| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Institute of Genetics and Biophysics "A. Buzzati-Traverso", CNR, Naples, Italy.
Department of Biochemistry and Molecular Biology, Genes and Development Research Group, University of Calgary, Calgary, Alberta, Canada T2N 4N1
1 Correspondence: Institute of Genetics and Biophysics "A. Buzzati-Traverso", CNR, Via Pietro Castellino 111, 80131 Naples, Italy. E-mail: persico{at}igb.cnr.it
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Key Words: growth factors • evolutionary conservation • Flt-1/KDR
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
2
3
). Besides the major role that PDGF/VEGF factors play in vascular development and vascular maintenance, additional roles have recently been revealed in neural development (for reviews, see refs 2
, 4
, 5
). PDGF/VEGF signaling has been well studied at the molecular level: four different forms of the PDGF ligand (A–D) bind to two cognate receptors (PDGFR 
and ß) (2)
while six different forms of VEGF ligand (A–E and PlGF) bind to three cognate receptors VEGFR-1 (Flt-1), VEGFR-2 (KDR), and VEGFR-3 (6)
. VEGF-A, VEGF-B, and PlGF have also been shown to bind to the neuropilin-1 and neuropilin-2 coreceptors (7
, 8)
. A subset of the PDGF/VEGF factors binds to heparan sulfate in the extracellular matrix (for reviews, see refs 2
, 3
).
While the vast majority of interest in PDGF/VEGF-like growth factors and their receptors has been in vertebrates, candidate homologs have also been identified in invertebrates. In Drosophila, three PDGF/VEGF-like factors (called pvf1, 2, and 3) and one receptor (pvr) have been implicated in migration of cells bordering the oocyte (9)
, maintenance of the wing disc monolayered epithelium (10)
, and rotation of male terminalia (11)
; they have also been implicated in the migration of early hemocytes (12
13
14)
, possibly pointing to an evolutionarily conserved role in blood formation. In the nematode C. elegans, analysis of the genomic sequence has identified four possible homologs of PDGF/VEGF receptors (VER-1 to VER-4) but no ligand has yet been identified (15
, 16)
. In the present paper, we describe just such a potential PDGF/VEGF-like ligand from C. elegans. Our major conclusions are that this ligand (named PVF-1) has the biochemical properties expected for a bona fide member of the vertebrate PDGF/VEGF family of growth factors. Most important, PVF-1 has the ability to bind to human receptors VEGFR-1 and VEGFR-2 and to induce angiogenesis in two model systems derived from vertebrates.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Vector constructions
The plasmid yk305 g9 containing the cDNA coding for PVF-1 was obtained from Dr. Y. Kohara (National Institute of Genetics, Japan). The cDNA insert (lacking sequences corresponding to the signal peptide and the 3'-UTR) was PCR-amplified and transferred into plasmid pGEX-5X-1 (Pharmacia, Piscataway, NJ, USA) and plasmid pET-22b(+) (Novagen, Madison, WI, USA) for production of protein in E. coli, as well as into plasmid pcDNA3-His-Tag (Invitrogen, San Diego, CA, USA) for production of PVF-1 in eukaryotic cells. To overexpress PVF-1 in C. elegans, a 4859 bp region of genomic sequence containing the pvf-1 gene (beginning 2 kb upstream of the initiation codon and ending with the termination codon) was amplified by PCR and inserted into plasmid pPD95.75 (kindly provided by A. Fire, Carnegie Institute, Baltimore). For all the above steps, details of primers and cloning strategies will be provided upon request. Transgenic C. elegans were produced by standard methods (17)
, using rescue of lin-15(n765ts) to identify transformants; constructs were injected at concentrations of 50–100 µg/mL. The stably transformed lines of transgenic worm overexpressing PVF-1 do not show any obvious phenotype.
Antibody production
The pET-PVF-1 plasmid was used to transform the BL21 (DE3) codon-plus strain of E. coli (Stratagene, San Diego, CA, USA). Recombinant protein, after induction for 3 h at 37°C in the presence of 1 mM isopropyl-D-thiogalacto-pyranoside, was purified from inclusion bodies using preparative SDS/12% PAGE. A total of 1 mg of bacterially produced PVF-1 was used to immunize two rabbits as described previously (18)
. Recombinant PVF-1 for antibody purification was produced by a similar method but using the pGEX-PVF-1 plasmid. Anti-PVF-1 serum was incubated for 15 h at 4°C with the corresponding antigen bound to nitrocellulose filters. After washing in PBS buffer, the purified antibodies were eluted with a 100 mM glycine buffer pH 2.8, neutralized with 1M Tris-HCl buffer to pH 7.5, and stored in aliquots at –20°C.
Cell culture
Growth of HEK 293 and HUVEC cells (Clonetics, San Diego, CA, USA), transient transfection of HEK 293T cells, selection of stably transfected HEK 293 cells, and collection of conditioned media were performed as described previously (19)
. Among the stable clones isolated, one cell line (293-PVF-1) was selected for high-level PVF-1 expression (300 ng/mL), as determined by Western blot and quantitative ELISA assays. A control cell line (293-pcDNA3) stably transfected with the empty vector was also established.
Western blot analysis of worm extracts and cell culture media
N2 (wild-type) or PVF-1-overexpressing worms were grown on standard NGM agar with OP50 bacteria as food (20)
. Mixed stage cultures were collected, washed by centrifugation in PBS, suspended in 10 mM Tris-HCl pH 8, 150 mM NaCl, 1% Triton X-100, and lysed by sonication. Following centrifugation at 4°C for 5 min at 13,000 rpm, supernatants were collected and the protein concentration was determined by the Bradford assay (Bio-Rad, Hercules, CA, USA). One hundred micrograms of total protein from each worm extract was used for Western blot analysis as described below. Aliquots of conditioned cell media were concentrated, suspended in SDS loading buffer with or without 10% 2-mercaptoethanol (for reducing or nonreducing conditions, respectively), and loaded onto 12% SDS-PAGE as described previously (19)
. Gels were blotted onto PVDF membranes, which were incubated for 1 h at room temperature with 300 ng/mL of purified anti-PVF-1 polyclonal antibodies and analyzed as described (19)
.
Protein purification, N-glycosidase F digestion, and heparin binding assay
PVF-1-His protein was purified from media of 293-PVF-1 cells using the QIAexpress protein purification system (Qiagen, Chatsworth, CA, USA). Two hundred microliters of conditioned media were treated with N-glycosidase F according to the manufacturer’s instructions and as described previously (18)
. To determine the binding ability of the PVF-1 protein to heparin, 2 mL of conditioned media from the stable 293-PVF-1 cell line was mixed with 100 µL of heparin-Sepharose beads and treated as described (18)
.
ELISA assays
To determine the binding activity of PVF-1 to VEGFR-1 (Flt-1), VEGFR-2 (KDR), or PDGFR ß, 96-well microtiter plates were coated with 100 µL/well of the soluble form of human Flt-1 (Flt-1/Fc chimera) or human KDR (KDR/Fc chimera) at 0.1 µg/mL or 0.5 µg/mL of soluble PDGFR ß receptor in PBS pH 7.5. Varying amounts, 3–100 ng/mL, of PVF-1 was used for binding. Incubation, washing, and absorbance measurements were performed as described (19)
. Positive controls for the binding assay of VEGFR-1 (Flt-1), VEGFR-2 (KDR), or PDGFR ß were run in parallel using 1–20 ng of VEGF and 1–20 ng of PDGFB (kindly provided by C. Heldin, Ludwig Institute for Cancer Research, Uppsala, Sweden). For the competitive binding assay, a 96-well microtiter plate was coated with a soluble form of human Flt-1 (Flt-1/Fc chimera) at 0.1 µg/mL in PBS pH 7.5 (100 µL/well, overnight at RT). The plate was washed five times with PBT (PBS containing 0.004% Tween-20). After blocking of nonspecific sites (for 3 h at RT with 1% bovine serum albumin (BSA) in PBS), 2 ng/mL of recombinant hVEGF-A165, diluted in PBET (PBS containing 0.1% BSA, 5 mM EDTA, 0.004% Tween 20), was added to the wells together with the competitor recombinant PVF-1, at concentrations ranging from 12.5 to 250 ng/mL. The binding reaction was performed for 1 h at 37°C and 1 h at RT. Wells were washed as described above and incubated (1 h at 37°C and 1 h at RT) with a biotinylated anti-human VEGF-A polyclonal antibody (300 ng/mL in PBET). Wells were washed again and incubated with a solution containing a preformed avidin-biotinylated-HRP macromolecular complex (Vectastain elite ABC kit) for 1 h at RT. The detection was performed as described previously (19)
. All the reported ELISA assays were performed in duplicate.
For quantitative determination of PVF-1, purified anti-PVF-1 polyclonal antibody, diluted to 1 µg/mL in PBS pH 7.5, was used to coat a 96-well plate (100 µL/well, overnight at 4°C). Recombinant PVF-1, at concentrations ranging from 25 to 500 ng/mL, or the conditioned media from PVF-1-producing HEK-293 cells, were then added. Washing, incubation and reading were performed as described above.
VEGFR-1 phosphorylation analysis
VEGFR-1 (Flt-1) phosphorylation analysis was performed as described previously (19)
. Subconfluent 293-hFlt-1 cells were starved ON in serum free medium. After treatment with Na3VO4, cells were incubated for 10 min at 37°C with conditioned medium from 293-pcDNA3 cells or with PVF-1-producing 293 cells (undiluted or diluted 1:1 with control medium, corresponding to 300 or 150 ng PVF-1/mL). After stimulation, cells were washed and treated as described (19)
. To immunoprecipitate VEGFR-1 receptor, 5 µg/mL of a goat polyclonal against human VEGFR-1, preincubated ON at 4°C with protein-G Sepharose, was added to 1 mg of protein extract and incubated ON at 4°C. Subsequently, the resin was recovered, washed, and loaded onto reducing 8.5% SDS PAGE. Phospho-tyrosine (PY) detection, followed by VEGFR-1 detection on the same filter after stripping, was performed as described (19)
.
Capillary-like tube formation in HUVEC cells
The formation of capillary-like structures was assayed as described previously (19)
in the presence of 1) one mL conditioned media produced by 293 cells expressing PVF-1 at a concentration of 300 ng/mL; 2) one mL of 200 ng/mL VEGF-A, or 3) 1 mL conditioned media produced by control 293-pcDNA3 cells. All assays were performed in triplicate wells. To confirm the specificity of PVF-1 activity, a purified neutralizing antibody (100 µg/mL) was preincubated (overnight at 4°C) with the conditioned media containing PVF-1.
PVF-1 induced angiogenesis in the chorioallantoic membrane (CAM) assay
Pellets of 3x106 cells from the stable cell line expressing PVF-1 (300 ng/mL) or hPlGF (30 ng/mL) were applied to the chorioallantoic membrane of 10 fertilized white Leghorn chicken eggs. Incubation, photography, and counting of allantoic vessels that enter the cell pellets were as described previously (19
, 21)
. Cell pellets from the cell line stably transformed with the empty vector were used as controls.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
using this consensus sequence identified a cDNA clone (yk305 g9) that corresponds to the gene Y119D3.663. We named this gene pvf-1 (for PDGF/VEGF-like growth factor), as have similar genes been named in Drosophila melanogaster (9
, 10)
; where necessary, the worm gene will be referred to as Cepvf-1. We sequenced the cDNA clone and confirmed the predicted gene structure present in the C. elegans database. The highly conserved central domain (see below) is encoded in exons 3, 4, and 5 of the overall 7-exon gene, just as it is in all members of this gene family (6
, 23)
. The amino acid sequence of PVF-1 is shown in Fig. 1
A (NCBI accession number NP_497461.1). The 304 residue open reading frame begins with a 19 residue hydrophobic sequence (underlined in Fig. 1A
) with characteristics of a signal peptide (24)
. After removal of this peptide, the mature (unmodified) molecular mass is predicted to be 32 kDa. Figure 1A
also indicates three potential N-glycosylation sites (25)
.
|
Figure 1B
shows an alignment between the amino acid sequences of the central 8-cysteine domain from PVF-1 and from selected members of the PDGF/VEGF family from vertebrates, Drosophila, and the closely related nematode Caenorhabditis briggsae. The most striking observation is the highly conserved positioning of the cysteine residues characteristic of this growth factor family and that are important for both dimerization and functional activity. Six cysteines are involved in intra-chain disulfide bridges, and two cysteine residues participate in inter-chain covalent bonds (19
, 26
, 27)
. Overall, the percentage of amino acid identity is in the range from 19% (hPDGF-B) to 93% (C. briggsae PVF-1).
In addition to the highly conserved central domain, the cysteine-rich C-terminal domain of PVF-1 shows similarity (Fig. 1C
) to C-terminal domains present in the members of the VEGF family (–A through –D), C. briggsae PVF-1 and Drosophila PVF1 (11)
. This region of the protein allows the binding to the extracellular matrix (1
, 6)
, i.e., C. elegans PVF-1 is predicted to bind to heparan sulfate (see below).
PVF-1 is a secreted glycosylated dimeric protein that binds to heparin
We now demonstrate that the C. elegans PVF-1 factor shows biochemical properties expected for members of the PDGF/VEGF family—namely, secretion, dimerization, glycosylation, and heparin binding.
To demonstrate that PVF-1 is a secreted protein, human 293 cells were transfected with a construct in which the pvf-1 cDNA was fused to a 6-His tag within the expression vector pcDNA3 (see Materials and Methods). As shown in Fig. 2
A, lane 1, PVF-1 protein is easily detected in the conditioned medium of the transfected cells, using an anti-PVF-1 antiserum (see Materials and Methods); no bands are detected with the same anti-PVF-1 antiserum in the conditioned medium of cells transfected with the empty pcDNA3 plasmid (Fig. 2A
, lane 2). The majority of the secreted PVF-1 migrates as two bands of an apparent size of 30–35 kDa, reasonably close to but slightly larger than the size predicted for the polypeptide lacking the signal peptide. At the moment, we do not know the origin of the several reactive minor bands detected in the vicinity of the main bands, whether these are due to postranslational modification (see below) or to cleavage, either specific or nonspecific; a similar set of bands was detected with PVF1 from Drosophila (10)
.
|
Judging from the highly conserved PVF-1 central domain (Fig. 1B
) with its cysteine residues involved in the dimerization of PDGF/VEGF growth factors, PVF-1 is predicted to be dimeric. As shown in Fig. 2B
, lane 1, when 293 cell-derived PVF-1 protein is electrophoresed under nonreducing conditions, immunoblots do indeed reveal a higher band of the 60–70 kDa size expected for a PVF-1 dimer and not seen under reducing conditions (Fig. 2A
). Clearly, the majority of the secreted PVF-1 protein does not migrate as this dimeric species, either because of adventitious reduction of disulfide bonds during isolation or because of the high level of expression not allowing efficient dimer formation. Nonetheless, we conclude that PVF-1 is capable of forming a dimer involving interchain disulfide bridges.
All members of the PDGF/VEGF family are N-glycosylated proteins (2
, 6)
, and the PVF-1 sequence contains three potential N-glycosylation sites (underlined in Fig. 1A
). As shown on Fig. 2C
, treatment of 293 cell-produced PVF-1 with the deglycosylating enzyme N-glycosidase F (see Materials and Methods) causes the PVF-1 protein bands to migrate more rapidly on reducing SDS gels, as expected if PVF-1 is N-glycosylated.
Alternative splicing pathways produce a variety of different isoforms from PDGF/VEGF family members. Forms that terminate after the central eight-cysteine motif are soluble but forms that contain a C-terminal arginine-rich or cysteine-rich domain bind to heparan sulfate in the extracellular matrix (for reviews, see refs 1
, 6
, 10
, and 28
). As shown on Fig. 1A, C
, PVF-1 contains a C-terminal cysteine-rich domain (corresponding to exons 6 and 7) that could potentially bind to heparin. Figure 2D
shows that a significant fraction of the PVF-1 (secreted from 293 cells) does indeed bind to a heparin-Sepharose column and is released by 300–500 mM NaCl, the same salt concentrations that release heparin binding forms of other PDGF/VEGF family members (6)
.
The same Western blot assay that can easily detect PVF-1 secreted from transfected 293 cells cannot detect PVF-1 in wild-type worms (Fig. 2E
, lane 1) but can detect the appropriate bands in extracts from a worm strain transformed with the wild-type pvf-1 gene (see Fig. 2E
, lane 2). This result is not unexpected. The fact that only five pvf-1 cDNAs have been identified in the C. elegans database and that SAGE libraries (http://elegans.bcgsc.ca/home/sage.html) prepared from different worm stages show a low number of pvf-1 tags (1 or 2 per 
100,000 tags) indicates that pvf-1 is expressed at a relatively low level in wild-type worms (29)
. It is well established that worm strains containing the usual multi-copy transgenic arrays invariably overexpress the transgene by many fold (see, for example, ref 30
).
There are, nonetheless, several points of interest associated with PVF-1 produced in the transgenic worms: 1) under nonreducing conditions, higher molecular weight species can be detected in Western blots, as expected if PVF-1 can form dimers within the worm (data not shown); 2) 10% of the protein migrates as several lower molecular weight bands that could be due to specific or nonspecific cleavage; and 3) PVF-1 produced in worms consistently migrates slightly slower on reducing SDS gels than does PVF-1 produced in 293 cells (data not shown). This difference could indicate more extensive post-translational modification in worms or perhaps that, under these overexpressing conditions, the signal peptide is not cleaved from the primary translation product. The fact that this slower migration is not altered by incubation with N-glycosidase (data not shown) would favor the latter possibility; on the other hand, it is not certain that all C. elegans glycosylation products can be removed by the enzyme treatment standard for vertebrate proteins (31)
.
PVF-1 binds to human VEGFR-1 and VEGFR-2 receptors
An ELISA-based assay was used to demonstrate that PVF-1 can bind specifically to the two human growth factor receptors VEGFR-1 (Flt-1) and VEGFR-2 (KDR). As described in more detail in Materials and Methods, a fixed concentration of a soluble form of human VEGFR-1 (Flt-1/Fc chimera) or human VEGFR-2 (KDR/Fc chimera) was used to coat a microtiter plate. Purified recombinant PVF-1 produced in HEK 293 cells was added to the wells, at concentrations ranging from 3 to 100 ng/mL. After extensive washing, bound PVF-1 was detected using biotinylated polyclonal anti-PVF-1 antibodies. As shown in Fig. 3
A, PVF-1 binding to both VEGFR-1 and -2 receptors can easily be detected. The binding is saturable and is of roughly the affinity expected for specific binding (19)
. No binding was detected in the same type of assay of PVF-1 to PDGFR ß (data not shown) or to human VEGFR-3 (K. Alitalo, personal communication). The specificity of the binding to the VEGFR-1 and -2 is further supported by the fact that PVF-1 can compete for binding of recombinant hVEGF-A165 to the VEGFR-1 receptor, as shown on Fig. 3B
. As far as can be judged from the properties of the competition curve, PVF-1 binds at least an order of magnitude more weakly to VEGFR-1 than does hVEGF-A (27)
.
|
PVF-1 causes the phosphorylation of human VEGFR-1
The in vitro data shown above indicate that PVF-1 is able to bind the soluble recombinant form of the human growth factor receptors VEGFR-1 (Flt-1) and VEGFR-2 (KDR). To investigate whether this binding leads to activation (phosphorylation) of the receptors on the cell surface, the stable cell line 293-Flt-1 was exposed to conditioned media containing PVF-1; VEGFR-1 phosphorylation was then evaluated by immunoprecipitation and Western blot assays. As shown in Fig. 3C
, incubation of starved 293-Flt-1 cells with PVF-1 (150–300 ng/mL for 10 min) clearly induces phosphorylation of the VEGFR-1 receptor.
PVF-1 is able to stimulate angiogenesis in two different vertebrate assay systems
We first show that PVF-1 stimulates capillary-like tube formation by human umbilical vein endothelial cells (HUVEC). HUVEC were grown on Matrigel and treated for 8 h with conditioned medium from human 293 cells expressing PVF-1 cDNA (300 ng/mL of PVF-1 protein). As shown in Fig. 4
A, PVF-1 is able to stimulate cell migration and the formation of capillary-like tubes. For comparison and as a positive control, the same Matrigel-grown cells were treated in parallel with 200 ng/mL of recombinant VEGF-A (Fig. 4B
). Two negative controls show that the effect is specific: 1) conditioned medium from human 293 cells transfected with the empty pcDNA3 plasmid does not induce capillary formation (Fig. 4C
), and 2) capillary tube formation induced by the condition medium containing PVF-1 (300 ng/mL) is largely abolished by preincubation with 100 µg/mL of anti-PVF-1 antibodies (Fig. 4D
).
|
We also show that PVF-1 can stimulate angiogenesis in the embryonic chorioallantoic membrane (CAM) assay. Human 293 cells (3x106 cells per pellet) expressing either PVF-1 cDNA or hPlGF or, as negative control, containing the empty pcDNA3 plasmid, were deposited onto the chorioallantoic membranes of 8-day-old chicken embryos (see Materials and Methods for details). The angiogenic response was quantitated by counting the number of blood vessels that develop radially toward and then enter the cell implant, immediately after the implantation and following a further 3-day incubation (see refs 19
, 21
for details). The magnitude of the PVF-1 induced angiogenesis is similar to that seen with hPlGF in the same assay (Fig. 5
, refs 19
, 21
).
|
|
SUMMARY AND CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Questions for the future are where the pvf-1 gene is expressed inside the worm, whether PVF-1 binds to the four neuronally expressed "VER" receptors described by Popovici and co-workers (16)
, whether it binds to some entirely new receptor, and finally, whether the loss of pvf-1 function confers a phenotype on the worms during development or during adult life (Tarsitano et al., unpublished results). The principal results of the current paper are the demonstration that the C. elegans PVF-1 protein can bind to mammalian VEGF receptors 1 and 2 and that PVF-1 has angiogenic properties in the experimental systems that have been used to define and explore vasculogenesis/angiogenesis in vertebrates. Furthermore, PVF-1 does not bind to human receptor VEGFR-3 (K. Alitalo personal communication) or to the PDGF receptor ß, perhaps suggesting that PVF-1 is more closely related to VEGF-A and PlGF, which are known to be involved in both angiogenesis and neuronal growth (4, 5, 32). These results should provide added significance to the genetic study of such growth factors in invertebrates, such as flies and worms.
|
ACKNOWLEDGMENTS |
|---|
Received for publication June 16, 2005. Accepted for publication October 13, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. Andrae, R. Gallini, and C. Betsholtz Role of platelet-derived growth factors in physiology and medicine Genes & Dev., May 15, 2008; 22(10): 1276 - 1312. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. G. Ball, C. A. Shuttleworth, and C. M. Kielty Vascular endothelial growth factor can signal through platelet-derived growth factor receptors J. Cell Biol., May 7, 2007; 177(3): 489 - 500. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
