HK1132177B - Solution phase biopanning method using engineered decoy proteins - Google Patents

Description

Liquid phase biopanning process using engineered trap proteins

This application claims priority to U.S. provisional application serial nos. 60/565,674 and 60/565,633 (application on 8/4/2004), the contents of which are all incorporated herein by reference in their entirety. The application filed hereby contains a sequence listing on a computer readable disk, the contents of which are incorporated herein by reference.

Background

Technical Field

The present invention relates to methods for selecting antibodies that bind to selected epitopes by using combinatorial antibody phage display libraries. The invention also relates to antibodies produced by such methods.

Prior Art

Drug development efforts in the post-genomic era have now focused on finding functional methods that specifically block key proteins that have been previously identified by analyzing the expression levels of mRNA under disease conditions, such as by microarray technology. Proteomics is a new discipline that involves looking at the way protein molecules interact in a synergistic pathway and as binding partners. The structure-activity relationship of proteins includes conventional domain mapping and identification of functionally relevant three-dimensional (3D) conformational information. Methods for understanding three-dimensional (3D) structural information of proteins have also been routinely practiced. For example, the NCBI maintains a publicly used tool, named VAST, structural similarity search service. It compares the three-dimensional structure of newly defined proteins to a Macromolecular Modeling Database (MMDB) and a Protein Database (PDB).

Phage display technology is an in vitro selection technique in which a peptide or protein encoded by a polynucleotide sequence is genetically fused to a coat protein of a phage, such that the fusion protein is displayed on the outside of the phage virion and the DNA encoding the fusion protein is located within the virion. The physical association between the displayed protein and its encoding DNA allows the screening of a large number of variant proteins, each associated with its corresponding DNA sequence, by a simple in vitro selection method called biopanning.

Phage, ribosome, yeast and bacterial display libraries are tools for screening large numbers of proteins or peptides. The ribosome display method is the translation of mRNA into its corresponding protein, while keeping the protein attached to the RNA. The nucleotide coding sequence was obtained by RT-PCR (Mattheakis, L.C.et al.1994.Proc.Natl.Acad.Sci.USA 91, 9022). Yeast display is based on the construction of membrane-associated a-lectin yeast binding receptor fusion proteins, aga1 and aga2, and a partial association (mate type) system (Broder, et al 1997.Nature Biotechnology, 15: 553-7). Bacterial display is based on the construction of fusion proteins of a target protein with bacterial export proteins (associated with cell membranes or cell walls) (Chen and Georgiou.2002.Biotechnol Bioeng, 79: 496-503).

In contrast to hybridoma technology, phage and other antibody display technologies offer the possibility of performing in vitro operational selections for antigens without potential limitations on the host effect of the antigen, and vice versa. One particular advantage of in vitro selection is the ability to select for antibodies that bind to the diversity site of the target protein.

Although phage libraries simplify the recovery of genetic material associated with functional attributes, strategies for multi-stage panning are still required to purify optimal candidate molecules from libraries. On the other hand, given the structural information of the functional domains of polypeptide ligands, there is a need for methods that allow the selection of antibodies or other binding partners (e.g., peptides or proteins) that bind to specific domains of the ligand. Domain or epitope directed panning has become a routine method of selecting antibodies that bind to a target protein. This selection is achieved primarily by stepwise selection of antibodies using a variety of known methods, such as selective panning, de-selective panning, ligand capture, subtractive panning or guided selection (pathfinderelection) (Hoogenboom, h.r. et al (2000) supra).

In subtractive panning, target molecules with overlapping regions (overlapping) but not identical binding sites can be used to cull out unwanted binders (bins). This method has been used to identify binders that bind to unknown antigens, for example, to eliminate binders that bind to cancer cells in normal cells. Natural proteins containing some common domains or structures may also be selected for sequential or competitive selection to obtain antibodies that bind to the common or exclusive sites in the relevant antigen. Typical such studies use native proteins associated with biochemical motility or with H-ras proteins (Horn, I.R.et al.1999, FEBS Lett.463: 115-120).

Ligand capture directed panning, similar to ELISA sandwich methods, uses immobilized antibodies of unrelated and non-adjacent epitopes to capture the target ligand and present it with a preferred binding surface for phage panning (US 6376170). Other investigators used competitive antibodies to selectively mask antigens on non-target domains (Tsui, P.et al.2002.J.Immunol. meth.263: 123-132). The guided selection method uses monoclonal and polyclonal antibodies, and also uses natural ligands coupled directly or indirectly to horseradish peroxidase (HRP). These molecules catalyze biotinylation of phage that bind tightly to the target antigen in the presence of tyramine biotin, allowing the specific recovery of "labeled" phage from all phage using streptavidin. This method selectively recovers phages bound to the target molecule itself or in its vicinity (Osborn, J.K.et al.1998. Immunotechnol.3: 293-302). The use of monoclonal antibodies directed to bind to alternating sites is also known as "epitope walking" (Osborn, J.K.et al 1998, supra).

These methods have the disadvantage that a lot of work has to be done to obtain and identify non-target binding partners before the work to obtain binding partners that bind to the domain of interest has to be done, and that no targeting of specific epitopes has been performed. The present invention provides novel methods for obtaining antibodies or ligand binding partners that bind to selected epitopes by using hybrid competitive proteins in the panning step.

Summary of The Invention

The present invention provides novel methods for selecting ligand-binding partners that bind to a predetermined domain by using engineered decoy ligands in the panning step. The decoy ligand is designed to differ from the target protein only in a predetermined domain which constitutes a recognized binding site. The decoy protein can be designed based on actually measured structural data (e.g., X-ray crystal diffraction data) or computational data (computer calculations from a three-dimensional structural model). When structural information is obtained, the design of the trap protein is simple. If there is no structural information or the information is incomplete, the sequence can be modified in key regions based on natural variant proteins (e.g., species homologous proteins) to produce decoys.

The invention further relates to nucleic acids encoding decoy proteins, which can be used to express the decoy proteins in a host cell or organism.

If the host cell is transfected, the decoy protein may be expressed on the cell surface, or the decoy protein may be a secreted free protein of the host and may be recovered from the cell culture medium. The trap protein may be purified or used in a heterologous environment (e.g., on the surface of a cell). In the biopanning step, the molar ratio of target molecule to decoy protein is maintained to eliminate non-specific and low affinity binders, and only the binders bound to the target molecule are recovered. In this method, protein binding partners fused to their cognate genetic material are selected from the library based on their ability to specifically bind to the target protein binding site (which is altered in the decoy protein and therefore known to interact with the domain of interest).

In another aspect, the methods of the invention for selecting antibodies that bind a predetermined epitope can be used to transfer the desired properties of a therapeutic target antibody or ligand-binding body that has proven effective in one species (e.g., an animal model) directly to a similar biotherapeutic agent to effectively treat the other species. The human biopharmaceuticals can be readily converted into homologues thereof for use in treating mammals of other interest, such as cattle, pigs, poultry, dogs, cats or other agriculturally important animals or domestic animals, with a similar mechanism of action in that animal genus or species. In one embodiment of the invention, the method is used to select antibodies that interact with homologous proteins having the same three-dimensional domain, which is used as a reference antibody. This is particularly useful where a monoclonal antibody is known to bind to a particular region of a human antigen and it is desired to recover a binding partner (e.g., an antibody to a human antigen) that binds to the same epitope. In another embodiment, this method is effective when there is an antibody that binds to an epitope of a human antigen, and a substitute antibody that acts on the same epitope of the corresponding protein in other animals of interest (e.g., mice) is required for the study. In this manner, an anti-mouse antibody with properties similar to those of the original anti-human antibody can be obtained.

Thus, in one aspect the present invention provides a method of selecting blocking polypeptide ligand binding partners from a library wherein the specific domain of the ligand to be bound is predetermined, the method comprising the steps of: a) determining a functional domain of the protein to be blocked, b) analyzing one or more species homologues or functional homologues of said ligand for common structural features between said ligand, c) constructing a decoy incorporating said common structural features of the ligand and the selected homologues, wherein said decoy has common structural features of regions other than the functional domain to be blocked d) using said decoy (in an amount greater than the ligand binding partner) to select a binding body which preferentially binds to the functional domain to be blocked.

In another aspect, the present invention provides a method of identifying a polypeptide that binds to a predetermined epitope of a target protein, comprising the steps of: a) creating a library of phage particles expressing polypeptides on their surface, b) preparing a decoy protein having a change in amino acid sequence corresponding to a predetermined epitope of the target protein, c) incubating the library of phage particles with the target protein to select phage having surface polypeptides bound to the target protein, d) adding a molar excess of the decoy protein as a competitor to negatively select for phage particles specific for the predetermined epitope, e) isolating phage bound to the target protein from phage bound to the decoy protein, f) recovering phage particles bound to the target protein.

Brief Description of Drawings

FIG. 1 is a schematic representation of the process of the present invention.

Figure 2 is the CDR sequence and framework arrangement of important candidate fabs that bind mTF.

FIG. 3 shows the concentration dependence of the binding of selected antibodies of the method of the invention to the target protein (mTF, solid line) and the decoy protein (hu/mTF (two amino acids in the predetermined epitope changed), dotted line).

FIG. 4 is a graph of concentration versus fluorescence units for two selected Fab's that bind mTF.

FIG. 5 is a multiple sequence alignment of IL-13 proteins from different species.

FIGS. 6A &6B are energy and area size (area) from crystallographic data for hIL-4.

FIGS. 7A &7B energy and area sizes of hIL-13 calculated from IL-4 crystallographic data.

Abbreviations

Abs: antibodies, polyclonal or monoclonal

bFGF: basic fibroblast growth factor

GM-CSF: granulocyte-macrophage colony stimulating factor

IL: interleukins

And Mab: monoclonal antibodies

TF: tissue factor

F IIV: factor IIV (Inactive)

F IIVa: factor IIVa (activation)

FX: factor X (inactive)

Fxa: factor Xa (activated)

Detailed Description

Concept definition

"antibody" refers to an immunoglobulin or binding fragment derived from an immunoglobulin. Although not all immunoglobulins bind to antigens, it is known that antibody fragments can bind to antigens, target polypeptides or proteins, and other molecules. Thus, as used herein, an "antigen-binding fragment" includes, but is not limited to: (i) fab fragments, which consist of the variable (V) regions of the heavy (H) and light (L) chains of the antibody, together with the respective constant (C) regions (VL-CL and VH-CH1 regions); (ii) fd fragment, which consists of VH and CH1 regions; (iii) fv fragments, which consist of the VL and VH regions of a single chain antibody; (iv) dAb fragments (Ward, E.S.et al, Nature 341: 544-546(1989)) which consist of a VH region; (v) an isolated CDR region; (vi) f (ab')₂A fragment; (vii) a single chain Fv molecule (scFV), wherein the VH and VL regions are linked by a polypeptide linker, such that the two regions combine to form an antigen binding site; (viii) bispecific single chain Fv dimers; (ix) recombinant and fusion proteins comprising the above proteins, including but not limited to diabodies, multivalent or multispecific fragments, or other engineered constructs capable of binding a target polypeptide and comprising an immunoglobulin-derived fragment.

"chimeras" or "chimeric proteins" refer to proteins that contain residues or domains of one or more homologous proteins. For example, chimeric antibodies comprising a variable region typically derived from a murine mAb fused to a human immunoglobulin constant region.

A "decoy" or "decoy protein" is a polypeptide designed to incorporate predetermined or engineered domains for negative or positive selection of a target ligand binding partner from a library of potential target binding partners.

An "epitope" is defined as a three-dimensional region of a target ligand that represents a structural unit that binds to a single chain antibody. Epitopes usually consist of chemically active surface groups of molecules such as amino acids or sugar side chains and usually have specific three-dimensional structural properties and specific charge properties. Conformational and non-conformational epitopes are distinguished in that binding to the former is cleaved in an inactivating solvent, while the latter is not. The epitope may comprise or be present in a functional unit or a structural characteristic protein domain as described above (e.g., a receptor binding domain or a fibronectin-like domain). Thus, when an epitope is a functional domain of a protein and binds to a selected binding partner, it causes the desired modification of the function of the target ligand, which modification is antagonistic or agonistic to the function of the target ligand.

"substituted" means having similar biological function. Alternative antibodies perform similar functions by agonizing or antagonizing the activity of the target ligand in an environment or animal that is heterologous to the original antibody.

"human" or any other species of antibody (e.g., human antibodies), including those antibodies that contain variable regions, or variable and constant regions, derived from human or other germline immunoglobulin sequences. The antibodies of the invention may comprise amino acid residues not encoded by immunoglobulin sequences (such as, but not limited to, mutations introduced by random or site-directed mutagenesis in vitro, or somatic mutations performed in vivo). Thus, as used herein, a "human antibody" refers to each portion of an antibody protein (e.g., CDR, framework, C)_L、C_HZone (e.g. C)_H1、C_H2、C_H3、C_H4) Hinge region (V)_L、V_H) All substantially similar to human germline antibodies. Human antibodies have been classified according to their amino acid sequence similarity, see e.g., http:// peple. cryst. bb k. ac. uk/-ubcg07 s/. Thus, using a sequence similarity search tool, antibodies with similar linear sequences can be selected as templates to make "human antibodies". Murine germline sequences are known and can be used in a similar manner. After collection and indexing of germline immunoglobulin-related data for other species, such sequences can also be used in a similar manner to prepare non-human antibodies of the invention from phage display libraries or other collections of antigen-binding fragments by methods known in the art.

In one aspect, the invention relates to phage display and the use of combinatorial peptide libraries. Phage display and combinatorial peptide libraries have been developed as powerful and suitable tools for exploring interactions between peptides and proteins. Phage libraries can be prepared by insertion of random oligonucleotide libraries or by insertion of polynucleotide libraries comprising sequences of interest (e.g., B cells from an immunized animal or human) (Smith, G.P.1985.science 228: 1315-1317). Antibody phage libraries contain pairs of variable regions of the heavy (H) and light (L) chains in one phage, allowing the expression of single chain Fv or Fab fragments (Hoogenboom, et al, 2000.immunol. today 21(8) 371-8). The diversity of the phagemid library can be manipulated to improve and/or alter the immunospecificity of the monoclonal antibodies in the library to generate and subsequently identify additional and desired human monoclonal antibodies. For example, genes encoding heavy (H) and light (L) chain immunoglobulin molecules can be randomly mixed (shuffled) to make new HL pairs during assembly of the immunoglobulin molecules. Alternatively, Complementarity Determining Regions (CDRs) of immunoglobulin variable regions of genes encoding heavy (H) and/or light (L) chains can be mutagenized and subsequently screened for the desired avidity and neutralizing (neutralizing) capabilities. Antibody libraries can also be synthesized by selecting one or more human framework sequences and introducing a set of CDR cassettes derived from all components of a human antibody or by designing variable regions (Kretzschmar and von Ruden 2000, Current Opinion in Biotechnology, 13: 598-. The positions where diversity is caused are not limited to CDR regions but may also include architectural segments of variable regions.

Other libraries useful in the present invention include phage display libraries from non-human animal antibody or engineered antibody libraries. Examples of the former include the use of immunoglobulins from camelid species which naturally lack a light chain (Hamers-Casterman et al, 1993, Nature 363: 446-; Gahroudi et al, 1997, FEBS Lett), and examples of the latter are for example the use of single domain antibodies with binding capacity from either a heavy or light chain (see e.g.U.S. Pat. No. 6248516).

In addition, various phage or other display systems, such as ribosomes, yeast, bacteria or animal cells, can be used in combination with peptide (or antibody) phage libraries in a variety of ways to conduct biological exploration or to search for novel drugs or drug targets. For example, a peptide phage display library can be used to search an antibody phage display library. The substrate phage display and substrate subtraction methods can be used in combination to search for differences in specificity between closely related enzymes by exclusion, and this information can be used to prepare highly selective inhibitors (Ke, S-H, et al 1997 J.biol. chem.272 (26): 16603-16609).

Binding between ligands and receptors (e.g., antibodies and antigens) relies on hydrogen bonding, hydrophobic interactions, electrostatic forces, and van der waals forces. These binding interactions are all non-covalent, weak bonds, but it is known that the binding between antigen and antibody is one of the strongest natural binding interactions. Antigens may also be multivalent like antibodies, either by multiple copies of the same epitope or by the presence of multiple epitopes recognized by multiple antibodies. Interactions involving polyvalency can result in stronger complexes, but polyvalency can also cause steric hindrance and thus reduce the likelihood of binding. All antigen-antibody binding interactions are reversible, but follow the basic laws of thermodynamics of any reversible biomolecular interaction:

wherein K_AIs the affinity constant, Ab and Ag are the molar concentrations of unoccupied binding sites on the antibody or antigen, respectively, and Ab-Ag is the molar concentration of the antibody-antigen complex. The forward reaction is known as the "on rate" and the dissociation or reverse reaction as the "off rate".

In order for the interaction between antigen and antibody to proceed efficiently, the epitope must be easy to bind. Because the antigenic molecule is present in space, the epitope recognized by the antibody may depend on the specific three-dimensional conformation of the antigen (e.g., a particular site formed by the interaction of two native protein subunits), or the epitope may correspond to a simple primary sequence region. Such epitopes are called "conformational" and "linear", respectively.

Method of the invention

We devised a method for isolating antibodies or other binding partners that bind to a predetermined epitope by targeted selection of antibodies displayed on phage using engineered competing proteins (fig. 1). The method relies on structural information of the target protein, which is used to design suitable decoy proteins. This decoy protein was used as a competitor in antibody phage display to isolate antibodies with the desired epitope-specificity (figure 1).

The binding specificity of antibodies obtained by traditional methods of immunizing animals is determined by the combination of the animal's immune system and the antigenic proteins. Antibodies from immunization therefore often interact with "immunodominant" epitopes, which differ from the desired target epitope. Existing selection methods using antibody phage display libraries do not allow for accurate targeted selection of epitopes of interest. The method disclosed herein has the advantage of enabling very accurate and efficient targeted selection of antibodies specific for a target epitope.

Methods of selecting antibodies that bind a predetermined epitope can be used to directly transform the desired properties of a therapeutic target antibody or ligand-binding body (biotherapeutic) that proves effective in one species (e.g., animal models) to a similar biotherapeutic to effectively treat other species. The human biopharmaceuticals can be conveniently converted into similar drugs and thus are effective in treating other mammals, such as cattle, swine, poultry, dogs, cats or other agriculturally important domestic animals, or rare animals or species endangered.

Of the 15 diseases most commonly affecting companion (companion) animals (dogs, cats and horses), many are associated with hormone secretion, including diabetes in canines and felines, thyroid disorders, hypothyroidism in canines, hyperthyroidism in felines, addison's disease and cushing's disease in canines. Other common conditions in companion and other animals include osteoarthritis and various cancers. It is therefore possible to convert effective human biotherapeutics (such as anti-cancer and anti-inflammatory antibody therapeutics) into other species-specific similar drugs. For example, infliximab drugs (REMICADE), which bind to specific epitopes on human α -TNF, can be converted using the methods of the invention into therapeutically effective drugs for effective use in treating disorders mediated by α -TNF dysregulation common in companion animals.

Selection of target site binding

Each lymphocyte produces an antibody that is not specific for an antigen, but rather for an epitope. An antigen is a part of a foreign cell, particle, protein or molecule that is recognized by the immune system and that serves as a target for antibodies and/or cytotoxic T cells, and an epitope is also a binding site corresponding to an antigenic determinant on a protein. Polypeptides, lipids, nucleic acids and many others may also be used as antigens. Some small molecule substances, referred to as "haptens," may also undergo immune reactions against large "carrier proteins" (such as bovine serum albumin or blue protein, or other artificial substrates) if chemically coupled to them. Haptens can be a variety of molecules, such as drugs, simple carbohydrates, amino acids, small peptides, phospholipids, or triglycerides. Antigens that elicit a strong immune response are referred to as "strong immunogens". It has been empirically demonstrated that antigenic determinants causing a clonal immune response can be as small as 1-8 amino acids or 1-6 monosaccharides. In practice, the epitope recognized by an immunoglobulin (monoclonal antibody) from a clone may contain larger non-contiguous (non-contiguous) sequences on the protein surface.

When an epitope is present within a functional region of a protein, the effect of binding of the antibody to the protein will neutralize the function imparted to the protein by the structural features of that region. The entire view has been shown to be the basis for therapeutic action of monoclonal antibodies. The ability to reproducibly select antibodies or other binding bodies that specifically bind to an antigen or protein domain may therefore represent an advance in the field of protein therapy.

Antibody epitope mapping (Antibody epitope mapping) is a method of identifying functional domains. Resolution with which high or low epitopes can be mapped, depending on the subject. Low resolution mapping involves the interaction of a set of monoclonal antibodies with the surface sequence of the native protein. The emphasis is on covering the entire target surface and identifying which sequence is an important functional sequence. Unlike the dominant candidate mabs, the antibodies used in epitope mapping can be low avidity and should include neutralizing and non-neutralizing mabs, while precise determination of epitopes is not always required in this approach. Once an epitope is identified at a particular desired resolution, other binding agents that bind to this region, such as human antibodies, can be identified using competition assays with antibodies that perform the desired function, usually neutralizing the function of the target protein, for example, human antibodies can compete with murine antibodies for binding to the human target protein.

Epitopes can be identified by other methods using sets of antibodies that bind to the target protein. For example, the antigen can be digested with a protease and binding of the fragments resulting from the digestion to the antibody determined using ELISA or mass spectrometry. The antigen-antibody complex can be digested with a protease and the protein digested fragments analyzed by mass spectrometry. In this case, the coverage of the proteolytic site by the antibody identifies the epitope.

There are other methods of identifying epitopes of antibodies. Peptides corresponding to overlapping fragments of the entire antigen sequence can be synthesized and binding of the antibody to these peptides determined by ELISA or Surface plasmon resonance spectroscopy (Surface plasmon resonance). NMR studies using isotopically labeled antigens can also identify which amino acids have changed their magnetic environment when the antibody binds. Another method is to measure the thermal melting transition temperature (thermal melting transition temperature). The crystal structure of the antigen-antibody complex can also be determined and used to identify the epitope. Among these methods, the crystal method is most authoritative, followed by the NMR method.

In the ELISA method, unbound material is removed by a washing step prior to detection. Where the epitope is linear (the antibody recognizes only one single-chain linear amino acid sequence), the avidity of the antibody for the antigen comprising the epitope may be high enough to detect binding events. When an epitope is conformational, consisting of two or more non-contiguous amino acid sequences within a protein, the affinity of each individual sequence for an antibody may be low and not sufficiently detectable. Binding to peptides constituting conformational epitopes may also be undetectable using surface plasmon resonance spectroscopy, as affinity for each peptide in the epitope may be too low. If the off-rate of the peptide is high, binding may not be detectable.

Nuclear Magnetic Resonance (NMR) can also be used to identify epitopes on proteins. The applicant's co-pending application (U.S. patent application Ser. No.10/393926) proposes the identification of a specific atom (usually H) by local environment¹、C¹³And N¹⁵) And thereby identifying amino acid residues. If there is sufficient time and high enough resolution of the device, the complete set-up of most or all resonances can be made for the macromolecular protein. This approach is based on the discovery that the local environment of some amino acids changes when an antibody binds to an antigen. Those amino acids that change most strongly are those amino acids that are most relevant to the action of the antibody. Epitopes can theoretically be identified by performing all NMR mapping analysis of antibodies and antigens in bound and unbound states and determining which amino acids have undergone atomic movement. The complexity of the NMR spectra of antigen-antibody complexes makes such analysis extremely difficult and thus cannot be a routine epitope identification method. However, applicants of the above patent identified protein epitopes by this method using proteins rich in amino acids of the C13 and N15 atoms, where accurate NMR signal identification is not always required. When the degree of resonance difference between different amino acids is sufficiently large, multiple labeling of two or more different amino acids of the same protein may be performed. For example, a-N15 alanine and epsilon-N15 lysine can be included in the same protein, and epsilon-N15 histidine and alpha-N15 leucine can be included in the same protein. The epitope may be a binding region of an antibody or ligand. In addition, molecular models or calculations that predict protein surface exposure sequences can aid epitope identification.

In the case where no physical measurement of the structure of the target molecule is performed, the epitope can be designed based on the primary amino acid sequence of the target molecule. For example, about 5-10 residues within a linear fragment of about 5-15 amino acid residues can be altered to produce a variant decoy or chimeric target protein and the binding capacity measured to identify the epitope.

The basis of protein homology is the similarity of the base sequences of genes or the similarity of the amino acid sequences of proteins, which represent a common evolutionary origin. Common evolutionary origins generally result in proteins with similar structures and functions. This is not always the case, and divergent evolution and mutations may cause proteins with similar structures to have different functions, or proteins with different structures to have similar functions, which are referred to as orthologs and paralogs, respectively. Homology scanning mutagenesis is a well-known method of identifying protein receptor binding by replacing similar regions of a homologous protein to maintain the three-dimensional structure of the original protein (e.g., replacing the corresponding region of human auxin with a region in porcine auxin, human prolactin, or human placental lactogen) and then determining the binding constant of the construct. These natural structural variants can be used to identify domains or epitopes within domains that can be used to construct suitable decoy proteins of the invention.

Construction of traps

As used herein, a "decoy protein" refers to a protein that differs from a target protein in one or more structural features in a specific domain that includes the site or region to be bound. Thus, the decoy protein may be a chimeric target protein or a native protein, such as an species homolog. And the target ligand may be an engineered sequence comprising a predetermined binding domain. In the method of the invention, the decoys bind low affinity and non-specific binders, and those that complex with the retained target molecules are selected. In one aspect, an analog having a suitable structure and which can serve as a scaffold for accepting a target epitope can be an ortholog of the target protein. Structural homologues may also be other members of the multigene family.

With the progress of X-ray crystallography, NMR and other techniques, and the large accumulation of information on three-dimensional structures brought about by these techniques, protein structural information can be conveniently obtained, and true structural information can also be obtained from a variety of ways. The bioinformatics research Center (bioinformatics research Center) at the University of Glasgow (University of Glasgow) used Protein Structure Topology (TOPS) software to create network sites for describing and comparing Protein structures (TP waters, DSoss and JM thornton.1994.Protein Engineering, 7: 31-37). Protein coordinates (coordinates) may be submitted to the server in a PDB document-like format. Protein structures are transformed in a simple cartoon representation, called a topological representation. And then compared to a non-redundant (redundant) subset of all known structures. The results are presented in a sorted list showing the results of the similarity of the structure for a pair of identical structures assigned a value of 1 and no identical features assigned a value of 0.

The basis of the MASS (Multiple Alignment by Secondary construction structures) technique is a two-level comparison using secondary structure and atomic representation. The basic principle of this technology is that proteins are inherently composed of Secondary Structural Elements (SSEs). These elements are regions within the protein that provide a stable backbone for the protein to which the functional site is attached. Thus SSE is highly evolutionarily conserved, while mutations often occur in flexible loops that are more difficult to compare. MASS is a highly efficient tool for performing structural comparisons of multiple protein molecules and detecting common structural motifs. The use of secondary structure information helps to filter out interfering information and helps to improve efficiency and energy savings. MASS has the advantage of being sequence order independent, thus enabling detection of non-topological motifs in multiple comparisons or subgroups. Protein-protein docking can be manipulated using MASS, which is a rather difficult problem. MASS can be obtained free of charge from the following sites. (http:// bioinfo3d.cs.tau.ac.il/MASS/(Dror, O.et. protein Science (2003), 12: 2492-

The methods of the invention combine the use of structural information with large libraries of protein-encoding nucleic acid expression systems to select antibodies that bind a particular epitope. For example, complex specific epitopes on murine human Tissue Factor (TF) homologues can be targeted. Antibodies currently available in the art do not inhibit mTF function, nor are specific competitive inhibitors of factor X binding to TF. The antibodies disclosed herein have these functions and therefore represent tools that have not previously been available to evaluate the therapeutic potential of anti-TF antibodies (inhibition of TF activity by inhibition of FX activation).

In addition, these antibodies are also advantageous agents for studying the role of TF in normal conditions and in the development of pathogenic thrombo-inflammatory, angiogenic, neoplastic transformations.

Purification of epitope-directed antibodies or other binding ligands

The present invention provides three general methods for the isolation of epitope-directed antibodies or other binding ligands: (1) competitive selection using a display library of antibodies or other potential binding ligands; (2) using the display library, and screening for differential binding activity to perform non-competitive selection; (3) animals were immunized and screened for differential binding activity.

Competitive selection using a decoy protein, in the presence of the decoy protein (at a molar concentration above the target protein)SelectingBinding to the target protein in the library is displayed. Selective passage of recovered antibody or bound ligandScreeningAn isolated antibody or binding partner that binds to the target protein but not to the decoy.

Thus in one embodiment of the invention, there is provided a method of identifying a polypeptide that binds to a predetermined epitope on a target protein, the steps include (a) providing a library of phage particles expressing polypeptides on their surface, (b) preparing a decoy protein, which has been altered in its amino acid sequence corresponding to a predetermined epitope of the target protein, (c) incubating the library of phage particles with the target protein, selecting phage particles having a polypeptide capable of binding to the target protein, (d) adding a decoy protein (in molar excess of the target protein) as a competitor, negatively selecting phage particles specific for the preselected epitope, (e) separating phage particles bound to the target protein from phage particles bound to the decoy protein, (f) recovering phage particles bound to the target protein (but not phage particles bound to the decoy protein).

A less preferred method is to use the natural protein as a "decoy", toSelectingBinding to chimeric or mutant proteins. In this method, a protein comprising the original scaffold protein is used in a molar concentration greater than that of the chimeric or mutant protein. Selective passage of recovered antibody or bound ligandScreeningAn isolated antibody or binding ligand that binds to the decoy protein and the target protein but not to the scaffold protein.

In the two-step selection process using a decoy protein, the display library is selected according to the target protein. The recovered antibodies or other binding partners that selectively bind the target protein but not the decoy protein are rescreened (usually separately).

To use the immunization with the decoy protein, animals suitable for purification of the stable hybridoma monoclonal antibodies are immunized with the target protein. Hybridomas are prepared and screened for expression of antibodies that bind to the target protein but not to the decoy protein. The immunization method may be used in combination with any of the display methods described above. Thus mRNA from immune cells (e.g.spleen or peripheral blood lymphocytes) can be used to generate antibody libraries, which can then be used according to any of the display methods described above. This method is not limited to use with only those animals suitable for purification of stable hybridomas.

Peptide libraries can be designed according to the methods detailed herein, or can be made according to methods generally known in the art (see, e.g., RAPIDLIB1 or GRABLIB', DGI Biotechnologies, Inc., Edison, N.J.; Ph.D.C. 7C Disulifide constructed Peptide Library, New England Biolabs).

Antibody libraries can be obtained from sources such as Cambridge Antibody Technology, Morphosys, Affymaxresearch Institute, Palo Alto, CA, and the like. There are many methods to select subsets of binders available for further analysis and affinity maturation (migration). These methods include: blocking immunodominant epitopes by competitive exclusion, using epitope-masking techniques to obtain a wider range of antibody specificities, capture transfer (capturelift) screening, antibody-guided selection using sandwich-capture (capture-sandwich) ELISA, antibody selection near guides (proximols), human mab-guided selection using mouse mab-guided selection, antibody selection binding to cell surface antigens using magnetic sorting techniques, surface antigen-binding single chain antibodies of human tumor-associated cells, single chain antibodies purified using tissue fragment subtraction, antibodies were selected according to Antibody binding kinetics and functional antibodies were selected according to valency (Antibody Phage display: methods and procedures, Antibody Phage display.methods and protocols.IN: David W.J.Coomber, Ed.methods in Molecular biology. Humana Press. Vol.178, December 2001 pps.133-145).

Increasing the affinity is based on a low dissociation rate of the target binding partner. A low dissociation rate usually indicates a high affinity. In these examples of increased affinity, the continued incubation of the target phage and the target-binding phage is performed in the presence of saturating amounts of the known target binding partner, or by increasing the volume of the incubation solution. In each case, the re-binding of the lysed target-binding phage is prevented and, over a longer period of time, high affinity target-binding phage are recovered.

The preincubation time and preincubation conditions were optimized for each target-binding entity of interest. To monitor the effect of changes in conditions on increased affinity, a pilot experiment panning was performed. After incubation of the target and target-binding phage together and transformation of the host cells, the host cells are plated on selective plating media and counted. Determining the change in the number of surviving clones provides a powerful tool for assessing the level of avidity improvement. As the number of surviving clones decreased, the number of weak binders retained decreased significantly, leaving less high affinity target binder. For example, a reduction in the number of surviving clones to only 1%, 0.1%, 0.001% indicates optimal conditions for increasing the number of target binders that bind high affinity targets. Under some conditions, the number of surviving clones can be limited to about 100 for sequencing analysis.

Depending on the diversity of the target binder library types used, the number of target binders with high affinity may be as low as less than 10.

The affinity enhancement techniques described above can be performed without the need for additional panning. Of course, if necessary, the present invention may also use multiple rounds of panning to increase affinity.

Quote: all publications and patents cited herein are incorporated herein by reference in their entirety to the same extent as if each individual publication or patent were specifically and individually indicated to be within the scope of the invention and/or to provide any further description or authorization. The publication refers to any available information in the form of any scientific or patent publication, or any other medium, including all recorded, electronic documents, or printed matter. The following references are incorporated herein by reference: ausubel, et al, ed., general methods in molecular biology, Current Protocols in molecular biology, John Wiley & Sons, Inc., NY, NY (1987-; sambrook, et al, molecular cloning: laboratory guidelines, Molecular Cloning: a Laboratory Manual, 2nd edition, Cold Spring Harbor, NY (1989); harlow and Lane, antibody Laboratory Manual, antibodies, a Laboratory Manual, Cold Spring Harbor, NY (1989); colligan, et al, eds., methods in general Immunology, Current Protocols in Immunology, John Wiley & Sons, Inc., NY (1994-2004); colligan et al, general methods of Protein research Current Protocols in Protein Science, John Wiley & Sons, NY, NY, (1997-2004).

Having thus described the invention in general terms, embodiments of the invention are further disclosed by way of example.

Example 1: design and preparation of human/mouse chimeric tissue factor protein

Mab named TF8-5G9 recognizes and binds to human tissue factor and prevents factor X from binding to TF or to the TF/V IIa factor complex (Ruf, W.and Edginton, T.S.1991.Thromb. Haemost.66: 529. 539). Based on the analysis of the crystal structure of the TF8-5G9Fab complexed with human TF, all residues forming the epitope recognized by the Fab were between residues 149-204 of human TF. This protein region also plays an important role in the interaction between TF and the Gla-domain FX (Ruf et al 1992). The appropriate positioning of the 15 specific residues between residues 149-204 of huTF contributes an important energy contribution to binding (Huang, et al. J. MoI. biol.275, 873-894). As revealed by the sequence alignment below, when the extracellular domain sequences of human (GenPept Access No. NP-001984) and murine TF (GenPept Access No. NP-034301) were compared at residue 149-204 of the human EC domain with residue 152-207 of the murine EC domain, 7 of the 15 important residues were identical (human residues K149, K165, K166, T167, T170, 171, Q190) and 8 were different (human residues were substituted as follows for Y156T, K169I, V192M, P194F, V198T, R200Q, K201N, D204G). Residue pair in bold TF8-5G 9: the stability of the huTF complex plays an important role, and these residues have a delta binding free energy of 1 to 4kcal/mol or more.

Human:

₁₄₉KDLIYTLYYWKSSSSGKKTAKTNTNEFLIDVDKGENYCFSVQAVIPSRTVNRKSTD₂₀₄

mouse:

₁₅₂KDLGYIITYRKGSSTGKKTNITNTNERSIDVEEGVSYCFFVQAMIFSRKTNQNSPG₂₀₇

based on this analysis, chimeric trap proteins can be constructed from the murine TF coding sequence by mutating single TF8-5G9 contact residues in the murine TF sequence to the corresponding residues in huTF according to sequence alignment. Although there are other amino acid residue differences between human and murine TF, these residues are considered to have no effect on the overall function or structure of the protein when the target epitope is considered. A chimeric protein containing mutations was constructed using the mTF gene as a template, and 8 single TF8-5G9 contact residues of mTF were mutated to the corresponding residues in huTF (SEQ ID NO. 1). The transmembrane region was deleted so that only the soluble extracellular domain of TF was expressed and a His-tag was added at the carboxy-terminus to facilitate purification. Soluble murine TF and chimeric proteins were expressed in HEK293E cells and purified. The purified protein was analyzed by SDS-PAGE, resulting in a band of the expected size (Hu/m TF (40kDa), mTF (35 kDa)).

HuCAL phage display libraries (Morphosys, Martinsreid, Germany) were subjected to liquid phase panning using biotinylated mTF proteins. Chimeric hu/mTF protein was added as a decoy in a 10-fold molar excess to eliminate all phages specific for epitopes other than the mTF target epitope. The phage bound to biotinylated mTF was recovered by capture with streptavidin coated magnetic beads. All binders from this round of panning were sequenced, yielding 23 single fabs: at the concentrations tested, 9 recognized only mTF, 3 preferentially recognized mTF over hu/mTF, and 11 recognized both proteins similarly (table 1).

Panning of mTF without chimeric protein competitors was performed to confirm that the Fab selected was the result of epitope directed selection, and not the result of hotspot directed selection on mTF. The conditions of the two elutriations were identical except for the competing antigens. The resulting combination was sequenced to give 7 single fabs. Only 1 isolated Fab specifically bound mTF in panning without competitor, suggesting that the addition of competing antigen allows Fab selection to specifically recognize mTF, but not hu/mTF with a change in the epitope of TF8-5G9 (table 1).

TABLE 1

Human anti-murine TF specific Fab was purified by affinity chromatography and its ability to bind mTF or hu/mTF was assessed by ELISA. The CDR sequences for these fabs are listed in fig. 2; and (5) making framework arrangement by comparing the morphological change HuCAL handbooks. The framework sequences are listed below in figure 2. All 9 mTF-specific fabs showed dose-dependent binding to mTF with minimal cross-reactivity to hu/mTF (figure 3). In the Fab format, PHD127 has the highest binding affinity for mTF, and PHD127 the lowest. Full-length immunoglobulins were prepared by selecting 5 fabs (PHDs 103, 104, 126, 127, 130) based on their affinity for mTF. The variable regions of these 5 fabs (PHDs 103, 104, 126, 127, 130) are listed in figure 2, SEQ ID NOS: 2-11 were cloned into vectors to express mlgG2a molecules in HEK293 cells.

Anticoagulation

The ability of selected anti-mTF surrogate fabs to inhibit human plasma clotting was evaluated using murine brain extract as a source of mTF. Based on the above experiments, it was predicted that Fab binding to the epitope of TF8-5G9 on mTF can interfere with the coagulation pathway and delay clot formation. In this experiment, the inhibition of human plasma fibrin clot formation was measured. 4 of the 8 fabs tested delayed or inhibited coagulation of human plasma in vitro: PHD103, PHD104, PHD126, PHD 127. PHDs 126 and 127 have a greater ability to resist coagulation in human blood. According to the clot formation time-Fab concentration curve, E50 values between 0.2. mu.g/ml and 63. mu.g/ml were measured.

TABLE 2

FabEC50 concentration (μ g/ml)

PHD102 ＞200

PHD103 63.3

PHD104 23.8

PHD109 ＞200

PHD126 0.23

PHD127 0.82

PHD128 ＞200

PHD129 ＞200

Factor X inhibition

Factor X inhibition by anti-mTF Fab was measured above with murine brain extracts as a source of mTF. The extracts were incubated with FVIIa and the ability to inhibit conversion of FX to FXa was measured by adding anti-mTF instead of Mab in the presence of FX. PHD103, 126 and 127Fab inhibit factor X activation (cleavage) to factor Xa. The ability to inhibit factor X activation was then measured again using full-length anti-mTF IgG. PHDs 103, 126 and 127 were found to have good inhibitory effects, whereas PHD104 had no inhibitory effect.

FACS analysis

As the most attractive candidate antibodies PHD126 and 127, their ability to bind to B6F10 melanoma cells (high levels of expression of mTF) was evaluated. PHDs 126 and 127 bound mTF-linked cells in a dose-dependent manner with EC50 values of 37.8nM and 4.35nM, respectively (fig. 4). The complete variable region sequences of the heavy and light chains of PHDs 126 and 127, respectively, are set forth in the corresponding regions of FIG. 2 (SEQ ID NOS: 6-9).

Summary of the invention

The above experiments demonstrate that a method for epitope-directed selection of phage-displayed antibodies using engineered competitor proteins is feasible. The method designs suitable competitors based on structural information of the target protein. Alternatively, the method may select antibodies that are reactive with an epitope on the protein of interest. The existing methods for antibody selection using phage-displayed antibody libraries do not allow for the accurate targeting of epitopes of interest. The methods disclosed herein have the advantage that antibodies specific for a target epitope can be selected efficiently and accurately. We have used this method to select antibodies against a single epitope on mTF.

TF is a complex molecule that acts both as a receptor and as a ligand, capable of forming a specific complex with FVIIa and FX. Thus, mabs that prevent this interaction must act on a single region on the TF molecule. Antibodies currently available in the art either fail to inhibit mTF function or are not specific competitive inhibitors of factor X binding to TF. The antibodies disclosed herein have these functions and therefore represent tools that have not previously been available to evaluate the therapeutic potential of anti-TF antibodies (which inhibit TF activity by inhibiting FX activation). In addition, these antibodies are valuable reagents for investigating the role of TF in normal conditions and in the development of pathogenic thrombo-inflammatory, angiogenic, neoplastic transformations.

Example 2: construction of chimeric decoy proteins to select binders that bind to a common domain capable of activating different receptor subunits

Interleukin-13 (IL-13) is a cytokine that is found in elevated levels in the respiratory tract of asthmatic patients. IL-13 activation by activated CD4⁺T cell production and plays an important role in B cell proliferation and IgE production in asthmatic patients, as well as goblet cell proliferation and mucus hypersecretion, eosinophilic bronchitis, airway hyperresponsiveness. It has been found that overexpression of IL-13 in transgenic mice causes an asthma-like phenotype, while inhibition of IL-13 with antagonists reduces the asthmatic response.

IL-13 binds to at least two receptors, one of which can be found on most cells other than T cells, while the other receptor can act as a decoy protein. The receptors involved in the pro-inflammatory response are identical to the IL-4 receptor and consist of two subunits, IL4R α 1 and IL13R β 1. IL-13 is a member of the short chain cytokine family, which also includes IL-4, IL-2, IL-3, and GM-CSF. These proteins have a 4-helix bundle structure and contain two or three disulfide bonds. The solution structure of IL-13 has been determined to demonstrate similarity to other members of the family (Eisenmesser, E.Z., et al.J.MoI.biol. (2001) 310: 231-. Although IL-13 and IL-4 have only 25% sequence identity, the overall structure of the two is similar, and therefore it is predicted that the interaction between IL-13 and its receptor should be similar to that between IL-4 and its receptor (recently identified). Indeed, given that the receptors for both cytokines share a common subunit, the structures of IL-13 and IL-4 interaction with IL4R α 1 may be similar. Three-dimensional structure and mutation studies of IL-13 suggest that the cytokine has two surfaces that play important roles in its interaction with the receptor. The model suggests that the protein surface consisting of helix a and helix C interacts with the IL4R α 1 subunit of the receptor, while the interface of helix a and helix D interacts with the IL13R α 1 subunit of the receptor.

According to the prediction of IL-13 structure and receptor interaction model, an antibody blocking the interaction between A, D helix and IL13R α 1, or an antibody blocking the interaction between A-C surface and IL4R α 1, may be an excellent candidate antibody with anti-IL-13 therapeutic effect. In selecting antibodies that bind to the receptor interaction site of IL-13, we plan to make chimeric cytokine molecules. In such a chimeric protein, the loop joining the C, D helices would be replaced by the corresponding sequence of the animal to be immunized. In the receptor interaction model, the ring between C-D forms the majority of the surface of the exposed portion of the molecule and does not interact with the IL-13 receptor. In addition, this ring exhibits good flexibility in solution structure, and thus mutations that do not destroy the overall structure of the protein may be tolerated. In the resulting chimeric protein, a portion of the molecule is very similar to the host itself and is thus unlikely to induce a significant immune response. However, the portion of the molecule that retains the full-length human sequence may appear foreign to the host and may elicit an immune response. It is predicted that antibodies selected from chimeric immunogens will exhibit the function of neutralizing human receptor activity in the experiment.

There is a need for potent IL-13 antagonists to evaluate the role of IL-13 inhibition in human diseases, particularly asthma, and thus to use such antagonists as therapeutic agents. The novel IL-13 variants described herein may be used as immunogens to enhance the production of antagonistic antibodies, as screening or selection agents to identify neutralizing antibodies, or directly as antagonists of native IL-13. In addition, the development of novel potent IL-13 agonists may also be helpful in targeting certain cancer cells that overexpress the IL-13 receptor on their cell surface (Hussain, S.R. and Puri, R.K., Blood (2000) 95: 3506-351).

Novel IL-13 analogs were constructed. These compounds can be considered chimeras of human IL-13 and IL-13 from other species because they use partial sequences of IL-13 from multiple species. These mutants can be rationally designed by incorporating regions of sequence differences in other species into the human IL-13 sequence.

Based on the structural homology of two cytokines, an IL-13: IL-13R1 complex model was proposed. IL-13 (coordinate file): 1GA3) NMR models and IL-13 sequences were used to construct IL-13 analogs that should function as human IL-13 agonists, human IL-13 antagonists, or as immunogens for the production of anti-human IL-13 antibodies, or as panning elements for the production of anti-human IL-13 antibodies.

The coordinate file 1GA3, available from http:// www.ncbi.nlm.nih.gov/, contains an overlay of the 20NMR structures of IL-13. Structural observations indicate that although the 4 helices are highly conserved, the loop between the N-and C-termini and C, D is highly flexible, as evidenced by a number of conformations. The first structure of this document is used to analyze designed IL-13 mutants that should retain structure and activity.

There is a large loop between the C, D helices, which is close to the B helix almost buried inside the molecule. This loop can tolerate a mutation because it is far from the A, C, D helix. Ring B consisting of Met⁴³To Asn⁵³By the amino acid definition of (5), the CD loop is formed by Cys⁷¹To Thr⁸⁸The amino acid definition of (1). The loop ends are difficult to locate, but the loop end at the beginning of the D helix is definitely Glu⁹¹. The amino acids involved in the interaction between the B helix and the CD loop in most structures are:

and (3) helix B: cys is⁴⁵、Leu⁴⁸、GIu⁴⁹、Leu⁵¹(possible) Asn⁵³And Val⁵⁴

C/D Ring: cys is⁷¹、Val⁷⁵And (possibly) Lys⁷⁴、Val⁸⁵And (possibly) Arg⁸⁶、Ile⁹⁰

In addition, there are no hydrogen bonds in this region. Pro⁷²Without participation but necessary for the angle of rotation, Trp³⁵And Arg⁸⁶And Lys⁸⁹The loop residues between them have important interactions.

The residue on the B helix that interacts with the C/D ring is Leu⁴⁸、Leu⁵¹、Val⁵⁴。

Ala⁴⁷In a pocket, may be replaceable. There is no hydrogen bond between the C/D loop and the B helix.

The residue on the B helix which interacts with the A/B loop is Met⁴³、Ala⁴⁷、Ser⁵⁰。

A Blast search was performed in NCBI to identify other species of IL-13, with the following results (http:// www.ncbi.nlm.nih.gov/entrez/query. fcgidb ═ Protein):

human IL-13

GPVPPSTALRELIEELVNITQNQKAPLCNGSMVWSINLTAGMYCAALESLINVSGCSA

IEKTQRMLSGFCPHKVSAGQFSSLHVRDTKIEVAQFVKDLLLHLKKLFREGRFN

Pig

GPVPPHSTALKELIEELVNITQNQKTPLCNGSMVWSVNLTTSMQYCAALESLINISDC

SAIQKTQRMLSALCSHKPPSEQVPGKHIRDTKIEVAQFVKDLLKHLRMIFRHG

Cattle

PVPSATALKELIEELVNITQNQKVPLCNGSMVWSLNLTSSMYCAALDSLISISNCSVI

QRTKKMLNALCPHKPSAKQVSSEYVRDTKIEVAQFLKDLLRHSRIVFRNERFN

Dog

PVTPSPTLKELIEELVNITQNQASLCNGSMVWSVNLTAGMYCAALESLINVSDCSAIQ

RTQRMLKALCSQKPAAGQISSERSRDTKIEVIQLVKNLLTYVRGVYRHGNF

Rat

GPVRRSTSPPVALRELIEELSNITQDQKTSLCNSSMVWSVDLTAGGFCAALESLTNIS

SCNAIHRTQRILNGLCNQKASDVASSPPDTKIEVAQFISKLLNYSKQLFRYG

Mouse

GPVPRSVSLPLTLKELIEELSNITQDQTPLCNGSMVWSVDLAAGGFCVALDSLTNISN

CNAIYRTQRILHGLCNRKAPTTVSSLPDTKIEVAHFITKLLSYTKQLFRHGPF

Human, bovine, porcine, canine, rat and mouse IL-13 sequences were aligned using the ClustalW algorithm and Vector NTi Suite (InforMax, Inc., Bethesda, Md.) (FIG. 5).

The B-helix sequence is given below, with amino acids different from human underlined. Residues predicted to be involved in B helix and C/D loop interactions are marked with asterisks. (Table 3)

TABLE 3

Human being	M YCAALESLINV
		Cattle	M YCAALDSLISI
Pig	MQYCAALESLINI
		Dog	M YCAALESLINV
Rat	GFCAALESLTNI
		Mouse	GFCVALDSLTNI
Interacting residues	＊＊＊＊＊

These alignments suggest that some amino acid residues on the B helix and C/D loops of human IL-13 can be replaced with corresponding residues from IL-13 of other species, while maintaining structural integrity and receptor binding activity. Using residues on the B-helix and the C/D loop predicted to interact, a chimeric protein was designed in which the human protein C/D loop residues were replaced with similar residues from other species. To maintain protein stability, it is necessary to replace the corresponding interacting residues on the B helix from the same source. If bovine, porcine or mouse IL-13 is used, the preferred embodiment must only replace one amino acid, namely Val⁵⁴Replacement by Ile⁵⁴Since these other interacting residues on IL-13 are identical to those of the human protein.

Another embodiment is a two-way substitution of the B helix from the mouse protein, Glu⁴⁹→Asp⁴⁹、Ala⁴⁶→Val⁴⁶。Tyr⁴¹Can also be replaced by Phe, Leu⁵¹Val may also be substituted. But the last alternative above is close to the C-helix and disturbs its structure. Table 4 lists the C/D loop sequences of 6 proteins, with residues predicted to interact with the B loop indicated by asterisks in the last row. (Table 4)

TABLE 4

Human being	CPHKVSAGQFSSLHVRDTKI
		Cattle	CPHKPSAKQVSSEYVRDTKI
Pig	CSHKPPSEQVPGKHIRDTKI
		Dog	CSQKPAAGQISSERSRDTKI
Rat	CNQKASDVASS PPDTKI
		Mouse	CNRKAPTTVSS LPDTKI
Interacting residues	＊＊＊＊＊＊＊＊

Many preferred variations on the C/D loop are possible depending on homology. Many of these substitutions are unlikely to affect the overall conformation of the loop, but Arg in the rat and mouse Il-13 proteins⁸⁶Pro substitution has a major impact on the structure of the C/D loop, as well as deletion of 3 amino acids in rat and mouse proteins. Similarly, Arg in bovine, porcine and canine proteins⁸⁶Mutations to Pro also result in significant conformational rearrangements. Additional design options in this region include Ala observed on the B helix of the mouse protein⁴⁶→Val、Glu⁴⁹→ Asp mutation, and Val in the C/D ring⁸⁵The → Leu mutation.

The substitution of the corresponding amino acids in the human protein with the amino acids of the B helix and C/D loops of different species, modeling each mutein using highly optimized Insight II, can also be performed using other modeling tools.

Dog model: the energy of all the 5 models constructed was similar. Examination of the models revealed that their CD loops and other structures were very similar. Thus, replacement of residues from the B helix and CD loops of canine IL-13 with corresponding residues from the human protein is predicted to result in a suitable chimeric protein.

Cattle model: the energy was similar for all 5 models. There was a large difference in position in the 81-85 side chains in all 3 models, but this difference was no greater than that observed in the 20NMR model. Addition of an additional proline to the CD loop did not significantly change the conformation. These substitutions are therefore predicted to yield useful chimeras.

Pig model: the energy was similar for all 5 models. Like the bovine model, there were differences in some amino acid positions of the loop side chain, but no significant differences in the backbone. The addition of amino acids at the beginning of the B helix gives good results and therefore it is predicted that this variant will be a useful chimera.

Mouse model: after 3 amino acids deletion in the loop, the loop conformations of all 5 models were significantly different from the human model. All models had low energy, and these were the least energy of all chimeras based on absolute energy comparison. The overall structure of the 4 helices was almost unchanged, and it was predicted that this variant would be a suitable chimera.

All of these chimeras have a usable conformation and have similar structures at A, C and the D helix. The recommended chimera sequences are as follows:

human native protein

GPVPPSTALR ELIEELVNIT QNQKAPLCNG SMVWSINLTA GMYCAALESL

Human-cow (SEQ ID NO: 12)

GPVPPSTALR ELIEELVNIT QNQKAPLCNG SMVWSINLTA GMYCAALESL

Human-porcine SEQ ID NO: 13)

GPVPPSTALR ELIEELVNIT QNQKAPLCNG SMVWSINLTA GMQYCAALESL

human-dog (SEQ ID NO: 14)

GPVPPSTALR ELIEELVNIT QNQKAPLCNG SMVWSINLTA GMYCAALESL

Human-mouse (SEQ ID NO: 15)

GPVPPSTALR ELIEELVNIT QNQKAPLCNG SMVWSINLTA GMYCAALESL

Human native protein

INVSGCSAIE KTQRMLSGFC PHKVSAGQFS SLHVRDTKIE VAQFVKDLLL

Human-cow

INISGCSAIE KTQRMLSGFC PHKPSAKQVS SEYVRDTKIE VAQFVKDLLL

Human-pig

INISGCSAIE KTQRMLSGFC SHKPPSEQVP GKHIRDTKIE VAQFVKDLLL

Human-dog

INVSGCSAIE KTQRMLSGFC SQKPAAGQIS SERSRDTKIE VAQFVKDLLL

Human-mouse

INISGCSAIE KTQRMLSGFC NRKAPTTV SSLP DTKIE

VAQFVKDLLLHuman HLKKLFREGR FN

Human HLKKLFREGR FN

Human-cow HLKKLFREGR FN

Human-pig HLKKLFREGR FN

Human-mouse HLKKLFREGR FN

Listing chimeric trap protein sequences: SEQ ID NO: 12 (human-bovine), SEQ id no: 13 (human-pig), SEQ ID NO: 14 (human-dog), SEQ ID NO: 15 (human-murine). One of the uses of these chimeric proteins is to select antibodies that can functionally neutralize human IL-13 activity.

Antibodies can be recovered by antibody library screening/selection techniques such as antibody phage display. In another aspect, these chimeric IL-13 proteins can be used to screen and/or select for antibodies with neutralizing effects. In one embodiment, animals are immunized with IL-13 or one (or more) of these chimeric proteins and hybridomas are recovered, which can be selected based on their binding to a chimera not used for immunization, thereby avoiding recognition of the C/D loop by the antibody. In another embodiment, these chimeric proteins can be used in combination in different ways to select and screen combinatorial antibody libraries, in particular phage display libraries. According to the design concepts described herein, the selection or screening process inhibits recognition of the C/D loop by the antibody. Thus, both of the above embodiments may facilitate the isolation of neutralizing antibodies, particularly those that recognize A, C and the D-helix (these structures are fully conserved in both species variants and the chimeric proteins constructed).

A second use of these mutants is as antagonists of human IL-13. IL-13 is known to bind two receptor subunits. The introduction of non-human amino acids into two regions of the molecule results in minor structural changes in human IL-13 that have an allosteric effect on its binding to two receptor subunits. Competitive antagonists are obtained by selectively inhibiting receptor subunit binding.

Example 3: engineering trap protein calculations using NMR data

The present invention describes a novel method for identifying protein epitopes using Nuclear Magnetic Resonance (NMR) techniques. NMR techniques identify a particular atom (typically H) based on its local environment¹、C¹³、N¹⁵) Thereby identifying the amino acid. Carbon and nitrogen NMR spectra are less complex than proton NMR spectra, and the natural abundance of the required nuclei limits sensitivity. For large proteins, there will be a large number of spectral overlaps between atoms in similar environments. If the time is sufficient and the device resolution is high enough, a complete set-up of most or all resonances of large proteins can be achieved.

When an antibody binds to an antigen, the local environment of some amino acids changes. Those amino acids that vary the most are those that participate in the greatest degree of antibody exposure. The strategy for identifying epitopes is to perform a complete NMR assignment of the antigen and antibody in the bound and unbound states to determine which amino acid has undergone atomic shift. Without the equipment and methods of today, the complexity of the NMR spectra of antigen-antibody complexes makes such analysis extremely difficult and impractical for routine epitope identification.

Protein epitope identification can be performed using proteins containing C13-rich or N15-rich amino acids, where accurate NMR signal identification is not always required. These epitopes may be binding regions of antibodies or binding regions of receptors.

Recombinant techniques are used to express modified proteins in which an amino acid is replaced with the same amino acid containing the N15 or C13 marker. The obtained protein has the same structure and activity as the unmodified protein. The protein was then subjected to N15 or C13NMR spectroscopy in the presence and absence, respectively, of the binding antibody to the modified protein. The low natural abundance of resonance nuclei in unlabeled amino acids will simplify the NMR spectrum, allowing decoupled NMR spectra to show a single peak for the N15 spectrum and a single peak for C13 or a simple scheme thereof, depending on whether the amino acids are uniformly or specifically labeled. Resonance shift (shift) is observed when the labeled amino acid is involved in binding to the antibody. For example, a protein 200 amino acids long and containing 10N 15-labeled alanines will show 10 single peaks in its N15NMR spectrum. If two of these single peaks move after binding of the antibody, that is caused by a change in the local environment, their position in the epitope can be deduced from this. And those two alanine positions in the sequence are not available in this unimodal spectrum. When the above procedure was repeated 20 times and the amino acid was labeled differently each time, the composition of the epitope was known. Since the protein is obtained by recombination, its sequence is known. The epitope position can be determined according to the epitope composition and the protein sequence. Molecular models or algorithms that predict protein surface exposure sequences can aid epitope identification.

There is no need to prepare 20 marker proteins. Multiple labels (labeling 2 or more different amino acids in the same protein) can be used when sufficient resonance of different amino acids is resolved. For example, a-N15 alanine and e-N15 lysine, or 3-N15 histidine and a-N15 leucine can be incorporated into the same protein.

The method is superior to the existing epitope identification method. Conformational epitopes may be missed by those identification methods using synthetic peptides (polyethylene pin peptide synthesis techniques, solid phase (spot) synthesis techniques or liquid phase synthesis techniques in combination with ELISA or competition methods (competition)) or identification methods using phages. This NMR method, however, can conveniently detect both linear and conformational epitopes by using the intact protein. Solid phase (spot) synthetic variations (e.g. peptide matrix) are said to better identify conformational epitopes, but the number of peptides required is such that the number of amino acids in the protein increases exponentially, in short about 2 million peptides are required for a protein with a molecular weight of 40 kD. Some conformational epitopes can be identified by proteolysis in combination with mass spectrometry, but this method destroys the protein and requires a greater amount of protein for mass spectrometry as the protein molecule becomes larger. While NMR methods are not destructive to proteins. If the amount of marker protein is small, it can be recovered after each experiment and reused when other antibodies are mapped. Point mutations or "alanine scans" of a protein can effectively identify linear and conformational epitopes, but the difficulty is that each protein requires its DNA for expression, not all point mutants are secreted, and it is also necessary to determine whether each point mutant protein is correctly folded. This NMR method uses the same DNA for all labeled proteins, and the secretion and folding of the labeled and unlabeled proteins are identical. Crystallography is the "gold standard" (gold standard) for epitope identification. Its disadvantages are the large amount of time required, the possibility of large amounts of protein required, the difficulty of crystal growth at the diffraction level, and the fact that every antibody to the same antigen requires a new protein and a new crystal.

Example 4: construction of chimeric trap proteins using crystal structures

Based on the crystal structure of IL-4 and its cognate analog, IL-13, a particular receptor binding domain is selected as an antibody target. A chimeric protein was designed and prepared to select for binding to this region of both proteins.

The crystal structure of IL-4 has been obtained. The crystal structure of IL-13 has not been determined, but a theoretical model has been constructed. Due to their biological functions, both IL-4 and IL-13 are important therapeutic proteins. IL-4 has been reported to inhibit autoimmune diseases, while both IL-4 and IL-13 have the ability to enhance anti-tumor immune responses. On the other hand, since both cytokines are involved in the pathogenesis of allergic diseases, antagonism of them would be beneficial in the treatment of allergy and allergic asthma.

Several muteins (e.g., (IL-4Y124D antagonist and IL-13R112D agonist) the IL-4Y124D antagonist and the IL-13R112D agogonist, J.biol.chem (2000), 275, 14375-. The following novel IL-4 and IL-13 agonistic mutants were designed using molecular modeling. Since these mutants are predicted to be more stable than the native protein, they are expected to be biologically more potent binders for binding to cytokine receptors and have potential as antitumor drugs. Alternatively, these proteins can be used in solution phase panning procedures as stable analogs of native cytokines, or to search for domain-selective binding agents, such as a receptor binding domain antagonist.

The molecular and crystal structures of IL-4 and the theoretical model of IL-13 in the Brookhaven crystallographical Database were used to examine the structures of IL-4 and IL-13. Some amino acids within the molecular structure were identified, substitutions were made to these amino acids, and substitutions were not expected to negatively affect the structure. In fact, energy calculations indicate that these substituted structures are more stable than the native sequence. These substitutions are, IL-4: thr (Thr)¹³＝＞Ser¹³、Thr＝＞Ser²²、Phe55＝＞Tyr⁵⁵、Phe55＝＞Tyr⁵⁵，IL-13：Ile48＝＞Val⁴⁸、Gln⁹⁰＝＞Glu⁹⁰、Leu⁹⁵＝＞lle⁹⁵、Leu⁹⁶＝＞Ile⁹⁶、Leu⁹⁹＝＞Ile⁹⁹、Phe¹⁰³＝＞Tyr¹⁰³。

An IL-4 database was established (FIGS. 6A &6B), the contents of which include calculations for exposed amino acids. The first column includes only side chain data and the second column includes side chain data and backbone data. Amino acids with little and no surface exposure are indicated in bold/blue and.

Residues and cysteines buried within the molecule are removed from the table because their substitution affects the structure. Possible alternatives are calculated, with the results as follows:

the structural range of IL-4 was minimized by a cycle of 100 conjugate gradients, dielectric (dielecrtic)100, all hydrogen determined by the action of the Tripos force field and the Kollman-Uni field. The single changes suggested above are then made and the energy calculated. Based on these calculations, the best replacement is Ser to Thr¹³Ser to Thr²²Replacement of Phe by Tyr⁴⁵Replacement of Phe by Tyr⁵⁵。

The crystal structure of IL-4 was searched, the energy before the structure was not minimized was calculated, and the energy after the above 4 substitutions was calculated, with the following results:

natural crystal structure

Key expansion energy: 231.645

Angular bending energy: 298.910

Torsion energy: 453.633

Out-of-plane bending energy: 46.674

1-4 van der waals force energy: 386.851

Van der waals force energy: 1134.199

1-4 electrostatic energy: 30.608

Electrostatic energy: 1043.112

-----------------------------------------

Total energy: 3625.631kcals/mol

Ser ¹³ 、Ser ²³ 、Tyr ⁴⁵ 、Tyr ⁵⁵ Structure of the product

Key expansion energy: 233.378

Angular bending energy: 299.147

Torsion energy: 449.605

Out-of-plane bending energy: 46.468

1-4 van der waals force energy: 384.334

Van der waals force energy: 1125.512

1-4 electrostatic energy: 30.677

Electrostatic energy: 1043.004

------------------------------------------

Total energy: 3612.126kcals/mol

The torsional energy, 1-4 van der waals forces, and van der waals forces are all reduced. From energy calculations, this structure is predicted to be more stable than the native sequence. The modified amino acid is buried inside the molecule, but since it is considered not to affect the secondary structure after the evaluation of the substitution, the surface structure and thus the functional activity are not changed.

A similar table was constructed for IL-13 (FIGS. 7A & 7B). The crystal structure is not published but has a theoretical model. Both the original and the modified structures were subjected to 10 rounds of minimization.

Internal residues may also be modified.

Replacement of Phe with Tyr⁶⁶The post energy is higher but it looks like a good alternative. The higher energy comes from higher inter-hydroxyl van der waals forces.

Phe⁶⁶And His⁶⁹Interactions (π - π), both of which must retain their aromaticity.

Ile replacement of Va^l92Bringing higher energy, but a small amount of minimization (minimization) greatly reduces this value. This is achieved byA single replacement is likely to be available.

Replacement of Phe103 by Tyr adds one to His⁶⁹May be a good alternative.

The IL-13 structure was retrieved, energy calculated, replaced and energy calculated after replacement, with the following results:

initial model structure

Key expansion energy: 290.14

Angular bending energy: 390.042

Torsion energy: 388.721

Out-of-plane bending energy: 30.217

1-4 van der waals force energy: 286.329

Van der waals force energy: 210.384

1-4 electrostatic energy: 27.906

Electrostatic energy: -1.547

----------------------------------------

Total energy: 1622.216kcals/mol

Val ⁴⁸ 、Glu ⁹⁰ 、Ile ⁹⁵ 、Ile ⁹⁶ 、Tyr ¹⁰³ Structure of the product

Key expansion energy: 288.604

Angular bending energy: 383.273

Torsion energy: 384.452

Out-of-plane bending energy: 29.889

1-4 van der waals force energy: 284.920

Van der waals force energy: 228.808

1-4 electrostatic energy: 27.989

Electrostatic energy: -1.594

----------------------------------------

Total energy: 1626.342kcals/mol

The bond stretching energy, the angle bending energy, 1-4 van der waals force and van der waals force energy decrease, and the twisting energy and the bond stretching energy increase. The latter can be reduced by repositioning the newly placed Ile side chain.

Decoration structure

Energy RMS force max force interaction Eval CPU time

Counting time of kcals/mol kcals/mol A kcals/mol A

1626.342 25.436 243.087 0 1 0

0:00:00.33

1326.768 14.851 134.040 1 8 0

0:00:01.11

1207.892 10.799 137.468 2 14 0

0:00:01.78

1121.012 10.155 171.991 3 20 0

0:00:02.44

1065.839 7.613 105.188 4 26 0

0:00:03.11

1031.513 6.361 73.260 5 32 0

0:00:03.76

995.717 6.643 64.166 6 38 0

0:00:04.42

965.486 5.399 54.345 7 44 0

0:00:05.09

945.615 5.429 61.839 8 50 0

0:00:05.76

924.480 4.655 50.214 9 56 0

0:00:06.42

907.908 4.448 55.866 10 62 0

0:00:07.08

Warning: the maximum number of iterations (10) has been reached.

Molecular energy of IL-13 model 1 (theoretical model)

Key expansion energy: 49.661

Angular bending energy: 297.149

Torsion energy: 328.222

Out-of-plane bending energy: 8.744

1-4 van der waals force energy: 189.249

Van der waals force energy: 8.139

1-4 electrostatic energy: 28.271

Electrostatic energy: -1.556

--------------------------------------

Total energy: 907.908kcals/mol

5677 average number of van der waals force + electrostatic pairs

The average number of 1-4 van der waals + electrostatic pairs is 3415

Average number of standard (scaled) van der waals + electrostatic pairs 248

	Number of	CPU time (sec)	% of the total
				Nonbond interactions (rebalds)	2	0.08	1.08
Energy assessment	64	7.33	98.92

Initial structure

IL-13 model 1 (theoretical model) molecular energy

Energy RMS force max force interaction Eva1 CPU time

Counting time of kcals/mol kcals/mol A kcals/mol A

1622.189 25.201 243.087 0 1 0

0:00:00.34

1328.927 14.694 132.377 1 8 0

0:00:01.12

1212.237 10.675 135.392 2 14 0

0:00:01.79

1127.031 10.061 169.922 3 20 0

0:00:02.47

1072.925 7.499 103.477 4 26 0

0:00:03.13

1039.540 6.289 73.212 5 32 0

0:00:03.81

1004.102 6.619 64.355 6 38 0

0:00:04.49

973.968 5.370 54.031 7 44 0

0:00:05.15

954.110 5.428 64.203 8 50 0

0:00:05.82

932.787 4.648 46.556 9 56 0

0:00:06.50

915.954 4.442 52.618 10 62 0

0:00:07.16

Warning: the maximum number of iterations has been reached (10).

IL-13 model 1 (theoretical model) molecular energy

Key expansion energy: 48.807

Angular bending energy: 300.150

Torsion energy: 331.002

Out-of-plane bending energy: 8.945

1-4 van der waals force energy: 192.067

Van der waals force energy: 8.307

1-4 electrostatic energy: 28.178

Electrostatic energy: -1.503

---------------------------------------------

Total energy: 915.954kcals/mol

The average number of van der waals + electrostatic pairs is 5717

The average number of 1-4 van der waals + electrostatic pairs 3426

Average number of standard van der waals + electrostatic pairs 247

	Number of	CPU time (sec)	% of the total
				Nonbond interactions (rebalds)	2	0.08	1.07
Energy assessment	64	7.41	98.93

The molecular and crystal structures of IL-4 and the theoretical model of IL-13 in the Brookhaven crystallographical Database were used to examine the structures of IL-4 and IL-13. Some amino acids within the molecular structure were identified, substitutions were made to these amino acids, and substitutions were not expected to negatively affect the structure. In fact, energy calculations indicate that these substituted structures are more stable than the native sequence. These substitutions are, IL-4: thr (Thr)¹³＝＞Ser¹³、Thr＝＞Ser²²、Phe⁴⁵＝＞Tyr⁴⁵、Phe⁵⁵＝＞Tyr⁵⁵，IL-13：Ile⁴⁸＝＞Val⁴⁸、Gln⁹⁰＝＞Glu⁹⁰、Leu⁹⁵＝＞Ile⁹⁵、Leu⁹⁶＝＞Ile⁹⁶、Leu⁹⁹＝＞Ile⁹⁹、Phe¹⁰³＝＞Tyr¹⁰³。

The complete sequence is:

IL-4 construct (SEQ ID NO: 16)

HKCDITLQEI IKSLNSLTEQ KSLCTELTVT DIFAASKNTT EKETYCRAAT VLRQYYSHHE

KDTRCLGATA QQFHRHKQLI RFLKRLDRNL WGLAGLNSCP VKEANQSTLE NFLERLKTIM

REKYSKCSS

IL-13 construct (SEQ ID NO: 17)

PPSTALRELI EELVNITQNQ KAPLCNGSMV WSINLTAGMY CAALESLVNV SGCSAIEKTQ

RMLSGFCPHK VSAGQFSSLH VRDTKIEVAE FVKDIILHIK KLYREGRFN

The replacement amino acids are underlined.

The energy of the few post-minimal modified structures was lower, indicating that this structure is more stable than the original sequence. These constructs and constructs prepared by other similar means may be used in the methods of the invention.

Sequence listing

<110>Centocor，Inc.

o′Neil，Karyn

Heavner，George

Sweet，Raymond

<120>Solution Phase Biopanning Method using Engineered Decouy Proteins

<130>CEN5055

<140>TBD

<141>2005-03-21

<150>60/565,633

<151>2004-04-26

<150>60/565,674

<151>2004-04-26

<160>17

<170>PatentIn version 3.3

<210>1

<211>221

<212>PRT

<213> Artificial

<220>

<223> murine TFw 8Hu residue

<400>1

Gly Ile Pro Glu Lys Ala Phe Asn Leu Thr Trp Ile Ser Thr Asp Phe

1 5 10 15

Lys Thr Ile Leu Glu Trp Gln Pro Lys Pro Thr Asn Tyr Thr Tyr Thr

20 25 30

Val Gln Ile Ser Asp Arg Ser Arg Asn Trp Lys Asn Lys Cys Phe Ser

35 40 45

Thr Thr Asp Thr Glu Cys Asp Leu Thr Asp Glu Ile Val Lys Asp Val

50 55 60

Thr Trp Ala Tyr Glu Ala Lys Val Leu Ser Val Pro Arg Arg Asn Ser

65 70 75 80

Val His Gly Asp Gly Asp Gln Leu Val Ile His Gly Glu Glu Pro Pro

85 90 95

Phe Thr Asn Ala Pro Lys Phe Leu Pro Tyr Arg Asp Thr Asn Leu Gly

100 105 110

Gln Pro Val Ile Gln Gln Phe Glu Gln Asn Gly Arg Lys Leu Asn Val

115 120 125

Val Val Lys Asp Ser Leu Thr Leu Val Arg Lys Asn Gly Thr Phe Leu

130 135 140

Thr Leu Arg Gln Val Phe Gly Lys Asp Leu Gly Tyr Ile Ile Tyr Tyr

145 150 155 160

Arg Lys Gly Ser Ser Thr Gly Lys Lys Thr Asn Lys Thr Asn Thr Asn

165 170 175

Glu Phe Ser Ile Asp Val Glu Glu Gly Val Ser Tyr Cys Phe Phe Val

180 185 190

Gln Ala Val Ile pro Ser Arg Lys Val Asn Arg Lys Ser Pro Asp Ser

195 200 205

Ser Thr Val Cys Thr Glu Gln Trp Lys Ser Phe Leu Gly

210 215 220

<210>2

<211>118

<212>PRT

<213> human

<220>

<221> binding

<222>(26)..(35)

<223>CDR1

<220>

<221> binding

<222>(50)..(67)

<223>CDR2

<220>

<221> binding

<222>(100)..(107)

<223>CDR3

<400>2

Gln Val Gln Leu Val Gln Ser Gly Ala Glu Val Lys Lys Pro Gly Glu

1 5 10 15

Ser Leu Lys Ile Ser Cys Lys Gly Ser Gly Tyr Ser Phe Ser Asn Ser

20 25 30

Trp Ile Ala Trp Val Arg Gln Met Pro Gly Lys Gly Leu Glu Trp Met

35 40 45

Gly Gly Ile Ile Gly Pro Gly His Ser Tyr Thr Lys Tyr Ser Pro Ser

50 55 60

Phe Gln Gly Gln Val Thr Ile Ser Ala Asp Lys Ser Ile Ser Thr Ala

65 70 75 80

Tyr Leu Gln Trp Ser Ser Leu Lys Ala Ser Asp Thr Ala Met Tyr Tyr

85 90 95

Cys Ala Arg Ile Asn Met Gly Tyr Phe Asp Tyr Trp Gly Gln Gly Thr

100 105 110

Leu Val Thr Val Ser Ser

115

<210>3

<211>111

<212>PRT

<213> human

<220>

<221> binding

<222>(23)..(34)

<223>CDR1

<220>

<221> binding

<222>(50)..(56)

<223>CDR2

<220>

<221> binding

<222>(89)..(99)

<223>CDR3

<400>3

Asp Ile Glu Leu Thr Gln Pro Pro Ser Val Ser Val Ala Pro Gly Gln

1 5 10 15

Thr Ala Arg Ile Ser Cys Ser Gly Asp Ser Leu Gly Leu Lys Lys Phe

20 25 30

Val Ser Trp Tyr Gln Gln Lys Pro Gly Gln Ala Pro Val Leu Val Ile

35 40 45

Tyr Asp Asp Ser Asn Arg Pro Ser Gly Ile Pro Glu Arg Phe Ser Gly

50 55 60

Ser Asn Ser Gly Asn Thr Ala Thr Leu Thr Ile Ser Gly Thr Gln Ala

65 70 75 80

Glu Asp Glu Ala Asp Tyr Tyr Cys Gly Thr Tyr Asp Gln Thr Ile Gly

85 90 95

His Asp Val Val Phe Gly Gly Gly Thr Lys Leu Thr Val Leu Gly

100 105 110

<210>4

<211>116

<212>PRT

<213> human

<220>

<221> binding

<222>(26)..(34)

<223>CDR1

<220>

<221> binding

<222>(49)..(65)

<223>CDR2

<220>

<221> binding

<222>(98)..(105)

<223>CDR3

<400>4

Gln Val Gln Leu Val Gln Ser Gly Ala Glu Val Lys Lys Pro Gly Glu

1 5 10 15

Ser Leu Lys Ile Ser Cys Lys Gly Ser Tyr Ser Phe Thr Ser Asn Trp

20 25 30

Ile Gly Trp Val Arg Gln Met Pro Gly Lys Gly Leu Glu Trp Met Gly

35 40 45

Trp Ile Tyr Pro Ser Asp Ser Met Thr Arg Tyr Ser Pro Ser Phe Gln

50 55 60

Gly Gln Val Thr Ile Ser Ala Asp Lys Ser Ile Ser Thr Ala Tyr Leu

65 70 75 80

Gln Trp Ser Ser Leu Lys Ala Ser Asp Thr Ala Met Tyr Tyr Cys Ala

85 90 95

Arg Tyr Leu Phe Gly Leu Phe Asp Asn Trp Gly Gln Gly Thr Leu Val

100 105 110

Thr Val Ser Ser

115

<210>5

<211>107

<212>PRT

<213> human

<220>

<221> binding

<222>(23)..(33)

<223>CDR1

<220>

<221> binding

<222>(49)..(55)

<223>CDR2

<220>

<221> binding

<222>(88)..(95)

<223>CDR3

<400>5

Asp Ile Glu Leu Thr Gln Pro Pro Ser Val Ser Val Ala Pro Gly Gln

1 5 10 15

Thr Ala Arg Ile Ser Cys Ser Gly Asp Asn Leu Gly Ser Tyr Tyr Val

20 25 30

Ser Trp Tyr Gln Gln Lys Pro Gly Gln Ala Pro Val Leu Val Ile Tyr

35 40 45

Asn Asp Asn Asn Arg Pro Ser Gly Ile Pro Glu Arg Phe Ser Gly Ser

50 55 60

Asn Ser Gly Asn Thr Ala Thr Leu Thr Ile Ser Gly Thr Gln Ala Glu

65 70 75 80

Asp Glu Ala Asp Tyr Tyr Cys Ala Thr Tyr Asp Ser Ser Thr Asp Val

85 90 95

Phe Gly Gly Gly Thr Lys Leu Thr Val Leu Gly

100 105

<210>6

<211>119

<212>PRT

<213> human

<220>

<221> binding

<222>(26)..(34)

<223>CDR1

<220>

<221> binding

<222>(49)..(65)

<223>CDR2

<220>

<221> binding

<222>(98)..(108)

<223>CDR3

<400>6

Gln Val Gln Leu Val Gln Ser Gly Ala Glu Val Lys Lys Pro Gly Glu

1 5 10 15

Ser Leu Lys Ile Ser Cys Lys Gly Ser Tyr Ser Phe Ser Asn Tyr Trp

20 25 30

Ile Gly Trp Val Arg Gln Met Pro Gly Lys Gly Leu Glu Trp Met Gly

35 40 45

Phe Ile Asp Pro Asp Asp Ser Asp Thr Asn Tyr Ser Pro Ser Phe Gln

50 55 60

Gly Gln Val Thr Ile Ser Ala Asp Lys Ser Ile Ser Thr Ala Tyr Leu

65 70 75 80

Gln Trp Ser Ser Leu Lys Ala Ser Asp Thr Ala Met Tyr Tyr Cys Ala

85 90 95

Arg Ala Leu Tyr Met Gln Gly Gly Ser Phe Asp Ser Trp Gly Gln Gly

100 105 110

Thr Leu Val Thr Val Ser Ser

115

<210>7

<211>109

<212>PRT

<213> human

<220>

<221> binding

<222>(23)..(33)

<223>CDR1

<220>

<221> binding

<222>(49)..(55)

<223>CDR2

<220>

<221> binding

<222>(88)..(97)

<223>CDR3

<400>7

Asp Ile Glu Leu Thr Gln Pro Pro Ser Val Ser Val Ala Pro Gly Gln

1 5 10 15

Thr Ala Arg Ile Ser Cys Ser Gly Asp Asn Leu Gly Ser Tyr Tyr Val

20 25 30

Ser Trp Tyr Gln Gln Lys Pro Gly Gln Ala Pro Val Leu Val Ile Tyr

35 40 45

Arg Asp Thr Asp Arg Pro Ser Gly Ile Pro Glu Arg Phe Ser Gly Ser

50 55 60

Asn Ser Gly Asn Thr Ala Thr Leu Thr Ile Ser Gly Thr Gln Ala Glu

65 70 75 80

Asp Glu Ala Asp Tyr Tyr Cys Gln Ser Tyr Asp Tyr Gly Val Ser Asn

85 90 95

Gln Val Phe Gly Gly Gly Thr Lys Leu Thr Val Leu Gly

100 105

<210>8

<211>119

<212>PRT

<213> human

<220>

<221> binding

<222>(26)..(34)

<223>CDR1

<220>

<221> binding

<222>(49)..(65)

<223>CDR2

<220>

<221> binding

<222>(98)..(108)

<223>CDR3

<400>8

Gln Val Gln Leu Val Gln Ser Gly Ala Glu Val Lys Lys Pro Gly Glu

1 5 10 15

Ser Leu Lys Ile Ser Cys Lys Gly Ser Tyr Ser Phe Thr Asn Ser Trp

20 25 30

Ile Ser Trp Val Arg Gln Met Pro Gly Lys Gly Leu Glu Trp Met Gly

35 40 45

Ile Ile Asp Pro Asp Asp Ser Tyr Thr Ser Tyr Ser Pro Ser Phe Gln

50 55 60

Gly Gln Val Thr Ile Ser Ala Asp Lys Ser Ile Ser Thr Ala Tyr Leu

65 70 75 80

Gln Trp Ser Ser Leu Lys Ala Ser Asp Thr Ala Met Tyr Tyr Cys Ala

85 90 95

Arg Gly Ala Gly Tyr Gly Arg Met Phe Gly Asp Val Trp Gly Gln Gly

100 105 110

Thr Leu Val Thr Val Ser Ser

115

<210>9

<211>108

<212>PRT

<213> human

<220>

<221> binding

<222>(23)..(33)

<223>CDR1

<220>

<221> binding

<222>(49)..(55)

<223>CDR2

<220>

<221> binding

<222>(88)..(96)

<223>CDR3

<400>9

Asp Ile Glu Leu Thr Gln Pro Pro Ser Val Ser Val Ala Pro Gly Gln

1 5 10 15

Thr Ala Arg Ile Ser Cys Ser Gly Asp Asn Leu Gly Ser Tyr Tyr Ala

20 25 30

Ser Trp Tyr Gln Gln Lys Pro Gly Gln Ala Pro Val Leu Val Ile Tyr

35 40 45

Gln Asp Asp Asn Arg Pro Ser Gly Ile Pro Glu Arg Phe Ser Gly Ser

50 55 60

Asn Ser Gly Asn Thr Ala Thr Leu Thr Ile Ser Gly Thr Gln Ala Glu

65 70 75 80

Asp Glu Ala Asp Tyr Tyr Cys Gly Ala Tyr Thr Tyr Ser Thr Ser Trp

85 90 95

Val Phe Gly Gly Gly Thr Lys Leu Thr Val Leu Gly

100 105

<210>10

<211>118

<212>PRT

<213> human

<220>

<221> binding

<222>(26)..(37)

<223>CDR1

<220>

<221> binding

<222>(52)..(67)

<223>CDR2

<220>

<221> binding

<222>(100)..(107)

<223>CDR3

<400>10

Gln Val Gln Leu Lys Glu Ser Gly Pro Ala Leu Val Lys Pro Thr Gln

1 5 10 15

Thr Leu Thr Leu Thr Cys Thr Phe Ser Gly Leu Ser Leu Ser Thr Sar

20 25 30

Gly Val Gly Val Gly Trp Ile Arg Gln Pro Pro Gly Lys Ala Leu Glu

35 40 45

Trp Leu Ala Leu Ile Tyr Ser Asn Asp Asp Lys Arg Tyr Ser Thr Ser

50 55 60

Leu Lys Thr Arg Leu Thr Ile Ser Lys Asp Thr Ser Lys Asn Gln Val

65 70 75 80

Val Leu Thr Met Thr Asn Met Asp Pro Val Asp Thr Ala Thr Tyr Tyr

85 90 95

Cys Ala Arg Tyr Lys Gln Glu Thr Ile Asp Tyr Trp Gly Gln Gly Thr

100 105 110

Leu Val Thr Val Ser Ser

115

<210>11

<211>105

<212>PRT

<213> human

<220>

<221> binding

<222>(23)..(33)

<223>CDR1

<220>

<221> binding

<222>(49)..(53)

<223>CDR2

<220>

<221> binding

<222>(86)..(93)

<223>CDR3

<400>11

Asp Ile Glu Leu Thr Gln Pro Pro Ser Val Ser Val Ala Pro Gly Gln

1 5 10 15

Thr Ala Arg Ile Ser Cys Ser Gly Asp Asn Leu Gly Glu Lys Tyr Ala

20 25 30

Tyr Trp Tyr Gln Gln Lys Pro Gly Gln Ala Pro Val Leu Val Ile Tyr

35 40 45

Asp Asp Asn Asn Arg Gly Ile Pro Glu Arg Phe Ser Gly Ser Asn Ser

50 55 60

Gly Asn Thr Ala Thr Leu Thr Ile Ser Gly Thr Gln Ala Glu Asp Glu

65 70 75 80

Ala Asp Tyr Tyr Cys Gln Ser Tyr Asp Ile Glu Ile Thr Val Phe Gly

85 90 95

Gly Gly Thr Lys Leu Thr Val Leu Gly

100 105

<210>12

<211>112

<212>PRT

<213> Artificial sequence

<220>

<223> chimeric protein based on substitution of human sequence and bovine homologous residue

<400>12

Gly Pro Val Pro Pro Ser Thr Ala Leu Arg Glu Leu Iie Glu Glu Leu

1 5 10 15

Val Asn Ile Thr Gln Asn Gln Lys Ala Pro Leu Cys Asn Gly Ser Met

20 25 30

Val Trp Ser Ile Asn Leu Thr Ala Gly Met Tyr Cys Ala Ala Leu Glu

35 40 45

Ser Leu Ile Asn Ile Ser Gly Cys Ser Ala Ile Glu Lys Thr Gln Arg

50 55 60

Met Leu Ser Gly Phe Cys Pro His Lys Pro Ser Ala Lys Gln Val Ser

65 70 75 80

Ser Glu Tyr Val Arg Asp Thr Lys Ile Glu Val Ala Gln Phe Val Lys

85 90 95

Asp Leu Leu Leu His Leu Lys Lys Leu Phe Arg Glu Gly Arg Phe Asn

100 105 110

<210>13

<211>113

<212>PRT

<213> Artificial sequence

<220>

<223> chimeric protein based on substitution of human sequence and porcine homologous residue

<400>13

Gly Pro Val Pro Pro Ser Thr Ala Leu Arg Glu Leu Ile Glu Glu Leu

1 5 10 15

Val Asn Ile Thr Gln Asn Gln Lys Ala Pro Leu Cys Asn Gly Ser Met

20 25 30

Val Trp Ser Ile Asn Leu Thr ALa Gly Met Gln Tyr Cys Ala Ala Leu

35 40 45

Glu Ser Leu Ile Asn Ile Ser Gly Cys Ser Ala Ile Glu Lys Thr Gln

50 55 60

Arg Met Leu Ser Gly Phe Cys Ser His Lys Pro Pro Ser Glu Gln Val

65 70 75 80

Pro Gly Lys His Ile Arg Asp Thr Lys Ile Glu Val Ala Gln Phe Val

85 90 95

Lys Asp Leu Leu Leu His Leu Lys Lys Leu Phe Arg Glu Gly Arg Phe

100 105 110

Asn

<210>14

<211>112

<212>PRT

<213> Artificial sequence

<220>

<223> chimeric protein based on substitution of human sequence with dog homologous residue

<400>14

Gly Pro Val Pro Pro Ser Thr Ala Leu Arg Glu Leu Ile Glu Glu Leu

1 5 10 15

Val Asn Ile Thr Gln Asn Gln Lys Ala Pro Leu Cys Asn Gly Ser Met

20 25 30

Val Trp Ser Ile Asn Leu Thr Ala Gly Met Tyr Cys Ala Ala Leu Glu

35 40 45

Ser Leu Ile Asn Val Ser Gly Cys Ser Ala Ile Glu Lys Thr Gln Arg

50 55 60

Met Leu Ser Gly Phe Cys Ser Gln Lys Pro Ala Ala Gly Gln Ile Ser

65 70 75 80

Ser Glu Arg Ser Arg Asp Thr Lys Ile Glu Val Ala Gln Phe Val Lys

85 90 95

Asp Leu Leu Leu His Leu Lys Lys Leu Phe Arg Glu Gly Arg Phe Asn

100 105 110

<210>15

<211>109

<212>PRT

<213> Artificial sequence

<220>

<223> chimeric protein based on substitution of human sequence and mouse homologous residue

<400>15

Gly Pro Val Pro Pro Ser Thr Ala Leu Arg Glu Leu Ile Glu Glu Leu

1 5 10 15

Val Asn Ile Thr Gln Asn Gln Lys Ala Pro Leu Cys Asn Gly Ser Met

20 25 30

Val Trp Ser Ile Asn Leu Thr Ala Gly Met Tyr Cys Ala Ala Leu Glu

35 40 45

Ser Leu Ile Asn Ile Ser Gly Cys Ser Ala Ile Glu Lys Thr Gln Arg

50 55 60

Met Leu Ser Gly Phe Cys Asn Arg Lys Ala Pro Thr Thr Val Ser Ser

65 70 75 80

Leu Pro Asp Thr Lys Ile Glu Val Ala Gln Phe Val Lys Asp Leu Leu

85 90 95

Leu His Leu Lys Lys Leu Phe Arg Glu Gly Arg Phe Asn

100 105

<210>16

<211>129

<212>PRT

<213> Artificial sequence

<220>

<223> human IL-4 with improved stability

<400>16

His Lys Cys Asp Ile Thr Leu Gln Glu Ile Ile Lys Ser Leu Asn Ser

1 5 10 15

Leu Thr Glu Gln Lys Ser Leu Cys Thr Glu Leu Thr Val Thr Asp Ile

20 25 30

Phe Ala Ala Ser Lys Asn Thr Thr Glu Lys Glu Thr Tyr Cys Arg Ala

35 40 45

Ala Thr Val Leu Arg Gln Tyr Tyr Ser His His Glu Lys Asp Thr Arg

50 55 60

Cys Leu Gly Ala Thr Ala Gln Gln Phe His Arg His Lys Gln Leu Ile

65 70 75 80

Arg Phe Leu Lys Arg Leu Asp Arg Asn Leu Trp Gly Leu Ala Gly Leu

85 90 95

Asn Ser Cys Pro Val Lys Glu Ala Asn Gln Ser Thr Leu Glu Asn Phe

100 105 110

Leu Glu Arg Leu Lys Thr Ile Met Arg Glu Lys Tyr Ser Lys Cys Ser

115 120 125

Ser

<210>17

<211>109

<212>PRT

<213> Artificial sequence

<220>

<223> human IL-13 with enhanced stability

<400>17

Pro Pro Ser Thr Ala Leu Arg Glu Leu Ile Glu Glu Leu Val Asn Ile

1 5 10 15

Thr Gln Asn Gln Lys Ala Pro Leu Cys Asn Gly Ser Met Val Trp Ser

20 25 30

Ile Asn Leu Thr Ala Gly Met Tyr Cys Ala Ala Leu Glu Ser Leu Val

35 40 45

Asn Val Ser Gly Cys Ser Ala Ile Glu Lys Thr Gln Arg Met Leu Ser

50 55 60

Gly Phe Cys Pro His Lys Val Ser Ala Gly Gln Phe Ser Ser Leu His

65 70 75 80

Val Arg Asp Thr Lys Ile Glu Val Ala Glu Phe Val Lys Asp Ile Ile

85 90 95

Leu His Ile Lys Lys Leu Tyr Arg Glu Gly Arg Phe Asn

100 105

Claims (6)

1. A method of identifying a polypeptide capable of binding to a preselected epitope of a target protein, comprising the steps of:

(a) providing a library of phage particles expressing polypeptides on the surface of phage particles,

(b) preparing a decoy protein having a change in the amino acid sequence of about 5 to 10 within a linear fragment of about 5 to 15 amino acid residues of a target protein corresponding to a preselected epitope of the target protein,

(c) incubating the library of phage particles with the target protein to select for phage particles having a polypeptide capable of binding to the target protein,

(d) adding excess molarity trap protein as competition agent to select out bacteriophage particle specific to pre-selected epitope,

(e) separating phage particles bound to the target protein from phage particles bound to the decoy protein,

(f) recovering the phage particles bound to the target protein.

2. The method of claim 1, wherein the decoy protein is used to select phage particles from the library for testing that have polypeptides capable of binding to the decoy protein and then to reject polypeptides that bind to the target protein.

3. The method of claim 1, wherein the target protein is used to select phage particles from the library that have polypeptides capable of binding to the target protein and then to reject those polypeptides that bind to the decoy protein.

4. The method of claim 1, wherein the polypeptide capable of binding to a preselected epitope is an antibody or antibody fragment.

5. The method of claim 4, wherein the polypeptide is an antibody fragment, including a Fab, Fab 'or F (ab') 2 fragment, or a derivative of such fragments.

6. The method of claim 5, wherein the antibody is a monoclonal antibody as a surrogate antibody that binds to a similar epitope on murine tissue factor.