D L C 8 2 C S H P F R C T H Q D L C 9 1 C S W P Y R C T H Q D L C 10 3 C R F P F K C T H Q D L C 11 1 C S F P F R C T H Q D L C 12 2 C S L P F R C T H Q D L C 13 3 C S W P F R C T H Q D L C 14 7 C S F P I R C T H Q D L C 15 1 C S L P F K C T H Q D L C 16 1 C S L P F R C T H Q D L C 17 4 C S Y P I R C T H Q D L C 18 2 C S W S A R C T H Q D L C 19 10 C S L P F K C T H Q D L C 20 1 C S Y P F R C T H Q D L C 21 1 C S F P Y K C T H Q D L C 22 1 C S F P W R C T H Q D L C 23 1 C S W H A R C T H Q D L C 24 1 C S L P F R C T H Q D L C 25 2 C S Y P Y R C T H Q D L C 26 2 C S W W A R C T H Q D L C 27 1 C S W P Y K C T H Q D L C 28 1 C S F L Y K C T H Q D L C 29 2 C S L P I R C T H Q D L C 30 1 C S M P Y R C T H Q D L C 31 2 C S I P F K C T H Q D L C 32 1 C S Y P W R C T H Q D L C 33 1 C S F P F W C T H Q D L C 34 1 C S F S Y K C T H Q D L C 35 1 C S W S Y R C T H Q D L C 36 1 C S F M Y K C T H Q D L C 37 1 C S Q V V G C T H Q D L C 38 1 C R W P Y H C T H Q D L C 39 1 C S L F D H C T H Q D L C 40 1 C S H R R W C T H Q D L C 41 1 C S W Q A R C T H Q D L C 42 1 Table 18 A. nce of particular motif in each output. Output sequences were analysed according to the stringency of selection. The % of a particular motif in the output from a particular stringency selection was calculated. Also see figure 17.

A Relative abundance of species from different selection encies WPSR WPAR WSAR WPFR WPYR FPYR FPFR High 34 17 3 7 3 7 7 Stringency 31 19 3 3 3 3 19 41 13 3 0 0 3 9 26 22 7 0 4 7 4 23 6 6 6 0 16 13 Low 11 4 11 4 0 4 11 B Target binding assay signal WPSR WPAR WSAR WPFR WPYR FPYR FPFR 178 264 57 67 80 127 171 Binding assay signal 300 Table 19 Top 50 Kallikrein binders ning a WPAR motif. The WPAR domain was fixed in a e library with positions 1, 6, 7, 8, 9 & 10 randomised. Kallikrein selection output sequences were isolated and assayed for Kallikrein binding. The sequences were ranked according to Kallikrein binding. The sequence of e 0603 was isolated from selection and is highlighted in red. Trends are visible in the second loop of WPAR- containing Kallikrein-binding peptides.

Rank Sequence Binding assay signal 1 C N W P A R C T H Q D L C 116 2 C S W P A R C T H Q D L C 110 3 C H W P A R C T H Q D L C 110 4 C P W P A R C T H Q D L C 107 C S W P A R C T H A D L C 100 6 C S W P A R C T H D D L C 93 7 C A W P A R C T H T D L C 92 8 C Q W P A R C L H T D L C 91 9 C L W P A R C T H Q D L C 90 C T W P A R C T H T D L C 88 11 C H W P A R C T H Q E L C 85 12 C A W P A R C L H D D L C 84 13 C S W P A R C L H T D L C 83 14 C A W P A R C T H V D L C 82 C A W P A R C T H T D F C 80 16 C M W P A R C M H Q D L C 79 17 C A W P A R C T H A D L C 79 18 C Q W P A R C M H Q D M C 75 19 C Q W P A R C T H S D L C 74 C L W P A R C T H A D L C 74 21 C R W P A R C T H Q D L C 73 22 C Q W P A R C M H Q E L C 73 23 C T W P A R C L H Q D L C 73 24 C S W P A R C T H S H L C 72 C V W P A R C T H Q D L C 71 26 C T W P A R C T H A D L C 71 27 C H W P A R C M H Q D L C 71 28 C P W P A R C T H T D L C 70 29 C A W P A R C T H Y D L C 70 C P W P A R C T H Q N L C 69 31 C S W P A R C T H T E L C 69 32 C A W P A R C M H D D L C 69 33 C S W P A R C L H T E L C 68 34 C S W P A R C I H Q D L C 68 C T W P A R C T H T D M C 67 36 C A W P A R C T H T H L C 66 37 C A W P A R C L H A D M C 66 38 C A W P A R C L H Q D W C 63 39 C D W P A R C M H Q E F C 63 40 C A W P A R C T H Q T M C 61 41 C T W P A R C L H Q H M C 61 42 C S W P A R C V H Q D M C 61 43 C E W P A R C L H T D L C 60 44 C L W P A R C L T T E L C 59 45 C S W P A R C T H A E M C 59 46 C R W P A R C T H T D L C 59 47 C T W P A R C T H Q A F C 59 48 C S W P A R C T H S D L C 59 49 C S W P A R C L H D D L C 59 50 C P W P A R C L H T D L C 58 Table 20 The output sequences from Kallikrein selections with fixed-WPAR in the 1st loop, and their associated Kallikrein binding assay signals, were grouped according to their derivation from ‘THxxL’ motif (A), or ‘xHxDL’ motif (B). The e binding assay signal for all members of a given group was calculated. Groups containing precisely the given motif are highlighted green; examples of groups with either one more or one less change away from the motif are also shown.

A B Groups based on xHxDL motif Groups based on THxxL motif Motif Bi nding Moti f Bindi ng Motif Binding Motif Binding Motif Bindi ng Moti f Binding THxDL 78.5 xHxDL 70.9 xHxxL 56.1 THxDL 78.5 THxxL 61.0 MHxxx 56.5 MHxDL 66.7 xHxDM 52.8 xHxxM 46.3 MHxDL 66.7 MHxxL 55.0 THxxx 53.4 LHxDL 58.0 xHxEL 51.7 xHxxF 43.8 LHxDL 58.0 LHxxL 47.9 LHxxx 44.4 THxEL 55.9 xHxHM 46.4 xHxxW 33.2 THxEL 55.9 THxxM 46.3 LTxxx 33.8 THxHL 52.5 xHxNL 46.2 xTxxL 31.1 THxHL 52.5 LHxxM 39.0 LHxEL 52.3 xHxDW 44.2 xHxxE 19.3 LHxEL 52.3 LTxxL 38.4 THxNL 51.9 xHxHL 44.2 THxNL 51.9 THxDW 44.9 xHxFL 40.4 THxFL 40.4 LHxDW 44.7 xTxEL 39.9 LTxEL 39.9 MHxEL 42.4 xTxDL 39.4 LTxDL 39.4 LTxEL 39.9 xHxQL 39.0 LHxSL 37.4 LTxDL 39.4 xHxAL 35.5 THxAL 35.5 LHxQL 39.0 xHxSL 35.0 LHxHL 34.0 LHxDM 38.0 xTxSL 33.1 LTxSL 33.1 LHxSL 37.4 xHxYL 29.2 THxSL 28.9 THxAL 35.5 LHxHL 34.0 LTxSL 33.1 THxSL 28.9 Table 21 Affinities and stabilities of WPAR motif variants.

Observable Peptide Ki (nM) (human in rat , kallikrein) for days Ac-(0618) WPAR 0.17 2 Ac-(0618) FPAR 6.28 2 18) WPYR 0.41 2 Ac-(0618) WPSR 0.44 2 Ac-(0618) FPYR 0.46 2 Table 22 Effect of substitutions on Phe2/Tyr4 with hydrophobic analogues Peptide Ki (nM) (human kallikrein) Ac-(0618) Phe2 Tyr4 0.46 Ac-(0618) Phe2 Cha4 0.91 Ac-(0618) Phe2 3Pal4 2.57 Ac-(0618) Phe2 4Pal4 2.20 Ac-(0618) Phe2 1Nal4 13.5 Ac-(0618) Phe2 2Nal4 7.27 Ac-(0618) Phe2 4,4-BPal4 10.5 Ac-(0618) 3Pal2 Tyr4 0.91 18) 4Pal2 Tyr4 3.56 Ac-(0618) Cha2 Tyr4 1.87 Table 23 Summary of the effect of Arg5 substitutions and Aze3 on Ac34-18(TMB)-NH2 Phe2Tyr4. 1 Comparative stabilities estimated according to method 1. 2 The true halflife of e stabilities in rat plasma was determined according to method 3. 3 IC50 values are relative, not absolute.

IC50 Ki (nM) Observable t1/2 (hrs) in Peptide (human in rat plasma, rat (nM) rein) for days1 plasma2 (rat kallikrein)3 Ac-(0618) Phe2 Tyr4 0.46 2 0.9 13.8 Ac-(0618) Phe2 Tyr4 HomoArg5 0.77 >10 12.2 19.7 Ac-(0618) Phe2 Tyr4 4GuanPhe5 0.40 >10 5.0 2.5 Ac-(0618) Phe2 Tyr4 NMeArg5 3.56 >10 >20 60.2 Ac-(0618) Phe2 Aze3 Tyr4 HomoArg5 0.12 nd nd nd 18) Phe2 Aze3Tyr4 4GuanPhe5 0.36 nd nd nd Ac-(0618) Phe2 Aze3Tyr4 NMeArg5 0.97 nd nd nd

Claims (29)

Claims

1. A peptide ligand specific for human Kallikrein comprising a polypeptide comprising at least three reactive groups, separated by at least two loop sequences, and a molecular scaffold which forms covalent bonds with the reactive groups of the polypeptide such that at least two polypeptide loops are formed on the lar ld, n the loops of the e ligand comprise three, four or five, but less than six, amino acids.

2. A e ligand ing to claim 1, wherein the loops of the peptide ligand comprise three amino acids and the polypeptide has the consensus sequence GrFxxGrRVxGr, wherein Gr is a reactive group.

3 A peptide ligand according to claim 2, which comprises one of the polypeptides set forth in Table 3.

4. A peptide ligand according to claim 1, wherein the loops of the peptide ligand comprise five amino acids and a first loop comprises the consensus sequence GrGGxxNGr, wherein Gr is a reactive group.

5. A peptide ligand according to claim 4, wherein two adjacent loops comprise the consensus sequence GrGGxxNGrRxxxxGr.

6. A peptide ligand according to claim 4 or claim 5, which comprises one of the es set forth in Table 4.

7. A peptide ligand according to claim 1, wherein the loops of the peptide ligand comprise five amino acids and a first loop ses the motif GrxW/FPxK/RGr, wherein Gr is a reactive group.

8. A e ligand according to claim 7, further comprising a second loop which comprises the motif GrT/LHQ/TxLGr.

9. A peptide ligand according to claim 1, wherein the loops of the peptide ligand comprise five amino acids and a first loop comprises the motif GrxHxDLGr, wherein Gr is a reactive group.

10. A peptide ligand according to claim 1, wherein the loops of the peptide ligand comprise five amino acids and a first loop comprises the motif GrTHxxLGr, wherein Gr is a reactive group.

11. A peptide ligand ing to any one of claims 8 to 10, wherein two adjacent loops comprise the motif GrxW/FPxK/RGrT/LHQ/TDLGr.

12. A peptide ligand according to any one of claims 7 to 11, which comprises one of the polypeptides set forth in Table 4, Table 5 or Table 6.

13. A peptide ligand according to any one of claims 7 to 11, wherein the first loop comprises the sequence GrxWPARGr.

14. A peptide ligand according to any one of claims 7 to 11, wherein the first loop ses the sequence GrxWPSRGr.

15. A peptide ligand according to any one of claims 7 to 11, wherein the first loop comprises the sequence GrxFPFRGr.

16. A peptide ligand according to any one of claims 7 to 11, wherein the first loop comprises the sequence GrxFPYRGr.

17. A peptide ligand according to any one of claims 13 to 16, wherein x is S or R.

18. A peptide ligand according to any ing claim, wherein the reactive group is cysteine.

19. A peptide ligand according to any preceding claim which ses one or more non-natural amino acid substituents and is resistant to protease degradation.

20. A peptide ligand according to claim 19, wherein the modified amino acid is selected from N-methyl Arginine, homo-arginine, hydroxyproline, and guanidylphenylalanine , and azetidine carboxylic acid.

21. A e ligand according to claim 20, n the polypeptide comprises a first loop which comprises the motif GrxWPARGr, wherein Pro is replaced with azetidine carboxylic acid; and/or R is replaced with N-methyl arginine or homoarginine or guanidylphenylalanine.

22. A peptide ligand according to claim 20, wherein the ptide ses a first loop which comprises the motif GrxFPYRGr, wherein Pro is replaced with azetidine carboxylic acid; and/or R is replaced with N-methyl neor homoarginine, or guanidylphenylalanine.

23. A peptide ligand according to claim 22, wherein the polypeptide comprises a first loop which comprises the motif RGr, wherein Pro is replaced with azetidine carboxylic acid; and/or R is replaced with homoarginine.

24. A peptide ligand according to claim 23, wherein the polypeptide comprises a first loop which comprises the motif GrxFPYRGr, wherein Pro is replaced with azetidine carboxylic acid and R is replaced with homoarginine.

25. A e ligand according to any one of claims 21 to 24, wherein x is S.

26. A peptide ligand according to any one of the preceding claims, which onally comprises an N-terminal alanine residue.

27. A peptide ligand according to any one of ing claims, which comprises N- terminal acetylation and C-terminal amidation.

28. A peptide ligand according to claim 1 or claim 19, which comprises one of the polypeptides set forth in Table 9, Table 10, Table 11, Table 12, Table 13, Table 15, Table 16, Table 17, Table 19, Table 21, Table 22 or Table 23.

29. A peptide ligand as d in any one of claims 1 to 28, substantially as herein described and with reference thereof.