A Three-Kilobase Fragment of the Human Robo4 Promoter Directs Cell Type–Specific Expression in Endothelium
Robo4, a member of the roundabout family, is expressed exclusively in endothelial cells and has been implicated in endothelial cell migration and angiogenesis. Here we report the cloning and characterization of the human Robo4 promoter. The 3-kb 5′-flanking region directs endothelial cell–specific expression in vitro. Deletion and mutation analyses revealed the functional importance of two 12-bp palindromic DNA sequences at −2528 and −2941, 2 SP1 consensus motifs at −42 and −153, and an ETS consensus motif at −119. In electrophoretic mobility shift assays using supershifting antibodies, the SP1 motifs bound SP1 protein, whereas the ETS site bound a heterodimeric member of the ETS family, GA binding protein (GABP). These DNA–protein interactions were confirmed by chromatin immunoprecipitation assays. Transfection of primary human endothelial cells with small interfering RNA against GABP and SP1 resulted in a significant (≈50%) reduction in endogenous Robo4 mRNA expression. The 3-kb Robo4 promoter was coupled to LacZ, and the resulting cassette was introduced into the Hprt locus of mice by homologous recombination. Reporter gene activity was observed in the vasculature of adult organs (particularly in microvessels), tumor xenografts, and embryos, where it colocalized with the endothelial cell–specific marker CD31. LacZ mRNA levels in adult tissues and tumors correlated with mRNA levels for endogenous Robo4, CD31, and vascular endothelial cadherin. Moreover, the pattern of reporter gene expression was similar to that observed in mice in which LacZ was knocked into the endogenous Robo4 locus. Together, these data suggest that 3-kb upstream promoter of human Robo4 contains information for cell type–specific expression in the intact endothelium.
There is increasing evidence that migration and patterning of axons and blood vessels share similar guidance mechanisms. Among the guidance systems involved in axonal and vascular networks are the ephrin-Eph, netrin-unc5b, semaphorin-plexin, and slit-Robo molecules.1
Robo (roundabout) is a member of the neural cell adhesion molecule family. Robo was originally isolated from Drosophila melanogaster.2 The ligand for Robo, Slit, was first identified in Drosophila as an extracellular molecule involved in axonal branching and neural migration.3 In vertebrates, 3 Robo receptor family members (Robo1 to -3) and 3 Slit ligands (Slit1 to -3) have been implicated in guiding axon growth via repulsive signaling.2 A fourth Robo receptor family member, Robo4, was cloned and shown to be restricted in its tissue distribution to endothelial cells.4,5 Robo4 is expressed in areas of in vivo angiogenesis. For example, the receptor is present in the endothelial lining of blood vessels in the developing embryo,6 placenta,4 and tumors.4,7 Robo4 has also been detected in the endothelium of normal nonangiogenic tissues, including the heart and lung.6,7 Recent studies support a role for Robo4 in endothelial cell migration, proliferation, and angiogenesis.6,8–10 The goal of the present study was to dissect the mechanisms of cell type–specific expression of the Robo4 promoter.
Materials and Methods
Human umbilical vein endothelial cells (HUVECs), human coronary artery endothelial cells (HCAECs), human pulmonary artery endothelial cells (HPAECs), human microvascular endothelial cells (HMVECs), and human coronary artery smooth muscle cells (HCASmCs) were purchased from Cambrex (Walkersville, Md). Primary endothelial cells were cultured in endothelial growth medium-2-MV. HCASmCs were cultured in SmBM-2 medium. The human embryonic kidney cell line (HEK293) and HepG2, U937, LL/2, and B16-F1 cells were grown in DMEM supplemented with 10% FCS (Hyclone, Logan, Utah). Plasmid constructions are detailed in the online data supplement at http://circres.ahajournals.org. Transient transfection of primary endothelial cells and HEK293 cells was performed using FuGENE 6 reagent (Roche Molecular Biochemicals, Mannheim, Germany). Nuclear extracts were prepared from HCAECs using Nuclear Extract Kit (Active Motif, Carlsbad, Calif) according to the instructions of the manufacturer. Chromatin immunoprecipitation (ChIP) assays were performed using a Chromatin Immunoprecipitation Assay Kit (Upstate, Lake Placid, NY) according to the instructions of the manufacturer. Immunoprecipitated genomic DNA fragments were quantified by real-time PCR using the primers to amplify the Robo4 proximal promoter region. Hprt-targeted mice were generated as described previously.11 The generation of the Robo4-lacZ knock-in mouse is detailed in the online data supplement. Expression level of Robo4, CD31, vascular endothelial (VE)-cadherin, and LacZ from mouse tissues and or cultured cells was measured by real-time PCR. All animal experiments were performed according to protocols approved by the Institutional Committee for Use and Care of Laboratory Animals.
An expanded Materials and Methods section is available in the online data supplement.
Cloning and Sequence Analysis of the Human Robo4 Gene
Under in vitro conditions, endogenous Robo4 was expressed at high levels in primary human endothelial cells, but not HCASmCs or HEK293 cells (please see Figure I in the online data supplement). 5′ Rapid amplification of cDNA ends revealed 2 common transcriptional start sites (supplemental Figure II). The most frequent transcriptional start site is designated as base pair number +1. The sequence of the upstream promoter region of the human Robo4 gene was determined (Figure 1). The promoter lacks a TATA box. The upstream region includes unique long direct-repeat sequences and 12-bp and 26-bp palindromic sequences (Figure 1A). The sequence between −285 and +40 is 80% conserved between mouse and human and 76% conserved between rat and human (Figure 1B). Among the consensus binding sites that are conserved among all 3 species in the 300-bp proximal promoter are ETS and 2 SP1 motifs (discussed below).
Functional Analysis of the Human Robo4 Promoter in Cultured Cells
To compare the activity of the Robo4 promoter with that of other endothelial cell–specific promoters, HUVECs and/or HCAECs were transiently transfected with pGL3 containing the upstream promoter regions of intercellular adhesion molecule-2 (0.37 kb), E-selectin (3 kb), P-selectin (3 kb), ephrinB2 (2.8 kb), Flt-1 (1.4 kb), Robo4 (3 kb) (termed pGL3-Robo4), Tie1 (0.8 kb), or Tie2 (0.72 kb). The Robo4 promoter demonstrated higher activity compared with promoters for intercellular adhesion molecule-2, P-selectin, E-selectin, ephrinB2, Flt-1, and Tie2 and similar activity to the Tie1 promoter (Figure 2A shows HCAECs). To determine whether the Robo4 promoter contains information for endothelial-specific expression in vitro, transient transfection assays were also performed in HEK293, HepG2 and U937 cells. As shown in Figure 2B, Robo4 promoter activity was significantly higher in HCAECs compared with HEK293 cells (5-fold), HepG2 cells (29-fold), and U937 cells (18-fold). Together, these findings suggest that the 3-kb upstream promoter of human Robo4 directs high-level cell type–restricted expression in vitro.
To delineate the functional elements within the upstream promoter region, a series of deletion and mutant promoter fragments were fused to the luciferase reporter gene in pGL3, and the resulting constructs were transiently transfected into HCAECs. Sequential 5′ deletions resulted in stepwise reduction of promoter activity, with Del1 (−2450) demonstrating 55%, Del2 (−1635) demonstrating 44%, Del3 (−1173) demonstrating 32%, Del4 (−930) demonstrating 25%, and Del5 (−329) demonstrating 15% of wild-type level (Figure 3⇓A). Deletion of sequences between −329 and −228 (Del6) resulted in a slight increase in activity (29%), suggesting the presence of a repressor in that region.
To further delineate the enhancing region between −2450 and −3000, additional 5′ deletions were generated and assayed for luciferase activity. As shown in Figure 3⇑B, Del1–1 (−2867), Del1–2 (−2745), Del1–3 (−2644), and Del1–4 (−2550) demonstrated activity comparable to the full length 3-kb promoter. The Del1–5 promoter, containing a 2515 bp 5′-flanking sequence, resulted in a significant (70%) decrease in activity, whereas further deletions had no such effect. These data suggest that a 35-bp DNA region between −2550 and −2515 possesses enhancing activity. This region (which we term the Robo4 enhancer element 1 [REn1]) contains consensus binding sites for nuclear factor-κB, nuclear factor of activated T cells (NF-AT), SP1, glucocorticoid receptor, and activator protein 2γ. However, single mutations of these motifs in the context of the 3-kb promoter failed to alter luciferase activity (Figure 3⇑C). To further address the role of this region in mediating expression of Robo4, the REn1 was removed from the full-length promoter, inverted, or replaced with 2 different heterologous sequences. As shown in Figure 3⇑D, none of these manipulations strongly affected promoter activity.
Based on these findings, we inferred that additional upstream DNA sequences between −3000 and −2551 must compensate for promoter activity in the absence of REn1. To test this hypothesis, 6 new internal–deletion constructs were generated in which the REn1 and progressive lengths of 5′ sequence were removed from the full-length promoter. As shown in Figure 3⇑E, there was little change in promoter activity with deletions of 100, 200, 300, and 400 bp. However, deletions of 435 bp (Del 435bp) or 465 bp (Del 465bp) resulted in 80% and 70% reduction in promoter activity, respectively, similar to that observed with Del1–5 (the same construct used in Figure 3⇑B). These data support the existence of a second enhancer between −2950 and −2916 (which we term REn2) (Figure 3⇑F).
To confirm the dual role for REn1 and REn2 in mediating Robo4 promoter activity, we generated new deletion constructs. Single deletions of REn1 or REn2 resulted in a 15% and 36% reduction in promoter activity, respectively, whereas a double deletion of REn1 and REn2 resulted in 42% decrease in activity (Figure 3⇑G). Interestingly, REn1 and REn2 contain 12-bp palindromic DNA sequences (5′CAGAGCCCAGA in REn1; 5′TCTGGGCTCTG in REn2) (Figure 3⇑F). To determine whether these sequences were responsible for the enhancing activity of REn1 and REn2, we deleted the two 12-bp elements from the full-length promoter (Figure 3⇑G). The resulting construct demonstrated a 56% reduction in promoter activity, similar to that observed with Del1–5. Taken together, these data support a role for the 12-bp palindromic sequences in mediating Robo4 promoter activity.
We next focused on the proximal region of the human Robo4 promoter because the sequence of the immediate upstream 300-bp region is highly conserved between species, and because the upstream 600-bp region contains putative binding motifs for ETS, SP1, TAL1, E2A, GATA, and nuclear factor-κB (Figure 3⇑H). A candidate ETS binding site located at −119 is identical to the known consensus sequence for ETS-1 and ETS-2. This element was designated ETS(1) to distinguish it from other ETS motifs in the promoter. The above sites were mutated alone or in combination (5 Mut and SP1[1,2]) in the context of the 3-kb Robo4 promoter, and the resulting mutants were assayed for activity in transient transfection assays. As shown in Figure 3⇑H, a single mutation of the −119 ETS(1), −153 SP1(1), or −42 SP1(2) site resulted in a 90%, 40%, or 50% reduction in promoter activity, respectively, whereas mutations of the other sites had no significant effect. A double mutation of the SP1 sites (SP1[1,2]) resulted in a 65% reduction in promoter activity. These findings suggest that the ETS(1), SP1(1), and SP1(2) sites are critical determinants of Robo4 promoter activity.
SP1 Binds to the Human Robo4 Promoter and Induces Promoter Activity
To investigate whether SP1 binds to the SP1(1) and SP1(2) sites, electrophoretic mobility shift assay (EMSA) was performed. Incubation of a radiolabeled probe spanning the −153 SP1 site with nuclear extract from HCAECs resulted in a strong DNA–protein complex (Figure 4A, lane 2). The complex was inhibited by addition of cold wild-type SP1 competitor, but not a mutant SP1 competitor (Figure 4A, lanes 3 to 6). Preincubation with anti-SP1 antibody resulted in a partial supershift of the specific DNA–protein complex, whereas control antibody had no such effect (Figure 4A, lanes 7 and 8). (The partial nature of the supershift may be explained by limiting amounts of antibody or the existence of a second DNA–protein complex that lacks SP1.) Finally, incubation of radiolabeled probe with in vitro–translated SP1 resulted in a DNA–protein complex of similar size to that obtained with nuclear extracts (Figure 4A, compare lanes 10 and 2). The latter complex was supershifted with anti-SP1 antibody (Figure 4A, lane 11). The same result was obtained in EMSA using a probe spanning the −42 SP1 site (Figure 4B). To determine whether SP1 transactivates the Robo4 promoter through these 2 SP1 sites, cotransfection assays were performed in HEK293 cells using an SP1 expression vector. Overexpression of SP1 resulted in significant (8-fold) induction of Robo4 promoter activity (Figure 4C). A single mutation of SP1(1) or SP1(2) reduced the promoter activity to 6.5- or 5.5-fold, respectively. A double mutation of the −42 and −153 SP1 motifs (SP1[1,2]) led to a further reduction in activity (4-fold). Taken together, these data suggest that SP1 regulates the Robo4 promoter activity through both the −42 and −153 SP1 sites.
GABP Binds to the Human Robo4 Promoter and Induces Promoter Activity
To identify the factor that binds to the −119 ETS(1) motif, EMSA was performed as described above using a probe that contains this site. A strong DNA–protein complex was detected (Figure 4D, lane 2) and was inhibited by addition of wild-type, but not mutant cold ETS(1) competitor (Figure 4D, lanes 3 and 4). Supershift assays were performed with antibodies to ETS factors that have been previously implicated in endothelial cell gene regulation, including ETS-1, ETS-2, ELF-1, FLI-1, ERG, NERF, and PEA3. None of these antibodies resulted in a supershift or inhibited the specific DNA–protein complex (Figure 4D, lane 5 shows ETS-1). As a positive control for ETS-1 binding and supershifting activity of the ETS-1 antibody, a radiolabeled probe spanning the consensus ETS-1 binding motif was incubated with recombinant ETS-1. As shown in Figure 4E, ETS-1 protein bound to the classical ETS motif (but not ETS from the Robo4 promoter), and the resulting DNA–protein complex was supershifted by ETS-1 antibody.
Based on the above results, we explored the potential role of another ETS factor that has not been previously described in endothelial cells, namely GA binding protein (GABP). GABP binds as a complex consisting of heterodimers of GABPα and GABPβ or γ. GABPγ is an alternative splice form of GABPβ (see review12). In supershift assays, preincubation with antibodies against GABPα or GABPβ/γ resulted in complete inhibition of the DNA–protein complex (Figure 4F, lanes 3 and 5). In contrast, anti-GABPβ antibody had minimal effect on DNA binding, and anti-SP1 antibody resulted in loss of a more slowly migrating DNA–protein complex (Figure 4F, lanes 4 and 6). In vitro translated protein consisting of GABPα, GABPα/β, or GABPα/γ resulted in specific DNA–protein complexes compatible with those observed with nuclear extracts (Figure 4G).
In cotransfection assays, overexpression of GABPα, alone or together with GABPβ or GABPγ, resulted in significant induction of Robo4 promoter activity, whereas GABPβ or GABPγ alone had no such effect (Figure 4H). Mutation of the ETS(1) site significantly attenuated GABP-mediated transactivation of the promoter (Figure 4H). Together, these data suggest that GABP plays an important role in mediating Robo4 expression.
SP1 and GABP Bind to the Endogenous Human Robo4 Promoter in Primary Endothelial Cells
To investigate whether SP1 and GABP bind to the Robo4 proximal region in endothelial cells, ChIP assay was performed. Formalin-fixed genomic DNA–protein complexes from HCAECs and HCASmCs were sheared by sonication (Figure 5A). Resulting small DNA–protein complexes were immunoprecipitated with antibodies to SP1, GABPα, or Egr-1 (control IgG), and the resulting products were used as template in a PCR reaction containing primers specific for the immediate upstream promoter of Robo4. Real-time PCR was used to calculate binding intensities. As shown in Figure 5B, SP1, and GABPα, but not Egr-1 (control IgG), bound to the proximal promoter region. Similar results were obtained with primary vascular smooth muscle cells (Figure 5B). These results demonstrate that SP1 and GABP bind to the Robo4 proximal region in vivo and that this interaction is not specific to endothelial cells.
Small Interfering RNA–Mediated Knockdown of GABPα and SP1 Results in Significant Reduction of Endogenous Robo4 mRNA Expression
To determine whether SP1 and GABPα play a role in mediating the endogenous expression of Robo4, small interfering RNA (siRNA) against these transcription factors were transfected into HCAECs. As shown in Figure 5C, siRNA against SP1 resulted in a 5.3-fold reduction in SP1 and a 1.9-fold reduction of Robo4 expression, whereas siRNA against GABPα resulted in 4.9-fold reduction in GABPα and 2.2-fold decrease in Robo4.
The Three-Kilobase Human Robo4 Promoter Contains Information for Endothelial-Specific Expression in Mice
To determine whether the Robo4 promoter directs lineage-specific expression in vivo, the 3-kb promoter region was coupled to LacZ. A single copy of the transgenic cassette was targeted to the Hprt locus of mice using homologous recombination as previously described. High percentage chimeric males were bred to wild-type females. Resulting female agouti offspring were bred to generate stable lines and F2 males were assayed for LacZ expression.
Whole-mount staining of organs (brain, heart, lung, skeletal muscle, aorta, trachea, diaphragm, and esophagus) revealed widespread, although not uniform, β-galactosidase activity in the vasculature (Figure 6). Tissue sections revealed LacZ staining in the endothelial lining of vessels in all organs examined (Figure 7B). Expression was greater in the microvessels compared with macrovessels. In the kidney, β-galactosidase activity was highest in the glomeruli. In serial sections, LacZ colocalized with CD31. LacZ was not observed in any other cell type or lineage including peripheral blood cells and bone marrow (supplemental Figure III).
We compared expression of LacZ mRNA with that of endogenous Robo4 and 2 endothelial markers, CD31 and VE-cadherin, using real-time PCR of adult mouse tissues. As shown in Figure 7C, Robo4 transcripts were detected in all organs, according to the following rank order: lung>heart>kidney>skeletal muscle>liver>spleen=brain. Importantly, LacZ mRNA expression in Hprt-targeted mice followed a similar pattern. To control for vascular density, tissue samples were also assayed for CD31 and VE-cadherin. The pattern of LacZ and Robo4 expression mirrored that of VE-cadherin and CD31 (Figure 7C). (In the case of CD31, expression was relatively higher in spleen, presumably owing to positivity in hematopoietic cells.) These data suggest that expression of the transgene mimics that of the endogenous Robo4 gene, which in turn correlates with the degree of vascularization. LacZ expression was also detected in the endothelium of embryos and tumor xenografts (supplemental Figures IV and V).
As a second strategy for comparing the expression of the Hprt-targeted transgene and the endogenous gene, we knocked LacZ into the endogenous Robo4 locus (Figure 8A). Heterozygous F2 adult males demonstrated endothelial cell–specific expression of LacZ in the vasculature (Figure 8B and 8C). One exception was the brain, where mounts and tissue sections revealed a weak nonvascular distribution in the pia mater (data not shown). Compared with the Hprt locus–targeted mice, the LacZ knock-in animals demonstrated lower β-galactosidase activity in the vasculature. Moreover, reporter gene expression in the aorta was more restricted to branch orifices and tributaries and was undetectable in the large arteries of the brain. Otherwise, the pattern of expression was similar between the two lines of mice, with predominant staining in the microvascular endothelium.
We have cloned and characterized the human Robo4 promoter. Similar to the endogenous Robo4 gene, a 3-kb fragment of the upstream promoter directed high-level lineage-specific expression in cultured endothelial cells. 5′-Deletion analyses revealed several positive regulatory regions. Using a series of mutational analyses, EMSA, ChIP, and siRNA-mediated knockdown experiments, we have demonstrated an important role for SP1 and GABP in governing basal expression of Robo4. In addition, we have identified 12-bp palindromic DNA sequences that are important for Robo4 promoter activity in endothelial cells. This DNA sequence does not conform to established cis-regulatory motifs and thus represents a potentially novel regulatory element. The mechanism by which the palindromic sequences mediate Robo4 expression is the focus of ongoing studies.
Members of the ETS family of transcription factors share an evolutionarily conserved DNA-binding domain of 85 aa with a winged-helix-turn-helix configuration.13 ETS factors bind to GGAA/T core sequences. Consensus ETS binding motifs have been identified within the promoters of several other endothelial cell genes, including Flt-1, Tie1, Tie2, and VE-cadherin.14–18 The functional relevance of ETS motifs in mediating endothelial cell gene expression has been demonstrated both in vitro and in vivo. Several ETS factors have been shown to mediate gene expression in endothelial cells, most notably ETS-1, ETS-2, ESE-1, NERF2, and ELF-1.19
GABP (also known as nuclear respiratory factor [NRF]-2 and adenovirus E4 transcription factor [E4TF]-1) is unique among the ETS family of transcription factors in that it forms multimers, consisting of 2 structurally unrelated subunits: GABPα and GABPβ. GABPα contains the ETS DNA-binding region, whereas GABPβ is required for nuclear translocation and transactivation. GABPβ stabilizes the GABPα–DNA interaction more than 100-fold.20 Mammalian GABP is ubiquitously expressed in all tissues and has been implicated in several critical cellular processes including cellular respiration in mitochondria, differentiation, cell cycle, cell survival, and neuromuscular function.12 Mice that are null for GABPα are embryonic lethal and die before implantation.21 In addition to controlling the expression of housekeeping genes, GABP has been shown to regulate the expression of cell type–specific genes in several distinct lineages, including myeloid cells, lymphocytes, neuromuscular cells, hepatocytes, and mast cells.12 To our knowledge, this is the first study to demonstrate a role for GABP in promoting the expression of an endothelial cell–specific target gene.
Our data are consistent with the notion that SP1 and GABP are necessary for full basal expression of Robo4 in endothelial cells. They do not prove that these transcription factors mediate cell type–specific gene expression. Indeed, SP1 and GABP are expressed in other cell types, and ChIP assays in vascular smooth muscle cells revealed binding of both transcription factors to the Robo4 promoter in an otherwise nonexpressing cell type. Thus, other mechanisms must be responsible for cell type–specific gene expression. One possibility is that GABP interacts with cell type–specific transcription factors or coactivators to promote cell-specific responses.
Previous studies using standard transgenic mouse assays or Hprt locus targeting have demonstrated that the majority of endothelial-specific promoters direct expression to specific vascular beds (reviewed previously22). In the current study, the Hprt-targeted 3-kb Robo4 promoter directed expression in the embryonic and adult vasculature. Reporter gene expression was restricted to the endothelium. Expression was more prominent in microvessels compared with macrovessels. Even within microvessels, LacZ expression was nonuniform. In real-time PCR analyses, LacZ mRNA expression correlated with expression of endogenous Robo4. Moreover, the expression pattern was similar to that observed when the LacZ reporter gene was knocked into the Robo4 locus. Because the knock-in strategy involved deletion of Robo4 DNA sequences, including potential regulatory elements in the first 3 introns, it is formally possible that the LacZ expression does not precisely reflect the endogenous of the Robo4 gene. That caveat notwithstanding, our data suggest that the 3-kb human Robo4 promoter contains information for near-authentic expression in the endothelium.
Based on its exquisite cell type specificity, and its expression in the neovasculature, the Robo4 gene (and promoter) represents a powerful tool for dissecting the molecular basis of lineage determination and new blood vessel growth.
We thank Dr Takahiro Sato and Masako Akiyama for helpful suggestions. We also thank Dan Li for technical help with real-time PCR measurement.
Sources of Funding
This work was supported by NIH grant HL076540.
Original received January 14, 2006; resubmission received June 23, 2006; second resubmission received March 13, 2007; revised second resubmission received April 26, 2007; accepted May 1, 2007.
Huminiecki L, Bicknell R. In silico cloning of novel endothelial-specific genes. Genome Res. 2000; 10: 1796–1806.
Suchting S, Heal P, Tahtis K, Stewart LM, Bicknell R. Soluble Robo4 receptor inhibits in vivo angiogenesis and endothelial cell migration. FASEB J. 2005; 19: 121–123.
Bedell VM, Yeo SY, Park KW, Chung J, Seth P, Shivalingappa V, Zhao J, Obara T, Sukhatme VP, Drummond IA, Li DY, Ramchandran R. roundabout4 is essential for angiogenesis in vivo. Proc Natl Acad Sci U S A. 2005; 102: 6373–6378.
Minami T, Donovan DJ, Tsai JC, Rosenberg RD, Aird WC. Differential regulation of the von Willebrand factor and Flt-1 promoters in the endothelium of hypoxanthine phosphoribosyltransferase- targeted mice. Blood. 2002; 100: 4019–4025.
Dube A, Akbarali Y, Sato TN, Libermann TA, Oettgen P. Role of the Ets transcription factors in the regulation of the vascular- specific Tie2 gene. Circ Res. 1999; 84: 1177–1185.
Dube A, Thai S, Gaspar J, Rudders S, Libermann TA, Iruela-Arispe L, Oettgen P. Elf-1 is a transcriptional regulator of the Tie2 gene during vascular development. Circ Res. 2001; 88: 237–244.
Iljin K, Dube A, Kontusaari S, Korhonen J, Lahtinen I, Oettgen P, Alitalo K. Role of ets factors in the activity and endothelial cell specificity of the mouse Tie gene promoter. FASEB J. 1999; 13: 377–386.
Minami T, Kuivenhoven JA, Evans V, Kodama T, Rosenberg RD, Aird WC. Ets motifs are necessary for endothelial cell-specific expression of a 723-bp Tie-2 promoter/enhancer in Hprt targeted transgenic mice. Arterioscler Thromb Vasc Biol. 2003; 23: 2041–2047.
Gory S, Dalmon J, Prandini MH, Kortulewski T, de Launoit Y, Huber P. Requirement of a GT box (Sp1 site) and two Ets binding sites for vascular endothelial cadherin gene transcription. J Biol Chem. 1998; 273: 6750–6755.
Rudders S, Gaspar J, Madore R, Voland C, Grall F, Patel A, Pellacani A, Perrella MA, Libermann TA, Oettgen P. ESE-1 is a novel transcriptional mediator of inflammation that interacts with NF-kappa B to regulate the inducible nitric-oxide synthase gene. J Biol Chem. 2001; 276: 3302–3309.
LaMarco K, Thompson CC, Byers BP, Walton EM, McKnight SL. Identification of Ets- and notch-related subunits in GA binding protein. Science. 1991; 253: 789–792.
Ristevski S, O’Leary DA, Thornell AP, Owen MJ, Kola I, Hertzog PJ. The ETS transcription factor GABPalpha is essential for early embryogenesis. Mol Cell Biol. 2004; 24: 5844–5849.