Muteins of human lipocalin 2 (Lcn2, hNGAL) with affinity for a given target

Abstract

The present invention relates to a novel library for the generation of muteins and to novel muteins derived from human lipocalin 2 (Lcn2, hNGAL) and related proteins that bind a given target with detectable affinity. The invention also relates to corresponding nucleic acid molecules encoding such a mutein and to a method for their generation. The invention further relates to a method for producing such a mutein. For example, such muteins may serve to bind and deplete pathological forms of natural biomolecules such as the amyloid beta peptide in Alzheimer's disease or may target the fibronectin extra-domain B, which is associated with tumor neovasculature.

Claims

1. A human neutrophil gelatinase-associated lipocalin (hNGAL) mutein, comprising a mutated amino acid residue at one or more amino acid sequence positions selected from the group consisting of amino acid sequence positions 96, 100, and 106 of mature hNGAL (SEQ ID NO: 44) and at nine or more amino acid sequence positions selected from the group consisting of amino acid sequence positions 36, 40, 41, 49, 52, 68, 70, 72, 73, 77, 79, 81, 103, 125, 127, 132, and 134 of mature hNGAL (SEQ ID NO: 44), wherein the mutein demonstrates binding to a target with detectable affinity, which is a non-natural ligand in that it does not bind to the mature hNGAL (SEQ ID NO: 44) with detectable affinity under physiological conditions, and wherein the mutein comprises with respect to the mature hNGAL (SEQ ID NO: 44) one or more amino acid replacements that are: (A) selected from the group consisting of Leu36.fwdarw.Val or Cys or Ala: Ala40.fwdarw.Tyr or Lys or Val: Ile41.fwdarw.Thr or Ser or Leu: Gln49.fwdarw.Leu or Trp; Leu70.fwdarw.Gly; Arg72.fwdarw.Gly or Asp; Lys73.fwdarw.Leu or Thr or Asp; Asp77.fwdarw.Asn or His or Leu; Trp79.fwdarw.Lys; Asn96.fwdarw.Ile or Arg; Tyr100.fwdarw.Gln or Arg or Glu; Leu103.fwdarw.Met or Arg or Gly: Tyr106.fwdarw.Tyr or Ala or Trp; Lys125.fwdarw.Thr or Val or Glu; Ser127-Gly or Gin or Ala; Tyr132.fwdarw.Met or Ser or Thr; and Lys134.fwdarw.Asn; or (B) selected from the group consisting of Leu36.fwdarw.Lys or Glu or Arg or Ala; Ala40.fwdarw.His or Met or Thr or Ser, Ile41.fwdarw.Asp or Arg or Trp or Leu; Gln49.fwdarw.Arg or Ala or Tyr; Leu70.fwdarw.Leu or Arg or Met; Arg72.fwdarw.Val or Arg or Gln or Met; Lys73.fwdarw.His or Arg or Ser, Asp77.fwdarw.Asn or His or Lys or Arg; Trp79.fwdarw.Arg or Met or Leu; Asn96.fwdarw.Lys or Ala or Ser; Tyr100.fwdarw.Trp or Pro or Lys; Leu103.fwdarw.His or Pro; Tyr106.fwdarw.Phe or Trp or Thr; Lys125.fwdarw.Arg or His or Thr; Ser127.fwdarw.Tyr or Phe or Ala; Tyr132.fwdarw.Leu or Phe; Lys134.fwdarw.Glu or His or Gly or Phe; and Ser146.fwdarw.Asn.

2. The mutein of claim 1, wherein the mutated amino acid residue at one or more amino acid sequence positions selected from the group consisting of amino acid sequence positions 96, 100, and 106 of mature hNGAL (SEQ ID NO:44) comprises a mutated amino acid residue at two or more of the amino acid sequence positions.

3. The mutein of claim 1, wherein the mutated amino acid residue at one or more amino acid sequence positions selected from the group consisting of 96, 100, and 106 of mature hNGAL (SEQ ID NO:44) comprises a mutated amino acid residue at each of the amino acid sequence positions.

4. The mutein of claim 1, wherein the mutein comprises with respect to the mature hNGAL (SEQ ID NO: 44) one or more amino acid replacements that are selected from the group consisting of Leu36.fwdarw.Val or Cys or Ala; Ala40.fwdarw.Tyr or Lys or Val; Ile41.fwdarw.Thr or Ser or Leu; Gln49.fwdarw.Leu or Trp; Leu70.fwdarw.Gly; Arg72.fwdarw.Gly or Asp; Lys73.fwdarw.Leu or Thr or Asp; Asp77.fwdarw.Asn or His or Leu; Trp79.fwdarw.Lys; Asn96.fwdarw.Ile or Arg; Tyr100.fwdarw.Gln or Arg or Glu; Leu103.fwdarw.Met or Arg or Gly; Tyr106.fwdarw.Tyr or Ala or Trp; Lys125.fwdarw.Thr or Val or Glu; Ser127.fwdarw.Gly or Gln or Ala; Tyr132.fwdarw.Met or Ser or Thr; and Lys134.fwdarw.Asn.

5. The mutein of claim 1, wherein the mutein comprises with respect to the mature hNGAL (SEQ ID NO: 44) one or more amino acid replacements that are selected from the group consisting of Leu36.fwdarw.Lys or Glu or Arg or Ala; Ala40.fwdarw.His or Met or Thr or Ser; Ile41.fwdarw.Asp or Arg or Trp or Leu; Gln49.fwdarw.Arg or Ala or Tyr; Leu70.fwdarw.Leu or Arg or Met; Arg72.fwdarw.Val or Arg or Gin or Met; Lys73.fwdarw.His or Arg or Ser; Asp77.fwdarw.Asn or His or Lys or Arg; Trp79.fwdarw.Arg or Met or Leu; Asn96.fwdarw.Lys or Ala or Ser; Tyr100.fwdarw.Trp or Pro or Lys; Leu103.fwdarw.His or Pro; Tyr106.fwdarw.Phe or Trp or Thr; Lys125.fwdarw.Arg or His or Thr; Ser127.fwdarw.Tyr or Phe or Ala; Tyr132.fwdarw.Leu or Phe; Lys134.fwdarw.Glu or His or Gly or Phe; and Ser146.fwdarw.Asn.

6. The mutein of claim 1, wherein said non-natural ligand is selected from the group consisting of a peptide, a protein, a fragment or a domain of a protein, and a small organic molecule.

7. The mutein of claim 1, wherein the mutein comprises with respect to the mature hNGAL (SEQ ID NO: 44) one or both of the amino acid replacements selected from the group consisting of Gln28.fwdarw.His and Cys87.fwdarw.Ser.

8. The mutein of claim 1, wherein the mutein comprises with respect to the mature hNGAL (SEQ ID NO: 44) one or more amino acid replacements selected from the group consisting of Tyr52.fwdarw.Gln or Val; Ser68.fwdarw.Lys or Asn; and Arg81.fwdarw.Trp or Asn or His.

9. The mutein of claim 1, wherein the mutein is conjugated to a compound selected from the group consisting of an organic molecule, an enzyme label, a radioactive label, a colored label, a fluorescent label, a chromogenic label, a luminescent label, a hapten, digoxigenin, biotin, a cytostatic agent, a toxin, a metal complex, metal, and colloidal gold.

10. The mutein of claim 1, wherein the mutein is fused at its N-terminus and/or its C-terminus to a fusion partner which is a protein, a protein domain or a peptide.

11. The mutein of claim 10, wherein the fusion partner of the mutein is a protein domain that extends the serum half-life of the mutein.

12. The mutein of claim 11, wherein the protein domain is an Fc part of an immunoglobulin, a CH3 domain of an immunoglobulin, a CH4 domain of an immunoglobulin, an albumin binding peptide, or an albumin binding protein.

13. The mutein of claim 1, wherein the mutein is conjugated to a compound that extends the serum half-life of the mutein.

14. The mutein of claim 13, wherein the compound that extends the serum half-life is selected from the group consisting of a polyalkylene glycol molecule, hydroxyethyl starch, an Fc part of an immunoglobulin, a CH3 domain of an immunoglobulin, a CH4 domain of an immunoglobulin, an albumin binding peptide, and an albumin binding protein.

15. The mutein of claim 14, wherein the polyalkylene glycol is polyethylene glycol (PEG) or an activated derivative thereof.

16. The mutein of claim 1, wherein the mutein has at least 70% sequence identity to the linear polypeptide sequence of mature hNGAL (SEQ ID NO: 44).

17. A nucleic acid molecule comprising a nucleotide sequence encoding a mutein of claim 1.

18. The nucleic acid molecule of claim 17, wherein the nucleic acid molecule is operably linked to a regulatory sequence to allow expression of said nucleic acid molecule.

19. A host cell containing a nucleic acid molecule of claim 18.

20. A method for the production of a product that is the mutein of claim 1, a fragment of the mutein or a fusion protein of the mutein and another polypeptide, wherein the method comprises: subjecting a nucleic acid molecule encoding mature hNGAL to mutagenesis at a nucleotide triplet coding for the at least one amino acid at sequence positions 96, 100, and 106 of the mature hNGAL (SEQ ID NO:44), and for the at least nine amino acid at sequence positions selected from the group consisting of amino acid at sequence positions 36, 40, 41, 49, 52, 68, 70, 72, 73, 77, 79, 81, 103, 125, 127, 132, and 134 of the mature hNGAL (SEQ ID NO:44), resulting in one or more nucleic acid molecule(s) encoding the mutein; and expressing the product from the one or more nucleic acid molecule(s) encoding the mutein.

21. The method of claim 20, wherein the product is produced in a bacterial or eukaryotic host organism and is isolated from said host organism or its culture.

Description

(1) The invention is further illustrated by the following non-limiting Examples and the attached drawings in which:

(2) FIG. 1 illustrates the PCR assembly strategy for the simultaneous random mutagenesis of the 20 amino acid positions 36, 40, 41, 49, 52, 68, 70, 72, 73, 77, 79 81, 96, 100, 103, 106, 125, 127, 132, and 134 (underlined and numbered) in the amino acid sequence of the mature Lcn2. These 20 positions were divided into four sequence subsets. For randomization of the amino acids in each subset an oligodeoxynucleotide was synthesized (SEQ ID NO: 1, SEQ ID NO:60, SEQ ID NO:3, SEQ ID NO:4) wherein NNK mixtures of the nucleotides were employed at the mutated codons. N means a mixture of all four bases A, C, G, and T while K means a mixture of only the two bases G and T; hence such a triplet encodes all 20 natural amino acids as well as the amber stop codon TAG, which is translated as glutamine in the E. coli supE-strains XL1-blue (Bullock et al., BioTechniques 5 (1987), 376-378) or TG1 (Sambrook et al., Molecular Cloning. A Laboratory Manual (1989), Cold Spring Harbor Press) that were used for phagemid production and gene expression. Four additional oligodeoxynucleotides (SEQ ID NO:61, SEQ ID NO:6, SEQ ID NO:7, SEQ ID NO:8) with fixed nucleotide sequences corresponding to the non-coding strand (written below the DNA double strand sequence in 3′-5′ direction) and filling the gaps between the aforementioned oligodeoxynucleotides were also used in the assembly reaction. Two shorter flanking oligodeoxynucleotides (SEQ ID NO:9 and SEQ ID NO:10), which were added in excess and carried biotin groups, served as primers for the PCR amplification of the assembled, entirely synthetic gene fragment. The two flanking primers each encompassed a BstXI restriction site (CCANNNNNNTGG) (SEQ ID NO: 56), giving rise to mutually non-compatible overhangs upon enzyme digest. This special arrangement of restriction sites enabled a particularly efficient ligation and cloning of the synthetic gene. Substitution of the amino acid Gln28 to His with respect to the original Lcn2 sequence was necessary to introduce the first BstXI site, while the second one naturally occurs in the cDNA of Lcn2. Furthermore, the unpaired residue Cys87 was replaced by Ser during the gene assembly. After one pot PCR the resulting gene fragment was inserted into a vector providing the missing parts of the Lcn2 structural gene. This illustration also depicts two short primers (SEQ ID NO: 45 and SEQ ID NO: 46) upstream and downstream, respectively, of the cassette flanked by the two BstXI restriction sites, which served for double stranded DNA sequencing. FIG. 1 discloses the full-length DNA sequence as SEQ ID NO: 62 and the full-length protein sequence as SEQ ID NO: 63.

(3) FIG. 2 illustrates the nucleotide sequence of a library of synthetic Lcn2 genes (only the central cassette flanked by the two BstXI restriction sites, as in FIG. 1, is shown). This gene fragment was prepared by Sloning BioTechnology GmbH. Compared with the DNA library described in FIG. 1 there are two differences. First, whenever possible codons optimized for E. coli expression were used throughout for the non-mutated amino acid positions. Second, a mixture of 19 different triplets (GAC, TTC, CTG, CAC, AAT, AGC, ACC, GCA, ATG, CCT, GTT, TGG, GAG, CAA, ATC, GGA, CGT, GCA, TAC), each encoding a different amino acid except Cys, was employed at the 20 randomized positions, which are identical to the ones depicted in FIG. 1. Numbering of amino acids corresponds here to an internal scheme employed by Sloning BioTechnology GmbH, whereby Gly no. 1 is the first amino acid codon directly following the upstream BstX1 restriction site. FIG. 2 discloses SEQ ID NOS 64-66, respectively, in order of appearance.

(4) FIG. 3 shows the SEC elution profiles of the recombinant Aβ fusion proteins Trx-Aβ28 (A) and MBP-Aβ40 (B) after expression in E. coli. The purity of both the synthetic Aβ40 and the recombinant AB fusion proteins was assessed by SDS-PAGE analysis (C). Both fusion proteins were expressed in the cytoplasm of E. coli JM83. After disruption of the cells with a French Pressure Cell the proteins were purified via His.sub.6-tag (SEQ ID NO: 54) affinity chromatography and then further applied to a Superdex 75 HR 10/30 column in the case of Trx-Aβ28 or to a Superdex 200 HR 10/30 column in the case of MBP-Aβ40. Both proteins eluted mainly as monomer in the size exclusion chromatography and were essentially homogeneous after purification. Aβ40 was obtained from the Keck Foundation, treated with HFIP, evaporated in a SpeedVac and finally dissolved in distilled H.sub.2O. The 15% gel shows samples of Aβ40 after the HFIP treatment and Trx-Aβ28 and MBP-Aβ40 after IMAC purification and SEC. M depicts the molecular weight marker with the corresponding band sizes in kDa displayed on the left of the gel.

(5) FIG. 4 represents an overlay of the SEC elution profiles of the Lcn2 muteins H1-G1, S1-A4, and US7 (A) and the analysis of the purified proteins via 15% SDS-PAGE (B). The three Lcn2 muteins H1-G1, S1-A4, and US7 were expressed in the periplasm of E. coli JM83 or TG1-F.sup.− (as shown here) and purified via Strep-tag II affinity chromatography. All three muteins eluted predominantly as monomeric proteins from the Superdex 75 HR 10/30 column with a retention volume of 10 to 11 ml. Peak intensities differed depending on the expression yield of the particular mutein. (B) shows a 15% SDS-PAGE analysis of recombinant wild-type Lcn2 (lanes 1, 5) and the muteins US7 (lanes 2, 6), H1-G1 (lanes 3, 7), and S1-A4 (lanes 4, 8) after Strep-tag II affinity purification and SEC. Lanes 1-4 show the Lcn2 muteins reduced with 2-mercaptoethanol, lanes 5-8 show the proteins under non-reducing conditions. M represents the molecular weight marker with the corresponding band sizes in kDa displayed on the left of the gel.

(6) FIG. 5 depicts the binding activity of the Lcn2 muteins H1-G1, S1-A4, and US7 in capture ELISAs with different biotinylated Aβ targets. StrepMAB-Immo at 10 g/ml was immobilized on microtitre plates and employed for capturing of 1 μM Lcn2 muteins via the Strep-tag II. In figures A to C the binding of the Lcn2 muteins S1-A4 and US7 to (A) biotinylated A340, (B) biotinylated Trx-Aβ28, and (C) biotinylated MBP-Aβ40 are shown. Binding was detected by incubation with ExtrAvidin/AP and a subsequent chromogenic reaction with pNPP. Ovalbumin (Ova), thioredoxin (Trx), and maltose binding protein (MBP) were used as negative controls. Additionally, binding of wild-type (wt) lipocalin was tested. The data were fitted to a monovalent binding model. Figure D shows the same ELISA setup performed with Lcn2 H1-G1 and biotinylated Aβ40 and Trx-Aβ28 as targets. K.sub.D values determined for Aβ40 were 2.7 nM for S1-A4, 6.8 nM for US7, and 16.2 nM for H1-G1, corresponding values for Trx-Aβ28 were 1.9 nM, 2.4 nM, and 24.3 nM. Binding to MBP-A040 showed a K.sub.D value of 4.7 nM for S1-A4 and 11.4 nM for US7. Accordingly K.sub.D values for A316-27 were 2.6 nM and 2.1 nM.

(7) FIG. 6 shows the binding activity of the Lcn2 muteins S1-A4 and US7 in a direct ELISA. The Aβ targets Trx-Aβ28 and MBP-A(340 as well as the control protein maltose binding protein (MBP) were immobilized at 2 μM in PBS over night. Bound Lcn2 muteins were detected via their Strep-tag II using Streptavidin/AP followed by a chromogenic reaction with pNPP. As a control, the binding of wild-type (wt) lipocalin was tested. The data were fitted to a monovalent binding model. K.sub.D values determined for Trx-Aβ28 were 16.2 nM for S1-A4 and 9.6 nM for US7, corresponding values for MBP-Aβ40 were 149 nM and 49.7 nM.

(8) FIG. 7 shows the binding activity of the Lcn2 muteins S1-A4 (A) and US7 (B) in a competitive ELISA. StrepMAB-Immo was immobilized at 10 μg/ml on microtitre plates and employed for capturing of 1 μM Lcn2 mutein via the Strep-tag II. For competition, a constant concentration of biotinylated Trx-Aβ28 as tracer was mixed with varying concentrations of unlabeled Trx-Aβ28. Bound biotinylated Trx-Aβ28 was subsequently detected via incubation with ExtrAvidin/AP followed by a chromogenic reaction with pNPP. The data were fitted using a sigmoidal equation. The K.sub.D values were 21.7 nM for S1-A4 and 76.9 nM for US7.

(9) FIG. 8 depicts the kinetic real time analysis of Lcn2 muteins S1-A4 (A) and US7 (B) measured on a Biacore instrument at a flow rate of 10 μl/min. The MBP-Aβ40 fusion protein was coupled via amine chemistry on a CMD 2001 chip (ΔRU=1455) and each of the purified Lcn2 muteins was applied in a series at different concentrations. The measured signal is shown as a grey line whereas the curve fit is depicted as a black line in each case. The k.sub.on and k.sub.off rates of both muteins differed significantly (k.sub.on(S1−A4)=0.48.Math.10.sup.5 M.sup.−1 s.sup.−1, k.sub.on(US7)=2.02.Math.10.sup.5 M.sup.−1s.sup.−1, k.sub.off(S1−A4)=8.36.Math.10.sup.−5 s.sup.−1, k.sub.off(US7)=25.2.Math.10.sup.5 s.sup.−1). However, the overall K.sub.D values were quite similar with 1.74 nM for S1-A4 and 1.25 nM for US7.

(10) FIG. 9 shows the functional activity of the Lcn2 muteins S1-A4 and US7 tested in a thioflavin T aggregation assay. 100 μM Aβ40 was incubated with 10 μM of US7, S1-A4, wild-type (wt) lipocalin or with PBS at 37° C. without shaking. At indicated time points 20 μl of each sample were mixed with 180 μl of 5 μM thioflavin T and fluorescence was measured with an excitation wavelength of 450 nm and an emission wavelength of 482 nm. The Lcn2 mutein US7 showed potent inhibition of aggregation at a ratio of 1:10 (US7:Aβ40). S1-A4 was not able to significantly inhibit aggregation at this ratio but exerted clear inhibition at a ratio of 1:2 (S1-A4:Aβ40) (not shown). The wild type Lcn2 did not reveal an inhibitory effect on Aβ aggregation.

(11) FIG. 10 shows SDS-PAGE analysis, after staining with Coomassie brilliant blue, of recombinant ED-B (lane 1), FN7B8 (lane 2), and FN789 (lane 3) after ion exchange chromatography. All samples were reduced with 2-mercaptoethanol. M: molecular weight marker (Fermentas, St. Leon-Rot, Germany).

(12) FIG. 11 shows SDS-PAGE analysis, after staining with Coomassie brilliant blue, of Lcn2 muteins N7A (lane 1), N7E (lane 2). N9B (lane 3), and N10D (lane 4) after Strep-tag 11 affinity purification. All samples were reduced with 2-mercaptoethanol. M: molecular weight marker (Fermentas, St. Leon-Rot, Germany).

(13) FIG. 12 depicts binding activity in the ELISA. A microtiter plate was coated with the purified proteins FN7B8 and FN789 and incubated with a dilution series of the selected Lcn2 muteins, followed by detection with Streptavidin-alkaline-phosphatase conjugate and pNPP substrate. Recombinant Lcn2 muteins N7A. N9B, and N10D revealed negligible signals in this assay for FN789, which lacked the ED-B and served as a negative control.

(14) FIG. 13 depicts the kinetic real time analysis of Lcn2 mutein N9B measured on a Biacore instrument. FN7B8 was coupled via amine chemistry to a CMD 200 m sensor chip (ΔRU=500) and the purified Lcn2 mutein was applied at varying concentrations. The measured signal is shown as a grey line whereas the curve fit is depicted as a black line in each case. The kinetic constants determined from this set of curves are listed in Table 2 (Example 17).

(15) FIG. 14 depicts the kinetic real time analysis of Lcn2 mutein N7E measured on a Biacore instrument. FN7B8 was coupled via amine chemistry to a CMD 200 m sensor chip (ΔRU=500) and the purified Lcn2 mutein was applied at varying concentrations. The measured signal is shown as a grey line whereas the curve fit is depicted as a black line in each case. The kinetic constants determined from this set of curves are listed in Table 2 (Example 17).

(16) FIG. 15 depicts the kinetic real time analysis of Lcn2 mutein N7A measured on a Biacore instrument. FN7B8 was coupled via amine chemistry to a CMD 200 m sensor chip (ΔRU=500) and the purified Lcn2 mutein was applied at varying concentrations. The measured signal is shown as a grey line whereas the curve fit is depicted as a black line in each case. The kinetic constants determined from this set of curves are listed in Table 2 (Example 17).

(17) FIG. 16 depicts the kinetic real time analysis of Lcn2 mutein N10D measured on a Biacore instrument. FN7B8 was coupled via amine chemistry to a CMD 200 m sensor chip (ΔRU=500) and the purified Lcn2 mutein was applied at varying concentrations. The measured signal is shown as a grey line whereas the curve fit is depicted as a black line in each case. The kinetic constants determined from this set of curves are listed in Table 2 (Example 17).

(18) FIG. 17 illustrates a sequence alignment of human Lcn2 muteins designated S1-A4 (SEQ ID NO: 39), US-7 (SEQ ID NO: 41), H1-G1 (SEQ ID NO: 43), N7A (SEQ ID NO: 20), N7E (SEQ ID NO: 22), N9B (SEQ ID NO: 24) and N10D (SEQ ID NO: 26) with Lcn2 (SEQ ID NO: 44). The last line below the sequences indicates secondary structural features of the lipocalin (Schönfeld et al. (2009) Proc. Natl. Acad. Sci. USA 106, 8198-8203). FIG. 17 discloses Lib as SEQ ID NO: 67.

(19) FIG. 18 represents an overlay of the SEC elution profiles of the Lcn2 muteins H1GA and H1GV (A), and the analysis of the purified proteins via 15% SDS-PAGE (B). The two Lcn2 muteins H1GA and H1GV were expressed in the periplasm of E. coli JM83 and purified via Strep-tag II affinity chromatography. Both muteins eluted predominantly as monomeric proteins from the Superdex 75 HR 10/30 column with a retention volume of 10 to 11 ml. Peak intensities differed depending on the expression yields of the muteins. (B) shows a 15% SDS-PAGE analysis of recombinant Lcn2 muteins H1-G1, H1GA, and H1GV after Strep-tag II affinity purification and SEC. All three Lcn2 muteins are shown under reduced conditions using 2-mercaptoethanol. M represents the molecular weight marker.

(20) FIG. 19 depicts the binding activity of the Lcn2 muteins H1GA, H1GV, and H1-G1 in capture ELISAs with the biotinylated Aβ targets A40 (A) and Trx-Aβ28 (B). StrepMAB-Immo was immobilized at 10 μg/ml on microtitre plates and employed for capturing of 1 JAM Lcn2 muteins via the Strep-tag II. Binding of the biotinylated targets, applied in a dilution series, was then detected by incubation with ExtrAvidin/AP and a subsequent chromogenic reaction with pNPP. Biotinylated ovalbumin (Ova) and thioredoxin (Trx) served as negative controls and exhibited no binding to the Lcn2 muteins (not shown). The data were fitted to a monovalent binding model. K.sub.D values determined for Aβ40 were 4.2 nM for H1GA, 4.9 nM for H1GV, and 21.5 nM for H1-G1; corresponding values for Trx-Aβ28 were 3.8 nM, 4.2 nM, and 24.4 nM, respectively.

(21) FIG. 20 depicts the kinetic real time analysis of Lcn2 muteins H1GA (A) and H1GV (B) measured on a Biacore X instrument at a flow rate of 20 μl/min. The MBP-A340 fusion protein was coupled via amine chemistry to a CMD 2001 chip (ΔRU=1316) and each of the purified Lcn2 muteins was applied in a series at varying concentrations. The measured signal is shown as a grey line while the curve fit is depicted as a black line in each case. The kon and koff rates of both muteins were calculated as follows: kon(H1GA)=5.77.Math.104 M−1s−1, kon(H1GV)=6.84-104 M−1s−1, koff(H1GA)=2.75.Math.10 −5 s−1, koff(H1GV)=7.26.Math.10 −5 s−1. The overall KD values were 0.476 nM for H1GA and 1.06 nM for H1GV.

(22) FIG. 21 depicts the kinetic real time analysis of the Lcn2 mutein H1GA measured on a Biacore T100 instrument with immobilized Aβ40 using an extended dissociation time for a trace measured at high concentration. The target peptide Aβ40 was coupled via amine chemistry onto the Biacore CM5 chip (ΔRU=325), and the purified Lcn2 mutein H1GA was applied at a flow rate of 30 μl/min. For exact determination of the low koff rate, dissociation was performed for 7200 s at the highest concentration tested. kon rates were determined using the dilution series of the Lcn2 mutein H1GA with both association and dissociation times of 300 s. The kinetic values for H1GA were determined as follows: kon=1.25.Math.105 M−1s−1, koff=1.18.Math.10 −5 s−1, and KD=0.095 nM.

(23) FIG. 22 (A) shows the functional activity of the Lcn2 mutein H1GA tested in a thioflavin T aggregation assay. 500 μl of 1 mg/ml monomeric A3 was incubated in the absence or presence of different molar ratios of Lcn2 mutein H1GA in 0.5×PBS at 37° C. with stirring. Aggregation reactions were prepared in triplicates. For fluorescence measurement 20 μl of a sample taken at periodic intervals was each mixed with 180 μl ThT at a final concentration of 50 μM in 0.5×PBS and analysed at an excitation wavelength of 450 nm and an emission wavelength of 482 nm. The Lcn2 mutein H1GA showed potent inhibition of aggregation at an equimolar ratio. At a subequimolar ratio of 10:2 (Aβ40:H1GA) aggregation was still significantly reduced. (B) In contrast, equimolar amounts of neither BSA nor Lcn2 as negative controls had an inhibitory effect on the aggregation of Aβ40.

(24) FIG. 23 summarizes the KD values determined in the different ELISA setups as well as in surface plasmon resonance measurements for the Lcn2 muteins H1GA and H1GV.

(25) FIG. 24 illustrates a sequence alignment of the Lcn2 muteins S1-A4 (SEQ ID NO: 39), US7 (SEQ ID NO: 41). H1-G1 (SEQ ID NO: 43). H1GA (SEQ ID NO: 50), H1GV (SEQ ID NO: 52). N7A (SEQ ID NO: 20), N7E (SEQ ID NO: 22), N9B (SEQ ID NO: 24), and N10D (SEQ ID NO: 26) with Lcn2 (SEQ ID NO: 44). FIG. 24 discloses Lib as SEQ ID NO: 67

(26) FIG. 25 illustrates a mutual sequence alignment of the muteins H1GA (SEQ ID NO: 50) and H1GV (SEQ ID NO: 52) with their precursor H1-G1 (SEQ ID NO: 43) and the wt Lcn2 (SEQ ID NO: 44)). FIG. 25 discloses Lib as SEQ ID NO: 67.

(27) FIG. 26 illustrates specific binding activity of the Lcn2 Variant N7E for FN7B8 in an ELISA. A microtiter plate was coated with FN7B8, FN789, BSA or ovalbumin and incubated with a dilution series of N7E, followed by detection with streptavidin-alkaline-phosphatase conjugate and pNPP substrate. The Lcn2 variant N7E revealed negligible signals in this assay for FN789, BSA and ovalbumin.

(28) FIG. 27 depicts the kinetic real time analysis of Lcn2 variants N7A (A), N7E (B), N9B (C) and N10D (D) measured on a Biacore X instrument. The single fibronectin domain ED-B was coupled via amine chemistry to a CMD 200 m sensor chip (ΔRU=180) and the purified Lcn2 variants were applied at varying concentrations. The measured signals are shown as traces together with the curve fits. The kinetic constants determined from these sensorgrams are listed in Table 3 (Example 24).

(29) FIG. 28 depicts the analytical size exclusion chromatography of Lcn2 variants N7A (A), N10D (B), N9B (C) and N7E (D). Affinity-purified protein was applied to a Superdex S75 10/30 column equilibrated with TBS. The arrows indicate the exclusion volume of the column (6.7 ml). Analytical gel filtration resulted in a prominent peak for each of the four Lcn2 variants. The apparent molecular weight of the proteins was 21.0 kDa for N7A, 21.3 kDa for N10D, 21.7 kDa for N9B and 21.7 kDa for N7E, indicating the presence of only monomeric species.

EXAMPLE 1: CONSTRUCTION OF A MUTANT LCN2 PHAGE DISPLAY LIBRARY

(30) A combinatorial library of Lcn2 variants was generated on the basis of the cloned cDNA (Breustedt et al. (2006) Biochim. Biophys. Acta 1764, 161-173), which carried the amino acid substitutions Cys87Ser, to remove the single unpaired thiol side chain (Goetz et al. (2000) Biochemistry 39, 1935-1941), as well as Gln28His to introduce a second BstXI restriction site. Mutagenesis and polymerase chain reaction (PCR) assembly of this region was essentially performed according to a published strategy (Beste et al. (1999) Proc. Natl. Acad. Sci. USA 96, 1898-1903; Skerra (2001) J. Biotechnol. 74, 257-275), this time using a one pot amplification reaction with oligodeoxynucleotides (SEQ ID NO: 1-10 AND 60-61) as illustrated in FIG. 1. Oligodeoxynucleotides were designed such that the primers with SEQ ID NO: 1-4 and 60 corresponded to the coding strand and carried degenerate codons at the amino acid positions 36, 40, 41, 49, 52, or 68, 70, 72, 73, 77, 79, 81, or 96, 100, 103, 106, or 125, 127, 132, 134 respectively, while primers with SEQ ID NO: 5-8 and 61 corresponded to the non-coding strand and did not carry degenerate codons or anticodons. The two flanking primers with SEQ ID NO: 9 and SEQ ID NO: 10 were used in excess and served for the amplification of the assembled randomized gene fragment. All PCR steps were performed using Go-Taq Hot Start DNA polymerase (Promega, Mannheim, Germany) as described (Schlehuber et al. (2000) J. Mol. Biol. 297, 1105-1120).

(31) Oligodeoxynucleotides that did not carry degenerate codons were purchased in HPLC grade from Metabion (Munich, Germany). NNK-containing oligodeoxynucleotides were purchased desalted from the same vendor and further purified by urea PAGE. The resulting DNA library was cut with BstXI (Promega, Mannheim, Germany) and cloned on the phagemid vector phNGAL102 (SEQ ID NO:11), which is based on the generic expression vector pASK111 (Vogt and Skerra (2001) J. Mol. Recognit. 14 (1), 79-86) and codes for a fusion protein composed of the OmpA signal peptide, the modified mature Lcn2, followed by an amber codon, and the C-terminal fragment of the gene 111 coat protein of the filamentous bacteriophage M13, i.e. similar as previously described for the bilin-binding protein (Beste et al., supra: Skerra, supra). After electroporation of E. coli XL1-Blue (Bullock et al. (1987) Biotechniques 5, 376-378) with the ligation mixture of 8.4 μg digested PCR product and 94 μg digested plasmid DNA, 1×10.sup.10 transformants were obtained.

(32) Alternatively, a cloned synthetic Lcn2 random library, which is described in FIG. 2, was obtained from Sloning BioTechnology GmbH (Puchheim, Germany). The central gene cassette flanked by the two BstXI restriction sites was amplified via PCR in 20 cycles using appropriate primers (SEQ ID NO: 9 and SEQ ID NO: 10) and subcloned on phNGAL108 (SEQ ID NO:12) yielding a library with a complexity corresponding to 1.7×10.sup.10 independent transformants. phNGAL108, which is based on the generic expression vector pASK75 (Skerra (1994) Gene 151, 131-135), codes for a fusion protein composed of the OmpA signal peptide, the modified mature Lcn2, the Strep-tag followed by an amber codon, and the full length gene III coat protein of the filamentous bacteriophage M13 (Vogt and Skerra (2004) ChemBioChem 5, 191-199).

(33) The following steps in library generation were performed identically for both Lcn2 libraries. 100 ml of the culture, containing the cells which were transformed with the phasmid vectors on the basis of phNGAL102 or phNGAL108, respectively, coding for the library of the lipocalin muteins as phage pIII fusion proteins, were transferred to a sterile Erlenmeyer flask and incubated for one hour at 37° C. 160 rpm in 2YT medium without antibiotic selection pressure. Before infection with VCS-M13 helper phage the culture was diluted in 2YT medium to an OD550 of 0.1 with the corresponding antibiotic added and further grown under identical conditions until an OD550 of 0.6 was reached. After infection with VCS-M13 helper phage (Agilent Technologies. La Jolla. USA) at a multiplicity of infection of approximately 10 the culture was shaken for additional 30 min at 37° C., 100 rpm. Then the incubator temperature was lowered to 26° C. and the shaker speed was increased again to 160 rpm, after 10 min kananmycin (70 μg/ml) was added, followed by induction of gene expression via addition of anhydrotetracycline (ACROS Organics, Geel, Belgium) at 25 μg/l (125 μl of a 200 μg/ml stock solution in dimethylformamide. DMF per liter of culture). Incubation continued for another 12-15 h at 26° C., 160 rpm.

(34) Cells from the complete culture were sedimented by centrifugation (30 min, 18000 g, 4° C.). The supernatant containing the phagemid particles was sterile-filtered (0.45 μm), mixed with ¼ volume 20% w/v PEG 8000, 15% w/v NaCl, and incubated on ice for at least 2 h. After centrifugation (30 min, 18000 g, 4° C.) the precipitated phagemid particles from 1 liter of culture were dissolved in 30 ml of cold BBS/E (200 mM Na-borate, 160 mM NaCl, 1 mM EDTA pH 8.0) containing 50 mM benzamidine (Sigma) and Pefabloc 1 μg/ml (Roth, Karlsruhe. Germany). The solution was incubated on ice for 1 h. After centrifugation of undissolved components (10 min, 43000 g. 4° C.) each supernatant was transferred to a new reaction vessel.

(35) Addition of ¼ volume 20% w/v PEG 8000, 15% w/v NaCl and incubation for 60 min on ice served to reprecipitate the phagemid particles until the phagemids were aliquoted and frozen at −80° C. for storage. For the first selection cycle phagemids were thawed and centrifuged (30 min, 34000 g, 4° C.), the supernatant was removed, and the precipitated phagemid particles were dissolved and combined in a total of 400 μl PBS containing 50 mM benzamidine. After incubation for 30 min on ice the solution was centrifuged (5 min, 18500 g, 4° C.) in order to remove residual aggregates and the supernatant was used directly for the phage display selection.

EXAMPLE 2: PREPARATION OF AD TARGETS IN DIFFERENT FORMATS

(36) Aβ40 peptide (SEQ ID NO: 29), corresponding to amino acids 1 to 40 of the mature beta-amyloid sequence (Dodel et al. (2003) Lance Neurology 2, 215-220), was purchased as synthetic, lyophilized peptide either with a C-terminal biotin group attached via a lysine spacer (Peptide Speciality Laboratory, Heidelberg, Germany) or in a non-labeled form (W. M. Keck Laboratory, New Haven, USA). Homogeneously monomeric A340 was obtained by dissolving up to 5 mg of the peptide in 0.5 mL of 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP; Sigma-Aldrich, Steinheim, Germany) at room temperature for at least 0.5 h. Subsequently, HFIP was evaporated in a SpeedVac concentrator, and A040 was dissolved in a suitable volume of distilled H.sub.2O followed by sonication (Bandelin, Sonorex, RK100, Germany) in cold water for 15 min. After filtration with a 0.22 μm filter (Spin-X Centrifuge Tube Filter; Corning, USA), protein concentration was determined by absorption measurement at 280 nm using the calculated extinction coefficient of 1490 M.sup.−1cm.sup.−1 (Gasteiger et al. (2003) ExPASy: the proteomics server for in-depth protein knowledge and analysis. Nucleic Acids Res. 31, 3784-3788, http://www.expasy.ch/tools/protparam.html).

(37) Shorter versions of the beta-amyloid peptide, such as the N-terminal peptide Aβ1-11 (SEQ ID NO: 30) and the central peptide Aβ16-27 (SEQ ID NO: 31), were also purchased from the Peptide Speciality Laboratory. These shorter versions were dissolved straight in the buffer of choice without previous HFIP treatment.

(38) In addition to the synthetic peptides, two vectors encoding different recombinant Aβ fusion proteins were constructed for the purpose of bacterial expression: First, a fusion protein with the maltose binding protein (MBP, pMBP-His, SEQ ID NO: 32) was constructed according to Hortschansky et al. ((2005) Protein Sci. 14, 1753-1759) resulting in pASK75-MBP-Abeta40 (SEQ ID NO: 33). Second, pASK75-TrxAbeta28H6 (SEQ ID NO: 34), in which amino acids 1 to 28 of the mature amyloid peptide were inserted into the active site loop of thioredoxin (Trx, pASK75-TrxH6, SEQ ID NO: 35), was constructed according to Moretto et al. ((2007) J. Biol. Chem. 282, 11436-11445).

(39) Sequential cloning of human Aβ40 carrying an extended N-terminus in frame with the gene of the maltose binding protein yielded the new plasmid pASK75-MBP-A340 a derivative of the vector pASK75 (Skerra (1994) Gene 151, 131-135). This vector encodes a fusion protein of the maltose binding protein followed by a His.sub.6-tag (SEQ ID NO: 54) for facile protein purification, a tobacco etch virus (TEV) recognition site for cleavage of the fusion protein as well as the sequence of human Aβ40. The vector is under tight control of the tetracycline promoter/operator system and allows expression of the fusion proteins in high yields in the cytoplasm of E. coli.

(40) The sequence coding for amino acid 1 to 28 of the mature amyloid peptide (Aβ28) was inserted into the thioredoxin loop via a unique CpoI site in the active site loop of thioredoxin (nucleotide position 99-105, corresponding to amino acid residues 34 and 35) yielding the vector pASK75-Trx-Abeta28H6.

(41) Both Aβ fusion proteins, Trx-Aβ28 and MBP-Aβ40, as well as the unfused control proteins Trx and MBP were expressed in E. coli JM83 (Yanisch-Perron et al. (1985) Gene 33, 103-119) at 37° C. Protein expression was induced at an optical density OD.sub.550 of 0.5 by adding 200 μg/l anhydrotetracycline (aTc; Acros, Geel, Belgium) dissolved in water-free dimethylformamide (DMF: Sigma-Aldrich, Steinheim. Germany) and continuing incubation with agitation for 3 h. Cells were harvested by centrifugation at 4200 g at 4° C. for 20 min. The cell pellets from a 2 L culture were resuspended in 20 ml lysis buffer (100 mM Tris/HCl pH 8.0, 50 mM NaCl, 1 mM EDTA) and the resulting suspension was homogenized by three passages through a French Pressure Cell. Insoluble material was removed by centrifugation (34500 g, 4° C., 20 min), the supernatant was filtered with a 0.45 μm filter (Filtropur S 0.45; Sarstedt, Nuembrecht, Germany) and used for affinity purification via the His.sub.6-tag (SEQ ID NO: 54) of each protein. After immobilized metal affinity chromatography (IMAC) the fusion proteins were further purified via size exclusion chromatography (SEC).

(42) Protein concentrations were determined by absorption measurement at 280 nm using calculated extinction coefficients of 66350 M.sup.−1cm.sup.−1 for MBP (SEQ ID NO: 32) 69330 M.sup.−1cm.sup.−1 for MBP-Aβ40 (SEQ ID NO: 33). 15470 M.sup.−1cm.sup.−1 for Trx-Aβ28 (SEQ ID NO: 34), and 13980 M.sup.−1cm.sup.−1 for Trx (SEQ ID NO: 35) (Gasteiger et al., supra).

(43) For the following experiments both Aβ fusion proteins, Trx-A328 and MBP-A340, as well as ovalbumin (Ova Sigma-Aldrich, Steinheim, Germany), Trx, and MBP, which served as control proteins, were labeled with either biotin or digoxigenin (DIG) at a molar ratio of 2:1 (labelling reagent: target protein).

(44) To this end. D-biotinoyl-ε-aminocaproic acid-N-hydroxysuccinimide ester (Roche Diagnostics, Mannheim, Germany) or digoxigenin-3-O-methylcarbonyl-ε-aminocaproic acid-N-hydroxysuccinimide ester (Roche Diagnostics, Mannheim, Germany) dissolved in water-free dimethylformamide (DMF; Sigma-Aldrich, Steinheim. Germany) or dimethylsulfoxide (DMSO; Sigma-Aldrich, Steinheim, Germany) was added in a 2-fold molar ratio to the protein in PBS (4 mM KH.sub.2PO.sub.4, 16 mM Na.sub.2HPO.sub.4, 115 mM NaCl, pH 7.4). The mixture was rotated for 1 h at room temperature. Then, 1 M Tris/HCl pH 8.0 was added to a final concentration of 10 mM and incubated for 10 min to saturate remaining active NHS ester groups, and the labeled proteins were purified via SEC on a Superdex 75 HR 10/30 column (Amersham-Pharmacia. Freiburg, Germany). To be used as target in phage display selection, Trx-Aβ28 was first labeled with digoxigenin for 1 h, then—to block any unpaired Cys residues—a 50-fold molar excess of iodoacetamide (Sigma-Aldrich, Steinheim, Germany) was added, followed by incubation for 1 h. Subsequently, the modified protein was treated with 1 M Tris/HCl pH 8.0 and purified as above.

EXAMPLE 3: SELECTION OF LCN2 MUTEINS WITH AFFINITY TO THE Aβ40 PEPTIDE BY PHAGE DISPLAY

(45) For each panning cycle about 10.sup.12 recombinant phagemids in PBS (4 mM KH.sub.2PO.sub.4, 16 mM Na.sub.2HPO.sub.4, 115 mM NaCl, pH 7.4) were blocked with 2% (w/v) BSA in PBS/T (PBS containing 0.1% (v/v) Tween 20 [polyoxyethylene sorbitan monolaurate; AppliChem, Darmstadt, Germany]) for 1 h. Portions of 50 and of 25 μL of Streptavidin-coated magnetic bead suspension (Dynabeads M-280 Streptavidin; Dynal Biotech, Invitrogen, Karlsruhe, Germany and Streptavidin Magnetic Particles; Roche Diagnostics, Mannheim. Germany) were separately washed with PBS/T and blocked with 2% (w/v) BSA in PBS/T for 1 h. The 25 μL of blocked beads were used for preadsorption of the blocked phagemids to remove phagemids specific for the beads, the 50 μL were used later for the selection cycle.

(46) The blocked phagemids were incubated for 30 min with the 25 μL washed and blocked Streptavidin-coated magnetic beads. The beads were then pulled down with a single tube magnetic stand (Promega, Mannheim, Germany) for 2 min, and the supernatant containing phagemids not bound to the beads was incubated for 1-2 h with 100 nM biotinylated A0340 from Example 2 in a total volume of 400 μL. The mixture of phagemids and peptide target was then incubated for 0.5 h with the 50 μL blocked beads and subsequently pulled down with a single tube magnetic stand for 2 min. The supernatant containing unbound phagemids was discarded. The target/phagemid complexes bound to magnetic beads were washed 10 times with 400 μL PBS/T, and then bound phagemids were eluted under rotation for 10 min with 350 μL 0.1 M glycine/HCl, pH 2.2, followed by immediate neutralization with 55 μL 0.5 M Tris base. Alternatively, elution was performed under denaturing conditions with 400 μL 4 M urea in PBS for 30 min, followed by dilution with 1 mL PBS. Elution under denaturing conditions with either acid or urea was applied for selection cycles 1 and 2 whereas for cycles 3 and 4 elution was performed under conditions of competition by adding 400 μL 100 μM non-biotinylated Aβ40 to the beads with the bound phagemids and rotation for 1 h. In total, 4 selection cycles were performed.

(47) For amplification of eluted phagemids an exponentially growing culture of E. coli XL-1 Blue was infected for 30 min at 37° C. Remaining bead-bound phagemids were eluted by addition of E. coli XL-1 Blue in the exponential growth phase directly to the beads. After centrifugation at 4° C. the bacterial pellet was resuspended in a suitable volume of 2×YT medium (16 g/L Bacto Tryptone, 10 g/L Bacto Yeast Extract, 5 g/L NaCl, pH 7.5), plated onto LB-Cam plates (10 g/L Bacto Tryptone, 5 g/L Bacto Yeast Extract, 5 g/L NaCl, 15 g/L Bacto Agar, 35 mg/L chloramphenicol, pH 7.5), and incubated for 14-16 h at 32° C. Cells were then scraped off the plates and employed for rescue and reamplification of the recombinant phagemids.

(48) Screening of the enriched phagemid pools was performed by a screening ELISA (Example 5) after the fourth panning step.

EXAMPLE 4: SELECTION OF LCN2 MUTEINS WITH AFFINITY TO THE TRX-A1328 FUSION PROTEIN BY PHAGE DISPLAY

(49) For each panning cycle about 10.sup.12 recombinant phagemids in PBS (4 mM KH.sub.2PO.sub.4, 16 mM Na.sub.2HPO.sub.4, 115 mM NaCl, pH 7.4) were first blocked for 1 h with 2% (w/v) BSA in PBS/T (PBS containing 0.1% (v/v) Tween 20) for cycles 1 and 2. From selection cycle 3 onwards, 2% (w/v) skim milk (Sucofin, TSI, Zeven, Germany) in PBS/T was used as blocking reagent. 50 and 25 μL anti-DIG-IgG coated magnetic bead suspension (Europa Bioproducts Ltd, Cambridge, UK) were washed with PBS/T, and separately blocked for 1 h with 2% (w/v) BSA in PBS/T or skim milk.

(50) The blocked phagemids were incubated for 30 min with the washed and blocked 25 μL anti-DIG-IgG coated magnetic beads. The beads were then pulled down with a single tube magnetic stand (Promega, Mannheim, Germany) for 2 min, and the supernatant containing phagemids not bound to the beads was incubated for 30 min with 15 μM non-labeled Trx from Example 2 to remove phagemids specific for Trx. Digoxigenated and carboxymethylated Trx-Aβ28 from Example 2 was then added to a final concentration of 100 nM, and the mixture was incubated for 1-2 h in a total volume of 400 μL. The mixture of phagemids, Trx, and Trx-Aβ28 was then incubated with the drained blocked beads from the 50 μL portion for 0.5 h. The beads were then pulled down with a single tube magnetic stand for 2 min. The supernatant containing unbound phagemids was discarded. The magnetic beads with the bound phagemids were washed 10 times with 500 μL PBS/T. Subsequently, bound phagemids were eluted under rotation for 30 min in 400 μl 4 M urea in PBS, followed by dilution with 1 mL PBS. In total, 6 selection cycles were performed.

(51) For amplification of eluted phagemids an exponentially growing culture of E. coli XL-1 Blue was infected for 30 min at 37° C. Remaining bead-bound phagemids were eluted by addition of E. coli XL-1 Blue in the exponential growth phase directly to the beads. After centrifugation at 4° C. the bacterial pellet was resuspended in a suitable volume of 2×YT medium (16 g/L Bacto Tryptone, 10 g/L Bacto Yeast Extract, 5 g/L NaCl, pH 7.5), plated onto LB-Cam plates (10 g/L Bacto Tryptone. 5 g/L Bacto Yeast Extract, 5 g/L NaCl, 15 g/L Bacto Agar, 35 mg/L chloramphenicol, pH 7.5), and incubated for 14-16 h at 32° C. Cells were then scraped off the plates and employed for rescue and reamplification of the recombinant phagemids.

(52) Screening of the enriched phagemid pools was performed by a filter-sandwich colony screening assay (Example 6) after the sixth panning step.

EXAMPLE 5: IDENTIFICATION OF LCN2 MUTEINS SPECIFIC FOR Aβ VIA SCREENING ELISA

(53) After four cycles of phagemid selection with A040 as described in Example 3, the enriched pool of Lcn2 muteins was subcloned on phNGAL98 (SEQ ID NO: 27), used for transformation of the E. coli supE strain TG1-F.sup.− (a derivative of E. coli K12 TG1 (Kim et al. (2009) J. Am. Chem. Soc. 131, 3565-3576), and subjected to screening ELISA.

(54) For this purpose, single colonies from the enriched pool were grown in 96-well plates (Multiple Well Plate 96 round bottom with lid; Sarstedt, Nuembrecht, Germany) in 100 μL TB-Amp medium (12 g/L Bacto Tryptone, 24 g/L Bacto Yeast Extract, 55 mM glycerol, 17 mM KH.sub.2PO.sub.4, 72 mM K.sub.2HPO.sub.4, 100 mg/L ampicillin) at 37° C. over night. A new plate with 100 μL TB-Amp medium was inoculated with the overnight cultures, and grown to exponential phase at 22° C. or 37° C. Periplasmic expression of Lcn2 muteins was induced with 20 μL 0.2 μg/mL anhydrotetracycline (aTc; Acros. Geel, Belgium) dissolved in water-free dimethylformamide (DMF; Sigma-Aldrich, Steinheim. Germany) for 13-17 h at 20° C. Periplasmic proteins were released with 40 μL BBS (800 mM Na-borate, 640 mM NaCl, 8 mM EDTA, pH 8.0) including 1 mg/mL lysozyme by incubation for 1 h at 4° C. and agitation at 750 rpm (Thermomixer Comfort, Eppendorf, Hamburg, Germany). After blocking with 40 μL 10% (w/v) BSA in PBS/T (4 mM KH.sub.2PO.sub.4, 16 mM Na.sub.2HPO.sub.4, 115 mM NaCl, pH 7.4 with 0.05% (v/v) Tween 20) for 1 h at 4° C. and 750 rpm the plates were centrifuged for 10 min at 4° C. and 3000 g. The supernatant was used for ELISA.

(55) For selective capturing of the Lcn2 muteins carrying the C-terminal Strep-tag 11 (Schmidt and Skerra (2007) Nat. Protoc. 2, 1528-1535), a 96-well MaxiSorp polystyrene microtitre plate (Nunc, Langenselbold, Germany) was coated with 10 μg/mL StrepMAB-Immo (IBA, Göttingen, Germany) in PBS over night at 4° C. and blocked with 3% (w/v) BSA in PBS/T (PBS with 0.1% (v/v) Tween 20) at room temperature for 1 h. After 3 washing steps with PBS/T, 120 μL of the cell extract from above was applied per well and incubated for 1.5 h at 300 rpm. After washing, biotinylated Aβ40 from Example 2 was added at a concentration of 0.5 μM and incubated for 1 h. As a control, biotinylated Ova instead of Aβ40 was used. The wells were washed again, and bound biotinylated peptide or protein was detected with 50 μL of ExtrAvidin/AP conjugate (Sigma-Aldrich, Steinheim, Germany) diluted 1:5000 in PBS/T for 1 h, followed by signal development in the presence of 100 μL 0.5 mg/mL p-nitrophenyl phosphate (AppliChem, Darmstadt, Germany) in 100 mM Tris/HCl, pH 8.8, 100 mM NaCl, 5 mM MgCl.sub.2 for up to 1 h. Absorption at 405 nm was measured in a SpectraMax 250 reader (Molecular Devices, Sunnyvale, USA).

(56) Alternatively, 0.5 μM non-biotinylated Aβ40 from Example 2 was directly immobilized onto a 96-well MaxiSorp polystyrene microtitre plate (Nunc, Langenselbold, Germany), followed by blocking. After incubation with the cell extract from above and washing, bound Lcn2 muteins were detected via the Strep-tag II using a 1:1500 dilution of Streptactin/AP conjugate (IBA, Göttingen, Germany). As a control, 0.5 μM Ova was immobilized onto the ELISA plate.

EXAMPLE 6: IDENTIFICATION OF LCN2 MUTEINS SPECIFIC FOR TRX-Aβ28 VIA FILTER-SANDWICH COLONY SCREENING ASSAY

(57) After six cycles of phagemid selection with Trx-Aβ28 as described in Example 4, the mutagenized Lcn2 gene cassette was subcloned via BstXI (Fermentas, St. Leon-Rot, Germany) on the plasmid phNGAL124 (SEQ ID NO: 42), which encodes a fusion of the OmpA signal peptide, the Lcn2-coding region with the C-terminal Strep-tag II (Schmidt and Skerra (2007) Nat. Protoc. 2, 1528-1535) followed by an amber stop codon as well as a gene for the albumin-binding domain (ABD) from streptococcal protein G (Schlehuber et al. (2000) J. Mol. Biol. 297, 1105-1120).

(58) Then, a filter-sandwich colony screening assay was performed, whereby the Lcn2-ABD fusion proteins were released from the colonies plated on a hydrophilic filter membrane (PVDF type GVWP, 0.22 μm; Millipore, Schwalbach, Germany) and functionally captured on an underlying second membrane (Immobilon-P membrane, 0.45 μm: Millipore, Schwalbach, Germany) coated with human serum albumin (HSA, Sigma-Aldrich, Steinheim. Germany) following a procedure described in detail by Schlehuber et al. ((2000) J. Mol. Biol. 297, 1105-1120). This membrane was probed with 100 nM of DIG-labeled Trx-Aβ28 from Example 2 in PBS/T for 1 h. The bound target conjugate was detected with an anti-DIG Fab/alkaline phosphatase (AP) conjugate (Roche Diagnostics, Mannheim. Germany) followed by a chromogenic reaction with 5-bromo-4-chloro-3-indolyl phosphate, 4-toluidine salt (BCIP: AppliChem, Darmstadt, Germany) and nitro blue tetrazolium (NBT; AppliChem, Darmstadt, Germany). Having identified spots with intense colour signals on this membrane, the corresponding colonies were picked from the first filter and propagated for side by side comparison in a secondary colony screen.

(59) For this purpose isolated bacteria as well as bacteria expressing control Lcn2 muteins, such as the wild-type lipocalin, were stippled onto a fresh hydrophilic membrane and grown until small colonies became visible. Binding was tested both for the DIG-labeled Trx-Aβ28 as well as for the DIG-labeled control proteins Trx and Ova. Lcn2 muteins that gave rise to specific signals for the Trx-Aβ28 target in this secondary screen were further propagated for plasmid isolation and subsequent sequence analysis.

EXAMPLE 7: SOLUBLE PRODUCTION AND PURIFICATION OF LCN2 MUTEINS SPECIFIC FOR Aβ AND ED-B

(60) The recombinant Lcn2 and its muteins were produced by periplasmic secretion in E. coli BL21 (Studier and Moffat (1986) J. Mol. Biol. 189, 113-130), E. coli W3110 (Bachmann (1990) Microbiol. Rev. 54, 130-197), E. coli JM83 (Yanisch-Perron et al. (1985) Gene 33, 103-119) or the E. coli supE strain TG1-F.sup.− (a derivative of E. coli K12 TG1 [Kim et al. (2009) J. Am. Chem. Soc. 131, 3565-3576] that was cured from its episome using acridinium orange). For soluble protein expression the plasmid phNGAL98 (SEQ ID NO: 27) was used, encoding a fusion of the OmpA signal peptide with the mature Lcn2 protein (SEQ ID NO: 28) and the C-terminal Strep-tag II, whereby the plasmid carries the two non-compatible BstXI restriction sites for unidirectional subcloning of the mutated gene cassette.

(61) The soluble protein was affinity-purified by means of the Strep-tag II (Schmidt and Skerra (2007) Nat. Protoc. 2, 1528-1535), followed by size exclusion chromatography (SEC) on a Superdex 75 HR 10/30 column (Amersham-Pharmacia. Freiburg, Germany) using PBS buffer. Protein purity was checked by SDS-PAGE (Fling and Gregerson (1986) Anal. Biochem. 155, 83-88). Protein concentrations were determined by absorption measurement at 280 nm using calculated extinction coefficients of 31400 M.sup.−1cm.sup.−1 for wt Lcn2 (SEQ ID NO: 44) and of 26930 M.sup.−1 cm.sup.−1 for the Aβ specific muteins H1-G1 (SEQ ID NO: 43), 24410 M.sup.−1cm.sup.−1 for S1-A4 (SEQ ID NO: 38), and 26930 M.sup.−1cm.sup.−1 for US7 (SEQ ID NO: 41). The calculated extinction coefficients of the ED-B specific Lcn2 muteins were 37930 M.sup.−1cm.sup.−1 for mutein N7A (SEQ ID NO: 20), 22920 M.sup.−1cm.sup.−1 for N7E (SEQ ID NO: 22), 21430 M.sup.−1cm.sup.−1 for N9B (SEQ ID NO: 24), and 39420 M.sup.−1 cm.sup.−1 for N10D (SEQ ID NO: 26) (Gasteiger et al., supra).

EXAMPLE 8: MEASUREMENT OF BINDING ACTIVITY FOR DIFFERENT Aβ TARGETS IN ELISA EXPERIMENTS

(62) For selective capturing of the Lcn2 muteins carrying the C-terminal Strep-tag II (Schmidt and Skerra (2007) Nat. Protoc. 2, 1528-1535), a 96-well MaxiSorp polystyrene microtitre plate (Nunc, Langenselbold, Germany) was coated with 10 μg/mL StrepMAB-Immo (IBA, Göttingen, Germany) in PBS over night at 4° C. and blocked with 3% (w/v) BSA in PBS/T at room temperature for 1 h. After 3 washing steps with PBS/T, 50 μL of a 1 μM solution of the purified Lcn2 muteins from Example 7 were applied to all wells for 1 h. After washing, 50 μL of a dilution series of the biotinylated targets Aβ40, MBP-Aβ40 or Trx-Aβ28 from Example 2 were added and incubated for 1 h whereby the biotinylated forms of Ova, MBP and Trx from Example 2 served as negative control proteins. The wells were washed again and bound conjugate was detected with 50 μL of ExtrAvidin/AP conjugate (Sigma-Aldrich, Steinheim. Germany) diluted 1:5000 in PBS/T for 1 h, followed by signal development in the presence of 100 μL 0.5 mg/mL p-nitrophenyl phosphate (AppliChem. Darmstadt, Germany) in 100 mM Tris/HCl, pH 8.8, 100 mM NaCl, 5 mM MgCl.sub.2. The time course of absorption ΔA/Δt at 405 nm was measured in a SpectraMax 250 reader (Molecular Devices, Sunnyvale. USA) and the data were fitted with KaleidaGraph software (Synergy software, Reading. PA) to the equation
ΔA=ΔA.sub.max×[L].sub.tot/(K.sub.D+[L].sub.tot)
whereby [L].sub.tot represents the concentration of the applied ligand conjugate and K.sub.D is the dissociation constant (Voss and Skerra (1997) Protein Eng. 10, 975-982).

(63) Alternatively, for a “direct” ELISA, 2 μM unlabeled A340 target from Example 2 was adsorbed onto a 96-well MaxiSorp polystyrene microtiter plate (Nunc, Langenselbold, Germany) and incubated with the purified Lcn2 mutein, which was detected via the Strep-tag II using a 1:1500 dilution of Streptavidin/AP conjugate (GE Healthcare UK Ltd, Buckinghamshire, UK). As a control, 2 μM MBP from Example 2 was adsorbed onto the ELISA plate. Absorption at 405 nm was measured in a SpectraMax 250 reader (Molecular Devices, Sunnyvale, USA).

(64) Alternatively, a “competitive” ELISA was performed in a similar manner as the “capture” ELISA described above. Again, after immobilizing StrepMAB-Immo and washing, 50 μL of a 1 μM solution of the purified Lcn2 mutein from Example 7 was applied to all wells for 1 h. After washing, the biotinylated target Trx-Aβ28 from Example 2 was applied at a fixed concentration of 5 nM (for S1-A4) or 40 nM (for US7) in the presence of varying concentrations of the free non-labeled Trx-A328 target as competitor. Starting with a 100-fold excess, the competitor concentration was decreased by the factor 3 in each step. In this case, the data were fitted to the sigmoidal equation
ΔA=(ΔA.sub.max−ΔA.sub.min)/(1+([L].sub.tot.sup.free/K.sub.D)).sup.p+ΔA.sub.min
with curve slope p (Hill coefficient) as a further parameter.

(65) The following table I summarizes the K.sub.D values determined in the different ELISA setups as well as in surface plasmon resonance (see Example 9).

(66) TABLE-US-00001 TABLE 1 K.sub.D [nM] S1-A4 US7 H1-G1 Aβ40 2.7 ± 0.1 6.8 ± 0.4 16.2 ± 1.9 (Capture ELISA, see FIG. 5) Trx-Aβ28 1.9 ± 0.1 2.4 ± 0.2 24.3 ± 3.8 (Capture ELISA, see FIG. 5) MBP-Aβ40 4.7 ± 0.1 11.4 ± 0.6  n.d. (Capture ELISA, see FIG. 5) Aβ16-27 2.6 ± 0.3 2.1 ± 0.1 n.d. (Capture ELISA, see FIG. 5) Trx-Aβ28 16.2 ± 4.4  9.6 ± 1.7 290 ± 62 (Direct ELISA, see FIG. 6) MBP-Aβ40 149 ± 31  49.7 ± 8.1  n.d. (Direct ELISA, see FIG. 6) Trx-Aβ28 21.7 ± 1.5  76.9 ± 4.5  n.d. (Comp. ELISA, see FIG. 7) MBP-Aβ40 1.7 1.3 n.d. (SPR, see FIG. 8, Example 9)

EXAMPLE 9: MEASUREMENT OF BINDING ACTIVITY FOR Aβ VIA SURFACE PLASMON RESONANCE (SPR)

(67) Real time analysis of Lcn2 muteins was performed on a Biacore X system (Biacore, Uppsala, Sweden) using PBS/T (PBS containing 0.005% (v/v) Tween 20) as running buffer. A 50 μg/mL solution of MBP-Aβ40 from Example 2 in 10 mM Na-acetate, pH 4.5 was immobilized onto a CMD 2001 chip (Xantec, Disseldorf, Germany) using standard amine coupling chemistry, resulting in a ligand density of 1455 resonance units (RU). The purified Lcn2 muteins S1-A4 and US7 from Example 7 were applied in concentrations ranging from 2 nM to 125 nM at a flow rate of 10 μL/min. The sensorgrams were corrected by subtraction of the corresponding signals measured for the control channel, which had been activated and blocked with ethanolamine. Kinetic data evaluation was performed by fitting with BIAevaluation software V 3.0 (Karlsson et al. (1991) J. Immunol. Methods 145, 229-240). During this step the k.sub.on and k.sub.off rates were fitted globally, while the maximum response difference R.sub.max was fitted locally for each curve.

EXAMPLE 10: DETERMINATION OF THE Aβ EPITOPE VIA EPITOPE MAPPING ON A SPOT MEMBRANE

(68) The SPOT membrane (Amino-PEG500-UC540 sheet, 100×150 mm, Intavis, Köln, Germany) for mapping the Aβ epitope of the Lcn2 muteins S1-A4 and US7 was prepared in an automated procedure as described by Frank ((2002) J. Immunol. Methods 262, 13-26) using a MultiPep RS instrument (Intavis, Köln, Germany) and activated amino acids from the same vendor. The A340 amino acid sequence (SEQ ID NO: 29) was synthesized on the membrane in successive hexamers, decamers, and pentadecamers, each with a dislocation of 1 amino acid, thus covering the entire sequence. The C-termini of these peptides became covalently attached to the membrane whereas their N-termini were acetylated. Immobilized Strep-tag II served as a positive control for the detection method.

(69) Side chain deprotection was performed with 95% trifluoroacetic acid (Sigma-Aldrich, Steinheim, Germany), 3% triisopropylsilane (Sigma-Aldrich, Steinheim, Germany), and 2% distilled H.sub.2O for 2 h. After washing the membrane 4 times with dichloromethane (Sigma-Aldrich, Steinheim, Germany), twice with dimethylformamide (Sigma-Aldrich, Steinheim, Germany), and twice with ethanol, the membrane was air-dried.

(70) Prior to use, the membrane was washed once with ethanol, 3 times with PBS (4 mM KH.sub.2PO.sub.4, 16 mM Na.sub.2HPO.sub.4, 115 mM NaCl, pH 7.4), and blocked with 3% (w/v) BSA in PBS/T (PBS containing 0.1% (v/v) Tween 20) for 1 h. After washing 3 times with PBS/T, the membrane was incubated with the Lcn2 mutein S1-A4 (50 nM), US7 (100 nM) or wild-type Lcn2 (100 nM) in PBS/T for 1 h and washed 3 times with PBS/T. Then, the membrane was incubated with a 1:1500 dilution of Streptavidin/AP conjugate (GE Healthcare, Buckinghamshire, UK) in PBS/T for 1 h for the protein detection via the Strep-tag 11. Subsequently, the membrane was washed 3 times with PBS/T and once in Aβ buffer (100 mM Tris/HCl, pH 8.8, 100 mM NaCl, 5 mM MgCl.sub.2), followed by a chromogenic reaction with 5-bromo-4-chloro-3-indolyl phosphate, 4-toluidine salt (BCIP; AppliChem, Darmstadt, Germany) and nitro blue tetrazolium (NBT; AppliChem, Darmstadt, Germany) in Aβ buffer as described by Schlehuber et al. ((2000) J. Mol. Biol. 297, 1105-1120).

(71) Comparison of the membranes incubated with S1-A4 and wild-type Lcn2, respectively, yielded prominent spots that only occurred upon incubation of the membrane with S1-A4 and not with wild-type Lcn2. With the pentadecameric A0340 fragments the motif LVFFAED (SEQ ID NO: 57) appeared to be essential for binding of S1-A4. With the shorter hexameric and decameric peptide sequences binding of S1-A4 to the minimal motives VFFAED (SEQ ID NO: 58) and FFAEDV(SEQ ID NO: 59) was detected. The Lcn2 mutein US7 showed a very similar epitope profile in this assay.

EXAMPLE 11: FUNCTIONAL ANALYSIS OF LCN2 MUTEINS WITH AFFINITY TO Aβ IN A THT AGGREGATION ASSAY

(72) Solubilized and homogenously monomeric, non-biotinylated Aβ40 from Example 2 was subjected to a thioflavin T aggregation assay in the presence or absence of various Lcn2 muteins. Thioflavin T (ThT; Sigma-Aldrich, Steinheim, Germany) specifically binds to β-sheet-rich amyloid fibrils and oligomeric precursors but not to the monomeric Aβ peptide, which is accompanied by an increase in ThT fluorescence (Khurana et al. (2005) J. Struct. Biol. 151, 229-238).

(73) To this end, ThT was dissolved in distilled H.sub.2O to a concentration of 1 mM. The working solution was prepared by diluting the 1 mM stock solution to a final concentration of 5 μM with 5 mM glycine NaOH, pH 8.5. 500 μL of a 1:1 mixture of 200 μM Aβ40 in distilled H.sub.2O (dissolved according to Example 2) and 20 μM of the Lcn2 mutein from Example 7 in PBS was incubated at 37° C. without agitation. Before each fluorescence measurement at different time points, samples were vortexed shortly 10 times and 20 μL of the respective sample were mixed with 180 μL of the ThT working solution. Fluorescence intensity was measured with an excitation wavelength of 450 nm and an emission wavelength of 482 nm in a luminescence spectrometer (LS 50 B, Perkin Elmer, Waltham. USA).

(74) As illustrated in FIG. 9, the Lcn2 mutein US7 showed potent inhibition of Aβ40 aggregation at a ratio of 1:10.

EXAMPLE 12: PREPARATION OF RECOMBINANT FIBRONECTIN FRAGMENTS CONTAINING THE EXTRA DOMAIN B (ED-B)

(75) Three different recombinant fragments of human fibronectin were used as targets for selection: the extra domain-B alone (termed ED-B: Zardi et al. (1987) EMBO Journal, 6, 2337-2342), the same domain in context of its adjacent domains 7 and 8 (referred to as FN7B8), and FN789 comprising the conventional domains 7, 8 and 9, thus lacking ED-B (Camemolla et al. (1996), Int. J. Cancer, 68, 397-405; Leahy et al. (1994) Proteins 19, 48-54).

(76) Coding sequences of FN7B8, FN789, and ED-B were cloned on the vector pASK75 (Skerra (1994) Gene 151, 131-135) or its derivative pASG-IBA-33 (IBA, Göttingen, Germany), yielding pASK75-FN7B8 (SEQ ID NO: 13), pASK75-FN789 (SEQ ID NO: 17), and pASG-IBA-33-EDB (SEQ ID NO: 15), respectively. All constructs provide a His.sub.6-tag (SEQ ID NO: 54) at the carboxy terminus for the purification via immobilized metal affinity chromatography (IMAC) and were produced as soluble proteins in the cytoplasm of E. coli TG1/F.sup.− [a derivative of E. coli K12 TG1 (Gibson (1984) Studies on the Epstein-Barr virus genome, Cambridge University, England)] or BL21 (Studier and Moffatt (1986). 189, 113-130). Therefore, a 2-L culture of E. coli harboring the expression plasmid was grown to mid-log phase at 37° C. in LB-medium containing 100 μg/ml ampicillin. After addition of 200 μg/L anhydrotetracycline (Acros, Geel, Belgium) growth was continued for 5 to 7 h at 37° C. Cells were concentrated by centrifugation, resuspended in 35 ml ice-cold chromatography buffer (40 mM Hepes/NaOH, 1 M NaCl, pH 7.5) and lysed by sonification (S250D, Branson, Danbury CT, USA).

(77) The purification protocol given below was the same for all three recombinant fibronectin fragments. The cleared lysate was applied to a Zn.sup.2+/IDA-Sepharose column (GE Healthcare, Munich, Germany) charged with 10 mM ZnSO.sub.4 and equilibrated with 40 mM Hepes/NaOH, 1 M NaCl, pH 7.5. Bound Protein was eluted with a shallow gradient between 0 to 300 mM imidazole (pH adjusted with HCl) in chromatography buffer. Fractions containing recombinant protein were identified by SDS-PAGE, mixed with EDTA to a final concentration of 1 mM, and dialysed against 20 mM Hepes/NaOH, pH 7.4 overnight. Subsequently, the protein was loaded on an ion-exchange chromatography column (Resource Q, GE Healthcare, Munich. Germany) equilibrated with 20 or 40 mM Hepes/NaOH, pH 7.4. To elute bound fibronectin fragments a gradient between 0 and 300 mM NaCl was used.

(78) Finally, purity of the proteins was confirmed by SDS-PAGE analysis and size exclusion chromatography. To determine the concentration of proteins by absorption measurement at 280 nm, the following calculated extinction coefficients (Gasteiger et al. (2003) Nucleic Acids Res. 31, 3784-3788) were used: 31,400 M.sup.−1cm.sup.−1 for FN7B8, 28420 M.sup.−1cm.sup.−1 for FN789, and 11460 M.sup.−1cm.sup.−1 for ED-B. Typically, about 10 to 20 mg purified protein was obtained per 2-L shake flask culture. Purified proteins were stored at 4° C. and used for all experiments.

(79) Purified fibronectin fragments were digoxigenin-labeled with an excess (4:1 molar ratio) of digoxigenin-3-O-methylcarbonyl-ε-aminocaproic acid-N-hydroxysuccinimide ester (Roche Diagnostics, Mannheim, Germany) for 1 h at room temperature according to the supplier's manual. To remove free digoxigenin the protein solution was applied to a PD-10 desalting column (GE Healthcare, Munich, Germany) equilibrated with 40 mM Tris/HCl, 115 mM NaCl and 1 mM EDTA, pH 7.5. Successful digoxigenation and integrity of the protein was checked by SDS-PAGE, ESI mass spectrometry or dot blot via staining with Anti-Digoxigenin-AP, Fab fragment (Roche Diagnostics, Mannheim, Germany).

EXAMPLE 13: SELECTION OF LCN2 VARIANTS WITH AFFINITY TO ED-B BY PHAGE DISPLAY

(80) Altogether four phagemid display selection cycles were carried out using the Lcn2 random phagemid library based on the synthetic gene collection originally synthesized by Sloning BioTechnology GmbH as described in Example 1. For the first selection cycle about 5×10.sup.12 recombinant phagemids dissolved in 300 μL TBS/E (40 mM Tris/HCl, 115 mM NaCl and 1 mM EDTA, pH 7.4) supplemented with 50 mM benzamidine were blocked for 30 min by adding 100 μL 8% (w/v) BSA (Sigma-Aldrich, Munich, Germany) in TBS containing 0.4% (v/v) Tween 20 [polyoxyethylene sorbitan monolaurate; AppliChem, Darmstadt, Germany]). Then, this solution was incubated for 1 h with Anti-Digoxigenin IgG Magnetic Beads (Europa Bioproducts, Cambridge, UK) that had been blocked for 1 h with 2% (w/v) BSA in TBS/T (TBS containing 0.1° % (v/v) Tween 20) and charged with 400 μl of 0.1 μM digoxigenin-labeled recombinant FN7B8 (see Example 12).

(81) After collecting the beads via a magnet and discarding the supernatant, 10 washing steps with TBS/T were performed and remaining bound phagemids were first eluted with 400 μl of 0.1 M triethylamine (pH not adjusted) for 6 min, followed by a second elution with 350 μl of 0.1 M glycine/HCl, pH 2.2 for 8 min. Eluates were collected and immediately neutralized with an appropriate amount of 2 M acetic acid or 50 μl 0.5 M Tris base, respectively, combined, and used to infect exponentially growing E. coli XL1-Blue. The phagemids were titered and reamplified prior to the next panning step following published protocols (Beste et al. (1999) PNAS 96, 1898-903: Schlehuber et al. (2000) J Mol Biol 297 1109-20).

(82) For the second round of selection about 2×10.sup.12 amplified phagemids were used and elution of bound phagemids was performed by competition with 400 μl of 231 nmol/ml free recombinant ED-B for 75 min.

(83) In the third and fourth selection cycles about 2×10.sup.12 of the amplified phagemids were first incubated for 1 h with 100 nM digoxigenin-labeled FN7B8. Then, the phagemid-antigen complexes were captured on Anti-Digoxigenin IgG Magnetic Beads for 20 min. After washing the beads ten times with TBS/T, bound phagemids were eluted by competition with 400 μl of 140 to 231 nmol/ml of free recombinant ED-B for 75 min at room temperature.

EXAMPLE 14: SELECTION OF LCN2 VARIANTS WITH AFFINITY TO ED-B BY SCREENING ELISA

(84) Enrichment of Lcn2 muteins resulting from the phage display selection described in Example 13 which specifically bind to the target protein ED-B was monitored by screening ELISA. Using the pooled phasmid preparation from the last panning step, the mutagenized gene cassette was subcloned via BstXI on the expression plasmid phNGAL98, which encodes a fusion of the OmpA signal peptide for the periplasmatic production in E. coli and the Lcn2 coding region with the C-terminal Strep-tag II (Schmidt and Skerra (2007) Nat. Protoc. 2, 1528-1535). Soluble expression of individual Lcn2 muteins in 96-well plates was performed as follows: 100 μl TB medium (Tartof and Hobbs (1987) Bethesda Research Laboratory Focus 9, 12) containing 100 μg/ml ampicillin was inoculated with a single, randomly picked colony and incubated for 5 h at 37° C. with shaking (500 to 800 rpm; Thermomixer comfort; Eppendorf, Hamburg, Germany). For each clone 100 μl of fresh medium was inoculated with 10 μl of this culture and incubated for 1 to 2 h at 37° C. with shaking followed by lowering the temperature to 22° C. After further incubation for 2-4 h expression of Lcn2 muteins was induced in exponentially growing cells for 12-14 h with 0.2 μg/ml anhydrotetracycline (Acros, Geel. Belgium). Periplasmic release of proteins was effected by the addition of 40 μl 2×BBS (0.2 M borate/NaOH, 160 mM NaCl, 1 mM EDTA, pH 8.0) containing 1 mg/ml lysozyme (Roche Diagnostics, Mannheim, Germany) for 1 h at room temperature with shaking. Lysates were blocked with 40 μl 10% w/v BSA (Applichem, Darmstadt, Germany) in TBS and 0.5% Tween-20 for 1 h and cell debris was removed from the crude extract by centrifugation for 10 min.

(85) To adsorb FN7B8 or FN789 on the surface of a 96-well Nunc Maxisorp plate (Thermo Fisher Scientific, Langenselbold, Germany) 50 μl of a 100 μg/ml protein solution in TBS was added per well and incubated overnight at 4° C. After three washing steps with TBS/T, wells were blocked with 2% (w/v) BSA in TBS/T for 3 h at room temperature and washed repeatedly before exposition to crude extract from E. coli. For ELISA, 50 μl of the cleared lysate was applied per well, incubated for 1 h, followed by washing three times with TBS/T. Bound Lcn2 muteins were detected with Streptavidin-alkaline phosphatase conjugate (1:1500 in TBS/T; GE Healthcare, Munich, Germany) using the substrate 4-nitrophenyl phosphate (pNpp, 0.5 mg/ml; AppliChem GmbH, Darmstadt, Germany) in 0.1 M Tris/HCl, 0.1 M NaCl, 5 mM MgCl.sub.2, pH 8.8 (see Example 16).

(86) Four clones were identified in this screening ELISA. In subsequent experiments (see Examples 16-18) these four Lcn2 muteins were found to also show specific binding activity for FN7B8 in ELISA. SPR-analysis, and in immunofluorescence microscopy of ED-B positive human colon cancer cells. These Lcn2 muteins were designated as N7A (SEQ ID NO: 20). N7E (SEQ ID NO: 22). N9B (SEQ ID NO: 24), and N10D (SEQ ID NO: 26).

EXAMPLE 15: SOLUBLE PROTEIN PRODUCTION AND PURIFICATION OF LCN2 AND ITS VARIANTS

(87) See Example 7 for details.

EXAMPLE 16: MEASUREMENT OF BINDING ACTIVITY FOR ED-B IN AN ELISA

(88) To adsorb FN7B8 or FN789 onto the surface of a 96-well Nunc Maxisorp plate (Thermo Fisher Scientific, Langenselbold, Germany) 50 μl of a 100 μg/ml protein solution in TBS/E was added per well and incubated at room temperature for 2 h. Additionally, to include control proteins, blank wells were exposed to 120 μl 2% (w/v) BSA (AppliChem, Darmstadt, Germany) in TBS/ET. After three washing steps, wells were blocked with 120 μl 2% (w/v) BSA (AppliChem. Darmstadt, Germany) in TBS/ET for 2 h at room temperature and washed repeatedly before 50 μL of a dilution series of the purified Lcn2 mutein was added and incubated for 1 h. The wells were washed again and bound Lcn2 muteins were detected with 50 μL of Streptavidin-alkaline phosphatase conjugate (GE Healthcare, Munich, Germany) diluted 1:1500 in TBS/T for 1 h, followed by signal development in the presence of 50 μl 0.5 mg/ml p-nitrophenyl phosphate in 100 mM Tris/HCl. 100 mM NaCl, 5 mM MgCl.sub.2, pH 8.8. The time course of absorption ΔA/Δt at 405 nm was measured in a SpectraMax 250 reader (Molecular Devices, Sunnyvale, CA) and the data were fitted with KaleidaGraph software (Synergy software, Reading, PA) to the equation
ΔA=ΔA.sub.max×[L].sub.tot/(K.sub.D+[L].sub.tot)
whereby [L].sub.tot represents the concentration of the applied ligand conjugate and K.sub.D is the dissociation constant (Voss and Skerra (1997) Protein Eng. 10, 975-982). Lcn2 variants N7A, N9B, and N10D, respectively, were found to specifically bind to FN7B8 with K.sub.D values of 14.8 nM (N7A), 40.1 nM (N9B), 30.0 nM (N7E), and 51.2 (N10D) but not to FN789 or BSA.

EXAMPLE 17: MEASUREMENT OF BINDING ACTIVITY FOR RECOMBINANT FIBRONECTIN FN7B8 VIA SURFACE PLASMON RESONANCE (SPR)

(89) Real time analysis of Lcn2 muteins was performed on a BIAcore X system (BIAcore, Uppsala. Sweden) using HBS/ET (20 mM Hepes, pH 7.5, 150 mM NaCl, 1 mM EDTA containing 0.005% (v/v) Tween 20) as running buffer. 200 μg/ml solution of recombinant FN7B8 in 10 mM Na-acetate. pH 4.0 was immobilized on a CMD 200 m Sensorchip (XanTec bioanalytics, Duesseldorf, Germany) using standard amine coupling chemistry (Biacore, Uppsala, Sweden), resulting in a ligand density of about 500 resonance units (RU). The purified Lcn2 mutein was applied at a flow rate of 25 μl/min at concentrations of 2.5 up to 160 nM. The sensorgrams were corrected by subtraction of the corresponding signals measured for the control channel, which had been activated and blocked with ethanolamine. Kinetic data evaluation was performed by global fitting with BIAevaluation software V 4.1 (Karlsson et al. (1991) J. Immunol. Methods 145, 229-240).

(90) TABLE-US-00002 TABLE 2 Kinetic binding data of selected Lcn2 muteins for FN7B8 determined by surface plasmon resonance. Lcn2 variant k.sub.on [M.sup.−1s.sup.−1] k.sub.off [s.sup.−1] K.sub.D [M] N7A 4.6 × 10.sup.6 2.6 × 10.sup.−2 5.8 × 10.sup.−9 N7E 4.1 × 10.sup.6 3.0 × 10.sup.−2 7.2 × 10.sup.−9 N9B 2.1 × 10.sup.6 7.5 × 10.sup.−2 3.6 × 10.sup.−8 N10D 1.5 × 10.sup.6 5.8 × 10.sup.−2 3.8 × 10.sup.−8

EXAMPLE 18: IMMUNOSTAINING OF ED-B POSITIVE CACO2 CELLS

(91) Human cancer colon cells (CaCo2 cells kindly provided by H. Daniel, Technische Universität München, Germany: Pujuguet et al., Am. J. Pathol. 148, 579-592) were cultured on the Nunc Lab-Tek™ II chamber Slide™ system (4 chambers per slide; Thermo Fisher Scientific, Langenselbold. Germany) in MEM with Earle's Salts and L-Glutamine, supplemented with 10% Fetal Bovine Serum, 1×MEM non-essential amino acids, and 50 μg/ml gentamycin (PAA Laboratories, Pasching, Austria) at 37° C. in a humidified atmosphere until cell confluence was about 50 to 70%. Cell monolayers attached to the cover slide were washed with PBS (Dulbecco's without Ca.sup.2+ and Mg.sup.2+; PAA Laboratories, Pasching. Austria), followed by distilled water, and then fixed and counterstained with ice-cold methanol containing 5 μg/ml DAPI (4′,6-diamidino-2-phenylindol; Sigma-Aldrich, Munich, Germany) for 5 min.

(92) All subsequent incubations were carried out in the dark. Fixed cells were washed and incubated with 500 μl 1 μM wild type Lcn2. Lcn2 mutein, PBS or ED-B specific antibody scFv-L19 (Pini et al. (1998) J. Biol. Chem. 273, 21769-21776) for 1 h. All of these reagents were purified as Strep-tag II fusion proteins (Schmidt and Skerra, supra). Cells were washed with PBS and incubated with unlabeled Antibody StrepMABimmo (5 μg/ml in PBS; IBA. Göttingen, Germany) for 1 h, followed by 2 washing steps. Finally, specific binding to CaCo2 cells was detected using a fluorescence labeled anti-mouse IgG (H+L) F(ab′).sub.2 fragment Dylight-488 conjugate (Cell Signaling Technology, Danvers, USA), diluted 1:200 in PBS, as a secondary antibody.

(93) All Lcn2 variants and also the antibody scFv-L19 showed specific cell staining when observed under a Axiovert 40 CFL microscope (Carl Zeiss, Göttingen, Germany) whereas recombinant wild type Lcn2 and PBS revealed negligible signals in this assay.

EXAMPLE 19: GENERATION OF THE A3-SPECIFIC LCN2 MUTEINS H1GA AND H1GV VIA EXCHANGE OF A FREE CYSTEINE RESIDUE IN H1-G1 AT POSITION 36

(94) Cys36 was replaced with Ala or Val by site-directed mutagenesis. To this end, a PCR was conducted with the degenerate oligodeoxynucleotide DH-4 (SEQ ID NO: 45) and the second oligodeoxynucleotide J08rev (SEQ ID NO: 48) as well as plasmid DNA encoding the Lcn2 mutein H1-G1 as template (SEQ ID NO: 37). The amplified fragments were subcloned via BstXI on the expression plasmid phNGAL98 and sequenced to identify the individual substitution variants. Depending on the introduced amino acid exchange the resulting Lcn2 variants were named H1GA (SEQ ID NO: 49, 50) and H1GV (SEQ ID NO: 51, 52).

EXAMPLE 20: SOLUBLE PRODUCTION AND PURIFICATION OF THE NEW AB-SPECIFIC LCN2 MUTEINS H1GA AND H1GV USING E. COLI CULTURES WITH A HIGH OPTICAL DENSITY

(95) The recombinant Lcn2 and its muteins were produced by periplasmic secretion in E. coli JM83. For soluble protein expression the plasmid phNGAL98 with the corresponding BstXI insert encoding either H1GA (SEQ ID NO: 49) or H1GV (SEQ ID NO: 51) was used.

(96) Cultures were grown under agitation at 22° C. in 2 L LB medium containing 100 mg/L ampicillin (Amp). Gene expression was induced at a cell density of OD.sub.550=2.5 by adding anhydrotetracycline (aTc) to a final concentration of 0.2 mg/L. After incubation for 5 h the cells were harvested by centrifugation, resuspended in 40 mL ice-cold periplasmic fractionation buffer (0.5 M sucrose, 1 mM EDTA, 100 mM Tris-HCl pH 8.0) containing 0.1 mg/mL lysozyme and incubated on ice for 30 min. The resulting spheroplasts were sedimented by repeated centrifugation, and the supernatant containing the soluble recombinant protein was recovered.

(97) The soluble protein was affinity-purified by means of the Strep-tag II, followed by size exclusion chromatography (SEC) on a Superdex 75 HR 10/30 column using PBS buffer (4 mM KH.sub.2PO.sub.4, 16 mM Na.sub.2HPO.sub.4, 115 mM NaCl, pH 7.4). Protein purity was checked by SDS-PAGE. Protein concentrations were determined by absorption measurement at 280 nm using a calculated extinction coefficient of 27055 M−1cm−1 for the Aβ specific muteins H1GA (SEQ ID NO: 50) and H1GV (SEQ ID NO: 52).

EXAMPLE 21: MEASUREMENT OF BINDING ACTIVITY FOR MBP-Aβ40 VIA SURFACE PLASMON RESONANCE ON A BIACOREX INSTRUMENT

(98) Real time analysis of the interaction between Lcn2 muteins H1GA or H1GV and MBP-Aβ40 (SEQ ID NO: 33) was performed on a Biacore X system using PBS/T (4 mM KH.sub.2PO.sub.4, 16 mM Na.sub.2HPO.sub.4, 115 mM NaC, pH 7.4 containing 0.005% (v/v) Tween 20) as running buffer. A 15 μg/mL solution of MBP-Aβ40 from Example 2 in 10 mM Na-acetate, pH 4.5 was immobilized onto a CMD 2001 chip (Xantec, Düsseldorf, Germany) using standard amine coupling chemistry, resulting in a ligand density of 1316 resonance units (RU). The purified Lcn2 muteins H1GA and H1GV from Example 20 were applied in concentrations ranging from 4 nM to 128 nM at a flow rate of 20 μL/min. The data were double-referenced by subtraction of the corresponding signals measured for the control channel, which had been activated and blocked with ethanolamine, as well as subtraction of the measured signals for an average of buffer injections. Kinetic data were fitted globally using the BIAevaluation software V 3.0 (Karlsson et al. (1991) J. Immunol. Methods 145, 229-240).

EXAMPLE 22: MEASUREMENT OF BINDING ACTIVITY FOR A1340 VIA SURFACE PLASMON RESONANCE ON A BIACORE T100 INSTRUMENT

(99) Real time analysis of the interaction between the Lcn2 mutein H1GA and Aβ40 (SEQ ID NO: 29) was performed on a Biacore T100 system (Biacore, Uppsala, Sweden) using PBS/P (4 mM KH.sub.2PO.sub.4, 16 mM Na.sub.2HPO.sub.4, 115 mM NaCl, pH 7.4 containing 0.005% (v/v) Surfactant P20) as running buffer. A 10 μg/mL solution of Aβ40 from Example 2 in 10 mM sodium acetate pH 4.5 was immobilized onto a CM5 chip (Biacore, Uppsala, Sweden) using standard amine coupling chemistry, resulting in a ligand density of 325 RU. The purified Lcn2 mutein H1GA from Example 20 was applied in concentrations ranging from 1 nM to 32 nM at a flow rate of 30 μL/min. The dilution series of H1GA was injected with both association and dissociation times of 300 s to obtain k.sub.on information. For exact determination of the low k.sub.off rate the highest concentration was analysed using a dissociation time of 7200 s. The data were double-referenced as in Example 21. k.sub.on and k.sub.off for the binding reaction were determined from the entire data set using Biacore T100 Evaluation Software V2.0.3 for data processing and kinetic fitting. The data was globally fitted using a 1:1 binding model.

EXAMPLE 23: FUNCTIONAL ANALYSIS OF THE LCN2 MUTEIN H1GA IN A THT AGGREGATION ASSAY

(100) For the Thioflavin T (ThT) aggregation assay synthetic Aβ peptide (SEQ ID NO: 29) was dissolved in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP; Sigma-Aldrich, Steinheim. Germany) for 12 h. HFIP was evaporated under vacuum, and Aβ was dissolved in a suitable volume of H.sub.2O dd, sonicated for 15 min at 4° C., and filtrated (Costar Spin-X centrifuge tube filter cellulose acetate membrane, 0.45 μm; Corning Inc., Corning, NY). The solubilized monomeric Aβ was then immediately used for the aggregation assays.

(101) 500 μl of 1 mg/ml Aβ was incubated in the absence or presence of various Lcn2 muteins at different molar ratios or BSA in 0.5×PBS at 37° C. with stirring. Aggregation reactions were prepared in triplicates. For fluorescence measurement 20 μl of periodically removed samples was mixed with 180 μl ThT at a final concentration of 50 μM in 0.5×PBS and analysed at an excitation wavelength of 450 nm and an emission wavelength of 482 nm using a FluoroMax-3 spectrofluorimeter (HORIBA Jobin Yvon, Grasbrunn, Germany).

EXAMPLE 24: MEASUREMENT OF BINDING ACTIVITY FOR THE RECOMBINANT FIBRONECTIN SINGLE DOMAIN ED-B VIA SURFACE PLASMON RESONANCE (SPR)

(102) Real time interaction analysis of Lcn2 muteins was performed on a BIAcore X instrument using HBS/ET (20 mM Hepes, pH 7.5, 150 mM NaCl, 1 mM EDTA containing 0.005% (v/v) Tween 20) as running buffer. 100 μg/ml solution of recombinant ED-B in 10 mM Na-acetate, pH 4.0 was immobilized on a CMD 200 m Sensorchip using standard amine coupling chemistry, resulting in a ligand density of about 180 resonance units (RU). The purified Lcn2 mutein was applied at a flow rate of 25 μl/min at concentrations of 2.5 up to 160 nM. The sensorgrams were corrected by subtraction of the corresponding signals measured for the control channel, which had been activated and blocked with ethanolamine. Kinetic data evaluation was performed by global fitting with BIAevaluation software V 4.1 (Karlsson et al. (1991) J. Immunol. Methods 145, 229-240). In the case of N10D the heterogeneous analyte competing reaction model was used for data fitting, resulting in two sets of kinetic constants.

(103) TABLE-US-00003 TABLE 3 Kinetic binding data of Lcn2 muteins for ED-B determined by surface plasmon resonance. Lcn2 variant k.sub.on [M.sup.−1s.sup.−1] k.sub.off [s.sup.−1] K.sub.D [M] N7A     1.8 × 10.sup.6 0.138 1.27 × 10.sup.−7 N7E     3.1 × 10.sup.6 0.054 1.73 × 10.sup.−8 N9B   1.45 × 10.sup.6 7.62 × 10.sup.−3 5.26 × 10.sup.−9 N10D k.sub.1 = 6.21 × 10.sup.5 k.sub.−1 = 0.064   .sup.  k.sub.2 = 2.94 × 10.sup.4 k.sub.−2 = 3.35 × 10.sup.−3