Vascular Endothelial Growth Factor Receptor-2

The vascular endothelial growth factor receptor-2 (VEGFR-2/KDR/flk-1) functions as the primary mediator of vascular endothelial growth factor activation in endothelial cells. Regulation of VEGFR-2 expression appears critical in mitogenesis, differentiation, and angiogenesis. Transcriptional regulation of the VEGFR-2 is complex and may involve multiple putative upstream regulatory elements including E boxes. Transcript initiation is dependent on an initiator (Inr) element flanking the transcriptional start site. The transcription factor, TFII-I, enhances VEGFR-2 transcription in an Inr-dependent fashion. TFII-I is unusual both structurally and functionally. The TFII-I transcription factor family members contain multiple putative DNA binding domains. Functionally, TFII-I acts at both the basal, Inr element as well as at several distinct upstream regulatory sites. It has been postulated that the structure of TFII-I might allow simultaneous interaction with both basal and regulatory sites in a given promoter. As TFII-I is known to act at regulatory sites including E boxes as well as at the basal Inr element, we evaluated the possibility of Inr-independent TFII-I activation of the VEGFR-2 promoter. We found that an Inr-mutated VEGFR-2 reporter construct retains TFII-I-stimulated activity. We demonstrated that TFII-I binds to both the Inr and to three regulatory E boxes in the human VEGFR-2 promoter. In addition, reduction in TFII-I expression by siRNA results in decreased VEGFR-2 expression. We also describe counter-regulation of the VEGFR-2 promoter by TFII-IRD1. We found that TFII-I is capable of acting at both basal and regulatory sites in one promoter and that the human VEGFR-2 promoter is functionally counter-regulated by TFII-I and TFII-IRD1.

The vascular endothelial growth factor receptor-2 (VEGFR-2/KDR/flk-1) functions as the primary mediator of vascular endothelial growth factor activation in endothelial cells. Regulation of VEGFR-2 expression appears critical in mitogenesis, differentiation, and angiogenesis. Transcriptional regulation of the VEGFR-2 is complex and may involve multiple putative upstream regulatory elements including E boxes. Transcript initiation is dependent on an initiator (Inr) element flanking the transcriptional start site. The transcription factor, TFII-I, enhances VEGFR-2 transcription in an Inr-dependent fashion. TFII-I is unusual both structurally and functionally. The TFII-I transcription factor family members contain multiple putative DNA binding domains. Functionally, TFII-I acts at both the basal, Inr element as well as at several distinct upstream regulatory sites. It has been postulated that the structure of TFII-I might allow simultaneous interaction with both basal and regulatory sites in a given promoter. As TFII-I is known to act at regulatory sites including E boxes as well as at the basal Inr element, we evaluated the possibility of Inr-independent TFII-I activation of the VEGFR-2 promoter. We found that an Inr-mutated VEGFR-2 reporter construct retains TFII-I-stimulated activity. We demonstrated that TFII-I binds to both the Inr and to three regulatory E boxes in the human VEGFR-2 promoter. In addition, reduction in TFII-I expression by siRNA results in decreased VEGFR-2 expression. We also describe counter-regulation of the VEGFR-2 promoter by TFII-IRD1. We found that TFII-I is capable of acting at both basal and regulatory sites in one promoter and that the human VEGFR-2 promoter is functionally counter-regulated by TFII-I and TFII-IRD1.
Elaboration of the complex mammalian vasculature requires precise spatial and temporal signaling. Central to this process is the vascular endothelial growth factor receptor-2 (VEGFR-2/flk-1/KDR) 1 (1). This transmembrane receptor tyrosine ki-nase transduces signals from extracellular vascular endothelial growth factor (VEGF). Other receptors for VEGF have been identified, but their roles in vasculogenesis are less well defined (2). The VEGFR-2 is an early embryonic marker of endothelial cells with expression detected in mouse embryo as early as day 7 after conception (3). The critical role of VEGFR-2 in development was dramatically demonstrated through targeted disruption of the gene in mice (3). Homozygous deficient embryos died in utero between days 8.5 and 9.5 after conception due to severe abnormalities in development of the hematopoietic and endothelial systems. In addition to its role in development, VEGFR-2 has been shown to be involved in multiple pathologic conditions including tumor-induced angiogenesis (4,5) and increased vascular permeability following ischemic insult (6).
Common to its role in both normal development and pathologic states is the process of transcriptional regulation of the VEGFR-2. The human and mouse VEGFR-2 5Ј-promoters have been identified (7). A number of regulatory sites have been identified including a GATA binding site (8), Sp1 and Sp3 binding sites (7,9), as well as a NF-B site (10) shown to be involved in basal transcriptional activity. Deletional analysis of the 4-kb upstream region of human VEGFR-2 demonstrated that full promoter activity was dependent on the region between Ϫ225 and the transcriptional start site (7). Further, 5Ј-deletion to Ϫ164 resulted in almost 50% loss of activity. This deletion resulted in the loss of an E box at Ϫ170 (Fig. 1B). The minimal promoter was further defined by 3Ј-deletions that identified the region up to ϩ268 as necessary for maximal promoter activity (7). The region from Ϫ225 to ϩ268 includes a number of putative DNA binding motifs including three E boxes at Ϫ170, ϩ70, and ϩ184 (Fig. 1B).
Further, Wu and Patterson (11) identified a functional initiator (Inr) element in the VEGFR-2 gene that is transactivated by TFII-I. The Inr element is a poorly conserved 7-base pair sequence overlapping the transcription start site that can function in TATA-less genes, such as VEGFR-2, to direct transcription from a single start site (12). In addition, the Inr can enhance transcription in TATA-containing genes (12). These studies showed loss of transcriptional activity in reporter systems with disruption of the Inr sequence. Thus, TFII-I has been implicated in regulating the expression of the VEGFR-2 gene.
TFII-I, a ubiquitously expressed, multifunctional transcription factor, was originally described based on its in vitro binding to the core promoter element, Inr, as well as to the upstream E box element (13). TFII-I was independently identified as Bruton's tyrosine kinase-associated protein, BAP135, a substrate of Bruton's tyrosine kinase (14,15). Alternative splicing of two exons in TFII-I results in production of four isoforms (16) that can form dimers with each other, with a related transcription factor, TFII-IRD1 (17), and with several other helix-loophelix transcription factors (18). Dimerization, DNA binding, and transactivation domains of TFII-I have been defined (19). The protein sequence includes six repeated motifs (so-called "I" repeats) that are putative helix-loop-helix (HLH) domains ( Fig.  1A) (15). The amino-terminal sequence preceding the first repeat contains a modified leucine zipper critical for dimerization (19).
TFII-I binds to and activates the promoters of several genes. These include xanthine dehydrogenase (20), mullerian-inhibiting substance (21), and mouse ribonuclease reductase R1 promoter (22), as well as the Rous sarcoma virus long terminal repeat (23). In addition, three TFII-I binding sites were identified in the c-fos promoter (24). TFII-I appears to be unusual in having both basal transcriptional activity (via the Inr) and activity at upstream sequences such as the E box, serumresponse element, c-sis platelet-derived growth factor-inducible element (S1E) (18). Indeed, the structure of TFII-I with six loosely conserved repeats that resemble HLH domains has led to speculation that TFII-I acts as an adapter capable of binding with multiple partners (18). The structure of TFII-I might make it particularly suited to a role in development by allowing it to link basal activity to various dynamically changing environmental signals.
TFII-IRD1 (BEN/WBSCR11/MusTRD1/Cream1) protein shares similarity with TFII-I in the 95-amino-acid HLH-like I-repeat domain (hence, RD1) (25). The I domain is significantly larger than in other described basic HLH proteins (26). The divergence of this motif from the HLH consensus has suggested that the I domain defines a distinct family of transcription factors (26). The gene for TFII-IRD1 is located close to that of TFII-I on human chromosome 7q11.23 in the Williams syndrome critical region (25,27), and the two proteins may have arisen as a result of partial gene duplication (27). This family also includes the recently described TFII-IRD2 (28). The similarities in protein structure among the known family members have been detailed (26). TFII-IRD1 has been shown to associate with retinoblastoma protein (29). TFII-IRD1 has been shown to act as both a positive and a negative regulator of gene function depending on both promoter (30) and context (26).
In addition to their structural similarities, there is growing evidence that TFII-I and TFII-IRD1 may functionally interact. The ability of TFII-IRD1 to interact physically with TFII-I was demonstrated by co-immunoprecipitation experiments (17), contrasting with earlier reports that these structurally similar proteins could not heterodimerize (19,31). In the IgH promoter, both TFII-I and TFII-IRD1 bind to the downstream immunoglobulin control element, although only TFII-IRD1 was shown to functionally regulate this promoter element (17). Further, in COS 7 cells, TFII-IRD1 was reported to inhibit translocation of TFII-I to the nucleus (32). Thus, TFII-I and TFII-IRD1 may function in a counter-regulatory fashion.
We hypothesized that TFII-I might act both in the basal regulation of VEGFR-2 expression as well as at upstream regulatory sites. We further hypothesized that TFII-I and TFII-I RD1 might act as counter-regulators of the VEGFR-2 promoter. Here we demonstrated that TFII-I activation of the VEGFR-2 occurs by interaction with both the Inr and the multiple E box elements. We showed that depletion of TFII-I in endothelial cells results in decreased VEGFR-2 expression. We also demonstrated that TFII-IRD1 is a negative regulator of the VEGFR-2 promoter. This negative regulation of the VEGFR-2 may, in part, be mediated by histone deacetylase (HDAC) activity.
Cell Culture-Bovine pulmonary artery endothelial (BPAE) cells were purchased from Cell Systems Corp. Cells were maintained in tissue culture medium containing 80% Dulbecco's modified Eagle's medium, low glucose (Invitrogen), 20% heat-inactivated fetal bovine serum (HyClone), 1% minimal Eagle's medium non-essential amino acid solution 100ϫ (Sigma), and 0.1% of initial mix endothelial cell growth factor (reconstituted with Dulbecco's modified Eagle's medium) (Upstate Biotechnology). Cells were maintained at 37°C in 5% CO 2 . All cells were used between passage numbers 5 and 8.
Transfection-For luciferase assays, BPAE were seeded in 24-well dishes such that the cells were 50 -80% confluent at the time of transfection. Transfection was performed using FuGENE 6 transfection reagent (Roche Diagnostics) as recommended by the manufacturer. For 24-well plates, 0.6 l of FuGENE 6 reagent/well and 200 ng of DNA/well were optimal. For most experiments, transfection efficiency was determined by FACS of cells transfected in parallel with GFP. Cells were assayed for luciferase activity at 24 -48 h after transfection. Lysates were retained for Western analysis.
For siRNA and FACS, HPAE were seeded in 6-cm dishes such that Amino acid numbering is indicated. B, schematic of minimal human VEGFR-2 promoter. Portion of human VEGFR-2 promoter from Ϫ225 toϩ268 relative to transcriptional start site (arrow) is sufficient to confer maximal activity in reporter assays. This region includes three E boxes (blue) and an Inr (green) flanking the transcriptional start site. C, consensus sequences for Inr (green) and E box (blue). Sequences from hVEGFR-2 are indicated in black. Mutations of each site are shown below in red.
they were 50 -80% confluent at the time of transfection. Transfection was performed using FuGENE 6 transfection reagent (Roche Applied Science) as recommended by the manufacturer. SiRNAs were co-transfected with 1 g of pCIS2 as carrier DNA. Cells were assayed by FACS 24 -48 h after transfection.
For experiments with trichostatin A (TSA), BPAE were transfected as above in 24-well plates. Six hours after transfection, TSA diluted in Me 2 SO or an equivalent amount of Me 2 SO was added to wells. Luciferase activity was assayed 24 h after the addition of TSA.
Luciferase Assays-Human VEGFR-2 (KDR)-pGL2 luciferase reporter constructs (4 kb, Ϫ225 to ϩ268 and Ϫ225 to ϩ268 Inr mutant) were a kind gift of Dr. Cam Patterson and are as described previously (11). Cells were transfected as described above and assayed for luciferase activity using Dual-Luciferase reporter assay (Promega). Each transfection was assayed for firefly luciferase (test reporter) and Renilla luciferase (pRL-CMV) to normalize for transfection efficiency. Some assays were performed with ␤-galactosidase (pCSK-LacZ) to normalize for transfection efficiency. Luciferase assays were performed as recommended by the manufacturer following lysis of cells with passive lysis buffer. ␤-Galactosidase activity was assayed following lysis as above using ␤-galactosidase reporter gene assay (Roche Applied Science) as recommended by the manufacturer. Luciferase and ␤-galactosidase activity were measured using a Tropix microplate luminometer (Applied Biosystems). All data are expressed as relative luciferase activity (firefly/Renilla or firefly/␤-galactosidase, as indicated) and represent the mean of three individual transfections in a given experiment. Each experiment was performed at least three times. Data shown are representative of the replicate experiments. Data are expressed as mean and standard deviation.
Electrophoretic Mobility Shift Assay-Oligonucleotides representing the Inr and E box domains of the human VEGFR-2 promoter (Table I) as well as the reverse complement of each oligonucleotide were synthesized. Complementary oligonucleotide pairs were hybridized by heating to 95°C for 5 min and cooling slowly to room temperature. Probes were end-labeled using T4-polynucleotide kinase. His-tagged hTFII-I was affinity-purified on a nickel column. Purified TFII-I was incubated with labeled oligonucleotide probe in buffer containing 20 mM Tris (pH 7.4), 100 mM NaCl, 1 mM EDTA, 10% glycerol, 5 mM dithiothreitol, 1 mg/ml bovine serum albumin, 0.7 mM phenylmethylsulfonyl fluoride, and 50 or 70 g/ml poly dG-dC. Unlabeled competitors were added as indicated. Sequences of the site-mutated competitors are indicated in Table I. After 20 (Inr, E (Ϫ170)) or 60 (E (ϩ70), E (ϩ184)) min at room temperature, the reaction was resolved by electrophoresis on a 4% non-denaturing acrylamide gel. After drying, the gel was placed on a Phospho-rImager screen for visualization. Images were obtained after 12-48 h.
Staining for FACS-Cells were harvested using 0.5 mM EDTA and counted. For intracellular TFII-I staining, cells were fixed in 1% paraformaldehyde for 10 min at room temperature in the dark. After washing with phosphate-buffered saline, cells were resuspended in permeabilization buffer (0.5% saponin in phosphate-buffered saline) with anti-TFII-I antibody or normal rabbit serum as control. After 30 min at room temperature in the dark, cells were washed, and Cy5-antirabbit antibody (Jackson ImmunoResearch) was added at 1:100 as recommended by the manufacturer. After 30 min at room temperature in the dark, cells were washed and resuspended in phosphate-buffered saline with 1% bovine serum albumin. Staining for VEGFR-2 was performed with monoclonal anti-human VEGFR-2 (KDR)-phycoerythrin antibody (R&D Systems) exactly as recommended by the manufacturer.
Transfected Cell Selection-HPAE were transfected essentially as described above. In addition to the plasmids indicated, pMACS K k .II (Miltenyi Biotech) plasmid was included. After harvest, cells were incubated with MACS Select K k MicroBeads as recommended by the manufacturer. Transfected cells were enriched by magnetic separation.
Immunofluorescence-BPAE were plated on sterile glass cover slips and maintained in culture as above. When cells were 40 -60% confluent, transfection with FuGENE 6 was performed as above. Cells were transfected with either GFP and TFII-IRD1 or GFP and pCIS. After 36 h in culture, cells were fixed and immunostained for TFII-I as described previously (33). Cells were imaged using a Zeiss Axiovert S100 TV microscope and Metamorph 4.5.6 software (Universal Imaging). Confocal images were obtained using a Zeiss LSM10 confocal microscope and LSM10 software (Zeiss).

RESULTS
The VEGFR-2 promoter has been previously demonstrated to have a functional Inr that is transactivated by TFII-I (11). The promoter contains multiple putative TFII-I binding sites in addition to the Inr (7). We hypothesized that the structure of TFII-I with multiple DNA binding domains might support a more complex interaction with the VEGFR-2 promoter DNA. To validate the use of the previously described (7) minimal promoter (human VEGFR-2 Ϫ225 to ϩ268), we determined the TFII-I responsiveness of the complete promoter (4 kb) relative to that of the minimal promoter ( Fig. 2A). We found the absolute activity of the minimal promoter to be greater than the complete promoter, suggesting that the minimal promoter lacks an inhibitory site. We determined that the fold activation by TFII-I was similar in the minimal and complete reporter constructs. In comparison, the promoter-less pGL2 Basic reporter showed no TFII-I response. We thus used the minimal promoter construct as the basis for further mutagenesis studies. The effect of TFII-I isoform was also tested. Transfection with TFII-I, isoform 1 (includes both alternative exons), and TFII-I, isoform 4 (includes neither alternative exon) showed no significant difference in luciferase activity (data not shown).
Mutagenesis of the VEGFR-2 Inr has been shown to result in decreased promoter activity (11). We noted that although the absolute activity was greatly diminished (Fig. 2B), the promoter retains TFII-I responsiveness (Table II). Because the Inr sequence overlaps the transcriptional start site, we hypothesized that the decreased activity was related to disruption of the conserved A(ϩ1)at the transcriptional start site. The TFII-I-enhanced activity suggested that TFII-I might be acting at a second putative binding site in the VEGFR-2 promoter.
To begin to address this possibility, we performed PCRmediated mutagenesis of the E box sequences in the minimal promoter. Mutations (Fig. 1C) were based on prior studies of critical E box residues (15). The resulting promoter constructs were assayed for TFII-I-stimulated activity (Fig. 2C). The constructs showed varying absolute activities, but all displayed a relative TFII-I activation similar to that of the intact minimal promoter and the Inr mutant promoter (Fig. 1C).
TFII-I DNA Binding to VEGFR-2 Promoter Inr and E Box Motifs-Luciferase reporter assays suggested that TFII-I may interact functionally at sites in the VEGFR-2 in addition to the Inr. We performed DNA binding studies using labeled probes spanning the Inr (Fig. 3A) and three E box motifs (Fig. 3, B-D) in the human VEGFR-2 promoter. Purified His-tagged TFII-I bound specifically to labeled probes spanning the Inr as well as 5Ј-GCTCCCTAGCCCTGTGCGCTCAACTGTCCTGCGCTGCGGGGTGCCG-3Ј CCGTAT E boxes (Ϫ170), (ϩ70), and (ϩ184). This binding was decreased by the addition of an anti-His antibody to the binding reactions (Fig. 3D). The binding to E box (ϩ70) could only be detected on longer exposures and appears to be lower affinity. The binding to the Inr could be competed by cold E (ϩ184) but not by cold Inr mutant probe ( Fig. 3A and data not shown). Similarly, binding to E box (Ϫ170) and (ϩ184) was competable by unlabeled Inr probe but not by cold E box (Ϫ170) mutant or cold E box (ϩ184) mutant ( Fig. 3B and data not shown). Binding to E box (ϩ184) was more efficiently competed by cold E box (ϩ184) than by unlabeled Inr probe (Fig. 3D). Binding to E box (ϩ70) was competed similarly by unlabeled E box (ϩ70), E box (ϩ184), and Inr (see Fig. 7C).
Counter-regulatory Effect of TFII-IRD1-TFII-IRD1 acts as an antagonist of TFII-I at the c-fos promoter (32). We sought to characterize this effect at the VEGFR-2 promoter. Co-transfection with reporter and TFII-IRD1 resulted in the dramatic reduction of TFII-I-induced luciferase activity (Fig. 4A). In addition, TFII-IRD1 decreased basal VEGFR-2 reporter activity, suggesting that the effect also altered native bovine TFII-I activity. Western blots of cell lysates do not indicate a significant effect of TFII-IRD1 on TFII-I protein expression (Fig. 4A). TFII-IRD1 appears to block TFII-I enhancement of luciferase activity by a mechanism other than a one-to-one interaction based on the near maximal inhibition of TFII-I activation by significantly lower molar quantities of transfected TFII-IRD1.
We sought to further define the mechanism by which TFII-IRD1 alters TFII-I effect on the VEGFR-2. In COS 7 cells, overexpressed TFII-IRD1 was reported to decrease the translocation of overexpressed TFII-I to the nucleus. We addressed this possible mechanism by transfection of BPAE with either TFII-IRD1 or an equivalent molar quantity of the empty pCIS2 expression vector. Cells were co-transfected with GFP as a marker of transfection. Cells were visualized after staining for TFII-I with a rhodamine labeled anti-rabbit secondary antibody following incubation with a rabbit polyclonal anti-TFII-I antibody. We compared the intensity of rhodamine labeling in TFII-IRD1-transfected and pCIS2-transfected cells (Fig. 4B).  We noted decreased nuclear TFII-I staining but did not detect increased cytoplasmic staining in these cells expressing only native TFII-I. This observation was confirmed by confocal imaging of these cells (Fig. 4C). This finding could be consistent with TFII-IRD1 decreasing nuclear TFII-I through a mechanism other than blocking nuclear translocation. Alternatively, the lack of apparent increase in cytoplasmic TFII-I may be a function of the distribution of the TFII-I over a large cytoplasmic volume. Regardless of mechanism, these data supported the interpretation that the inhibitory effect of TFII-IRD1 is exerted at least in part through decreasing TFII-I expression in the nucleus. A second mechanism for TFII-IRD1 antagonism might be through the recruitment of HDACs to the promoter. TFII-I has been shown to co-purify with the HDAC3 complex (34,35). In overexpression studies, TFII-IRD1 can also interact with HDAC3 (35). To investigate the possible role of HDAC recruitment in the TFII-IRD1-mediated inhibition of the VEGFR-2, we made use of the potent and general HDAC inhibitor, TSA. BPAE were transfected with VEGFR-2 luciferase reporter construct and pCIS, TFII-I, or TFII-IRD1. Six hours after transfection, cells were treated with TSA ranging from 0.1 to 20 nM. Luciferase assays were performed after 36 h with TSA (Fig.  4D). Treatment with TSA at 20 nM nearly completely reversed antagonism of VEGFR-2 minimal promoter by TFII-IRD1. These data lent support to the hypothesis that TFII-IRD1 acts, in part, via recruitment of HDACs.
VEGFR-2 Protein Expression Is Decreased by TFII-I siRNA and TFII-IRD1-Regulation of VEGFR-2 protein expression is critical in normal vasculogenesis as well as in pathological angiogenesis. We have demonstrated that TFII-I can up-regulate the VEGFR-2 promoter in reporter assays. To determine whether VEGFR-2 protein expression is altered in response to changes in the amount of TFII-I, we made use of two human TFII-I siRNA constructs. Human pulmonary artery endothelial cells were transfected either with an active TFII-I siRNA construct or with a control siRNA. Expression of TFII-I was determined by FACS using rabbit anti-TFII-I antibody and Cy5labeled anti-rabbit secondary in fixed and permeabilized cells (Fig. 5A). Two independent siRNAs both resulted in a significant decrease in TFII-I staining. Cells transfected in parallel experiments were incubated with phycoerythrin-antiVEGFR-2 antibody (Fig. 5B). Cells transfected with active TFII-I siRNA showed significantly decreased VEGFR-2 cell surface expression in comparison with cells transfected with control siRNA. The effect on VEGFR-2 expression is also noted in immunoblots of cells sorted for a marker of transfection (H2K k ) (Fig. 6A) as described under "Experimental Procedures." When normalized to actin level, siRNA 1 reduces TFII-I and VEGFR-2 immunoreactivity to about 60% of the control siRNA (Fig. 6B). SiRNA 2 is more potent at decreasing TFII-I level as well as VEGFR-2 level (Fig. 6B). The effect of the human TFII-I siRNAs on VEGFR-2 could be rescued by co-transfection with bovine TFII-I (data not shown).
We have demonstrated that TFII-IRD1 can antagonize TFII-I activation of the VEGFR-2 promoter. To determine whether this results in altered VEGFR-2 protein levels, we transfected BPAE with TFII-IRD1 or the empty pCIS vector. Twenty-four hours after transfection, cell lysates were separated by SDS-PAGE and immunoblotted for VEGFR-2, TFII-IRD1, TFII-I, and actin as indicated (Fig. 7). Decreased VEGFR-2 immunoreactivity is noted in the lysates transfected with TFII-IRD1. Of note, TFII-I levels do not appear to be altered.

DISCUSSION
The dynamic regulation of VEGFR-2 expression is critical in normal vascular development. The transcription factor, TFII-I, is known to play a role in this regulation by acting at the basal transcriptional Inr site (11). TFII-I, however, has an unusual structure with multiple potential DNA binding domains. In other promoters, it has been shown to bind at both basal and upstream regulatory sites as well as to interact with a number of other HLH transcription factors. These characteristics have led to the hypothesis that TFII-I may function in a novel fashion to bridge basal and upstream regulatory sites and to integrate signals from multiple pathways (18). This function would make TFII-I particularly capable of the type of dynamic regulation of VEGFR-2 that occurs during vascular development.
Earlier experiments (11) demonstrated a role for TFII-I in regulating of the human VEGFR-2 promoter via a poorly conserved Inr motif flanking the transcriptional start site. In this study, we have examined the possibility that TFII-I is capable of regulating the VEGFR-2 promoter by interaction with other sites. Specifically, we have demonstrated that a luciferase re-FIG. 3. Electrophoretic mobility shift assay of purified TFII-I with Inr and E box oligonucleotide probes. Purified TFII-I was incubated with end-labeled Inr (A), E (Ϫ170) (B), E (ϩ70) (C), or E (ϩ184) (D) oligonucleotide with or without unlabeled competitor (as indicated). Probes for Inr, E (Ϫ170) E (ϩ70), and E (ϩ184) show a band of restricted mobility (arrow) that is competed by an identical cold oligonucleotide competitor at 5-(ϩ) or 10-(ϩϩ) fold excess. Binding to the Inr is competed to a lesser degree by cold E (ϩ184) competitor. Binding to E (Ϫ170), E (ϩ70), and E (ϩ184) are competed by cold Inr. The major band of binding to E (ϩ184) is decreased by the addition of anti-His antibody (anti-His Ab) to the binding reaction. The residual binding to E (ϩ184) may reflect binding to TFII-I, which has lost the amino-terminal His. Both E (ϩ70) and E (ϩ184) also show a slower migrating band, which under some binding conditions is more prominent. This may reflect a higher order complex formation, but this complex is not affected by the addition of anti-His.
porter driven by an Inr-mutated VEGFR-2 promoter construct retains TFII-I-dependent enhancement of expression. The reduction in absolute activity seen for the Inr mutant promoter may result from the disruption of the conserved A at the transcriptional start site. The VEGFR-2 Inr mutation would be predicted based on studies of mutagenesis of the terminal deoxynucleotidyltransferase Inr to virtually abolish Inr-mediated activity (12). We also showed that TFII-I binding to an Inr probe cannot be competed by a cold Inr mutant competitor, suggesting that the mutation disrupts binding. We nevertheless observed a consistent TFII-I-dependent increase in luciferase activity with the Inr mutant. This supported the idea that TFII-I interacts functionally with other sites in the VEGFR-2 promoter.
TFII-I has been demonstrated to interact with the E box in the adenovirus major late promoter (13,15). The VEGFR-2 promoter contains three such motifs (7). We introduced mutations in each E box motif based on published mutagenesis studies identifying a critical core sequence (15). The E box mutants also showed a wide range in absolute luciferase activity, but each showed similar enhancement with TFII-I. This was suggestive that TFII-I might interact with multiple sites in the promoter. We were further able to demonstrate TFII-I DNA binding oligonucleotide probes spanning E box (Ϫ170), (ϩ70), and (ϩ184). As might be expected, a cold Inr competitor diminished TFII-I binding to the E box probes. Based on competition by cold E box and Inr oligonucleotides, we would expect that the affinity of the Inr site for TFII-I is greater. We did not directly address whether a single TFII-I protein interacts with multiple sites in the VEGFR-2 promoter simultaneously, but our data did imply that loss of a single binding site is not sufficient to eliminate the functional interaction.
Since regulation of VEGFR-2 protein is critical to its function, we examined cell surface expression in HPAE transfected with TFII-I siRNA. These siRNAs produced dramatic downregulation of TFII-I in HPAE based on FACS analysis of fixed and permeabilized cells. HPAE transfected with either of the two TFII-I siRNAs showed significant decreases in VEGFR-2 expression. This effect could be rescued by bovine TFII-I, which differs in sequence at the target site for the siRNA. We were not able to co-stain cells for VEGFR-2 and TFII-I because VEGFR-2 staining was lost after the cells were fixed.
The ability to modulate VEGFR-2 in both a positive and a negative direction is clearly critical in its regulation. The transcription factor, TFII-IRD1, is expressed in the vascular compartment over a brief period in development (36). It is undetectable in vascular tissue at day 8.5 after conception. It is briefly up-regulated and then declines by day 13.5 after conception (36). This timing corresponds with a period of active vasculogenesis. The possibility that TFII-I and TFII-IRD1 are counter-regulatory had been raised by experiments with the c-fos promoter (32). Our data suggested that TFII-IRD1 does not inhibit TFII-I activation of the VEGFR-2 promoter with a one-to-one stoichiometry. This lack of one-to-one correspondence is demonstrated by the near maximal inhibition of TFII-I activation by 10-fold lower molar lower quantities of TFII-IRD1. Inhibition based on a single molecule of TFII-I binding to a molecule of TFII-IRD1 would be expected to result in a more linear inhibition of TFII-I activation by increasing amounts of TFII-IRD1. Western blot confirmed expression of both TFII-I and TFII-IRD1 in the expected ratios based on the amount of DNA transfected. Thus, the inhibition did not appear to be due to TFII-IRD1 altering TFII-I protein expression. We investigated several other potential mechanisms for the antagonism. TFII-IRD-1 was shown to block translocation of TFII-I to the nucleus in COS 7 cells overexpressing both factors (32). The proposed mechanism was competition between TFII-IRD1 and TFII-I for a dedicated nucleo-cytoplasmic shuttle. It is of note that in some cell types, there is a large fraction of TFII-I protein present in the cytoplasm, including in neurons in which TFII-I is expressed in the dendritic tree of Purkinje cells (33). By contrast, BPAE show little cytoplasmic TFII-I staining at baseline. In comparing cells transfected with TFII-IRD1 with those transfected with empty vector, we did observe an apparent decrease of nuclear staining but with no apparent increase in cytoplasmic staining. These results could be consistent with TFII-IRD1 blockade of translocation; the lower levels of endogenous TFII-I as compared with levels in TFII-I-transfected cells might make detection of an increase in the cytoplasm difficult. On the other hand, overexpression of TFII-I in COS 7 cells might have resulted in altered distribution from the native state.
Both TFII-I and TFII-IRD1 have been described to interact with HDAC3 (34,35). As TFII-IRD1 has been demonstrated to repress a number of promoters (30,32), this seemed a plausible mechanism for the inhibition of TFII-I activation of the VEGFR-2 promoter. In fact, the general HDAC inhibitor, TSA, reversed the TFII-IRD1-mediated inhibition of the VEGFR-2 driven reporters. Thus, the effect of TFII-IRD1 might be indirect and might involve binding to a distinct site in the VEGFR-2 promoter. Recent studies, however, have demonstrated co-immunoprecipitation of TFII-I and TFII-IRD1 (17).
The process of vasculogenesis requires the integration of multiple signals to modulate the patterns of expression of VEGFR-2. This study did not exclude the possibility of other FIG. 7. TFII-IRD1 expression decreases VEGFR-2 protein levels. BPAE were transfected with TFII-IRD1 plasmid or empty vector. Cells were lysed, separated on SDS-PAGE, and immunoblotted with the antibodies indicated. Cell transfected with TFII-IRD1 demonstrated reduced VEGFR-2 immunoreactivity in comparison with cells transfected with empty vector.

FIG. 5. TFII-I and VEGFR-2 expression in HPAE transfected
with TFII-I siRNA or control siRNA. HPAE were transfected with human TFII-I siRNAs (hTFII-1 siRNA 1 (2020) or hTFII-2 siRNA 1 (3390)) or control siRNA (equal length, no homology in human). Upper, detection of TFII-I in cells transfected with siRNA. Cells were fixed, permeabilized, and stained with anti-TFII-I antibody and Cy5-antirabbit secondary. Cells were excited at 633 nm, and emission was detected with a 660 nm band-pass filter (allophycocyanin). Lower, detection of VEGFR-2 in cells transfected with TFII-I siRNA. Cells were incubated with phosphatidylethanolamine (PE)-VEGFR-2. Cells were excited at 488 nm, and emission was detected with a 575 nm band-pass filter. Traces from three separate experiments were overlaid using FACS Calibur software.
FIG. 6. VEGFR-2 and TFII-I protein are decreased by TFII-I siRNA. A, HPAE were transfected with human TFII-I siRNAs (hTFII-1 siRNA 1 (2020) or hTFII-2 siRNA 1 (3390)) or control (Ctrl) siRNA (equal length, no homology in human) as above. In addition, cells were transfected with H2K k plasmid as described under "Experimental Procedures." Transfected cells were enriched by adhesion to MACSelect MicroBeads and magnetic separation as recommended by the manufacturer. SDS lysates from the selected cells were separated on SDS-PAGE and immunoblotted with the antibody indicated. The band of VEGFR-2 immunoreactivity is broader than in lysates from cells not selected using MACSelect Microbeads. This is likely secondary to an effect on the entry of this large protein into the gel. B, the immunoblot was quantified using NIH Image J software (Gel Analyzer Plugin) and normalized to actin as a loading control. VEGFR-2 and TFII-I levels for siRNA 1 and 2 are shown as a fraction of the expression of the respective protein with control siRNA. SiRNA 2 is more potent at the inhibition of both TFII-I and VEGFR-2 expression than siRNA 1. Data are representative of three independent experiments. factors interacting with the VEGFR-2 promoter directly. It provided evidence, however, that at the VEGFR-2 promoter, TFII-I may subserve a more complex function than simply as a basal transcription factor at the Inr. This is a paradigm for which there is emerging evidence in a number of promoter systems (18).
The interaction of TFII-I with the VEGFR-2 promoter might be modeled in several ways. In one model, a TFII-I homodimer interacts simultaneously with multiple sites in the VEGFR-2 promoter. The affinity of each of the sites for TFII-I differs. In the second model, different molecules of TFII-I (heterodimers) interact with unique sites in the VEGFR-2 promoter. The differences in interaction might result from the known splicing variations of TFII-I. A third model reflects the known ability of TFII-I to interact with other HLH factors. Interaction with different factors might alter the affinity of the TFII-I complex for specific binding sites in the VEGFR-2 promoter. All or some of these models might operate simultaneously within a cell depending on the intracellular milieu. This complexity of interaction affords TFII-I a unique capacity to modulate VEGFR-2 expression and to integrate multiple signaling pathways. The counter-regulatory effect of TFII-IRD1 provides a potent brake to the activation pathway. We speculate that disruption of TFII-I expression in endothelial cells during development will result in significant pathology and might explain the failure to identify individuals with Williams Syndrome with homozygous deletion or mutation in TFII-I.