JP2016123342A - Vegf-binding peptides - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide VEGF-binding peptides useful for treating diseases accompanied with angiogenesis such as malignant tumor.SOLUTION: Provided is a VEGF-binding peptide comprising an amino acid sequence of three finger-like scaffold, in which the sequence of a region from the second amino acid from the C1-Cys toward the C-terminal side to the 7th amino acid from C2-Cys toward the N-terminal side in Loop I is PLTRVV, the sequence of a region from the second amino acid from C3-Cys toward the C-terminal side to the 4th amino acid from C4-Cys toward the N-terminal side in Loop II is HGDHHTLSEW, and the sequence of a region from the 4th amino acid from C5-Cys toward the C-terminal side to the first amino acid from C6-Cys toward the N-terminal side in Loop III is EEPTAHV.SELECTED DRAWING: Figure 16

Description

  The present invention relates to VEGF-binding peptides useful for the treatment of malignant tumors and other diseases associated with angiogenesis.

  Vascular Endothelial Growth Factor (VEGF) is a glycoprotein having a molecular weight of 40,000 to 45,000, and is a growth factor that promotes angiogenesis and angiogenesis that are mainly expressed in vascular endothelial cells (non-) Patent Document 1). In particular, it is essential for angiogenesis during the embryonic period and plays an important role in the construction of many tissues such as the circulatory system. In adults, it is important not only for luteinization and placenta formation, but also for wound healing. On the other hand, it is known to play a central role in pathological angiogenesis as seen in solid cancer, rheumatoid arthritis, diabetic retinopathy and the like. In particular, in solid cancer, it is known that cancer cells overexpress VEGF, promote angiogenesis for supplying nutrients necessary for proliferation, and metastasize through the blood vessels (Non-patent Document 2). ~ 4 etc.).

In addition, antitumor effects of anti-VEGF neutralizing antibodies have been known from previous administration experiments to tumor-bearing mice transplanted with various cancer cells (Non-patent Document 5), and is an intraocular disease of an ischemic retinal disease model. An angiogenesis inhibitory effect (Non-patent Document 6) is also known. Therefore, in recent years, VEGF inhibitors that inhibit the binding of VEGF and its receptors, especially anti-VEGF antibodies, are expected to have a powerful inhibitory effect on pathological angiogenesis and an effective anticancer agent, and are actively researched and developed. Has been made.
In order to reduce the antigenicity of anti-VEGF mouse antibody, in addition to the development of human-derived anti-VEGF human antibody (Patent Document 1), anti-VEGF mouse antibody chimerization (Non-Patent Document 7), anti-VEGF mouse antibody CDR In addition, a humanized antibody using non-patent document 8 (non-patent document 8) and a method for immunization using a transgenic mouse producing a human antibody itself (non-patent document 9) have also been proposed. The humanized anti-VEGF antibody has already been put to practical use as an anticancer agent for metastatic colorectal cancer and metastatic breast cancer under the name bevacizumab (Genentech, registered trademark Avastin). The affinity Fab fragment of bevacizumab is also sold for the treatment of intraocular neovascularization by Rabinizumab (Genentech, registered trademark Lucentis). Furthermore, research and development to acquire anti-VEGF antibodies with higher efficacy and fewer side effects by introducing mutations into a part of the CDR region using phage display technology etc. based on bevacizumab and other humanized antibodies Is also actively carried out (Patent Documents 2 to 4).

  However, in the case of a humanized antibody, all the concerns about side effects based on antigenicity (Non-Patent Document 10) have not been solved. Antibodies are proteins with large molecular weight (150 kDa) that cannot work in cells, and post-translational modifications that involve glycosylation and disulfide bond formation are not suitable and are not suitable for mass production in microorganisms. There are problems such as high costs. In particular, when used for the treatment of solid cancer, it has been reported that polypeptides having a large molecular weight have poor invasion into affected areas (Non-patent Document 11).

  The extracellular domain of the VEGF receptor has a structure in which a plurality of immunoglobulin-like domains are linked, and specifically binds to VEGF with high affinity. Moreover, because it is originally a human-derived polypeptide and is considered to have low antigenicity, soluble VEGF receptor extracellular domain is expected to be an excellent VEGF inhibitor, but it is rapidly metabolized in the body. Not practical. Therefore, by constructing a fusion polypeptide with an antibody molecule using a partial region such as the first to third immunoglobulin-like domains instead of the full length of the extracellular domain, the affinity with VEGF is not impaired in the blood. VEGF-binding polypeptides having an extended half-life have been provided (Patent Documents 5 to 7). However, these VEGF-binding polypeptides have an amino acid residue of 200 or more even in the extracellular domain partial region polypeptide alone, and the molecular weight of the antibody becomes higher when used as a fusion protein. It has the same problem as the case.

In general, the probability that a molecule has antigenicity (immunogenicity) is said to decrease with decreasing molecular size. Antigenicity appears in molecules of 10 kDa or more, and those containing heterologous proteins as structural elements Risk increases. With the expectation of such a decrease in antigenicity, research and development of a VEGF inhibitor having a smaller molecular weight is being actively pursued.
SU6668 (Non-patent document 12), PTK787 / ZK222584 (Non-patent document 13), etc. have been reported as orally administrable low molecular weight compounds exhibiting VEGF inhibitory action, and patent applications relating to quinoline-indeneindoline derivatives have been made (patents) Literature 8-10).
In addition, peptide pharmaceuticals have a tendency to reduce the molecular weight, and research and development using non-immunoglobulin scaffolds having a small molecular weight are also active. Traditionally, like an immunoglobulin-like domain, a natural non-immunoglobulin domain structure having a plurality of rigid frameworks and a loop region that specifically binds to a target substance arranged on the framework is utilized. A variety of peptide libraries in which loop regions are randomized are known. Several attempts have been made to search for VEGF-binding peptides using these non-immunoglobulin scaffold peptide libraries, and include a type III fibronectin domain (FN3) structure (Patent Document 11) and a helix-loop-helix structure (Patent Document 12). VEGF-binding peptides having) have been proposed. FN3 is a scaffold consisting of about 100 amino acids that does not contain cysteine in the molecule, so it has the advantages of being highly productive in E. coli and stable to heat and protease, but even with the lowest dissociation rate constant (Kd value) of 0.1 It has a VEGF binding activity on the order of μM, which is very low compared to anti-VEGF antibody Avastin (Kd = 1.1 nM). On the other hand, in the case of a VEGF-binding peptide having a helix-loop-helix structure, a peptide showing Kd = 2.9 nM (Patent Document 12, Table 18) can be obtained by looking at the binding activity alone, The peptide alone does not show VEGF inhibitory activity or cell growth inhibitory action. In order to obtain a VEGF inhibitory activity and a cell growth inhibitory effect, it is necessary to construct a fusion protein with thioredoxin (105 amino acids, molecular weight 12 kDa). Therefore, it is considered that the merit of low molecular weight cannot be fully utilized.

  Therefore, it is desired to provide a highly stable peptide having a low molecular weight and a high half-life in serum, which is a low molecular weight VEGF binding peptide having a high VEGF binding activity and a VEGF inhibitory activity alone. It was.

JP 9-316099 A WO2010 / 148223 WO2009 / 073160 WO98 / 45331 JP 9-255700 A JP-A-9-154588 WO98 / 31794 JP-A-2005-145928 JP-A-2005-298437 WO2002 / 094809 WO2009 / 023184 JP2014-47156 Patent No. 4959226

Science, 1989 Dec 8; 246 (4935): 1306. Cancer Res., 1995 Sep 15; 55 (18): 3964-8. Cancer, 1996 Mar 1; 77 (5): 858-63. Clin Cancer Res., 1997 Jun; 3 (6): 861-5. Nature, 1993 Apr 29; 362 (6423): 841-4. Arch. Ophthalmol., 114: 66-71 (1996). S. L. Morrison et al., Proc. Natl. Acad. Sci. U. S. A., 81: 6851 (1989). Cancer Res., 1997 Oct 15; 57 (20): 4593-9. S. Wagner et al., Nucleic Acid Res., 22: 1389 (1994). Blower 2009 Br. J. Nurs., 18 (6): 351-358. Br. J. Cancer, 1994 Sep; 70 (3): 521-5. Cancer Res., 2000 Aug 1; 60 (15): 4152-4160. Cancer Res., 2000 Apr 15; 60 (8): 2178-2189. Naimuddin, M., et al, Mol. Brain 2011, 4, 2. Kubo, T., et al., (2002), Program No. 137. 16., 32nd Annual Meeting of Society for Neuroscience, Washington, DC, 3rd November, 2002. Chen et al. JMB (1999) 293: 865-881. Yamaguchi, J., et al, Nucleic Acids Res. 2009, 37, e108. J Biol Chem. 1991 Jun 25; 266 (18): 11947-54.

  The present invention provides a highly stable peptide having a low molecular weight, having a high VEGF binding activity and a VEGF inhibitory activity, low immunogenicity, and a long half-life in serum. It is what.

The present inventors searched for VEGF-binding peptides using cDNA display technology using a three-finger-like scaffold (3F) peptide library that we previously developed for screening protein binding to target substances (patents) I recalled using Document 13 and Non-Patent Document 14).
The three-finger-like scaffold (3F) derived from snake venom is composed of 60-70 amino acids and forms 4-5 disulfide bonds, 3-5 antiparallel beta sheets, and a three-finger structure It is a very stable structure consisting of three protruding loops. In particular, it shows thermal stability that can withstand temperatures of 60 to 70 ° C. In addition, α-type neurotoxin, which is representative of 3F, binds to nicotinic acetylcholine receptor (nAChR) and acetylcholine binding protein (AChBP) and has been observed to inhibit ligand binding to both. In addition to its small size, it is a snake venom-derived peptide that is hydrophilic and resistant to proteases, so it is more difficult to recognize as an antigen and is also less prone to protease degradation in serum, so it has a longer half-life in serum. It has the characteristics.
In the previous patent document 13 using a 3F peptide library, three types of highly binding activity peptides having a Kd value of 55 to 115 nM were obtained as IL6R binding peptides, and at the same time, these peptides did not form a fusion with thioredoxin. However, it has been confirmed that all have high inhibitory activity (IC ₅₀ = 105 to 250 nM).

The present inventors have developed a new puromycin linker for use in cDNA display while conducting research and development of an efficient cDNA display method. The resulting puromycin linker (Linker-N) has high translation efficiency in a cell-free translation system, and can synthesize a high-yield and high-yield cDNA display, so that an efficient cDNA display method can be established. did it. The linker-N and a cDNA display method using the linker-N have been applied for a patent on the same date as the present application.
In the present invention, the cDNA display method is applied to the 3F peptide library developed by the present inventors, and Kd = 2.1 nM after seven selections using biotinylated VEGF on the surface of streptavidin beads. A high binding activity VEGF-binding peptide could be obtained with a high probability of 60%.
By obtaining the above knowledge, the present invention could be completed.

That is, the present invention is as follows.
[1] In the amino acid sequence of the three-finger-like scaffold (3F),
(I) the sequence of the region from the second amino acid on the C-terminal side of C1-Cys to the seventh amino acid on the N-terminal side of C2-Cys contained in Loop I is “PLTRVV (SEQ ID NO: 6)”;
(Ii) the sequence of the region from the second amino acid on the C-terminal side of C3-Cys to the fourth amino acid on the N-terminal side of C4-Cys contained in Loop II is “HGDHHTLSEW (SEQ ID NO: 7)”; iii) An amino acid sequence in which the sequence from the fourth amino acid on the C-terminal side of C5-Cys to the first amino acid on the N-terminal side of C6-Cys contained in Loop III is “EEPTAHV (SEQ ID NO: 8)” A VEGF-binding peptide comprising.
[2] A VEGF-binding peptide comprising the amino acid sequence represented by SEQ ID NO: 16.
[3] A nucleic acid encoding the amino acid sequence according to [1] or [2].
[4] A VEGF-binding peptide derived from the amino acid sequence of three-finger-like scaffold (3F) and comprising any one of the following amino acid sequences (1) to (3);
(1) The amino acid sequence forming loop I, and the sequence of the region from the second amino acid on the C-terminal side of C1-Cys to the seventh amino acid on the N-terminal side of C2-Cys contained in loop I is “ PLTRVV (SEQ ID NO: 6) ", an amino acid sequence in which Cys that does not participate in cyclization is substituted with another amino acid,
(2) The amino acid sequence forming loop II, and the sequence of the region from the second amino acid on the C-terminal side of C3-Cys to the fourth amino acid on the N-terminal side of C4-Cys contained in loop II is “ HGDHHTLSEW (SEQ ID NO: 7) ", an amino acid sequence in which Cys not involved in cyclization is substituted with another amino acid,
(3) The amino acid sequence forming loop III, and the sequence of the region from the fourth amino acid on the C-terminal side of C5-Cys to the first amino acid on the N-terminal side of C6-Cys contained in loop III is “ An amino acid sequence which is “EEPTAHV (SEQ ID NO: 8)”.
[5] The amino acid sequence forming loop I of (1) is the amino acid sequence shown in SEQ ID NO: 17, and the amino acid sequence forming loop II of (2) is the amino acid sequence shown in SEQ ID NO: 18. The VEGF-binding peptide according to [4] above, wherein the amino acid sequence forming the loop III of (3) is the amino acid sequence shown in SEQ ID NO: 19.
[6] A nucleic acid encoding the amino acid sequence according to [4] or [5].

  The VEGF-binding peptide of the present invention is a peptide that is stable against heat and protease degradation, and is considered to have no antigenicity and a long serum half-life. In addition, it has a binding activity (Kd = 2.1 nM) that is comparable to that of anti-VEGF antibody (Kd = 1.1 nM), and forms a fusion with thioredoxin in view of the characteristics of the 3F scaffold, which is the basic skeleton. It can be expected to have high VEGF inhibitory activity alone.

The structure of linker-N of the present invention, and a method for producing the same. (A) Four functional properties of Linker-N; ligation site for ligation with T4 RNA ligase, puromycin site for covalent binding of the synthesized protein, spacer sequence and oligo dA site, and synthetic cDNA Priming site for reverse transcription. (B) A method for producing linker-N. C in the figure indicates 5-Me-dC. Schematic diagram of (1) mRNA protein fusion and (2) cDNA protein fusion (cDNA display) formation. A gene construct introduced into a cDNA display. In the figure, SP6 represents the SP6 promoter and cap site, UTR represents the 5 ′ untranslated region of Xenopus β-globin, Spc represents the spacer sequence, and the Y-tag sequence has a sequence for further hybridization with the linker at the 3 ′ end. Including. Ligation reaction with PDO (3F, 6R14, BDA) mRNA and linker-N (ratio 1: 1). (A) FITC fluorescence detection after ligation of PDO-mRNA and linker-N. In the figure, M represents a size marker, and 10 min, 20 min, and 40 min represent ligation reaction times, respectively. (B) Detection of all nucleic acid components by the Sybr Gold. (C) Sybr Green staining of all nucleic acid components after ligation of 3F, 6R14 and BDA mRNA and linker-N for 10 minutes, 3F on the left in the figure is 3F-mRNA used as a size marker. Ligation efficiency with PDO (3F, 6R14, BDA) mRNA and linker-N Examination of the length of oligo dA in linker-N. (A) Linker-N (18 A residues) (b) Linker-N-10A (10 A residues) (c) Linker-N-0A (0 A residues). Comparison of the ability to form protein fusions based on the length of oligo dA in the rabbit reticulocyte translation system. (A) Electrophoretic analysis (Fluorescein fluorescence image). (B) Comparison of fusion rate (%). Comparison of the ability to form protein fusions with different oligo dA lengths in the wheat germ extract translation system. (A) Maturation effect in the presence of a high concentration salt on the ability to form a protein fusion when linker-N is used. In the figure, 0: post-translational sampling for 10 minutes, 1; maturation for 1 hour in the absence of high concentration of salt, 2; maturation for 30 minutes in the presence of high concentration of salt, 3; maturation for 1 hour in the presence of high concentration of salt The fractions obtained are shown. (B) The ability to form a protein fusion when linker-N-10A and linker-N-0A are used. (C) Comparison of fusion rate (%). The conceptual diagram which shows the effect of presence of oligo dA in a linker. (A) Ribosome size. (B) When there is no oligo dA (linker-N-0A). (C) When oligo dA is present (linker-N). (D) A schematic diagram showing the effect of the “stem” structure of the oligo dA in the linker-N increasing the range of motion of the spacer flexible structure in response to a change in the position of the ribosome in the solution after protein synthesis. Purification of cDNA protein fusion using His-tag. In the figure, F.T. represents the flow-through fraction, E1 represents eluate 1 and E2 represents eluate 2, and PDO corresponds to the mRNA protein (PDO) fusion that has not undergone reverse transcription. The mRNA protein fusion is slower to migrate than the cDNA protein fusion. Therefore, the PDO lane is shifted downward for comparison of concentration. Demonstration of high-throughput cDNA display using linker-N and its in vitro selection. (A) Schematic diagram showing essential steps and required time for functioning a cDNA display using wheat germ lysate. Transcription of template DNA into mRNA, purification and quantification (˜1.5 hours). Ligating mRNA to linker-N (0.2 hours). A process of translation into a wheat germ lysate system and maturation under high salt conditions to form a display protein (mRNA protein fusion) (˜1 hour). Step of elution after purification of mRNA protein fusion on Ni-NTA magnetic beads, reverse transcription, and purification of cDNA protein fusion (0.5 hours). A process of selecting and elution from a target protein immobilized on a streptavidin (SA) matrix, and amplifying DNA corresponding to the target protein (˜2 hours). PCR amplification analysis step (~ 1 hour). The whole process can be completed in 6-8 hours. (B) In vitro selection using PDO and 3F, 6R14, BDA display proteins. Bind 20: 1 mixture of non-binding PDO and binding BDA or 3F or 6R14 to (1) IgG binding to BDA, (2) AChBP binding to 3F, and (3) IL6R binding to 6R14 The selection fraction using the SA beads was subjected to electrophoresis after PCR. (C) The selection efficiency was expressed as “the number of fold enrichments” from the change in the ratio before and after the selection of each selection fraction obtained in (b). FLAG-tag sequence selection from 8-mer peptide library. (A) DNA construct for preparation of 8-mer peptide library. (B) Selection cycle. An 8-mer peptide library complex mixture containing 0.1% FLAG-tag sequence is subjected to a selection cycle process of transcription, ligated to linker-N, translated in WGE to synthesize mRNA fusion protein, and His-tag The cDNA fusion protein is synthesized after purification on the solid phase using reverse transcription and reverse transcription, followed by selection, selection on an anti-FLAG M2 antibody matrix, and PCR amplification. The amplified product is further subjected to two cycles of the same process for cloning and sequencing. (C) Ratio of FLAG-tag sequence in the PCR amplification product after 3 selection cycles. In “puro-Linker” shown as a comparison, the same composite mixture was applied to the selection procedure of Non-Patent Document 13 and selection was performed three times. Three finger (3F) structure and gene construct of 3F library. Binding assay of cDNA protein fusion (R7) after 7 rounds of in vitro selection of VEGF affinity protein using 3F library targeting VEGF. (A) (1) 1/3 fraction of R7 to streptavidin beads (SA), (2) 1/3 fraction of R7 to VEGF immobilized beads, (3) R7 cDNA (R7 not translated) (4) After binding 1/3 fraction of R7 to Fc-immobilized beads, the eluted fraction was PCR amplified and analyzed by gel electrophoresis. . (B) ELISA was applied to each of the elution fractions after the binding steps (1), (2) and (4), and binding to beads was detected at a 450 nm wavelength absorption. Evolutionary engineering process with a new linker that enables shorter time The fully automated high-throughput system for finding lead compounds for drug discovery is about 8 hours per round, and lead compounds in about 2 weeks even if 8-10 rounds are performed That you can get. Estimated 3D structure of VEGF-binding peptide

1. About the VEGF-binding peptide of the present invention (1-1) VEGF-binding peptide having the 3F skeleton of the present invention The VEGF-binding peptide of the present invention is a 3F random using “three-finger-like scaffold (3F)” described later. Since it is a peptide molecule evolved by using an evolutionary engineering in vitro selection method for a peptide library, “three finger-like scaffold (3F)” described in (Patent Document 13, Non-Patent Document 14) and It has the same basic skeleton. The amino acid sequences of specific portions of the protruding loop regions Loop I, Loop II and Loop III are “PLTRVV (SEQ ID NO: 7)”, “HGDHHTLSEW (SEQ ID NO: 8)” and “EEPTAHV (SEQ ID NO: 9)”, respectively. is there.
That is, one embodiment of the VEGF-binding peptide of the present invention is a VEGF-binding peptide having a 3F skeleton and can be expressed as follows.

In the amino acid sequence of the three finger-like scaffold (3F),
(I) the sequence of the region from the second amino acid on the C-terminal side of C1-Cys to the seventh amino acid on the N-terminal side of C2-Cys contained in Loop I is “PLTRVV (SEQ ID NO: 6)”;
(Ii) the sequence of the region from the second amino acid on the C-terminal side of C3-Cys to the fourth amino acid on the N-terminal side of C4-Cys contained in Loop II is “HGDHHTLSEW (SEQ ID NO: 7)”; iii) An amino acid sequence in which the sequence from the fourth amino acid on the C-terminal side of C5-Cys to the first amino acid on the N-terminal side of C6-Cys contained in Loop III is “EEPTAHV (SEQ ID NO: 8)” A VEGF-binding peptide comprising.
Here, in the amino acid sequence of the three finger-like scaffold (3F), the conserved cysteines that form disulfide bonds are referred to as C1-C8 in order from the N-terminal side, between C1 and C2, between C3 and C4, And the region between C5 and C6 is referred to as Loop I, Loop II, and Loop III, respectively.

Three-finger-like scaffold (3F) is commonly found in some α-neurotoxins derived from snake venom, as described in (Patent Document 13 and Non-Patent Document 14). It is a structure having a small size of 60 to 70 amino acids, including 4 to 5 disulfide bonds, 3 to 5 antiparallel β-sheets, and 3 protruding loops forming a three-finger structure. Extremely stable. In particular, it has high thermal stability and can withstand 60 to 70 ° C. (Biochemistry. 2000 Aug 1; 39 (30): 8705-10). In addition, since it is small in size, it is unlikely to be recognized as an antigen and prone to protease degradation, so it is predicted that the half-life in serum is long.
In the present invention, when referring to “3F peptide library” or “3F random peptide library”, loop I and loop of three-finger-like scaffold (3F) constructed in (Patent Document 13, Non-Patent Document 14) A random peptide library in which amino acids in specific regions of II and Loop III are simultaneously randomized. Typically, a library in which each peptide constituting a 3F peptide library exists in a form corresponding to a polynucleotide (cDNA) encoding each peptide so that it can be used as a combinatorial library (Also referred to as cDNA display). However, a 3F polynucleotide library composed of mRNA or cDNA encoding the randomized 3F peptide library used to produce the cDNA display is referred to as a “3F peptide library”.
Methods for randomizing specific regions of Loop I, Loop II, and Loop III in the “3F peptide library” are known and described in detail in, for example, (Patent Document 13 and Non-Patent Document 14). Yes.

Even though the three-finger-like scaffold (3F) is composed of different amino acids, the three-dimensional structure is highly conserved. For example, there are various types listed in [Table 1] of (Patent Document 13). Although it may be a library constructed using a three finger-like scaffold (3F) derived from the origin organism, a CTx3 peptide library derived from Micrurus corallinus represented by the following (formula 1) (SEQ ID NO: 15) is preferred.
(Formula 1)
LVCY (X) ₆ PGTLETCPDDFTCV (X) ₁₀ TQYCSHACAIP (X) ₇ CCQTDKCNG (SEQ ID NO: 15)
(In the formula, X is an arbitrary amino acid residue.)

That is, the most preferred amino acid sequence of the VEGF-binding peptide of the present invention is represented by the following (Formula 2).
(Formula 2)
LVCYPLTRVVPGTLETCPDDFTCVHGDHHTLSEWTQYCSHACAIPEEPTAHVCCQTDKCNG (SEQ ID NO: 16)

(1-2) Optimization of VEGF-binding peptide size The three-finger-like scaffold (3F) used in the present invention is composed of a total length of 61 amino acid residues in the case of 3F derived from CTx3, for example. There are three loops, loop I, loop II, and loop III, and each loop has a structure that is firmly fixed by an SS bond in the terminal region. Therefore, the peptide “1F scaffold” having a small size of 12 to 25 amino acid residues cleaved leaving Cys for forming the terminal SS bond of each loop structure also retains the original three-dimensional structure. In Non-Patent Document 14, it is demonstrated that the inhibitory activity, agonist activity, and antagonist activity similar to those shown in the original 3F are maintained in the 1F peptide containing loop I.
The VEGF-binding peptide of the present invention does not need to be a full-length peptide that can constitute a 3F structure, but includes any one of loop I, loop II, and loop III, and can constitute at least a “1F scaffold”. If present, it has the same VEGF-binding activity as the full-length VEGF-binding peptide having a 3F structure and can be expected to have VEGF inhibitory activity. In addition to the ease of synthesis, the small size of the full-length peptide provides advantages such as small antigenicity and serum half-life.

From the above, the VEGF-binding peptide of the present invention is preferably optimized to further reduce the size, and other aspects of the present invention are any of the following (Formula 3) to (Formula 5) It can be said that the peptide comprises the amino acid sequence of In particular, it is a peptide containing any one of the amino acid sequences of (Formula 3) to (Formula 5), and more preferably a partial peptide of (Formula 2). At that time, among the amino acids constituting 1F, Cys that is not involved in cyclization is preferably changed to another amino acid. According to Non-Patent Document 14, (Formula 3) and (Formula 4) Cys is replaced with Gly. In addition, amino acids outside Cys at both ends used for cyclization, for example, “Leu-Val” on the N-terminal side and “Val” on the C-terminal side in (Formula 3) can be replaced with other amino acids.
(Formula 3)
LVCYPLTRVVPGTLET G PDDFTCV (SEQ ID NO: 17)
(Formula 4)
ETCPDDFT G VHGDHHTLSEWTQYCS (SEQ ID NO: 18)
(Formula 5)
ACAIPEEPTAHVC (SEQ ID NO: 19)

2. “VEGF” as a target substance for the “VEGF-binding peptide” of the present invention
VEGF (vascular endothelial growth factor) is a growth factor involved in angiogenesis, angiogenesis, etc. In mammals including humans, five types of VEGF-A, VEGF-B, VEGF-C, VEGF-D and PIGF Is known, and VEGF-A most potently enhances angiogenesis. Human VEGF-A consists of 189 amino acid residues in the full length of the mature body, but multiple isofarms are known based on different splicing of the VEGF-A gene, and VEGF consists of 165 amino acid residues lacking exon 6. 165 "has a biological activity equivalent to the full length" VEGF 189 "(Non-patent Document 18).
In addition, “VEGF 165” is a VEGF that is mainly expressed in the body and is known to have a large amount of expression in solid tumors. Therefore, “VEGF 165” is most suitable for screening of “VEGF-binding peptide” in the present invention. It can be said to be a preferred “target VEGF”. In the examples, commercially available VEGF (PeproTech, London) consisting of a dimer of “VEGF 165” shown in SEQ ID NO: 20 was used to biotinylate by a known method, and then applied to streptavidin-coated magnetic beads. Immobilized.

3. 3. Production method of 3F-cDNA protein fusion library used as a cDNA display by the cDNA display method of the present invention (3-1) Structure and synthesis method of linker-N The linker-N used in the present invention is constituted as follows. The As in the case of a normal T-shaped puromycin linker oligonucleotide, when a horizontal DNA strand is referred to as a main chain and a DNA strand having a vertical puromycin is referred to as a side chain, from the 5 ′ end side of the main chain, It consists of a “ligation site” with mRNA, a “hybridization region” with mRNA, and a “primer site” for reverse transcription, and the side chains from the base part to “oligo dA”, “spacer sequence” and puromycin The biotin molecule is not contained in the linker (FIG. 1a).
The “spacer sequence” of the side chain uses the “Sp.18 (Spacer 18 Phosphoramidite)” molecule represented by the chemical formula “PO ₄ (CH ₂ CH ₂ O) ₅ CH ₂ CH ₂ PO ₄ ” and binds FITC. The sequence of “(Sp.18 × 1) (T-FITC) (Sp.18 × 3)” was used.
The linker-N used in the present invention was synthesized by the following method.
“5′-CCCCCCCGCCGCCCCCCG (5-Me-dC) A18” described in FIG. 1B in which the 5′-ligation site / hybridization region and puromycin-oligo dA-containing spacer sequence are linked by 5-Me-dC. (Spec18) (Spec18) (Spec18) (F-dT) (Spec18) CC (Puro) -3 '"is synthesized with a DNA synthesizer or the like. Here, (F-dT) represents (T-FITC). After removing the levulinyl protecting group of 5-Me-dC, a primer sequence, for example, “5′-CCTTG-3 ′”, is linked to 5-Me-dC to easily produce the linker-N of the present invention. it can. The generation of “Linker-N” can be confirmed by TOF-MS and gel electrophoresis.

(3-2) 3F-mRNA-Linker-N Conjugate Next, a 3F-mRNA library to be linked to Linker-N is prepared as follows.
The template DNA includes a sequence encoding a DNA gene encoding a 3F polypeptide library (Patent Document 13), a sequence necessary for promoting a translation reaction, a His tag sequence, and other sequences. The terminal side includes a primer site corresponding to the main chain of linker-N and a base sequence complementary to the hybridization region (FIG. 3).
The mRNA and linker are linked by ligation using T4 RNA ligase after annealing by hybridization by a known method.

(3-3) Preparation of 3F-mRNA protein fusion The resulting mRNA-linker-N conjugate was contacted with wheat germ extract (WGE) to synthesize proteins in ribosome A, The protein translated via puromycin and the mRNA-linker-N conjugate are fused to construct an mRNA-linker-N-protein fusion (hereinafter also referred to as 3F-mRNA protein fusion).
As a wheat germ extract (WGE), a kit commercially available from Promega (Madison, Wis., USA) can be used.

(3-4) Production and purification of 3F-cDNA-linker-N-protein fusion (hereinafter referred to as 3F-cDNA protein fusion) Purification of mRNA protein fusion of the present invention, cDNA protein fusion by reverse transcription reaction The same commercially available Ni-NTA magnetic beads (Qiagen) are all used for the preparation of and the purification of the obtained cDNA protein fusion.
In the examples of the present invention, mRNA / cDNA protein fusion is used as it is without degrading mRNA. However, mRNA may be used after degrading with RNaseH (Ambion, Austin, TX, USA), etc. The case can be decomposed on Ni-NTA magnetic beads. In general, the term “cDNA display” refers to both cases of mRNA / cDNA protein fusion and cDNA protein fusion. Therefore, in the present invention, they are referred to as cDNA protein fusion without distinguishing both.
The obtained random peptide library of 3F-cDNA protein fusion is used as a cDNA display, and is applied to the cDNA display method and subjected to evolutionary engineering screening.

4). In vitro selection by cDNA display method used in the present invention (4-1) Preparation and immobilization of “VEGF” as target substance Using random peptide library of 3F-cDNA protein fusion by cDNA display method of the present invention The prepared cDNA display is screened based on the strength of interaction with VEGF, and VEGF high affinity protein is screened at high throughput.
Here, since the target substance “VEGF” needs to be able to obtain a peptide that binds to VEGF and inhibits binding to the VEGF receptor, it is the above-mentioned human VEGF, particularly human VEGF-A, and is VEGF-specific. A typical epitope sequence, in particular, an amino acid sequence comprising part or all of a region involved in binding to a VEGF receptor. These are simply referred to as “VEGF” in the present invention. In this example, commercially available human VEGF 165 (PeproTech, London) consisting of 165 amino acid residues represented by typical SEQ ID NO: 20 was used.
In the present invention, the binding reaction with the cDNA display of the present invention immobilized on a solid surface such as magnetic beads is observed in an aqueous solution.
For binding between the target substance and the magnetic beads, biotinylated VEGF is bound using the affinity with streptavidin previously bound to the magnetic beads.

As a method for biotinylating VEGF, a known method can be applied. A base capable of binding biotin (for example, deoxythymine (dT)) or a base to which biotin is bound (for example, biotin-deoxythymine (Biotin-dT)), amino Base modified by a group (eg, amino-modified deoxythymine (eg, Amino-Modifier C6-dT: manufactured by Glen Research Search)), base modified by a carboxy group (eg, carboxy-modified deoxythymine (Carboxy-dT) ), A base modified with a thiol group (for example, thiol-modified deoxythymine (4-Thio-dT)) and the like. Commercially available EZ-Link Sulfo-NHS-SS-biotin (Pierce, Rockford, USA) can also be used.
Magnetic beads bound with streptavidin can be prepared by coating streptavidin on the surface of beads such as Ni-NTA magnetic beads by a known method, but commercially available streptavidin-coated beads (Magnotex-SA (Takara Bio) or Magnosphere ^™ ) MS300 / Streptavidin (Takara Bio) may be used.

(4-2) Measurement Method of Interaction and Identification Method of Obtained Display Protein In the cDNA display method of the present invention, whether or not the displayed protein and VEGF interact with each other is measured by both molecules. This is done by measuring and detecting changes in the signal generated based on the interaction between the two.
Examples of such measurement techniques include surface plasmon resonance (Cullen DC, et al., Biosensors, 3 (4), 211-225 (1987-88)), evanescent field molecular imaging (Funatsu T., et al. al., Nature, 374, 555-559 (1995)), fluorescence imaging analysis, Enzyme Linked Immunosorbent Assay (ELISA): Crowther JR, Methods in Molecular Biology, 42 (1995)), fluorescence Depolarization (Perran J., et al., J. Phys. Rad., 1, 390-401 (1926)), and Fluorescence Correlation Spectroscopy (FCS): Eigen M., et al., Proc. Natl. Acad. Sci. USA, 91, 5740-5747 (1994)).

  In the cDNA display method of the present invention, identification of a display protein in a protein-VEGF conjugate determined to interact with VEGF is performed by PCR amplification of the cDNA in the cDNA protein fusion and analysis of the base sequence. . Identification of a target substance can be performed using NMR, IR, various mass spectrometry, etc., using a label that has been labeled in advance. If the labeling substance is a protein, it can also be performed with a normal amino acid sequencer.

(4-3) Whole process diagram of in vitro selection (reference)
In vitro selection using the cDNA display method using the linker-N of the present invention is not limited to the following procedure, but will be described with reference to FIG. 11a.
(A) Transcribing template DNA (for example, random peptide library) to mRNA, purifying it, and quantifying it as appropriate (mRNA library): ~ 1.5 hours.
(B) Ligating mRNA to linker-N (mRNA-linker-N): 0.2 hours.
(C) Process of maturation under high salt condition (mRNA-protein fusion) after translation in wheat germ lysate translation system: ~ 1 hour.
(D) A step of purifying an mRNA protein fusion on Ni-NTA magnetic beads, performing reverse transcription, and purifying the cDNA protein fusion and then eluting it (cDNA protein fusion): 0.5 hours.
(E) A step of binding a cDNA protein fusion to a target protein immobilized on SA beads or the like, washing, and eluting: ~ 2 hours.
(F) PCR amplification of the eluted cDNA protein fusion cDNA and determining the DNA sequence, or preparing the template DNA and subjecting it to step (a) again:
The whole process can be completed in 6-8 hours.

EXAMPLES The present invention will be described below in more detail with reference to examples, but the present invention is not limited to these examples.
Other terms and concepts in the present invention are based on the meanings of terms that are conventionally used in the field, and various techniques used to implement the present invention include those that clearly indicate the source. Except for this, it can be easily and reliably carried out by those skilled in the art based on known documents and the like. In addition, various analyzes were performed by applying the methods described in the analytical instruments or reagents used, kit instruction manuals, catalogs, and the like.
In addition, the description content in prior art documents, patent publications, and patent application specifications cited in the present specification shall be referred to as the description content of the present invention.

Reference Example 1 Synthesis of Nucleic Acid Linker of the Present Invention (Linker-N) 5′-CCCCCCCGCCGCCCCCCG (5-Me-dC) A18 (Spec18) (Spec18) (Spec18) (F- A puromycin-oligonucleotide consisting of dT) (Spec18) CC (Puro) -3 ′ was synthesized by DNA synthesizer ABI394 (Applied Biosystem, Japan) outsourced to BEX (Japan). The 5 ′ end of the puromycin-oligonucleotide was protected by acetylation. In the formula, the symbol (5-Me-dC) represents 5′-dimethoxytrityl-N4- (O-levulinyl-6-oxyhexyl) -5-methyl-2′-deoxycytidine (Glen Research). (Spec18) indicates C18 spacer phosphoramidite, (F-dT) indicates fluorescein-dT, and (Puro) indicates puromycin CPG.
The levulinyl protecting group of 5-Me-dC was removed with 0.5 M hydrazine hydrate in 1: 1 pyridine / acetic acid and the column was further washed with pyridine / acetic acid (1: 1) and then with acetonitrile. Primer sequence 5′-CCTTG-3 ′ was linked to 5-Me-dC in puromycin-oligonucleotide to branch the oligonucleotide chain. The resulting puromycin-branched oligonucleotide was cleaved from the column with K ₂ CO ₃ in methanol to deprotect the 5 ′ terminal acetyl group. In order to complete this deprotection, the reaction was carried out in 25% ammonium hydroxide at 65 ° C. for 1.5 hours and purified by reverse phase HPLC to purify the puromycin-linker oligonucleotide of the present invention (hereinafter “linker- N ”) was obtained (FIG. 1). The generation of “Linker-N” was confirmed by TOF-MS and gel electrophoresis. All phosphoramidite reagents used for modification were obtained from Glen Research (Sterling, VA, USA).

Reference Example 2 Ligation Efficiency of Linker-N (2-1) Preparation of Template DNA Construct In order to examine the ligation efficiency of Linker-N, the template DNA of mRNA linked to Linker-N was used as a known protein A B- Domain (BDA: Gouda, H., et al, Biochemistry 1992, 31, 9665.), Oct-1 Pou domain (PDO: Dekker, N., et al, Nature 1993, 362, 852-855.), Three It was prepared using DNA encoding a finger (3F: Non-Patent Document 14) and an IL6 receptor (6R14: Non-Patent Document 14).
“BDA (B-domain of protein A)” is amplified from the pEFZZ18 protein A fusion vector (GE Healthcare) described in Non-patent document 20, and “PDO (Pou domain of Oct-1)” is non-patent document. Synthesized according to 21. According to Non-Patent Document 14, “3F (three finger)” is cloned from the South American coral snake Micrurus coralinus, and “6R14” (3F peptide having binding activity with human IL6 receptor α) is described in Non-Patent Document 22. PBADThio / TOPO vector IL6R10-14.
The 5 ′ terminal side of these DNAs is linked to “an oligonucleotide containing SP6 promoter / cap site / Xenopus β-globin 5 ′ untranslated region (UTR)” and a translation initiation codon (ATG) described in Non-Patent Document 22, and On the 3 ′ end side, “spacer (G3S) ₂ / His-tag sequence (6 × His) / spacer (G3S) / Y-tag sequence” was added and amplified by PCR to obtain a template DNA construct ( FIG. 3). Here, the spacer (G3S) is “GGGS (G = glycine; S = serine)”. PCR amplification was performed with a 20 second denaturation (94 ° C) step, a 15 second annealing (60 ° C) step and a 30 second extension step (72 ° C).

(2-2) Four types of template DNA constructs obtained by ligation of linker-N and various mRNAs (2-1) were respectively performed at 37 ° C. in RiboMAX Large Scale Production System (Promega, Madison, WI, USA). Transcripted with SP6 RNA polymerase for 1 hour to produce capped transcripts and purified with RNeasy purification kit (Qiagen) to obtain BDA-mRNA, PDO-mRNA, 3F-mRNA, 6R14-mRNA, with absorbance at 260 nm Quantified.
Next, each mRNA and linker-N were annealed at a ratio of 1: 1 (50 pmol) at 94 ° C. for 15 to 30 minutes, and then gradually cooled to 4 ° C. Ligation was performed at 25 ° C. for 10 minutes, 20 minutes and 40 minutes using 3U T4 kinase (NEB, USA) and 20U T4 RNA ligase (Takara, Japan). The formation of an RNA-linker complex can be confirmed by degrading mRNA hybridized with DNA with RNase H (Ambion, Austin, TX, USA) and performing electrophoresis on a denaturing polyacrylamide gel.
Fig. 4 shows an example of ligation reaction of PDO-mRNA and linker in the presence or absence of RNase H for 10 minutes, 20 minutes, and 40 minutes, and ligase reaction for 10 minutes. These were separated by denaturing polyacrylamide gel electrophoresis (6% gel and 8M urea). The results were obtained by using a fluorescence image (FIG. 4a) by FITC incorporated in linker-N and a nucleic acid staining image (FIG. 4b) by Sybr Gold (Molecular Probes, USA) with fluoroimagers (Bio-Rad, Hercules, CA, USA). FIG. 4c shows similarly the ligation result after 10 minutes of 3F-mRNA other than PDO, 6R14 mRNA and BDA mRNA and linker-N (Sybr Gold staining).
Next, the ligation efficiency of each was calculated by the following formula, and the following results were obtained: PDO 96 ± 4%, 3F 91 ± 5%, 6R14 95 ± 3%, BDA 96 ± 2% (FIG. 5).
Ligation% = (A / A + B) × 100
(A in the formula is the band intensity of the ligated product and B is the intensity of the remaining mRNA).
The above results indicate that Linker-N achieves both a low linker: mRNA ratio and high ligation efficiency compared to conventional nucleic acid linkers for cDNA display or mRNA display such as puromycin linker. Yes.

(Reference Example 3) Optimization of the length of “oligo dA” in linker-N from the viewpoint of protein translation efficiency in various cell-free translation systems (3-1) Synthesis of mRNA-linkers having different oligo dA lengths The length of oligo dA in N was changed to 18, 10, 0, and linker-N (18A), linker-N-10A (10A) and linker-N-0A ( 0A) was synthesized (FIG. 6). Using each linker, an mRNA (PDO, BDA, 3F, 6R14) -linker fusion was prepared for each linker in the same manner as in Reference Example 2 (2-2).

(3-2) Rabbit Reticulocyte Translation System Each mRNA (PDO, BDA, 3F, 6R14) -linker fusion is used as a rabbit reticulocyte lysate IVT kit (Ambion, Austin, TX, USA) (hereinafter referred to as “Rabbit Reticulocyte Lysate”). , And referred to as RRL), each mRNA was translated, and mRNA (PDO, BDA, 3F, 6R14) -protein fusion was prepared.
Specifically, according to Non-Patent Documents 13 and 22, 100 nM of mRNA-linker fusion was added to 25 μl of RRL reaction system (16.75 μl of lysate, 80 mM potassium acetate, 0.5 mM magnesium acetate, 10 mM creatine phosphate, 0.05 mM amino acid (methionine). And leucine were each 0.025 mM)) and the protein synthesis reaction was carried out at 30 ° C. for 10 minutes. The protein fusion (ie, covalent binding to the puromycin portion of the protein) was then matured in the presence of 65 mM MgCl ₂ and 750 mM KCl for 2 hours at 37 ° C. Sampling for electrophoresis analysis was performed at 10 minutes (FIG. 7a).
The efficiency of fusion formation was calculated according to the following formula, and the values for linker-N (18A), linker-N-10A (10A) and linker-N-0A (0A) for each mRNA were PDO (18A, 60 ± 5; 10A, 30 ± 2; OA, 20 ± 3), BDA (18A, 65 ± 3; 10A, 35 ± 3; 0A, 22 ± 2), 3F (18A, 50 ± 4; 10A, 30 ± 3) 0A, 18 ± 3) and 6R14 (18A, 55 ± 5; 10A, 35 ± 4; 0A, 20 ± 2).
Fusion formation% = (A / A + B) × 100
(Here, A is the band intensity of the protein fusion (mRNA-linker-protein), and B is the band intensity of the remaining mRNA-linker.)
The ability of each mRNA-linker to form a fusion is shown in FIG. 7b.
In any mRNA, it was found that the longer the oligo dA in the linker, the higher the translation efficiency.

(3-3) Wheat germ extract translation system About the fusion of each mRNA (PDO, BDA, 3F, 6R14) and linker-N, wheat germ extract (Wheat Germ Extract (Promega, Madison, WI, USA) ( Hereinafter, each mRNA was translated in the same manner as in (3-2) above, and an mRNA (PDO, BDA, 3F, 6R14) -protein fusion was prepared.
For translation in WGE, 100 nM of the mRNA-linker fusion was performed in 25 μl of WGE reaction system (12.5 μl of lysate, according to other product protocols), followed by protein synthesis reaction at 25 ° C. for 10 minutes, and then at 65 mM for 1 hour at 25 ° C. It matured protein fusions with the addition of KCl MgCl ₂ and 750 mM. In order to verify the effect of high-concentration salt, it was also carried out as a control when it was matured for 1 hour in the absence of high-concentration salt.
Electrophoretic analysis was sampled at 10 minutes (ie before the addition of high-concentration salt), 30 minutes and 1 hour (after addition of high-concentration salt) (FIG. 8a). In FIG. 8a, the 0 fraction represents the fraction sampled after 10 minutes of translation, the 1 fraction represents the control fraction cultured for 1 hour without high concentration salt, and the 2 fraction represents the high concentration salt and 30 fraction. The fractions and 3 fractions represent the fractions cultured for 1 hour, respectively. As a result, it can be seen that the addition of a high concentration of salt in the translation process is effective, and the one-hour maturation process is also effective.
Subsequently, using each mRNA (PDO, BDA, 3F, 6R14), a fusion of linker-N-10A and linker-N-0A, similarly in WGE at 25 ° C. for 10 minutes, Maturated for 1 hour in the presence (Figure 8b).
The percentage formation of each mRNA protein fusion was calculated according to the above formula, and the values for Linker-N (18A), Linker-N-10A (10A) and Linker-N-0A (0A) for each mRNA were PDO (18A , 90 ± 5; 10A, 65 ± 3; 0A, 30 ± 4), BDA (18A, 95 ± 4; 10A, 60 ± 6; 0A, 35 ± 4), 3F (18A, 85 ± 4; 10A, 60 ± 4; 0A, 30 ± 4) and 6R14 (18A, 90 ± 6; 10A, 60 ± 5; 0A, 35 ± 5) (FIG. 8c).

When the% of protein fusion formation in the case where WGE is selected as the translation system (FIG. 8c) and the case where RRL is used as the translation system (FIG. 7b) are compared, the value at linker-N is RRL. Is 50 to 65%, whereas WGE shows an extremely high value of 90 to 95%.
The difference is as much as one order in terms of the formation amount of 2.5 × 10 ¹⁵ per 1 ml (4 μM) and 1.2 × 10 ¹⁴ per 1 ml (0.2 μM) in the RRL translation system.
From these results, it has been demonstrated that the translation efficiency increases as the length of the oligo dA in the linker increases in the WGE translation, and at least 10 or more, preferably 12 or more dA (deoxyadenine) as the oligo dA. It turns out that the number is necessary.
It was confirmed that “oligo dA” in linker-N contributes to proper placement of puromycin near the P site of ribosome A via a flexible spacer sequence, as shown in FIG. 9d.

(Reference Example 4) mRNA (PDO, BDA, 3F, 6R14) each protein fusion purification and reverse transcription invention protein fusions, prior to reverse transcription (RT), the previously provided His-tag protein Utilizing this, it was fixed on Ni-NTA magnetic beads (Qiagen), which is a kind of IMAC (Immobilized metal affinity chromatography) carrier, at 25 ° C. for 10 minutes for purification.
The beads were washed with 1X PBS buffer, 20 U of M-MLV reverse transcriptase (Takara, Japan) was added, RT was performed at 42 ° C. for 10 minutes, and placed on ice.
After several washes, the fusion was eluted twice with 250 mM imidazole at 25 ° C. for 5 minutes.
Each of the obtained cDNA protein fusions was reacted with RNaseH (Ambion, Austin, TX, USA) at 37 ° C. for 30 minutes to decompose and remove mRNAs forming mRNA / cDNA duplexes. As an example, the case of using a PDO-mRNA protein fusion is shown in FIG. Each sample was separated by 12% SDS-PAGE containing 8M urea, and the band was confirmed with a fluoroimager. It can be seen that the product obtained by removing mRNA from the reverse transcripts from the first (E1) and second (E2) eluates from the magnetic beads has extremely high purity (FIG. 10).
This means that mRNA protein fusion is bound to the magnetic bead surface by simple IMAC using His-tag, purification of mRNA protein fusion, reverse transcription, mRNA degradation and removal, and cDNA protein fusion. This means that the entire purification process can be carried out efficiently and in a short time, and that the cDNA protein fusion can be purified with high purity and high yield.
In the figure, FT is the filtered fraction, E1 and E2 are those obtained by reverse transcription of eluates 1 and 2 and then removing mRNA by RNaseH treatment. PDO corresponds to the mRNA protein fusion, but cDNA protein fusion. In order to show the quantity ratio, the position of the PDO band is shifted upward.

Reference Example 5 In Vitro Selection In this example, IgG recognized by BDA, AChBP recognized by 3F, and IL6R recognized by 6R14 were immobilized on magnetic beads as models, and each target protein was BDA, 3F of the present invention. The PDO cDNA protein fusion is used as a control together with the 6R14 cDNA protein fusion, and the in vitro selection efficiency of the present invention is evaluated.

(5-1) Preparation of Magnetic Beads Immobilized with Biotin-Labeled Target Protein Among the target proteins used for selection, immunoglobulin G (IgG) was obtained from Sigma from Interleukin-6 receptor (IL6Rα) and VEGF (Human VEGF165) was purchased from Pero Tech (London, UK). Acetylcholine binding protein (AChBP) cDNA was cloned from Aplysia kurodai [manuscript in preparation], expressed in recombinant E. coli and purified.
In order to immobilize these target proteins on magnetic beads, they were biotinylated using EZ-Link Sulfo-NHS-SS-biotin (Pierce, Rockford, USA) at a molar ratio of 1:10. Unreacted biotin was inactivated with 1M Tris-HCl (pH 8) and removed by dialysis at 4 ° C. for 16 hours. Each biotinylated target protein was confirmed by SDS-PAGE and Western blot using streptavidin-HRP (GE Healthcare). Quantitative immobilization was checked by titrating at 25 ° C. for 30 minutes against a fixed amount of magnetic beads coated with streptavidin (Takara, Japan). The target protein immobilized by storing the supernatant was eluted with 100 mM dithiothreitol (DTT; Wako, Japan) at 25 ° C. for 10 minutes.

(5-2) In vitro selection Subsequently, each cDNA protein fusion of PDO and BDA, 3F, and 6R14 was immobilized on the magnetic beads immobilized with the target protein (IgG, IL6R, AChBP) prepared in (5-1) above. To react.
Non-binding PDO and binding BDA (or 3F, 6R14) mRNA- (linker-N) complex were mixed at a ratio of 20: 1, and translated and modified by the same steps as in Example 4 to obtain cDNA. Display protein.
Here, in the case of a ligand rich in disulfide such as 3F and 6R14, it is assumed that a fusion of 70% or more of protein disulfide isomerase (PDI) (input template is formed according to Non-Patent Document 23). (1: 1 molar ratio to cDNA display) in the presence of 10 mM oxidized glutathione (GSSG) and 1 mM reduced glutathione (GSH) and subjected to translation, refolding and maturation steps.
As in Example 4, 50 μl of each sample eluted with His-tag was mixed with 150 μl of selection buffer (SB; 1 × PBS, 100 mM NaCl and 0.1% Tween-20), and 30 minutes at 25 ° C. IgG, IL6R and AChBP were incubated with immobilized magnetic beads. All dithiothreitol (DTT) is removed from the buffer used for each selection.
The beads were washed 10 times with 200 μl of the same buffer, then eluted with 100 mM DTT (micro-spin column, GE Healthcare) and desalted (10 mM TrisHCl pH 7).
The “before selection” fraction and the eluted “after selection” fraction were amplified for 25 cycles using the following 0.2 mM PS and Pv3A primers (combined to Greiner), respectively (denaturation (94 ° C.)) 20 seconds; annealing (60 ° C), 15 seconds; extended (72 ° C), 20 seconds).
PS: 5'-ATTTAGGTGACACTATAGAATACAAGCTTGCT-3 '(SEQ ID NO: 1) and
Pv3 A: 5'-TTTCCCCGCCGCCCCCCGTCCTGCTTCCGCCGTGATGAT-3 '(SEQ ID NO: 2)
Denaturing gel electrophoresis (4.5% gel and 8M urea) was performed, and the amount of DNA encoding each model protein was quantitatively analyzed by Sybr Gold staining.
A series of these steps is shown as FIG. 11a. The average time required from the initial template DNA to mRNA transcription step to the final analysis step is also included. Thus, each process is performed in a very short time, and it is also proved that the screening method of the present invention is performed with high throughput.
The model (1) PDO / BDA and IgG combined with BDA, (2) PDO / 3F and AChBP combined with AChBP, and (3) PDO / 6R1 and IL6R combined with IL6R. It was selectively amplified compared to PDO (FIG. 11b). (1)-(3) “Fold Enrichment Number”, which is the ratio of change in quantitative PCR before and after each selection, is (1) 40 ± 5%, (2) 30 ± 3 %, (3) 35 ± 2% (FIG. 11c).

(Reference Example 6) Selection using library of 8-mer peptide mixture corresponding to FLAG-tag sequence
An 8-mer peptide library was prepared as follows, which was a randomized mixture of 8-mer peptides corresponding to “DYKDDDDK (SEQ ID NO: 3)” required for binding of the FLAG-tag to the anti-FLAGM2 antibody. .
The 8-mer peptide library was constructed using two oligonucleotides with complementary “overlap sequences”. (1) upstream primer; oligonucleotide having SP6 promoter and cap on the most 5 'side, followed by 5'-UTR, Kozak and ATG followed by "overlap sequence";
(2) Downstream primer; a mixture of oligonucleotide fragments constructed so that an “overlap sequence” is provided on the 3 ′ end side of the random region and has a His-tag and Y-tag sequence at the 5 ′ end ( FIG. 12a).
Here, “overlapping sequence” means that two or more bases on the 3 ′ side of a synthetic oligonucleotide that is an antisense primer containing two sense primers and a random sequence are complementarily overlapped with each other by PCR. This is a method of constructing a library while extending from the 5 'side to the 3' side.
The two types of primers used here are as follows.
Upstream sense primer: (SEQ ID NO: 4)
5'-ATTTAGGTGACACTATAGAATACAAGCTTGCTTGTTCTTTTTGCAGAAGCT
CAGAATAAACGCTCAACTTTGGCAGATCTACCATGG-3 '
Downstream antisense primer: (SEQ ID NO: 5)
5'-TTTCCCCGCCGCCCCCCGTCCTGCTTCCGCCGTGATGATGATGGTGGTGAG
ACCCTCCGCCTGAGCCTCCACC SNNSNNSNNSNNSNNSNNSNNSNN TGAACCTCCCATGGT-3 '
The random region was encoded by the NNS codon (N = A, T, C or G; S = C or G) and included in the downstream primer.
The two primer fragments were ligated in 15 cycles of overlapping PCR with denaturation (94 ° C) for 20 seconds, annealing (60 ° C) for 15 seconds and extension (72 ° C) for 30 seconds.
The tag construct was prepared in the same way except that the random region was replaced with a FLAG-tag sequence.

The 8-mer peptide library mRNA-linker-N was prepared by ligating the 8-mer peptide library mRNA to the linker-N obtained in Reference Example 1 in the same manner as in Reference Example (2-2). 10 pmol of 8-mer peptide library mRNA-linker-N was mixed with 10 fmol of FLAG-tag mRNA-linker prepared similarly.
In the same manner as in Example (3-3), an mRNA protein fusion was obtained by translation and maturation in a wheat germ extract translation system (WGE), and reverse-transcribed after purification with a His-tag as in Reference Example 4. Elution gave a library of cDNA protein fusions.
On the other hand, for comparison, a “puro-linker” described in Non-Patent Document 17 was applied to the 8-mer peptide library, and a library of cDNA protein fusions was prepared according to the protocol of the document.

Each of the above-mentioned two cDNA protein fusion libraries was combined with 200 nM anti-FLAG M2 antibody (manufactured by Sigma) immobilized on magnetic beads in the same manner as in Example (5-1), and 200 μl of DTT. Was mixed in a buffer solution (SB) from which was removed, and reacted at 25 ° C. for 30 minutes. As in Example (5-2), the beads were washed 5 times (Round 1), 8 times (Round 2) and 10 times (Round 3) with an equal volume of buffer (SB) and eluted with 100 mM DTT. And then desalted.
The eluted fraction was PCR amplified and proceeded to the next round, and a total of 3 cycles were selected. After round 3, PCR products were cloned (TA cloning, Invitrogen) and 20 clones were selected and sequenced by random picking.
A conceptual diagram of these series of steps is shown in FIG. 12b.
FIG. 12 c shows the selection results when the linker-N of the present invention is used and when the Puro-linker of Non-Patent Document 17 is used.
When linker-N was used, 80% of the FLAG-tag sequence of DYKDDDDK (SEQ ID NO: 3) could be selected after 3 cycles, but when the conventional Puro-linker (Non-patent Document 17) was used. The selected FLAG-tag sequence remained at 20%.

  This indicates that, during the selection process of the present invention, surplus mRNA-linker-N conjugate, mRNA protein fusion, etc., which become “noise”, could be efficiently removed from the system, and the S / N ratio could be reduced. ing.

(Example 1) Screening of VEGF binding protein from 3F library (Ex. 1-1) Preparation of 3F cDNA protein fusion library First, according to Non-Patent Document 14, the following three fingers (3F) Loop1, Loop2 and An mRNA library was prepared from a 3F library obtained by randomizing Loop3 (MicTx-3 template: FIG. 13).
The amino acid sequence of “3F” is as follows, and the random region is referred to as Loop1, Loop2, and Loop3 in order from the beginning.
LVCY (X) ₆ PGTLETCPDDFTCV (X) ₁₀ TQYCSHACAIP (X) ₇ CCQTDKCNG (SEQ ID NO: 15)
Next, the same procedure as in the above example was applied to the linker-N of the present invention, and the library mRNA- (linker-N) fusion (200 nM) was translated in 0.5 ml of WGE lysate reaction solution to give 3F- A cDNA fusion was made. The composition of the reaction solution is as follows.
Library mRNA-linker fusion 200 nM
WGE lysate (PROMEGA) 250μl
Amino acid mixture (PROMEGA) 40μl
Potassium acetate solution 50mM
RNase inhibitor (PROMEGA) 200unit
Protein Disulfide Isomerase (Takara Bio) 5μg
Oxidized glutathione 10 mM
Reduced glutathione 1mM
The reaction was carried out at 25 ° C. for 15 minutes.

As a result, approximately 3 × 10 ¹² molecules were obtained as a 3F-cDNA fusion.
In order to estimate the number of molecules of the cDNA protein fusion in advance, SDS-PAGE is performed using a known amount of the 3F-mRNA-linker as a concentration standard.

(Ex. 1-2) In vitro selection
VEGF (PeproTech, London) was immobilized on magnetic beads (Magnosphere MS300 / Streptavidin, Takara Bio) as in Reference Example (5-1).
The cDNA protein fusion obtained in (Act 1-1) is reacted with immobilized VEGF in 200 μl of SB (selection buffer) buffer at 25 ° C. for 1 hour, followed by SB buffer and WB (wash). buffer) was washed several times with a buffer solution (1 × PBS, 100 mM NaCl and 0.5% Tween-20). A 100 mM DTT solution was administered to the washed beads, and allowed to stand at 25 ° C. for 10 minutes to obtain a supernatant containing cDNA.
The resulting supernatant was desalted with Micro-spin column (GE Health Healthcare) and amplified by PCR (denaturation (94 ° C) for 20 seconds, annealing (60 ° C) for 15 seconds and extension (72 ° C). 30 seconds). In order to apply the selection pressure, the concentration of VEGF was changed from 1 μM to 1 nM, and the number of washes was gradually increased with the WB wash. After 7 rounds (R7), the PCR product was cloned into pCR2.1 vector (TA cloning, Invitrogen) and 20 clones were randomly selected and sequenced.

(Act 1-3) Analysis of clone sequence As a result of analyzing the sequence of the obtained 20 clones, when grouped by the amino acid sequence encoded by each, the G-1 sequence was 60%, the G-2 sequence and the G -3 sequences were 15% each. The amino acid sequences of Loops 1 to 3 of each group are as follows.
G-1: 60% (12/20)
Loop1 = PLTRVV (sequence number 6), Loop2 = HGDHHTLSEW (sequence number 7), Loop3 = EEPTAHV (sequence number 8)
G-2: 15% (3/20)
Loop1 = WEVVLL (SEQ ID NO: 9), Loop2 = AHSVTLAHGH, (SEQ ID NO: 10),
Loop3 = TGPGAER (SEQ ID NO: 11)
G-3: 15% (3/20)
Loop1 = TLWLSY (SEQ ID NO: 12), Loop2 = DVPQSGTNLA (SEQ ID NO: 13),
Loop3 = ELTHPVD (SEQ ID NO: 14)

(Act 1-4) Measurement of dissociation constant The dissociation constant (Kd) was measured according to the method reported in Non-Patent Document 14. A certain amount of 3F display protein (˜1 nM) was reacted with various amounts of VEGF (0.1 nM to 1 μM) for 1 hour. The mixture was applied to a constant amount of VEGF (˜100 nM) and incubated for another 30 minutes. After washing several times, anti-His-antibody (R & D systems) was added at 1/2000 dilution and allowed to react for 1 hour. After the reaction mixture was washed several times, substrate and TMB were added. After appropriate color development, the reaction was stopped by adding 0.5 M H ₂ SO ₄ and the absorbance at 450 nm was measured.
The typical sequences and their dissociation constants (Kd values) of G-1, G-2 and G-3 groups of 20 clones are shown in (Table 1). In particular, the G-1 group has an extremely low value of 2.1 ± 0.5 nM as a peptide, and its affinity is comparable to the Kd value of 2.9 nM of a commercially available anti-VEGF antibody (Non-patent Document 16).

The peptides of the G-1 group (peptide G-1) are VEGF-binding peptides with high specificity to VEGF, and have a CTx3-3F sequence derived from CTx3 (Non-patent Document 15) as a basic skeleton. Therefore, it is represented by the following amino acid sequence (Formula 2).
(Formula 2)
LVCYPLTRVVPGTLETCPDDFTCVHGDHHTLSEWTQYCSHACAIPEEPTAHVCCQTDKCNG (SEQ ID NO: 16)

(Example 2) Evaluation of binding specificity (Act 2-1)
Using the R7 mRNA- (linker-N) (200 nM) obtained in Example 1, translation was performed in the same manner as in (Ex. 1-1) to prepare a cDNA protein fusion.
Dividing these fusions into three parts,
(1) Streptavidin beads (when nothing is bound),
(2) VEGF beads (200 nM) and (4) Fc beads (200 nM), (Fc-immobilized beads were prepared by the method described in Example (5-1) as with VEGF beads). The reaction was performed in 200 μl of SB (selection buffer) buffer at 25 ° C. for 30 minutes. As “Fc”, human IgG1-derived Fc (R & D Systems) was used.
As a control for (3), an equivalent amount (200 nM) of R7 cDNA- (linker-N) obtained by reverse transcription of R7 mRNA- (linker-N) without a translation step into protein was reacted under the same conditions as for VEGF beads. .
Each bead was washed 10 times with the same buffer and eluted with 100 mM DTT. The eluted fraction was desalted and PCR amplified under the following conditions (PCR conditions: denaturation (94 ° C) for 20 seconds, annealing (60 ° C) for 15 seconds and extension (72 ° C) for 30 seconds).

For the R7 cDNA fusion in the cases (1) to (4) above, the eluted fraction was PCR amplified and analyzed by gel electrophoresis (FIG. 14a). Other conditions such as electrophoresis conditions are the same as in Non-Patent Document 14.
Next, analysis by ELISA was performed according to Non-Patent Document 14.
The R7 cDNA fusion was divided into three portions and incubated with SA beads, 200 nM VEGF beads and 200 nM Fc beads in PBS-BSA (0.01% BSA) at 25 ° C. for 1 hour. The mixture was washed with PBS-T (0.1% Tween) and subsequently reacted with anti-His antibody (R & D system) in PBS-T at 25 ° C. for 30 minutes. The mixture was then washed several times with PBS-T, followed by the addition of the substrate, 3,3 ′, 5,5′-tetramethylbenzidine (TMB; Sigma). After color development, 0.5 M H ₂ SO ₄ was added to stop the reaction, the sample was centrifuged, and the absorbance at 450 nm was measured.
As a result, as described in detail in (Ex. 2-2) below, it was clearly shown that the R7 cDNA fusion was highly bound to the VEGF beads (FIG. 14b).
About the obtained VEGF-binding peptide (peptide G-1: SEQ ID NO: 16), the three-finger peptide AdTx1 (derived from Dendraspis angusticeps, which has high amino acid sequence homology with peptide G-1 and has already been subjected to a three-dimensional structure analysis, PDB = 4IYE) was used as a template, and an estimated three-dimensional structure diagram was created by structure modeling using molecular structure construction software “Prime” (Schroedinger) (FIG. 16).

(Act 2-2) Results of PCR analysis
In the binding assay analyzed by PCR, the R7 library protein fusion binds less than 0.2% to the negative control without VEGF (FIG. 14a (1)). On the other hand, significant binding (2% or more; FIG. 14a (2)) was shown with VEGF beads. To verify the potential for library enrichment via nucleic acid aptamers, ie, mRNA / cDNA hybrids, R7 mRNA / cDNA hybrids were prepared by reverse transcription of mRNA-linkers and tested for binding potential. As a result, binding to VEGF via mRNA / cDNA hybrid was not observed (less than 0.2%; FIG. 14a (3)). The R7 protein fusion also showed no affinity for Fc (less than 0.2%; Figure 14a (4)), suggesting target (VEGF) specific enrichment of the 3F library. In the binding assay by ELISA, the R7 protein fusion was 5 times higher for VEGF-immobilized beads compared to SA beads without VEGF (1) and Fc-immobilized beads as a negative control (4). A strong signal was shown (Figure 14b). This result also suggests that this library was strongly specific for VEGF.

1. Sequence number 1: PS primer
2. Sequence number 2: Pv3 A primer
3. Sequence number 3: FLAG-tag
4). SEQ ID NO: 4: 8-mer peptide library primer (F)
5). SEQ ID NO: 5: 8-mer peptide library primer (R)
6). Sequence number 6: G-1 Loop1
7). Sequence number 7: G-1 Loop2
8). Sequence number 8: G-1 Loop3
9. Sequence number 9: G-2 Loop1
10. SEQ ID NO: 10: G-2 Loop2
11. Sequence number 11: G-2 Loop3
12 Sequence number 12: G-3 Loop1
13. SEQ ID NO: 13: G-3 Loop2
14 SEQ ID NO: 14: G-3 Loop3
15. SEQ ID NO: 15: CTx3-3F peptide
16. SEQ ID NO: 16: VEGF binding peptide
17. SEQ ID NO: 17: 1F-Loop I
18. SEQ ID NO: 18: 1F-Loop II
19. SEQ ID NO: 19: 1F-Loop III
20. SEQ ID NO: 20: hVEGF-A

Claims (6)

In the amino acid sequence of the three finger-like scaffold (3F),
(I) the sequence of the region from the second amino acid on the C-terminal side of C1-Cys to the seventh amino acid on the N-terminal side of C2-Cys contained in Loop I is “PLTRVV (SEQ ID NO: 6)”;
(Ii) the sequence of the region from the second amino acid on the C-terminal side of C3-Cys to the fourth amino acid on the N-terminal side of C4-Cys contained in Loop II is “HGDHHTLSEW (SEQ ID NO: 7)”; iii) An amino acid sequence in which the sequence from the fourth amino acid on the C-terminal side of C5-Cys to the first amino acid on the N-terminal side of C6-Cys contained in Loop III is “EEPTAHV (SEQ ID NO: 8)” A VEGF-binding peptide comprising.   A VEGF-binding peptide comprising the amino acid sequence shown in SEQ ID NO: 16.   A nucleic acid encoding the amino acid sequence according to claim 1 or 2. A VEGF-binding peptide derived from the amino acid sequence of a three-finger-like scaffold (3F) and comprising any one of the following amino acid sequences (1) to (3);
(1) The amino acid sequence forming loop I, and the sequence of the region from the second amino acid on the C-terminal side of C1-Cys to the seventh amino acid on the N-terminal side of C2-Cys contained in loop I is “ PLTRVV (SEQ ID NO: 6) ", an amino acid sequence in which Cys that does not participate in cyclization is substituted with another amino acid,
(2) The amino acid sequence forming loop II, and the sequence of the region from the second amino acid on the C-terminal side of C3-Cys to the fourth amino acid on the N-terminal side of C4-Cys contained in loop II is “ HGDHHTLSEW (SEQ ID NO: 7) ", an amino acid sequence in which Cys not involved in cyclization is substituted with another amino acid,
(3) The amino acid sequence forming loop III, and the sequence of the region from the fourth amino acid on the C-terminal side of C5-Cys to the first amino acid on the N-terminal side of C6-Cys contained in loop III is “ An amino acid sequence which is “EEPTAHV (SEQ ID NO: 8)”.   The amino acid sequence forming loop I of (1) is the amino acid sequence shown in SEQ ID NO: 17, and the amino acid sequence forming loop II of (2) is the amino acid sequence shown in SEQ ID NO: 18. The VEGF-binding peptide according to claim 4, wherein the amino acid sequence forming the loop III of (3) is the amino acid sequence represented by SEQ ID NO: 19.   A nucleic acid encoding the amino acid sequence according to claim 4 or 5.