Nitrosative stress uncovers potent β₂-adrenergic receptor-linked vasodilation further enhanced by blockade of clathrin endosome formation

Abstract

This study investigated the effect of sodium nitroprusside (SNP) preexposure on vasodilation via the β-adrenergic receptor (BAR) system. SNP was used as a nitrosative/oxidative proinflammatory insult. Small arterioles were visualized by intravital microscopy in the hamster cheek pouch tissue (isoflurane, n = 45). Control dilation to isoproterenol (EC₅₀: 10⁻⁷ mol/l) became biphasic as a function of concentration after 2 min of exposure to SNP (10⁻⁴ M), with increased potency at picomolar dilation uncovered and decreased efficacy at the micromolar dilation. Control dilation to curcumin was likewise altered after SNP, but only the increased potency at a low dose was uncovered, whereas micromolar dilation was eliminated. The picomolar dilations were blocked by the potent BAR-2 inverse agonist carazolol (10⁻⁹ mol/l). Dynamin inhibition with dynasore mimicked this effect, suggesting that SNP preexposure prevented BAR agonist internalization. Using HeLa cells transfected with BAR-2 tagged with monomeric red fluorescent protein, exposure to 10⁻⁸−10⁻⁶ mol/l curcumin resulted in internalization and colocalization of BAR-2 and curcumin (FRET) that was prevented by oxidative stress (10⁻³ mol/l CoCl₂), supporting that stress prevented internalization of the BAR agonist with the micromolar agonist. This study presents novel data supporting that distinct pools of BARs are differentially available after inflammatory insult.

NEW & NOTEWORTHY Preexposure to an oxidative/nitrosative proinflammatory insult provides a “protective preconditioning” against future oxidative damage. We examined immediate vasoactive and molecular consequences of a brief preexposure via β-adrenergic receptor signaling in small arterioles. Blocked receptor internalization with elevated reactive oxygen levels coincides with a significant and unexpected vasodilation to β-adrenergic agonists at picomolar doses.

INTRODUCTION

Nitrosative and/or oxidative stress are associated with many acute and chronic proinflammatory pathological states leading to cardiovascular disease. The hallmark sign of cardiovascular disease is endothelial dysfunction (5, 11, 12, 20). Accompanying these states, a number of receptors are desensitized, including the β₂-adrenergic receptor (BAR-2) (4). Desensitization is an expected means to regulate any receptor-mediated signal. After normal agonist binding to BAR-2, the first step toward desensitization is phosphorylation of the receptor with dissociation of G_s and continuation with signal transduction (19, 28). Phosphorylation occurs via either PKC (with low-dose agonist) or G protein-coupled receptor kinase (GRK; with high-dose agonist) (15, 23, 24). Phosphorylation via GRK (likely GRK-2) recruits β-arrestin to the BAR agonist complex and then recruits dynamin and results in clathrin-mediated endosome formation to internalize the complex with the release of β-arrestin. This internalization mechanism requires nitrosylation of β-arrestin and is accelerated by nitrosylation of dynamin (19, 25, 27). Nitrosylation of GRK instead blocks phosphorylation of the BAR-agonist complex, thus leaving the active complex at the membrane (19). Phosphorylation via PKC, especially with very low agonist doses, does not necessarily lead to receptor internalization (15). Instead, dephosphorylation can occur at the membrane, thus recycling the receptor, and at agonist concentrations well below those required to activate GRK. Although nitrosylation of PKC has been demonstrated, it is unclear how this affects desensitization of the BAR-agonist complex (2). How desensitization occurs with inflammation or stress is not clear, but it appears to occur before internalization (26).

In the present study, after exposure to the nitric oxide (NO) donor sodium nitroprusside (SNP), dilation to micromolar BAR-2 agonists is attenuated and after oxidative stress BAR-2 internalization is reduced, thus supporting that desensitization occurs before internalization. Furthermore, after NO exposure, picomolar agonist-induced dilation via BAR-2 is uncovered. This is mimicked by deliberately blocking clathrin endosome formation with dynasore [to inhibit dynamin (17)]. Our study raises further questions regarding how inflammation differentially impacts the initial steps toward desensitization.

MATERIALS AND METHODS

Intravital microscopy model.

With permission from the Stony Brook University Animal Care and Use Committee, adult male hamsters (n = 45, 92 ± 22 days, 116 ± 16 g) were anesthetized [pentobarbital sodium (70 mg/kg ip)] and tracheostomized. Animal core temperature was maintained at 38–39°C using conductive and convective heat sources. The left cheek pouch tissue was prepared for intravital microscopy and superfused with physiological saline solution flowing at a constant rate of 5 ml/min over the tissue at a temperature of 36°C (7, 8, 10, 16). Small arterioles were observed using a modified Nikon upright microscope with fluorescence capability, ×60 salt water immersion objective (Nikon, numerical aperture: 0.90), and an intensified charge-coupled device camera (Dage MTI, Gen/Sys II, Indianapolis, IN). Experiments were video recorded (SVHS, Panasonic AG7350). Diameters were measured offline and calibrated with a scale micrometer, with a final magnification of ×400 and an optical resolution of 0.36 µm. Criteria for inclusion in the study were dilation to topical adenosine (10⁻⁴ mol/l) and constriction to topical phenylephrine (10⁻⁴ mol/l). The observed microvascular location was the junction of an arcade arteriole (18 ± 8 µm, means ± SD, 6–43 µm, baseline diameter, range) and the entrance to a terminal arteriolar network (7.3 ± 2.5 µm, 2.5–15 µm). Terminal arteriolar networks are defined by unidirectional flow proceeding to capillaries and arising from an arcading network system, which has by definition bidirectional flow (8, 10, 16). In the protocols outlined below, drugs were applied via micropipette (tip diameter: 10–15 µm, model 740, David Kopf Instruments, Tunjunga, CA) using a minimal ejection pressure (MPPI-2 Applied Scientific Instrumentation, Eugene, OR). Flow out of the micropipettes was confirmed by adding 50 µM BSA conjugated to FITC (Sigma-Aldrich, molecular wt: 4000) and observing the fluorescent dye flow path. All chemicals were obtained from Sigma-Aldrich unless otherwise noted.

SNP-induced stress.

SNP (10⁻⁴ mol/l) was added to the tissue bath for 60 s and then washed out; total exposure time based on the fluid exchange rate is ∼2 min. In this model, 15 min after SNP exposure, the responses at the arcade terminal arteriole junction are altered in specific, reproducible ways (10, 16).

Confirmation of endothelial cell dysfunction after SNP.

Before and after SNP exposure, the site was tested for endothelium-dependent dilation to confirm that tissue bath SNP had altered the responses. Diameter responses were obtained from dilators that act through the endothelium (10⁻⁷ mol/l bradykinin and 10⁻⁵ mol/l acetylcholine), through the vascular smooth muscle (10⁻⁴ mol/l adenosine) or to the constrictor phenylephrine (10⁻⁶ mol/l). Exposure times were 30 s.

Concentration response to BAR-linked events.

Before and after SNP exposure, the concentration response to curcumin (10⁻¹⁵−10⁻⁶ mol/l) or isoproterenol (10⁻¹⁴−10⁻⁵ mol/l) was obtained using only four doses per network. The restriction to only four doses per network is based on prior work showing that when single networks are similarly stressed, repeated exposure to more than five challenges after SNP exposure results in a return to control responses (10, 16). Nitrosative stress here is via tissue bath, affecting all networks, and we have never seen reversal to control responses using this procedure; limiting to four doses reflects caution in characterizing this response. Because of the biphasic nature of the responses as a function of concentration seen in the initial experiments, we randomly altered the dose range as well as the doses tested each day. Exposure times were 30 s. In select animals, inhibitors were applied using a second micropipette directed at the arcade terminal arteriole junction for 5 min before and during testing curcumin or isoproterenol, as outlined below. To confirm specificity of the dilation for BAR-2, carazolol (10⁻⁹ mol/l), a potent BAR inverse agonist, was used with curcumin or isoproterenol. Carazolol was a gift of Dr. S. O. Smith (Stony Brook University). To confirm that the curcumin-induced constriction was α-adrenergic mediated, phentolamine (10⁻⁵ mol/l) was used with curcumin. To test whether inhibition of clathrin endosome formation mimicked the changes in vasoactive responses seen after SNP exposure, dynasore [30 s at 8 × 10⁻⁵ mol/l or at 1.6 × 10⁻⁴ mol/l, inhibiting dynamin (17)] was used with curcumin or isoproterenol. To evaluate an additive versus synergistic effect of preventing endosome formation after SNP stress, dynasore was used without and with SNP preexposure.

Colocalization and FRET analysis of curcumin and BARs in HeLa cells.

HeLa cells were obtained from Dr. D. Brown (Stony Brook University), and BAR-monomeric red fluorescent protein (mRFP) was a gift from Dr. C Berlot (Geisinger Research). HeLa cells were maintained in DMEM (GIBCO) and supplemented with 10% FBS and 1% penicillin-streptomycin at 37°C and 5% CO₂. Cells were transfected with Fugene (Promega) according to the manufacturer’s instructions. Briefly, 1 × 10⁵ cells were incubated with 1 µg DNA-2 µl of Fugene complex in antibiotic-free media and were seeded on glass bottom dishes (MatTek). Forty-eight hours posttransfection, cells were either treated with 1 mM of CoCl₂ and with SNP (10⁻⁴ mol/l) or media were changed to HBSS. Curcumin stock (1 mM) was dissolved in absolute ethanol and diluted with HBSS to obtain the desired concentration (1 × 10⁻⁹−1 × 10⁻⁶ M). Confocal imaging was performed on a Zeiss LSM 510 equipped with a ×40 (numerical aperture: 1.2) Apochromat objective. For colocalization analyses, curcumin, which is a naturally fluorescent molecule (Fig. 5D), was excited using a 488-nm argon ion laser, and its emission was collected from 500 to 530 nm. BAR-mRFP was excited using a 543-nm HeNe laser, and its emission was collected from 535 to 590 nm. Images were acquired sequentially to minimize cross-talk between the two channels. Uptake of curcumin (1 × 10^–8−1 × 10^–6 mol/l) and BAR-2-mRFP were monitored from 15 s to 1 h after the addition of curcumin. Colocalization was assessed using Mander’s colocalization analysis (18) implemented in the MacBiophotonics version of ImageJ.

Fig. 5.

HeLa cells transfected with β₂-adrenergic receptor-monomeric red fluorescent protein. A: phase contrast image for B and C. B: no fluorescence is seen at 453 nm until addition of curcumin (C). D: excitation and emission spectra of curcumin in HBSS (pH 7.2). E: without curcumin, receptor location was fully seen at 543 nm.

Sensitized FRET emission between curcumin and BAR-2-mRFP was performed using the following settings: curcumin and BAR-2-mRFP were excited using 488 and 543 nm, respectively. Emission was collected from 505 to 530 nm (donor channel) and from 560 to 615 nm (FRET and acceptor channels). Normalized FRET was calculated using the FRET module package of Zeiss LSM software, as previously described (3, 14).

ROS kit experiments.

We used a commercial kit (ab139476, Abcam) to determine changes in the level of reactive oxygen species (ROS). This kit included the oxidative stress detection reagent (453 nm, green), the superoxide detection reagent (543 nm, orange), the ROS inducer (pyocyanin), the ROS inhibitor (N-acetyl-l-cycteine), and wash buffer salts. All components were stored at −80°C. Based on the protocol information regarding this assay kit, the oxidative stress detection reagent (green) and superoxide detection reagent (orange) were each reconstituted in 60 µl of anhydrous dimethylformamide for a final concentration of 5 mM stock solution. {Although all data were collected at both wavelengths, data at 453 nm [reactive nitrogen species (RNS)] overlap with the curcumin signal and thus is not reported in this study.} The positive control included the ROS inducer (pyocyanin), which was reconstituted in 100 µl of anhydrous dimethylformamide for a stock solution of 10 mM. A final concentration of 300 µM was used. The ROS inhibitor (N-acetyl-l-cysteine) was reconstituted in 123 µl of deionized water for a stock solution concentration of 0.5 M and used at a final concentration of 5 mM. The wash buffer salts were dissolved in 1 liter of deionized water. Wash buffer was stored in 4°C and warmed to room temperature before use. Protocols tested for ROS include curcumin and SNP alone and together, SNP with dynasore, pretreatment with SNP followed by curcumin, and pyocyanin (positive control) and N-acety-l-cysteine (negative control).

Statistics.

Typically, two or three observation sites were tested per animal; n refers to numbers of animals and N refers to the number of sites tested per condition. Diameters (in µm) are reported for baseline values. Diameter change was reported as percent baseline: [peak change – baseline]/[baseline] × 100. Differences between groups were determined by ANOVA, using standard equations as found in Snedecor and Cochran (22). Because only four concentrations of curcumin or isoproterenol were tested per site and the dose ranges and concentrations tested per animal were not identical from animal to animal, comparisons are unpaired. Mander’s coefficient was analyzed using a Mann-Whitney test. The significance level for hypothesis testing was P = 0.05.

RESULTS

The observation site was the arcade arteriole and terminal arteriole, which were equally exposed to micropipette contents. We first confirmed that brief tissue bath exposure to SNP induced endothelial dysfunction; responses were equally diminished in the arcades and terminal arterioles. Dilation to acetylcholine (10⁻⁴ mol/l) was reduced from 75 ± 11% (means ± SE, arcades and terminal arterioles pooled before SNP exposure) to −7 ± 3.7% (P < 0.05 before vs. after SNP exposure). Dilation to bradykinin (10⁻⁷ mol/l) was reduced from 22 ± 6% to 6 ± 1% (P < 0.05). Additionally, dilation to adenosine (10⁻⁴ mol/l) was reduced from 75 ± 5% to 50 ± 5% (P < 0.05), as previously described (5). Constriction to phenylephrine (10⁻⁴ mol/l) was not affected, from −56 ± 2% to −56 ± 3%.

The control dilation to isoproterenol had a single component with an EC₅₀ value of 7 ± 1 × 10^–7 mol/l (arcades) and 4 ± 2 × 10–7 mol/l (terminal arterioles). After SNP preexposure, the concentration response to isoproterenol became biphasic over the concentration range tested; at very low doses dilation was uncovered, and at higher doses dilation was attenuated (Fig. 1). After a full dilation was achieved, the dilation response was constant over the time of exposure (30 s). Thus, SNP preexposure uncovers potent picomolar dilation to the BAR agonist isoproterenol, with an attenuation effect to the “classic” dilation to isoproterenol.

Fig. 1.

Isoproterenol concentration response before and after exposure to exogenous nitric oxide [after sodium nitroprusside (SNP)] for arcade arterioles (A) and terminal arterioles (B). Control data (□) and data after SNP (■) are the peak dilation responses within 60 s of micropipette exposure to isoproterenol (n = 9). *Differs from control, P < 0.05.

The highly potent BAR-dependent response after SNP preexposure was further examined using the adrenergic agonist curcumin. Exposure to SNP likewise altered the response to curcumin. Previously, we have shown that the vasoactive response to a single dose of curcumin is biphasic over time in these arterioles, with an initial dilation peaking by 15–20 s that is BAR dependent followed by a sustained constriction peaking by 50–60 s that is α-adrenergic receptor dependent (6). Figure 2 shows the initial dilation in response to curcumin, peaking by 15–20 s during the 30-s exposure period. Unlike the control biphasic response over time, after SNP (10^–4 mol/l), the response to curcumin showed a single component over time that was concentration dependent. Lower concentrations induced only dilation, and higher concentrations induced only constriction. SNP preexposure induced the emergence of a bimodal response with decreased efficacy at high doses and increased potency at low doses (Fig. 2). This occurred similarly for the arcades and terminal arterioles.

Fig. 2.

Curcumin concentration response before (control) and after exposure to exogenous nitric oxide [after sodium nitroprusside (SNP)] for arcade arterioles (A) and terminal arterioles (B). Control data (□) show the peak dilation with 60 s of micropipette exposure to curcumin (n = 11). Data after SNP (■) show the peak diameter change with 60 s of micropipette exposure to curcumin (unpaired comparison, n = 11). *Differs from control, P < 0.05.

Using carazolol (10⁻⁹ mol/l), we confirmed that the dilation seen with low (10⁻¹³ mol/l) curcumin is BAR-2 dependent (Fig. 3). The action of carazolol was verified by showing that the control dilation to isoproterenol (10⁻⁶ mol/l, +44 ± 26%, means ± SD, N = 8) was absent in the presence of carazolol (−4.8 ± 18%, P < 0.05). Using phentolamine (10⁻⁵ mol/l), we confirmed that the constriction seen with high (10⁻⁷ mol/l) curcumin was α-adrenergic receptor dependent (Fig. 3), as previously described (6). Inhibition was verified by showing that the control constriction to phenylephrine (10⁻⁴ mol/l, −64 ± 19%, means ± SD, N = 10) was absent in the presence of phentolamine (+3 ± 10%, P < 0.05). Thus, after SNP preexposure, curcumin-induced BAR-2-dependent dilation was significantly more potent and paradoxically absent at high concentrations of curcumin. The residual constriction did not appear to be affected by SNP preexposure.

Fig. 3.

Diameter change for arcade and terminal arteriolar segments in response to curcumin after sodium nitroprusside exposure. Dilation with 10^–13 M curcumin was blocked with carazolol (10^–9 M, n = 8). Constriction with 10^–5 M curcumin was blocked with phentolamine (10^–5 M, n = 10). *Differs from control.

To explore a mechanism for this change in potency for BAR-dependent responses, we considered that SNP preexposure may have altered a fundamental ligand-receptor interaction involving clathrin endosome formation (i.e., desensitization). We hypothesized that by preventing endosome formation, the high-dose dilation would be restored. Dynasore alone did not significantly change baseline diameter (8 × 10⁻⁵ mol/l, 24 ± 16%, means ± SD, N = 6; and 1.6 × 10⁻⁴ mol/l, 18 ± 17%, N = 6). Dynasore had no effect on the control constriction to phenylephrine (10⁻⁴ mol/l) for arcades (from −61 ± 4% to −62 ± 4%) or terminal arterioles (from −55 ± 4% to −57 ± 2%, N = 12). Dynasore exposure 5 min before and with curcumin or isoproterenol exposure uncovered the low-dose dilation and attenuated dilation at the high dose (Fig. 4 and Table 1). SNP preexposure followed by dynasore further increased efficacy for the dilation to curcumin or isoproterenol at picomolar doses (Fig. 4). Significant changes in dilation across these conditions are compared in Table 1.

Fig. 4.

Diameter change in response to curcumin (n = 6; A and B) or isoproterenol (n = 6; C and D) during dynasore (80 μM, 160 μM pooled) exposure before (○) and after sodium nitroprusside (SNP) exposure (●). *Significant dilation for dynasore alone; **differs from dynasore alone.

Table 1.

Statistical comparison for data shown in Figs. 1,2, and 4

Arteriole (Dose, 10−13 M)	Control	After SNP	Dynasore Alone	SNP + Dynasore
Curcumin
Arcade
−15	−0.75 ± 1.3	7.76 ± 8.1
−14	−1.19 ± 1.3	24.60 ± 6.1*†	9.32 ± 0.8	42.55 ± 10.1*†‡
−13	−2.18 ± 2.4	32.64 ± 8.6*
−12	4.44 ± 2.7	15.46 ± 8.3	10.05 ± 1.2	41.95 ± 7.7*†‡
−11	9.13 ± 5.3	−5.34 ± 2.1
−10	13.99 ± 7.0	−11.95 ± 3.8*†	7.67 ± 3.1	24.95 ± 8.5†‡
−9	17.75 ± 6.7	−13.21 ± 2.0*
−8	23.84 ± 5.7	−14.74 ± 2.6*	−3.45 ± 2.4*	8.01 ± 3.9*†‡
−7	23.33 ± 5.5	−20.87 ± 4.2*
−6	23.05 ± 7.1	−20.78 ± 2.0*†	−6.78 ± 0.5*	−14.05 ± 5.1*
−5	27.74 ± 6.1	−20.66 ± 1.5*
−4	27.08 ± 5.9	−20.17 ± 2.2*
Terminal
−15	0.97 ± 1.8	7.99 ± 3.2
−14	3.26 ± 1.9	29.48 ± 5.2*†	7.50 ± 1.9	37.3 ± 6.8*†
−13	−1.48 ± 2.9	35.55 ± 5.3*
−12	5.78 ± 5.5	33.45 ± 5.3*	22.62 ± 5.5*	39.7 ± 11.5*†
−11	13.87 ± 3.9	12.79 ± 10.1
−10	26.63 ± 8.5	8.21 ± 10.3*	11.47 ± 8.5*	34.5 ± 9.9†‡
−9	37.21 ± 8.7	0.60 ± 4.0*
−8	36.77 ± 9.4	−7.77 ± 5.8*	−11.27 ± 9.4*	11.3 ± 1.0*†‡
−7	35.35 ± 6.2	−10.56 ± 3.3*
−6	37.74 ± 6.9	−15.70 ± 4.1*	−3.75 ± 6.9*	−16.93 ± 2.1*†
−5	35.90 ± 8.0	−18.59 ± 2.5*
−4	36.22 ± 4.4	−23.18 ± 3.2*
Isoproterenol
Arcade
−14	−3.63 ± 1.6	−0.57 ± 2.3	1.32 ± 0.7	11.51 ± 3.4*†‡
−12	−1.9 ± 0.6	4.21 ± 1.8*	5.55 ± 0.6*	22.37 ± 2.9*†‡
−10	−0.78 ± 1.1	25.7 ± 2.8*†	11.74 ± 4.1*	8.56 ± 4.0‡
−9	−0.12 ± 0.9	21.15 ± 3.3*†	−0.39 ± 0.3	6.16 ± 2.2‡
−8	7.49 ± 2.9	−5.69 ± 4.3	1.99 ± 1.2	0.32 ± 3.4
−7	11.71 ± 3.8	−0.4 ± 4.0	4.38 ± 2.3	3.89 ± 8.4
−6	41.24 ± 7.4	17.94 ± 3.1*	17.10 ± 3.6*	6.61 ± 8.4*
−5	57.09 ± 11.5	32.83 ± 12.5	20.94 ± 4.0*	40.82 ± 8.4†
Terminal
−14	−0.71 ± 2.3	2.46 ± 0.9	−0.36 ± 1.3	10.14 ± 2.7*†‡
−12	0.24 ± 1.6	13.06 ± 4.7*	7.02 ± 2.3	33.6 ± 1.6*†‡
−10	−0.16 ± 1.1	33.09 ± 3.3*†	14.69 ± 4.1*	9.68 ± 5.6‡
−9	0.59 ± 1.5	23.66 ± 4.0*†	0.58 ± 1.9	3.17 ± 4.2‡
−8	6.91 ± 2.4	−1.76 ± 7.3	2.96 ± 2.1	2.28 ± 4.5
−7	13.64 ± 3.0	−3.30 ± 6.4	3.89 ± 0.8	11.49 ± 11.1
−6	54.79 ± 5.9	35.27 ± 4.7*	24.38 ± 6.2*	8.04 ± 10.3
−5	67.92 ± 12.0	38.77 ± 5.6	32.73 ± 10.7*	45.72 ± 7.5

Entries are means ± SE. Control and sodium nitroprusside (SNP) data are from Figs. 1 and 2. Dynasore and SNP + dynasore data are from Fig. 4.

^*Differs from control;

^†differs from dynasore alone;

^‡SNP + dynasore differs from SNP alone.

Based on the in situ data, a cell culture system was used to address whether curcumin (naturally fluorescent; Fig. 5) and BAR-2 (with a fluorescent tag, mRFP) were internalized together and colocalized and that the two signals from curcumin (453 nm) and the mRFP tag (543 nm) did not overlap (Fig. 6).

Fig. 6.

Colocalization of curcumin and β₂-adrenergic receptor (BAR-2)-monomeric red fluorescent protein (mRFP) in HeLa cells. Shown are merged images of HeLa cells taken at 453 nm (curcumin; green) and 543 nm (β₂-AR-mRFP; red). Colocalization is shown in yellow. Without oxidative stress, colocalization increased over 5 min [<30 s (A), >2 min (B), and >5 min (C)] after the addition of curcumin (100 nM). After oxidative stress by 1 mM CoCl₂ for 1 h [<30 s (D) and 5 min (E)] or 12 h [5 min (F)], there was immediate internalization and colocalization at early times (D) that appeared to be reduced with longer oxidative stress (F). Scale bars = 20 μm.

Figure 6A shows that without oxidative stress, curcumin and BAR-2-mRFP are readily internalized within seconds and colocalized within HeLa cells. Over 2 to 5 min (Fig. 6, B and C), there was an increase in colocalized signal. Initially, a harsh oxidative stress, 1 mM CoCl₂, was used. After 1 h of oxidative stress with 1 mM CoCl₂, the initial internalization (Fig. 6D) was not as pronounced as without oxidative stress (Fig. 6A). Comparing Fig. 6, C (>5 min without stress) and D (>5 min with stress), there was less internalization/colocalization. Finally, with 12 h of oxidative stress, the receptors were not internalized, although curcumin was still internalized. The early time point images (Fig. 6, A and D) are of most relevance to the in situ data, as curcumin or isoproterenol exposure was for 30 s only. Figure 7 confirms these trends with the Mander’s coefficient of colocalization.

Fig. 7.

Mander’s colocalization values between curcumin (10 nM) and β₂-adrenergic receptor-monomeric red fluorescent protein shown as mean ± SD values for control (n = 19) and with 1 h (n = 11) or 12 h (n = 6) of exposure to oxidative stress (1 mM CoCl₂). Data ≥5 min after the addition of curcumin were pooled. *Differs from control (without CoCl₂, P < 0.009 by Mann-Whitney test). No difference was found between 1 and 12 h of CoCl₂ treatment.

Using the same stress protocol as for the in situ case, we tested whether SNP or dynasore altered the internalization and colocalization (Fig. 8). Figure 8A shows that curcumin alone is internalized over several minutes, as expected. Figure 8B shows that pretreatment with SNP plus dynasore results in internalization of curcumin plus receptor with colocalization. Pretreatment with SNP alone and with dynasore alone likewise resulted in internalization with colocalization. We noted that with both SNP plus dynasore pretreatment, there was an additional finding that HeLa cells spontaneously formed multiple vesicles (membrane blebbing; Fig. 8C, red arrows) not seen with SNP alone or with dynasore alone. The combination of SNP plus dynasore pretreatment and curcumin exposure was required. Thus, it appears that SNP and/or dynasore pretreatment enhances curcumin plus receptor internalization and colocalization.

Fig. 8.

Curcumin and β₂-adrenergic receptor (BAR-2)-monomeric red fluorescent protein (mRFP) colocalization with sodium nitroprusside (SNP; 10⁻⁴ M) and dynasore (160 μM) pretreatment. A: control. HeLa cells were exposed to curcumin at time 0, and over 12 min, curcumin gradually entered the cells (green channel only). B and C: SNP + dynasore. HeLa cells were pretreated with SNP and dynasore for 2 min and then only dynasore for 15 min before the addition of curcumin at time 0. The colocalization of curcumin and BAR-2-mRFP occurred immediately after the addition of curcumin instead of gradually over time (green and red channels merged). C: marked area of B. Arrows show membrane vesicle formation (blebbing) not seen with SNP + curcumin or dynasore + curcumin.

To next examine whether SNP or curcumin affected ROS state, we measured ROS in HeLa cells. Note that the green (453 nm) channel to measure RNS overlaps with the curcumin signal and was not analyzed. Figure 9 shows ROS signaling when SNP was added at the onset of imaging (0 min) and then followed for 12 min. Twelve minutes is the minimum time course for the in situ vasoactive changes to manifest after preexposure to SNP. In HeLa cells, SNP alone did not alter measured ROS levels over 12 min (Fig. 9A). However, curcumin alone increased ROS signaling (Fig. 9B); of importance is the curcumin dosage (10⁻⁹ M). SNP and curcumin applied together (Fig. 9C) likewise increased ROS over time. We next preexposed HeLa cells to SNP for 2 min followed by a 15-min waiting period (identical to the in situ experiments). Figure 10A shows that the ROS signal was stable from 15 to 18 min after SNP exposure (0–3 min in Fig. 10A). When curcumin was added at 0 min (Fig. 10B), there was a rapid increase in ROS signaling by 1.5 min, with evidence of ROS concentration within subcellular structures. Thus, preexposure to SNP followed by exposure to curcumin increases ROS within HeLa cells.

Fig. 9.

ROS increased with curcumin/sodium nitroprusside (SNP) over time: A: HeLa cells were exposed to SNP (at 0 min, time 0) and observed for ROS (543 nm, orange) over 12 min; no appreciable increase in ROS was seen. Note that reactive nitrogen species (RNS; 453 nm, green) overlap with curcumin and, therefore, were not useful in distinguishing RNS versus more rapid curcumin uptake. B: HeLa cells exposed to curcumin (at 0 min) showed increased ROS over time. C: HeLa cells exposed to both SNP and curcumin at 0 min showed increased ROS with evidence of concentration within cellular structures.

Fig. 10.

ROS increased with sodium nitroprusside (SNP) preexposure followed by curcumin: A: HeLa cells were preexposed to SNP (for 2 min, followed by a 15-min wait period, imaging began at 0 min) and observed for ROS (543 nm, orange) over 3 min; no appreciable increase in ROS was seen (15–18 min after SNP exposure). B: HeLa cells preexposed to SNP (same as for A) and then exposed to curcumin at 0 min showed increased ROS with evidence of concentration within cellular structures.

Figure 11 shows positive control experiments with preexposure to the inducer, pyocyanin, followed by an immediate exposure to SNP alone (Fig. 11A) or immediate exposure to SNP + curcumin (Fig. 11C) and preexposure to SNP alone (Fig. 11B) or preexposure to SNP with a subsequent exposure to curcumin (Fig. 11D). The data shown in Fig. 11D follow the protocol used in situ. The inducer clearly increased ROS signaling, and yet preexposure to SNP (Fig. 11B) further increased the ROS signal, which was further affected by curcumin (Fig. 11D). The data are consistent with SNP preexposure elevating ROS within the cell and curcumin modulating the ROS signal. The negative control (N-acetyl-l-cysteine) did not show a time-dependent or SNP/curcumin-dependent change in the signal (data not shown).

Fig. 11.

ROS levels increased with inducer were further increased by sodium nitroprusside (SNP) preexposure. A: HeLa cells were preexposed to inducer (pyocyanin, 30 min) and then to SNP (at 0 min, time 0) and observed for ROS (543 nm, orange) over 3 min. B: HeLa cells were preexposed to inducer and then preexposed to SNP (for 2 min, followed by a 15-min wait period, imaging began at 0 min) and observed for ROS (543 nm, orange) over 3 min. C: conditions in A with curcumin also added at 0 min. D: conditions in B with curcumin also added at 0 min. Preexposure to SNP increased ROS above inducer alone, with a markedly different ROS pattern.

DISCUSSION

This study shows that brief exogenous NO exposure as SNP (nitrosative and/or oxidative stress) fundamentally alters the BAR-2 concentration-response relationship by inducing the emergence of a bimodal response with decreased efficacy at high doses and increased potency at low doses. Prevention of clathrin endosome formation mimicked this stress treatment, and, together, stress with prevention of endosome formation shifted the response further to the left, indicating a further increase in potency.

We found that SNP preexposure induced oxidative stress within the tissue, whereas immediate exposure to SNP was not as potent in altering oxidative stress. We were not able to determine whether RNS were involved because curcumin and the RNS signal overlap. Based on the data (not shown), there was a rapid increase in the 453-nm signal, but this could be attributed to either increased RNS, rapidly increased internalization of curcumin, or both. Instead, the data definitively show that SNP preexposure elevated ROS levels above those seen with the inducer alone.

The elevated oxidative stress is a likely causative explanation for endothelial dysfunction. Furthermore, we have previously shown that exposure to a range of NO donors, peroxide, or other factors also induces endothelial dysfunction and modifies vasodilatory responses; we have also called this a “preconditioning effect” (10, 16), which is consistent with “tachyphylaxis.” Furthermore, initiation of this effect requires convergence on and activation of mitochondrial ATP-sensitive K⁺ channels so that the responses shift from control to the dysfunctional state. This is accompanied by a fundamental shift in cyclic nucleotides responsible for dilatory tone from a control case where there appears to be a balance between cAMP and cGMP to the dysfunctional case where cGMP predominates (10).

Curcumin was used as a model ligand for BAR. Curcumin is known to induce an immediate BAR-linked vasodilation in the micromolar range using isolated blood vessel models (1, 6, 13, 30). In situ, we have shown that curcumin [alcohol-extracted curcumin I (21)] is a potent vasodilator at nanomolar doses and is an efficacious vasoconstrictor at micromolar doses (6). The response in the micromolar range is biphasic over time, with dilation peaking by 15–20 s of exposure followed by constriction peaking by 45–60 s of exposure. We have shown that adrenergic receptor-linked events are required for both the dilation and the constriction responses to curcumin in normal healthy tissue (6).

Our data are consistent with at least two scenarios (not mutually exclusive) to explain the decreased dilation at micromolar doses and uncovered dilation at picomolar doses: internalization of some but not all BARs (decreased dilation at micro-molar doses) and an alternate desensitization of select BARs that have increased affinity (uncovered picomolar dilation). Supportive details from other studies are as follows. Phosphorylation is the first step in desensitization of BARs. At high agonist concentrations, GRK is activated and BARs are phosphorylated at Ser³⁵⁵ and Ser³⁵⁶ (15, 23, 24). Nitrosylation of GRK prevents phosphorylation of BARs, thus preventing internalization and extending the activity of bound BARs (19, 25). We mimicked this state by applying dynasore to prevent dynamin-mediated endosome formation. Finally, evidence from the literature supports that BAR, being a class A receptor, once internalized, does not support further signal transduction via the MAPK cascade, as has been described for class B receptors such as angiogensin (26). This suggests that, after exposure to SNP, some BAR are not internalized and remain active at the plasma membrane. Our data in HeLa cells support this by showing that cells exposed to oxidative stress no longer internalize BARs. However, we hypothesized that by preventing endosome formation we would enhance the high-dose dilation, and we were wrong. High-dose dilation remained suppressed with isoproterenol and absent with curcumin. Nitrosylation of dynamin enhances internalization of bound BARs [desensitization (19, 27)]. If both occurred, then high-dose dilation could be both suppressed (due to dynamin inhibition) and protected (due to GRK inhibition). This explains the data with isoproterenol but not curcumin. Thus, GRK and dynamin nitrosylation may be occurring but cannot be the only events to explain all of the data.

A different scenario is required to explain the uncovered dilation with picomolar concentrations of agonist. At low-agonist concentrations, PKA is activated, phosphorylating BAR at Ser²⁶² (15, 23, 24). Although nitrosylation of PKA has been reported, and this activates PKA independent of receptor binding [i.e., cAMP increases (2)], it is unclear how this affects BAR desensitization or resensitization. Furthermore, nitrosylation of PKA is not achieved by NO donors that readily shed NO (2). Thus, this is not a likely means by which the low-dose dilation is uncovered after SNP exposure. Instead, a simpler scenario is that after SNP exposure, BARs rapidly cycle between the phosphorylated and unphosphorylated state at concentrations well below that required to activate GRK. It is known that cycling does occur while BARs remain in the plasma membrane (15). In cell systems that have not been stressed, phosphorylation occurs at 10–11 mol/l with 40–50% efficiency by 1 min (24). Dephosphorylation (in the absence of agonist) is slower, dropping 60% by 10 min (23). Our exposure time was chosen to be 30 s because that was sufficient to provide a full dilation in response to the agonists we used. Our standard waiting time between successive exposures was 5 min. We speculate that exposure to SNP may have shortened the dephosphorylation time. We do not feel that repeated exposures have reduced the dilation by not permitting dephosphorylation because the doses for each day varied (see materials and methods). Thus, rapid cycling could explain the uncovered dilation at low doses for both isoproterenol and curcumin.

The findings of attenuated dilation to isoproterenol and absent dilation to curcumin at high doses remain a puzzle. Curcumin as a BAR agonist shows a single response at low doses (dilation only) and a biphasic dilation followed by constriction in the micromolar range (6). After exposure to SNP, only the low-dose dilation remains. PKA is activated by picomolar agonist binding to BARs, whereas GRK only becomes active at micromolar agonist binding to BARs and is faster than PKA (24). Together, this suggests that in the control state curcumin has access to two distinct pools of receptors, BARs regulated by PKA at low doses and by GRK at high doses, whereas isoproterenol only has access to BARs regulated by GRK. After SNP exposure, curcumin apparently accesses only the pool of receptors regulated by PKA. What this implies is that after inflammatory nitrosative and/or oxidative stress, isoproterenol can now access the pool of BARs regulated by PKA to initiate dilation.

Conclusions.

The key finding of this study is that exposure to SNP uncovers a picomolar dilation to BAR-2 agonists. The implication is that BARs occur as distinct pools regulated by PKA versus GRK that are differentially available after inflammatory (oxidative) insult.

GRANTS

This work was supported by National Institutes of Health Grant GM-116187 (to S. F. Scarlata).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.D.F. and S.F.S. conceived and designed research; M.D.F., A.M.D., R.C.C., and A.Q. performed experiments; M.D.F., A.M.D., R.C.C., A.Q., and S.F.S. analyzed data; M.D.F. and S.F.S. interpreted results of experiments; M.D.F., R.C.C., A.Q., and S.F.S. prepared figures; M.D.F., A.Q., B.C.C., and S.F.S. drafted manuscript; M.D.F., A.M.D., R.C.C., and S.F.S. edited and revised manuscript; M.D.F., A.M.D., R.C.C., A.Q., and S.F.S. approved final version of manuscript.