A fluorescence activated cell sorter analysis of the relationship between protein kinase G and endothelial nitric oxide synthase

Abstract

The process of regulation of NOS after production of nitric oxide is not yet delineated. Protein kinase G may exert a feedback regulation of this enzyme. We used diaminofluorescein assays to detect changes in basal nitric oxide production caused by modulators of protein kinase G activity in freshly isolated ovine lung microvascular endothelial cells. We also used fluorescence activated cell sorter analysis (FACS) to determine molecular and phosphorylation changes caused by PKG activation with 8-Br-cGMP. The PKG activator, 8-Br-cGMP (100 μM) produced a shift in the basal NO production curve downward. The inhibition began within 5 minutes and was sustained over 4.5h. The two protein kinase G inhibitors 100 μM Rp-8-Br-PET-cGMPS and 50 nM guanosine 3′-5′-cyclic monophosphoro thionate-8-Br-Rp isomer Na salt and the cGMP inhibitor 4 μM Rp-8-pCPT-cGMPS all enhanced NO production as seen by the upward shift in the basal NO curve. Conversely, the PKG activator drug, 100 μM guanosine-3′-5′-cyclic monophosphate-β-phenyl-1NF-ethano-8-bromo sodium salt decreased NO production causing a downward shift in the basal curve. FACS analysis revealed that 5 μM 8-Br-cGMP in <5 min caused an increase in N-terminal labeling of NOS and a decrease in both C-terminal and serine 1177 labeling of NOS. 8-Br-cGMP appeared to increase PKG 1α and to decrease PKG 1β labeling. Changes in other phosphorylation sites were less consistent but overall mean channel fluorescence increased from 19.92 to 217.36 for serine 116 and decreased from 329.27 to 254.03 for threonine 495 phosphorylation. Data indicated that PKG caused both molecular and phosphorylation changes in NOS.

INTRODUCTION

Constitutive nitric oxide synthase in endothelial cells (eNOS, NOS-3, NOS) is localized to caveolae (27, 12) where it docks into the intracellular domain 4 of the bradykinin B2 receptor (16). The structural protein of caveolae, caveolin-1, also binds to NOS keeping it inactive (8). Activation of NOS leading to its dissociation from the complex is calcium dependent (19, 8). A further activation on serine 1177/1179 is produced by kinase activity (21). Other negative regulators of NOS are NOSIP (eNOS interacting protein) (6) and NOSTRIN (nitric oxide synthase traffic inducer) (29). Both interfere with the association of NOS with caveolae and cause its redistribution from the plasma membrane to intercellular compartments with a decrease in nitric oxide (NO) production. Three positive regulators of NOS have been identified. The protein kinase aKt (Protein kinase B) phosphorylates NOS on serine 1177/1179, enhancing NOS activation (10). Protein kinase A also phosphorylates NOS to increase its activity (3). Heat shock protein 90 (HSP90) is a molecular scaffold that facilitates the interaction of kinases and substrates including NOS. It facilitates the dissociation of NOS from caveolae in response to calcium-calmodulin (11, 13).

The process of regulation of NOS after production of nitric oxide is not yet delineated (21, 22) and may be governed by subcellular translocation involving the Golgi network (20). The nucleus has not been considered as playing a prominent role in the metabolism of NOS but recently we have localized serine 116 phosphorylated NOS (pSer116-NOS) in distinct vesicles in ovine neonatal lung microvascular endothelial cell nuclei as well as in the endoplasmic reticulum using fluorescence immunohistochemistry (15). At both sites, we found pSer116-NOS colocalized with protein kinase G1β.

We have shown that 8-Br-cGMP which activates protein kinase-G, a down stream component of the NO signaling pathway, decreased NO production (15). We have also observed that while caveolin-1 is colocalized with NOS in the plasmalemma and golgi, PKG is colocalized with NOS in the cytosol, endoplasmic reticulum and nucleus (unpublished). Thus PKG appears to be directly involved in inactivation of NOS after NO production and to be chaperoned with spent NOS.

In the present analysis, we sought to determine further the relationship between protein kinase G and NOS using fluorescence activated cell sorter analysis (FACS analysis). We compared control cells with their sibling cells treated with 8-Br-cGMP or its analogues using the following parameters: 1) basal nitric oxide production; 2) the expression of serine 1177, threonine 495 and serine 116 phosphorylated NOS; 3) the expression of protein kinase G 1α and 1β isoforms; 4) NOS C-terminal and N-terminal specific antibody binding.

METHODS

This work was reviewed and approved by the Animal Care and Use Review Committee of Los Angeles Biomedical Research Institute.

Primary culture of microvascular endothelial cells

Endothelial cell isolation was done as previously reported (15). Briefly, newborn lambs aged <2 d were obtained from Nebeker Ranch (Lancaster, CA). Each animal was anesthetized with 30 mg/kg ketamine HCl (Phoenix, MO) and sacrificed with pentobarbital (Virbac, TX), and then the lung was excised. Primary microvascular endothelial cell cultures were derived from distal lung parenchyma explants. Tissue strips (<1 mm wide) were cut from the edges of the lung, cleansed with PBS plus 1% Gibco® antibiotic/antimycotic mixture (Invitrogen, NY) and cut with fine scissors under sterile conditions in a tissue culture θ100mm plate. The cut strips were placed in cell growth feed medium (minimal volume to prevent dryness) and then gently crushed with a hemostat to displace endothelial cells. The tissue pieces were removed and the volume of the crowded cell suspension was expanded for redistribution at a density of 50,000 cell/cm² in DMEM (Invitrogen, NY) plus 10% FBS (Atlanta Biologicals, GA), 1% antibiotic/antimycotic mixture and endothelial cell growth supplement (20–80 μg/ml, Sigma, MO) in θ100 mm dishes. Plates were placed in a Hepa filtered incubator (Thermo Electron, Ohio) at 37 °C with 5% CO₂ mixed in air for 48 h to allow viable cells to be anchored and then fresh medium was supplied. To characterize the derived monolayers, 5 × 10³ of passage 1 cells in 0.5 ml were seeded per chamber on Lab-Tek®II slides (Nalge Nunc, IL). The monolayers grown were fixed with freshly prepared 4 % paraformaldehyde for 10 min, permeabilized with 0.2 % triton X-100 in PBS for 2 min, and blocked with Chemicon® blocking reagent by 30 min incubation and then probed for von Willebrand Factor (14) and CD31 (1) using fluorescence immunohistochemistry. The mounted slides were examined under a fluorescence microscope (Zeiss Axioskop 40) and their endothelial characteristics confirmed before cells were used for the following experiments.

Timecourse of effect of cGMP modulators on basal nitric oxide production

Nitric oxide measurements were done using 4,5-diaminofluorescein derivative, DAF-FM, fluorescence (Molecular Probes, OR) as previously reported (15). Briefly, stably confluent microcultures of LMVECs grown in CELLSTAR®96-well clusters (Greiner Bio-one, NC) were used in groups of N=6. Cells incubated with 1% antibiotic/antimycotic mixture were studied in an incubator at 37 °C with room air plus 5 % CO₂. Cells were examined under a Leica® light microscope for morphologic normalcy and washed with warm PBS before the incubation medium was added. In control experiments, one set of microcultures was used for measurements of any cell autofluorescence, one set for measurement of basal NOS activity and one set for measurement of basal NOS activity in cells preincubated for 2h with NOS inhibitor 200 uM L-NAME (Sigma, St. Louis MO). Equivalent culture wells without cells were used to measure autofluorescence of the DAF reaction mixtures with or without test drugs. These controls were use to confirm the validity of the DAF signal for measurement of nitric oxide and to eliminate extraneous fluorescence from experimental data as done in previous report (15).

To show the basic effect of PKG activation, cells were incubated with 100 μM 8-Br-cGMP, a cell permeable drug that stimulates PKG activity and compared with control untreated cells. Other experimental groups were preincubated with a modulator drug: PKG inhibitors 100 μM Rp-8-Br-PET-cGMPS or 50 nM guanosine 3′-5′-cyclic monophosphoro thionate-8-Br-Rp isomer Na salt; cGMP inhibitor: 4 μM Rp-8-pCPT-cGMPS; or PKG activator: 100 μM guanosine-3′-5′-cyclic monophosphate-β-phenyl-1NF-ethano-8-bromo sodium salt. A combination of 0.5 μM DAF-FM and 0.8 μM DAF FM diacetate (Calbiochem EMD, San Diego, CA) was added in the reaction medium to measure both extracellular and intracellular NO. Low doses of fluorescence probe were used to enable detection of minute differences between groups with minimal DAF autofluorescence (26). Inhibitors were added 30 min before DAF and activators were added with DAF. Total NO produced after 5, 90, 180 and 270 min of incubation with DAF at 37 °C was measured using a Victor 1420 multilabel counter (Waltham, Massachusetts) with a fluorescein setting (λex/em 485/535 nm).

Fluorescence of NO produced by cells was calculated as the difference above fluorescence detected in the presence of NOS inhibitor 200 μM L-NAME (which generally was numerically equivalent to measurements for DAF autofluorescence) and plotted against time of measurement. Mean corrected fluorescence ± SE of NO produced by the control and experimental groups were plotted against time of measurement.

Fluorescence activated cell sorter analysis of the effect of 8-Br-cGMP on NOS and PKG expression

LMVECs were grown into confluence in 75cm² flasks. They were then trypsinised and the detached cells were washed with PBS. Cell suspensions in PBS containing 5% antibiotic/antimycotic mixture were split equally into control and treatment groups. Treatment was 5 μM 8-Br-cGMP given <5 min before fixation. To fix cells, 8% paraformaldehyde (PFA) in PBS was added to make a final concentration of 4% PFA. Suspensions were incubated for 10 min with intermittent vortexing and then washed thrice with PBS. Cells in the final pellet were re-suspended in PBS and permeabilized with 100% ice cold methanol dropwise with vortexing. They were left with the methanol for 15 min and then pelleted. They were re-suspended in blocking solution: 20 mg/ml bovine serum albumin (BSA) in PBS with added 0.1 % triton X-100, intermittently vortexed and left on a shaker in between. After 1h of blocking, cells pelleted were resuspended in PBS plus 2 mg/ml BSA plus 0.1 % triton (PBT buffer). The control and cGMP treated suspensions were then equally split for specific antibody probing (Figure 1). The antibodies used (1: 100) were: mouse anti NOS (Sigma, St Louis MO, immunogen was bovine eNOS c-terminal amino acids 1185–1205); rabbit anti NOS (N-terminal moiety); rabbit anti NOS-CT (Santa Cruz, CA, immunogen was a peptide mapping at the C-terminus of human eNOS), rabbit anti-NOS-NT (Santa Cruz, CA, immunogen was N-terminal amino acids 2–160 of human eNOS); rabbit anti serine 1177 phosphorylated NOS (Santa Cruz, CA), rabbit anti serine 116 phosphorylated NOS (Sigma, MO), rabbit anti threonine 495 phospohorylated NOS (Santa Cruz), goat anti PKG 1α and goat anti PKG 1β (Santa Cruz, CA). Probing was for 2h with intermittent brief vortexing and shaking in between for the antibody to permeate the cells. The cells were washed with PBT thrice with increasing volume and spun at 700 rpm for 5 min each time. Cells were then suspended in PBT containing sheep anti-mouse conjugated with FITC secondary antibody, sheep anti-rabbit conjugated with FITC or rabbit anti goat conjugated with FITC secondary antibodies (5 μl per 100 μl PBT) in microcentrifuge tubes. After 1 h incubation with shaking, they were washed thrice and re-suspended in 1 ml PBS and run through a FACS analysis system comprised of a BD FACSCalibur and a MacIntosh Q3 FACS workstation.

Figure 1. Schema of experimental protocol for fluorescence activated cell sorter analysis of the effect of cGMP on antibody labeling of NOS regions and protein kinase G isoforms.

Sibling cells were pooled and equally distributed for controls and tests.

Quantitative Immunofluorescence measurement of In-Cell pSer1177-NOS expressions modified by 8-Br-cGMP in intact monolayers

To quantitatively determine the effect of PKG activation on in-cell pSer177-NOS expression, stably confluent monolayers of LMVECs grown in 96-well clusters were treated with NOS inhibitor 1 mM L-NNA 1h before or PKG activator 20 μM 8-Br-cGMP (Calbiochem, CA) 10 min before fixation. At the end of the incubation period with the drugs, monolayers were washed once with PBS and fixed with freshly prepared 4% paraformaldehyde for 10 min, permeabilized with 0.2 % triton X-100 in PBS for 2 min, blocked with Chemicon® blocking reagent by 30 min incubation, and then incubated with rabbit anti-pSer1177-NOS primary antibody (Santa Cruz) overnight at 4C. After washing, the cells were incubated for 1h with an anti rabbit fluorescein (FITC)-tagged secondary antibody (Sigma, MO). Negative controls without primary antibodies were used to confirm non-specific blocking. The fluorescence was read with a Victor 1420 multilabel counter with a fluorescein setting (λex/em 485/535 nm).

Western blot analysis of the effect of PKG activation on serine 1177 and threonine 495 phosphorylation of nitric oxide synthase

To examine a physiological role of PKG in inhibiting basal NOS, stably confluent monolayers of LMVECs in Ø100mm dishes were treated with low dose 300 nM 8-Br-cGMP (Calbiochem, CA) to activate PKG. Routine immunoblot analysis of cell lysates for expression of pSer1177-NOS and pThreo495-NOS were then carried out to observe the effect of 8-Br-cGMP on the known serine 1177 activation phosphorylation site (7) and the threonine 495 NO-synthesis-uncoupling site (18). Treated monolayers and controls placed on ice were washed once with ice cold PBS. Cell proteins were immediately harvested in 500 μl of complete lysis buffer (50 mM Tris HCl, 1 % Triton X, and 0.5 % SDS) plus 50 mM sodium fluoride (serine/threonine protein phosphatase inhibitor), 5 μM PMSF (serine proteinase inhibitor) and a protease inhibitor cocktail. Lysates were spun at 13000 rpm in a Beckman benchtop centrifuge at 4 C for 10 min. Supernatants were analyzed for protein content with a Pierce BCA Protein Assay Kit. Uniform samples prepared in sample buffer with added disulfide bond reducing agent 1 mM dithiothreitol (DTT) were boiled for 3 min and loaded on NuPage 4–12% Bis-Tris gels (Invitrogen) and electrophoresed under 100 V until the marker front reached the limit. Separated proteins were transferred to nitrocellulose membranes under 30 V over 2 h or 90 V over 1–2 h. Membranes were blocked with 5% nonfat milk and then incubated with a rabbit anti pSer1177-NOS antibody or a rabbit anti pThreo495-NOS (Santa Cruz) (1: 200) at 4C overnight with gentle rocking. Membranes were stripped and re-probed with mouse anti actin primary antibody. Protein bands were visualized each time by horseradish peroxidase conjugated anti rabbit or anti mouse secondary antibodies (Amersham, Buckinghamshire, UK) using Super Signal® West Pico Chemiluminescent Substrate kit (Pierce) to prepare the developing solution. Membranes were exposed to Blue Lite Autorad Films (ISC BioExpress, Kaysville, UT), and the films were processed in a HOPE developer.

RESULTS

Regulation of NOS function by cGMP analogues

The PKG activator, 8-Br-cGMP (100 μM) produced an instantaneous decrease in NO production seen within 5 minutes and sustained over 4.5h with the continued presence of 8-Br-cGMP. 8-Br-cGMP produced a shift in the basal NO production curve downwards indicating that less NO is produced per unit time in the presence of added 8-Br-cGMP. The curve shift was parallel with the basal NO curve indicating a competitive interaction of the 8-Br-cGMP signaling with a NOS activation component (Figure 2A). Rp-8-Br-PET-cGMPS (100 μM), a PKG inhibitor produced an increase in NO production with a parallel shift of the basal NO curve upwards indicating that more NO is produced per unit time in the presence of added PKG inhibitor and this was sustained over 4.5h with the continued presence of the PKG inhibitor (Figure 2A). Another PKG inhibitor, 50 nM guanosine 3′-5′-cyclic monophosphoro thionate-8-Br-Rp isomer Na salt reproduced the effect observed with Rp-8-Br-PET-cGMPS (Figure 2B). The cGMP inhibitor, 4 μM Rp-8-pCPT-cGMPS also produced an increase in NO production with an upward shift of the basal NO curve and this was sustained over 4.5h with the continued presence of the cGMP inhibitor (Figure 2C). The PKG activator, 100 μM guanosine-3′-5′-cyclic monophosphate-β-phenyl-1NF-ethano-8-bromo sodium salt shifted the basal NO curve downwards showing that less NO is produced per unit time in the presence of added PKG activator (Figure 2D). The two PKG activators 8-Br-cGMP and guanosine-3′-5′-cyclic monophosphate-β-phenyl-1NF-ethano-8-bromo sodium salt exhibited synergism in lowering NO production (Figure 2D). Using the 90 min time point, the data plotted as mean ± SE show that 8-Br-cGMP reduces NO production in controls and in all the tests with cGMP/PKG inhibitor drugs as well as in tests with the PKG activator (p<0.05^# in all cases). The cGMP/PKG inhibitor drugs (A), (B) and (C) significantly increased NO production (p<0.05* in all cases) and the PKG activator guanosine-3′-5′-cyclic monophosphate-β-phenyl-1NF-ethano-8-bromo sodium salt (D) significantly decreased NO production (p<0.05⁺) (Figure 2E).

Figure 2. Timecourse of the inhibition of nitric oxide production by protein kinase G: modifications by PKG inhibitor and activator drugs.

Nitric oxide measurements were done using 4,5-diaminofluorescein derivative, DAF-FM, (0.5 μM DAF-FM and 0.8 μM DAF FM diacetate) added to stably confluent microcultures of LMVECs in 96-well clusters. To show the basic effect of PKG activation, cells were incubated with 100 μM 8-Br-cGMP. Other experimental groups were preincubated with a modulator drug: PKG inhibitors 100 μM Rp-8-Br-PET-cGMPS or 50 nM guanosine 3′-5′-cyclic monophosphoro thionate-8-Br-Rp isomer Na salt; cGMP inhibitor: 4 μM Rp-8-pCPT-cGMPS; or PKG activator: 100 μM guanosine-3′-5′-cyclic monophosphate-β-phenyl-1NF-ethano-8-bromo sodium salt. Inhibitors were added 30 min before DAF and activators were added with DAF. Total NO produced after 5, 90, 180 and 270 min of incubation with DAF at 37 °C was measured using a Victor 1420 multilabel counter (Waltham, Massachusetts) with a fluorescein setting (λex/em 485/535 nm). Using the 90 min time point, the data plotted as mean ± SE show that 8-Br-cGMP reduces NO production in controls and in all the tests with cGMP/PKG inhibitor drugs as well as in tests with the PKG activator (p<0.05^# in all cases). The cGMP/PKG inhibitor drugs (A), (B) and (C) significantly increased NO production (p<0.05* in all cases) and the PKG activator guanosine-3′-5′-cyclic monophosphate-β-phenyl-1NF-ethano-8-bromo sodium salt (D) significantly decreased NO production (p<0.05⁺) (lower figure).

Changes in in-cell immunolabeling of NOS and PKG produced by added 8-Br-cGMP to live MVECs

In the FACS analysis, cells with positive signal of fluorescence were analysed using a FITC (fluorescein) filter (FL1). The illustrations presented (Figures 3–4) show a y-axis representing the fluorescein channel FL1-H and indicating the fluorescence range produced by each sample. Ten thousand events were taken per control or test sample. For all controls and cGMP treated groups, generally ~90% of the cells labeled with FITC-conjugated secondary antibody gave fluorescence of less than 5 fluorescence units (FU). To compare any change in fluorescence caused by 8-Br-cGMP, a gate was set for counting cells with greater than 10 FU (Figure 3A). We observed both change in number of cells with high fluorescence intensity (% gated) and change in overall (100% of the cells) mean channel fluorescence intensity (MCF) of 10,000 cells per sample. Data were automatically generated by a Cell Quest® software. For control cells unlabeled with fluorescent antibody, 0.08% were gated and for control 8-Br-cGMP treated cells that were unlabelled 0.03% were gated. The computer was set to adjust experimental results using the background obtained with untreated, unlabeled cells.

Figure 3. Effect of PKG activation on percentage of cells with high fluorescence intensity for various immunolabeling.

Cell suspensions in PBS containing 5% antibiotic/antimycotic mixture were split equally into control and 5 μM 8-Br-cGMP treatment groups before fixation with 4% paraformaldehyde. Cell were permeabilized with methanol, blocked with 20 mg/ml bovine serum albumin (BSA) and then labeled with 1: 100 primary antibodies against NOS N-terminus, NOS C-terminus, or serine 1177, serine 116 or threonine 495 phosphorylated NOS for 2h. They were counterlabelled with FITC conjugated secondary antibodies and ran through a BD FACSCallibur. Cells with positive signal of fluorescence were analysed using a FITC (fluorescein) filter (FL1). Upper figure shows gating of cells for 10 fluorescence units (y-axis) and 10 forward scatter units (x-axis). Middle table shows actual Cell Quest read out values for % gated. The mean values of the three determinations were plotted in the lower figure.

Figure 4. FACS analysis of the effect of PKG activation on NOS N-terminal and C-terminal specific rabbit antibody labeling and pSer1177 NOS specific antibody labeling.

Cells were immunolabeled as described for Figure 3. The fluorescence intensities of sample populations were captured with a Cell Quest software. Upper channels show control cells and lower channels show PKG activator 5 μM 8-Br-cGMP treated cells. Left panels were labeled with a rabbit anti N-terminal of NOS antibody. Middle panels were labeled with a rabbit anti C-terminus of NOS antibody. Right panels were labeled with a rabbit anti serine 1177 phosphorylated NOS antibody. The fluorescence intensity range (y-axis) was increased by PKG activation for N-terminal labeling and decreased by PKG activation for C-terminal labeling. The fluorescence intensity range was decreased by PKG activation for serine 1177 phosphorylation.

Figure 3 shows FACS output data obtained for % of cells gated (> 10 FU) in 3 different determinations. 8-Br-cGMP consistently increased percentage of cells with high NOS N-terminal fluorescence (NOS-NT, >10FU) in 3 separate experiments. 8-Br-cGMP also consistently decreased both percentage of cells with >10 FU for NOS C-terminal fluorescence and serine 1177 phosphorylated NOS fluorescence in 3 separate experiemnts. 8-Br-cGMP did not affect serine 116 and threonine 495 phosphorylation consistently. Using Microsoft Excel 2007, the experiment to experiment variance was 2.98912, 1.82203, 0.50288, 0.81599, 0.41199 respectively for NOS-NT, NOS-CT, pS1177-NOS, pS116-NOS, and pT495-NOS. The p values for a one tailed paired t-test were >0.05 in all cases (p=0.07 for pSer1177-NOS). A single factor ANOVA to determine a real effect of 8-Br-cGMP using QI Macros 2009 within Microsoft Excel gave p>0.05 in all cases.

The data presented in Figure 4 are from a single experimental protocol as outlined in Figure 1 showing the cells were all from the same source and identically treated at the same time. Control unlabeled cells did not show fluorescence; overall MCF was 23.88. Control 8-Br-cGMP treated cells that were unlabelled with antibody did not show fluorescence; overall MCF was 11.63.

FACS print-outs are shown in Figures 4 and 5. Figure 4 indicates that the anti NOS-N-terminal antibody labeled more NOS after 8-Br-cGMP (left panels); the anti NOS-C-terminal antibody labeled less NOS after 8-Br-cGMP treatment (middle panels); the antibody against serine 1177 phosphorylated NOS labeled less NOS, the fluorescence intensity range decreased after 8-Br-cGMP treatment (right panels). Figure 5 indicates that PKG activator, 8-Br-cGMP, caused increase in total labeling intensity and increase in total number of cells with high fluorescence, >10¹ FU (y-axis), with PKG 1α labeling (left panels) but decreased the number of cells with high fluorescence for PKG 1β labeling.

Figure 5. FACS analysis of the effect of PKG activation on antibody labeling of protein kinase G isoforms.

Cells were immunolabeled as described for Figure 3. The fluorescence intensities of sample populations were captured with a Cell Quest software. Upper channels show control cells and lower channels show PKG activator 5 μM 8-Br-cGMP treated cells. Left panels were labeled with a goat anti PKG 1α antibody. Right panels were labeled with a goat anti PKG 1β antibody. PKG activation increased total number of cells with >10¹ FU (y-axis) labeling for PKG 1α and decreased the same labeling for PKG 1β.

Data from overall MCF were similar but slightly different from data from gated cells. We used data from gated cells. Data of MCF for gated cells (FU>10) were pooled from three experiments. The rabbit anti NOS (N-terminus moiety) antibody gave MCF of 24.24 for controls and 77.16 for 8-Br-cGMP treated. The mouse anti NOS (C-terminus moiety) antibody gave MCF of 32.83 for controls and 12.87 for 8-Br-cGMP treated. The anti NOS-N-Terminus antibody gave MCF of 103.5 for controls and 197.09 for 8-Br-cGMP treated. The anti NOS-C-terminus antibody gave MCF of 120.59 for controls and 110.55 for 8-Br-cGMP treated. Mean of collective data from anti phospho-NOS antibodies were: for pSer1177-NOS, 61.97 for controls and 28.69 for 8-Br-cGMP treated; for pSer116-NOS, 19.92 for controls and 217.36 for 8-Br-cGMP treated; and for pThreo495-NOS, 329.27 for controls and 254.03 for 8-Br-cGMP treated.

PKG activation decreases pSer1177-NOS In-Cell Immunofluorescence

A mean background autofluorescence value of 1758 ± 13 was obtained from primary antibody blank wells and this was subtracted from all raw fluorescence values of individual wells. Mean corrected fluorescein fluorescence value for controls was 1447.67 ± 195.1; for 1 mM L-NNA was 268 ± 52.8 and for 20 μM 8-Br-cGMP was 322.7 ± 40.4. Both L-NNA and 8-Br-cGMP significantly decreased fluorescence produced by the presence of fluorescein-tagged serine 1177 phosphorylation of NOS in intact cells (p<0.01 in all cases) (Figure 6).

Figure 6. PKG activation decreases pSer1177-NOS Immunofluorescence.

To quantitatively determine the effect of PKG activation on pSer1177-NOS expression, stably confluent monolayers of LMVECs grown in 96-well clusters were treated with NOS inhibitor 1 mM L-NNA 1h before or PKG activator 20 μM 8-Br-cGMP (Calbiochem, CA) 10 min before fixation. Monolayers fixed with 4% paraformaldehyde were permeabilized with 0.2 % triton X-100 in PBS for 2 min, blocked with Chemicon® blocking reagent by 30 min incubation, and then incubated with rabbit anti-pSer1177-NOS primary antibody (Santa Cruz) overnight at 4C. The cells were counter-labeled with an anti rabbit fluorescein (FITC)-tagged secondary antibody (Sigma, MO). The fluorescence was read with a Victor 1420 multilabel counter with a fluorescein setting (λex/em 485/535 nm). Both L-NNA and 8-Br-cGMP significantly* decreased fluorescence produced by the presence of fluorescein-tagged serine 1177 phosphorylation of NOS (p<0.01 in both). Figure is a representative of two studies.

PKG activation decreases serine 1177 phosphorylated NOS protein expression

The treatment of cells with low dose 300 nM 8-Br-cGMP caused a visual decrease in protein bands generated by blotting with anti-body against pSer1177-NOS (Figure 7). PKG activation did not alter threonine 495 phosphorylation of NOS (Figure not shown).

Figure 7. PKG activation decreases serine 1177 phosphorylated NOS protein expression.

To examine a physiological role of PKG in inhibiting basal NOS, stably confluent monolayers of LMVECs in Ø100mm dishes were treated with low dose 300 nM 8-Br-cGMP. Routine immunoblot analysis of expression of pSer1177-NOS were then carried out to observe the effect on the known serine 1177 activation phosphorylation site. Membranes were blocked with 5% nonfat milk and then incubated with a rabbit anti pSer1177-NOS antibody (Santa Cruz) (1: 200) at 4C overnight. Membranes were stripped and re-probed with mouse anti actin primary antibody. The treatment of cells with low dose 8-Br-cGMP caused a visual decrease in pSer1177-NOS protein bands. Figure is a representative of 2 studies.

DISCUSSION

Measurement of basal NO in intact endothelial cells under physiologic conditions is difficult because of the short half life (3 s) and nanomolar level of NO produced but DAF can be reliably used (26) when extraneous fluorescence is taken care of. The protein kinase G activator, 8-Br-cGMP, inhibited NO production in ovine lung MVECs as seen by a downward parallel shift in the DAF-fluorescence-time curve for basal measurements (Figure 2A). The two protein kinase G inhibitors and the cGMP inhibitor both enhanced NO production as seen by the upward shift in the basal NO curve (Figure 2A, 2B, 2C). Conversely, the PKG activator drug, like 8-Br-cGMP, decreased NO production causing a downward shift in the basal NO production curve (Figure 2D). We used the 90 min timepoint to test the statistical significance of the effects of the various drugs on basal NO production. 8-Br-cGMP reduces NO production in controls and in all the tests with cGMP/PKG inhibitor drugs as well as in tests with the PKG activator (p<0.05^# in all cases, Figure 2E). The cGMP/PKG inhibitor drugs (A), (B) and (C) (Figure 2E) significantly increased NO production (p<0.05* in all cases) and the PKG activator guanosine-3′-5′-cyclic monophosphate-β-phenyl-1NF-ethano-8-bromo sodium salt (D) significantly decreased NO production (p<0.05⁺) (Figure 2E). The effect of PKG activation was immediate and may be a direct molecular alteration of a component of the NOS activation. The data are in consonance with other data we reported (15).

If we know how NO synthesis is terminated and what happens to NOS enzyme after NO synthesis, we would be able to enhance or limit the enzyme function directly. Phosphorylation (21) and translocation (22) of endothelial NOS have been studied with respect to pharmacological influences on NOS activity but we lack clarity about how they relate to the physiological or metabolic processing of the enzyme molecule under basal conditions. Evidence exists for the proteolytic degradation of nNOS and iNOS through the calpain and proteasome pathways. Some NOS inhibitors are known to decrease NO production through increased nNOS or iNOS degradation (24, 28).

Protein kinase G is a down-stream effector of NO signaling and from Figure 2 appears to play a feed back role in inhibiting NOS. Because of the almost instantaneous effect of protein kinase G activation in inhibiting NO production and because we observed that protein kinase G and NOS colocalize in distinct vesicles in the endoplasmic reticulum and in the nucleus (15) we speculated that PKG may be complexed with NOS and may affect the enzyme positioning on a chaperone and therefore its activity state.

In some preliminary experiments, two antibodies from different sources raised against nitric oxide synthase, one in the mouse and the other in the rabbit gave opposite results of changes in NOS immunofluorescence expression after treatment of cells with 8-Br-cGMP under identical conditions. The only difference between the antibodies was the specific moiety of the NOS molecule they were raised against: the mouse antibody being raised against a moiety near the C-terminus and the rabbit antibody being raised against a moiety near the N-terminus. Hence we obtained other antibodies specifically raised against the C-terminus and the N-terminus of NOS from an entirely different source and used these in further FACS analysis of the effect of 8-Br-cGMP.

Interestingly, in the Figure 3 bar chart, comparing cells with high fluorescence intensity (above 10 FU on the y-axis), the effects of 8-Br-cGMP on the percentage of cells gated after labeling with N-terminal or C-terminal specific antibodies against NOS were opposite. Statistical analyses did not portray a significance level. This may be in part due to large in-group (experiment to experiment) variations possibly because of difference in in-cell antibody permeation or other experimental details which might have been improved by a large number of experiments. Nevertheless, the consistency of the trend in the effect of 8-Br-cGMP from experiment to experiment in the Table is worth noting. The bar chart shows that N-terminal NOS labeling was increased by 217.14% and C-terminal NOS labeling was decreased by 48.28%. Like C-terminal labeling, phosphorylated serine 1177 labeling was also decreased by 44.09% by 8-Br-cGMP treatment. This data suggested that PKG may promote carboxy terminal cleavage of NOS or reorientation of this terminal from the cytosol to make it less available for antibody labeling (thus active NOS may be borne or orientated on a molecular chaperone).

Endothelial NOS is known to be regulated by HSP90-based chaperones (11, 4). HSP90 may affect heme reconstitution of nNOS, allosteric modulation of the enzyme and monomer to dimer interconvertion of NOS. The dimer interface of NOS is located near the N-terminal (17) and the fact that our data show a greater intensity of antibody labeling of the N-terminal after PKG activation may depict a conversion from the more active homodimer state to the less active monomer structure of NOS (2) which is a plausible explanation for the 8-Br-cGMP induced decreased NO (Figure 2). We do need to examine in detail if PKG is complexed with NOS and how it affects NOS on a chaperone. So far, there is no evidence that PKG is included in the NOS chaperone complex (13, 9) or that it can displace another kinase such as Akt from this complex.

We compared the overall fluorescence intensities of 10,000 cells from each sample pictorially. Interestingly, for both antibodies against N-terminal regions of NOS, 8-Br-cGMP increased the range of fluorescence intensity (y-axis, Figures 4 left panels show data for one specific anti-NOS-N-terminal antibody). Conversely, for both antibodies against C-terminal regions of NOS, 8-Br-cGMP decreased the range of fluorescence intensity (Figures 4, middle panels show data for one specific anti-NOS-C-terminal antibody). These data support our speculation that PKG mediates molecular conversion of NOS.

The numerical values for mean channel fluorescence were also indicative that PKG activation led to molecular change of NOS resulting in a change in the respective availability of NOS C-terminus and N-terminus for specific antibody labeling. Clearly the degree of fluorescence obtained is dependent on the specificity of the antibody used. In basal NOS activity cells, the C-terminal antibody labeling was greater than the C-terminal labeling of 8-Br-cGMP treated cells. Conversely, in basal NOS activity cells, the N-terminal antibody labeling was less than the N-terminal labeling of 8-Br-cGMP treated cells. Therefore, we speculate that 8-Br-cGMP causes a molecular rotation of NOS upon a biologic axis that resulted in change in subcellular compartmental exposure of the N- and C- terminuses of NOS (Figure 8).

Figure 8.

Summary of effect of 8-Br-cGMP on NOS immunolabeling sites.

The PKG activator also had the effect of increasing the number of cells with high fluorescence for PKG 1α labeling and decreasing the number of cells for high fluorescence with PKG 1β labeling (Figure 5). It appears that cGMP may be causing interconvertion of the PKG 1 isoforms suggesting that PKG 1 is a metabolism–based inactivator of NOS. Such inactivators are well described for the cytochrome P450 liver microsomal enzymes (23).

The PKG activator, 8-Br-cGMP, produced visible changes in NOS phosphorylation. The fluorescence intensity range was reduced for Serine 1177 phosphorylation (Figure 4) but not for both serine 116 and threonine 495 phosphorylation (not shown). The effect of PKG activation in reducing both the % gated for high fluorescence (Figure 3) and the fluorescence intensity range (Figure 4) for serine 1177 phosphorylated NOS labeling may explain the decrease in NO production caused by 8-Br-cGMP (Figure 1). Serine 1177 phosphorylation is associated with increased or prolonged NOS activity (21). In Figures 6 and 7, 8-Br-cGMP caused decrease in pSer1177 antibody labeling of NOS. Both the in-cell labeling and the western blotting show that the pSer1177-NOS expression was down-regulated within a short time frame (< 15 min) therefore it is possibly not a genomic effect. It appears that structural changes and changes in phosphorylation of NOS are linked (Figures 3, 4 and 8) and may be both induced by PKG activation.

Threonine 495 phosphorylation is thought to reduce activity of endothelial NOS (21) and this may partly explain the reduction of NO caused by 8-Br-cGMP (Figure 1) but none of the FACS or immunoblotting data supported this. By comparison, for serine 116 phosphorylated NOS labeling, 8-Br-cGMP did not affect the % gated for high fluorescence (Figure 3), but the automated readout for overall mean channel fluorescence intensity for serine 116 NOS labeling was 19.92 for controls and 217.36 for 8-Br-cGMP treated cells. The role of serine 116 phosphorylation of NOS is unknown but there are suggestions that it may be involved in inactivation of NOS (21). In our previous observations using freshly isolated ovine lung microvascular endothelial cells, we noted intense colocalization of PKG 1β and serine 116 phosphorylated NOS in distinct vesicles in the endoplasmic reticulum and nucleus apart from their diffuse cytosolic colocalization (15). This observation plus the current data suggests that the action of PKG on NOS molecular and phosphorylation changes may be linked to its metabolism. An established parallel is the metabolism-based inactivators of the liver cytochrome P450 which have been shown to increase the proteolytic recognition of the cytochrome P450 enzyme (24). Both processes of inactivation and degradation of NOS may be linked and PKG may be involved under basal regulation of NOS. PKG modulation of NOS has not been extensively studied by other workers.

PKG modulating drugs have been studied for their effect on vascular contractility. In the intact vessel, the PKG modulating drugs may affect the cGMP-PKG signaling within endothelial cells and smooth muscle cells simultaneously and we could observe a net effect. Some published data show that while NO mimetics such as DETANONOate and nitroglycerine cause relaxation, cGMP inhibitors increase vascular tension (25) thus contractile function of vessels is altered by PKG modulating drugs. These effects appear to be directly on smooth muscle cGMP or PKG rather than on endothelial NOS. The effect we observe in endothelial cells without smooth muscle cell connection (15) therefore needs further investigation. Combining the two reports, exogenous inhibition of PKG could inhibit smooth muscle relaxation directly as well as enhance NO release from the endothelium directly and the latter could result in compensatory crosstalk of cyclic nucleotides in smooth muscle cells. We therefore need to address whether PKG functions in the endothelial cell and smooth muscle cell are compensatory mechanisms.

In general, for therapeutic purposes, actual targeting of NOS enzyme has not been done. Drugs are used to antagonize or compete with the enzyme’s substrates or cofactors. This is because we lack knowledge of the fate of nitric oxide synthase itself that could be manipulated. It is not known whether the enzyme action is terminated by degradation or recycling of the enzyme. Such metabolic targets are worth studying because of the importance of nitric oxide in the cardiovascular system. The role of protein kinase G may be important in this respect. If we know how NO synthesis is terminated and what happens to NOS enzyme after NO synthesis, we would be able to enhance or limit the enzyme function directly.