The Ca2+ mobilizing peptide, bradykinin (BK), stimulates endothelial nitric oxide synthase (eNOS)-derived cellular nitric oxide (NO)production in association with altering the subcellular distribution of the enzyme. In the present study we examine the influence of cellular GTPases,particularly the large GTPase dynamin, on BK-mediated eNOS localization and cellular NO production. BK stimulation of ECV cells, which were stably transfected with eNOS-GFP (eNOS-GFP ECV304), increased NO production. This was associated with the mobilization of eNOS-GFP protein into Triton X-100-insoluble fractions of cell lysates, and an internalization of plasmalemmal eNOS-GFP in live and fixed ECV 304 cells. Incubation of digitonin-permeabilized ECV304 cells with the non-hydrolyzed GTP analog,GTP-γ-S, abrogated the BK-mediated internalization of eNOS-GFP as assessed by confocal microscopy. Conversely, inhibition of clathrin-dependent endocytosis, via overexpression of AP 180 or pretreatment of cells with chlorpromazine, did not influence BK-mediated eNOS redistribution. Furthermore, specific inhibition of dynamin-2 GTPase function by overexpression of a dominant negative construct, K44A, prevented the BK-mediated enrichment of eNOS-GFP within low buoyant density,caveolin-enriched fractions of eNOS-GFP ECV304 cell lysates. Dynamin-2 K44A overexpression also markedly impaired BK-dependent, L-NAME-inhibited NO production as did incubation of permeabilized cells with GTP-γ-s. These studies demonstrate that disruption of dynamin- and GTP-dependent, but clathrin-independent, vesicle trafficking pathways impairs BK-dependent cellular NO production, via inhibition of the internalization of eNOS-containing plasmalemmal vesicles.
Endothelial nitric oxide synthase (eNOS) and its enzymatic byproduct,nitric oxide (NO), are integral to numerous vascular functions including vasodilation and vascular remodeling(Palmer et al., 1987; Rudic, 1998; Shaul, 2002). In this regard,eNOS derived NO production is regulated at multiple cellular levels including transcription, post-transcription and post-translation(Papapetropoulos et al.,1999). With regard to the last of these mechanisms, the appropriate subcellular targeting of eNOS within the membrane compartment of cells is necessary for optimizing cellular NO production(Blair et al., 1999; Fulton et al., 2002; Shaul et al., 1996). Membrane targeting, predominantly to plasmalemmal caveolae and Golgi membranes, has been ascribed to a number of processes including amino-terminal acylation,sequence-specific binding of eNOS protein with specific membrane-spanning proteins such as caveolin and chaperone molecules(Blair et al., 1999; Fulton et al., 2002; Liu et al., 1995; Prabhakar et al., 2000; Shaul et al., 1996). As an extension to this concept, previous studies have demonstrated that cellular eNOS localization is influenced by both cellular homeostatic conditions such as cell confluency state and receptor-mediated inhibitory and excitatory signaling stimuli such as oxidized low density lipoproteins and bradykinin(BK) (Michel et al., 1993; Blair et al., 1999; Prabhakar et al., 1998; Sowa et al., 1999; Thuringer et al., 2002).
BK is a soluble peptide that acts via NO to promote vasodilation and capillary permeability. Soluble BK binds the membrane-bound BK2 receptor with downstream activation of eNOS through divergent and convergent signaling pathways that include Hsp 90, MAP kinase, receptor tyrosine kinases, and eNOS dephosphorylation (Bernier et al.,2000; Harris et al.,2001; Thuringer et al.,2002; Venema et al.,1996). In addition to these complimentary signaling pathways, it has been observed that treatment of cells with BK is associated with changes in the subcellular distribution of eNOS(Prabhakar et al., 1998; Thuringer et al., 2002; Venema et al., 1996). Owing to the prominent influence of GTPases, particularly the large GTPase, dynamin-2,on internalization of plasmalemmal vesicles and downstream signaling pathways such as that described for the extracellular signal-regulated kinase cascade(Pierce et al., 1999), we sought to determine the influence of perturbation of dynamin-2 and GTP hydrolysis on cellular eNOS localization and NO production. We hypothesized that the specific targeting of eNOS, as a passive cargo within cellular vesicles, may influence the ability of the enzyme to produce and release NO. We demonstrate, using two complimentary cell systems, that BK-stimulated NO production is associated with a redistribution of the plasmalemmal pool of eNOS protein, with enrichment within Triton X-100 insoluble, low buoyant density cell fractions. Furthermore, we demonstrate that disruption of dynaminand GTP-dependent endocytosis, but not clathrin-dependent endocytosis,abrogates the BK-mediated redistribution of eNOS within cells. Also,disruption of dynamin GTP hydrolysis attenuates BK-dependent cellular NO production. These studies implicate a role for GTPase-dependent,clathrin-independent, trafficking of eNOS-containing vesicles in the mechanism of BK-mediated NO generation.
Materials and Methods
Materials and antibodies
All tissue culture and transfection reagents were obtained from GIBCO BRL. eNOS mAb, caveolin pAb, were obtained from Transduction Laboratories(Lexington, KY, USA), β-COP pAb was from Affinity Bioreagents, Inc(Golden, CO, USA), Golgi 58K mAb and plasma membrane marker,Na+-K+ ATPase mAb were from Sigma (St Louis, MO), and myc and V5 mAb were from Invitrogen Life Technologies (Carlsbad, CA, USA). Texas Red-conjugated transferrin was from Rockland Immunochemicals(Gilbertsville, PA). A polyclonal antibody that specifically recognizes dynamin-2 (dyn2T) was a generous gift from Dr Mark McNiven(Henley et al., 1998). Optiprep, was obtained from Nycomed Pharma AS (Oslo, Norway). The K44A dominant negative dynamin-2 cDNA, which contains a point mutation in the GTP binding element, thereby preventing GTP hydrolysis, a kind gift from Dr Mark McNiven, was subcloned into pcDNA3 in frame with a V5 epitope tag. The myc-tagged AP 180 cDNA in pCMV was a generous gift from Dr H. McMahon(Ford et al., 2001). The probes, DAF-2 and DAF-2DA, for the fluorimetric detection of NO, were obtained from Calbiochem (La Jolla, CA)
Complimentary techniques were performed in both bovine aortic endothelial cells (BAEC), which express eNOS endogenously, and in ECV-304 cells stably transfected with eNOS-GFP (eNOS-GFP ECV 304 cells)(Sowa et al., 1999). Although the latter cells are probably derived from a human bladder tumor(Brown et al., 2000), their phenotype has been documented to partially overlap with that of endothelial cells; and owing to their stable and uniform expression of eNOS-GFP at relatively high levels (Paxinou et al.,2001; Sowa et al.,1999), these cells have served as a useful model in which to study eNOS trafficking. BAEC were from Clonetics (San Diego, CA) and used between P2 and P5, while eNOS-GFP-ECV 304 cells were a generous gift from Dr William Sessa (Sowa et al., 1999). Both cell types were cultured in Dulbecco's modified Eagle's medium (DMEM),supplemented with 10% fetal bovine serum; 1 mM L-glutamine, and 100 IU/ml penicillin. eNOS-GFP ECV 304 cells were additionally supplemented with 400μg/ml G418. For transfection experiments, eNOS-GFP ECV 304 cells were plated on glass coverslips in 12- or 24-well plates, or in 100 mm culture dishes and transfected using a calcium phosphate method with pcDNA3-dyn 2 K44A, pCMV-myc AP 180, or appropriate control vectors. Twenty-four hours after transfection of eNOS-GFP ECV 304 or 24 hours after plating of BAEC, the cells were incubated with 10 μM BK or vehicle for 10 minutes at 37°C. After BK stimulation, cells were prepared for various subcellular fractionation procedures, microscopic analyses, or NO measurements as individually described below. For permeabilization experiments, cells were pretreated with either 1μM digitonin, 10 μM GTP-γ-s, 1 μM digitonin plus 10 μM GTP-γ-s, or vehicle (PBS) for 5 minutes, then washed with PBS prior to BK stimulation. In some experiments 10 μM ATP-γ-s was used in place of GTP-γ-s and in some experiments, cells were treated with 15 μM chlorpromazine (CP), an agent that blocks clathrin-dependent endocytosis, for 10 minutes prior to BK stimulation (Petris et al., 2002). To assess transferrin uptake in response to AP 180 transfection (Ford et al.,2001), cells were incubated in low serum medium (0.2% BSA) for 30 minutes at 37°C and Texas Red-conjugated transferrin was added from a stock of 5 mg/ml to the medium at a ratio of 1:1000. Cells were washed with PBS after 10 minutes incubation at 37°C. Cells were washed further with low serum medium (pH 3.5) to reduce background signals and fixed with 2%formaldehyde solution. Uptake of transferrin was assessed by confocal laser scanning microscopy.
Confocal laser scanning microscopy
To study eNOS-GFP localization in living cells in response to BK stimulation, a cell culture system with stable temperature control (POC-R;Zeiss) was mounted to the laser scanning confocal microscope (Pasqual LSM 5;Zeiss). Briefly, eNOS-GFP ECV 304 cells plated and cultured on 32 mm coverslips were mounted in fresh medium in the 37°C POC-R chamber on the microscope. BK was added to the chamber and the cells were scanned and single confocal images were captured at 4-minute intervals using a 63×C-apochromat lens. Micrographs of live cells were digitized and exported to Adobe Photoshop, version 5.0, for calculation of the intensity of the fluorescence in the perinuclear and plasmalemmal regions at each time interval, using data pooled from three independent experiments. For immunofluorescence microscopic analysis, BAEC were fixed in acetone/methanol 1:1 for 3 minutes at -20°C after BK stimulation (10 minutes), while ECV 304 cells were fixed with 2% paraformaldehyde. Fixed cells were incubated with the appropriate dilution of indicated Ab or blocking serum for 2 hours at room temperature, and primary Ab was detected with an FITC- or Texas Red-conjugated secondary antibody and mounted in Anti-fade (Molecular Probes, Oregon). Cells were visualized using a confocal laser scanning microscope with a 63×lens. Owing to the previously noted heterogeneity in subcellular distribution of eNOS between neighboring cells(Prabhakar et al., 1998; Sowa et al., 1999), we classified cells into two groups for the purpose of quantifying changes in NOS distribution in association with BK: (1) plasma membrane pattern (PM:cells with distinct plasma membrane and perinuclear localization of eNOS-GFP)and (2) non-plasma membrane pattern (Non PM: cells with perinuclear localization of eNOS-GFP in the absence of plasmalemmal staining). Quantification of eNOS redistribution in fixed cells was performed in a blind fashion from 200 individual cells in each group from two independent experimental preparations.
Measurement of cellular NO production
NO production was measured using the complimentary fluorescent NO indicator probes, DAF 2DA and DAF-2, the former detects intracellular NO accumulation and the latter detects NO production by assessing NO released from cells(Fulton et al., 2002; Goetz et al., 1999; Kojima et al., 1998), as well as NO-specific chemiluminescence (Shah et al., 1997). For the DAF-2DA assay, BAEC were cultured on glass coverslips in 12-well plates. After the permeabilization procedure described above, cells from each of the experimental groups were washed twice with PBS,preincubated with 0.1 mM NOS substrate, L-arginine at 37°C, or 1.0 mM of the NOS inhibitor, L-NAME for 10 minutes, then loaded with DAF-2DA (10 μM). After 10 minutes, BK or vehicle was added to the cells. Intracellular DAF-2DA fluorescence was visualized using a conventional fluorescence microscope(5100TV, Zeiss, Germany) and fluorescence was quantified from ten selected regions of the digitized micrographs, each taken at 1 minute time intervals,using Adobe Photoshop 5.0 software. Specificity of DAF-2DA fluorescence was established in preliminary experiments, which demonstrated a linear relationship between DAF-2DA fluorescence intensity in response to the NO-stimulating calcium ionophore A23187 and parallel control experiments using DAF-4, a nonfluorescent analog (data not shown). To measure cellular NO release using DAF-2, eNOS-GFP ECV304 cells were transfected as described above and 24 hours later cells were incubated in a solution containing 1 μM DAF-2 with either 0.1 mM L-arginine or 1 mM L-NAME at 37°C. After 30 minutes, BK(10 μM) was added to the cells for an additional 10 minutes at which point the cell supernatant was transferred to microcuvettes for fluorescence quantification with excitation of 485 nm and emission of 538 nm using a fluorimeter (VersaFluor Flurometer, BioRad). Fluorimetry was performed with duplicate readings from duplicate wells and run along side a standard curve generated using known amounts of sodium nitroprusside (0-500 nM) and normalized for cellular protein.
NO-specific chemiluminescence was performed using a Seivers NOA,essentially as previously described (Shah et al., 1997), except that glacial acetic acid containing sodium iodine was substituted for vanadium/hydrochloric acid as the refluxing agent. A standard curve was generated using known nitrite standards. Samples of medium were measured in duplicate and normalized for protein concentration.
Subcellular fractionation and western blotting
For separation of cell membranes from cytosol, BAEC or eNOS-GFP ECV 304 cells were scraped in a buffer containing 50 mM Tris-HCl pH 7.4, 0.1 mM EDTA,0.1 mM EGTA, plus a protease inhibitor cocktail (Roche Diagnostics, Germany),and sonicated on ice at 200 W/cm2 for 3 cycles with 10 seconds intervals, then centrifuged at 100,000 g for 1 hour at 4°C. The pellet constituted the membrane fraction and the supernatant was utilized as the cytosolic fraction. Equal amounts of protein from the membrane and cytosol fractions were boiled in Laemmli buffer for 5 minutes and subjected to SDS-PAGE and western blotting. In experiments examining eNOS segregation in response to Triton X-100, the membrane pellet was resuspended in 0.5% Triton X-100 in 50 mM Tris-HCl buffer for 15 minutes at 4°C and centrifuged at 10,000 g for 10 minutes. The resulting pellet constituted the Triton insoluble fraction (TIF) while the supernatant represented the Triton soluble fraction (TS). The TIF membrane pellet was homogenized with 5 cycles of 20 strokes in a Dounce homogenizer. Protein concentrations of these fractions were determined by BCA assay and equal aliquots were boiled in Laemmli buffer and subjected to SDS-PAGE and western blotting. Cellular proteins were also fractionated by buoyant density, using a well established protocol (Smart et al.,1995) that we have slightly modified as we previously described(Peterson et al., 1999). For separation of low buoyant density membranes, four 150 mm dishes of confluent eNOS-GFP ECV 304 cells were washed and scraped in 5 ml of fractionation buffer(0.25 M sucrose, 6 mM EDTA, 120 mM Tricine, pH 7.8) and pelleted by centrifugation (1400 g). Cells were resuspended in 1 ml of the identical buffer and homogenized with 20 strokes of a Dounce homogenizer. A postnuclear supernatant was collected by centrifugation for 10 minutes at 1000 g. The supernatant was layered on 30% Percoll in fractionation buffer and centrifuged at 84,000 g for 30 minutes. A distinct visible band comprising plasma membranes was collected and sonicated three times (200 W/cm2). Sonicated samples were mixed with Optiprep to a final concentration of 23%, in a Sorvall Ultrafuge tube. Two additional layers of 20% and 10% Optiprep were added prior to centrifugation at 52,000 g for 90 minutes. Eight gradient fractions were collected and 30 μl of each fraction were boiled in Laemmli buffer for 5 minutes and used for SDS-PAGE. After cell fractionation and SDS-PAGE, gels were prepared for Coommassie staining or alternatively for transfer to nitrocellulose membranes for western blotting. For western blotting, membranes were blocked with 5% dry milk for 1 hour at room temperature prior to incubation with antibodies specific for eNOS, caveolin, β-COP, dynamin-2, myc and V5 epitope as previously described (Shah et al.,1999).
Where appropriate, analysis of variance (two-tailed, paired values) was used to evaluate the differences in the levels of parameters studied, using Statgraff, Version 3. A P value of 0.05 was considered statistically significant.
Bradykinin-induced cellular NO production is associated with a temporal redistribution of eNOS
BK acts through the membrane-bound BK2 receptor to stimulate NO production through well delineated signaling pathways, which are associated with a redistribution of eNOS protein (Bernier et al., 2000; Harris et al.,2001; Prabhakar et al.,1998; Thuringer et al.,2002; Venema et al.,1996). To examine this model, we stimulated eNOS-GFP ECV 304 cells with BK (10 μM) and examined NO production and subcellular localization of eNOS. First, we measured BK-induced production of NO in live cells using the complimentary fluorescence compounds DAF-2 and DAF-2DA, which assess NO production by estimating extracellular NO release and intracellular accumulation of NO, respectively. eNOS-GFP ECV 304 cells generated NO in a BK-dependent manner, and L-NAME-inhibitable manner, as assessed by DAF-2 fluorimetry (Fig. 1, left panel). Corroborative results were obtained in BAEC loaded with DAF-2DA, as assessed by quantification of fluorescence intensity from digitized images of cells stimulated with BK (10 μM) in the presence of L-arginine or alternatively, L-NAME (Fig. 1,right). These studies establish the NO stimulatory effect of BK and demonstrate the specificity of the DAF fluorescent compounds in these experimental conditions and cell systems.
To examine the influence of BK on eNOS subcellular distribution, we performed biochemical and microscopic analyses. First, eNOS-GFP ECV 304 cells were stimulated with BK (10 μM) or vehicle for 10 minutes, and cell lysates were separated into cytosolic and membrane fractions. Western blot analysis of these fractions did not demonstrate differences in distribution of eNOS between these two fractions in cells treated with BK as compared to vehicle,as previously reported (Liu et al.,1995). However, further separation of the membrane pellet by Triton X-100 solubility revealed an enrichment of eNOS protein within the Triton-insoluble fraction of cells treated with BK as compared to vehicle(Fig. 2, right).
We next assessed the influence of BK on eNOS localization by performing studies in live eNOS-GFP ECV 304 cells using a cell culture system adapted for a laser scanning confocal microscope. Previous studies in these cells have established that tracking of GFP accurately assesses, and does not confound,the subcellular localization of the fused eNOS protein(Sowa et al., 1999). Treatment of eNOS-GFP ECV 304 cells with BK promoted the time-dependent internalization of eNOS from the plasmalemmal eNOS location that is characteristic of these cells when grown to confluence (Fig. 3A, bottom panel micrographs). Temporal quantification of fluorescence intensity of both Golgi and plasmalemmal pools of eNOS demonstrate that Golgi distribution of eNOS did not decrease in parallel with the plasmalemmal pool of eNOS, reducing the likelihood of a photobleaching effect (Fig. 3A, graph in bottom panel). Cells treated with vehicle showed no change in the fluorescence pattern of eNOS-GFP over a similar time period indicating that eNOS localization was not influenced by the cell culture conditions within the experimental system (Fig. 3A,top panel, micrographs and graph). The graphs were generated by compiling cumulative data points from 3 cells from each group (mean ±s.e.m.).
Next, to further demonstrate that eNOS does indeed leave the plasma membrane in response to BK, immuno-localization studies of plasma membrane and Golgi marker proteins were performed in conjunction with eNOS localization in fixed cells after BK stimulation. As seen in Fig. 3B (top panels),Na+-K+ ATPase immuno-localization is detected on the plasma membrane of cells (top left panel), and substantive pools of eNOS are situated on the plasma membrane as well (top middle panel). The merged image(top right panel) indicates that pools of the two proteins co-distribute on discrete regions of plasma membrane (arrows, indicated in yellow). As seen in Fig. 3B (lower panels), after stimulation of cells with BK, Na+-K+ ATPase distribution(Mobasheri et al., 1997)remains largely unchanged (bottom left panel), while plasma membrane pools of eNOS are no longer prominent (bottom middle panel). Subsequently,colocalization of the two proteins on plasma membrane is no longer detected(bottom right panel). Next, the Golgi membrane pool of eNOS was examined in an analogous manner (Fig. 3C). Under basal conditions, substantive pools of eNOS (top middle panel; arrows)reside within a perinuclear pattern reminiscent of the location of the Golgi marker, 58K protein (top left panel). Indeed, merging of these images (top right panel, yellow) demonstrates colocalization of eNOS-GFP with 58K protein. After stimulation of cells with BK, neither 58K protein (bottom left panel) or eNOS-GFP (bottom middle panel) redistribute from the Golgi membrane, as further substantiated by persistent colocalization of eNOS-GFP with Golgi 58K protein in a perinuclear distribution (bottom right panel, yellow). These observations suggest that BK stimulation of cells promotes the selective dissociation of eNOS from plasma membrane but does not promote dissociation of eNOS from Golgi.
Inhibition of GTP-dependent endocytosis, but not clathrin-dependent endocytosis, influences the subcellular redistribution of eNOS in response to BK
To test the role of dynamin and retrograde vesicle trafficking on BK-mediated internalization of eNOS, we first interrupted GTP hydrolysis within eNOS-GFP ECV 304 cells using cell permeabilization and GTP-γ-S. Cells were preincubated with either the cell permeant, digitonin (1 μM for 5 minutes); digitonin with the nonhydrolyzed GTP analog, GTP-γ-S; or digitonin with the non-hydrolyzed ATP analog, ATP-γ-S. Cells were stimulated with BK or vehicle and prepared for confocal imaging. The left micrograph in Fig. 4Ademonstrates the plasmalemmal and perinuclear distribution of eNOS observed in cells treated with digitonin alone, which was similar to that observed in the absence of digitonin (seen in Fig. 3C, top middle panel). BK stimulation (10 μM) was associated with a decrease in the fluorescence intensity of plasmalemmal eNOS (middle micrograph), consistent with the analyses in Fig. 3. Interestingly, the influence of BK on plasmalemmal eNOS was abrogated by 5 minutes preincubation of cells with GTP-γ-s (right micrograph), though not so prominently by ATP-γ-s (not shown). Similar results were obtained in fixed BAEC, and this data was utilized for quantification of imaging as depicted in the graph in Fig. 4B. For this analysis,quantification was performed in a blind manner by assessing the percentage of cells with and without plasmalemmal eNOS after the various experimental treatments (n=200 BAEC per group in two independent experiments). Consistent with that observed qualitatively in the representative micrographs from ECV304 cells in Fig. 4A,BK stimulation of BAEC was associated with a reduction in the number of cells that expressed plasmalemmal eNOS (open bars), and this effect was abrogated by preincubation of cells with GTP-γ-s(Fig. 4B). These studies indicate that inhibition of GTP hydrolysis abrogates eNOS internalization after BK stimulation.
GTP-γ-s may inhibit both clathrin and caveolae vesicle-mediated endocytosis (Oh et al., 1998). However, as plasmalemmal eNOS resides largely within caveolae(Shaul et al., 1996), we anticipated that the inhibitory influence of GTP-γ-s on BK-mediated eNOS redistribution from the plasma membrane was largely mediated through a caveolae-dependent process rather than via a clathrin-dependent process. To further examine this, we used complimentary pharmacological and molecular approaches, to examine a role for clathrin-dependent endocytosis in this process. First, CP, a cationic amphiphilic drug that inhibits the assembly of the clathrin adapter protein AP2 on clathrin-coated pits(Petris et al., 2002), was incubated with eNOS-GFP ECV 304 cells prior to BK stimulation. As seen in Fig. 4C, in cells preincubated with CP, internalization of eNOS-GFP is still detected in response to BK(image c; compare with cells in the absence of BK and CP in image a, and cells in the presence of BK and absence of CP in image b). Consistent with this observation, inhibition of clathrin-mediated endocytosis via the overexpression of AP 180 protein, which limits the size and distribution of clathrin cages (Ford et al.,2001), also did not prevent BK-mediated eNOS internalization(image d; compare to cells in the absence of BK in image a, and cells in presence of BK and absence of AP 180 in image b). However, AP 180 did reduce the uptake of Texas Red-conjugated transferrin from transfected cells by greater than 50% compared to cells transfected with the empty vector,indicating that this reagent was indeed functional under our experimental conditions (n=250 cells per group, data not shown). These studies suggest that the inhibitory influence of GTP-γ-s on BK-mediated eNOS internalization is likely a caveolae-dependent process rather than a clathrin-dependent process.
We next dissected the role of the specific cellular GTPase, dynamin-2, as this GTPase is essential to retrograde vesicle trafficking events from the plasma membrane (Dessy et al.,2000; Henley et al.,1998; Oh et al.,1998). We tested the role of dynamin in this process by overexpressing a dominant negative dynamin-2 K44A construct in eNOS-GFP ECV 304 cells, and measuring eNOS distribution within the varying buoyant density membrane fractions after BK stimulation. This construct contains a point mutation in the dynamin GTPase domain that impairs the ability of dynamin to hydrolyze GTP (Cao et al.,2000). Fig. 5Ademonstrates the validity of the density gradient used in these experiments,to distinguish low and high buoyant density proteins prepared from eNOS-GFP ECV 304 cells as shown by the enrichment of the 22 kDa protein, caveolin, in the low buoyant density fractions (fractions 1-4), and the converse enrichment of the Golgi and trans-Golgi compartment marker protein, β-COP, in the high buoyant density fractions (fractions 6-7), consistent with prior studies by us and others using this technique(Peterson et al., 1999; Smart et al., 1995). After BK stimulation of eNOS-GFP ECV 304 cells transfected with empty vector, eNOS was enriched within low buoyant density fractions[Fig. 5B; rectangular box, see eNOS signal in lanes labeled E in fractions #1-4; (-) vs (+)],while in cells overexpressing K44A, BK-mediated enrichment of eNOS in low buoyant density fractions was not detected[Fig. 5B; rectangular box, see eNOS signal in lanes labeled K in fractions #1-4; (-) vs (+)]. Control experiments confirm overexpression of K44A as assessed by immunodetection of the V5 epitope tag (see V5 western blot band in Fig. 5B) and similar level of total proteins within each fraction as assessed by Coommassie Blue staining of gels (not shown). These studies indicate that the biochemical redistribution of eNOS in response to BK stimulation requires the GTPase function of dynamin-2.
Perturbation of GTP hydrolysis inhibits cellular NO production
We next sought to determine whether inhibition of dynamin-2-dependent vesicle trafficking might influence cellular NO production as well. First, to examine the influence of GTP-γ-s on the cellular production of NO in cells that express eNOS endogenously, BAEC were permeabilized with digitonin and pretreated with GTP-γ-s (10 μM) or vehicle. Cells were then loaded with the NO indicator dye, DAF-2DA, and BK-stimulated NO production was assessed by fluorescence microscopy. As seen in Fig. 6A (top panel micrographs), there was a time-dependent increase in cellular fluorescence intensity in response to BK. However, BK-induced NO production was markedly abrogated in cells pretreated with GTP-γ-s (middle panel micrographs in Fig. 6A), and further diminished in the presence of the NOS inhibitor, L-NAME, as well (bottom panel micrographs in Fig. 6A). Quantification of the fluorescent intensity in cells from digitized micrographs revealed a 60% decrease in NO production by GTP-γ-s in comparison to digitonin alone, in response to BK stimulation (lower graph in Fig. 6A). Next, experiments were performed to assess the effect of K44A overexpression on cellular NO production. eNOS-GFP ECV cells were transfected with empty pcDNA vector or K44A-V5 epitope-tagged vector, and cellular NO release was measured from the media using the fluorescent NO probe, DAF-2. Transfected cells were treated with either L-arginine or L-NAME and stimulated with BK or vehicle. As seen in Fig. 6B, overexpression of K44A markedly inhibited the cellular production of NO in response to BK. L-NAME inhibited NO production in all groups, indicating specificity of effect. Western blot control experiments from cell lysates, confirmed the expression of V5 epitope and overexpression of dyn-2 in cells transfected with the vector encoding K44A-V5 as compared to cells transfected with the empty vector, as well as similar eNOS protein levels throughout the groups. To further confirm the inhibitory influence of K44A on cellular NO production, we performed complimentary studies using NO-specific chemiluminescence, which assesses levels of the NO metabolite, nitrite. As seen in Fig. 6C, L-NAME-inhibited,BK-stimulated nitrite accumulation was reduced in cells transfected with K44A as compared to empty vector. Thus, studies using complimentary NO measurement,indicate that inhibition of cellular GTPases, particularly dynamin-2, abrogate BK-mediated NO production.
Microscopic and biochemical analyses have previously established that eNOS protein targets to specific membrane compartments within cells and is enriched within membrane fractions of cellular lysates(Blair et al., 1999; Fulton et al., 2002; Liu et al., 1995; Prabhakar et al., 2000; Sessa et al., 1995; Shaul et al., 1996). Further membrane fractionation analyses and co-localization studies indicate that these membrane domains include plasmalemmal caveolae and Golgi membranes(Blair et al., 1999; Fulton et al., 2002; Liu et al., 1995; Prabhakar et al., 2000; Sessa et al., 1995; Shaul et al., 1996). Furthermore, the subcellular localization of eNOS appears to be a dynamic event (Michel et al., 1993). For example, photobleaching studies by Sowa et al., indicate that eNOS quickly redistributes within a specific membrane domain and that alterations in homeostatic factors, such as cell confluency, markedly influence the cellular distribution of eNOS between different compartments(Sowa et al., 1999). Other studies have demonstrated that the cellular distribution of eNOS within distinct subcellular domains is influenced by specific extracellular stimuli,some, such as BK, capable of promoting NO production(Prabhakar et al., 1998; Thuringer et al., 2002; Venema et al., 1996). For example, Prabhaker et al. and Thuringer et al., independently observed that BK promoted the distribution of cellular eNOS from plasmalemmal caveolae into a more diffuse intracellular localization(Prabhakar et al., 1998; Thuringer et al., 2002). Consistent with a prior study by Liu et al.(Liu et al., 1995), we did not observe a reduction in the membrane distribution of eNOS as assessed by cell fractionation analysis, but did observe a change in the biochemical characteristics of eNOS characterized by enrichment of NOS protein into Triton X-100-insoluble fractions after BK treatment, as previously noted by Venema et al. (Venema et al., 1996). In the present studies, performed with both eNOS-GFP ECV 304 cells and BAEC,which express eNOS heterologously and endogenously, respectively, BK promotes the internalization of plasmalemmal eNOS. Our studies using CP and AP 180 overexpression indicate that BK-mediated eNOS internalization is a clathrin-independent event. These studies suggest that eNOS protein accumulates within internalized caveolin-coated vesicles which have budded from the plasma membrane. However, neither the current study nor the aforementioned earlier studies on this topic have been able to fully characterize the destination of these internalized vesicles. Furthermore,stimulus-dependent eNOS internalization does not appear to be a response limited to NOS agonists as incubation of endothelial cells with OxLDL, which,conversely to BK, inhibits cellular NO production, also promotes the internalization of eNOS from its basal plasmalemmal location(Blair et al., 1999). More recent studies suggest that the NOS associated protein, NOSTRIN, may play a role in plasmalemmal eNOS internalization as well(Zimmerman et al., 2002).
The precise role of nucleotide hydrolysis and dynamin-2 in the molecular steps of membrane scission, internalization, and ensuing retrograde vesicle trafficking remain an area of active investigation(Henley et al., 1998; McNiven, 1998; Oh et al., 1998; Schnitzer et al., 1996; Schekman and Orci, 1996). In the present studies, we chose to exploit the energy requirement of vesicle trafficking to study the mechanism by which perturbation of vesicle trafficking might influence eNOS distribution and NO production by using both pharmacological and molecular approaches, in conjunction with complimentary techniques, to measure cellular NO and assess NOS localization. Indeed, a similar approach was recently utilized to dissect the role of vesicle trafficking in the desensitization of G protein-coupled receptor signaling and eNOS-caveolin complex stability in cardiac myocytes(Dessy et al., 2000). We find that inhibition of GTP hydrolysis by GTP-γ-S impairs BK-dependent eNOS redistribution and NO production. Furthermore, perturbation of cells with dominant negative form of dynamin-2 mimic the inhibitory effects of GTP-γ-S on NO generation, indicating a particular importance of dynamin-2 in NOS trafficking and function. Dynamin-2 is of particular relevance as this protein colocalizes with caveolin and eNOS within caveolae and Golgi vesicles, and is essential in the scission and internalization of plasmalemmal vesicles (Cao et al.,2001; Henley et al.,1998; Oh et al.,1998). Our current observations are consistent with several prior studies, which have demonstrated that inhibition of dynamin GTPase function prevents internalization of caveolin-coated membrane vesicles(Henley et al., 1998; Oh et al., 1998; Pierce et al., 1999), and add to these studies by demonstrating that this process influences downstream signaling pathways, in this case relating to NOS activation. However, it is important to recognize that the influence of GTP-γ-S on eNOS trafficking may reflect inhibition of cellular GTPases in addition to dyn-2, which are essential to the processes of vesicle trafficking, such as Rab 3, Rab 5 and Rab 7 (Chavrier and Goud,1999). Several converging and diverging BK-mediated signaling pathways that culminate in NOS activation have been well delineated (Bernier et al., 275; Harris et al.,2001; Thuringer et al.,2002; Venema et al.,1996), and the interplay of these pathways with dynamin-dependent,BK-mediated eNOS internalization will require further elucidation.
We have previously demonstrated that dynamin-2 binds in a specific manner with eNOS and promotes the catalytic activity of recombinant eNOS protein in vitro (Cao et al., 2001). In the current studies, we anticipate that K44A dynamin is likely influencing NO production through effects on vesicle internalization rather than through direct binding actions. This concept is supported by the observation that the eNOS binding domain within dynamin, resides within the carboxy-terminal proline-rich domain of dynamin-2, rather than the amino-terminal GTPase domain, wherein lies the K44A mutation (Cao et al., 2003). Furthermore, colocalization of eNOS and dynamin in cells is most prominent in the Golgi membranes(Cao et al., 2001), where a significant and bioactive pool of both proteins reside(Fulton et al., 2002), rather than within plasma membrane, where the influence of K44A appears to be most prominent. It is possible that direct binding regulation between eNOS and dynamin occurs in the Golgi membranes while dynamin-dependent endocytosis selectively influences the plasmalemmal pool of eNOS.
In conclusion, these studies, using complimentary cell lines and methodologies, demonstrate that perturbation of specific vesicle trafficking pathways has the capability to influence cellular eNOS distribution and the ensuing NO production. These studies add to the current understanding of eNOS biology by demonstrating that BK-dependent eNOS trafficking is a GTPase-dependent process, with particular importance of dynamin-2 GTPase, and that perturbation of this process impairs the cellular production of NO. Additionally, they provide further evidence that dynamin-dependent endocytosis is intimately linked to downstream cell signaling events.
These studies were supported by grants from the NIH to V.S. (R01 DK-59388,R01 DK-59615), grants from the American Heart Association to V.S. and S.C.,and the Mayo Clinic Foundation. The authors acknowledge Dr M. McNiven and Dr H. McMahon for providing dynamin-2 anti-sera and the dynamin-2 K44A cDNA, and AP 180 cDNA, respectively, as well as Dr W. Sessa for providing the eNOS-GFP ECV 304 cells.