| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,2
* Molecular Cell Biology and
Structural Biology, Laboratory of Biomolecular Research, Paul Scherrer Institut, Villigen, Switzerland;
Department of Genetics and Pathology, Rudbeck Laboratory, and
Department of Medical Biochemistry and Microbiology (IMBIM), The Biomedical Center, Uppsala University, Uppsala, Sweden; and
|| Virus Research Unit, Department of Microbiology and Immunology, University of Otago, New Zealand
3Correspondence: Paul Scherrer Institut, Laboratory of Biomolecular Research, Molecular Cell Biology, Bldg. OFLC 102, 5232 Villigen-PSI Switzerland. E-mail: kurt.ballmer{at}psi.ch
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: neuropilin • angiogenesis • heparin sulfate • pox virus • vascular endothelial growth factor
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. VEGF family proteins are expressed in multiple isoforms generated by proteolytic processing and alternative splicing (3
, 4)
. In addition, orf virus and other parapoxviruses encode a set of VEGF homologues called VEGF-E, which share 25–35% amino acid identity with VEGF-A and a high degree of structural identity (5
6
7
8
9
10
11
12
13)
. Orf virus VEGF-Es are potent mitogens stimulating proliferation of human endothelial cells in vitro and vascularization of sheep skin in vivo with potencies equivalent to VEGF-A (9)
.
The biological functions of VEGF polypeptides result from binding to several cellular receptors: type V receptor tyrosine kinases (RTKs), VEGFR-1 (Flt-1), VEGFR-2 (KDR/Flk-1), and VEGFR-3 (Flt-4) (14
15
16)
; neuropilin-1 and -2 (NRP-1, -2) (17)
; and heparan sulfate (HS) proteoglycans (18)
. Some VEGFs interact with multiple receptors while others show very distinct receptor binding properties. So, for instance PlGF and VEGF-B are specific for VEGFR-1 (19
, 20)
, and VEGF-Es and some snake venom VEGF homologues (VEGF-Fs) bind VEGFR-2 (7
, 9
10
11
, 21
, 22)
. VEGF-A121 encoded by exons 2–5 and 8 binds VEGFR-1 and -2, the longer VEGF-A variants that also contain exons 6 and/or 7, such as VEGF-A165, VEGF-A183, VEGF-A189, and VEGF-A203, also interact with NRP-1 and HS (23
24
25
26)
. VEGF-A165 presumably also binds NRP-1 and VEGFR-2 when the receptors are expressed separately on adjacent cells promoting endothelial cell migration and cell guidance, for instance when vessels form along tracks defined by neural cells (27
, 28)
or during endothelial tip cell guidance (29
, 30)
. More recently partially defective VEGF-A splice variants carrying a different carboxyterminus encoded by exon 8b instead of exon 8a have been described (3
, 31
, 32)
. These variants do not bind NRP-1 and have reduced affinity for HS (33)
. In addition, the carboxyterminal peptide of VEGF-A, DKPRR, and a related peptide, TKPPR derived from the proinflammatory peptide tuftsin (TKPR), have been shown to bind NRP-1 directly, confirming that the exon 8a domain is required for NRP-1 binding (34
, 35)
. Finally, viral and snake venom VEGF-E and -F homologues differ in their potential to form complexes with HS and NRP-1, rendering these molecules ideal tools to study the distinct roles of HS and NRP-1 in VEGF signaling (5
, 6
, 12
, 21
, 22)
.
The present study aimed to identify and characterize the binding sites of VEGF-A and -E for NRP-1 and to deduce the contribution of NRP-1 and HS in modulation of VEGF signaling by comparing the properties of VEGF-A with those of orf virus VEGF-Es. A short carboxyterminal peptide, RPPR, was identified as the critical determinant for VEGF-E binding to NRP-1. High-affinity NRP-1 binding VEGFs promoted the formation of higher-order complexes between VEGFR-2 and NRP-1 in cells expressing both receptors and in cultures where these receptors were expressed on separate cells. The functionality of VEGF variants was determined in a mouse model of differentiating embryonic stem (ES) cells, embryoid bodies (EBs) (36)
.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 38)
. The specific activities were 
2.5 x 104 cpm/ng polysaccharide, 3.0 x 104 cpm/ng, 2.0 x 104 cpm/ng, and 2.7 x 104 cpm/ng, respectively, for the four preparations. VEGFR-2 antibody 11939 was from Abcam (Cambridge, UK), mouse VEGFR-2 antibody A3 sc6251 and goat NRP-1 antibody sc19 were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Rabbit phosphopeptide-specific and phospholipase C
-1 antibodies were from Cell Signaling Technology (Danvers, MA USA).
The expression vectors for producing VEGFs in Pichia pastoris were generated using the pPICZ
A vector system from Invitrogen (Carlsbad, CA, USA) as previously described (39)
. All expression vectors were made with the PCR subcloning technology, and all VEGF proteins except VEGF-A165 carried a hexahistidine tag at the amino terminus as described before (40)
. Pichia pastoris strain X33 was transfected by electroporation, and Zeocin-resistant clones were picked and tested for transgene expression on methanol induction. Secreted protein was purified from the yeast culture supernatant by immobilized metal affinity chromatography (IMAC) and polished on Superdex 200 (GE Healthcare, Basel, Switzerland). The correct identity of the proteins was confirmed by SDS-PAGE, immunoblotting, and Edman sequencing. Peptides were synthesized by the Protein and Peptide Chemistry Facility of the University of Lausanne (UNIL; Lausanne, Switzerland).
Cloning and expression of NRP-1 proteins
The sequences encoding a truncated NRP-1 ectodomain (amino acids 1–641; NRP-1–641) and the NRP-1 a1a2b1b2 domains (amino acids 1–587; NRP-1–587) were amplified by PCR. For NRP-1–641 the following primers were used: forward primer, 5'-GGGGACAAGTTTGTACAAAAAAGCAGGCTTCGAAGGAGATAGAACCATGGA GAGGGCTGC-3'; reverse primer: 5'-GGGGACCACTTTGTACAAGAAAGCTGGGT CCTATTAGTGATGGTGATGGTGATGCTGGAAGTAGAGGTTCTCTGATTGTATGGTGCTGTCTATGACC-3'. For NRP-1–587 the same forward primer and the reverse primer 5'-GGGGACCACTTTGTACAAGAAAGCTGGGTCCTATTAGTGATGGTGA TGGTGATGCTGGAAGTAGAGGTTCTCGGCTTCCACTTCACAGCCCAG-3' were used. The PCR fragments were introduced into the pDONR221 and pcDNA-DEST40 vector to generate stably transfected human embryonic kidney (HEK) 293 cells. The protein products carry a tobacco etch virus protease cleavage site followed by a hexahistidine tag at their carboxyterminal ends. In vitro recombination reactions were performed using the Gateway® technology according to the manufacturer’s instructions (Invitrogen Corporation).
Soluble NRP-1 fragments were expressed by growing cells in Dulbecco’s modified Eagle’s medium supplemented with 6 mM sodium butyrate and 0.1% fetal bovine serum at near confluency. After 7 days, secreted protein was purified by IMAC. Bound protein was eluted with 500 mM imidazole in PBS and further purified on a SuperdexTM 200 column (GE Healthcare) equilibrated with PBS containing 1 mM EDTA. The correct identity of the proteins was confirmed by SDS-PAGE, immunoblotting, and Edman sequencing and the oligomeric state was determined by static light scattering.
Surface plasmon resonance (SPR)
Binding experiments were performed by SPR measurements on a BIACore T100 instrument (GE Healthcare) at 25°C using PBS at pH 7.4 supplemented with 0.05% surfactant P20 as running and sample buffer. NRP-1–641 was immobilized at high (1400 RU) and low (200 RU) level, respectively, on series S sensor chips CM5 using an amine coupling kit (GE Healthcare). Briefly, the surface was activated by a 7 min injection of a mixture of 0.2 M 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDC) and 0.05 M N-hydroxysuccinimide in water. NRP-1 was injected as a 5 or 10 µg/ml solution in 10 mM acetate buffer at pH 4.5 until the desired immobilization level was reached. Excess reactive groups were deactivated by a 7 min injection of 1 M ethanolamine-HCl, pH 8.5.
Biotinylated peptides were coupled to a series S sensor chip SA (GE Healthcare) by 5 s injections of 5 µg/ml solutions in PBS at 5 µl/minute to immobilization levels of 150 RU (Biotin-GSGSTEPPR), 212 RU (Biotin-GSGSTRPPR), and 228 RU (Biotin-GSGSTDKPRR), respectively. Ligands were diluted in running buffer to give series of concentrations from 0 to 1024 nM (NRP-1–587), 0 to 2000 nM (VEGFs), and 0 to 50 µM (peptides) and injected at 30 µl/min until equilibrium was reached. Data were double referenced, and dissociation constants were calculated by steady-state analyses with Biacore T100 1.1.1 evaluation software. Competition assays were performed by injecting VEGFs at constant concentrations near their KD (10 µM VEGF-A121, 750 nM VEGF-A165, 300 nM VEGF-E NZ2 
R3, or 300 nM VEGF-A-NZ2), supplemented with a series of 0–50 µM peptide until equilibrium was reached. Regeneration of the sensor chips was achieved by a 30 s injection of 4 M NaCl followed by a 10 s injection of 5 mM NaOH (CM5 chip) or a 5 s injection of 100 mM Na2CO3 (SA chip).
In solution HS filter binding assay
To determine VEGF binding to glycosaminoglycans (GAGs), [3H] GAGs (4 pmol) were incubated for 2 h with protein (0–0.5 nmol) in a final volume of 100 µl of 50 mM Tris/HCl, pH 7.4; 0.15 M NaCl; and 0.1 mg/ml bovine serum albumin (BSA). Protein along with bound polysaccharide was then trapped by filtration over a nitrocellulose filter (Sartorius, Edgewood, NY, USA), unbound polysaccharide was washed off with buffer, and bound carbohydrate was released with 2 M NaCl followed by scintillation counting as previously described (37)
.
VEGF stimulation and immunoprecipitation of receptor complexes
VEGFR-2 and VEGFR-2/NRP-1 expressing porcine aortic endothelial cells (PAE) and NRP-1 expressing PAE cells generated by transfection of PAE cells with pcDNA3-NRP-1 were maintained in Ham’s F12 supplemented with 10% fetal bovine serum and processed as described before (33)
. Cells (4x106) were plated into 15 cm tissue culture plates, grown for 2 days in Ham’s F12 medium, and starved in medium with 0.1% fetal bovine serum overnight before stimulation with 1.5 nM VEGF in medium containing 1 mg/ml BSA and 1 µg/ml heparin for 15 min on ice followed by incubation at 37°C for 10 min. Cells were washed once with cold PBS containing 100 µM Na3VO4 and lysed in 500 µl lysis buffer (HEPES, 20 mM, pH 8; 1% Triton-X-100; 10 mM EGTA; 5 mM MgCl2; 10% glycerol; 20 µg/ml leupeptin; 1% aprotinin; 1 mM PMSF; 20 µM phenyl arsine oxide; 20 mM NaPPi; 50 mM NaF; 100 µM Na3VO4; 1 mM DTT). Lysates were cleared by centrifugation at 13,000 rpm and incubated overnight with VEGFR-2 antibody Santa Cruz A3 or Abcam 11939. Protein A Sepharose (60 µl; GE Healthcare) was added to each sample for an additional hour at 4°C, and the beads were washed once with TNET (50 mM Tris, pH 7.5; 140 mM NaCl; 5 mM EDTA; 1% Triton X-100), TNE (50 mM Tris, pH 7.5; 140 mM NaCl; 5 mM EDTA) and H2O. Immunoprecipitates were boiled in Lämmli buffer and resolved on 8% SDS gels and blotted to polyvinylidine difluoride (PVDF) membranes and immunodecorated with phospho-specific, VEGF receptor-specific or phospholipase C -1 antibodies.
In vitro differentiation of stem cells and angiogenic sprouting assay
R1/SVJ 129 mouse ES cells (a kind gift of Dr. Andras Nagy, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, ON, Canada) were cultured as previously described (41)
. At day 0, ES cells were trypsinized and resuspended in ES medium without leukemia inhibitory factor to induce differentiation. Cells were aggregated in hanging drops (1200 cells/drop) on the lid of a nonadherent culture dish to form EBs. On day 4, EBs were seeded onto a solidified collagen matrix containing 1.5 mg/ml purified collagen I, Ham’s F12 medium (Invitrogen), 0.12% NaHCO3, 50 mM HEPES, and 5 mM NaOH. Subsequently, a second layer of collagen gel was added. EBs were maintained in the absence and presence of 1 µg/ml VEGF-A165 (PeproTech, Rocky Hill, NJ, USA), VEGF-A121, and VEGF-A-NZ2. At day 12, EBs in collagen gels were fixed in 4% p-formaldehyde (Sigma, St. Louis, MO, USA) in PBS for 30 min at room temperature, followed by blocking and permeabilization using 0.2% Triton X-100 in PBS with 3% BSA, for 2 h. After overnight incubation at 4°C with primary rat anti-mouse CD31 antibody (Becton Dickinson, Stockholm, Sweden), diluted in PBS with 3%BSA and 0.1% Tween 20, samples were washed several times in PBS with 0.1% Tween 20, and thereafter incubated with secondary antibodies (Alexa donkey anti-rat 594; Invitrogen, Molecular Probes) and with Hoechst 33342, to visualize nuclei. Samples were analyzed using a Nikon Eclipse E1000 microscope with a Nikon Eclipse DXM 1200 camera (Nikon, Tokyo, Japan). Quantification of the CD31-positive sprout area (obtained by subtracting the staining of the EB core area) was done using the Easy Image Analysis software (Tekno Optik, Skärholmen, Sweden).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Competitive binding studies performed with PAE cells expressing NRP-1 (Supplemental Fig. 1) showed that VEGF mutants carrying a carboxyterminal RPPR or DKPRR motif bound NRP-1. All VEGFs used in this study also bound VEGFR-2 (Supplemental Fig. 2) and immunoblots of cell lysates performed with phospho-specific antibodies recognizing tyrosine 1175 of VEGFR-2 and tyrosine 783 of phospholipase C-1, respectively, show that all VEGFs were functional and activated VEGFR-2 (Supplemental Fig. 3).
|
We have shown earlier that the six amino acids encoded by exon 8a are essential, but not sufficient, for VEGF-A165 binding to NRP-1. The naturally occurring splice variant VEGF-A165b did not bind NRP-1 and showed reduced binding to HS (33)
. This finding is confirmed by the data given in Fig. 2
, showing that VEGF-A165 promotes formation of stable coreceptor complexes between VEGF-A165, VEGFR-2, and NRP-1, while VEGF-A121 and VEGF-A165b do not.
|
Among the many variants of orf virus-expressed VEGF-Es, none have been identified so far that bind both HS and NRP-1, and we therefore used such VEGF-E variants to further characterize their receptor binding properties. VEGF-E NZ2 does not bind HS, presumably because it lacks a basic domain similar to that encoded by exon 7 in VEGF-A. However, VEGF-E NZ2 and NZ10, but not VEGF-E NZ7, interact with NRP-1 (5
, 9)
. On the basis of sequence comparison of the NZ2, NZ7, and NZ10 variants of VEGF-E, we concluded that the carboxyterminal sequence RPPR is the most likely NRP-1 binding site. This was tested with mutant proteins in coimmunoprecipitation assays with cells expressing VEGFR-2 together with NRP-1 (Fig. 2)
. VEGFs containing a carboxyterminal RPPR sequence as well as VEGF-A165 were precipitated together with NRP-1 by VEGFR-2 antibody. Mutation of arginine 126 to glutamate in VEGF-E NZ2 abolished NRP-1 binding, while deletion of the last three arginines had no effect. These data suggest that RPPR is the minimal sequence required for NRP-1 binding of VEGF-E.
To corroborate that RPPR is responsible for NRP-1 binding we engineered a recombinant VEGF-A in which the exon 8a-encoded domain was replaced with 14 amino acids from the carboxyterminus of VEGF-E NZ2. Figure 2
shows that this protein, VEGF-A NZ2, efficiently formed stable coreceptor complexes with VEGFR-2 and NRP-1. The mutant VEGF-A QEN, carrying RPPR internally rather than at the carboxyterminal end, binds NRP-1 (Supplemental Fig. 1), yet does not form stable receptor complexes (Fig. 2)
. The immunoprecipitation data also show that the carboxyterminus of VEGF-A121, DKPRR, is not sufficient for stable coreceptor complex formation.
VEGF-E has higher affinity for NRP-1 than VEGF-A165
VEGF-A165, but not VEGF-A121 or VEGF-A165b, forms stable coreceptor complexes with VEGFR-2 and NRP-1, suggesting that the exon 8a-encoded as well as the HS binding domain encoded by exon 7 are required for complex formation. In contrast, HS binding is dispensable for VEGF-E NZ2 mediated coreceptor complex formation.
Synthetic RPPR peptide did not disrupt receptor complex formation in coimmunoprecipitation assays at concentrations as high as 10 nM (data not shown); therefore, we determined the affinities of different VEGF variants for NRP-1 by SPR. VEGF-E NZ2 and the chimeric VEGF-A-NZ2 bound NRP-1 with higher affinity than VEGF-A165 (Fig. 3
A; Table 1
), suggesting that the RPPR motif might be both necessary and sufficient for stable coreceptor complex formation. The SPR data also show that amino acid 126 preceding the diproline motif is critical for VEGF-E NZ2 binding to NRP-1, since mutation to glutamate significantly reduced binding, while removal of the carboxyterminal three arginines had little effect.
|
|
NRP-1 binding through RPPR is superior to DKPRR-mediated binding
The affinity of a series of immobilized peptides representing putative NRP-1 binding sites was further analyzed by SPR using soluble NRP-1-587 protein. The peptide RPPR and the exon 8a peptide DKPRR bound NRP-1 with high and intermediate affinity, respectively, whereas the affinity for the sequence EPPR was distinctly lower, which showed that glutamate amino-terminal to the first proline prevents interaction with NRP-1 (Fig. 3B
). To corroborate that the peptides indeed bound to the VEGF binding site of NRP-1 and to exclude non-specific interaction, we competed VEGF binding to immobilized NRP-1 with each of the three peptides (Fig. 4
). RPPR efficiently inhibited VEGF binding at micromolar concentrations, while DKPRR and EPPR showed intermediate and weak inhibition, respectively.
|
HS binding of VEGF variants
HS/heparin binding of VEGF was tested with two assays, affinity chromatography on heparin columns (Supplemental Fig. 4) and a nitrocellulose filter binding assay using purified, 3H-radiolabeled polysaccharides. The only VEGF that bound heparin or HS with high affinity was VEGF-A165 (Fig. 5
). Interestingly, VEGF-A165b, which still carries the basic domain encoded by exon 7, or the mutant VEGF-A165 C160S, in which cysteine 160 is replaced by serine, also did not bind heparin (Supplemental Fig. 4). This finding suggests that a specific conformation of the exon 7 and 8a-encoded carboxyterminal domain is required for high-affinity interaction with HS. Binding of NRP-1 and HS/heparin to VEGF-A165 are, therefore, mediated by distinct but overlapping domains. Porcine skin CSB and squid cartilage CSE showed similar binding properties as HS or heparin, with exclusive binding to VEGF-A165 but to none of the other VEGF isoforms (Fig. 5B
).
|
VEGF also forms complexes with VEGFR-2 and NRP-1 expressed on separate cells
Spatially restricted deposition of VEGF-A, for instance upon immobilization in the extracellular matrix or on the surface of NRP-1-expressing cells, determines signal output in endothelial cells and is essential for vessel organization. Figure 6
shows that VEGF-A165 and VEGF-E NZ2 promoted the formation of paracellular VEGFR-2/NRP-1 complexes in cocultures of cells expressing either VEGFR-2 or NRP-1. This might reflect the situation in vivo where the spatiotemporal distribution of VEGF is critical for coordinating signal output in endothelial cells during vessel development and remodeling.
|
NRP-1 binding is essential for VEGF-mediated angiogenesis in differentiating ES cells
To determine that the chimeric VEGF-A-NZ2 was biologically active, we tested selected VEGF variants in a model of differentiating mouse ES cells, EBs. In this in vitro model, angiogenesis is faithfully recapitulated in 3D collagen matrix cultures. As shown in Fig. 7
, VEGF-A165 and VEGF-A-NZ2, which bind NRP-1 with high affinity, induced vessel-like structures, while VEGF-A121 that binds NRP-1 with low affinity had significantly reduced activity. These data suggest that stable coreceptor assembly by VEGF is an indicator for the potential of VEGF to promote vessel formation. The data imply that signal output by VEGFR-2/NRP-1 complexes differs from that induced by VEGFR-2 alone.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and transiently disrupt endothelial cell GAP junction permeability (42)
. VEGF RTKs activate distinct signaling networks following ligand binding. VEGFs bind a second type of transmembrane receptors, the neuropilins, which also interact with class 3 semaphorins regulating for instance neuronal cell guidance. VEGF signaling by neuropilins results from interaction of a short carboxyterminal motif with PDZ domain-containing proteins such as the adapter molecule GIPC (43
, 44)
. Distinct domains of VEGF are responsible for interaction with these two types of receptors, and only a subpopulation of the VEGF family proteins binds both receptors. The exon 7- and 8-encoded carboxyterminal domain was shown to be responsible for VEGF-A165 binding to NRP-1 (24
, 25
, 45)
. Based on data obtained with new VEGF-A splice variants, we recently assigned NRP-1 binding to the carboxyterminal six amino acids encoded by exon 8a (33)
. Furthermore, VEGF-A165 was shown to bind HS through a basic cluster of amino acids encoded by exon 7 (46)
.
Here we investigated the interaction of VEGF-A and orf virus VEGF-Es with NRP-1 and HS/heparin. More than 20 orf virus isolates differing in their ability to bind NRP-1 and/or HS have been described so far, which allow us to distinguish between the roles of NRP-1 and HS in VEGF signaling (5
, 6)
. HS binding of VEGF was determined by heparin affinity chromatography and in a nitrocellulose membrane binding assay. Both assays showed that VEGF-A165, but not VEGF-A121 and none of the VEGF-E variants, bound HS with high affinity. Earlier results based on peptides derived from VEGF-E NZ2 indicated that this protein might interact with HS (12)
. However, biological data supporting such an interaction using full-length proteins are not available and our data clearly show that this protein does not bind HS. Mutations of the carboxyterminal domain of VEGF-A165, such as replacement of exon 8a with 8b or replacing cysteine 160 with alanine in the mutant VEGF-A165 C160S, destroy both NRP-1 and HS binding, suggesting that in this protein NRP-1 and HS binding are interdependent. This finding is particularly surprising for the mutant VEGF-A165 C160S in which a single point mutation prevents the formation of a cross-link between cysteine 146 and 160, which normally stabilizes the carboxyterminal domain of VEGF-A165 as described earlier (47)
. In this mutant protein the carboxyterminal NRP-1 binding domain is apparently not accessible, and the HS binding domain encoded by exon 7 might assume a highly flexible structure kinetically unfavorable for binding.
Receptor coprecipitation assays performed with cells expressing NRP-1 and VEGFR-2 or with cocultured cells expressing either one of these receptors were used to further study NRP-1 binding of VEGF. VEGF-A121 has been shown to bind NRP-1 with low affinity, presumably through its exon 8a-encoded carboxyterminus. However, this protein does not form stable coreceptor complexes with VEGFR-2 and NRP-1 (35)
. VEGF-A165 and VEGF-E NZ2, however, bind NRP-1 with high affinity and allow immunoprecipitation of VEGFR-2/NRP-1 coreceptor complexes. To identify the sequence of VEGF-E responsible for NRP-1 binding, we created mutant proteins. Data for VEGF-E NZ2 and NZ10, which bind NRP-1 and contain a carboxyterminal RPPR motif, suggest that RPPR is the binding site for NRP-1. RPPR was clearly superior to the exon 8a-encoded peptide DKPRR of VEGF-A as shown by SPR analysis. Introduction of a glutamic residue to replace arginine 126 amino-terminal to PPR completely abolished NRP-1 binding, while other amino acid substitutions were shown earlier to be less disruptive (34
, 48)
. In the absence of structural information for an RPPR/NRP-1 complex the exact role of this arginine residue remains elusive.
To corroborate the functionality of the RPPR sequence, we created a chimeric VEGF-A, called VEGF-A-NZ2, carrying the carboxyterminal domain of VEGF-E NZ2 in place of the exon-8a-encoded sequence. In this molecule the VEGFR-2 and NRP-1 binding domains were separated by a 7 amino acid linker. VEGF-A-NZ2 efficiently formed stable coreceptor complexes with VEGFR-2 and NRP-1, and its activity in the EB angiogenesis assay was almost as high as that of VEGF-A165. In the longer VEGF-As the linker separating the carboxyterminal NRP-1 binding peptide from the exon 5-encoded domain might improve concomitant interaction of VEGF with NRP-1 and VEGFR-2. However, based on data obtained by Gluzman-Poltorak et al. (49)
for the splice variant VEGF-A145, in which the exon 5 domain is separated by a linker of 25 amino acids from the DKPPR motif and which does not bind NRP-1, we propose that linker length is not the determining factor in coreceptor complex formation. NRP-1 binding to VEGF-A depends on the structural context in which DKPRR is presented to the receptor as shown with peptides representing the carboxyterminal exon 7/8-encoded domain. The intramolecular crosslinks between cysteines 146 and 160 and 139 and 158 seem to be essential for high-affinity NRP-1 binding (50)
. Our SPR analysis of the binding affinities of VEGFs for NRP-1 showed that RPPR containing VEGFs bind NRP-1 with significantly higher affinity than VEGF-A165. We therefore conclude that HS binding might be required to further stabilize the interaction between VEGF-A165 and NRP-1, resulting in the formation of stable coreceptor complexes. This idea is also supported by the structure of the b1/b2 domain of NRP-1 published recently (51)
, which suggests that HS binding to a continuous heparin binding surface exposed on the surface of this dimer might stabilize the interaction with VEGF-A165. VEGF-E NZ2, however, binds NRP-1 through a high-affinity binding site apparently independent of cooperation with HS. Taken together, we propose that RPPR is both necessary and sufficient for NRP-1 binding. The inability of the NRP-1 binding motif of VEGF-A121, DKPRR, to promote stable coreceptor assembly in the absence of HS might, therefore, be due to its low binding affinity.
Our binding and coimmunoprecipitation data were complemented with functional data using the EB angiogenesis assay. EBs treated with VEGF recapitulate many aspects of natural vessel formation, such as sprouting and interaction of endothelial cells with smooth muscle cells. Our data clearly show that VEGF-A165 and VEGF-A-NZ2 induce vessel-like structures in EB cultures. VEGF-A121, however, has reduced angiogenic activity, indicating that high-affinity VEGF binding to both receptors is required for angiogenic signaling in EBs.
The role of VEGF-E and, in particular, of NRP-1 binding to VEGF-E, in orf virus replication is still unclear. VEGFR-2 activation may enhance virus dissemination by increasing vascularization of infected tissues (52)
. Similarly, the role of NRP-1 in VEGF signaling during vessel development and maintenance is only poorly understood. Earlier published data show that VEGF-A165 binding to NRP-1 increases VEGF receptor signaling through an unknown mechanism resulting in increased receptor activation (23
24
25
26
, 33
, 45)
. It was also shown that a chimeric VEGF-E variant carrying the HS binding and the exon 8 DKPRR motif from VEGF-A165 instead of its authentic carboxyterminus promoted increased complex formation between VEGFR-2 and NRP-1 (53)
. This mutant protein also showed increased competition with VEGF-A165 for VEGFR-2 binding when compared with VEGF-E, suggesting a role for both HS and NRP-1 in coreceptor assembly. Recent data suggest that neuropilins also modulate VEGF-A121-mediated receptor signaling, apparently in the absence of stable coreceptor complex assembly. This may result from ligand-driven assembly of membrane-bound receptor complexes in live cells that stimulates signaling by VEGFR-2 (54)
.
We have shown earlier that NRP-1-binding VEGFs activate downstream targets such as MAP kinases with different kinetics than VEGFs lacking NRP-1 binding (33)
. Whether this is the consequence of altered turnover of VEGFR-2 in the absence or presence of associated NRP-1 or whether NRP-1 association leads to structural changes in VEGFR-2 resulting in the phosphorylation of distinct sites activating specific downstream signals remains to be established.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 These authors contributed equally to this work.
Received for publication February 2, 2008. Accepted for publication April 10, 2008.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  V.-M. Leppanen, A. E. Prota, M. Jeltsch, A. Anisimov, N. Kalkkinen, T. Strandin, H. Lankinen, A. Goldman, K. Ballmer-Hofer, and K. Alitalo Structural determinants of growth factor binding and specificity by VEGF receptor 2 PNAS, February 9, 2010; 107(6): 2425 - 2430. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Tremmel, A. Matzke, I. Albrecht, A. M. Laib, V. Olaku, K. Ballmer-Hofer, G. Christofori, M. Heroult, H. G. Augustin, H. Ponta, et al. A CD44v6 peptide reveals a role of CD44 in VEGFR-2 signaling and angiogenesis Blood, December 10, 2009; 114(25): 5236 - 5244. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Lambert, M. Bouttier, R. Vassy, M. Seigneuret, C. Petrow-Sadowski, S. Janvier, N. Heveker, F. W. Ruscetti, G. Perret, K. S. Jones, et al. HTLV-1 uses HSPG and neuropilin-1 for entry by molecular mimicry of VEGF165 Blood, May 21, 2009; 113(21): 5176 - 5185. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Kawamura, X. Li, K. Goishi, L. A. van Meeteren, L. Jakobsson, S. Cebe-Suarez, A. Shimizu, D. Edholm, K. Ballmer-Hofer, L. Kjellen, et al. Neuropilin-1 in regulation of VEGF-induced activation of p38MAPK and endothelial cell organization Blood, November 1, 2008; 112(9): 3638 - 3649. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
