SIRT1 promotes endothelium-dependent vascular relaxation by activating endothelial nitric oxide synthase

Abstract

Reduced caloric intake decreases arterial blood pressure in healthy individuals and improves endothelium-dependent vasodilation in obese and overweight individuals. The SIRT1 protein deacetylase mediates many of the effects of calorie restriction (CR) on organismal lifespan and metabolic pathways. However, the role of SIRT1 in regulating endothelium-dependent vasomotor tone is not known. Here we show that SIRT1 promotes endothelium-dependent vasodilation by targeting endothelial nitric oxide synthase (eNOS) for deacetylation. SIRT1 and eNOS colocalize and coprecipitate in endothelial cells, and SIRT1 deacetylates eNOS, stimulating eNOS activity and increasing endothelial nitric oxide (NO). SIRT1-induced increase in endothelial NO is mediated through lysines 496 and 506 in the calmodulin-binding domain of eNOS. Inhibition of SIRT1 in the endothelium of arteries inhibits endothelium-dependent vasodilation and decreases bioavailable NO. Finally, CR of mice leads to deacetylation of eNOS. Our results demonstrate that SIRT1 plays a fundamental role in regulating endothelial NO and endothelium-dependent vascular tone by deacetylating eNOS. Furthermore, our results provide a possible molecular mechanism connecting the effects of CR on the endothelium and vascular tone to SIRT1-mediated deacetylation of eNOS.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.

Endothelial SIRT1 regulates endothelium-dependent vasodilation and bioavailable NO through NOS. (A) A replication-deficient adenovirus encoding the myc-tagged catalytically inactive dominant negative mutant of SIRT1 (AdSIRT1(H363Y)) (a and c) was used to inhibit endogenous SIRT1 in the endothelium of rat aortic rings ex vivo. Control rings were infected with an adenovirus encoding the inert E. coli lacZ gene (AdLacZ) (b and d). Rings were fixed and immunostained with anti-myc antibody 24 h after adenoviral infection. Photomicrographs of representative rings at ×100 (a and b) and ×400 (c and d) are shown. (B) Endothelium-dependent vasodilation was determined by measuring relaxation of rings preconstricted with phenylephrine to the vasodilator acetylcholine. AdSIRT1(H363Y) (■) and AdLacZ (◆) are shown. *, P < 0.001 compared with AdLacZ (n = 7). (C) Bioavailable NO was determined by measuring the difference in phenylephrine-induced vasoconstriction in the presence and absence of the NOS inhibitor l-NAME. AdSIRT1(H363Y) (■) and AdLacZ (◆) are shown. *, P < 0.01 compared with AdSIRT1(H363Y) (n = 7). (D) Endothelium-independent vasodilation was determined by measuring relaxation of rings preconstricted with phenylephrine to the NO donor sodium nitroprusside. AdSIRT1(H363Y) (△) and AdLacZ (■) (n = 4) are shown. (E) Rat aortic rings were infected ex vivo with AdLacZ (■ and □) or AdSIRT1 (△ and ▲). AdSIRT1 encodes wild-type SIRT1. Endothelium-dependent vasodilation, in the presence (□ and △) and absence (■ and ▲) of the NOS inhibitor l-NAME, was determined by measuring relaxation of rings preconstricted with phenylephrine, to the vasodilator acetylcholine. n = 4 rings in each group.

Fig. 2.

SIRT1 deacetylates eNOS in cells and in vitro, and calorie restriction is associated with deacetylation of eNOS in mice livers. (A) Acetylation of endogenous eNOS on lysine residues, with and without RNAi-mediated knockdown of endogenous SIRT1, and with and without treatment of cells with the SIRT1 inhibitor nicotinamide. Rat aortic endothelial cells were transfected with SIRT1 RNAi or control RNAi, or treated with nicotinamide. Immunoprecipitates of eNOS from whole-cell lysates were immunoblotted with anti-eNOS and anti-acetyllysine antibodies (top two panels). Immunoprecipitates by using nonimmune IgG were used as a control. Whole-cell lysates were immunoblotted with anti-eNOS, anti-SIRT1, and anti-β-actin antibodies (bottom three panels). (B) Acetylation of endogenous eNOS on lysine residues, with and without treatment with the SIRT1 activator resveratrol. Immunoprecipitated eNOS from whole-cell lysates of rat aortic endothelial cells, with and without treatment with resveratrol, was immunoblotted with anti-acetyllysine and anti-eNOS antibodies. The whole-cell lysates were also immunoblotted with anti-eNOS, anti-SIRT1, and anti-β-actin antibodies. (C) Incorporation of [3H]acetate into eNOS expressed in COS7 cells. Cells transfected with eNOS, with and without SIRT1 overexpression, or treated with nicotinamide, were incubated in medium containing sodium [3H]acetate. Immunoprecipitated eNOS from cell lysates was autoradiographed, and immunoblotted with anti-eNOS antibody. Whole-cell lysates were immunoblotted with anti-eNOS and anti-SIRT1 antibodies. (D) Deacetylation of acetylated eNOS by SIRT1 in vitro. Purified full-length GST-eNOS and GST-eNOS acetylated with the p300 acetyltransferase in the presence of the acetyl group donor acetyl-CoA (Ac-CoA) were incubated with active SIRT1 enzyme, in the presence or absence of the SIRT1 cofactor NAD. Reaction mixes were immunoblotted with anti-acetyllysine and anti-eNOS antibodies. (E) Deacetylation of eNOS in mice livers by CR. Male and female mice were fed ad libitum or a calorie-restricted diet (57.4 kcal/week) for 3 weeks. Immunoprecipitates of eNOS from liver homogenates were immunoblotted with anti-acetyllysine and anti-eNOS antibodies (top two panels), and homogenates immunoblotted with anti-SIRT1 and anti-β-actin antibodies (bottom two panels). The bar graph shows the calculated acetylated/total eNOS ratio, expressed relative to the mouse fed ad libitum in each gender group. The weight of the mice at the end of 3 weeks (as a percentage of starting weight) is shown. AL, ad libitum.

Fig. 3.

SIRT1 stimulates eNOS activity and increases endothelial NO. (A) NOS catalytic activity (conversion of arginine to citrulline) in lysates of COS7 cells expressing eNOS, with and without SIRT1 overexpression, was measured. Lysates were immunoblotted with anti-SIRT1 and anti-eNOS antibodies. *, P < 0.05 compared with eNOS without SIRT1. (B) NOS catalytic activity in vitro, in the absence and presence of p300, SIRT1, and NAD. Total and acetyllysine eNOS is shown at bottom. *, P < 0.05 compared with eNOS + p300. (C) Metabolites of NO (nitrite and nitrate) were measured in media of COS7 cells expressing eNOS, with and without overexpression of SIRT1, or inactive SIRT1 (H363Y). Whole-cell lysates were immunoblotted with anti-eNOS and anti-SIRT1 antibodies. Endogenous SIRT1 is not seen in the lysates because of underexposure of the blots. *, P < 0.05 compared with eNOS without SIRT1. (D) Nitrite and nitrate were measured in media of rat aortic endothelial cells that were transfected with SIRT1 RNAi or control RNAi, or treated with the SIRT1 activator resveratrol. *, P < 0.05 compared with control RNAi. **, P < 0.01 compared with control RNAi. #, P > 0.05 compared with SIRT1 RNAi. (E) Nitrite and nitrate were measured in COS7 cells expressing eNOS (WT), eNOS (K496R), or eNOS(K506R), with and without SIRT1 overexpression. *, P < 0.05 compared with eNOS (WT).

Fig. 4.

SIRT1 and eNOS localize and associate with each other in endothelial cells. (A) Double immunofluorescence for endogenous eNOS (red) and SIRT1 (green) in human umbilical vein endothelial cells. Fixed, permeabilized cells were stained with anti-eNOS antibody and Texas Red-labeled (red) secondary antibody, followed by staining with anti-SIRT1 antibody and FITC-labeled (green) secondary antibody (c–e). Control cells were stained with the Texas Red-labeled (a) and FITC-labeled (b) secondary antibodies without the anti-eNOS and anti-SIRT1 primary antibodies. Representative red, green, and merged images (×40) captured on a fluorescence microscope are shown. Arrow points to colocalization (yellow) of eNOS and SIRT1 in merged image. (B) Coimmunoprecipitation of endogenous eNOS and SIRT1 from rat aortic endothelial cells. Whole-cell lysates were immunoprecipitated with anti-eNOS or anti-SIRT1 antibodies. Immunoprecipitates were immunoblotted with anti-eNOS and anti-SIRT1 antibodies. Immunoprecipitates with nonimmune IgG were used as controls.