Targeting angiogenesis: Structural characterization and biological properties of a de novo engineered VEGF mimicking peptide Luca Domenico D’Andrea*, Guido Iaccarino†, Roberto Fattorusso‡, Daniela Sorriento†, Concetta Carannante*, Domenica Capasso*, Bruno Trimarco†, and Carlo Pedone*§ *Istituto di Biostrutture e Bioimmagini, Consiglio Nazionale delle Ricerche, Via Mezzocannone 16, 80134 Napoli, Italy; †Dipartimento Medicina Clinica Scienze Cardiovascolari ed Immunologiche, Universita` degli Studi di Napoli ‘‘Federico II’’, Via Pansini 5, 80131 Napoli, Italy; and ‡Dipartimento di Scienze Ambientali, Seconda Universita` di Napoli, Via Vivaldi 43, 81100 Caserta, Italy
Modulating angiogenesis is an attractive goal because many pathological conditions depend on the growth of new vessels. Angiogenesis is mainly regulated by the VEGF, a mitogen specific for endothelial cells. In the last years, many efforts have been pursued to modulate the angiogenic response targeting VEGF and its receptors. Based on the x-ray structure of VEGF bound to the receptor, we designed a peptide, QK, reproducing a region of the VEGF binding interface: the helix region 17–25. NMR conformation analysis of QK revealed that it adopts a helical conformation in water, whereas the peptide corresponding to the ␣-helix region of VEGF, VEGF15, is unstructured. Biological assays in vitro and on bovine aorta endothelial cells suggested that QK binds to the VEGF receptors and competes with VEGF. VEGF15 did not bind to the receptors indicating that the helical structure is necessary for the biological activity. Furthermore, QK induced endothelial cells proliferation, activated cell signaling dependent on VEGF, and increased the VEGF biological response. QK promoted capillary formation and organization in an in vitro assay on matrigel. These results suggested that the helix region 17–25 of VEGF is involved in VEGF receptor activation. The peptide designed to resemble this region shares numerous biological properties of VEGF, thus suggesting that this region is of potential interest for biomedical applications, and molecules mimicking it could be attractive for therapeutic and diagnostic applications. helical peptides 兩 peptide design 兩 VEGF 兩 angiogenesis 兩 NMR spectroscopy
A
ngiogenesis is a phenomenon intimately associated with endothelial cell (EC) migration and proliferation. During embryonic development, ECs rapidly proliferate, thereby forming new blood vessels. In adult life, however, EC turnover is very low but under a variety of pathological conditions such as chronic ischemia, cancer, proliferative retinopathy, and rheumatoid arthritis, these cells detach from their neighbors, migrate, proliferate, and subsequently form a new branch vessel with a lumen (1). The mediators of this proliferation have been identified in a series of vascular growth factors that can be released in response to many intermediates such as cytokines. VEGF is a potent angiogenic factor, a mitogen specific for vascular ECs and plays a major role in angiogenesis. VEGF and its receptors are overexpressed in pathological angiogenesis, making this system a potential target for therapeutic and diagnostic applications (2, 3). VEGF is a homodimeric protein belonging to the cystine knot growth factor family. It is encoded by a single gene that is expressed in four different isoforms because of different splicing events. VEGF165, the most abundant isoform, is a glycoprotein that binds to heparin with high affinity. The biological function of VEGF is mediated through binding to two tyrosine kinase receptors, the kinase domain receptor (KDR) and the Fms-like tyrosine kinase (Flt-1). VEGF induces receptor dimerization
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0505047102
that stimulates EC mitogenesis. KDR and Flt-1 are localized on the cell surface of various EC types (4). Increased expression of these receptors occurs in response to several stimuli and results in priming of EC toward proliferation, migration, and angiogenesis (5). Oxygen tension plays a major role in the regulation of VEGF. Its mRNA expression is rapidly and reversibly induced by hypoxia in a variety of normal and transformed cultured cell types (6). Several VEGF structures have been reported (7–11) included the VEGF bound to the extracellular domain 2 of Flt-1 (12). Two VEGF monomers, linked by disulfide bonds, bind to two receptor molecules that are localized at the poles of the VEGF antiparallel homodimer. The analysis of structural and mutagenesis data allowed to identify the residues involved in the binding to the receptors. They are distributed over a discontinuous surface that include residues from the N-terminal helix 17–25. KDR and Flt-1 share the VEGF binding region, in fact five of the seven most important VEGF binding residues are present in both interfaces (8, 12, 13). Many approaches have been pursued to modulate the VEGF– receptors interaction, and new molecular entities as peptides (14–20) and antibodies (21–23) have been reported to bind to the extracellular region of the VEGF receptors. A large number of them show an antagonist activity and only few behave as agonists (15). Remarkably, the peptides modulating the VEGF–receptors interaction are mainly derived by phage display libraries screening, and only few examples of rational design approaches have been reported so far (19). The aim of this work was to develop small peptides able to recognize VEGF receptors to modulate both endothelial cell proliferation and propensity toward angiogenesis. We designed, by a structure-based approach, a peptide, QK, reproducing the VEGF 17–25 helix region. In this paper, we report the design, the conformational analysis, and the biological properties of QK revealing its ability to assume a helical conformation in pure water and to modulate the angiogenic response mediated by VEGF. Materials and Methods Peptide Synthesis. Peptides were synthesized on solid phase by
using Rink Amide MBHA resin (Novabiochem) by standard Fmoc (N-(9-Fluorenyl)methoxycarbonyl) chemistry. The side chain of the N-terminal lysine was protected with the methyltrytil group to allow selective deprotection and peptide labeling. This paper was submitted directly (Track II) to the PNAS office. Abbreviations: BAEC, bovine aorta endothelial cells; EC, endothelial cell; ERK, extracellular signal-regulated kinase; Flt-1, Fms-like tyrosine kinase; Flt-1D2, Flt-1 domain 2; KDR, kinase domain receptor. §To
whom correspondence should be addressed. E-mail:
[email protected].
© 2005 by The National Academy of Sciences of the USA
PNAS 兩 October 4, 2005 兩 vol. 102 兩 no. 40 兩 14215–14220
BIOCHEMISTRY
Edited by Charles A. Dinarello, University of Colorado Health Sciences Center, Denver, CO, and approved August 8, 2005 (received for review June 16, 2005)
Cleavage from the resin were achieved by treatment with trifluoracetic acid, triisopropyl silane, and water (95:2.5:2.5) at room temperature for 3 h. Purity and identity of the peptides were assessed by HPLC and MALDI-TOF mass spectrometry. Circular Dichroism Spectroscopy. CD spectra were collected on a
Jasco 715 instrument by using a 1 mm path-length quartz cuvette (Hellma) at 20°C in 10 mM phosphate buffer, pH 7.1. Peptide concentrations were 21.8 M (QK) and 43.4 M (VEGF15). The aggregation state of QK was checked, over a concentration range of 1 M to 1.5 mM, by UV spectroscopy (absorbance at 280 nm) and NMR (line widths and chemical shift variations). All of the experimental data (data not shown) indicate the QK peptide does not aggregate up to 1.5 mM. NMR Spectroscopy. The NMR samples were prepared dissolving the QK peptide at a concentration of 1.0 ⫻ 10⫺3 M either in a H2O兾2H2O 90:10 mixture or in pure 2H2O at pH 5.5. The NMR experiments were recorded on a Varian Inova 600 spectrometer at a temperature of 298 K. All of the spectra were processed with the software PROSA (24) and analyzed with the program XEASY (25). Structure Calculation. Experimental distance restraints for struc-
ture calculations were derived from the cross-peak intensities in NOESY spectra recorded in H2O and 2H2O. Structure calculations, which used the torsion angle dynamics protocol of CYANA (46), were started from 100 randomized conformers. The 20 conformers with the lowest CYANA target function were further refined by means of restrained energy minimization with the GROMOS 96 (26) force field with the program SPDB VIEWER (27). The color figures and the structure analysis have been performed with the program MOLMOL (28). Cell Culture. ECs from bovine aorta, immortalized with SV40, were cultured in DMEM (Sigma) and supplemented with 10% FBS (Invitrogen) at 37°C in 95% air兾5% CO2. In all of the experiments, VEGF165 (Alexis) was used at 100 ng兾ml. VEGF Receptors Binding Assay. Cells were homogenized in lysis buffer (12.5 mM Tris, pH 6.8兾5 mM EDTA兾5 mM EGTA), and membranes were separated from the cytosol fraction by centrifugation. Membranes were suspended in binding buffer (75 mM Tris兾12.5 mM MgCl2兾2 mM EDTA), and an equal amount of membrane protein (1 g) was plated in 96-well plates with QK (10⫺13 to 10⫺8 M) and [125I]-VEGF (Amersham Biosciences). VEGF binding was evaluated with a ␥-counter. Western Blot. Cells were plated on six-well dishes and serum starved overnight. On the next day, cells were treated with a different amount of peptide in the absence or in presence of VEGF165 for 15 min at 37°C and then dissolved in radioimmunoprecipitation assay-SDS buffer (50 mM Tris䡠HCl, pH 7.5兾150 mM NaCl兾1% Nonidet P-40兾0.25% deoxycholate兾9.4 mg/50 ml sodium orthovanadate兾20% SDS). In some experiments, total KDR and Flt-1 were immunoprecipitated from an equal amount of whole-cell protein extracts by using protein A兾G agarose beads conjugated with antibodies raised against total KDR or Flt-1 (R & D Systems). Proteins from whole-cell extracts or immunocomplexes were resolved by PAGE and transferred to nitrocellulose. Total extracellular signal-regulated kinase 1 and 2 (ERK1兾2), serine-tyrosin phosphorylated ERK1兾2, phosphotyrosine (Cell Signaling Technology) and phospho-retinoblastoma protein (p-RB) (Santa Cruz Biotechnology) were visualized by specific antibodies, anti-rabbit horseradish peroxidaseconjugated secondary antibody (Santa Cruz Biotechnology) and standard chemiluminescence (Pierce). 14216 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0505047102
[3H]Thymidine Incorporation. Cells were serum starved for 24 h and
then incubated in DMEM with [3H]thymidine (Amersham Pharmacia) and QK alone (10⫺12 to 10⫺6 M) or with a combination of QK and VEGF165. After 24 h, cells were fixed with trichloracetic acid (0.05%) and dissolved in 1M NaOH. Scintillation liquid was added and thymidine incorporation was evaluated with a beta counter. Cells Proliferation Assay. Cells were seeded at a density of 10,000
per well in six-well plates, serum starved overnight, and then stimulated with QK (10⫺12 to 10⫺6 M) in the absence or presence of VEGF165. Cell number was determined at 24 h after stimulation. The phospho-retinoblastoma cyclin was evaluated by Western blot 12 and 18 h after stimulation with QK (10⫺6 M), VEGF165, and VEGF15 (10⫺6 M). Angiogenesis in Vitro Assay. Human endothelial cells were cocul-
tured with other human cells in a specially designed medium (Angiokit, TCS CellWorks, Buckingham, U.K.), in 24-well plates. Every 3 days, QK in the absence or presence of VEGF165 was added to the cultures. VEGF and suramine (20 M) were used as positive and negative controls, respectively. Cells subsequently begin to proliferate and then enter a migratory phase, during which they move through the matrix to form thread-like tubule structures. On the 11th day, cells were fixed with ice cold 70% ethanol, and tubule formation was visualized by staining for anti-human CD31 (PECAM-1). Results were scored with the image analysis software, ANGIOSYS (TCS CellWorks). Results Peptide Design. Based on the x-ray structure of the VEGF兾Flt-1
domain 2 (Flt-1D2) complex (12), we designed and synthesized a peptide reproducing the VEGF binding region spanning the amino acid sequence Phe-17-Tyr-25. This region contains 5 (Phe-17, Met-18, Tyr-21, Gln-22, and Tyr-25) of 21 residues situated at ⬍4.5 Å from the receptor and it assumes, in the natural protein, an ␣-helix conformation. The design strategy we adopted was to keep fixed the three-dimensional arrangement of the residues interacting with the receptor and stabilize the secondary structural motif. Mutagenesis data indicate that when Phe-17 is mutated to Ala, the affinity toward KDR is reduced by 90-fold, whereas mutations of the other four residues only slightly affect the binding (8, 13). All of the five interacting residues occupy a face of the helix, and they make hydrophobic interaction with the receptor. Residues on the opposite face protrude toward the protein interior and, in an isolate peptide, they would be solvent exposed. The helix conformation of the QK peptide was stabilized introducing N- and C-capping sequences (29), amino acid with intrinsic helix propensity, and favorable electrostatic interactions (30). The N- and C-capping residues (L15兾T16 and K26兾G27兾I28, respectively) were chosen based on statistical preference for each capping position (29). Phe-17 was replaced by Trp to introduce a spectroscopic probe and to increase the hydrophobic interactions; Met-18, which is close to the residue Asn-219 of Flt-1, was substituted with a Gln residue, present in the VEGF homolog protein, Placenta Growth Factor, more suited to form favorable hydrogen bond interaction. Asp-19 was replaced by Glu because of its higher helix propensity, and Ser-24 was substituted with Lys to increase helix propensity and solubility. An extra Lys residue was appended at the N-terminal to allow selective labeling. The peptide was acetylated and amidate to avoid electrostatic repulsion between peptide terminal charges and helix dipoles. The sequences of the designed peptide, QK, and the peptide corresponding to the ␣-helix region of VEGF, VEGF15, are reported in Fig. 1a. Conformational Analysis in Solution. CD spectra of QK peptide
(Fig. 1b) were recorded at pH 5.5 (data not shown) and 7.1. The D’Andrea et al.
peptide showed a high helical content, which did not vary in the experimented pH range. Notably, VEGF15 is unstructured in the same experimental condition. A solution conformational analysis of the peptide QK was then undertaken, through the acquisition of a complete set of homonuclear NMR spectra in pure water at pH 5.5. By a careful inspection of total correlated spectroscopy, NOESY, and double-quantum filtered (DQF)-COSY spectra, a virtually complete proton assignment of the QK peptide was obtained (Table 1, which is published as supporting information on the PNAS web site) by following the standard procedures (31). The H␣ chemical shifts analysis, performed by using the Chemical Shift Index (32) (Fig. 7, which is published as supporting information on the PNAS web site), strongly indicated the presence of a helical structure that includes the core region (residue 17–26) of the QK amino acid sequence. The NOESY assignments of the QK peptide showed an extensive HN–HN(i, i⫹1), H␣–HN(i, i⫹3), and H␣-H(i, i⫹3) net of cross peaks along the sequence of the peptide (Fig. 8, which is published as supporting information on the PNAS web site), which confirmed the high helical propensity of the QK peptide. Totally, 184 NOE cross peaks were assigned and integrated; stereospecific assignments for CH2 protons of Glu-19, Tyr-21, and Leu-23 were derived from the input data by using the software CYANA. Moreover, 13 3JHNH␣ coupling constants were extracted from the DQF-COSY spectrum (Table 2, which is published as supporting information on the PNAS web site); temperature coefficients of the amide protons of QK peptide were also measured and indicated that the backbone amide protons of residues 19, 23, and 24 could be well involved in hydrogen bonding (data not shown). The final input file for the CYANA structure calculation software contained 104 meaningful distance constraints (40 intraresidue, 42 short-range, and 22 medium-range) and 57 angle constraints that were derived from intraresidue and sequential NOEs and the 3JHNH␣ coupling constants. The backbone superposition of the best 20 CYANA conformers with the lowest target function values (0.19 ⫾ 0.002 Å2) is reported in Fig. 2a. The NMR solution structure of QK peptide is quite well defined (the rms deviation values of the backbone and of the all heavy atoms of the 17–25 region are 0.163 Å and 0.988 Å, respectively) and mainly consists of a helix, whose core is the region 17–25, with less-ordered N and C termini. These data are in very much agreement with the H␣ chemical shift index analysis that already showed the presence of a helical conformation in the central region of QK peptide. Taken together, H␣ chemical shift and NMR structure determination, which is mainly based on D’Andrea et al.
Fig. 2. NMR structure of QK. (a) Superposition of the backbone of the best 20 CYANA QK structures. (b) Backbone superposition of the QK representative structure (yellow) and VEGF helix (red) bound to Flt-1D2. Side chain of the interacting residues and the Flt-1D2 electrostatic surface are shown.
NOE cross-peaks, appear to describe very accurately the aqueous solution conformation of QK peptide. The helical region of QK exactly corresponds to the Nterminal helix (residues 17–25) of VEGF in complex with Flt-1D2, resulting in a very good structural similarity between the designed peptide and its cognate natural sequence (Fig. 2b). VEGF Receptors Binding Assay. To verify the biological behavior of
QK peptide, we tested its ability to compete for the binding sites of VEGF on cell membranes (Fig. 3a). We competed membranes, obtained from isolated bovine aorta endothelial cells (BAEC), with iodinated VEGF and then with increasing amounts of QK. Competition curves showed a displacement of iodinated VEGF by QK with an estimated apparent dissociation constant of ⬇10⫺9 M, thus suggesting the interaction with receptors localized on particulate cellular fraction. To show that indeed VEGF receptors are involved in the binding to QK and to evaluate the ability of QK to initiate early events of signal transduction, we immunoprecipitated total KDR and Flt-1 from BAEC whole extracts and visualized tyrosin phosphorylation by Western blot. As expected, 15 min of exposure to VEGF165, used as control, caused the reduction in the levels of phospho-KDR at the membrane and increased Flt-1 phosphorylation (Fig. 3 b and c) (33). QK exerted similar effects on these receptors, because it reduced phospho-KDR below the levels in unstimulated cells and increased the levels of phosphorylation of Flt-1. Together with ligand binding data, these results suggest that QK recapitulated the effects of VEGF165 on VEGF receptors. Activation of the Proliferative Intracellular Pathways. We explored whether QK is able to start the pathways of EC activation. It is PNAS 兩 October 4, 2005 兩 vol. 102 兩 no. 40 兩 14217
BIOCHEMISTRY
Fig. 1. Peptide sequences and circular dichroism spectroscopy. (a) Amino acid sequences of the peptides VEGF15 and QK. The numbering is referred to the VEGF sequence. (b) CD spectra of QK (solid line) and VEGF15 (dashed line) in 10 mM phosphate (pH 7.1) at 20°C. CD spectra are converted and displayed in molar residue ellipticity [].
Fig. 3. VEGF receptors binding and activation. (a) VEGF competitive binding on BAEC. One microgram of membrane protein was plated with QK and [125I]-VEGF (500,000 cpm, 10⫺10 M). (b) KDR activation. After stimulation, total KDR was immunoprecipitated from a whole-cell protein extracts and phospho-tyrosine was visualized by a specific antibody, anti-rabbit horseradish peroxidase-conjugated secondary antibody and standard chemiluminescence. (c) Flt-1 activation. After stimulation, total Flt-1 was immunoprecipitated from a whole-cell protein extract, and phospho-tyrosine was visualized by a specific antibody, anti-rabbit horseradish peroxidase-conjugated secondary antibody, and standard chemiluminescence.
well established that angiogenesis modulated by VEGF is largely ERK1兾2 dependent, leading to DNA synthesis and cell proliferation (34). QK led to ERK1兾2 activation in a dose-dependent fashion, and this response was additive to VEGF, indicating that low doses of QK facilitate VEGF signaling (Fig. 4a). Instead, the peptide reproducing the natural helix (VEGF15) had no effect on ERK activation, proving that it is unable to start intracellular signaling (Fig. 4b). To verify whether ERK1兾2 activation to QK results in cell proliferation, we studied cell proliferation indicators such as cell number, DNA synthesis, and cyclin activation. QK increased DNA synthesis at any dosage, and the effect was enhanced in the presence of VEGF (Fig. 5a). Cell proliferation studies likewise indicated that QK produces cell proliferation per se and enhances VEGF response (Fig. 5b). Finally, QK and VEGF165, but not VEGF15, enhanced phosphorylation of the protein RB, thus indicating cell cycle progression from G1 to S (Fig. 5c). In Vitro Angiogenesis Assay. To investigate whether QK recapitulates the overall angiogenic properties of VEGF, we studied the ability of the peptide to induce EC network formation on a matrigel substrate (Fig. 6). Tubule formation was evaluated by positive staining for CD31兾PECAM-1, an intercellular adhesion molecule involved in leucocytes diapedesis. We determined the number of cell junctions corrected by the total tubules length. As a positive control, we used VEGF, which caused an increase in the number of connections that each EC extend to the neighborhood cell (from 0.1 ⫾ 0.1 to 2.14 ⫾ 0.17). QK induced the formation of new connections in a dose-dependent manner and enhanced the response to VEGF165 (Fig. 6e).
Discussion Modulating angiogenesis in the adult life is a very attractive goal because it is involved in relevant pathological conditions. Therapeutic angiogenesis is sought as the ultimate intervention to solve chronic ischemia in those conditions that cannot be treated
alternatively. Its converse, the anti-angiogenic treatment, is a promising therapy in oncology. Because the angiogenic response strictly depends on VEGF activity, this protein is considered a very attractive pharmacological target and, in the last year, it has been the object of intense investigations. The x-ray structure of the complex VEGF兾Flt-1D2 shows that the binding interface is mainly localized in three regions (12). One of them is the ␣-helix spanning the amino acid sequence 17–25. We focused on this region because it comprises some of the key residues involved in receptors recognition and because new molecules interacting with the receptors reproducing this region have not been developed so far. Moreover, the design of helical peptide represents a tractable target for peptide engineering because the folding and stability rules of helical peptides have been elucidated in the last decade (30). It is well known that peptide fragments spanning the helices, turns, and -hairpins of natural proteins show little propensity, with very few exceptions, to reproduce their natural secondary structure under physiological conditions (30). Nevertheless, the stabilization of suitable conformation in aqueous solutions is a condition to gain the binding of designed peptides to their targets. We reasoned that if appropriate tools were introducing in the natural sequence, such as to stabilize the helical conformation, then the key residues will be displayed in the threedimensional arrangements suitable for the receptor binding. We designed by a structure-based approach a linear peptide, QK, which should interact with the VEGF receptors. All of the data collected on the structural preferences of the QK peptide in aqueous solution strongly indicated that it mainly folds in helical conformation. In particular, the first indication derives from the CD spectrum, which is well confirmed by the H␣ chemical shift analysis and the NMR structure determination (Fig. 2a). These two latter analyses defined the QK helical region because that included residues between 17 and 25, which correspond to the VEGF helical region (Fig. 2b) and represent an important prerequisite for the QK biological activity. The stabilization of
Fig. 4. Effect of QK and VEGF15 on ERK1兾2 activation. Serum-deprived BAEC were treated with QK (a) or with VEGF15 (b) in the absence or presence of VEGF165 (100 ng兾ml) for 15 min at 37°C and then dissolved in radioimmunoprecipitation assay-SDS buffer. Total ERK1兾2 and the phosphorylated form of ERK1兾2 were visualized by specific antibodies. 14218 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0505047102
D’Andrea et al.
QK helical conformation is not a trivial result, because VEGF15 assumes in solution a random coil conformation and, because typically, short peptides, composed of natural amino acids, are rarely helical in solution, mainly because of inherent thermodynamic instability. The basis of the QK helical fold seems to reside on the presence of amino acids with intrinsic helix preference
Fig. 6. In vitro angiogenic properties of QK. Human endothelial cells were cocultured with other human cells in a specially designed medium in a 24-well plate. Every 3 days, QK alone (c) or a combination of QK and VEGF165 (100 ng兾ml) (d) was added. On the 11th day, cells were fixed with ice cold 70% ethanol, and tubule formation was visualized by staining for anti-human CD31 (PECAM-1). (a–d) Sample images are reported. Suramine (20 M) (a) and VEGF165 (b) were used as negative and positive controls, respectively. (e) The number of cellular connections and the total tubule length were determined by using software that analyzes the images after digitalization.
D’Andrea et al.
and on the amphipathic nature of the helix, which allows a number of medium range ionic, polar, and hydrophobic interactions on opposite faces of the peptide. Moreover, QK peptide, which is composed by only 15 natural amino acids and whose structure in pure water has been derived with a good backbone resolution, could represent a model for further folding studies. Most of the biological function of VEGF are mediated by its receptors KDR and Flt-1 (2, 35, 36). VEGF interaction with KDR or Flt-1 induces receptor dimerization and activation. Therefore, VEGF dimers are considered the only active form. In this paper, we report evidence that the peptide QK binds to VEGF receptors. Binding studies showed that QK competes with VEGF for a binding site on EC membranes. These cells express both KDR and Flt-1 receptors, two tyrosine receptors that undergo autophosphorylative events upon binding to their agonist. To evaluate whether QK has any preference toward one of the two receptors, we immunoprecipitated the receptors and evaluated the tyrosin phosphorylation by Western blot. Data reported in Fig. 3 showed that QK binds and activates both receptors similarly to VEGF. KDR phosphorylation decreases after ligand binding because KDR is internalized and digested, whereas Flt-1 remains exposed on the membrane (33). The agonist-like behavior of QK is confirmed by cell proliferation experiments and by the downstream activation of VEGFdependent intracellular pathway (ERK1兾2). It has been reported that VEGF stimulates DNA synthesis and proliferation in a variety of EC types (37–40). VEGF strongly induces the activity of ERK1兾2, and the activation of this pathway presumably plays a central role in the stimulation of EC proliferation (6, 41). Our data showed that QK led to ERK1兾2 activation (Fig. 4a) and cell proliferation (Fig. 5 a and b) in a dose-dependent fashion and, in both experiments, QK enhanced the VEGF activity. Moreover, we checked, as a marker of cell proliferation, the phosphorylation of RB, the protein that regulates proliferation by controlling progression through the restriction point within the G1 phase of the cell cycle. QK and VEGF165, but not VEGF15, enhanced RB phosphorylation, thus indicating cell cycle progression from G1 to S (Fig. 5c). Finally, we tested the biological properties of our peptide in a functional assay performing an in vitro angiogenesis assay by using a matrigel substrate. VEGF is a potent angiogenic factor in vivo, which induces cell proliferation and migration through extracellular matrix to form thread-like tubule structures that join up to create a network of tubules (42, 43). QK, as shown in Fig. 6, induced the formation of newly formed connections in a dose-dependent manner and enhanced the VEGF response. This experiment confirmed that our peptide QK recapitulates many of the features in signal transduction that are reported for VEGF. PNAS 兩 October 4, 2005 兩 vol. 102 兩 no. 40 兩 14219
BIOCHEMISTRY
Fig. 5. Effect of QK on cell proliferation. (a) DNA synthesis. BAEC were incubated in DMEM with [3H]thymidine and QK in the absence or presence of VEGF165 (100 ng兾ml). After 24 h, cells were fixed and lysed. Scintillation liquid was added, and [3H]thymidine incorporation was evaluated. (b) Cell proliferation. BAEC were stimulated with the indicated amount of QK in the absence or presence of VEGF165 (100 ng兾ml). Cell number was determined at 24 h after stimulation. (c) RB phosphorylation. phospho-retinoblastoma protein (p-RB) was evaluated at 12 and 18 h after stimulation with QK (10⫺6 M), VEGF165 (100 ng兾ml), and VEGF15 (10⫺6 M).
Overall, our results demonstrate that QK binds to VEGF receptors in vitro and show that it is a potential agonist for angiogenesis. A stable helical structure of the core region of the QK region appears to be a key requisite for its ability to bind the VEGF receptor. In fact, the natural fragment, VEGF15, which as expected is unstructured in water, did not show any appreciable biological activity alone or in combination with VEGF. Although QK was designed to bind to KDR and Flt-1, its agonist-like activity is surprising because receptor dimerization is necessary for receptor activation. QK is a small peptide and probably is unable to induce receptor dimerization by itself. To explain the activity of QK and its additive effects in presence of VEGF, we can speculate that the QK binding possibly induces a conformational change of the receptor that either triggers, in an unknown way, the dimerization or induces the receptor activation through a different兾previously uncharacterized mechanism that may not require dimerization. Further structural and biochemical studies are needed to address this issue. Therapeutic angiogenesis in cardiovascular conditions such as chronic ischemia or heart failure is sought as a promise of modern biotechnology. Indeed, the hypothesis that VEGF administration may result in therapeutically significant angiogenesis in humans has been already tested in a gene therapy trial in patients with severe limb ischemia (44). Major limitations to the use of growth factors such as VEGF are associated to their ability
to promote uncontrolled neoangiogenesis and lymphatic edema. Recent findings, furthermore, propose VEGF as a factor promoting asthma (45), a side effect that could preclude the use of this molecule in a large share of the ischemic population. Our data, albeit not produced in vivo settings, are suggestive that either QK or improved analogues might fulfill the request for a safer pro-angiogenic drug. To this aim, it will be intriguing to see whether QK preserves these biological effects also in vivo, in models of pathological angiogenesis, such as chronic limb ischemia. In conclusion, based on x-ray structure of the VEGF兾receptor complex, we designed a small peptide that mimics the VEGF helix region 17–25. It adopts in pure water a helical conformation and binds to VEGF receptors. We showed that the structural preorganization is needed for its biological function. QK peptide acts as an agonist and induces angiogenesis in vitro. Interestingly, it potentiates the EC response to VEGF. Potential applications for this peptide are in the diagnostic field and in therapy of cardiovascular diseases.
Carmeliet, P. (2003) Nat. Med. 9, 653–660. Ferrara, N. & Davis-Smyth, T. (1997) Endocr. Rev. 18, 4–25. Ferrara, N. (2000) Curr. Opin. Biotechnol. 11, 617–624. Ferrara, N., Gerber, H. P. & LeCouter, J. (2003) Nat. Med. 9, 669–676. Brogi, E., Schatteman, G., Wu, T., Kim, E. A., Varticovski, L., Keyt, B. & Isner, J. M. (1996) J. Clin. Invest. 97, 469–476. Abedi, H. & Zachary, I. (1997) J. Biol. Chem. 272, 15442–15451. Muller, Y. A., Christinger, H. W., Keyt, B. A. & de Vos, A. M. (1997) Structure (London) 5, 1325–1338. Muller, Y. A., Li, B., Christinger, H. W., Wells, J. A., Cunningham, B. C. & de Vos, A. M. (1997) Proc. Natl. Acad. Sci. USA 94, 7192–7197. Muller, Y. A., Chen, Y., Christinger, H. W., Li, B., Cunningham, B. C., Lowman, H. B. & de Vos, A. M. (1998) Structure (London) 6, 1153–1167. Wiesmann, C., Christinger, H. W., Cochran, A. G., Cunningham, B. C., Fairbrother, W. J., Keenan, C. J., Meng, G. & de Vos, A. M. (1998) Biochemistry 37, 17765–17772. Pan, B., Li, B., Russell, S. J., Tom, J. Y., Cochran, A. G. & Fairbrother, W. J. (2002) J. Mol. Biol. 316, 769–787. Wiesmann, C., Fuh, G., Christinger, H. W., Eigenbrot, C., Wells, J. A. & de Vos, A. M. (1997) Cell 91, 695–704. Keyt, B. A., Nguyen, H. V., Berleau, L. T., Duarte, C. M., Park, J., Chen, H. & Ferrara, N. (1996) J. Biol. Chem. 271, 5638–5646. An, P., Lei, H., Zhang, J., Song, S., He, L., Jin, G., Liu, X., Wu, J., Meng, L., Liu, M. & Shou, C. (2004) Int. J. Cancer 111, 165–173. Scheidegger, P., Weiglhofer, W., Suarez, S., Console, S., Waltenberger, J., Pepper, M. S., Jaussi, R. & Ballmer-Hofer, K. (2001) Biochem. J. 353, 569–578. Jia, H., Jezequel, S., Lohr, M., Shaikh, S., Davis, D., Soker, S., Selwood, D. & Zachary, I. (2001) Biochem. Biophys. Res. Commun. 283, 164–173. Binetruy-Tournaire, R., Demangel, C., Malavaud, B., Vassy, R., Rouyre, S., Kraemer, M., Plouet, J., Derbin, C., Perret, G. & Mazie, J. C. (2000) EMBO J. 19, 1525–1533. Hetian, L., Ping, A., Shumei, S., Xiaoying, L., Luowen, H., Jian, W., Lin, M., Meisheng, L., Junshan, Y. & Chengchao, S. (2002) J. Biol. Chem. 277, 43137–43142. Zilberberg, L., Shinkaruk, S., Lequin, O., Rousseau, B., Hagedorn, M., Costa, F., Caronzolo, D., Balke, M., Canron, X., Convert, O., et al. (2003) J. Biol. Chem. 278, 35564–35573. El-Mousawi, M., Tchistiakova, L., Yurchenko, L., Pietrzynski, G., Moreno, M., Stanimirovic, D., Ahmad, D. & Alakhov, V. (2003) J. Biol. Chem. 278, 46681–46691. Prewett, M., Huber, J., Li, Y., Santiago, A., O’Connor, W., King, K., Overholser, J., Hooper, A., Pytowski, B., Witte, L., et al. (1999) Cancer Res. 59, 5209–5218.
22. Cooke, S. P., Boxer, G. M., Lawrence, L., Pedley, R. B., Spencer, D. I., Begent, R. H. & Chester, K. A. (2001) Cancer Res. 61, 3653–3659. 23. Lu, D., Shen, J., Vil, M. D., Zhang, H., Jimenez, X., Bohlen, P., Witte, L. & Zhu, Z. (2003) J. Biol. Chem. 278, 43496–43507. 24. Gu ¨ntert, P., Dotsch, V., Wider, G. & Wu ¨thrich, K. (1992) J. Biomol. NMR 2, 619–629. 25. Bartels, C., Xia, T., Billeter, M., Guntert, P. & Wu ¨thrich, K. (1995) J. Biomol. NMR 6, 1–10. 26. Van Gunsteren, W. F., Billeter, S. R., Eising, A. A., Hunenberger, P. H., Kruger, P., Mark, A. E., Scott, W. R. P. & Tironi, I. G. (1996) Biomolecular Simulation: The GROMOS96 Manual and User Guide (BIOMOs, Zurich). 27. Guex, N. & Peitsch, M. C. (1997) Electrophoresis 18, 2714–2723. 28. Koradi, R., Billeter, M. & Wuthrich, K. (1996) J. Mol. Graphics 14, 29–32. 29. Aurora, R. & Rose, G. D. (1998) Protein Sci. 7, 21–38. 30. DeGrado, W. F., Summa, C. M., Pavone, V., Nastri, F. & Lombardi, A. (1999) Annu. Rev. Biochem. 68, 779–819. 31. Wu ¨thrich, K. (1986) NMR of Proteins and Nucleic Acids (Wiley, New York). 32. Wishart, D. S., Sykes, B. D. & Richards, F. M. (1992) Biochemistry 31, 1647–1651. 33. Duval, M., Bedard-Goulet, S., Delisle, C. & Gratton, J. P. (2003) J. Biol. Chem. 278, 20091–20097. 34. Berra, E., Milanini, J., Richard, D. E., Le Gall, M., Vinals, F., Gothie, E., Roux, D., Pages, G. & Pouyssegur, J. (2000) Biochem. Pharmacol. 60, 1171–1178. 35. Shibuya, M. (2001) Cell Struct. Funct. 26, 25–35. 36. Zachary, I. & Gliki, G. (2001) Cardiovasc. Res. 49, 568–581. 37. Thakker, G. D., Hajjar, D. P., Muller, W. A. & Rosengart, T. K. (1999) J. Biol. Chem. 274, 10002–10007. 38. Pedram, A., Razandi, M. & Levin, E. R. (1998) J. Biol. Chem. 273, 26722–26728. 39. Xia, P., Aiello, L. P., Ishii, H., Jiang, Z. Y., Park, D. J., Robinson, G. S., Takagi, H., Newsome, W. P., Jirousek, M. R. & King, G. L. (1996) J. Clin. Invest. 98, 2018–2026. 40. Doanes, A. M., Hegland, D. D., Sethi, R., Kovesdi, I., Bruder, J. T. & Finkel, T. (1999) Biochem. Biophys. Res. Commun. 255, 545–548. 41. Wheeler-Jones, C., Abu-Ghazaleh, R., Cospedal, R., Houliston, R. A., Martin, J. & Zachary, I. (1997) FEBS Lett. 420, 28–32. 42. Conway, E. M., Collen, D. & Carmeliet, P. (2001) Cardiovasc. Res. 49, 507–521. 43. Ferrara, N. (2001) Am. J. Physiol. 280, C1358–C1366. 44. Isner, J. M. & Asahara, T. (1999) J. Clin. Invest. 103, 1231–1236. 45. Lee, C. G., Link, H., Baluk, P., Homer, R. J., Chapoval, S., Bhandari, V., Kang, M. J., Cohn, L., Kim, Y. K., McDonald, D. M., et al. (2004) Nat. Med. 10, 1095–1103. 46. Herrmann, T., Guntert, P. & Wuthrich, K. (2002) J. Mol. Biol. 319, 209–227.
1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
11. 12. 13. 14. 15. 16. 17.
18.
19.
20.
21.
14220 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0505047102
We thank Dr. Giuseppe Perretta and Leopoldo Zona for technical assistance and Dr. Alessandra Romanelli for careful reading of the manuscript. This work was supported by the Centro Regionale di Competenza in Diagnostica e Farmaceutica Molecolari della Regiene Campania.
D’Andrea et al.
