Neuropilin-2 and Neuropilin-1 Are Receptors for the 165-Amino Acid Form of Vascular Endothelial Growth Factor (VEGF) and of Placenta Growth Factor-2, but Only Neuropilin-2 Functions as a Receptor for the 145-Amino Acid Form of VEGF*

Neuropilin-1 (np-1) and neuropilin-2 (np-2) are receptors for axon guidance factors belonging to the class 3 semaphorins. np-1 also binds to the 165-amino acid he-parin-binding form of VEGF (VEGF 165 ) but not to the shorter VEGF 121 form, which lacks a heparin binding ability. We report that human umbilical vein-derived endothelial cells express the a17 and a22 splice forms of the np-2 receptor. Both np-2 forms bind VEGF 165 with high affinity in the presence of heparin ( K D 1.3 3 10 2 10 M ) but not VEGF 121 . np-2 also binds the heparin-binding form of placenta growth factor. These binding charac-teristics resemble those of np-1. VEGF 145 is a secreted heparin binding VEGF form that contains the peptide encoded by exon 6 of VEGF but not the peptide encoded by exon 7, which is present in VEGF 165 . VEGF 145 binds to np-2 with high affinity ( K D 7 3 10 2 10 M ). Surprisingly, VEGF 145 did not bind to np-1. Indeed, VEGF 145 does not bind to MDA-MB-231 breast cancer cells, which predom-inantly express np-1. By contrast, VEGF 145 binds to hu- man umbilical vein-derived endothelial cells, which express both np-1 mg/ml). np-2-expressing cell co-transfecting PAE cells with the PECE/np-2(a17) or PECE/np-2(a22) expression vectors and pBabePuro plasmid (49), by selection with 0.5 m g/ml puromycin (49). Transfection was carried out using LipofectAMINE according to manufacturer’s instruc- tions. subsequently puromycin or G418. MDA-MB-231 cancer as

The various growth factors belonging to the VEGF 1 family (VEGF, PlGF, VEGF-B, VEGF-C, and VEGF-D) act as modulators and inducers of angiogenesis in vivo (1)(2)(3). The active forms of VEGF are synthesized as homodimers (4,5) or as heterodimers with other VEGF family members such as PlGF (6). Targeted disruption of the VEGF gene has shown that even in animals containing a single allele of the VEGF gene angiogenesis is severely disrupted, indicating that the maintenance of exact VEGF levels in vivo is critical for the correct development of the cardiovascular system (7,8). Five splice forms of human VEGF ranging in length from 121 to 206 amino acids (VEGF 121 -VEGF 206 ) have been characterized (4,5,9,10). These forms differ primarily in the presence or absence of the heparin binding domains encoded by exons 6 and 7, giving rise to forms that differ in their heparin and heparan-sulfate binding ability (11). The VEGF 121 , VEGF 165 , and VEGF 189 forms appear to be abundant and are usually produced simultaneously by VEGF-producing cells (1). VEGF 145 is a rarer and much less studied VEGF form, which was reported to be expressed by cells derived from the female reproductive system (9) as well in other organs such as the skin, penis, and kidney (12)(13)(14). It contains the heparin binding domain encoded by exon 6 and binds tightly to the extracellular matrix (9). VEGF 165 contains the heparin binding domain included in exon 7 and binds to heparin with an affinity that is similar to that of VEGF 145 . However, VEGF 165 binds much less tightly than VEGF 145 to extracellular matrix. VEGF 121 lacks both exons and has no affinity for either heparin or for extracellular matrixes. Other VEGF family members such as PlGF and VEGF-B are also expressed in several forms that differ in their heparin binding ability. For example, the peptide encoded by exon 6 of PlGF is found only in PlGF-2 and confers a heparin binding ability to this form while PlGF-1 and PlGF-3 do not bind to heparin (15,16).
All the VEGF isoforms bind to the tyrosine-kinase receptors VEGFR-1 (flt-1) (17) and VEGFR-2 (KDR/flk-1) (18). The binding of VEGF to VEGFR-2 initiates intracellular signal transduction (1, 19 -22) and is correlated with the induction of endothelial cell proliferation and migration, angiogenesis, and permeabilization of blood vessels (1,23,24). In contrast the activation of VEGFR-1 does not seem to result in the induction of cell proliferation, angiogenesis, or permeabilization of blood vessels but enhances cell migration (24 -26). However, there have also been other reports that indicate that the activation of VEGFR-1 can induce cell proliferation and angiogenesis (27). Both of these receptors have also been shown to play critical roles in embryonic vasculogenesis and angiogenesis (28,29).
Endothelial cells also contain another type of VEGF receptors possessing a lower mass than either VEGFR-2 or VEGFR-1 (30,31). It was subsequently found that these smaller VEGF receptors of the endothelial cells are isoform specific receptors that bind VEGF 165 but not VEGF 121 (32). It was therefore recognized that these receptors are not related to * This work was supported by grants from the Israel Academy of Sciences and from the Israel Cancer Research Fund (to G. N.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ The first two authors contributed equally to this work. § To whom correspondence should be addressed. Tel.: 972-294-216; Fax: 972-225-153; E-mail: gera@techunix.technion.ac.il. 1 The abbreviations used are: VEGF, vascular endothelial growth factor; BS 3 , bis(sulfosuccinimidyl) suberate; HUVEC, human umbilical vein-derived endothelial cell; PAE, porcine aortic endothelial; np-1, neuropilin-1; np-2, neuropilin-2; PAGE, polyacrylamide gel electrophoresis; VEGF 165 , 165-amino acid form of vascular endothelial growth factor; VEGF 121 , 121-amino acid form of vascular endothelial growth factor; VEGF 145 , 145-amino acid form of vascular endothelial growth factor; VEGFR-1, tyrosine kinase vascular endothelial growth factor receptor-1 (flt-1); VEGFR-2, tyrosine-kinase vascular endothelial growth factor receptor-2 (KDR/flk-1); PCR, polymerase chain reaction; PlGF, placenta growth factor. the VEGFR-1 or to the VEGFR-2 receptors, which bind to both VEGF isoforms. An additional search revealed several types of prostate and breast cancer cell lines, which express unusually large amounts of these isoform-specific receptors (33). A VEGF 165 affinity column was used to purify the receptors from MDA-MB-231 breast cancer cells, and sequencing of the receptor revealed it to be the product of the gene for np-1 (34). NP-1 is likely to play an important role in the development of the cardiovascular system. Gene disruption studies have indicated that np-1 is probably an important regulator of blood vessel development since mouse embryos lacking a functional np-1 gene die because their cardiovascular system fails to develop properly (35). Subsequent experiments have shown that np-1 also serves as a receptor for the heparin-binding form of placenta growth factor (PlGF), PlGF-2, and for VEGF-B (36,37). np-1 was previously identified as a receptor for semaphorin-3A (38,39) and it was demonstrated that the VEGF 165 and semaphorin-3A binding sites on np-1 overlap (40). Semaphorins act as repellents of growing tips of axons, and it was recently observed that semaphorin-3A inhibits migration of endothelial cells causing a collapse of the actin cytoskeleton (41).
NP-1 is part of a receptor family that also contains np-2, a receptor that displays highly similar structural features and is a receptor for semaphorin-3C and for semaphorin-3F. Interestingly, np-1 and np-2 can form complexes (40,42). NP-2 shares a 44% identity at the amino acids level with np-1. Several alternatively spliced forms of np-2 that can be divided into two broad groups have been identified. Mouse group A variants have insertions of 5, 17, or 22 amino acids at amino acid 809 (42). In humans only one splice form of this group was identified having a 17-amino acid insertion at position 808 (42). Mouse group B isoforms differ in the sequence of the transmembrane and intracellular parts starting from amino acid 809, and two such isoforms have been identified (42). The expression pattern of np-2 differs from that of np-1. Although there are broad overlaps, there are regions in which there is np-1 expression but no np-2 expression. For example, in contrast to np-1, np-2 expression was not detected in the heart or in capillaries but was found in the dorsal aorta colliculus (40). The neuropilins have short intracellular domains and are therefore unlikely to function as independent signaling receptors, and it was indeed shown recently that plexins form signaling complexes with both neuropilins (43,44).
Since np-1 functions as a VEGF receptor, we sought to determine whether np-2 is also a receptor for VEGF family members. We have cloned the np-2 cDNA from HUVEC. We show that np-2 functions as a splice variant specific VEGF receptor that binds VEGF 165 but not VEGF121. However, the two receptors do not behave equally with regard to their interactions with VEGF since np-2 is able to bind VEGF 145 , a VEGF splice form that does not bind to np-1.

EXPERIMENTAL PROCEDURES
Materials-The VEGF splice forms and PlGF-2 were produced in sF9 cells using appropriate baculoviruses and purified as described previously (9,36,45,46). LipofectAMINE was bought from Life Technologies, Inc. The pBabePuro plasmid was kindly given to us by Dr. Eyal Ben-Gal from the Technion School of Medicine (Haifa, Israel). BS 3 was obtained from Pierce. Tissue culture media, sera, and cell culture supplements were from Beth-Haemek Biological Industries (Kibbutz Beth-Haemek, Israel), or from Life Technologies, Inc. All other chemicals were purchased from Sigma. The pre-stained protein size markers used were purchased from Bio-Rad.
Preparation of np-1 and np-2 cDNA-The np-1 and np-2 cDNA were reverse transcribed using Moloney murine leukemia virus reverse transcriptase (U. S. Biochemical Corp.) from total RNA prepared from HU-VEC as described (47). The cDNA was amplified using the expand high fidelity PCR system kit (Roche Molecular Biochemicals). The primers used were ACGCGGATCCACCATGGAGAGGGGGCTGCCGCT and AGCGGATCCGTCGACTCATGCCTCCGAATAAGTACTC for np-1 and GCTCTAGAACCATGGATATGTTTCCTCTCACC and GCTCTAGAGT CGACTCATGCCTCGGAGCAGCACTT for np-2. The primers for np-2 generated two cDNA products corresponding to the previously described a17 and a22 forms of np-2 (42). Both strands of the PCR products were sequenced and the nucleotide sequence compared with published sequences. The np-1 cDNA contained one consistent deviation from published sequences. Nucleotide 2197 contains an A instead of a G, giving rise to an isoleucine residue instead a valine residue. The np-1 cDNA was subcloned into the BamHI site of pcDNA3/neo expression vector to generate the pcDNA3/np-1 expression vector, and the two np-2 cDNAs were subcloned into XbaI site of the PECE expression vector (48) to generate PECE/np-2 expression vectors containing the two splice forms (42).
Cell Lines and cDNAs-Human umbilical vein-derived endothelial cells (HUVEC) were isolated and cultured in M199 medium containing 20% fetal calf serum, as described previously (32). The PAE cells used throughout this report were cultured as described in F-12 medium containing 10% fetal calf serum, glutamine, and antibiotics (24). The PAE cells were kindly provided by Dr. Carl Heldin. The np-1-expressing PAE cell lines were generated by transfecting PAE cells with the pcDNA3/np-1 expression vector followed by selection of stable transfectants using G418 (0.5 mg/ml). The np-2-expressing cell lines were generated by co-transfecting PAE cells with the PECE/np-2(a17) or PECE/np-2(a22) expression vectors and the pBabePuro plasmid (49), followed by selection with 0.5 g/ml puromycin (49). Transfection was carried out using LipofectAMINE according to manufacturer's instructions. Stable cell lines were subsequently cultured without puromycin or G418. MDA-MB-231 breast cancer cells were cultured as described previously (33).
Binding and Cross-linking-VEGF 121 , VEGF 145 , and VEGF 165 were produced using the baculovirus expression system and iodinated as described (9,32,45). The binding of the VEGF isoforms to cells and cross-linking was carried out essentially as described previously (30,32). All the experiments were repeated at least twice. 165 but Not for VEGF 121 -NP-2 was originally characterized as a receptor for semaphorins 3C and 3F, whereas np-1 was originally identified as a semaphorin-3A receptor (39,42). NP-1 was subsequently found to be a receptor for VEGF 165 (34). We hypothesized, therefore, that np-2 may also act as a VEGF receptor. Northern blot analysis revealed np-2 mRNA transcripts in HUVEC (data not shown). RT-PCR led to the isolation of two cDNA species encoding two splice forms of human np-2 from HUVEC-derived mRNA. These correspond to the previously identified a17 and a22 splice forms (42). The full-length cDNAs encoding these np-2 forms were stably expressed in PAE cells. 125 I-VEGF 165 was subsequently bound and cross-linked to the PAE/np-2(a22) cells. The binding was strongly enhanced by heparin, and two 125 I-VEGF 165 /np-2 cross-linked complexes of ϳ140 and ϳ160 kDa could be observed (Fig. 1A, lane 2). The binding was specific since 125 I-VEGF 165 did not bind to parental PAE cells, even though heparin was added during the binding (Fig. 1B,  lane 3). Similar results were obtained when PAE/np-2(a17) cells were used (data not shown). It was previously observed that the related receptor encoded by the np-1 gene binds VEGF 165 but not VEGF 121 (34,36). It seems that np-2 behaves similarly since cells expressing the recombinant np-2 forms were not able to bind 125 I-VEGF 121 (Fig. 1A, lane 4), even though the 125 I-VEGF 121 used bound efficiently to recombinant VEGFR-1 receptors that were expressed in PAE cells (Fig. 1B,  lane 2). The affinities of np-1 and np-2 toward VEGF 165 were compared using Scatchard analysis. The dissociation constant of VEGF 165 from np-2 found to be 1.3 ϫ 10 Ϫ10 M, while the dissociation constant of VEGF 165 from np-1 was 1.8 ϫ 10 Ϫ10 M (data not shown). The dissociation constants are thus very similar.

NP-2 Is a Receptor for VEGF
NP-2 Is a Receptor for PlGF-2-NP-1 was also found to function as a VEGF-B receptor and a receptor for the heparin-binding form of placenta growth factor, PlGF-2 (36,37). We have therefore asked whether PlGF-2 can also bind to np-2. A complex corresponding in size to a PlGF-2/np-2 complex was observed when 125 I-PlGF-2 was bound and cross-linked to PAE cells expressing either the a22 splice form of np-2 (Fig. 2, lane  1) or the a17 form (Fig. 2, lane 3). Similar complexes were absent when 125 I-PlGF-2 was bound and cross-linked to parental PAE cells (Fig. 2, lanes 2 and 4). These experiments indicate that np-2 can also function as a receptor for PlGF-2.
VEGF 145 Binds to np-2 but Not to np-1-np-2 and np-1 bind to different ligands in the central nervous system. Our initial experiments did not reveal any differences between np-1 and np-2 with regard to their interactions with VEGF 165 and PlGF-2. However, when VEGF 145 was bound and subsequently cross-linked to np-1 and to np-2, it was found that VEGF 145 is able to bind to np-2 but not to np-1 (Fig. 3). VEGF 145 is a VEGF splice form that lacks exon 7 and contains instead the peptide encoded by exon 6 of the VEGF gene (9). These results were verified by competition experiments, which have shown clearly that VEGF 145 competes with VEGF 165 for binding to np-2 but not for binding to np-1 (Fig. 3B).
The affinity of VEGF 145 toward np-2 is about 5 fold lower than that of VEGF 165 as is revealed by binding/competition experiments (Fig. 4B). Thus the dissociation constant of VEGF 145 is around 7 ϫ 10 Ϫ10 M as compared with 1.3 ϫ 10 Ϫ10 M for VEGF 165 . However, even 4 g/ml VEGF 145 were not sufficient to inhibit significantly the binding of 125 I-VEGF 165 to np-1, indicating that the affinity of np-1 to VEGF 145 is at least 100-fold lower (Fig. 4A).
Because VEGF 145 differentiates between the two neuropilin types, we have also looked at the expression of np-2 on the surface of several cell types using VEGF 145 binding to differentiate between the two receptors. Our experiments indicate that HUVEC express functional np-2 receptors on their surface since VEGF 145 is able to bind to a receptor with a mass corre-sponding to that of np-2 on their cell surface (Fig. 5, lane 1). HUVEC apparently express both np-1 and np-2 on their cell surface since VEGF 145 inhibits only partially the binding of  2 and 4) in the presence of 1 g/ml heparin. Binding was carried out as described (36), and bound 125 I-PlGF-2 was subsequently cross-linked to the cells. Cross-linked complexes were detected as described in the legend to Fig.  1.   FIG. 3. VEGF 145 binds to np-2 but not to np-1. A, 125 I-VEGF 145 (10 ng/ml) was bound to PAE/np-1 cells (lanes 1 and 2) or to PAE/np-2 cells (lanes 3 and 4) in the presence of 10 g/ml heparin and in the absence (lanes 1 and 3) or presence (lanes 2 and 4) of 1 g/ml unlabeled VEGF 165 . Binding was carried out 2 h at 4°C. Bound 125 I-VEGF 145 was cross-linked to the cells, which were subsequently lysed. Equal amounts of protein were chromatographed on an SDS-PAGE gel, which was subsequently dried and autoradiographed. B, 125 I-VEGF 165 (5 ng/ml) was bound to PAE/np-1 cells (lanes 1-3) or to PAE/np-2 cells (lanes 4 -6) in the presence of 10 g/ml heparin in the absence of competitors (lanes  1 and 4) or in the presence of either 1 g/ml unlabeled VEGF 165 (lanes 2 and 5) or 1 g/ml unlabeled VEGF 145 (lanes 3 and 6). Subsequent cross-linking and autoradiography were carried out as described in the legend to Fig. 1 and under "Experimental Procedures." 125 I-VEGF 165 to neuropilin receptors on the surface of these cells (9). This conclusion is also supported by Northern blot experiments showing that the np-2 mRNA is expressed in HUVEC (data not shown). In contrast, MDA-MB-231 cells express high levels of np-1 and therefore bind 125 I-VEGF 165 efficiently (Fig. 5, lane 6), but do not express detectable levels of functional np-2 receptors since 125 I-VEGF 145 does not bind to these cells (Fig. 5, lane 3). It is thus possible, using 125 I-VEGF 145 binding, to differentiate between functional np-1 and np-2 receptors expressed on the surfaces of cells.

FIG. 2. NP-2 is a receptor for PlGF-2. 125 I-PlGF-2 (15 ng/ml) was bound to PAE cells expressing the a22 form of np-2 (lane 1), to PAE cells expressing the a17 form of np-2 (lane 3) or to PAE cells (lanes
Our experiments suggest that the binding of VEGF 165 and VEGF 145 to np-2 may initiate signal transduction in np-2expressing cells. We have therefore examined the effects of VEGF 165 on the proliferation and migration of PAE/np-2 cells. These initial experiments indicated that VEGF 165 does not induce or inhibit cell proliferation or cell migration in these cells (data not shown). However, it is possible that np-2 may signal in collaboration with other VEGF receptors such as VEGFR-1 or VEGFR-2. These possibilities are currently under investigation. DISCUSSION np-1 had been characterized as a receptor for semaphorin-3A/collapsin-1, while np-2 was initially characterized as a receptor for the related semaphorin-3F. Both receptors play an important role in nerve guidance during embryonic development and repel growing axons in response to the binding of their respective ligands (40,50,51). The identification of receptors that interact with VEGF 165 but not with VEGF 121 in endothelial cells (32) was followed by the identification of tumorigenic cells that express large amounts of such receptors (33). Affinity purification of membrane extracts from MDA-MB-231 cells on VEGF 165 affinity columns identified np-1 as the splice variant specific receptor (34). NP-1 does not bind VEGF 121 , a VEGF form lacking a heparin binding ability, but the heparin-binding VEGF 165 and the heparin-binding form of PlGF (36) as well as VEGF-B (37) bind to this receptor.
We have noticed that the np-2 mRNA is also expressed in HUVEC. We have therefore suspected that np-2 may also function as a VEGF receptor. Our experiments reveal that np-2 does indeed function as a splice variant-specific VEGF receptor. NP-2 binds VEGF 165 and VEGF 145 with high affinity, but is unable to bind VEGF 121 . Furthermore, np-2 is also a receptor to PlGF-2. Thus, except for the VEGF 145 binding ability, which is unique to np-2, these two receptors appear to behave very similarly with regard to their interaction with various VEGF family members. We have also shown here that the selective binding ability of VEGF 145 can be used to distinguish between np-1 and np-2 receptors on various cell types. The neuropilin receptors of MDA-MB-231 cells bind VEGF 165 and PlGF-2 (34,36). By using VEGF 145 binding, we see clearly that these cells have a very low or no functional np-2 receptors at all on their surface, and that their splice variant-specific receptors are almost exclusively np-1 receptors, indicating independently that np-1 is indeed a PlGF-2 receptor.
VEGF 165 , VEGF 145 , and PlGF-2 are all heparin-binding proteins, and their heparin-binding domains are well defined (1,3). The heparin-binding domains of VEGF 165 , VEGF 145 and PlGF-2 contain clusters of highly charged basic amino acids that presumably interact with charged sulfate groups on heparin-like molecules. However, it is difficult to detect other common sequence motifs in the heparin-binding peptides encoded by exon 6 of PlGF and exons 6 and 7 of VEGF. This observation argues for the participation of heparin-like molecules in the mechanism that regulates the binding of these VEGF and PlGF forms to neuropilins. However, this is not the only cue that regulates binding to neuropilins since heparin-binding factors such as bFGF fail to bind to neuropilins (data not shown) and since VEGF 145 , despite its heparin binding ability fails to bind to np-1. Nevertheless, when a heparin-binding factor binds to a neuropilin, heparin potentiates the binding. It is possible that heparin-like molecules may directly assist the binding of heparin-binding VEGF forms to neuropilins by changing the conformation of VEGFs so as to enable their binding to neuropilins. This hypothesis predicts that heparin-binding VEGF forms bind to neuropilins through a domain that is distinct from the heparin-binding site. Indeed, if neuropilins were to bind to the heparin-binding forms of VEGF by interacting directly with their heparin-binding domain, than heparin would have been expected to inhibit the binding of VEGF 165 to neuropilins through steric interference. This hypothesis needs to be tested. However, because iodinated VEGFs had been used, it is also clear that part of the observed effect of heparin on 125 I-VEGF 165 binding is the result of a chaperone-like effect. VEGF 165 loses part of its biological activity following exposure to oxidants such as those used during iodination. The binding of oxidized 125 I-VEGF 165 to heparin or to cell surface heparansulfate glycosaminoglycans such as glypican-1 can restore its receptor binding ability (32,52). Thus, it is logical to assume that at least a part of the observed effect of heparin is due to this effect.
What are the biological roles of np-1 and np-2 in the vascular system? Targeted disruption of the np-1 gene has revealed that np-1 participates in embryonic vasculogenesis and angiogenesis and plays an important role in the maturation of blood vessels (35). However, similar data are not yet available with regard to np-2. Furthermore, this type of data does not reveal much about the mechanism by which np-1 affects blood vessel maturation. It was observed that the binding of VEGF 165 to np-1 does not induce cell migration or cell proliferation by itself (34). It was further reported that np-1 enhances the binding of 125 I-VEGF 165 to the VEGFR-2 receptor and it was suggested, based on these observations, that the np-1-enhanced binding of VEGF 165 to VEGFR-2 potentiates VEGFR-2-mediated cell migration (34). However, recent experiments performed in our laboratory failed to confirm these data.
There is therefore no clear-cut evidence indicating that the binding of VEGF family members to np-1 or to np-2 results in signal transduction in endothelial cells or in any other cell type. However, VEGF 165 competes with semaphorin-3A for binding to the semaphorin-3A binding site of np-1 (40). It is thus possible that the function of VEGF binding is mainly inhibitory. It was shown that the binding of semaphorin-3A to PAE cells expressing np-1 inhibits their spontaneous migration. Such inhibition was also observed when primary endothelial cells were used (53). It was also shown that semaphorin-3A inhibits the migration of VEGF 165 -stimulated PAE cells expressing both VEGFR-2 and np-1 (53). However, these results do not necessarily mean that the inhibition was the outcome of competition between semaphorin-3A and VEGF 165 for the common binding site on np-1. It is equally reasonable to assume that VEGF 165 stimulated migration by activation of VEGFR-2 (1) and that semaphorin-3A, by interacting with np-1 and possibly with an associated plexin (43), produces an independent inhibitory effect which overcomes the VEGFR-2-mediated stimulation of migration. Thus, these experiments (53) leave open the question of whether the binding of VEGF family members to neuropilins results in any direct biological effects.
The intracellular domain of np-1 is apparently not required for the np-1-mediated activity of axon repulsion, indicating that neuropilins form signaling complexes with other cell surface proteins (54). Furthermore, it was recently shown that np-1 and np-2 form signaling complexes with plexins (43,44). It is possible that neuropilins may also be able to associate with the tyrosine kinase receptors of VEGF, VEGFR-1 or VEGFR-2 (1), to form signaling complexes responsive to VEGF binding. This possibility is currently under investigation. These observations do not mean that the intracellular domains of neuropilins are superfluous. The discovery of NIP, a PDZ domain containing intracellular protein that binds to the tail end of np-1 (55), and the discovery of np-2 splice forms that differ in their intracellular domains (42) indicate that the intracellular domains of neuropilins are likely to be important for their full function.
To conclude, we have shown here that np-2 behaves like a splice variant VEGF receptor. It behaves like np-1 in that it is able to bind PlGF-2 and VEGF 165 but not VEGF 121 . However, np-2 binds VEGF 145 , a heparin-binding VEGF form that does not bind to np-1. This last result indicates that a heparin binding ability may be required, but not sufficient, for the np-1 binding ability of VEGF splice forms.