WO2012146729A1 - Lasso peptides as scaffolds for peptide grafting - Google Patents

Abstract

The present invention provides novel templates for peptide grafting. The template is a modified lasso peptide wherein an amino acid sequence of the wildtype lasso peptide comprising one to five amino acids is substituted by another amino acid sequence comprising one to five amino acids. In the modified lasso peptides according to the present invention, any site of the wildtype lasso peptide can be substituted by another amino acid sequence com- prising one to five amino acids. The term "another amino acid sequence" refers to any amino acid sequence that differs from the amino acid sequence of the respec- tive site in the wildtype lasso peptide. The modified lasso peptides according to the present invention are produced by applying the natural maturation machinery for the processing of lasso peptide pre- cursor variants. These variants are generated on the DNA level of the precursor genes by site-directed mutagenesis. The lasso peptide precursor proteins are het- erologously coexpressed with the two processing enzymes and the ex- port/immunity protein which transform the precursor into the modified lasso variant that is secreted into the culture supernatant from which the modified lasso peptide is extracted and purified. The modified lasso peptides according to the present invention can be used for peptide grafting.

Description

Patent Application

Lasso peptides as scaffolds for peptide grafting

Description and technical field of the invention

Lasso peptides represent an emerging class of stable bacterial peptides with unique characteristics encouraging their application as molecular scaffolds for drug design. Herein we show that epitope grafting of the integrin binding motif RGD onto the lasso structure of microcin J25 converts the knotted peptide into a nanomolar integrin antagonist. Engineered lasso peptides can therefore be added to the toolbox for pharmacophore presentation.

State of the art

Peptides combine a high specificity for their target receptor with a generally low toxicity and are therefore a promising source for drug leads (Sato, A.K., Viswana- than, M., Kent, R.B. & Wood, C.R. Curr. Opin. Biotechnol. 17, 638-42 (2006); An- tosova, Z., Mackova, M., Krai, V. & Macek, T. Trends Biotechnol. 27, 628-35 (2009)). However, their use has been limited due to undesirable physicochemical and pharmacokinetic properties, such as variable solubility, low bioavailability and poor stability under physiological conditions (Antosova, Z., Mackova, M., Krai, V. & Macek, T. Trends Biotechnol. 27, 628-35 (2009)). To overcome these obstacles N- or C-terminal modifications (e.g. PEGylation), cyclization, introduction of unnatural or D-amino acids, N-methylation or retro inversion have been applied (Werle, M. & Bernkop-Schnurch, A. Amino Acids 30, 351 -67 (2006); Huber, T., Manzenrieder, F., Kuttruff, C.A., Dorner-Ciossek, C. & Kessler, H. Bioorg. Med. Chem. Lett. 19, 4427-31 (2009); Biron, E. et al. Angew. Chem. Int. Ed. Engl. 47, 2595-9 (2008); Chatterjee, J., Gilon, C, Hoffman, A. & Kessler, H. Acc. Chem. Res. 41 , 1331 -42

(2008) ). Besides these synthetic chemical strategies protein scaffolds can be used to present biologically active peptide epitopes benefiting from the overall stability of the scaffold (Gebauer, M. & Skerra, A. Curr. Opin. Chem. Biol. 13, 245-55 (2009)). Furthermore ultrastable ribosomally assembled peptides sharing the cyclic cystine knot (CCK) motif have been recently used as molecular frameworks exemplified by the conversion of the cyclotide kalata B1 into an vascular endothelial growth factor-A antagonist by insertion of the hexapeptide RRKRRR epitope into an intercysteine loop (Jagadish, K. & Camarero, J. A. Biopolymers 94, 61 1 -6 (2010); Gunasekera, S. et al. J. Med. Chem. 51 , 7697-704 (2008)). As the biosyn- thetic machinery of the plant-derived cyclotides is not understood in detail and therefore inaccessible to genetic manipulation and heterologous production this epitope grafting approach relies either on solid phase peptide synthesis (SPPS) or on expressed protein ligation (EPL) in vivo to generate the circular peptide back- bone followed by oxidative folding (Craik, D.J., Mylne, J.S. & Daly, N.L. Cell Mol. Life Sci. 67, 9-16 (2010); Berrade, L. & Camarero, J.A. Cell Mol. Life Sci. 66, 3909-22 (2009)).

In addition to cyclotides, bacterial lasso peptides are under discussion as molecu- lar scaffolds for drug design (Rosengren, K.J. & Craik, D.J. Chem. Biol. 16, 121 1 -2

(2009) ; Knappe, T.A., Linne, U., Robbel, L. & Marahiel, M.A. Chem. Biol. 16, 1290-8 (2009); Pavlova, O., Mukhopadhyay, J., Sineva, E., Ebright, R.H. & Seve- rinov, K. J. Biol. Chem. 283, 25589-95 (2008)). These ribosomally assembled peptides consist of 16 - 21 amino acids and share an N-terminal 8/9-residue mac- rolactam ring through which the C-terminal linear tail is threaded and trapped by steric hindrance of bulky side chains (Rebuffat, S., Blond, A., Destoumieux- Garzon, D., Goulard, C. & Peduzzi, J. Curr. Protein Pept. Sci. 5, 383-91 (2004); Rosengren, K.J. et al. J. Am. Chem. Soc. 125, 12464-74 (2003); Bayro, M.J. et al. J. Am. Chem. Soc. 125, 12382-3 (2003); Wilson, K.A. et al. J. Am. Chem. Soc. 125, 12475-83 (2003); Knappe, T.A. et al. J. Am. Chem. Soc. 130, 1 1446-54 (2008)). The currently known gene clusters of the lasso peptides microcin J25 (MccJ25) and capistruin consist of four genes, one coding for the precursor protein, two for the processing enzymes and one for the export and immunity protein (Knappe, T.A. et al. J. Am. Chem. Soc. 130, 1 1446-54 (2008); Solbiati, J.O. et al. J. Bacteriol. 181 , 2659-62 (1999)). Mutational analysis of the precursor proteins McjA of MccJ25 and CapA of capistruin revealed a high promiscuity of the biosyn- thetic machineries as well as the feasible heterologous production of various vari- ants in Escherichia coli (Knappe, T.A., Linne, U., Robbel, L. & Marahiel, M.A. Chem. Biol. 16, 1290-8 (2009); Pavlova, O., Mukhopadhyay, J., Sineva, E., Ebright, R.H. & Severinov, K. J. Biol. Chem. 283, 25589-95 (2008)). Therefore lasso peptides combine unique characteristics relevant for the application as robust scaffolds for epitope grafting (Fig. 1 ): (I) extraordinary stability against prote- olytic degradation, temperature and chemical denaturants; (II) gene-encoded lasso peptide precursor proteins; (III) gene cluster of bacterial origin allowing heterologous production in E. coli; (IV) promiscuous biosynthetic machinery tolerating various amino acid substitutions within the lasso peptide sequence. To prove the applicability of lasso structured peptides as molecular scaffolds we chose the tripeptide integrin binding motif RGD as peptide epitope to be grafted onto a stable lasso peptide structure. Integrins are a large class of heterodimeric cell surface receptors and among the RGD-binding integrins _νβ₃ and _νβ_δ subtypes have received increasing interest as therapeutic targets due to their role in tumor growth and angiogenesis (Meyer, A., Auernheimer, J., Modlinger, A. & Kessler, H. Curr. Pharm. Des. 12, 2723-47 (2006); Avraamides, C.J., Garmy- Susini, B. & Varner, J.A. Nat. Rev. Cancer 8, 604-17 (2008)).

Technical problem

The aim of the present invention is to provide novel templates for peptide grafting and a method of making said templates.

Solution of the technical problem

The aim to provide novel templates for peptide grafting is solved by modified lasso peptides wherein an amino acid sequence of the wildtype lasso peptide compris- ing one to five amino acids is substituted by another amino acid sequence comprising one to five amino acids.

Lasso peptides are ribosomally synthesized and posttranslationally modified peptides with a length of 16 to 21 amino acids. The free amino group of the first amino acid (glycin or cystein) forms an isopeptide bond to an aspartate or glutamate side chain in position 8 or 9, thus generating a lactame ring. The "tail" comprising 7 to 13 amino acids is led through this ring. In most cases, the tail is sterically fixed by large residues on both sites of the ring. In the modified lasso peptides according to the present invention, any site of the wildtype lasso peptide can be substituted by another amino acid sequence comprising one to five amino acids. The term "another amino acid sequence" refers to any amino acid sequence that differs from the amino acid sequence of the respective site in the wildtype lasso peptide.

In a preferred embodiment of the present invention, the modified lasso peptide is represented by the triple mutant MccJ25 RGD of the lasso peptide MccJ25 wherein the tripeptide sequence Gly12-lle13-Gly14 is substituted by Arg-Gly-Asp (RGD). MccJ25 RGD provides integrin antagonist properties.

As the RGD epitope should be presented on the surface of the lasso structure facilitating productive interaction with the surface exposed binding site between the two integrin subunits, the turn motif inside the threading tail was chosen as insertion site (Xiong, J. P. et al. Science 296, 151 -5 (2002)). Moreover, the installation within the turn motif should allow a presentation of the epitope in a conformation- ally restricted, kinked geometry being responsible for the superior activity and se- lectivity profiles of cyclic RGD peptides in comparison to their linear representatives (Aumailley, M. et al. FEBS Lett. 291 , 50-4 (1991 ); Heckmann, D. & Kessler, H. Methods Enzymol. 426, 463-503 (2007)). Therefore only MccJ25 was suitable as molecular scaffold, since the biosynthetic machinery of capistruin is not tolerating substitutions within the noose (Knappe, T.A., Linne, U., Robbel, L. & Marahiel, M.A. Chem. Biol. 16, 1290-8 (2009)). Consequently the tripeptide sequence Gly12-lle13-Gly14 (numbering according to MccJ25) was substituted by Arg-Gly- Asp via site-directed mutagenesis of the precursor protein McjA, which is encoded alongside the processing enzymes McjB/McjC and the export and immunity protein McjD on the pTUC202 plasmid (Fig. 2) (Solbiati, J.O. et al. J. Bacteriol. 181 , 2659-62 (1999)). HPLC-HRMS analysis of the culture supernatant of E. coli DH5cc harboring the mutated pTUC202 RGD plasmid revealed the successful maturation of the mutated precursor protein into the MccJ25 RGD triple mutant being consistent with the processing of the single mutants described by Pavlova et al. (Pavlova, O., Mukhopadhyay, J., Sineva, E., Ebright, R.H. & Severinov, K. J. Biol. Chem. 283, 25589-95 (2008)). By using a combination of XAD-16-based solid phase extraction and preparative reversed phase HPLC MccJ25 RGD could be purified to homogeneity with a yield of 0.7 mg L^"1 (Fig. 3), which is comparable to the production rate of the single mutants (Pavlova, O., Mukhopadhyay, J., Sineva, E., Ebright, R.H. & Severinov, K. J. Biol. Chem. 283, 25589-95 (2008)).

To confirm the lasso structure of the MccJ25 RGD variant, gas phase fragmentation studies were performed in comparison to the wild type lasso peptide. MS² fragmentation of MccJ25 yields in addition to the b and y fragment ions, which originate from peptide bond cleavages within the threading tail, binary peptide complex fragment ions composed of b and y ions (Fig. 6, Table 1 ). These b+y fragment ions are characteristic for the lasso structure of MccJ25 and result from the stable trapping of the tail by Phe19 and Tyr20 located on opposite sides of the macrolactam ring (Fig. 8) (Rosengren, K.J. et al. J. Am. Chem. Soc. 125, 12464- 74 (2003)). In comparison, the tandem mass spectrometric fragmentation spec- trum of MccJ25 RGD shows b and y fragment ions consistent with the mass shift resulting from the RGD substitution as well as the prototypical binary peptide complex fragment ions (Fig.7, Table 2). This fragmentation behavior in combination with the insensitivity towards carboxypeptidase Y (Fig. 4) proves unequivo- cally the lasso structure of MccJ25 RGD.

To analyze the affinity of MccJ25 RGD towards _νβ3 and _νβδ integrins the ability of the grafted lasso peptide to antagonize the vitronectin binding to the integrins was evaluated. In addition, the ability to inhibit the binding of fibronectin to ₅βι and of fibrinogen to a_{\\b 3} was assayed to gain insights into the selectivity of the engineered lasso peptide. The wild type peptide MccJ25, the linear heptapeptide P1 (Ac-FVRGDTP-NH₂) corresponding to the sequence of the turn motif in MccJ25 RGD and the cyclic pentapeptide cilengitide (cyclo[RGDf-N(Me)V-]), a known ocv integrin inhibitor currently in clinical phase III for the treatment of glioblastoma, as well as the non-peptide ccnb integrin inhibitor tirofiban served as con- trols (Dechantsreiter, M.A. et al. J. Med. Chem. 42, 3033-40 (1999); Mas-Moruno, C, Rechenmacher, F. & Kessler, H. Anticancer Agents Med. Chem. (201 1 ); Hart- man, G.D. et al. J. Med. Chem. 35, 4640-2 (1992)). MccJ25 did not show biological activity towards α_νβ3, «νβδ, «δβι and

integrins proving the molecular framework to be inactive (Tab. 7). In contrast, the installation of the RGD epitope in the turn motif of the MccJ25 scaffold resulted in a remarkable increase of its binding affinity with IC₅₀ values of 17 nM (0^3), 170 nM (θνβδ), 855 nM ( ₅βι) and 29.7 nM (cciibP3)- The linear peptide P1 displayed a high affinity towards the _νβ3 receptor (IC₅o of 43 nM), moderate affinity towards ₅βι and

integrins (654 nM and 185 nM) but no inhibitory activity towards the _νβδ integrin (Fig. 15-18). Cilengitide and tirofiban showed IC₅o values being consistent with published data and thereby proving the reliability of the assay (Dechantsreiter, M.A. et al. J. Med. Chem. 42, 3033-40 (1999); Hartman, G.D. et al. J. Med. Chem. 35, 4640-2 (1992); Goodman, S.L., Holzemann, G., Sulyok, G.A. & Kessler, H. J. Med. Chem. 45, 1045-51 (2002)). Taken together, the grafting of the bioactive RGD epitope onto the inactive MccJ25 scaffold generated a nanomolar integrin inhibitor with an at least tenfold higher affinity for _νβ3 compared to _νβδ and ₅βι integrins. However, the absence of selectivity for the platelet receptor

indicates that the grafted lasso peptide requires further modifications (e.g. mutations of the flanking amino acids) to become a candidate for clinical applications.

θνβ3 and θνβδ integrins play a crucial role in endothelial cell migration and blood vessel formation promoting tumor growth by removing waste products and by pro- viding nutrients (Avraamides, C.J., Garmy-Susini, B. & Varner, J. A. Nat. Rev. Cancer 8, 604-17 (2008)). Upon showing nanomolar affinities in vitronectin binding inhibition assays (Table 7), MccJ25 RGD was analyzed towards its influence on capillary formation of human umbilical vein endothelial cells (HUVECs) in comparison to the linear peptide P1 and cilengitide. MccJ25 RGD did not show any inhibitory effect at 1 .2 μg ml_^"1 as the expected tube formation of HUVECs on Ma- trigel substrate was observed (Fig. 20 A). However, increasing concentrations of MccJ25 RGD suppressed the formation of capillaries in a dose-responsive manner with a minimal effective concentration in the range of 5 - 10 μg ml_^"1 (Fig. 20). The linear peptide P1 , which showed nanomolar affinity towards the _νβ3 integrin (IC₅o of 43 nM) in vitro, had no effect on HUVEC tube formation up to 100 μg ml_^"1. This discrepancy points towards the stabilizing effect of the lasso peptide scaffold. To analyze this further, stability studies in human serum were conducted showing that the linear peptide P1 is completely degraded after 4 h whereas more than 50% of MccJ25 RGD are present in the serum after 30 h (Fig. 5). Thus, due to the stability of the lasso fold against proteolytic degradation the affinity of MccJ25 RGD towards _νβ3 and _νβδ integrins observed in ligand inhibition assays in vitro can be transformed into an inhibitory effect on capillary formation in cell culture assays, whereas the linear heptapeptide P1 being susceptible to proteolytic degradation only shows a productive interaction in the absence of proteases. Interest- ingly, cilengitide did not prevent the formation of capillaries up to 10 μg ml_^"1 and consequently was not more potent than the grafted lasso peptide, although significantly higher affinities for _νβ3 and _νβδ integrins (20- and 7-fold, respectively) were observed (Tab. 7). Moreover, HUVEC proliferation was not compromised by MccJ25 RGD up to a concentration of 125 μg ml_^"1. A reduction in viability was only observed with higher concentrations resulting in an extrapolated IC₅o of 500 μg ml_^"1. Thus, the suppression of tube formation by MccJ25 RGD is a specific integrin inhibitory effect without influencing the vitality of the cells. In addition, antibacterial activity studies showed that the grafted lasso peptide is in contrast to MccJ25 not able to inhibit the growth of E. coli K12 MC4100 (Fig. 19), which is in agreement with the previously described inactive single mutant MccJ25 G14D (Pavlova, O., Mukhopadhyay, J., Sineva, E., Ebright, R.H. & Severinov, K. J. Biol. Chem. 283, 25589-95 (2008)).

To investigate the influence of the RGD substitution on the lasso scaffold, the three dimensional structure of MccJ25 RGD was determined by NMR spectroscopy. The superposition of the 20 lowest energy structures in Fig. 14 A shows that according to the mass fragmentation behavior the grafted peptide adopts a well defined lasso fold confirmed by the low RMSD for the backbone (0.2 A) of the structure ensemble (Tab. 3). As intended, the RGD epitope is presented on the surface of the lasso structure in a kinked conformation of the peptide backbone enabling productive interaction with the binding site of the heterodimeric receptors and locking the most important functionalities (guanidine group of Arg12 and car- boxyl group of Asp14) in an optimized distance (Fig. 14 B). This conformational restraint and the resulting reduced flexibility of the RGD motif are most likely the explanation for the higher affinity of MccJ25 RGD compared to the linear peptide P1 which is devoid of any conformational restriction (Heckmann, D. & Kessler, H. Methods Enzymol. 426, 463-503 (2007)). Hydrogen exchange studies of MccJ25 and MccJ25 RGD in CD₃OD revealed an identical behavior, since the amide hy- drogens of Glu8, Phe19 and Tyr20 did not exchange after 4 weeks incubation in deuterated solvent indicating strong hydrogen bonding (Fig. 9 + 10). In addition, the ¹ H chemical shifts of MccJ25 and MccJ25 RGD show almost no differences (Tab. 4 + 5) and the structural alignment of the wild type and the grafted lasso peptide (Fig. 14 C) illustrates that the RGD substitution did not significantly alter the overall structure of the molecular framework demonstrating the robustness of the lasso fold of MccJ25 for epitope grafting of short peptide sequences.

In conclusion, the conversion of the lasso peptide microcin J25 into a nanomolar integrin inhibitor by RGD substitution proves that lasso structured peptides are a promising molecular scaffold for the presentation of bioactive peptide epitopes. These privileged templates are devoid of cytotoxicity and display remarkable stability under physiological conditions (Lopez, F.E., Vincent, P. A., Zenoff, A.M., Salomon, R.A. & Farias, R.N. J. Antimicrob. Chemother. 59, 676-80 (2007)). Espe- daily the accessibility of the grafted natural products via a fermentative route is a major advantage compared to synthetic strategies as it is time-saving, environment-friendly and economical. Future investigations will involve the grafting of other peptide epitopes to prove a universal application of lasso peptides as molecular scaffolds as well as further optimization of MccJ25 RGD on the basis of the NMR structure applying disulfide formation of introduced cysteine residues to rigidity the environment of the insertion site as found in class I and class III lasso peptides (Knappe, T.A., Linne, U., Xie, X. & Marahiel, M.A. FEBS Lett. 584, 785-789 (2010)). The reduction of the conformational space may also improve the selectiv- ity of the engineered lasso peptide towards distinct integrin subtypes. In addition, the introduction of nonproteinogenic amino acids into the lasso structure using orthogonal aminoacyl-tRNA synthetase/tRNA pairs remains to be explored and should expand the chemical space accessible via ribosomally assembled lasso peptides (Liu, C.C. & Schultz, P.G. Annu. Rev. Biochem. 79, 413-44 (2010)).

The modified lasso peptides are generated as follows (Fig. 1 ): The peptide epitope that is to be grafted onto the lasso peptide is introduced on DNA level into the lasso peptide precursor gene by site-directed mutagenesis. The modified lasso peptide precursor protein is expressed in E. coli and converted into the modified lasso peptide by the maturation enzymes encoded in the corresponding lasso peptide gene cluster that are coexpressed. The modified lasso peptide is exported out of the producing cell by the corresponding export and immunity protein that is also coexpressed. The modified lasso peptides can be extracted and purified from the culture supernatants of the producing cells by standard solid or liquid phase extraction in combination with reversed phase chromatography..

The modified lasso peptides according to the present invention are suitable tools for peptide grafting. Embodiments

1. Generation of plasmid pTUC202 RGD by site-directed mutagenesis Plasmid pTUC202 carrying the MccJ25 biosynthetic gene cluster was used as the template for site-directed mutagenesis (Sato, A.K., Viswanathan, M., Kent, R.B. & Wood, C.R. Curr. Opin. Biotechnol. 17, 638-42 (2006). Mutagenesis was performed using Phusion High-Fidelity DNA-Polymerase (New England Biolabs) in the presence of 8% DMSO and the RGD sequence was inserted into the microcin J25 precursor gene mcjA by the following pair of primers:

- RGD_FP: 5'-GCAGGACATGTGCCTGAGTATTTTGTGCGGGGTGATACA-

CCTATATCTTTCTATG G CTG-3 '

- RGD_RP: 5'-CAGCCATAGAAAGATATAGGTGTATCACCCCGCACAAAAT-

ACTCAGGCACATGTCCTGC-3'

Mutagenized plasmid DNA was transformed into chemically competent Escherichia coli NEB 10-beta cells (New England Biolabs). Transformed cells were plated on LB agar plates containing 34 μg ml_^"1 chloramphenicol and incubated overnight at 37 °C. Plasmids were prepared from individual transformants and analyzed via restriction mapping and subsequent DNA dideoxy sequencing (GATC Biotech).

2. Heterologous production of MccJ25 and MccJ25 RGD

Heterologous production of MccJ25 and the grafted analog MccJ25 RGD was performed as described in Chiuchiolo, M.J., Delgado, M.A., Farias, R.N. & Salomon, R.A. Growth-phase-dependent expression of the cyclopeptide antibiotic microcin J25. J. Bacteriol. 183, 1755-64 (2001 ). In brief, transformants of Escherichia coli DH5cc harboring the pTUC202 or pTUC202 RGD plasmid were cultured for 24 h at 37 °C in 500 ml_ M63 medium in 2 L baffled flasks containing 2 μg ml_^"1 thia- mine/biotin, 0.2% glucose and 34 μg ml_^"1 chloramphenicol. The insertion site of the integrin binding epitope RGD within the lasso scaffold of MccJ25 is shown in Fig. 2.

3. Extraction and purification of MccJ25 and MccJ25 RGD

The culture supernatant of a 5 L fermentation of E. coli DH5cc [pTUC202] or E. coli DH5cc [pTUC202 RGD] was obtained by centrifugation for 45 min at 4 °C at 6000 rpm. Afterwards the culture supernatant was applied to solid phase extraction us- ing ~ 5 g XAD16 resin (Sigma-Aldrich) per liter of supernatant. After incubation of the culture supernatant with XAD16 resin for 30 min the supernatant was removed by filtration and the resin was washed with 300 ml_ water and eluted with 300 ml_ methanol. The methanol eluate was evaporated to dryness and dissolved in 10 ml_ 20% methanol. 2.5 ml_ of the prepared extract were filtered using a 0.45 μιτι filter and applied onto a RP-HPLC preparative Nucleodur C18ec column (250 mm x 21 mm). In the case of MccJ25 RGD, elution was performed by applying the following gradient of water/0.1 % TFA (solvent A) and methanol/0.1 % TFA (solvent B) at a flow rate of 18 mL min^-1 : linear increase from 40% to 55% B within 30 min followed by a linear increase from 55% to 95% B within 5 min followed by holding 95% B for 1 min. The retention time of MccJ25 RGD was 28.7 min. The grafted lasso peptide could be purified to homogeneity using the aforementioned two step protocol from the culture supernatant with a yield of 0.7 mg L^"1. In the case of MccJ25, the first linear gradient increased from 60% to 70% B within 30 min. The retention time of the wild type peptide was 29.3 min and 1 mg could be purified from 1 L of the culture supernatant.

An UV chromatogram of a preparative RP-HPLC purification of MccJ25 RGD is shown in Fig. 3.

4. Carboxypeptidase Y digestion assays of MccJ25 RGD

10 - 20 μg of purified MccJ25 RGD or wild type lasso peptide were incubated in a 50 reaction in 20 mM Tris/HCI, pH 8.0, in the presence of 0.01 - 1 U car- boxypeptidase Y for 15 - 120 min at 25 °C. Subsequently the assay was stopped by heat inactivation of the protease for 5 min at 95 °C. Afterwards the assays were analyzed via HPLC-HRMS using a Nucleodur 125/2 C18ec column applying the following gradient of water/0.05% formic acid (solvent A) and acetonitrile/0.045% formic acid (solvent B) at a column temperature of 40 °C and a flow rate of 0.2 mL min^"1 : Linear increase from 10% to 95% B with 35 min followed by holding 95% B for 2 min. MccJ25 RGD showed a retention time of 16.1 min, whereas the wild type lasso peptide eluted at 17.8 min.

The stability of MccJ25 RGD against carboxypeptidase Y is shown in Fig. 4.

5. Stability studies of MccJ25 RGD and the linear heptapeptide P1 in human serum The enzymatic degradation was studied using male human serum type AB (Al- drich) and Hanks' Balanced Salt solution (HBSS) (Aldrich). The peptides were dissolved in a serum-HBBS solution (9:1 ) at a concentration of 2 mg mL^"1 and incubated at 37 °C. 50 μ\- aliquots were periodically taken at different time points and poured into 200 μί. of cold methanol to precipitate the proteins and stop the deg- radation process. Samples were cooled on ice for 30 min and next centrifuged at 10000 rpm for 15 min at 4 °C. Since the peptides were completely soluble in methanol, the supernatant was analyzed by HPLC.

The stability of the grafted lasso peptide MccJ25 RGD in human serum in comparison to the linear peptide P1 is shown in Fig. 5.

6. Mass spectrometric analysis of MccJ25 and MccJ25 RGD

The mass spectrometric characterization of the grafted lasso peptide MccJ25 RGD and the wild type peptide MccJ25 was performed with an LTQ-FT instrument (Thermo Fisher Scientific, Germany) connected to a microbore 1 100 HPLC system (Agilent, Germany). Separation of MccJ25 RGD and MccJ25 from contaminants was achieved using a 125/2 Nucleodur C18ec column (Macherey-Nagel, Germany). Following gradient of water/0.05% formic acid (solvent A) and acetoni- trile/0.045% formic acid (solvent B) was applied at a column temperature of 40 °C and a flow rate of 0.2 ml_ min^-1 : linear increase from 10% to 95% B within 35 min followed by holding 95% B for 2 min.

CID fragmentation studies within the linear ion trap were performed with purified lasso peptide samples, which were analyzed using a syringe pump at a flow rate of 10 L min_i . Usually, the doubly charged ions were selected for fragmentation in the ion trap applying normalized collision energy of 35.

Tandem mass spectrometric studies of MccJ25 and MccJ25 RGD are depicted in Figs. 6 and 7, respectively.

Table 1 : Experimental fragmentation series of MccJ25

Peptide sequences corresponding to the fragment ions are highlighted.

Fragment ion [M + H]⁺ _ca|_CU|_ated [M + H]⁺ _{o served}

C-terminal y-series

1 GGAGHVPEYFVGIGTPISFYG y 1 1 10.6 1 1 10.8

2 GGAGHVPEYFVG IGTPISFYG ys 841 .4 841 .4

3 GGAGHVPEYFVG IGTPISFYG Ί 784.4 784.6

4 GGAGHVPEYFVG IGTPISFYG Ye 683.3 683.5

5 GGAGHVPEYFVG IGTPISFYG ys 586.3 586.3

6 GGAGHVPEYFVG IGTPISFYG y4 473.2 473.3

7 GGAGHVPEYFVG IGTPISFYG ys 386.2 386.3

N-terminal b- and a-series (-H₂0)

8 GGAGHVPEYFVGIGTPISFYG is 1424.7 1424.8

9 GGAGHVPEYFVGIGTPISFYG b_M 1323.6 1323.5

10 GGAGHVPEYFVGIGTPISFYG bl3 1266.7 1267.0

11 GGAGHVPEYFVG IGTPISFYG l2 1 153.6 1 153.8

12 GGAGHVPEYFVGIGTPISFYG i i 1096.5 1096.7

13 GGAGHVPEYFVG IGTPISFYG bio 997.5 997.7

14 GGAGHVPEYFVG IGTPISFYG b₉ 850.4 850.3

15 GGAGHVPEYFVG IGTPISFYG a₈ 659.3 659.6

Sterically linked b+y-series

16 GGAGHVPEYFVGIGT(PI)SFYG bis+ 4 1896.9 1897.1

17 GGAGHVPEYFVGI(GTPI)SFYG bi3+y₄ 1738.8 1739.2

18 GGAGHVPEYFVG(IGTPI)SFYG bi2+y₄ 1625.7 1625.8

19 GGAGHVPEYFV(G IGTPI)SFYG bii +y₄ 1568.7 1568.9

20 GGAGHVPEYF(VG IGTPI)SFYG b₁₀+y₄ 1469.7 1469.8

21 GGAGHVPEY(FVG IGTPI)SFYG b₉+y₄ 1322.6 1322.9

22 GGAGHVPEYFV(G IGT)PISFYG bn +y₆ 1778.9 1779.2

23 GGAGHVPEYF(VG IGT)PISFYG bio+ye 1679.8 1680.0

24 GGAGHVPEY(FVG IGTP)ISFYG b₉+y₆ 1532.7 1532.9 Table 2: Experimental fragmentation series of MccJ25 RGD

Peptide sequences corresponding to the fragment ions are highlighted.

Fragment ion [M + H]⁺ _ca|_cu|_at_ed [M + H]⁺observed

C-terminal y-series

25 G G AG H VP EYFVRG DTPISFYG yi2 1358.7 1359.0

26 GGAGHVPEYFVRGDTPISFYG yn 121 1 .6 121 1 .9

27 GGAGHVPEYFVRGDTPISFYG yio 1 1 12.5 1 1 12.7

28 GG AG H VP EYFVRG DTPISFYG y? 784.4 784.7

29 GG AG H VP EYFVRG DTPISFYG Ye 683.3 683.6

30 GGAGHVPEYFVRGDTPISFYG y4 473.2 473.2

31 GGAGHVPEYFVRGDTPISFYG ys 386.2 386.4

N-terminal b- und a-series (-H₂0)

32 GGAGHVPEYFVRGDTPISFYG bis 1525.7 1526.0

33 GGAGHVPEYFVRGDTPISFYG l4 1424.7 1424.9

34 GGAGHVPEYFVRGDTPISFYG bl3 1309.6 1310.0

35 GGAGHVPEYFVRGDTPISFYG l2 1252.6 1252.9

36 GGAGHVPEYFVRGDTPISFYG bii 1096.5 1096.2

37 GGAGHVPEYFVRGDTPISFYG bio 997.5 997.8

38 GGAGHVPEYFVRGDTPISFYG b₉ 850.4 850.6

39 GGAGHVPEYFVRGDTPISFYG a₈ 659.3 659.2

Sterically linked b+y-series

40 GGAGHVPEYFVRGD(TPI)SFYG b₁₄+y₄ 1896.9 1897.0

41 GGAGHVPEYFV(RGDTPI)SFYG bi i +y₄ 1568.7 1568.9

42 GGAGHVPEYF(VRGDTPI)SFYG bio+y₄ 1469.6 1469.8

43 GGAGHVPEY(FVRGDTPI)SFYG b₉+y₄ 1322.6 1322.8

44 GGAGHVPE(YFVRGDTPI)SFYG b₈+y₄ 1 159.5 1 159.7

45 GGAGHVPEYFV(RGDT)PISFYG bn +y₆ 1778.9 1779.1

46 GGAGHVPEYF(VRGDT)PISFYG bio+y₆ 1679.8 1680.0

47 GGAGHVPEY(FVRGDT)PISFYG b₉+y₆ 1532.7 1532.9

7. NMR spectroscopic analysis of MccJ25 RGD

Sample for NMR measurements contained 7.2 mg of MccJ25 RGD in 280 μΙ_ of CD₃OH at 298 K. Spectra were recorded on a Bruker Avance 600 MHz spec- trometer equipped with an inverse probe with z-gradient. Samples were filled into Wilmad 3mm tubes obtained from Rototec Spintec. Temperature effect on the structure was surveyed by recording ¹ H spectra at variable temperatures. ¹ H and TOCSY spectra were recorded in CD₃OD at 298 K sequentially 0.5 h, 2 h, 24 h and 4 weeks after sample preparation. For sequential assignment, DQF-COSY, TOCSY, and NOESY experiments were performed in phase-sensitive mode using States-TPPI. TOCSY spectra were recorded with mixing time of 80 ms. NOESY spectra were taken at 100 and 250 ms mixing times. Solvent suppression was fulfilled by using excitation sculpting with gradients for DQF-COSY, TOCSY, and NOESY experiments. 1 D spectra were acquired with 65 536 data points, while 2D spectra were collected using 4096 points in the F₂ dimension and 512 increments in the ϊ_Λ dimension. For 2D spectra 32 transients were used. Relaxation delay was 2.5 s. Chemical shifts were referenced to solvent signal. All spectra were processed with Bruker TOPSPIN 2.1 . NOE cross-peaks were analyzed within the program Sparky (Goddard, T.D. & Kneller, D.J. SPARKY 3, University of California, San Francisco).

Fig. 8 depicts the steric trapping of the linear C-terminal tail within the N-terminal macrolactam ring.

Fig. 9 shows the ¹ H variable delay spectra in NH region of MccJ25 in CD₃OD at 298 K.

Fig. 10 shows the ¹ H variable delay spectra in NH region of MccJ25 RGD in CD₃OD at 298 K.

Fig. 1 1 shows the fingerprint region of the DQF-COSY spectrum of MccJ25 RGD in CD₃OH at 298 K.

Fig. 12 shows a section of the NOESY spectrum of MccJ25 RGD in CD₃OH at 298 K. NOESY mixing time was 250 ms.

Fig. 13 shows a section of the TOCSY spectrum of MccJ25 RGD in CD3OH at 298 K. TOCSY mixing time was 80 ms. Table 3: Structural statistics for the family of 20 structures selected to represent the solution structure of MccJ25 RGD

Restraining Constraints Constraint Violations

Total: 128 Distance Violations, >0.5 A: 0

Distance, i=j: 4 RMS Deviations: 0.021 A

Distance, li-j l=1 : 53 Dihedral Violations, > 5°: 0

Distance, li-j l>1 : 71 RMS Deviation: 1 .9°

Dihedral: 19 Average pairwise RMS Deviation (Gly²-Tyr²⁰)

Hydrogen Bond: 0 Backbone Atoms: 0.20 A

Constraints/Residue: 7.0 All heavy Atoms: 0.80 A Table 4: ¹H chemical shifts of MccJ25 in CD30H at 298 K amino acid NH Ha Ηβ others

Gly1 7.960 4.32

3.54

Gly2 9.015 4.22

3.89

Ala3 8.562 4.68 1.306

Gly4 7.773 4.06

3.53

His5 7.600 4.63 3.30

2.93

Val6 8.848 4.70 1.79 γΟΗ₃: 1.099; 0.855

Pro7 4.23 pro-S 1.80 pro-S γ 2.25; pro-R γ 2.02 pro-R 1.70 pro-R 54.14; pro-S 53.83

Glu8 8.362 4.40 1.74 γΟΗ₂: 2.00; 1.90

1.65

Tyr9 7.350 4.56 2.88 H₂,₆: 6.94; H₃,₅: 6.59

2.66

PhelO 8.307 4.86 2.95 H₂,₆: 7.06; H₃,₅: 7.15

2.77 H₄: 7.07

Valll 8.277 4.30 2.14 yCH₃: 0.945; 0.912

Glyl2 8.160 4.00

3.75

Ilel3 8.329 4.16 1.95 γΟΗ₂: 1.55; 1.21

γΟΗ₃: 0.932

5CH₃: 0.889

Glyl4 8.456 4.06 1.907

3.63 1.566

Thrl5 7.911 4.75 4.12 γΟΗ₃: 1.210

Prol6 4.50 pro-S 1.85 pro-S γ 2.17; pro-R γ 1.98 pro-R 1.74 pro-R 54.03; pro-S 5 3.85

Ilel7 7.986 4.44 1.87 γΟΗ₂: 1.43; 1.08

γΟΗ₃: 0.847

5CH₃: 0.851

Serl8 7.365 4.40 4.10

3.86

Phel9 8.902 5.46 2.60 H_2j6: 6.87; H₃,₅: 7.15

2.60 H₄: 7.07

Tyr20 9.523 4.94 3.02 H_2j6: 6.95; H₃,₅: 6.70

3.02

Gly21 8.594 3.80

3.76 Table 5: ¹H chemical shifts of MccJ25 RGD in CD30H at 298 K amino acid NH Ha Ηβ others

Gly1 7.873 4.200

3.566

Gly2 9.034 4.232

3.856

Ala3 8.511 4.658 1.312

Gly4 7.761 4.050

3.484

His5 7.381 4.694 3.245 H₂: 8.706; H₄: 7.340

2.962

Val6 8.682 4.690 1.76 7CH3: 1.114; 0.852

Pro7 4.278 1.878 γΟΗ₂: 2.209; 1.965

1.676 5CH₂: 4.124; 3.796

Glu8 8.263 4.330 1.755 γΟΗ₂: 1.997; 1.852

1.652

Tyr9 7.562 4.479 2.897 H_¾6: 6.92; H_3,5: 6.614

2.690

Phe10 8.043 4.707 2.972 H_¾6: 7.075; H_3,5: 7.184

2.819

VaM 1 8.071 4.181 2.030 γΟΗ₃: 0.915; 0.909

Arg12 8.262 4.170 1.851 γΟΗ₂: 1.651

1.769 5CH₂: 3.217

NH: 7.418

Gly13 8.526 4.000

3.645

Asp 14 8.392 4.759 2.874

2.808

Thr15 7.965 4.761 4.182 1.242

Pro16 4.461 1.998 γΟΗ₂: 2.097; 1.991

1.805 5CH₂: 3.936; 3.848

Ile17 7.858 4.357 1.840 γΟΗ₃: 0.823

γΟΗ₂: 1.413

5CH₃: 1.056

Ser18 7.417 4.381 3.994

3.829

Phe19 8.815 5.447 2.575 H_¾6: 6.868; H_3,5: 7.085

2.554

Tyr20 9.493 4.918 3.024 H_¾6: 6.966; H_3,5: 6.707

2.977

Gly21 8.627 3.853 Table 6: J coupling constants of MccJ25 RGD in CD30H at 298 K amino acid peak positions assignment J (Hz)

Glyl 7.873-4.200 8.5

Glyl 7.873-3.566 5.7

Glyl 4.2000-3.566 18.5

Gly2 9.034-4.232 7.2

Gly2 9.034-3.856 4.6

Gly2 4.232-3.856 16.8

Ala3 8.511-4.658 8.6

Ala3 4.658-1.312 7.3

Gly4 7.761-4.050 8.2

Gly4 7.761-3.484 4.6

Gly4 4.050-3.484 18.5

His5 7.389-4.701 10.1

His5 4.701-3.244 3.5

His5 4.701-2.916 6.1

His5 3.324-2.916 16.1

Val6 8.682-4.690 8.9

Val6 4.690-1.760 not observed

Val6 1.760-1.114 8.2

Val6 1.760-0.852 7.9

Glu8 8.263-4.330 9.1

Tyr9 7.562-4.479 not observed

Tyr9 4.479-2.897 7.0

Tyr9 4.479-2.690 5.0

PhelO 8.043-4.707 9.0

PhelO 4.707-2.972 5.7

PhelO 4.707-2.819 5.1

Valll 8.071-4.181 9.5

Valll 4.181-2.030 not observed

Argl2 8.262-4.170 8.2

Glyl3 8.526-4.000 6.9 amino acid peak positions assignment J [Hz)

Glyl3 8.526-3.645 5.7

Glyl3 4.000-3.645 17.3

Aspl4 8.392-4.759 8.7

Aspl4 4.759-2.874 5.1

Aspl4 4.759-2.808 6.0

Thrl5 7.965-4.761 7.9

Thrl5 4.182-1.242 7.2

Ilel7 7.858-4.357 9.7

Serl8 7.417-4.381 8.8

Serl8 4.381-3.994 5.52

Serl8 4.381-3.829 5.03

Phel9 8.815-5.447 10.8

Phel9 5.447-2.575 not observed

Phel9 5.447-2.554 not observed

Tyr20 9.493-4.918 8.1

Tyr20 4.918-3.024 not observed

Tyr20 4.918-2.977 not observed

Gly21 8.627-3.853 6.5

Fig. 14 shows the NMR structure of the grafted lasso peptide MccJ25 RGD.

8. Synthesis of Ac-Phe-Val-Arg-Gly-Asp-Thr-Pro-NH₂ (P1)

The peptide was manually synthesized in solid-phase using the Fmoc strategy. Solvents and soluble reagents were removed by suction. Washings between de- protection, couplings, and subsequent deprotection steps were carried out with NMP and CH₂CI₂, using 10 mL of solvent/g of resin each time. The Fmoc group was removed by treatment with piperidine-NMP (1 :4, v/v) for 2 x 10 min. Couplings and washes were performed at 25 °C. Couplings were monitored using Kaiser or chloranil methods (Vazquez, J., Qushair, G. & Albericio, F. Qualitative col- orimetric tests for solid phase synthesis. Methods Enzymol. 369, 21 -35 (2003)). Fmoc-Rink Amide MBHA resin (200 mg, 0.65 mmol g^"1) was placed in a 10 ml_- polypropylene syringe fitted with a polyethylene filter disk. After Fmoc removal, Fmoc-amino acids (3 equiv) were coupled with TBTU (3 equiv), HOBt (3 equiv) and DIEA (6 equiv) in NMP for 2 h. Alternatively, couplings were performed using HATU (3 equiv), HOAt (3 equiv) and DIEA (6 equiv) in NMP for 1 h. Side chains of Fmoc-amino acids were protected as follows: Thr and Asp were protected with the tert-butyl group (£Bu) and Arg with the 2,2,4,6,7-pentamethyldihydrobenzofuran-5- sulfonyl group (Pbf). The peptide was finally /V-acetylated with Ac₂0-DIEA-NMP (1 :2:7) for 30 min. For the deprotection of side-chain groups and concomitant cleavage of the peptide from the support, the resin was washed with CH₂CI₂ (3 x 1 min), dried, and treated with TFA-H₂0-TIS (95:2.5:2.5) for 1 .5 h. TFA was then removed by evaporation with nitrogen, and the peptide was precipitated with cold diethylether, dissolved in H₂0-MeCN (1 :1 ) and then lyophilized. The crude peptide was purified by semi-preparative HPLC (from 10 to 50 % MeCN over 20 min, flow rate 8 mL min^"1). Characterization of P1 : HPLC (t_R 14.03 min, from 10 to 50% MeCN over 30 min, purity >99%), ESI-MS {m/z calcd. for C₃₇H₅₇N₁₁OH 831 .5, found 832.6 [M+H]⁺, 854.5 [M+Na]⁺, 870.3 [M+K]⁺). 9. In vitro inhibition of integrin - extracellular matrix (ECM) protein binding

The inhibiting activity and integrin selectivity of the integrin antagonists were determined in a solid-phase binding assay based on previously reported methods with some modifications (Marugan, J.J. et al. Design, synthesis, and biological evaluation of novel potent and selective alphavbeta3/alphavbeta5 integrin dual inhibitors with improved bioavailability. Selection of the molecular core. J. Med. Chem. 48, 926-34 (2005); Stragies, R. et al. Design and synthesis of a new class of selective integrin alpha5beta1 antagonists. J. Med. Chem. 50, 3786-94 (2007)). Human integrins _νβ3 and _νβδ were purchased from Chemicon Corpora- tion/Millipore, ₅βι from R&D Systems and αι^β₃ from Enzyme Research Laboratories. Vitronectin was purchased from Chemicon Corporation/Millipore, fibronectin from Sigma and fibrinogen from Calbiochem. For the integrins ₅βι and αι^β₃ the binding was visualized using antibodies from BD Biosciences (mouse anti-human CD49e for ₅βι and mouse anti-human CD41 b for α^β₃) and Sigma (anti-mouse IgG-peroxidase). Peroxidase development was performed using the substrate solution 3,3,5,5'-tetramethylethylenediamine (TMB from Seramun Diagnostic GmbH) and 3 M H₂S0₄ for stopping the reaction. Alternatively, for the integrins _νβ3 and θνβ_δ the binding was detected using the conjugate neutravidin-horseradish peroxi- dase (HRP) from Pierce and the HRP substrate o-phenylenediamine hydrochloride (OPD) from Sigma. The absorbance (450, 492 nm) was recorded with a PO- LARstar Galaxy plate reader (BMG Labtechnologies). Every concentration was analyzed by duplicate and the resulting inhibition curves were analyzed using OriginPro 7.5G software, the turning point describes the IC₅o value. Each plate contained either cilengitide or tirofiban as internal controls (Dechantsreiter, M.A. et al. N-Methylated cyclic RGD peptides as highly active and selective al- pha(V)beta(3) integrin antagonists. J. Med. Chem. 42, 3033-40 (1999); Hartman, G.D. et al. Non-peptide fibrinogen receptor antagonists. 1 . Discovery and design of exosite inhibitors. J. Med. Chem. 35, 4640-2 (1992)). Blocking and binding steps were always performed with TS buffer (20 mM Tris-HCI pH 7.5, 150 mM NaCI, 1 mM CaCI₂, 1 mM MgCI₂, and 1 mM MnCI₂) containing 1 % BSA (TSB- buffer). Washings after each incubation step were done with PBST buffer (10 mM Na₂HP0₄ pH 7.5, 150 mM NaCI, and 0.01 % Tween 20). Table 7: Affinity of MccJ25 RGD for α_νβ₃, α_νβ₅, α₅β₁ and a_Mbp₃ integrins

Shown are the IC50 values determined in isolated receptor binding assays. For comparison the affinity of MccJ25, of cilengitide, of tirofiban and of the linear heptapeptide P1 corresponding to the peptide sequence of the replacement site were determined.

peptide θνβ₃ (ICso in nM) α_νβ₅ (ICso in nM) α₅βι (ICso in nM) o iibP3 (ICso in nM)

MccJ25 >10000 >10000 >10000 >10000

MccJ25 RGD 17 ± 9 170 ± 37 855 ± 191 29.7 ± 2.9

Ac-FVRGDTP-

43 ± 14 >10000 654 ± 122 185 ± 146 NH₂

cilengitide 0.92 ± 0.19 25 ± 8 8.2 ± 0.6 206 ± 92 tirofiban 0.6 ± 2.3

9.a) Vitronectin-a_v 3 assay

Flat-bottom 96-well ELISA plates (Brand) were coated overnight at room tempera- ture with 100 L/well of 0.4 g ml_^"1 human _νβ₃ in TS buffer. After emptying the plate the wells were blocked for 2 h at 30 °C with 150 pL/well of TSB. Plates were then washed three times with 200 L/well of PBST. Next, test compounds and cilengitide (Orpegen Pharma) as internal control (0.0003-10 μΜ) were mixed with 1 g ml_^"1 of human vitronectin, which had been biotinylated with sulfo-NHS-LC- LC-biotin (Pierce, 20:1 molar ratio), and 100 L/well of these solutions were incubated for 2 h at 30 °C. After washing the plates five times, 100 pL/well of 0.25 pg ml_^"1 of neutravidin-HRP were added to the plate and incubated for 1 h at 30 °C. After another 5-fold wash, the plate was developed by adding 100 L/well of OPD solution, obtained from dissolving substrate tablets in buffer (24 mM sodium cit- rate, 50 mM Na₂HP0₄ pH 5.0, 0.012% H₂0₂). After 15 min the reaction was quenched with 3 M H₂S0₄ and the binding analyzed at 492 nm as described above.

Fig. 15 shows the ELISA based in vitro inhibition of protein-protein binding for the c^-vitronectin assay. 9.b) Vitronectin-o v ₅ assay

The assay is similar to the vitronectin-c^ assay. In brief, flat-bottom 96-well ELISA plates were coated overnight at room temperature with 100 L/well of 1 .0 pg ml_^"1 human _νβ_δ in TS buffer. After blocking and washing the plates, test compounds and cilengitide were mixed with 1 μg mL^"1 of human biotinylated- vitronectin and were incubated for 2 h at 30 °C. The binding was visualized by incubation with neutravidin-HRP and further oxidation of the OPD substrate. IC₅o values were calculated as previously described.

Fig. 16 shows the ELISA based in vitro inhibition of protein-protein binding for the c^-vitronectin assay.

9.c) Fibronectin-a₅ i assay

Flat-bottom 96-well ELISA plates were coated overnight at 4 °C with 100 μυννβΙΙ of 0.50 pg mL^"1 of fibronectin in 15 mM of Na₂C0₃, 35 mM NaHC0₃, pH 9.6 (carbonate buffer). Plates were subsequently blocked for 1 h at room temperature. Next, 1 .0 μg mL^"1 of soluble integrin ₅βι and a serial dilution of integrin inhibitors and the control cilengitide were incubated in the coated wells for 1 h at room temperature. After washing three times, the plate was treated with 100 L well of primary antibody (CD49e) at 1 .0 pg mL^"1 (1 :500 dilution) and secondary antibody (anti-mouse IgG-peroxidase) at 2.0 g mL^"1 (1 :385 dilution) for 1 h each at room temperature. After this treatment the plate was washed three times and the bind- ing visualized with TMB. For this substrate the oxidation was left for only 5 min and the absorbance measured at 450 nm. IC₅o determination was done as explained above.

Fig. 17 shows the ELISA based in vitro inhibition of protein-protein binding for the α₅βι -fibronectin assay. 9.d) Fibrinogen-ociibP3 assay

The assay is similar to the fibronectin-cc₅ i assay. In this case, flat-bottom 96-well ELISA plates were coated overnight at 4 °C with 100 μΙ_Λ/νβΙΙ of 10.0 g mL-1 of fibrinogen in carbonate buffer. After blocking the plates, 2.5 μg mL-1 of soluble integrin l Ι β3 and a serial dilution of integrin inhibitors and the control molecules cilengitide and tirofiban were incubated in the coated wells for 1 h at room temperature. The plate was then washed three times and subsequently treated with 100 pL/well of primary antibody (CD41 b) at 2.0 pg mL-1 (1 :250 dilution) and sec- ondary antibody (anti-mouse IgG-peroxidase) at 1 .0 g mL-1 (1 :770 dilution) for 1 h each at room temperature. The binding was visualized as explained for ₅βι . Fig. 18 shows the ELISA based in vitro inhibition of protein-protein binding for the ciii_{b 3}-fibrinogen assay.

10. Analysis of the antibacterial activity of MccJ25 RGD

The antibacterial acitivity of MccJ25 RGD was analyzed in comparison to MccJ25 by a radial diffusion assay against E. coli K12 MC4100. A bacteria soft agar layer was prepared by inoculating 10 mL LB soft agar medium (6.5 g L^"1 agar) with 100 - 250 μ\- inoculum of an E. coli K12 MC4100 culture in the exponential growth phase to reach 10⁷ cfu mL^"1. The bacterial suspension was deposited onto a 20 mL LB agar layer (15 g L^"1 agar) in petri dishes. After solidification, 1 nmol MccJ25 and 1 - 10 nmol MccJ25 RGD diluted in Milli-Q water were laid over the bacterial layer. Plates were analyzed for the presence of inhibition halos after incubation for 16 h at 37 °C.

Fig. 19 shows the inhibitory effect of 1 nmol Mccj25 and 10 nmol MccJ25 RGD towards the growth of Escherichia co//^' K12 MC4100 by a radial diffusion assay. 11. Analysis of the influence of MccJ25 RGD on capillary formation

Inhibition of capillary formation was tested in vitro using human umbilical vein endothelial cells (HUVEC) from LONZA. Frozen Matrigel (BD Biosciences) was thawed at -4 °C overnight, and after mixing it with an equal volume of cold EBM-2 medium (LONZA), 35 μΙ_ aliquots were distributed in wells of a 96-well plate placed on ice. The plates were kept at 37 °C then for 1 h to let the Matrigel solidify. 25 μΙ_ of a suspension of HUVECs (8 x 10⁵ mL^"1) that were harvested from a grown flask by trypsinization were given to each well, to which serial dilutions of the compounds were added. Plates were incubated at 37 °C and 10% CO₂ overnight, and the tube formation judged under the microscope with a 5 x lens and documented using an attached CCD camera.

Fig. 20 shows the inhibitory effect on capillary formation of MccJ25 RGD.

12. Proliferation inhibition studies

60 μ\- of serial dilutions of the test compounds were added to 120 μί. aliquots of a cell suspension of HUVECs (50.000 mL^"1) in 96-well plates and incubated at 37 °C and 10% CO₂ for 5 d. MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] was used to measure growth and viability of cells which are capable of reducing it to a violet formazan product. 20 μί. MTT in phosphate buffered saline (PBS) were added to a final concentration of 0.5 mg mL^"1. After 2 h the precipitate of formazan crystals was centrifuged, and the supernatant discarded. The precipi- tate was washed with 100 μί. PBS and dissolved in 100 μί. isopropanol containing 0.4% hydrochloric acid. The microplates were measured at 590 nm using an ELI- SA plate reader. All experiments were carried out in two parallels; the percentage of metabolic activity per well was calculated as the mean with respect to the controls set to 100%. 13. Structure determination of MccJ25 RGD

Assignments were obtained by standard procedures (Wuthrich, K. NMR of Protein and Nucleic Acids, (Wiley, New York, 1 986)). A combination of DQF-COSY (Fig. 1 1 ) and NOESY (Fig. 1 2) produced sequential assignments (i.e. all ccH and NH and their sequence in the backbone), and a combination of DQF-COSY and TOCSY (Fig. 1 3) resulted in assignments of the side chains (Tab. 5). Structure calculations were performed with the program CYANA 2.1 (Herrmann, T., Guntert, P. & Wuthrich, K. Protein NMR structure determination with automated NOE as- signment using the new software CANDI D and the torsion angle dynamics algorithm DYANA. J. Mol. Biol. 319, 209-227 (2002)). The internal linkage was realized by putting a distance constraint between N of Gly¹ and C5 of Glu⁸ to be 1 .33 A. NOE cross-peaks of 1 00 ms NOESY spectrum were converted into distance constraints manually. In this way, 1 28 distance constraints were obtained, with 53 for backbone, 71 for long-range, and 75 are for the side-chains (Tab. 3). Thus there were average 7.0 distance constraints per residue. Torsion angles φ were restrained to -1 20 ° ± 30 ° for Ala³, His⁵, Val⁶, Glu⁸, Phe¹⁰, Val¹¹ , Arg¹², Asp¹⁴, lie¹⁷, Ser¹⁸, Phe¹⁹, and Tyr²⁰ with ³JHH_CC > 8 HZ (Tab. 6). The conformation and the range of the corresponding side chain torsion angle χ¹ were established using standard methods, shown as follows: His⁵ (t²g³), Tyr⁹ (g²g³), Phe¹⁰ (t²g³), Asp¹⁴ (g²t³), and Ser¹⁸ (g²g³) (Wagner, G. Nmr Investigations of Protein-Structure. Prog. Nucl. Magn. Reson. Spectrosc. 22, 1 01 -1 39 (1 990)). For t²g³, g²t³, and g²g³ conformations around the Ccc-C bond, the torsion angle χ¹ was constrained in the range of -60 ± 30 °, 1 50 ± 30 °, and 60 ± 30 °, respectively.

The above mentioned constraints were used in the simulated annealing protocol for CYANA 2.1 calculation. The calculation initiated with 200 random conformers and the resultant structures were engineered by the program package Sybyl 7.3 to include the covalent linkage between the nitrogen of Gly¹ and C5 of Glu⁸, followed by energy minimization under NMR constraints using TRI POS force field within Sybyl 7.3. Thus, on the basis of low energies and minimal violations of the experimental data, a family of 20 structures was chosen. These 20 energy- minimized conformers show an average root-mean-square deviation (RMSD) for the backbone of 0.2 A and are kept to represent the solution structure of the grafted lasso peptide (Tab. 3). The quality of these structures was evaluated using the program PROCHECK (Laskowski, R.A., Rullmann, J.A.C., MacArthur, M.W., Kaptein, R. & Thornton, J.M. AQUA and PROCHECK-NMR: Programs for checking the quality of protein structures solved by NMR. J. Biomol. NMR 8, 477-486 (1996)). The structure has been deposited to the Biological Magnetic Resonance Bank (BMRB) (accession number: 21000).

Figure legends

Fig. 1 Fig. 1 shows a biologically active peptide epitope (arrows, a) consisting of pro- teinogenic amino acids that is inserted into the lasso peptide scaffold by site- directed mutagenesis of the precursor gene {lpA^#). The mutated lasso peptide precursor protein (LpA^#) is converted into the grafted lasso peptide by the biosyn- thetic machinery (LpB/LpC) and subsequently secreted into the culture super- natant by the export and immunity protein (LpD) using Escherichia coli as host system. The grafted lasso peptide can be extracted from the culture supernatant and combines the stability of the scaffold with the biological activity of the peptide epitope.

Fig. 2

Insertion site of the integrin binding epitope RGD within the lasso peptide scaffold of MccJ25. (A) NMR structure of MccJ25 shown as sticks. The isopeptide bond between Gly1 and Glu8 (a) results in an 8-residue macrolactam ring (b) through which the linear C-terminal tail (c) is threaded. The Gly12-lle13-Gly14 sequence (d) located within the turn motif of the threading tail was chosen as insertion site for the grafting of the RGD epitope. (B) Primary structures of MccJ25 and of the grafted lasso peptide MccJ25 RGD.

Fig. 3

Preparative purification of MccJ25 RGD. UV chromatogram of a preparative RP- HPLC purification of MccJ25 RGD. The microcin J25 variant shows a retention time of 28.7 min. The ESI-FT-MS spectrum matches the calculated m/z ([M + 2H]²⁺caicuiated = 1 104.5298; [M + 2H]²⁺ ₀bse_rved = 1 104.5314; deviation 1 .4 ppm). Fig. 4

Stability of MccJ25 RGD against carboxypeptidase Y digestion. Incubation of MccJ25 RGD in the presence of 1 U carboxypeptidase Y for 2 h at 25 °C does not lead to a degradation of the peptide proving its insensitivity towards carboxytermi- nal degradation which is consistent with a lasso structure.

Fig. 5

Stability of the grafted lasso peptide MccJ25 RGD in human serum in comparison to the linear peptide P1 . The relative amounts of MccJ25 RGD (squares) and of the peptide P1 (triangles) are plotted versus time.

Fig. 6

Tandem mass spectrometric studies of MccJ25. MS² spectrum of the m/z 1 054 doubly protonated species of MccJ25. By way of illustration, the sterically linked bi₃+y₄ fragment ion (m/z 1 739.2) is depicted schematically.

Fig. 7 Tandem mass spectrometric studies of MccJ25 RGD. MS² spectrum of the m/z 1 104.5 doubly protonated species of MccJ25 RGD. By way of illustration, the sterically linked b-i₄+y₄ fragment ion (m/z 1 897.0) is depicted schematically. Fig. 8

Steric trapping of the linear C-terminal tail within the N-terminal macrolactam ring.

(A) NMR structure of MccJ25. Backbone is shown as cartoon and the side chains are depicted as sticks. The surface of Phe19 and Tyr20 are shown in light grey.

(B) Solvent accessible surface of the N-terminal macrolactam ring is additionally shown in medium grey.

Fig. 9

¹H variable delay spectra in NH region of MccJ25 in CD₃OD at 298 K

Fig. 10

¹H variable delay spectra in NH region of MccJ25 RGD in CD₃OD at 298 K

Fig. 11

Fingerprint region of the DQF-COSY spectrum of MccJ25 RGD in CD₃OH at 298 K

Fig. 12

Section of the NOESY spectrum of MccJ25 RGD in CD₃OH at 298 K. NOESY mixing time was 250 ms.

Fig. 13

Section of the TOCSY spectrum of MccJ25 RGD in CD3OH at 298 K. TOCSY mixing time was 80 ms. Fig. 14

NMR structure of the grafted lasso peptide MccJ25 RGD.

(A) Superposition of the 20 lowest energy structures of MccJ25 RGD shown as sticks. The isopeptide bond (a) between Gly1 and Glu8 generates an 8-residue macrolactam ring (b) through which the C-terminal tail (c) is threaded.

(B) Average structure of MccJ25 RGD in solution. The solvent accessible surface is shown in light grey (d) and the grafted RGD epitope is shown in medium grey (e).

(C) Structural alignment of MccJ25 (light grey, f) and MccJ25 RGD (medium grey, g). Arrows (h) depict the grafted RGD sequence, and arrows (i) point to the substituted GIG sequence in MccJ25 is shown in green. Structural alignment was performed using RAPIDO and yielded an RMSD of 1 .22 A.

Fig. 15

ELISA based in vitro inhibition of protein-protein binding - c^-vitronectin assay. Data points are shown for MccJ25 (a), peptide P1 (b) and MccJ25 RGD (c). The fit is presented as colored line and was obtained using Origin (sigmoidal fit). IC₅o values derived from the sigmoidal fit are presented in the figure.

Fig. 16

ELISA based in vitro inhibition of protein-protein binding - av s-vitronectin assay. Data points are shown for MccJ25 (a), peptide P1 (b) and MccJ25 RGD (c). The fit is presented as colored line and was obtained using Origin (sigmoidal fit). IC₅o values derived from the sigmoidal fit are presented in the figure. Fig. 17

ELISA based in vitro inhibition of protein-protein binding - a₅ i-fibronectin assay. Data points are shown for MccJ25 (a), peptide P1 (b) and MccJ25 RGD (c). The fit is presented as colored line and was obtained using Origin (sigmoidal fit). IC₅o values derived from the sigmoidal fit are presented in the figure.

Fig. 18

ELISA based in vitro inhibition of protein-protein binding - anb 3-fibrinogen assay. Data points are shown for MccJ25 (a), peptide P1 (b) and MccJ25 RGD (c). The fit is presented as colored line and was obtained using Origin (sigmoidal fit). IC₅o values derived from the sigmoidal fit are presented in the figure. For MccJ25 neither a sigmoidal nor a linear fitting were possible.

Fig. 19 Analysis of the antibacterial activity of MccJ25 RGD in comparison to MccJ25 by a radial diffusion assay. Shown is the inhibitory effect of 1 nmol Mccj25 and 10 nmol MccJ25 RGD towards the growth of Escherichia co//^' K12 MC4100.

Fig. 20

Inhibitory effect on capillary formation of MccJ25 RGD. Microscopic images of tube formation of HUVECs on Matrigel substrate. (A) Capillary formation of HU- VECs in the presence of 1 .2 μg ml_^"1 MccJ25 RGD. Red arrows point to capillaries formed. Red asterisks mark groups of single cells. (B) Reduced tube formation in the presence of 1 1 μg ml_^"1 MccJ25 RGD. (C) Formation of capillaries is completely inhibited at 98 μg ml_^"1 MccJ25 RGD. Only cell aggregation is observed.

Claims

Claims

A template for peptide grafting characterised in that the template is a modified lasso peptide wherein an amino acid sequence of the wildtype lasso peptide comprising one to five amino acids is substituted by another amino acid sequence comprising one to five amino acids.

A template according to claim 1 wherein the modified lasso peptide is represented by the triple mutant MccJ25 RGD of the lasso peptide MccJ25 wherein the tripeptide sequence Gly1 2-lle1 3-Gly14 is substituted by Arg-Gly-Asp (RGD).

A method for making the modified lasso peptide according to claim 1 , characterised by consecutively carrying out the following steps.

a) generation of a modified lasso peptide precursor gene by site directed mutagenesis that contains the amino acid substitutions/insertions on the DNA level

b) heterologous expression of the lasso peptide precursor variant together with the two processing enzymes and the export/immunity protein of the corresponding lasso peptide biosynthetic gene cluster resulting in the production and export of the modified lasso peptide

c) extraction and purification of the modified lasso peptide form the culture supernatant.

A method for making the lasso peptide MccJ25 RGD according to claim 2, characterised by consecutively carrying out the following steps:

a) generation of plasmid pTUC202 RGD by site-directed mutagenesis, b) heterologous production of MccJ25 RGD by cultivating Eschericia coli DH5a harboring the pTUC202 RGD plasmid,

c) extraction and purification of MccJ25 RGD.

5. Use of the lasso peptide MccJ25 RGD as an antagonist of the RGD-binding integrins α_νβ3, α_νβ5, α₅βι , and awbfa-

6. Use of modified lasso peptides according to one of the claims 1 to 5 for tide grafting.