Endothelial NO Synthase-Dependent S-Nitrosylation of β-Catenin Prevents Its Association with TCF4 and Inhibits Proliferation of Endothelial Cells Stimulated by Wnt3a

ABSTRACT

Nitric oxide (NO) produced by endothelial NO synthase (eNOS) modulates many functions in endothelial cells. S-nitrosylation (SNO) of cysteine residues on β-catenin by eNOS-derived NO has been shown to influence intercellular contacts between endothelial cells. However, the implication of SNO in the regulation of β-catenin transcriptional activity is ill defined. Here, we report that NO inhibits the transcriptional activity of β-catenin and endothelial cell proliferation induced by activation of Wnt/β-catenin signaling. Interestingly, induction by Wnt3a of β-catenin target genes, such as the axin2 gene, is repressed in an eNOS-dependent manner by vascular endothelial growth factor (VEGF). We identified Cys466 of β-catenin as a target for SNO by eNOS-derived NO and as the critical residue for the repressive effects of NO on β-catenin transcriptional activity. Furthermore, we observed that Cys466 of β-catenin, located at the binding interface of the β-catenin–TCF4 transcriptional complex, is essential for disruption of this complex by NO. Importantly, Cys466 of β-catenin is necessary for the inhibitory effects of NO on Wnt3a-stimulated proliferation of endothelial cells. Thus, our data define the mechanism responsible for the repressive effects of NO on the transcriptional activity of β-catenin and link eNOS-derived NO to the modulation by VEGF of Wnt/β-catenin-induced endothelial cell proliferation.

INTRODUCTION

The Wnt/β-catenin signaling pathway is involved in physiological and pathological angiogenesis, the process of blood vessel formation from preexisting vasculature, through its transcriptional regulation (1–3). Canonical Wnt proteins, namely, Wnt1, Wnt3a, Wnt7a, Wnt7b, and Wnt10b, have been demonstrated to activate β-catenin signaling in endothelial cells (ECs) and to regulate target genes that affect the angiogenic process (3–6). In addition to its essential structural role in cadherin-based adhesions, β-catenin is a regulator of Wnt-mediated gene expression through its association in the nucleus with the T-cell factor/lymphoid-enhancing factor (TCF/LEF) (7–9). In the absence of Wnt stimulation, free cytoplasmic β-catenin levels are kept low by a degradation complex that includes glycogen synthase kinase 3β (GSK3β), adenomatous polyposis coli (APC), and axin (10–12). Upon Wnt stimulation, β-catenin is stabilized in the cytoplasm and subsequently translocalized to the nucleus, interacts with TCF/LEF, and enhances the expression of genes involved in cell proliferation, differentiation, cell fate, and survival (6, 13–15).

In ECs, growth factors such as vascular endothelial growth factor (VEGF) regulate β-catenin activity to promote angiogenesis. It is well established that VEGF must induce signaling at adherens junctions of ECs in order to induce blood vessel formation (16, 17). VEGF activates the phosphatidylinositol 3-kinase (PI3K)/Akt/endothelial nitric oxide synthase (eNOS) signaling pathway and stimulates nitric oxide (NO) production from ECs, which is essential for VEGF-regulated angiogenesis (18–20). NO carries out many biological functions in ECs by binding to its receptor, guanylate cyclase, and inducing cyclic GMP (cGMP) (21, 22). In addition, NO is capable of regulating protein function and cellular signal transduction by modifying thiol groups on cysteine residues by means of S-nitrosylation (SNO) (23–25). We previously showed that VEGF-stimulated NO production in ECs induces SNO of β-catenin at Cys619 and promotes its dissociation from the adherens junction protein VE-cadherin. This results in the opening of intercellular contacts between ECs and in increased permeability to macromolecules in response to VEGF (26). We also identified other Cys residues on β-catenin that could be targets for SNO, namely, Cys300, -381, -439, -466, and -520 (26). Furthermore, Cys213 and -520 were identified as potential SNO sites in large-scale proteomics assays (27, 28). In addition to its effects at cell-cell junctions, it has been shown that NO can inhibit the transcriptional activity of β-catenin in the nucleus and reduce proliferation of cancer cells (29, 30). These studies used NO-donating aspirin or other NO-releasing chemical agents to inhibit the transcriptional activity of β-catenin, reduce proliferation, and induce apoptosis of transformed cells (30, 31). However, it is still unclear whether SNO of β-catenin is responsible for both effects of NO, at cell junctions and in the nucleus, and what the mechanisms responsible for these effects are.

Here, we demonstrate that eNOS-derived NO attenuates Wnt/β-catenin signaling in ECs by decreasing the transcriptional activity of β-catenin. This inhibitory effect of NO on β-catenin activity is caused by SNO of β-catenin on Cys466. We show that SNO of the Cys466 residue of β-catenin causes its dissociation from the transcriptional factor TCF4, which results in a reduction of Wnt3a-stimulated transcription of target genes. In addition, we show that SNO of β-catenin on Cys466 inhibits proliferation of ECs stimulated by Wnt3a and that VEGF-stimulated eNOS activation and NO production inhibit Wnt3a signaling, possibly forming an inhibitory feedback loop on Wnt/β-catenin signaling. These findings define SNO as a modulator of β-catenin transcriptional activity, which is important for Wnt/β-catenin signaling and in the control of EC proliferation.

RESULTS

eNOS-derived NO inhibits the transcriptional activity of β-catenin.

To determine if eNOS-derived NO inhibits the transcriptional activity of β-catenin, we expressed in COS-7 cells increasing amounts of myc–β-catenin and a constitutively active mutant of eNOS, S1179D-eNOS. Then, we examined the transcriptional activity of β-catenin using TOPFlash, a luciferase reporter construct that contains TCF/LEF binding sites upstream of luciferase (32). We found that expression of S1179D-eNOS significantly attenuated TOPFlash activity induced by β-catenin (Fig. 1A). To confirm that the inhibitory effect of eNOS on β-catenin is due to its enzymatic activity and NO production, we compared the effect of inactive eNOS, S1179A-eNOS, on the transcriptional activity of β-catenin with that of S1179D-eNOS. In contrast to S1179D-eNOS, expression of S1179A-eNOS did not reduce the activation of the β-catenin reporter, TOPFlash (Fig. 1B). We validated that S1179D-eNOS expression produced significant amounts of NO compared to S1179A-eNOS when cotransfected with β-catenin (see Fig. S1 in the supplemental material). In addition, treatment with the NOS inhibitor N^G-monomethyl l-arginine (l-NMMA) prevented the inhibitory effect of S1179D-eNOS on β-catenin transcriptional activity, and treatment with the soluble guanylate cyclase inhibitor ODQ (1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one) failed to interfere with the capability of eNOS to inhibit β-catenin transcriptional activity (Fig. 1C). Finally, the constitutively activated chimeric VP16-LEF1 protein, a fusion protein between the activation domain of VP16 and LEF1 (33) known to activate TOPFlash independently of β-catenin transactivation properties, was insensitive to S1179D-eNOS expression (see Fig. S2A in the supplemental material). Collectively, these results suggest that eNOS-derived NO directly inhibits the transcriptional activity of β-catenin and that this is done independently of cGMP generation.

FIG 1.

eNOS-derived NO inhibits the transcriptional activity of β-catenin. (A) β-Catenin luciferase reporter assay of COS-7 cells expressing TOPFlash and transfected with the indicated amounts of myc-tagged β-catenin and S1179D-eNOS. The FOPFlash luciferase reporter served as a negative control (n = 4). The transfection levels of myc-tagged β-catenin and eNOS were monitored by immunoblotting (IB), and annexin II was used as a loading control. (B) β-Catenin luciferase reporter assay of COS-7 cells expressing TOPFlash or FOPFlash and transfected with myc-tagged β-catenin and active (S1179D) or inactive (S1179A) eNOS (n = 3). (C) β-Catenin luciferase reporter assay of HEK293T cells stably expressing the TOPFlash reporter and transfected as indicated with myc-tagged β-catenin and/or S1179D-eNOS. The cells were treated with the NOS inhibitor l-NMMA (0.1 mM) or with the soluble guanylate cyclase inhibitor ODQ (10 μM) for 8 h where indicated (n = 3). The data are represented as means and SEM. *, P < 0.05.

NO inhibits Wnt/β-catenin signaling in endothelial cells.

Next, we investigated whether NO affects β-catenin transcriptional activity in ECs. We found that the NO donor S-nitrosoglutathione (GSNO) inhibited TOPFlash activity induced by β-catenin overexpression (Fig. 2A). Similarly, GSNO treatment of bovine aortic endothelial cells (BAECs) inhibited Wnt3a-induced expression of the axin2 gene, a known β-catenin target gene (Fig. 2B). Moreover, we examined whether NO could inhibit EC proliferation stimulated by Wnt/β-catenin signaling. BAECs were exposed to Wnt3a in the absence or presence of GSNO. Bromodeoxyuridine (BrdU) incorporation assays revealed that treatment of cells with GSNO inhibited EC proliferation induced by Wnt3a (Fig. 2C). In addition, GSNO also inhibited proliferation of BAECs that was stimulated by transient expression of myc-tagged β-catenin (Fig. 2D; see Fig. S2B in the supplemental material). GSNO can produce effects trough transnitrosylation reactions; thus, we confirmed the antiproliferative effects of the NO donor on ECs using a bona fide NO-releasing agent, NOC-18 [2,2′-(hydroxynitrosohydrazino)bis-ethanamine] (Fig. 2C). Finally, similar to the β-catenin transcriptional activity assays (see Fig. S2A in the supplemental material), proliferation of BAECs induced by expression of the constitutively active chimeric VP16-LEF1 protein was not affected by GSNO treatment, arguing against a nonspecific toxic effect of GSNO (see Fig. S2B in the supplemental material). Taken together, these results show that NO inhibits EC proliferation promoted by activation of Wnt/β-catenin signaling and that this can be mediated through decreased transcriptional activity of β-catenin by NO.

FIG 2.

NO inhibits β-catenin transcriptional activity and EC proliferation induced by Wnt3a. (A) β-Catenin luciferase reporter assay of BAECs expressing myc-tagged β-catenin in the presence or absence of GSNO (0.1 mM; 18 h; n = 3). The transfection levels of myc-tagged β-catenin were monitored by IB, and β-actin was used as a loading control. (B) qRT-PCR analysis of axin2 mRNA levels in BAECs treated with Wnt3a-conditioned medium and in the presence or absence of GSNO (0.1 mM; 24 h) (n = 4). (C) BrdU incorporation assay in BAECs treated with Wnt3a and in the presence or absence of GSNO (0.1 mM) or NOC-18 (25 μM) for 18 h. (Left) Representative immunofluorescence images of BrdU incorporation in BAECs. Cell nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole). (Right) The percentage of BrdU-positive cells for each treatment was normalized to that of nontreated cells (n = 3). (D) BrdU incorporation assay in BAECs expressing myc-tagged β-catenin in the presence or absence of GSNO (0.1 mM; 18 h; n = 3). The data are represented as means and SEM. *, P < 0.05.

VEGF-stimulated NO production inhibits Wnt3a-mediated activation of β-catenin.

Since eNOS-dependent NO production is central for the effects of VEGF in ECs, we investigated whether eNOS activation by VEGF affects Wnt/β-catenin signaling. BAECs were transfected with small interfering RNA (siRNA) against eNOS or with control (CT) siRNA. First, in CT-siRNA-transfected BAECs, treatment with Wnt3a, and to a lesser extent with VEGF, increased mRNA levels of the β-catenin axin2 target gene (Fig. 3A). Interestingly, when BAECs were treated with both VEGF and Wnt3a, this resulted in a reduction in axin2 mRNA compared to Wnt3a treatment alone (Fig. 3A). Remarkably, the inhibitory effect of VEGF on Wnt3a-stimulated induction of axin2 mRNA was completely abolished in eNOS-depleted BAECs (Fig. 3A). VEGF and Wnt3a are both known to promote proliferation of ECs; thus, we examined the effect of VEGF treatment on Wnt3a-stimulated BAEC proliferation and on cyclin D1 mRNA levels, a β-catenin target gene involved in cell cycle progression. We observed that treatment with VEGF or Wnt3a alone increased BrdU incorporation in BAECs. In contrast, proliferation of BAECs induced by cotreatment with Wnt3a and VEGF was reduced compared to Wnt3a treatment alone (Fig. 3B). Similarly, induction of cyclin D1 mRNA levels by Wnt3a was reduced by VEGF cotreatment (Fig. 3C). Taken together, these results suggest that VEGF-stimulated eNOS activation and NO production negatively regulate transcription of β-catenin target genes and cell proliferation induced by Wnt3a.

FIG 3.

VEGF inhibits Wnt/β-catenin signaling in an eNOS-dependent manner. (A) qRT-PCR analysis of axin2 mRNA levels in control or eNOS-depleted BAECs treated with Wnt3a-conditioned medium and/or VEGF (40 ng/ml; 24 h; n = 3) as indicated. eNOS was depleted in BAECs by transfection of siRNA against eNOS (eNOS-siRNA), and CT-siRNA was used for comparison. Depletion of eNOS was monitored by IB, and β-actin was used as a loading control. (B) BrdU incorporation assay in BAECs treated with Wnt3a-conditioned medium and/or VEGF (40 ng/ml; 24 h) as indicated. The percentage of BrdU-positive cells for each treatment was normalized to that of nontreated cells (n = 3). (C) qRT-PCR analysis of cyclin D1 mRNA levels in BAECs treated with Wnt3a-conditioned medium and/or VEGF (40 ng/ml; 6 h; n = 3) as indicated. The data are represented as means and SEM. *, P < 0.05; ns, not statistically significant.

SNO of β-catenin decreases its transcriptional activity.

We hypothesized that SNO is responsible for inhibition of the transcriptional activity of β-catenin by NO. We have previously shown by liquid chromatography-tandem mass spectrometry (LC–MS-MS) analyses that cysteine residues 300, 381, 439, 466, 520, and 619 of β-catenin are possible substrates for SNO by NO (26). In addition, Cys213 was identified as a potential SNO site in another proteomics study (28). To determine which cysteine residue of β-catenin the inhibitory effects of NO on the transcriptional activity of β-catenin are attributable to, we generated β-catenin mutants in which cysteine residues were replaced by nonnitrosylable serine. Thus, the following β-catenin mutants were generated: C213S-, C300S-, C381S-, C439S-, C466S-, C520S-, and C619S-β-catenins. We tested the sensitivity of these mutants to eNOS-mediated inhibition of β-catenin transcriptional activity (Fig. 4A). Wild-type (WT) β-catenin and the β-catenin mutants were coexpressed with S1179D-eNOS, and the transcriptional activity of β-catenin was monitored using the TOPFlash reporter. Our results showed that, among all the nonnitrosylable mutants tested, only C466S-β-catenin was insensitive to the inhibitory effects of NO (Fig. 4A). This suggests that NO inhibits the transcriptional activity of β-catenin through SNO of Cys466.

FIG 4.

SNO of Cys466 is responsible for the inhibitory effect of NO on β-catenin transcriptional activity. (A) β-Catenin luciferase reporter assay in HEK293T stably expressing the TOPFlash reporter and transfected with myc-tagged WT β-catenin or the indicated nonnitrosylable cysteine-to-serine-mutated β-catenin constructs. Where indicated, cells were cotransfected with S1179D-eNOS or the empty vector pcDNA3 (Mock) (n = 3). The data are represented as means and SEM. *, P < 0.05 versus mock transfected; ‡, P <0.05 versus the corresponding mutant without S1179D-eNOS; #, P < 0.05 versus cells transfected with WT β-catenin and S1179D-eNOS. The transfection levels of myc–β-catenin and eNOS were validated by IB, and β-actin was used as a loading control. (B) BAECs transfected with myc–WT, myc–C466S-, or myc–C619S-β-catenin were treated with VEGF (40 ng/ml; 15 min) or GSNO (0.1 mM; 30 min). SNO levels of β-catenin were determined by biotin switch assays and detected by anti-myc immunoblotting (n = 3). The transfection levels were validated by anti-myc, and eNOS was used as a loading control. The data are represented as means and SEM. *, P < 0.05. (C) qRT-PCR analysis of axin2 mRNA levels in BAECs expressing myc-tagged WT, C466S-, or C619-β-catenin treated or not with VEGF (40 ng/ml; 6 h; n = 4). The transfection levels of myc-tagged β-catenin were validated by immunoblotting, and β-actin was used as a loading control. The data are represented as means and SEM. *, P < 0.05 versus nontransfected and nonstimulated cells; ‡, P < 0.05 versus cells expressing myc–WT β-catenin without stimulation; #, P < 0.05 versus cells expressing myc–WT β-catenin and VEGF treated.

We have previously shown that Cys619 on β-catenin is susceptible to SNO and revealed that it participates in the VEGF-stimulated dissociation of β-catenin from VE-cadherin in ECs (26). To demonstrate that Cys466 is also a substrate for SNO, we expressed myc-tagged WT and C466S-β-catenin in ECs, and SNO levels following VEGF or GSNO treatment were determined by a biotin switch assay (Fig. 4B). As previously shown, VEGF and GSNO treatments of ECs increased SNO levels of WT catenin (26). The C466S mutation markedly reduced SNO of β-catenin induced by VEGF or GSNO. This indicates that Cys466 is an S-nitrosylation site on β-catenin in response to VEGF stimulation or following exposure of ECs to the NO donor GSNO. Interestingly, consistent with our previous study, C619S mutation decreased SNO levels induced by VEGF and GSNO (Fig. 4B), confirming that multiple cysteines on β-catenin are sites for SNO (26). Next, we examined whether Cys466 is important for mediating the inhibitory effects of VEGF on β-catenin signaling in ECs. myc-tagged WT, C466S-, or C619S-β-catenin was expressed in BAECs, and the effects of VEGF treatment on the induction of axin2 mRNA by β-catenin overexpression were determined (Fig. 4C). Similar to what is shown in Fig. 2A, overexpression of WT β-catenin did induce transcription of axin2, which was inhibited by VEGF treatment. In contrast, the induction of axin2 transcription by C466S-β-catenin expression in BAECs was resistant to VEGF treatment. Similar to the effect of VEGF on WT β-catenin, induction of axin2 mRNA levels by the C619S-β-catenin mutant was inhibited by VEGF treatment. These results obtained in BAECs suggest that residue Cys466 of β-catenin is a substrate for SNO and is responsible for the repression of β-catenin transcriptional activity by VEGF-stimulated NO production.

SNO of Cys466 disrupts the interaction between β-catenin and TCF4.

Previous crystallographic studies of the TCF4–β-catenin complex revealed that the N-terminal portion of TCF4 (residues 13 to 25) binds a positively charged groove created by the armadillo repeats 4 to 9 of β-catenin. Interestingly, the side chains of residues Ile19 and Phe21 of TCF4 are thought to form hydrophobic contacts with residues Cys466 and Pro463 and the aliphatic portion of Arg386 on β-catenin (34). Figure 5A highlights the proximal position of Cys466 of β-catenin relative to Ile19 of TCF4. Therefore, it is reasonable to hypothesize that addition of an NO group on Cys466 of β-catenin could affect the stability of the interaction with TCF4. To investigate this, TCF4 was immunoprecipitated from nuclear extracts of ECs stimulated with Wnt3a in the absence or the presence of GSNO, and the interaction with β-catenin was determined. Our results show that Wnt3a stimulation of BAECs induced the association of β-catenin with TCF4 (Fig. 5B). Interestingly, treatment of cells with GSNO diminished the interaction between β-catenin and TCF4 promoted by Wnt3a stimulation (Fig. 5B). Next, to determine if Cys466 of β-catenin is important for the effects of NO on disruption of the interaction between β-catenin and TCF4, we transfected HEK293T cells with WT or C466S-β-catenin, together with Flag-tagged TCF4, in the presence or the absence of constitutively active S1179D-eNOS. Cell lysates were immunoprecipitated with an anti-Flag antibody, and the TCF4–β-catenin complex was resolved. In the absence of eNOS, basal association of C466S-β-catenin with TCF4 was similar to that of WT β-catenin (Fig. 5C). Coexpression of S1179D-eNOS decreased the interaction between WT β-catenin and TCF4. In contrast, the association between C466S-β-catenin and TCF4 was not affected by the expression of S1179D-eNOS (Fig. 5C). Finally, the protein stability of β-catenin is also a determinant of its transcriptional activity; the half-lives of WT and C466S-β-catenin proteins expressed in COS7 in the presence of S1179D-eNOS cells were not different (15.7 ± 2.7 h and 14.8 ± 2.4 h, respectively). Together, these results suggest that SNO at residue Cys466 of β-catenin results in dissociation from TCF4 and that this may be responsible for the decrease of β-catenin transcriptional activity mediated by NO.

FIG 5.

Cys466 of β-catenin is necessary for NO-mediated disruption of the β-catenin–TCF4 complex. (A) Three-dimensional representation of human β-catenin (cyan) in complex with a TCF4 peptide (residues 13 to 25; yellow). The side chains of residues Cys466, Pro463, and Arg386 of β-catenin are shown in red, and residues Ile19 and Phe21 of TCF4 are highlighted in purple. The image was generated based on analysis of the PDB 1JPW protein structure (34) with the PyMol molecular graphics system (version 1.7.4). (B) Coimmunoprecipitation (IP) of β-catenin and TCF4 from the nuclear fraction of BAECs treated with Wnt3a-conditioned medium overnight and in the presence or absence of GSNO (0.1 mM; 30 min). Nonimmune IgG was used for control IP. The bar graph shows the normalized densitometric ratio, quantified by ImageJ, of β-catenin to TCF4 levels (n = 3). The data are represented as means and SEM. *, P < 0.05. (C) Co-IP of myc–WT or myc–C466S-β-catenin with Flag-TCF4 from HEK293 cell lysates transfected with S1179D-eNOS where indicated. The levels of β-catenin and TCF4 constructs in the IP were monitored using anti-myc and anti-Flag. The bar graph shows the normalized densitometric ratio, quantified by ImageJ, of myc–β-catenin to Flag-TCF4 levels (n = 3). The data are represented as means and SEM. *, P < 0.05; n/s, not statistically significant.

Cys466 of β-catenin is necessary for inhibition of EC proliferation by NO.

Next, we examined whether SNO of Cys466 of β-catenin could be responsible for the inhibitory effects of NO on Wnt/β-catenin-induced EC proliferation. To test this, we expressed WT or C466S-β-catenin in BAECs and determined by BrdU incorporation assays the inhibitory capacity of GSNO on β-catenin-dependent or on Wnt3a-stimulated cell proliferation. As shown previously, GSNO decreased proliferation of BAECs that overexpress WT β-catenin. In contrast, BAECs that express C466S-β-catenin were completely resistant to the inhibitory effects of GSNO (Fig. 6A). Furthermore, Wnt3a-induced proliferation of cells expressing WT β-catenin was decreased by GSNO treatment (Fig. 6B). Importantly, Wnt3a-stimulated proliferation of BAECs expressing C466S-β-catenin was not affected by GSNO. These results confirm that SNO of Cys466 of β-catenin is responsible for the inhibitory effects of NO on Wnt/β-catenin-stimulated proliferation of ECs.

FIG 6.

SNO of Cys466 is critical for the inhibitory effects of NO on EC proliferation. (A) BrdU incorporation assay in BAECs expressing myc-tagged WT or C466S-β-catenin in the presence or absence of GSNO (0.1 mM; 18 h; n = 3). (B) BrdU incorporation assay in BAECs expressing myc-tagged WT or C466S-β-catenin and treated with Wnt3a-conditioned medium in the presence or absence of GSNO (0.1 mM; 18 h; n = 3). The percentage of BrdU-positive cells for each treatment was normalized to that of nontreated cells. The data are represented as means and SEM. *, P < 0.05; n/s, not statistically significant.

DISCUSSION

Here, we investigated the influence of NO on β-catenin transcriptional activity and its consequences for Wnt/β-catenin-mediated EC proliferation. We elucidate a new mechanism that explains how eNOS-derived NO modifies Wnt/β-catenin signaling to regulate proliferation of ECs. The main findings are (i) that eNOS-derived NO inhibits β-catenin transcriptional activity and EC proliferation stimulated by Wnt3a, (ii) that VEGF may act as a feedback inhibitor for Wnt/β-catenin signaling through the generation of NO by eNOS, (iii) that Cys466 of β-catenin is a substrate for SNO in response to VEGF stimulation of ECs, (iv) that disruption of the β-catenin–TCF4 interaction by SNO of Cys466 may be responsible for inhibition of the transcriptional activity of β-catenin by NO, and (v) that it may be responsible for the repression of Wnt/β-catenin-stimulated EC proliferation by NO. Taken together, our findings have uncovered a mechanism used by NO to modulate Wnt/β-catenin signaling that may explain some of the important effects produced by eNOS in ECs and during angiogenesis.

We have previously shown, using LC–MS-MS, that many cysteine residues on β-catenin are susceptible to S-nitrosylation by GSNO in vitro and that SNO of Cys619 of β-catenin at cell junctions promotes its dissociation from VE-cadherin and is important for VEGF-induced EC permeability (26). It is also well established that, once released from cell junctions, β-catenin may act in concert with TCF/LEF transcription factors in the canonical Wnt signaling pathway to modulate gene expression in many cell types (5, 35–38). Here, we reveal that SNO of Cys466 can affect this function of β-catenin in ECs. Using the biotin switch assay, we show that mutation of Cys466 and Cys619 on β-catenin did reduce SNO levels, confirming that both residues on β-catenin are S-nitrosylated following VEGF stimulation of ECs. However, only SNO of Cys466 is responsible for inhibition of β-catenin transcriptional activity by NO. Thus, the two functions of β-catenin are regulated by SNO of different residues, Cys466 and Cys619.

Previous studies have shown that NO donors, including NO-donating aspirin or other NO-releasing chemical entities, could repress the transcriptional activity of β-catenin, and some have suggested that SNO of β-catenin could be responsible for this effect (29–31, 39–41). Chemical NO-donating agents have shown promise on cancer cells, where they attenuate Wnt/β-catenin signaling, reduce proliferation, and induce apoptosis (30, 31). However, the direct targets of NO responsible for attenuating Wnt/β-catenin were not identified. To our knowledge, this study provides the first demonstration that SNO of Cys466 of β-catenin by NO is responsible for disruption of the β-catenin–TCF4 association, which results in inhibition of the transcriptional activity of the complex. As stated above, previous structural studies have defined residues 13 to 31 of TCF4 as forming a minimal binding unit for the armadillo repeats 4 to 9 of β-catenin (42). Cys466 is located in the 8th armadillo repeat of β-catenin and interacts with Ile19 of TCF4 (34). Further analysis of the β-catenin–TCF4 complex (Protein Data Bank [PDB] file 1JPW) by molecular modeling revealed that modification of the thiol group of Cys466 by SNO causes a shift of Ile19 of TCF4, which results in decreased binding energy (−7.4 kcal/mol) between β-catenin and TCF4, calculated using the molecular mechanics-generalized born surface area (MM-GBSA) method (43). In addition, Cys466 is close to Lys435 of β-catenin, which has also been suggested to play an important role in the interaction with TCF4 via a salt bridge hydrogen bond with Asp16 of TCF4 (34, 42). Hence, it is likely that SNO of Cys466 disrupts the interaction between β-catenin and TCF4 by inducing conformational changes that affect binding to TCF4 and result in decreased transcriptional activation.

Cross-regulation between NO and Wnt/β-catenin signaling has been shown in other systems. Interestingly, it appears that the Wnt/β-catenin activation state determines the modulatory effects of NO. This study and others show that NO is inhibitory for β-catenin transcriptional activity when Wnt/β-catenin signaling is activated (29–31, 39–41). Consistent with these results, we revealed that VEGF-mediated eNOS activation in ECs acts as an inhibitor of Wnt3a-induced β-catenin activation and cell proliferation (Fig. 3) (44, 45). Conversely, it has been shown that NO may act as a positive modulator of β-catenin activity under low basal Wnt activation. For instance, expression of iNOS (NOS2) in embryonic stem cells influences cell differentiation and early lineage commitment through activation of β-catenin (46). Another study showed that β-catenin associates with eNOS in quiescent ECs and that eNOS activation causes cGMP-dependent nuclear translocation of β-catenin and influences angiogenesis (47). In agreement with these studies, our results show that, in resting ECs, VEGF activates β-catenin signaling and that this is prevented by knockdown of eNOS (Fig. 3A). In contrast to the inhibitory role of NO for β-catenin, the involvement of SNO in the positive effects of NO on β-catenin nuclear signaling remain to be defined. Perhaps when β-catenin is associated with VE-cadherin at cell-cell contacts, VEGF-induced NO dissolves adherens junctions through SNO of Cys619, and this could promote the transcriptional activity of β-catenin by allowing translocation from the plasma membrane to the nucleus (26).

VEGF was previously identified as one of the target genes of Wnt/β-catenin signaling (44, 45). Our results now reveal that VEGF, through the production of NO, may engage in a negative-feedback loop regulating Wnt signaling in ECs. We show that the inhibitory effect of VEGF on Wnt signaling depends on eNOS activation and NO-mediated dissociation of β-catenin from the transcription factor TCF4. Both VEGF and Wnt exert angiogenic properties; however, distinct outcomes of VEGF and Wnt in the regulation of EC specification and vascular remodeling have been reported (3, 48). During angiogenic sprouting, VEGF signaling is implicated in endothelial tip cell specification, maintenance, and migration, whereas Wnt-dependent intracellular signals are involved in the promotion of stalk cells (48, 49). Our results allow us to speculate that inhibition of Wnt3a-induced proliferation of ECs by VEGF-mediated eNOS activation may be relevant during angiogenesis in order for VEGF to suppress proliferative signals in tip cells by restraining Wnt/β-catenin signaling. This could contribute to maintaining the migratory phenotype of endothelial tip cells during sprouting. Therefore, it would make sense that Wnt-induced VEGF may somehow affect the outcomes of Wnt in the remodeling due to its inhibitory effect on Wnt/β-catenin in certain contexts.

In summary, our results demonstrate that NO directly modulates Wnt/β-catenin nuclear signaling, resulting in decreased β-catenin transcriptional activity and in reduced EC proliferation stimulated by Wnt3a. Our results shed light on a mechanism that explains the inhibitory effects of NO on β-catenin transcriptional activity. This mechanism involves VEGF-induced SNO of β-catenin at Cys466 by eNOS-derived NO, which disrupts the β-catenin–TCF4 complex and results in decreased transcription of β-catenin target genes and proliferation of ECs. Thus, NO affects the function of β-catenin at the plasma membrane through SNO of Cys619 and in the nucleus by modifying Cys466. This study further highlights the roles of eNOS-derived NO in EC biology and angiogenesis.

MATERIALS AND METHODS

Cell culture and reagents.

BAECs were obtained from VEC Technologies (Rensselaer, NY) and grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (vol/vol) fetal bovine serum (FBS) (HyClone, Logan, UT), 2 mM l-glutamine, 100 U/ml penicillin, and 100 mg/ml streptomycin. COS-7 and HEK293 cells were grown in DMEM supplemented with 10% FBS (Invitrogen), 2 mM l-glutamine, 100 U/ml penicillin, and 100 mg/ml streptomycin. HEK293T stable Wnt reporter cells that express the β-catenin (firefly) luciferase reporter TOPFlash and Renilla luciferase were provided by Stéphane Angers (University of Toronto, Toronto, Canada) (33).

The following primary antibodies were used: anti-myc-tag (9B11), anti-β-actin (8H10D10), and anti-BrdU (Bu20a) monoclonal antibodies (MAb) from Cell Signaling; anti-β-catenin, anti-eNOS, and anti-annexin II MAb from BD Transduction Laboratories; anti-TCF-4 (6H5-3) MAb from Upstate; and anti-Flag (M2) MAb from Sigma. Recombinant human VEGF-A (referred to as VEGF here), obtained from R&D Systems, was used for cell stimulation throughout this study. GSNO was from Sigma-Aldrich, and NOC-18 was from Santa Cruz.

Plasmids, siRNA, and cell transfections.

Bovine S1179D-eNOS and S1179A-eNOS (in pcDNA3) were provided by William C. Sessa (Yale University School of Medicine, New Haven, CT). Human Flag-TCF-4 was provided by Hans Clevers (University Medical Center, Utrecht, The Netherlands). M50 Super 8× TOPFlash and M51 Super 8× FOPFlash vectors were purchased from Addgene. The Renilla luciferase reporter vector pRL-TK was from Promega. Full-length human β-catenin was purchased from Open Biosystems (Huntsville, AL) and subcloned in pCMV-3Myc (Stratagene). Single point mutations of myc–β-catenin plasmid causing codon changes from cysteine to serine for residues Cys213, Cys300, Cys381, Cys439, Cys466, Cys520, and Cys619 were generated by using a QuikChange site-directed mutagenesis kit (Stratagene). The sequences of mutagenic sense primers can be found in Table S1 in the supplemental material. All the mutations were verified by DNA sequencing.

siRNA control and siRNA against eNOS were generated by Thermo Scientific Dharmacon. The sense sequences are as follows: 5′-CCAGGAAGAAGACCUUUAAUU-3′ and 5′-CCAACAUGCUGCUGGAAAUUU-3′ for eNOS-siRNA, as well as 5′-AUGAACGUGAAUUGCUCAAUU-3′ for CT-siRNA. Cell transfections were performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol.

Wnt3a-conditioned medium.

Wnt3a-conditioned medium was collected by culturing Wnt3a stable transfected mouse L cells (CRL-2647; ATCC) according to the manufacturer's instructions. Conditioned medium from parental mouse L cells not producing Wnt3a (CRL-2648) was also collected to use as a CT. The activity of Wnt3a-conditioned medium was monitored by luciferase assay in the TOPFlash stable HEK293 cell line. Wnt3a-conditioned medium that induced at least 50-fold luciferase activity compared to CT medium was used all experiments.

Luciferase assay.

COS-7 cells were transfected with the β-catenin-dependent luciferase (firefly) reporter plasmid TOPFlash or the negative-control FOPFlash and Renilla luciferase (pRL-TK) as an internal control. In some instances, HEK293T stable Wnt reporter cells expressing TOPFlash and Renilla luciferase were used. Luciferase activity was determined using a dual-luciferase reporter assay system (Promega, Madison, WI) 48 h after transfection of the indicated plasmid constructs.

BrdU proliferation assay.

Cells were incubated with 0.03 mg/ml BrdU at 37°C for 30 min, fixed with 70% ethanol for 5 min, and then denatured with 1.5 M HCl for 30 min at room temperature. After incubation with PBS–1% BSA–0.3%Triton X-100 to block nonspecific staining for 1 h, the cells were incubated with BrdU antibody overnight at 4°C. After three washes with PBS, the cells were then incubated with Alexa Fluor 568-conjugated goat anti-mouse antibody (Invitrogen) for another 2 h. The samples were then counterstained with DAPI (4,6-diamidino-2-phenylindole) to stain the nuclei and analyzed with a Zeiss Axio Observer.Z1 microscope.

RNA extraction and quantitative RT (qRT)-PCR.

Total RNA was extracted with an RNeasy minikit (Qiagen). After DNase I treatment, cDNA was synthesized using the SuperScript II reverse transcriptase kit for reverse transcription (RT)-PCR (Invitrogen) from 1 μg total RNA according to the manufacturer's instructions. Real-time PCR was performed with SYBR Select master mix (Applied Biosystems) and the Eco real-time PCR system (Illumina) or the ViiA 7 real-time PCR system (Life Technologies). Gene expression analysis was performed by using the comparative cycle threshold (ΔC_T) method, normalized using reference β-actin and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) gene expression, and presented as the mean fold change (and standard error of the mean [SEM]) compared with the control. The sequences of primers can be found in Table S2 in the supplemental material.

Biotin switch assay.

S-nitrosylated proteins were detected using the biotin switch assay (50). Lysates from transfected BAECs were incubated in HENS buffer containing 250 mM HEPES (pH 7.7), 1 mM EDTA, 0.1 mM neocuporine, SDS (2% final concentration), and 20 mM methylmethanethiosulfonate (MMTS) at 50°C for 20 min. Proteins were precipitated with cold acetone, washed twice, resuspended in HENS (1% SDS final concentration), and mixed with 0.2 mM biotin-HPDP {N-[6-(biotinamido)hexyl]-3′-(2′-pyridyldithio)propionamide} and 2.5 mM ascorbate at room temperature for 1 h. All the steps were carried out in the dark. Finally, the biotinylated proteins were purified using streptavidin-agarose beads, separated by SDS-PAGE, and detected by immunoblotting.

IP and immunoblotting.

For immunoprecipation (IP) from whole-cell lysate, cells were solubilized in lysis buffer containing 1% Triton X-100, 50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 20 mM sodium fluoride, 1 mM sodium pyrophosphate, 1 mM orthovanadate, and protease inhibitor cocktail (Roche Diagnostics). For TCF4 IP from EC nuclear lysates, nuclei were first separated as described previously (51). Soluble proteins were incubated with primary antibodies (2 μg) at 4°C overnight. Protein A-Sepharose (Sigma; 50 ml of a 50% slurry) or anti-Flag M2 affinity gel (for Flag IP) was added and incubated for an additional hour. The immune complexes were precipitated, separated by SDS-PAGE, transferred onto a nitrocellulose membrane (Hybond-ECL; GE Healthcare), and Western blotted. Detection and quantification were performed with an Odyssey infrared-imaging system (Li-Cor Biosciences) using the appropriate Alexa Fluor 680- or Alexa Fluor 800-labeled secondary antibodies (Invitrogen) or with an Image Quant chemiluminescence-based detection system (enhanced chemiluminescence [ECL]) (GE Healthcare).

NO release analysis.

Cell culture medium from COS-7 cells expressing S1179D- or S1179A-eNOS and myc-tagged β-catenin was collected and processed for the measurement of nitrite (NO₂⁻), the stable breakdown product of NO, in aqueous solution. NO-specific chemiluminescence was measured using an NO analyzer (NOA 280i; GE Ionics Instruments) (52).

Protein degradation assay.

COS-7 cells overexpressing myc–WT or myc–C466S-β-catenin were treated with 100 μg/ml cycloheximide (CHX) in serum-free medium for 24 h. The cells were washed twice with cold PBS and lysed in RIPA buffer (50 mM NaCl, 50 mM Tris, 0.1 mM EDTA, 0.1 mM EGTA, 0.1% SDS, 0.1% deoxycholic acid, 1% NP-40) supplemented with 20 mM sodium fluoride, 1 mM sodium pyrophosphate, 1 mM sodium orthovanadate, and complete EDTA-free protease inhibitor (Roche). Protein was separated by SDS-PAGE, transferred onto a nitrocellulose membrane, and Western blotted. Detection of the signal was performed with Image Quant, and densitometric analysis was performed using ImageJ. The results are expressed as a ratio of the levels of myc in treated cells versus nontreated cells (t = 0). The half-lives of proteins were calculated with linear regression equations.

Statistics.

Values are reported as means and SEM. Statistical comparisons were performed by analysis of variance (ANOVA), followed by Bonferroni's post hoc test, using GraphPad Prism 5.0. Student's t test was used to compare two data sets (Fig. 5B). A probability value of <0.05 was considered statistically significant.