Phosphorylation of GATA-6 is required for vascular smooth muscle cell differentiation after mTORC1 inhibition

See allHide authors and affiliations

Sci. Signal.  12 May 2015:
Vol. 8, Issue 376, pp. ra44
DOI: 10.1126/scisignal.2005482


Vascular smooth muscle cells (VSMCs) undergo transcriptionally regulated reversible differentiation in growing and injured blood vessels. This dedifferentiation also contributes to VSMC hyperplasia after vascular injury, including that caused by angioplasty and stenting. Stents provide mechanical support and can contain and release rapamycin, an inhibitor of the mechanistic target of rapamycin complex 1 (mTORC1). Rapamycin suppresses VSMC hyperplasia and promotes VSMC differentiation. We report that rapamycin-induced differentiation of VSMCs required the transcription factor GATA-6. Inhibition of mTORC1 stabilized GATA-6 and promoted the nuclear accumulation of GATA-6, its binding to DNA, its transactivation of promoters encoding contractile proteins, and its inhibition of proliferation. These effects were mediated by phosphorylation of GATA-6 at Ser290, potentially by Akt2, a kinase that is activated in VSMCs when mTORC1 is inhibited. Rapamycin induced phosphorylation of GATA-6 in wild-type mice, but not in Akt2−/− mice. Intimal hyperplasia after arterial injury was greater in Akt2−/− mice than in wild-type mice, and the exacerbated response in Akt2−/− mice was rescued to a greater extent by local overexpression of the wild-type or phosphomimetic (S290D) mutant GATA-6 than by that of the phosphorylation-deficient (S290A) mutant. Our data indicated that GATA-6 and Akt2 are involved in the mTORC1-mediated regulation of VSMC proliferation and differentiation. Identifying the downstream transcriptional targets of mTORC1 may provide cell type–specific drug targets to combat cardiovascular diseases associated with excessive proliferation of VSMCs.


Mature vascular smooth muscle cells (VSMCs) retain plasticity to undergo phenotypic modulation in response to growth factor stimuli or injury. VSMCs in the vessel wall normally exhibit a differentiated contractile phenotype, but can undergo phenotypic switching to a dedifferentiated, proliferative, and migratory phenotype with enhanced protein synthesis in response to extracellular cues (1, 2). This dedifferentiated or “synthetic” phenotype not only contributes to physiological processes such as vascular remodeling and angiogenesis, but it can also contribute to the pathogenesis of both atherosclerosis and intimal hyperplasia. Stents eluting rapamycin or rapamycin analogs have revolutionized coronary artery revascularization, reducing rates of restenosis compared to bare metal stents (3). Exploring the molecular basis for the actions of mechanistic target of rapamycin (mTOR) complex 1 (mTORC1) inhibitors has important implications for future vascular therapeutics.

mTOR is a ubiquitously distributed serine/threonine protein kinase. When associated with other proteins in mTORC1, it serves an important checkpoint function in regulating specific protein synthesis in response to mitogens, stress, energy, and nutritional signals (4). mTORC1 coordinates anabolic processes including cell growth, proliferation, and metabolism (5). mTORC1 activity can be inhibited by nutrient starvation or pharmacologically by the inhibitor rapamycin (4).

The mTORC1 pathway is activated in VSMCs in response to vascular injury in vivo. Rapamycin inhibits VSMC proliferation and migration in vitro (68). Moreover, we have demonstrated that rapamycin treatment induces VSMC differentiation through increasing the expression of contractile protein-encoding mRNAs (9). This is mediated by relief of the classical feedback loop in which mTORC1 and its substrate S6 kinase 1 (S6K1) promote insulin receptor substrate 1 (IRS-1) degradation to dampen signaling through insulin and insulin-like growth factors (10). We have shown that in VSMCs, Akt2 is specifically activated in response to mTORC1 inhibition, and that this induction of the activity of Akt2, but not Akt1, is required for the VSMC differentiation response (10). The key downstream transcriptional targets of Akt2 in vitro and in vivo are not yet known.

Whereas mTORC1 was initially appreciated for its role in regulating protein synthesis in mammalian cells, little is known regarding mTORC1-mediated regulation of cell type–specific transcription. Here, we demonstrate that rapamycin promotes VSMC differentiation through activation of GATA-binding protein 6 (GATA-6), and that this signaling may be mediated by Akt2-mediated phosphorylation of GATA-6. We identify a function of mTORC1 in regulation of cell type–specific transcription, a finding that has important implications for vascular therapeutics.


GATA-6 mediates the mTORC1-regulated modulation of SMC differentiation and proliferation

We have previously shown that the mTORC1 inhibitor rapamycin promotes VSMC differentiation through the classic feedback activation of the IRS-1–PI3K (phosphatidylinositol 3-kinase)–Akt pathway (10). mTORC1 inhibition induces expression of VSMC-specific markers including smooth muscle myosin heavy chain (SM-MHC), h-caldesmon, SM-α-actin, and calponin at the mRNA and protein levels (9), which requires activation of the Akt2 isoform (10). Because smooth muscle contractile proteins are transcriptionally regulated, we next sought to identify transcription factors downstream of Akt2 signaling. GATA-6 is present in mature, differentiated smooth muscle, but its abundance is rapidly decreased after vascular injury and growth factor stimulation (11, 12). Because GATA-6 plays a potent antiproliferative, prodifferentiation role in VSMC in vitro and in vivo (11, 12), we investigated whether GATA-6 could mediate rapamycin-induced differentiation in human coronary artery SMCs (hCASMCs). Consistent with our previous studies, rapamycin treatment induced MYH11 mRNA by more than fourfold in control transfected hCASMCs (Fig. 1A). Notably, GATA-6 knockdown significantly reduced the basal amount of MYH11 mRNA and prevented rapamycin induction of this gene, which is the most definitive marker of VSMC differentiation (13) (Fig. 1A). The same pattern was also observed at the protein level with multiple contractile markers (SM-MHC, calponin, and SM-α-actin) in hCASMCs from three different donors (Fig. 1, B and C, and fig. S1, A and B).

Fig. 1 GATA-6 is required for rapamycin-induced VSMC differentiation.

(A and B) hCASMCs were transfected with control or GATA-6 siRNAs, then treated with ethanol (vehicle) or rapamycin for 48 hours, and analyzed for (A) GATA-6 or MYH11 mRNA by quantitative reverse transcription PCR (RT-PCR) (n = 3 biological replicates; *P = 0.05; n.s., not significant) or (B) abundance of SMC contractile proteins and p21cip (Western blot). (C) Quantitation of three replicate experiments from each of three different donors as in (B) (**P < 0.01, ***P < 0.001, ****P < 0.0001). (D) hCASMCs were transfected with siRNAs as above and treated with either ethanol or rapamycin. Proliferation was monitored by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay after 72 hours (n = 8 biological replicates; ***P < 0.001, ****P < 0.0001). Data were presented as the fold change of cell index compared to input. (E) hCASMCs were transfected with siRNAs as above and treated with either ethanol or rapamycin. Cells were trypsinized and counted (n = 3 biological replicates; *P = 0.05).

Rapamycin inhibits VSMC proliferation in vitro and in vivo (68). We have shown that rapamycin induces the cyclin-dependent kinase (CDK) inhibitors p21cip and p27kip in VSMCs (9). The antiproliferative effect of GATA-6 in VSMCs depends on induction of p21cip (12), suggesting that p21cip may be a common antiproliferative effector of rapamycin and GATA-6. Consistent with this notion, GATA-6 knockdown reduced the basal amount of p21cip and prevented its induction by rapamycin (Fig. 1, B and C, and fig. S1, A and B). GATA-6 knockdown also significantly increased proliferation in hCASMCs and attenuated the inhibitory effects of rapamycin on proliferation (Fig. 1, D and E, and fig. S1, C and D). These data indicate that GATA-6 plays a role in rapamycin-induced differentiation and proliferation, regulating genes encoding both SMC contractile proteins and cell cycle progression factors.

mTORC1 inhibition promotes GATA-6 transactivation of an MHC-luciferase promoter reporter

Because GATA-6 was required for rapamycin-induced VSMC differentiation, we next assessed whether rapamycin alters GATA-6 transactivation of a smooth muscle–specific promoter. We used a truncated (836 base pairs) MYH11 promoter-luciferase reporter because it has few other regulatory elements and its activation is GATA-dependent (14). Rapamycin induced a dose-dependent activation of this MYH11 promoter-reporter in transfected hCASMCs (fig. S2A), suggesting that rapamycin could induce MYH11 transcription by activating endogenous GATA-6. Similar to rapamycin-treated cells, those overexpressing GATA-6 showed a twofold increase in reporter activity, an effect that was not significantly enhanced by rapamycin treatment, suggesting that GATA-6 overexpression had saturated reporter activity (fig. S2B). Notably, a GATA-6 mutant lacking the two DNA binding zinc finger domains (Δzf GATA-6) (12) not only failed to activate the luciferase reporter but also prevented rapamycin-induced promoter-reporter activity (fig. S2B). We further demonstrated that knockdown of endogenous GATA-6 in hCASMCs reduced basal activity and prevented rapamycin-induced expression from this reporter (fig. S2C). A reporter with a mutation of the GATA element showed reduced basal activity and was not responsive to GATA-6 overexpression, thus confirming the GATA dependence of this reporter (fig. S2D).

mTORC1 inhibition stabilizes GATA-6 protein, promotes its nuclear accumulation, and increases its DNA binding activity

GATA-6 protein abundance increased after rapamycin treatment (Fig. 1, B and C, and fig. S1, A and B). Rapamycin gradually increased endogenous GATA-6 protein abundance over 2 to 24 hours (Fig. 2A) without significantly altering GATA-6 mRNA abundance (Fig. 1A and fig. S3) in hCASMCs. GATA-6 protein abundance was also increased by rapamycin treatment in vivo, as intraperitoneal injection of rapamycin increased Gata-6 and SM-MHC abundance in mouse arteries in a time-dependent manner (Fig. 2B). Immunofluorescence staining for endogenous GATA-6 in rapamycin-treated hCASMCs similarly revealed an increase in GATA-6 abundance over time (Fig. 2C). Whereas the increase in GATA-6 abundance was apparent in both the cytosol and nucleus, we noted strong nuclear accumulation after rapamycin treatment (Fig. 2C and fig. S4). Rapamycin-induced nuclear accumulation of GATA-6 was confirmed by biochemical subcellular fractionation of hCASMCs (fig. S5).

Fig. 2 Rapamycin stabilizes GATA-6 protein and promotes its binding to DNA.

(A) hCASMCs were treated with rapamycin for the indicated times and immunoblotted for endogenous GATA-6 protein. The bar graph shows the quantitation of GATA-6 protein abundance from three individual experiments [*P = 0.05 compared with vehicle (Veh) control]. p, phosphorylated. t, total. (B) Mice were intraperitoneally injected with vehicle or rapamycin and euthanized at the indicated time points. Aortas and femoral arteries were pooled, and the lysates were immunoblotted for endogenous Gata-6 and SM-MHC proteins. The bar graph shows the quantitation of Gata-6 and SM-MHC protein abundance from three individual experiments (*P = 0.05 compared with vehicle control). (C) hCASMCs plated on cover slides were treated with ethanol or rapamycin, fixed at the indicated time points, and immunostained for GATA-6. Z-stack images are shown. Scale bars, 50 μm. Representative of two independent experiments. (D) hCASMCs were treated with ethanol or rapamycin for 4 hours and then with cycloheximide (CHX) for times indicated. Cells were harvested at the indicated time points and immunoblotted for the indicated proteins. A representative Western blot analysis is shown in fig. S6. The plot is the quantitation of three individual experiments (P < 0.05, paired t test). (E) hCASMCs were treated with vehicle or rapamycin and subjected to ChIP assay with primer sets flanking the GATA elements or control regions (lacking GATA elements) in the human ACTA2 and MYH11 gene promoters (n = 3 independent experiments; *P = 0.05 compared with time zero).

We next examined whether rapamycin affects GATA-6 protein half-life by measuring the rate of protein decay after inhibition of new protein synthesis with cycloheximide. IRS-1 is stabilized after rapamycin treatment in VSMCs (10) and served as a positive control. Rapamycin pretreatment before cycloheximide treatment increased IRS-1 protein abundance and half-life in hCASMCs (Fig. 2D and fig. S6). Similarly, rapamycin pretreatment increased GATA-6 protein abundance and increased its half-life from 4 to 7 hours (Fig. 2D and fig. S6), indicating that mTORC1 inhibition stabilizes GATA-6 protein.

We next assessed the effect of mTORC1 inhibition on GATA-6 DNA binding activity with chromatin immunoprecipitation (ChIP)–quantitative polymerase chain reaction (qPCR) assays performed on rapamycin-treated hCASMCs. In immunoprecipitates of endogenous GATA-6, the relative abundance of coprecipitated DNA was analyzed by qPCR with primer sets that amplified the GATA element–containing regions within the human SM-α-actin (ACTA2) and SM-MHC (MYH11) gene promoters. There was a time-dependent increase in GATA-6 binding at these GATA elements, indicating that rapamycin could increase GATA-6 DNA binding activity (Fig. 2E).

mTORC1 inhibition induces phosphorylation of GATA-6 in an Akt2-dependent manner

To elucidate the mechanism of how mTORC1 inhibition modulates the protein abundance and DNA binding activity of GATA-6, we next investigated posttranslational modifications of GATA-6 triggered by rapamycin treatment. We have previously found that the inhibition of mTORC1 and S6K1 by rapamycin or adiponectin relieves the feedback inhibition of IRS-1, resulting in increased PI3K and Akt2 activity (10, 15). VSMC differentiation induced by mTORC1 inhibition requires activation of Akt2 (10, 15). GATA-6 contains an evolutionarily conserved basophilic sequence motif (KPQKRVPS*), suggesting that Ser290 within this motif could potentially be targeted by Akt. To determine whether the mTORC1 pathway regulates GATA-6 through phosphorylation at this motif, GATA-6 immunoprecipitates from rapamycin-treated hCASMCs were immunoblotted with an antibody that detects the phosphorylated Akt substrate motif (RXRXXS*/T*). GATA-6 phosphorylation was detected after 0.5 hour and peaked between 2 and 4 hours of rapamycin treatment (Fig. 3A).

Fig. 3 Rapamycin induces phosphorylation of GATA-6 in an Akt2-dependent manner.

(A) GATA-6 immunoprecipitates (IP) from hCASMCs treated with rapamycin for the indicated times were Western-blotted with Akt phospho-substrate motif antibody (RXRXXS*/T*). (B) hCASMCs were transfected with control or siRNAs against Akt1 or Akt2 and treated with ethanol or rapamycin. GATA-6 immunoprecipitates from lysates were Western-blotted with Akt phospho-substrate motif antibody (RXRXXS*/T*). (C) Wild-type (WT) or Akt2 knockout (KO) mice were injected intraperitoneally with vehicle or rapamycin. Aortas and femoral arteries were harvested and lysed. Gata-6 immunoprecipitates from these lysates were Western-blotted with Akt phospho-substrate motif antibody (RXRXXS*/T*). (D) GATA-6 immunoprecipitates from HEK293 cells (upper panel) or recombinant GSK3 protein (bottom panel) were used as substrates in in vitro kinase assays with recombinant Akt1 or Akt2 and immunoblotted with Akt phospho-substrate motif antibody (RXRXXS*/T*). (E) Bacterial glutathione S-transferase (GST), GST-GATA-6_20aa_WT, or GST-GATA-6_20aa_S290A protein was used as substrate in in vitro kinase assays with recombinant Akt1 or Akt2 and [33P]adenosine triphosphate (ATP). The samples were subjected to SDS-PAGE and autoradiography. (A) to (D) are representative of at least three individual experiments. (E) is representative of two individual experiments.

Because we have previously found that VSMC differentiation in response to mTORC1 inhibition depends specifically on activation of the Akt2 isoform (10, 15), we tested whether there was a similar Akt isoform specificity for phosphorylation of GATA-6. Transfection of small interfering RNAs (siRNAs) directed against Akt2, but not against Akt1, prevented the rapamycin-induced phosphorylation of GATA-6 in hCASMCs (Fig. 3B). Consistent with this finding, transfection of Akt2 or Myr-Akt2, but not of Akt1, induced phosphorylation of GATA-6 in human embryonic kidney (HEK) 293 cells (fig. S7). Furthermore, intraperitoneal injection of rapamycin induced phosphorylation of Gata-6 as detected by the Akt phospho-substrate motif antibody in the arteries of wild-type mice but not in Akt2−/− mice (Fig. 3C). Although this evidence supports a role for Akt2 in phosphorylation of GATA-6, the motif recognized by the antibody is basophilic and could potentially be phosphorylated by protein kinase A (PKA) (16). However, treatment of hCASMCs with the PKA inhibitor efficiently inhibited phosphorylation of the PKA substrate cyclic adenosine monophosphate response element–binding protein (CREB) at Ser133, but did not diminish rapamycin-induced immunoreactivity with the Akt phospho-substrate motif antibody (fig. S8).

To determine whether Akt2 can directly phosphorylate GATA-6, we performed in vitro kinase assays using recombinant Akt1 or Akt2 proteins. Both recombinant Akt1 and Akt2 phosphorylated a glycogen synthase kinase 3β (GSK3β substrate peptide in vitro, but only Akt2 phosphorylated full-length GATA-6 immunoprecipitated from HEK293 cells, as detected by the Akt phospho-substrate antibody (Fig. 3D). To determine whether Ser290 was directly phosphorylated by Akt2, we used GST-fusion proteins encoding 20 amino acids of GATA-6, centered around either wild-type Ser290 or a S290A mutation, as substrates for in vitro kinase assays. The wild-type, but not the S290A sequence, was phosphorylated specifically by Akt2; there was minimal incorporation of 33P with Akt1, although Akt1 had greater activity toward the GSK3β control substrate in this assay (Fig. 3E). Mass spectrometry (MS) analysis of an Akt2 in vitro kinase assay using a synthetic substrate peptide encoding this region of GATA-6 confirmed phosphorylation of Ser290 because this generated a spectrum identical to that of a corresponding synthetic peptide containing phospho-serine at position 290 (fig. S9). We additionally assessed the Akt isoform specificity for Ser290 in intact cells by coexpressing Akt1 or Akt2 with GATA-6 wild type or S290A in HEK293 cells. GATA-6 was phosphorylated in cells coexpressing Akt2, but not Akt1, and this depended on Ser290 (fig. S10).

Akt1 and Akt2 are highly homologous (82% identity between mouse Akt1 and Akt2) (17) with a conserved four-domain structure that includes PH, linker, catalytic, and regulatory domains, with the highest degree of homology (91%) in the catalytic domain. To identify regions that mediate the Akt2-specific phosphorylation of GATA-6, we cotransfected GATA-6 into HEK293 cells with epitope-tagged chimeric Akt constructs in which one or more domains are swapped between Akt1 and Akt2 (17) (fig. S11A). Of the eight chimeras tested, only those containing the Akt2 linker domain (Akt1222, Akt2211, Akt2221, and Akt2212), but not those containing the Akt1 linker domain, induced phosphorylation of GATA-6 (fig. S11, A and B). These data suggest that the linker domain, which has the lowest identity (46%) between (mouse) Akt1 and Akt2 (17), confers GATA-6 substrate specificity to Akt2.

Phosphorylation of Ser290 promotes GATA-6 DNA binding and function

To investigate the effect of phosphorylation on GATA-6 activity, we used phospho-mimetic (S290D) and nonphosphorylatable (S290A) mutants. To assess the effects of this phosphorylation on DNA binding, we generated stable hCASMC clones that overexpressed Myc epitope–tagged wild-type GATA-6, S290D, or S290A. We verified Myc–GATA-6 overexpression in ~100% of the cells by immunostaining (fig. S12). ChIP-qPCR analysis using anti-Myc antibody revealed that the S290D mutant bound more target DNA per protein input than did the wild type or S290A, indicating that phosphorylation at Ser290 enhanced the DNA binding affinity. The binding of the S290D mutant to the MYH11 or ACTA2 promoters was two- to threefold greater than that of wild-type GATA-6. The binding of the S290A mutant was modestly but significantly reduced compared to that of wild-type GATA-6. Amplification with negative control primers confirmed the specificity of this assay (Fig. 4A).

Fig. 4 Phosphorylation of Ser290 promotes GATA-6 DNA binding and function.

(A) hCASMCs transduced to express Myc-tagged WT GATA-6 or the S290D or S290A mutants were subjected to ChIP-qPCR assay with Myc antibody and primer sets flanking the GATA elements or control regions (lacking GATA elements) in the promoters of human ACTA2 and MYH11 genes. Data were normalized to the expression of each mutant (n = 3 independent experiments; P = 0.05, indicated by brackets). (B) hCASMCs transfected with vector or WT, S290D, or S290A Myc–GATA-6 were immunoblotted for the indicated proteins. (C) Quantitation of (B) (n = 6 biological replicates; P < 0.05, indicated by brackets). (D) Proliferation in hCASMCs transduced as in (A) was monitored by MTT assay for 60 hours (n = 7 biological replicates; P < 0.01, indicated by brackets). Data were presented as the fold change of cell index compared to input. (E) hCASMCs transfected as in (B) were treated with cycloheximide for the indicated times before immunoblotting with Myc antibody to determine GATA-6 protein half-life. (F) Quantitation of (E) (n = 3 independent experiments; paired t test, P = 0.0625 between the WT GATA-6 and GATA-6 S290D groups, as well as between the GATA-6 S290D and S290A groups).

To determine whether the increased DNA binding of the GATA-6 S290D mutant leads to an increase in functional regulation of transcription, we transfected hCASMCs with plasmids encoding Myc–GATA-6 wild type, S290D, or S290A. Expression of the GATA-6 S290D mutant in hCASMCs increased the abundance of smooth muscle differentiation–specific contractile proteins, including SM-MHC, SM-α-actin, and SM-calponin by 1.5- to 2.0-fold compared to untransfected hCASMCs (Fig. 4, B and C). The S290D mutant was about twice as potent as GATA-6 wild type and nearly four times as potent as S290A (Fig. 4C). Expression of the S290D mutant also increased the abundance of the CDK inhibitor proteins p21cip and p27kip (Fig. 4, B and C). Lentiviral overexpression of GATA-6 S290D inhibited hCASMC proliferation to a greater extent than did wild-type or S290A GATA-6 (Fig. 4D), consistent with the greater increase in the abundance of p21cip and p27kip in S290D-expressing cells (Fig. 4, B and C). Collectively, these data support that phosphorylation of GATA-6 at Ser290 is an activating modification that increases GATA-6 DNA binding, prodifferentiation, and antiproliferative activities in VSMCs.

To determine whether phosphorylation at Ser290 might mediate the rapamycin-induced GATA-6 protein stabilization (as in Fig. 2D), we determined the protein half-life of wild-type or mutant GATA-6 expressed at low amounts in cycloheximide-treated hCASMCs. The S290D mutant had a longer protein half-life than wild-type or S290A GATA-6 (Fig. 4, E and F), suggesting that phosphorylation of GATA-6 at Ser290 also enhances the protein stability. It is likely that this posttranslational modification enhances GATA-6 functional activity by promoting both its protein stabilization and DNA binding.

GATA-6 expression attenuates neointima formation in the mouse femoral artery wire injury model

We have previously shown that Akt2 is a key effector in rapamycin-induced VSMC differentiation (10). To support that Akt2 mediates this effect through phosphorylation of GATA-6 in vivo, we determined the effects of Akt2 loss of function and GATA-6 phosphorylation site mutants in the mouse femoral artery wire injury model. Inserting a guide wire into the mouse femoral artery models the endothelial denudation and mechanical injury that leads to intimal hyperplasia in humans after angioplasty (18) (fig. S13). Wild-type C57BL/6 mice exhibited only modest neointima formation after this procedure (Fig. 5A). Consistent with a prodifferentiation function for Akt2 in VSMCs (10), Akt2 knockout mice exhibited more severe intimal hyperplasia than did wild-type mice as indicated by an ~80% increase in the intima/media area ratio (Fig. 5, A and B). Because GATA-6 overexpression rescues neointimal formation in rats (11), we next determined the effects of local overexpression of wild-type or mutant forms of GATA-6 in Akt2−/− mice. Control lentivirus or virus encoding Myc-tagged GATA-6 wild type, S290D, or S290A was applied in a pluronic gel to the adventitial side of the femoral artery at the time of wire injury. Overexpression of wild-type or mutant forms of GATA-6 significantly attenuated intimal hyperplasia in the Akt2−/− mouse compared to control virus; however, the mutants differed in potency (Fig. 5, C and D). The S290D mutant more potently inhibited intimal hyperplasia than did wild-type GATA-6, and the S290A mutant conferred only a partial rescue compared to control virus (Fig. 5, C and D). The trends in the effects of the GATA-6 constructs on intimal hyperplasia (Fig. 5, C and D) and proliferation (as measured by Ki67 staining in medial and neointimal cells) (Fig. 5E and fig. S14) were similar. Proliferation was increased to a greater extent in Akt2−/− arteries compared to the wild-type arteries, an effect that was inhibited by overexpression of wild-type GATA-6. The S290D mutant was significantly more potent than wild-type GATA-6 in inhibiting proliferation in Akt2−/− arteries, whereas the S290A mutant only partially rescued this phenotype (Fig. 5E and fig. S14). These findings were consistent with the in vitro efficacy of these mutants in their regulation of DNA binding, expression of contractile proteins, and proliferation in VSMCs (Fig. 4, A to D).

Fig. 5 GATA-6 expression attenuates neointima formation in the mouse femoral artery wire injury model.

(A) Mice were subjected to wire injury in the left femoral artery to denudate the endothelial layer and induce neointima formation. Neointima formation was analyzed 21 days after injury by light microscopy. (B) The neointima area and intima/media ratio in (A) were quantitated (n = 11 mice per group; *P < 0.05, **P < 0.01). (C) WT or Akt2 knockout mice received femoral artery injury as in (A). Immediately after injury, lentivirus encoding WT, S290D, or S290A mutant GATA-6 cDNA or control virus was loaded onto the injured artery. Neointima formation was analyzed as above. (D) Quantitation of intima/media ratio in (C) (n = 4 mice in each group; *P < 0.05). (E) Cryosections from (A) were subjected to Ki67 immunohistochemistry. The data are presented as the percentage of Ki67-positive cells of the total cells in the media and intima areas. Five different sections from each of two mice per treatment condition were quantitated (*P < 0.05, **P < 0.01). Scale bar, 50 μm. Arrowheads denote maximal neointimal thickness in each section.

Immunostaining for the Myc epitope tag indicates that the overexpressed proteins were expressed at similar amounts at 6 days after injury, and that whereas some adventitial expression was detected, the overexpressed GATA-6 proteins were abundant in the medial layer (fig. S15). The abundance of contractile proteins decreases after vascular injury (2). Notably, immunostaining for SM-α-actin colocalized with the overexpressed wild-type and mutant GATA-6 at 6 days after injury in Akt2−/− mice (fig. S15). Quantitation of these images indicates that the intensity of SM-α-actin staining per cell in the medial layer after injury correlates with the efficacy of the GATA-6 constructs in increasing contractile protein abundance in vitro (S290D > wild type > S290A) (fig. S15 and Fig. 4, B and C). These studies confirm the role of Akt2 in the modulation of VSMC phenotype in vivo as well as the effects of the Ser290 phosphorylation site on GATA-6 function in an in vivo model of intimal hyperplasia.


Akt2 is specifically activated in response to mTORC1 inhibition and promotes VSMC differentiation in vitro (10, 15). We now report that loss of Akt2 markedly exacerbates intimal hyperplasia, and that GATA-6 may be the substrate through which Akt2 promotes VSMC differentiation in vitro and in vivo (fig. S16). Furthermore, local vascular overexpression of GATA-6, particularly the S290D phospho-mimetic mutant, rescues the Akt2−/− injury phenotype. These findings provide insights into the therapeutic mechanisms of mTORC1 inhibitors as well as an important role of Akt2 in the regulation of VSMC phenotype.

Akt isoform–specific regulation of GATA-6

Although the highly homologous Akt isoforms share many common substrates (19), the differing phenotypes of the Akt1 and Akt2 knockout mice suggest that unique substrates exist. We have identified GATA-6 as a potential Akt2-specific substrate in VSMCs. In vivo, Akt1−/− mice display reduced cell and organism size, whereas Akt2−/− mice are prone to diabetes (20, 21). However, the double Akt1/Akt2 knockout dies shortly after birth, suggesting that there is also considerable functional redundancy between these isoforms (22). We now report that loss of Akt2 led to severe intimal hyperplasia after injury, and that overexpression of GATA-6 could rescue this phenotype. This is consistent with our previous in vitro findings that Akt2, but not Akt1, is necessary for rapamycin-induced VSMC differentiation, and that overexpression of Akt2, but not Akt1, is sufficient to promote VSMC differentiation (10). Although it is possible that Akt2 regulation of GATA-6 could be indirect, we demonstrate that Akt2 is required for rapamycin-induced phosphorylation of GATA-6 in vitro and in vivo, and contributes to the prodifferentiation and antiproliferative effects of rapamycin on SMCs. Akt1 stimulates different responses in VSMCs, namely, migration and proliferation (23). Thus, Akt1 and Akt2 appear not only to be nonredundant but also to serve opposing functions in VSMCs. There is precedence for opposing roles for Akt isoforms in a cell type–specific manner. For example, Akt1 inhibits but Akt2 promotes migration in breast cancer cells (2426), whereas Akt2 inhibits and Akt1 promotes migration in mouse embryonic fibroblasts (MEFs) (27). It is likely that these cell type differences arise because of differences in the abundance and localization of both Akt isoforms and substrates, some of which may be cell type–specific. This has important implications for the therapeutic potential of Akt isoform–specific inhibitors.

The Akt linker domain has been implicated in the isoform-specific regulation of migration in MEFs (17, 27). We determined that phosphorylation of GATA-6 may require the linker domain of Akt2, the region of greatest divergence between Akt isoforms that could potentially mediate a specific interaction between Akt2 and GATA-6. Whereas many Akt substrates are phosphorylated by all Akt isoforms in kinase assays in vitro, Akt2 appeared to be specific for GATA-6 in vitro and in intact cells. Similarly, Akt2, but not Akt1, regulates FoxO4 in response to mTORC1 inhibition in VSMCs (15).

Mechanisms and consequences of mTORC1 regulation of GATA-6

The important role of GATA-6 in intimal hyperplasia (11), together with intriguing data indicating that the TOR pathway could regulate GATA-family transcription factors in yeast (28), prompted us to investigate mTORC1-mediated regulation of GATA-6 in VSMCs. We found that whereas mTORC1 inhibition similarly activates GATA-6, the mechanisms in human VSMC and yeast appear to be distinct. In yeast, the GATA factor Gln3 is activated rapidly, translocating to the nucleus within 10 min of TOR inhibition (28), which requires dephosphorylation of Gln3 by the phosphatase Tap42. We found that GATA-6 activation in human VSMCs is slower, possibly due to the requirement for feedback activation of Akt2. Whereas it is highly probable that GATA-6 contains additional phosphorylation sites, we found that mutation of Ser290 has marked effects on protein function, especially in vivo. We note that trace 33P incorporation was detected in the in vitro kinase reaction with Akt2 and the GST-S290A substrate, suggesting that Ser291, the only other proximal serine within this GATA-6 sequence, could also be phosphorylated, but most likely only when the primary site, Ser290, is mutated. This scenario can indeed occur when there are multiple serine residues neighboring a basophilic site (29). Our MS analysis did not detect phosphorylation of Ser291 in the wild-type GATA-6 peptide. We detected a modest immunoreactivity with the RXRXXS*/T* phospho-substrate antibody in cells overexpressing GATA-6 S290A and Akt (fig. S10). This immunoreactivity could represent detection of phosphorylation at Ser291, or perhaps another basophilic site elsewhere in the protein. Although the contribution of other potential phosphorylation sites is not yet known, our data suggest that phosphorylation of Ser290 affects GATA-6 protein function and that Akt2 could be the kinase that phosphorylates this site.

We found that phosphorylation of GATA-6 at Ser290 affected multiple parameters that collectively led to increased GATA-6 activity: increased DNA binding, perhaps due to the location of this phosphorylation site between the two zinc finger domains, and increased stability. The S290D mutant also increased the expression of GATA-6 target genes, including those encoding contractile proteins and inhibitors of proliferation, in cultured cells. When locally overexpressed at the site of vascular injury, the wild type, S290D, and S290A forms of GATA-6 all diminished intimal hyperplasia and proliferation in Akt2−/− vessels. However, the S290D mutant was more effective than wild type, and both forms were substantially more effective than the S290A mutant. These effects are consistent with the differences in protein half-life observed in vitro. Whereas differences in antiproliferative activity between the mutants were not significant in vitro, likely due to the relatively short time point assessed, the phospho-mimetic mutant was the most potent inhibitor of proliferation in vivo, followed by wild type, then the S290A mutant. Our results suggest that inhibition of proliferation plays a major role in the inhibitory effect of GATA-6 on intimal hyperplasia. However, the increase in contractile protein abundance colocalized with the overexpressed GATA-6 proteins and also correlated with phosphorylation status, suggesting that promotion of VSMC differentiation may also contribute to this effect.

Role of GATA-6 in VSMC phenotype

Study of GATA-6 in vivo has been hindered by the very early (E6.5) embryonic lethality of the global knockout (30) and severe developmental defects in the smooth muscle– or neural crest–specific knockouts (31). GATA-6 exerts potent antiproliferative effects in VSMCs (12) and mesangial cells (32). Its abundance is decreased upon vascular injury, and local GATA-6 overexpression can rescue injury-induced intimal hyperplasia in vivo (11). Consistent with these findings, our study also supports a prodifferentiation, antiproliferative, and antihyperplastic role for GATA-6. Others have reported that the transcription factor CHF1/Hey2, implicated in intimal hyperplasia, directly interacts with and represses GATA-6 (33). In contrast, a microarray using an engineered dominant negative GATA-6 fusion protein suggested that GATA-6 regulates genes associated with the dedifferentiated VSMC phenotype (34).

The ability of VSMCs to reversibly differentiate has been intensely investigated. The current paradigm has focused on the interaction of the potent prodifferentiation transcriptional coactivator myocardin with serum response factor at CArG elements in smooth muscle promoters (1, 35). Several studies raise the possibility that GATA-6 may function in concert with myocardin to promote contractile gene expression in VSMCs. GATA-6 may have differential effects on contractile gene promoters, depending on the spacing between CArG and GATA elements in the promoters, resulting in GATA-6 inhibition of or synergy with myocardin: GATA-6 represses the promoter activity of the CArG-dependent smooth muscle gene telokin but induces the expression of SM-MHC, MLCK (myosin light chain kinase), and SM22α (36). Similarly, the closely related GATA-4 can stimulate or suppress myocardin activity in a gene-specific manner (37). We found that knockdown of GATA-6 alone reduced contractile protein abundance by up to 60% below baseline amounts. The magnitude of this effect is similar to the degree of inhibition observed with myocardin knockdown (38, 39). Because the major VSMC genes assessed in our study, including SM-MHC, calponin, and SM-α-actin, are regulated by CArG elements (1), our data suggest that in the context of mTORC1 inhibition, GATA-6 may cooperate with myocardin to regulate these promoters. This will be an important area for further study.

mTORC1-regulated transcription

Our work reveals that the mTORC1 pathway exerts tissue-specific transcriptional control, a finding with potential therapeutic relevance. The reversible differentiation of smooth muscle is transcriptionally regulated through incompletely understood mechanisms. We now implicate regulation of GATA-6 by mTORC1 and potentially Akt2 as a key mechanism underlying the efficacy of rapamycin in promoting VSMC differentiation. mTORC1 coordinates growth and metabolism through transcriptional mechanisms, including regulation of HIF1α (hypoxia-inducible factor 1α) and SREBP1c (sterol regulatory element–binding protein 1c) (40). Whereas knockout mouse studies have revealed specific mTORC1-regulated processes in different tissue types including the liver and skeletal muscle (5), little is known about cell type–specific transcriptional control downstream of mTORC1. mTORC1 is required for adipocyte differentiation through regulation of PGC1α [peroxisome proliferator–activated receptor γ (PPARγ) coactivator 1α] and PPARγ (5, 41, 42). We have shown that mTORC1 inhibition promotes VSMC differentiation (9, 10, 15). Together, these findings suggest that distinct cell type–specific transcription factor targets may be regulated by the mTORC1 pathway in different tissues.

Therapeutic relevance

The phenotypic modulation of vascular smooth muscle is critical not only for normal growth and healing in blood vessels but also in pathologies including atherosclerosis, restenosis, hypertension, and transplant arteriosclerosis. Although rapamycin is highly efficacious as a stent therapeutic, it cannot be used systemically for long-term treatment or prevention of vascular disease because the high doses required are associated with adverse effects including impaired wound healing, leukopenia, and thrombocytopenia (43). We showed that systemic rapamycin treatment increased the abundance of SM-MHC and GATA-6 in healthy, mature wild-type aorta, and that overexpression of GATA-6 with the S290D mutant rescued severe intimal hyperplasia. Elucidating the smooth muscle–specific signaling and downstream targets of mTORC1 inhibition such as Akt2 and GATA-6 may give rise to new therapeutic strategies for cardiovascular diseases.


Cell culture

hCASMCs (Cascade Biologics or Lonza) (passages 3 to 6) were propagated in M199 medium with 10% fetal bovine serum (FBS) as described (9, 10). SMCs were cultured with 2.5% FBS for 16 to 24 hours before drug treatment. [hCASMCs proliferate slowly and do not spontaneously differentiate under these conditions (9, 10).] Rapamycin (20 nM) or ethanol vehicle was added at the indicated time points, up to 48 hours. Human VSMCs from peripheral vessels were prepared by explant method (9). Vehicle was added for the maximum duration of treatment.

Cell lysis and Western blotting

Cells were washed with phosphate-buffered saline (PBS) and harvested in radioimmunoprecipitation (RIPA) assay buffer with protease and phosphatase inhibitors (Roche). Equal amounts of protein per lane were sep????arated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE), transferred onto a nitrocellulose membrane, and immunoblotted with primary antibodies against SM2-MHC, SM-α-actin, calponin-1 (Sigma), GATA-6 (Abcam), GAPDH (glyceraldehyde-3-phosphate dehydrogenase), IRS-1, Akt1, Akt2, pan-Akt, Myc tag, phospho-Akt (Ser473), phospho-S6 (Ser240/244), CREB, phospho-CREB (Ser133), phospho-(Ser/Thr) Akt substrate (RXRXXS*/T* motif), phospho-(Ser/Thr) PKA substrate (RRXS*/T* motif) (Cell Signaling), β-tubulin, p21cip, p27kip, DNA polymerase II (Santa Cruz Biotechnology), or hemagglutinin (HA) epitope and horseradish peroxidase–conjugated secondary antibodies (Pierce). Blots were developed using enhanced chemiluminescence reagents (Pierce), and signals were quantitated using ImageJ.

Quantitative RT-PCR

RNA was isolated from cells using the Qiagen RNeasy kit, and 1 μg of RNA was reverse-transcribed using an iScript cDNA synthesis kit (Bio-Rad). qPCR was performed using primers for GATA-6 (sense: 5′-GTGCCTTCATCACGGCGGCT, antisense: 5′-CACACGGGTTCACCCTCGGC) and MYH11 (Qiagen) using the ΔΔCT method. GATA-6 and MYH11 products were normalized to β-actin (ACTB) (sense: 5′-GACAGGATGCAGAAGGAGA, antisense: 5′-CCACATCTGCTGGAAGGTGG) product, a housekeeping gene control.

Transient transfection of siRNA and plasmid

Transient transfection of siRNA in hCASMCs was performed using Lipofectamine RNAiMAX (Life Technologies). Cells were plated in six-well plates or 10-cm dishes the day before transfection to reach a confluency of 70 to 80%. Transfection reagent and siRNA were mixed in Opti-MEM (Gibco) according to the manufacturer’s protocol and added to the cells cultured in Opti-MEM. M199 medium with 20% FBS was added to the cells after 6 hours. After 24 hours, the medium was changed to M199 medium with 2.5% FBS, and the cells were cultured overnight before drug treatment [vehicle or 20 nM rapamycin (Calbiochem)]. GATA-6 and nonsilencing siRNAs (siControl) were purchased from Dharmacon, and siRNAs for Akt1 and Akt2 were purchased from Qiagen. For overexpression, 8 μg of plasmid DNA was transfected into 1 million to 1.5 million hCASMCs using Nucleofector (Lonza) (10). Transient transfection of plasmid DNA to HEK293 cells was done using X-tremeGENE 9 (Roche).

Lentiviral expression of GATA-6 mutants

pLVX-IRES-puro plasmids containing complementary DNAs (cDNAs) encoding Myc-tagged GATA-6 mutants were cotransfected with packaging vectors (pMDLg/pRRE, pRSV-Rev, pMD2.G) into HEK293 cells. Lentivirus secreted into the media was isolated by ultracentrifugation, resuspended in PBS, and used to infect hCASMCs in the presence of polybrene (8 μg/ml; Santa Cruz Biotechnology). Seventy-two hours after infection, transduced cells were selected in puromycin (0.75 μg/ml) (InvivoGen) for 48 hours.

Luciferase reporter assay

VSMCs were transfected with a firefly luciferase reporter construct driven by a GATA element from the rat MYH11 promoter (14) and thymidine kinase–Renilla–luciferase construct (44) (transfection control). pcDNA vector or GATA-6 expression plasmids were cotransfected where indicated. Twenty-four hours after transfection, cells were treated with rapamycin. In the mutant MYH11 (SM-MHC) promoter, the GATA element GTAATCATCACA was mutated to GTAAcagcCACA using KOD Xtreme Hot Start DNA Polymerase (Novagen). Luciferase assays were performed using the Dual-Luciferase System (Promega).


hCASMCs were cultured on glass coverslips, treated, washed with PBS, fixed in ice-cold 4% paraformaldehyde (PFA), blocked with 10% FBS in PBS with 0.1% Triton X-100 (PBS-T), and incubated with anti–GATA-6 (Santa Cruz Biotechnology, clone H92) or anti–Myc epitope antibody in PBS-T at 4°C overnight. Cells were then incubated with Alexa Fluor 488– or Alexa Fluor 594–conjugated secondary antibodies (Molecular Probes Inc.) in PBS-T for 1 hour, and 4′,6-diamidino-2-phenylindole (DAPI) staining was performed before mounting and analysis using a spinning disc confocal microscope (Perkin-Elmer).

Nuclear-cytosolic fractionation

hCASMC cytosolic and nuclear fractions were separated using the Subcellular Protein Fractionation Kit (Pierce).


Cells were harvested in 1× cell lysis buffer (Cell Signaling) containing protease and phosphatase inhibitors. An equal amount of protein lysate per sample was incubated overnight with antibody to GATA-6 (R&D Systems Inc.) and then with protein G–Sepharose beads (Amersham Biosciences/GE Healthcare) at 4°C. Bound protein was washed, eluted by heating in 1× SDS sample buffer, and subjected to Western blotting.

Chromatin immunoprecipitation–quantitative polymerase chain reaction

Four confluent 10-cm dishes of hCASMCs (~4 × 106 cells) were treated with 1% formaldehyde to cross-link protein to DNA and were neutralized with glycine. Cells were scraped in cold PBS and subjected to ChIP assay using the SimpleChIP Plus Enzymatic Chromatin IP Kit (Cell Signaling) with antibody to GATA-6 (R&D Systems Inc.) or Myc tag (Cell Signaling). Precipitated DNA was analyzed by qPCR with the following primer sets: ACTA2 promoter, sense: 5′-AGGCCTCCGGCCACCCAGATTA and antisense: 5′-GCCTGCTCTCCTCCCACTTGCT; MYH11 promoter, sense: 5′-GACACAGAGACCAGAGACAAAG and antisense: 5′-TCACTCTGATCATTGCTGTCTC. The amount of coprecipitated DNA was normalized to the amount of PCR product amplified from input samples reserved before immunoprecipitation. The negative control primer sets were designed to amplify regions in the ACTA2 or MYH11 promoters that do not contain GATA elements: ACTA2, sense: 5′-CCAGGCAGTGTTCTAGGTGC and antisense: 5′-TCCCCTCTGTGCCTTAGCTT; MYH11, sense: 5′-ATTCCTGTTCCACCACTGCT and antisense: 5′-TCCCTCCCTACCCCCATTTT. For ChIP assays showing differences in DNA binding between GATA-6 Ser290 mutants, the qPCR data were normalized to expression of Myc–GATA-6, which was assayed in parallel by Western blotting in the same lysates.

In vitro kinase assay and peptide synthesis

GATA-6 overexpressed in HEK293 cells was immunoprecipitated and bound to protein G–Sepharose beads. Equal amounts of immunoprecipitated beads were mixed with recombinant active Akt1 or Akt2 (Abcam) protein or PBS or bovine serum albumin (BSA) as controls and incubated at 30°C for 30 min with 1× kinase buffer (Cell Signaling). 2× SDS sample buffer was added, and samples were heated at 95°C for 5 min and subjected to Western blotting with anti–phospho-RXRXXS*/T* antibody. In parallel experiments, recombinant GSK3β protein (Cell Signaling) was used as a substrate to verify the enzymatic activity of the recombinant kinases. For radioactive experiments, [33P]ATP (6 μCi per reaction) was included in the kinase assay using GST-GATA-6 substrates as described below or recombinant GSK3β. Reactions were subjected to SDS-PAGE, Coomassie staining, gel drying (Promega kit), and autoradiography.

To generate the GST-tagged GATA-6 peptide, oligos encoding amino acids 277 to 296 of human GATA-6 protein (LSRPLIKPQKRVPSSRRLGL) and the S290A mutant (LSRPLIKPQKRVPASRRLGL) were synthesized and subcloned into the pGEX-4T-1 vector. The recombinant plasmid was transformed to the BL21 (DE3) bacteria strain and induced by isopropyl-β-d-thiogalactopyranoside to produce GST alone or the GST fusion protein with the 20–amino acid (aa) peptide at the C-terminal end (named GST-GATA-6_20aa_wild-type or GST-GATA-6_20aa_S290A). The proteins were then purified with glutathione–Sepharose 4B beads and eluted with buffer containing glutathione.

MS analysis

Peptides containing amino acids 281 to 292 of human wild-type GATA-6 protein (LIKPQKRVPSSR) or a mutant with a phospho-serine substituted at the Ser290 position were synthesized by GenScript. The peptides were desalted using tC18 solid-phase extraction Sep-Pak columns (Waters). The unphosphorylated peptide (~1 pmol) was subjected to an in vitro kinase assay using purified Akt2 or no kinase as described above. The peptides from the reaction were concentrated and desalted using Stage Tips (45). Dried peptides were resuspended in 2.5% acetonitrile, 2.5% formic acid and subjected to liquid chromatography–tandem MS using a SIL-201 Prominence autosampler (Shimadzu), a Finnigan Surveyor MS Pump Plus (Thermo Fisher), and a linear ion trap–Orbitrap hybrid mass spectrometer (Thermo Fisher) essentially as described (46). In addition to two data-dependent MS2 scans per cycle, we collected MS2 scans targeted on the triply charged precursor phosphopeptide ion. Only after collecting data from the peptides extracted from the in vitro kinase assays was the synthetic phosphopeptide analyzed.

MTT assay

hCASMCs transfected with siRNA or puromycin-selected hCASMCs expressing GATA-6 lentiviral constructs were seeded at 6000 cells per well (seven or eight replicate wells per condition) in 96-well plates, starved, and then cultured in 10% FBS for 60 or 72 hours (as indicated), and analyzed with the CellTiter 96 Non-Radioactive Cell Proliferation Assay (MTT) (Promega). Cells seeded at the start of the experiment but harvested just after attachment served as an input control. The end-point reading was normalized to that of input control as fold change.

Animal experiments

All experiments were approved by the Institutional Animal Care and Use Committees of Yale University. For in vivo Gata-6 protein level or phosphorylation induction by rapamycin, wild-type C57BL/6 or Akt2−/− mice were injected intraperitoneally with vehicle (dimethylacetamide diluted 1:24 in 10% polyethylene glycol and 17% polyoxyethylene sorbitan monooleate) or rapamycin (1 mg/kg). Mice were sacrificed at the indicated time points for tissue harvest. Tissues were homogenized using TissueLyser II (Qiagen, 30 Hz, 3 min twice) in 1× cell lysis buffer (Cell Signaling) with protease and phosphatase inhibitors. The supernatant after centrifugation was collected for the following immunoprecipitation and/or Western blot analysis. For femoral wire injury model, 8- to 12-week-old C57BL/6 (wild-type) or Akt2−/− mice were anesthetized with ketamine (100 mg/kg) and xylazine (10 mg/kg), and femoral artery injury was performed as previously described (47). For GATA-6 overexpression, control virus or lentivirus carrying GATA-6 wild type, S290D, or S290A (1 × 108 plaque-forming units) was delivered to the adventitial side of the vessel in a 50-μl mixture containing 15 μl of lentivirus and 35 μl of pluronic-127 gel immediately after injury. The animals were sacrificed 6 or 21 days after surgery by cardiac perfusion with warm 4% PFA, and the injured and uninjured contralateral femoral arteries were collected. Fixed femoral arteries were embedded in optimum cutting temperature compound (Tissue-Tek). For 21-day injury samples, cryosections (5 μm) were obtained for elastic van Gieson staining, and morphometric analysis was performed on 10 sections per artery. For immunohistochemistry, sections were blocked with 2% BSA in 0.2% Triton X-100 (PBS-T) and then incubated with rabbit anti–Myc tag antibody and Cy3-conjugated anti–SM-α-actin at 4°C overnight. Cells were then incubated with Alexa Fluor 594–conjugated anti-rabbit secondary antibody in PBS-T for 1 hour, and DAPI staining was performed before mounting and analysis using a fluorescence microscope (Nikon Eclipse 80i).

Statistical analysis

All values are presented as mean ± SEM. For most data, comparisons were analyzed with the Mann-Whitney U test. For protein decay assays, a paired t test was used for comparison between two groups. P values were one-tailed. Analyses were performed with Prism 6.0 software (GraphPad).


Fig. S1. GATA-6 is required for rapamycin-induced VSMC differentiation.

Fig. S2. Rapamycin and GATA-6 transactivate a GATA-dependent MYH11 promoter reporter.

Fig. S3. Rapamycin treatment does not affect GATA-6 mRNA.

Fig. S4. GATA-6 accumulates in the nucleus after rapamycin treatment.

Fig. S5. Rapamycin promotes the nuclear accumulation of GATA-6.

Fig. S6. Rapamycin treatment stabilizes GATA-6 protein in the presence of cycloheximide.

Fig. S7. Phosphorylation of GATA-6 in cells is increased by coexpression of Akt2, but not Akt1.

Fig. S8. Inhibition of PKA activity does not block the rapamycin-induced phosphorylation of GATA-6.

Fig. S9. MS analysis shows that Akt2 could phosphorylate Ser290 in GATA-6.

Fig. S10. The phosphorylation of wild-type but not S290A mutant GATA-6 is increased by Akt2 coexpression in cells.

Fig. S11. The linker domain confers GATA-6 substrate specificity to Akt2.

Fig. S12. Myc–GATA-6 is expressed in nearly 100% of cells after lentivirus infection and puromycin selection.

Fig. S13. Schematic of the femoral artery wire injury model.

Fig. S14. GATA-6 inhibits SMC proliferation after injury in vivo.

Fig. S15. Lentivirally delivered Myc-tagged GATA-6 cDNAs in wire-injured femoral artery (early time point).

Fig. S16. Schematic of rapamycin signaling to Akt2 and GATA-6 in VSMCs.


Acknowledgments: We thank K. Walsh for the GATA-6 wild type and ΔZF plasmids and K. Hasegawa for the MYH11 promoter reporter plasmids. We thank B. Turk for helpful discussions and H.Y.-F. Chiu and N. Garg for technical contributions to this project.Funding: Supported by grants from the NIH National Heart, Lung, and Blood Institute to K.A.M. (RHL091013), J.H. (HL074190), J.Y. (R56HL117064), and R.J.P. (HL76612) and from the American Heart Association to J.Y. (SDG0930157N) and K.A.M. (SDG0230356N). B.A.B. was supported by the Vermont Genetics Network through NIH grant P20 RR16462 from the INBRE [Institutional Development Award (IDeA) Networks of Biomedical Research Excellence] program of the National Institute of General Medical Sciences. A.K.L. was supported by NIH Training Grant GM007324.Author contributions: Y.X. and B.L.M. performed experiments, interpreted results, and wrote the manuscript; Y.J., M.D., K.M.F., R.J.W., A.M., S.G., D.F.T, A.K.L., B.A.B., and J.Y. performed and interpreted experiments; M.J.B., W.C.S., E.M.R., R.J.P., and J.H. interpreted results and edited the manuscript; H.Z. and L.H. performed statistical analysis; K.A.M. supervised the project, performed and interpreted experiments, and wrote the manuscript.Competing interests: The authors declare that they have no competing interests.
View Abstract

Navigate This Article