Insights into the arginine paradox: evidence against the importance of subcellular location of arginase and eNOS

Abstract

Reduced production of nitric oxide (NO) is one of the first indications of endothelial dysfunction and precedes overt cardiovascular disease. Increased expression of Arginase has been proposed as a mechanism to account for diminished NO production. Arginases consume l-arginine, the substrate for endothelial nitric oxide synthase (eNOS), and l-arginine depletion is thought to competitively reduce eNOS-derived NO. However, this simple relationship is complicated by the paradox that l-arginine concentrations in endothelial cells remain sufficiently high to support NO synthesis. One mechanism proposed to explain this is compartmentalization of intracellular l-arginine into distinct, poorly interchangeable pools. In the current study, we investigated this concept by targeting eNOS and Arginase to different intracellular locations within COS-7 cells and also BAEC. We found that supplemental l-arginine and l-citrulline dose-dependently increased NO production in a manner independent of the intracellular location of eNOS. Cytosolic arginase I and mitochondrial arginase II reduced eNOS activity equally regardless of where in the cell eNOS was expressed. Similarly, targeting arginase I to disparate regions of the cell did not differentially modify eNOS activity. Arginase-dependent suppression of eNOS activity was reversed by pharmacological inhibitors and absent in a catalytically inactive mutant. Arginase did not directly interact with eNOS, and the metabolic products of arginase or downstream enzymes did not contribute to eNOS inhibition. Cells expressing arginase had significantly lower levels of intracellular l-arginine and higher levels of ornithine. These results suggest that arginases inhibit eNOS activity by depletion of substrate and that the compartmentalization of l-arginine does not play a major role.

Keywords: arginase; endothelial nitric oxide synthase; l-arginine; nitric oxide; urea.

Fig. 1.

The intracellular location of endothelial nitric oxide synthase (eNOS) does not modify the ability of l-arginine (l-Arg) or l-citrulline to stimulate nitric oxide (NO) production. A: COS-7 cells were cotransfected with cDNAs encoding eNOS and control plasmid of the red fluorescent protein (RFP). Cells were washed in l-Arg-free medium for 30 min and incubated in the indicated concentrations of l-Arg for 24 h. NO levels were assessed using a NO analyzer and are reported as %NO release in the presence of DMEM. The relative expression of eNOS and heat shock protein 90 (Hsp90; loading control) was obtained by immunoblotting with anti-eNOS and anti-Hsp90. Results are presented as %control of 3 mM l-Arg. Results are presented as means ± SE; n = 6 experiments. B–F: COS-7 cells were transfected with a calcium-insensitive eNOS (CIeNOS) targeted per wild-type eNOS or to the cytosol, Golgi, plasma membrane, or mitochondria. Cells were incubated in l-Arg-free DMEM in the presence of the indicated concentrations of l-Arg or l-citrulline for 24 h, and NO release was determined by NO-specific chemiluminescence. Results are presented as means ± SE; n = 4. Nonparametric 1-way ANOVA; Kruskal-Wallis test with a Dunns multiple comparison of all columns within each substrate addition. *P < 0.05 vs. 0 μM l-Arg; #P < 0.05 vs. 0 μM l-citrulline.

Fig. 2.

Arginase (Arg) I and Arg II target different subcellular compartments but do not have an additive effect on NO release. A: COS-7 cells were cotransfected with plasmids encoding Arg I and Arg II. Cells were fractionated into cytosolic (Cyto) or mitochondrial (Mito) fractions by differential centrifugation. β-Tubulin was used as a marker for the Cyto compartment and cytochrome oxidase subunit 4 (CO₄) for mitochondria. The relative expression of Arg I and Arg II was confirmed by immunoblotting using anti-Arg I and anti-Arg II antibodies. Results are representative of 4 independent experiments. B and C: COS-7 cells were cotransfected with plasmids encoding wild-type eNOS or a CIeNOS construct and either a control gene (RFP) or Arg I/Arg II. NO production was measured by NO analyzer and is reported as %control (eNOS + RFP or CIeNOS + RFP). The expression of relevant proteins was determined by immunoblotting. Hsp90 was used as a loading control. D: titration of Arg I expression against a fixed concentration of eNOS. Numbers indicate that the ratio of transfected DNA and the expression of relevant proteins are shown below. Results are presented as means ± SE; n = 6 experiments. *P < 0.05 vs. control.

Fig. 3.

The location of eNOS does not affect the relative inhibition by Arg I vs. Arg II. COS-7 cells were cotransfected with different eNOS constructs targeted to different intracellular locations and either RFP (transfection control), Arg I, or Arg II. CI S17, CIeNOS localized to the Golgi (A); CI NLS, CIeNOS localized to the nucleus (B); CI Mito, CIeNOS localized to the inner mitochondrial membrane (C); CI TOM, CIeNOS localized to the outer mitochondrial membrane (D); CI PM, CIeNOS localized to the plasma membrane (E); CI G2A, CIeNOS localized to the cytosol (F). NO release was measured by NO analyzer and is presented as %RFP cotransfection control (top of graphs). The expression levels of relevant proteins were determined by immunoblotting (bottom). Hsp90 was used as a loading control. Results are represented as means ± SE; n = 6 experiments. *P < 0.05 vs. control.

Fig. 4.

Arg I constructs targeted to different locations reduce eNOS-derived NO equally despite differences in relative activity. A: COS-7 cells were transfected with Arg I constructs that target specific subcellular domains by inframe fusion of targeting sequences. PM, plasma membrane; Mito, inner mitochondria; NLS, nucleus; e1–e75 (the targeting sequence from eNOS) represents the plasma membrane and Golgi. NO levels were measured and presented as %control (RFP). Results are represented as means ± SE; n = 12 experiments. *P < 0.05 vs. RFP. B: arginase activities in cell lysates from COS-7 cells transfected with targeted Arg I constructs. The expression levels of Arg I protein and the loading control GAPDH were determined by immunoblotting (green and red, respectively; bottom). Results are means ± SE; n = 4. *P < 0.05 vs. wild type using nonparametric 1-way ANOVA and Kruskal-Wallis test with Dunn's multiple comparison.

Fig. 5.

Extracellular l-arginine promotes greater NO production but cannot overcome the Arg I-mediated reduction in eNOS activity. A: COS-7 cells were transfected with either CIeNOS + RFP (control) or CIeNOS + Arg I. Twenty-four hours after transfection, cells were incubated in medium containing l-arginine (0–3,000 μM). NO release was measured by NO analyzer and presented as %NO release from cells cotransfected with CIeNOS + RFP control at 3,000 μM. Expression levels of eNOS, Arg I, and Hsp90 (loading control) were determined by immunoblotting (bottom). B: expression levels of eNOS, the cationic amino acid transporter (CAT-1), argininosuccinate synthase (ASS), and argininosuccinate lyase (ASL) vs. Hsp90 (loading control) in untreated bovine aortic endothelial cells (BAEC) and COS-7. C: expression levels of eNOS in BAEC and COS-7 grown in l-arginine replete and free DMEM. D and E: COS-7 cells cotransfected with CIeNOS + RFP or CIeNOS + Arg I and incubated for 24 h in the concentration of l-arginine notated below the bar graph. Cells were isolated, and relative amounts (nmol) of l-arginine (D) and ornithine (E) determined. Results are presented as means ± SE; n = 6 experiments. *P < 0.05 vs. control.

Fig. 6.

Reduced eNOS activity in the presence of arginase is unrelated to cofactor suppression or direct interaction. A: COS-7 cells were transfected with eNOS + Arg I; eNOS and arginase (Flag epitope) were immunoprecipitated (IP) vs. a nonimmune control (IgG), and associated proteins were immunoblotted (IB). B: COS-7 cells transfected with CIeNOS + RFP or CIeNOS + Arg I were treated with or without the tetrahydrobiopterin (BH₄) precursor sepiapterin (Sep; 100 μM) and with or without N^G-nitro-l-arginine methyl ester (l-NAME; 1 mM) and NO levels determined by NO-specific chemiluminescence. Results are represented as means ± SE; n = 6 experiments. *P < 0.05 vs. CIeNOS + RFP; #P < 0.05 vs. CIeNOS + RFP + Sep.

Fig. 7.

Catalytic activity is necessary for arginase-mediated inhibition of eNOS-derived NO but not ornithine metabolism. A: COS-7 cells were transfected with CIeNOS and either wild-type Arg I (WT Arg I) or the catalytically inactive Arg I (D128G). Top: NO was measured by NO-specific chemiluminescence and is presented as %NO release from cells transduced with RFP control. Bottom: the expression of relevant proteins was determined by Western blotting. B: NO release from COS-7 cells cotransfected with CIeNOS and either RFP or Arg I in the presence and absence of the arginase inhibitor nor-N-hydroxy-l-arginine (NOHA; 1 mM). C: NO release in the presence and absence of the ornithine decarboxylase (ODC) inhibitor difluoromethylornithine (DFMO; 50 μM). Results are represented as means ± SE; n = 6–12 experiments. *P < 0.05 vs. control (RFP); #P < 0.05, CIeNOS + Arg I vs. CIeNOS + Arg I + Nor-NOHA. D and E: NO release was measured in COS-7 cells expressing eNOS in the presence and absence of urea (0.1–10 mM) and l-ornithine (0.3–10 mM). There is no significant difference between NO release in urea or ornithine addition as determined by nonparametric 1-way ANOVA and Kruskal-Wallis test, with a Dunn's multiple comparison of all columns within urea addition or ornithine addition. Results are represented as means ± SE; n = 4. NS, not significant.

Fig. 8.

Transduction of endothelial cells with plasma membrane-targeted Arg I elicits equal eNOS inhibition compared with WT Arg I. BAEC were transduced with 20 MOI of RFP, catalytically inactive Arg I (D128G), WT Arg I, and PM-targeted arginase (PM Arg I) adenoviruses. A, top: NO release from BAEC was measured by NO-specific chemiluminescence and presented as %NO release from cells transduced with RFP control. A, bottom: expression of relevant proteins was determined by immunoblotting. Hsp90 was used as a loading control. Nonparametric 1-way ANOVA and Kruskal-Wallis test with Dunn's multiple comparison. Results are represented as means ± SE; n = 6 experiments. *P < 0.05 vs. RFP. B: confocal microscopy of BAEC transduced with PM Arg I shows the predominant PM location of targeted arginase. Green indicates the FLAG tag (PM Arg I), and blue indicates DAPI (nucleus). Results are representative of more than 3 independent experiments.

Fig. 9.

Extracellular l-arginine stimulates NO production in BAEC but does not fully reverse the effect of Arg I. BAEC were transduced with adenovirus encoding green florescent protein (GFP; control), the catalytically inactive arginase (D128G), or WT Arg I at 20 MOI. Twenty-four hours after transduction, cells were incubated in medium containing l-arginine (0, 300, or 3,000 μM). NO release was measured via NO analyzer and presented as %NO release from control (GFP; 3,000 μM) cells. Expression levels of eNOS, Arg I, FLAG (D128G and Arg I virus tag), and Hsp90 (loading control) were determined by immunoblotting (bottom). Results are represented as means ± SE; n = 8 experiments. *P < 0.05 vs. D128G and GFP controls at 3,000 μM.