PECAM-1 Interacts With Nitric Oxide Synthase in Human Endothelial Cells
http://www.100md.com
《动脉硬化血栓血管生物学》
N. Dusserre; N. L’Heureux; K.S. Bell; H.Y. Stevens; J. Yeh; L.A. Otte; L. Loufrani; J.A. Frangos
From the Department of Bioengineering (N.D., N.L., L.L., J.A.F.), University of California San Diego, and the Bioengineering Institute (K.S.B., H.Y.S., J.Y., L.A.O., J.A.F.), La Jolla, Calif.
ABSTRACT
Objective— We have previously shown that fluid shear stress (FSS) triggers endothelial nitric oxide synthase (eNOS) activity in endothelial cells and that the mechanotransduction mechanisms responsible for activation discriminate between rapid changes in FSS and FSS per se. We hypothesized that the particular sublocalization of eNOS at the cell–cell junction would render it responsive to activation by FSS temporal gradients.
Methods and Results— In human umbilical vein endothelial cells (HUVECs), immunofluorescence revealed strong eNOS membrane staining at the cell–cell junction colocalizing with platelet/endothelial cell adhesion molecule-1 (PECAM-1). In PECAM-1 –/– mouse aorta, eNOS junctional localization seen in the wild type was absent. Similarly, junctional staining was lost in wild-type aorta near intercostal artery branches. eNOS/PECAM-1 association in HUVECs was confirmed by coimmunoprecipitation. When HUVECs were subjected to a 0.5s impulse of 12 dynes/cm2, a transient disruption of the eNOS/PECAM-1 complex was observed, accompanied by an increase in eNOS activity (cGMP production). Ramped flow did not trigger complex dissociation or an increase in cGMP production. In a cell-free system, a direct inhibition of eNOS activity by PECAM-1 is shown.
Conclusions— These results suggest that eNOS is complexed with PECAM-1 at the cell–cell junction and is likely involved in the modulation of eNOS activity by FSS temporal gradients but not by FSS itself.
We hypothesized that the particular sublocalization of eNOS at the cell–cell junction would render it responsive to activation by FSS temporal gradients. Our results suggest that eNOS is complexed with PECAM-1 at the cell–cell junction and is likely involved in the modulation of eNOS activity by FSS temporal gradients.
Key Words: shear stress ? temporal gradients ? eNOS ? PECAM-1 ? endothelial
Introduction
Nitric oxide (NO) produced by the endothelial isoform of NO synthase (eNOS) is a fundamental component of cardiovascular homeostasis. It regulates systemic blood pressure, vascular remodeling, and angiogenesis.1,2 The most important physiological stimulus for the formation of NO is the fluid shear stress (FSS) continuously generated by the flow of blood on the endothelial cell (EC) layer. The importance of this mechanical regulation is particularly well illustrated by the preferential distribution of atherosclerotic lesions in regions where endothelium is exposed to rapidly changing levels of FSS (high temporal gradients ).3,4
In vitro studies have shown a biphasic production of NO in response to flow. Sudden onset of flow induces a burst of NO production which is both calcium- (Ca2+) and G-protein–dependent. In contrast, steady FSS that follows induces a lower but sustained release of NO and is both Ca2+- and G-protein–independent.5 Moreover, the initial and transient production of NO is directly related to the rate of change in FSS rather than to its absolute magnitude, whereas the sustained release of NO is directly related to the level of FSS.5 Hence, shear-induced NO production appears to be a superposition of two independent mechanochemical pathways.6 However, the initial molecular events enabling eNOS to respond discriminantly to steady and unsteady flow are still not understood.
We hypothesized that the plasma membrane is a site of FSS mechanochemical transduction and that different plasmalemmal molecular mechanisms are involved in the transduction of constant FSS versus TG in FSS. We speculated that rapid changes in FSS will deform the plasma membrane at the interface with rigid domains. Such membrane deformation would lead to membrane protein signaling.7 Interestingly, caveolae, which, in contrast to the major part of the plasma membrane, appear as more rigid liquid-ordered domains, have been described as sites of eNOS activation after FSS.8 The glycocalyx has also been implicated as a mechanosensor structure.9,10 However, the role of glycocalyx components in modifying the shear stress response in ECs is still unclear. The cell–cell junction is also such a rigid domain, characterized by a marked concentration of membrane-associated cytoskeletal elements, as well as numerous intercellular bridges that closely juxtapose the two plasma membranes of neighboring cells (platelet/endothelial cell adhesion molecule-1 junction, adherence junction, tight junction, and gap junction).11 Therefore, we hypothesized that eNOS involved in the mechanotransduction of rapid changes in FSS would be localized at or near the cell–cell junction. In contrast, we expect that steady FSS induces diffuse increases in membrane fluidity on the apical plasma membrane. Hence, eNOS involved in the mechanotransduction of steady FSS would be rather broadly distributed along the plasma membrane.
PECAM-1 is a transmembrane glycoprotein primarily localized at the cell–cell junction of ECs, where it is involved in homophilic interactions with PECAM-1 of neighboring ECs.12 Interestingly, PECAM-1 has been shown to be tyrosine-phosphorylated in response to FSS stimulation.13 Furthermore, it has been suggested that PECAM-1 may function as an inhibitory receptor, interacting with various signaling molecules through its Src homology (SH2) containing protein tyrosine phosphatase (SHP-2) binding domain.14 Therefore, we hypothesized that PECAM-1 might interact with eNOS in primary ECs, hence providing a compartmentalization of eNOS with other signaling proteins, while localizing the enzyme in an area prone to sense any sudden change in membrane tension after a rapid change in FSS
The present study describes a series of experiments which have explored eNOS cell–cell junction localization, eNOS interaction with PECAM-1, the dynamics of this association after different FSS stimulations, and concomitant eNOS activation.
Methods
Cell Culture and FSS
Harvested human umbilical vein endothelial cells (HUVECs) were plated onto slides as described.15 Cells were starved overnight before the experiment in DMEM plus 1% BSA.
Slides were assembled into parallel plate flow chambers.16 Starvation medium was used as perfusion medium. Exposure of the cells to laminar FSS was accomplished using a syringe pump (Harvard Apparatus) driven by a computer, as described.17 At the end of varying periods of FSS, cells were dipped into ice-cold PBS and immediately lysed. Sham control cells were mounted onto the chamber but not subjected to shear.
Animals
All procedures were in accordance with institutional guidelines for animal research. PECAM-1 knockout mice (B6;129S-PECAMtm1Lex) and wild-type mice (B6;129SF2/J) were purchased from the Jackson Laboratory (Bar Harbor, Me). Animals were genotyped at the Jackson Laboratory and any offspring were genotyped in our laboratory.
Immunofluorescence
HUVECs on slides were fixed in 4% PBS-buffered paraformaldehyde (PFA) for 45 minutes at 4°C, treated with 25 mmol/L NH4Cl, and then permeabilized with 0.1% Triton X-100. For double labeling, HUVECs were successively incubated with: (1) 10% normal serum, (2) goat anti–PECAM-1 IgG (Santa Cruz Biotechnology, Santa Cruz, Calif; 1/50) and rabbit anti-eNOS IgG (Santa Cruz Biotechnology, 1/25), and (3) donkey anti-goat IgG (Alexa 488, Molecular Probes, Eugene, Ore) and donkey anti-rabbit (Alexa 568, Molecular Probes). Cells were washed with PBS between incubations. Slides were mounted with "Slow-Fade" reagent (Molecular Probes).
For en face deconvolution microscopy, mice were anesthetized with Ketamine 90 mg/kg/Xylazine 10 mg/kg, IP, and then perfused with ice-cold 4% PFA (15 minutes, 4°C). Aortas were carefully dissected; cross sectional segments were cut and then opened longitudinally. Branches and blood flow direction were recorded. Permeabilization and staining steps were similar to those for HUVECs, except nuclei were stained with DAPI (Molecular Probes). Preparations were mounted in Gelvatol.
Slides were observed under a DeltaVision deconvolution microscope system (Applied Precision Inc.) (For deconvolution methodology, see Falk).18
Please see http://atvb.ahajournals.org for a description of similar studies performed using en face confocal microscopy on rat aorta.
Cellular Lysates
HUVECs were harvested in octyl glucoside (OG) buffer (60 mmol/L OG, 50 mmol/L Tris·HCl, pH 7.4, 125 mmol/L NaCl, 2 mmol/L dithiothreitol, 50 μmol/L EGTA) with protease inhibitors and sonicated. Lysates were centrifuged to remove insoluble material and stored at –80°C.
Protein Assay
Protein in cell lysates was determined by a BCA protein assay (BioRad).
Coimmunoprecipitation
Aliquots of cell lysates (500 μg protein) were precleared for 1.5 hour (4°C) with Protein G-Agarose beads (Santa Cruz Biotechnology). Supernatants were incubated overnight with goat anti–PECAM-1 (Santa Cruz Biotechnology) or rabbit anti-eNOS (Santa Cruz Biotechnology) polyclonal antibodies at a final concentration of 8 and 4 μg/mL, respectively. Antibody titration experiments (not shown) demonstrated that these concentrations led to the quantitative immunoprecipitation of PECAM-1 and eNOS, respectively, from cell lysates. Protein G–Agarose beads were added for an additional 1.5 hour incubation (4°C). Bound immune complexes were washed 3 times with OG buffer and once with 50 mmol/L Tris · HCl (pH 7.4), 150 mmol/L NaCl. Immunoprecipitated proteins were released by boiling in sodium dodecyl sulfate (SDS) sample buffer for 5 minutes.
Immunoblot
Denatured immunoprecipitated proteins were separated on 7.5% SDS-PAGE gels and transferred to polyvinylidene fluoride (PVDF). After blocking with 5% nonfat milk in Tris-buffered saline 0.1% Tween 20 (TBST), membranes were incubated with primary antibodies against eNOS (monoclonal, Transduction Labs) or PECAM-1 (Santa Cruz Biotechnology) for 1 hour in TBST/milk. Bound primary antibodies were detected with horseradish peroxidase–conjugated secondary antibodies: goat–anti-mouse IgG (Pierce, 1/5000) and rabbit anti-goat IgG (Santa Cruz Biotechnology, 1/2500), respectively, followed by chemiluminescent SuperSignal substrate (Pierce). Band intensity was quantified on unsaturated X-ray film by a digital image analyzer (Quantity-One; BioRad). All comparisons were made relative to sham controls.
cGMP Accumulation
NO production in response to FSS was assessed by measuring intracellular accumulation of cGMP. HUVECs were treated with 0.5 mmol/L isobutylmethylxanthine (IBMX) for 30 minutes (to inhibit phosphodiesterases) and 6 mmol/L L-Arginine (substrate of eNOS) for 1 hour including FSS stimulation. Control cells (2 μmol/L ionomycin in DMSO or DMSO only) were used in each experiment. Subsequent steps were performed on ice with 0.5 mmol/L IBMX to prevent cGMP degradation. Cells were scraped in 500 μL 65% EtOH. Resulting extracts were centrifuged to eliminate the insoluble fraction. Supernatants were dried at 60°C in a SpeedVac and stored at –80°C. cGMP levels were measured using a cGMP enzyme immunoassay, following the acetylation procedure as recommended by the supplier (Amersham Pharmacia Biotech). To normalize cGMP values, protein content was measured in each pellet.
eNOS Activity Assay
Please see http://atvb.ahajournals.org for a description of eNOS Activity Assay performed in a cell-free system.
Production of CD31 Cytoplamic Domain GST-Fusion Protein
Please see http://atvb.ahajournals.org for a description of the production of CD31 (PECAM-1) cytoplasmic domain GST-fusion protein, required for the eNOS Activity Assay.
Data Analysis
Data are presented as mean±SEM of the indicated number of observations or as percentage of control. Statistical comparisons between groups were performed using the Student t test with significance level of P<0.05.
Results
eNOS Colocalized With PECAM-1 at the Periphery of ECs In Vitro and In Vivo
To determine whether eNOS and PECAM-1 colocalize, a confocal double-label immunofluorescence study was performed on freshly isolated, confluent, serum-starved HUVECs. A striking eNOS staining was present along the cell periphery in addition to Golgi-like labeling (Figure 1A). This peripheral staining suggests a localization of eNOS at the cell–cell junction. Indeed, immunolocalization of PECAM-1, a major component of the ECs intercellular junction, demonstrates an almost identical peripheral pattern (Figure 1B), and PECAM-1/eNOS double staining demonstrates clear colocalization at the cell periphery (Figure 1C). No fluorescence was detected in negative control slides in which normal serum was substituted for the primary antibody or in cells incubated with secondary antibody alone (data not shown). Similar controls were included within each of the following sections of our immunostaining study with similar results (data not shown).
Figure 1. Major colocalization of PECAM-1 and eNOS in HUVECs in vitro. In serum-starved confluent primary HUVECs, double-immunofluorescence staining revealed peripheral eNOS (A) and PECAM-1 (B) expression. Superposition of eNOS and PECAM-1 staining demonstrated colocalization in yellow-orange (C). Bar=25 μm
As confluence increased, the emergence of a strong eNOS membrane staining was synchronous with the appearance of PECAM-1 staining at the cell–cell junction (data not shown).
To confirm the physiological relevance of the results obtained in our in vitro model, the same staining procedure was performed on rat and mouse aortas. Please see http://atvb.ahajournals.org for results on rat tissue.
eNOS Was Present at the Cell Periphery in PECAM-1 +/+ Mouse Aorta But Was Absent in the Knockout Mouse
Deconvolution microscopy was used to visualize eNOS/PECAM-1 immunolocalization on a straight section of aorta from PECAM-1 knockout and wild-type mice (Figure 2). Antibodies and fixative used were the same as above. In the wild-type mouse aorta, eNOS staining presented the same pattern as the confluent HUVECs: a Golgi labeling and a strong peripheral staining (Figure 2B). PECAM-1 was also detected at the cell periphery (Figure 2A) and a predominant colocalization of the two proteins was observed at the cell–cell junction (Figure 2C). In contrast, in PECAM-1 knockout mouse, the strong eNOS Golgi-like labeling remained but the peripheral staining was absent (Figure 2E and 2F). The absence of PECAM-1 protein in the aorta of PECAM-1 knockout mouse was confirmed (Figure 2D). These results clearly demonstrate that eNOS peripheral staining in mouse aorta endothelium is dependent on the presence of PECAM-1 at the cell–cell junction.
Figure 2. eNOS distribution is altered in PECAM-1 knockout mouse aorta. ECs located in the straight section of the mouse aorta are shown. Staining was performed on wild-type (WT) mouse aorta in A, B, and C, and on PECAM-1 knockout (KO) mouse aorta in D, E, and F. Double staining was performed to localize PECAM-1 in A and D and eNOS in B and E. Superposition of eNOS and PECAM-1 staining demonstrates colocalization in yellow-orange (C and F). Cell nuclei were labeled using DAPI (blue). Bar=15 μm
In Wild-Type Mouse Aorta, Peripheral eNOS Staining Was Absent From the Branch Point of the Intercostal Arteries
In straight sections of the aorta, endothelium is reportedly primarily exposed to unidirectional FSS with a high mean and small TG. In branched arteries, the ECs are believed to experience large TG in FSS with a low mean.3,4 Deconvolution microscopy was used to visualize eNOS/PECAM-1 immunolocalization near an intercostal artery branch in wild-type mouse aorta (Figure 3). Although anti–PECAM-1 stained the cell–cell junction strongly (Figure 3A), eNOS junctional staining was reduced (Figure 3B). However, the strong Golgi-like eNOS labeling was still present. Consequently, eNOS/PECAM-1 colocalization was barely detected at the cell–cell junction (Figure 3C). Such results suggest that chronic TGs in FSS disrupt eNOS/PECAM-1 colocalization.
Figure 3. eNOS distribution is altered in mouse aorta branched areas exposed to high TGs in shear. ECs located near an intercostal artery branch of the aorta are shown. Double staining was performed to localize PECAM-1 in A and eNOS in B. Superposition of eNOS and PECAM-1 staining demonstrates colocalization in yellow-orange (C). Cell nuclei are labeled using DAPI (blue) Bar=15 μm
eNOS Coimmunoprecipitated With PECAM-1 and PECAM-1 Interaction Decreased eNOS Activity In Vitro
Based on the prominent colocalization of eNOS and PECAM-1 within confluent HUVECs, we next investigated whether the two proteins interacted biochemically at the cell–cell junction. For this purpose, detergent-soluble lysates were prepared from confluent serum-starved HUVECs, and both eNOS and PECAM-1 were immunoprecipitated. OG detergent preserves only direct protein–protein interactions. To further enhance the stringency of the immunoprecipitation process, a high ionic strength wash buffer, containing 150 mmol/L NaCl, was used. As seen in Figure 4, immunoprecipitation of eNOS coimmunoprecipitated PECAM-1 from HUVECs extracts (Figure 4A), and conversely, immunoprecipitation of PECAM-1 coimmunoprecipitated eNOS (Figure 4B).
Figure 4. Association of PECAM-1 with eNOS and its effect on eNOS activity in a cell-free system. eNOS or PECAM-1 were immunoprecipitated from HUVECs using polyclonal antibodies. Samples were prepared for Western blot (WB) analysis. A, Under nondenaturing conditions, immunoprecipitation of eNOS coprecipitates PECAM-1 in the presence of 60 mmol/L OG detergent and 150 mmol/L NaCl wash. B, Under identical conditions, immunoprecipitation of PECAM-1 coprecipitates eNOS. C, eNOS activity was measured in the presence of varying concentrations of either GST or GST-PECAM-1 cytoplasmic domain fusion proteins at 37°C for 30 minutes. L-NAME, a competitive inhibitor of eNOS, was also used as control. The experiments were performed in duplicate, and eNOS activity was presented as net radioactivity after subtracting the value for the negative control. The effect of GST-PECAM-1 on eNOS activity was compared with GST at the same concentration using a t test assuming unequal variances. Significance at P<0.05 level, two-tailed, was achieved at 20 μmol/L GST-PECAM-1.
To determine the effects of PECAM-1 interaction with eNOS in vitro, we measured the effect of PECAM-1 on eNOS activity using recombinant proteins. We found that 20 μmol/L GST-PECAM-1 inhibited eNOS activity by 36% compared with 20 μmol/L GST alone (Figure 4C).
FSS Caused eNOS Dissociation From PECAM-1
To determine whether the eNOS/PECAM-1 complex is indeed affected by FSS, highly confluent serum-starved HUVECs were subjected to three different profiles of FSS. TGs in FSS were generated by a sudden onset of FSS (0.5s at 12 dynes/cm2 or 24 dynes/cm2), which we have referred to as an "impulse" in previous studies.6 To apply similar levels of FSS without significant TG, we used a slow ramping in flow over 30s, which reached 24 dynes/cm2 and stayed at this level for 0.5s before ramping down to 0 dynes/cm2 over 30s. Repeated high shear gradients experienced by ECs at branch points were simulated by "pulsatile flow" (repeated impulses; 0.5s every 3 seconds for 30 minutes at 12 dynes/cm2).
A 0.5s impulse of 12 dynes/cm2 resulted in a rapid disruption of the eNOS/PECAM-1 complex (Figure 5A and 5B) within 7s, which was maximal after 15s. After that time, the proteins rapidly reassociated. The reassociation of the complex reached a maximum between 45s and 1 minute. The exact time of the complex reformation varied slightly between these two time points from experiment to experiment, a variation which we attributed to differences in HUVEC batches, but was consistent within duplicate analyses of a single batch. The levels of eNOS immunoprecipitated with PECAM-1 subsequently returned to their prestimulation level. Sham controls (0s in Figure 5A and 5B) were used to normalize immunoprecipitation results. This allowed us to eliminate artifact due to any mechanical stimulation incidentally applied to the cells while mounting the slides on the flow chamber, as well as artifact possibly induced by postshear treatment of the cells. Detergent-soluble lysates were prepared at various times after FSS and PECAM-1 protein was quantitatively immunoprecipitated. As shown in Figure 5C, where eNOS and PECAM-1 binding was assessed 30s after the end of the FSS, the slowly ramping flow (24 dynes/cm2) did not trigger eNOS/PECAM-1 dissociation, although it generated much more shear than the impulse. This clearly demonstrates that eNOS/PECAM-1 dissociation is triggered by a TG of FSS rather than by the FSS itself.
Figure 5. Single FSS impulse and pulsatile FSS stimulations induce a transient dissociation of PECAM-1/eNOS, whereas slowly ramping the FSS does not. Confluent serum-starved HUVECs were subjected to an impulse in FSS of 12 dynes/cm2 for 0.5s. Cell lysates were prepared for PECAM-1 immunoprecipitation and gel electrophoresis. A, A densitometric analysis of five independent experiments reveals the association-dissociation kinetics of PECAM-1/eNOS after the impulse. eNOS densitometry results were normalized for PECAM-1 detected by a second immunostaining performed on the same blot. Results are percentage of sham control. Values are means±SEM (*P<0.05 from sham control). B, A representative Western blot (WB) is shown, demonstrating that eNOS transiently dissociates from PECAM-1 after the impulse in FSS. Dissociation of both molecules was maximal around 15s to 30s. The association was restored 1 minute after FSS. C, A densitometric analysis of four independent experiments reveals the absence of PECAM-1/eNOS dissociation, 30s after the stimulation, when the stimulation was a slow ramp in FSS as opposed to an impulse. D, A densitometric analysis of three independent experiments reveals a partial PECAM-1/eNOS dissociation 30s after pulsatile flow (12 dynes/cm2, 0.5s impulse, 3s intervals for 30 minutes). In C and D, values are means±SEM (*P<0.05 and **P>0.05 from sham control).
Cells subjected to pulsatile flow for 30 minutes showed a partial dissociation of the eNOS/PECAM-1 complex (Figure 5D). This is consistent with the in vivo observation of a reduced junctional eNOS staining near intercostal artery branches, where ECs are exposed to repeated high TGs in FSS. Results observed after a single impulse in FSS were similar to those obtained after a pulsatile FSS. This suggests that data obtained from single impulse studies are good indicators of the in vivo effects of high TGs in FSS. However, the response to the pulsatile FSS was less marked than the response to the single impulse. This could indicate some degree of acclimatization of the cells to a repeated impulse stimulus.
A 2-Fold Augmentation of cGMP Accumulation Was Observed After Impulse
The role of eNOS/PECAM-1 dissociation on eNOS activity was indirectly measured by assessing the cGMP amount accumulated within HUVECs at various times after FSS. A nearly 2-fold augmentation of cGMP accumulation was observed 1 minute after an impulse in FSS compared with sham controls (Figure 6A). No further FSS-induced cGMP production was detected after this time, indicating that cGMP synthesis is very rapidly triggered and downregulated within the first minute after FSS. This time frame is identical to the one of the eNOS/PECAM-1 complex dissociation–reassociation. No augmentation of cGMP production was observed during the first five minutes after slow ramping in flow (24 dynes/cm2), but large quantities of cGMP were produced when cells were exposed to an impulse of the same intensity (Figure 6B).
Figure 6. An impulse in FSS induces a concomitant augmentation of eNOS activity. Slowly ramping the FSS does not trigger such an increase in eNOS activity. NO production from HUVECs in response to FSS was assessed by the intracellular accumulation of cGMP. Confluent serum-starved HUVECs were incubated with 0.5 mmol/L IBMX for 30 minutes and 6 mmol/L L-Arginine for 1 hour. Cells were then subjected to an impulse in FSS of 12 dynes/cm2 for 0.5s in A, and to an impulse in FSS of 24dynes/cm2 for 0.5s or a slow ramping toward a FSS of 24dynes/cm2 in B. At different times after impulse or ramping (0, 30, 60, 120, or 300s), HUVECs were lysed and cellular extracts were generated. As shown in C, in each of these experiments, control cells were stimulated for 15 minutes with 2 μmol/L ionomycin or with the vehicle alone (DMSO). cGMP levels in the extracts were normalized for cellular proteins content. cGMP production values are percentages of sham control. A and C display means (±SEM) from 3 independent experiments (*P<0.05, **P>0.05). B shows one representative experiment of two.
Discussion
This study is the first to demonstrate, by means of coimmunoprecipitation and in situ immunolocalization, that eNOS and PECAM-1 are closely associated in human primary ECs in vitro and mouse and rat aortal tissue in vivo. This suggests that PECAM-1 is responsible for the localization of eNOS at the cell–cell junction. Additional compelling experimental evidence to support this conclusion is that eNOS peripheral staining was absent in the endothelium of PECAM-1 knockout mice in situ, whereas it was present in the wild-type controls. In addition, results showed that eNOS/PECAM-1 association is rapidly and transiently disrupted after a sudden change in FSS but not after steady FSS in ECs. These in vitro data were corroborated by in situ immunostainings demonstrating the absence of eNOS/PECAM-1 colocalization in wild-type mice endothelium in regions exposed to rapidly changing levels of FSS (high TGs). Moreover, the dissociation of the eNOS/PECAM-1 complex was paralleled by an augmentation in eNOS activity in freshly harvested ECs. Finally, in a cell-free system we were able to show a direct inhibition of NOS activity by PECAM-1. The fact that PECAM-1, in the absence of cells, is able to directly affect eNOS activity, suggests that these proteins not only colocalize but also interact and influence protein function. This is further evidence of the effect of PECAM-1 on eNOS protein localization and activity. Therefore, we provide the first evidence to suggest the involvement of an eNOS/PECAM-1 interaction in FSS-induced mechanotransduction.
The eNOS/PECAM-1 colocalization we observed in human primary endothelial cells mimics the staining pattern previously reported in immortalized murine microvascular ECs and native rat cardiac endothelium.19,20 Its physiological relevance is further strengthened here by the demonstration of eNOS/PECAM-1 colocalization in rat and mouse aorta. The conservation of this colocalization through three species suggests that it has an important function. The correlation between cell confluence, PECAM-1 appearance, and eNOS presence at the cell–cell junction confirms published observations using an immortalized murine EC line.19 In addition, the results obtained with PECAM-1 knockout mice are consistent with the notion that PECAM-1 presence at the cell–cell junction is necessary for eNOS peripheral localization. The colocalization of eNOS and PECAM-1 immunostaining is corroborated by the coimmunoprecipitation of these two molecules. This procedure was performed under highly stringent conditions which preserve only direct protein–protein bindings, eliminating any weak interactions. Together with the demonstration that PECAM-1 directly inhibits eNOS activity in the purified protein preparation (Figure 4), this strongly suggests that the two proteins form a direct and tight association.
The close association of eNOS with PECAM-1 is important in light of recent developments in PECAM-1 biology. PECAM-1 has recently been suggested to function as a mechanotransducing protein13 and has been shown to be tyrosine-phosphorylated within 1 minute in ECs exposed to a "step change" in FSS.13 Furthermore, the activation of extracellular signal regulated kinase (ERK), one of the early EC responses to FSS, has been shown to be dependent on PECAM-1 phosphorylation. The recruitment and binding of SHP-2 to phospho-PECAM-1 at the cell periphery eventually triggers activation of ERK.13 Interestingly, we have recently shown that ERK activation is more specifically sensitive to TG in FSS.15 The tyrosine phosphorylation of PECAM-1 occurs in a time frame similar to that of the eNOS/PECAM-1 dissociation observed in our study. A recent article by Jin et al suggests that flow-stimulated tyrosine kinases recruit phosphoinositide 3-kinase and mediate activation of Akt and eNOS.21 Whether, and to what extent, these events are intertwined is presently unknown.
The fact that the eNOS/PECAM-1 was not affected by a slow ramping FSS, even at high shear levels, strongly suggests that this mechanoreceptive system is specific to rapid changes in FSS. The localization of the complex at the cell–cell junction is consistent with this discriminatory ability because it is located in a region of the membrane presumably sheltered from the FSS itself. In fact, one would expect mechanotransductory molecules sensitive to the FSS (and not the TG of shear stress) to be concentrated at the apex of the apical surface of the cell, where the FSS is maximal.22
During recent years, several groups have demonstrated that activation of eNOS, both in conjunction with and independent of intracellular calcium flux, occurs through the allosteric binding of eNOS with neighboring regulatory proteins. In this regard, several eNOS-associated proteins have been identified, including calmodulin, Caveolin-1, Hsp 90, dynamin, NOSIP, and some G protein–coupled receptors (for detailed review see Fleming et al).23 In addition, most of the kinases known to phosphorylate eNOS, notably on serine 1177, are believed to be concentrated near eNOS through a scaffold protein. Our data suggest that PECAM-1 is also involved in an eNOS signaling complex at the cell–cell junction. Currently, we are investigating which other known eNOS-associated proteins might be present in the eNOS/PECAM-1 complex.
In conclusion, we have shown that eNOS is closely associated with PECAM-1 at the cell–cell junction and that this complex is dissociated by TG in FSS but not by FSS itself. Current studies focus on the role of intracellular calcium and G proteins on eNOS/PECAM-1 association as well as on eventual interactions of eNOS/PECAM-1 with previously described eNOS-associated proteins.
Acknowledgments
The authors thank the University of California, San Diego Cancer Center Digital Imaging Shared Resource staff for their assistance with the deconvolution studies. This work was supported by National Institutes of Health grant HL40696.
References
Ignarro LJ, Cirino G, Casini A, Napoli C. Nitric oxide as a signaling molecule in the vascular system: an overview. J Cardiovasc Pharmacol. 1999; 34: 879–886.
Davies PF. Flow-mediated endothelial mechanotransduction. Physiol Rev. 1995; 75: 519–560.
Zarins CK, Giddens DP, Bharadvaj BK, Sottiurai VS, Mabon RF, Glagov S. Carotid bifurcation atherosclerosis. Quantitative correlation of plaque localization with flow velocity profiles and wall shear stress. Circ Res. 1983; 53: 502–514.
Ku DN, Giddens DP, Zarins CK, Glagov S. Pulsatile flow and atherosclerosis in the human carotid bifurcation. Positive correlation between plaque location and low oscillating shear stress. Arteriosclerosis. 1985; 5: 293–302.
Kuchan MJ, Frangos JA. Role of calcium and calmodulin in flow-induced nitric oxide production in endothelial cells. Am J Physiol. 1994; 266: C628–C636.
Frangos JA, Huang TY, Clark CB. Steady shear and step changes in shear stimulate endothelium via independent mechanisms-superposition of transient and sustained nitric oxide production. Biochem Biophys Res Commun. 1996; 224: 660–665.
Gudi S, Nolan JP, Frangos JA. Modulation of GTPase activity of G proteins by fluid shear stress and phospholipid composition. Proc Natl Acad Sci U S A. 1998; 95: 2515–2519.
Rizzo V, McIntosh DP, Oh P, Schnitzer JE. In situ flow activates endothelial nitric oxide synthase in luminal caveolae of endothelium with rapid caveolin dissociation and calmodulin association. J Biol Chem. 1998; 273: 34724–34729.
Pohl U, Herlan K, Huang A, Bassenge E. EDRF-mediated shear-induced dilation opposes myogenic vasoconstriction in small rabbit arteries. Am J Physiol. 1991; 261: H2016–H2023.
Hecker M, Mulsch A, Bassenge E, Busse R. Vasoconstriction and increased flow: two principal mechanisms of shear stress-dependent endothelial autocoid release. Am J Physiol. 1993; 265: H828–H833.
Klymkowsky MW, Parr B. The body language of cells: the intimate connection between cell adhesion and behavior. Cell. 1995; 83: 5–8.
Newton JP, Buckley CD, Jones EY, Simmons DL. Residues on both faces of the first immunoglobulin fold contribute to homophilic binding sites of PECAM-1/CD31. J Biol Chem. 1997; 272: 20555–20563.
Osawa M, Masuda M, Kusano K, Fujiwara K. Evidence for a role of platelet endothelial cell adhesion molecule-1 in endothelial cell mechanosignal transduction: is it a mechanoresponsive molecule? J Cell Biol. 2002; 158: 773–785.
Newman PJ, Newman DK. Signal transduction pathways mediated by PECAM-1: new roles for an old molecule in platelet and vascular cell biology. Arterioscler Thromb Vasc Biol. 2003; 23: 953–964.
Bao X, Lu C, Frangos JA. Mechanism of temporal gradients in shear-induced ERK1/2 activation and proliferation in endothelial cells. Am J Physiol. 2001; 281: H22–H29.
Kuchan MJ, Jo H, Frangos JA. Role of G proteins in shear stress-mediated nitric oxide production by endothelial cells. Am J Physiol. 1994; 267: C753–C758.
Gudi SR, Clark CB, Frangos JA. Fluid flow rapidly activates G proteins in human endothelial cells. Involvement of G proteins in mechanochemical signal transduction. Circ Res. 1996; 79: 834–839.
Falk MM. Connexin-specific distribution within gap junctions revealed in living cells. J Cell Sci. 2000; 113: 4109–4120.
Govers R, Bevers L, de Bree P, Rabelink TJ. Endothelial nitric oxide synthase activity is linked to its presence at cell–cell contacts. Biochem J. 2002; 361: 193–201.
Andries LJ, Brutsaert DL, Sys SU. Nonuniformity of endothelial constitutive nitric oxide synthase distribution in cardiac endothelium. Circ Res. 1998; 82: 195–203.
Jin Z, Ueba H, Tanimoto T, Lungu AO, Frame MD, Berk BC. Ligand-independent activation of vascular endothelial growth factor receptor 2 by fluid shear stress regulates activation of endothelial nitric oxide synthase. Circ Res. 2003; 93: 354–363.
Barbee KA, Mundel T, Lal R, Davies PF. Subcellular distribution of shear stress at the surface of flow-aligned and nonaligned endothelial monolayers. Am J Physiol. 1995; 268: H1765–H1772.
Fleming I, Busse R. Molecular mechanisms involved in the regulation of the endothelial nitric oxide synthase. Am J Physiol. 2003; 284: R1–R12.(Implication for Flow-Indu)
From the Department of Bioengineering (N.D., N.L., L.L., J.A.F.), University of California San Diego, and the Bioengineering Institute (K.S.B., H.Y.S., J.Y., L.A.O., J.A.F.), La Jolla, Calif.
ABSTRACT
Objective— We have previously shown that fluid shear stress (FSS) triggers endothelial nitric oxide synthase (eNOS) activity in endothelial cells and that the mechanotransduction mechanisms responsible for activation discriminate between rapid changes in FSS and FSS per se. We hypothesized that the particular sublocalization of eNOS at the cell–cell junction would render it responsive to activation by FSS temporal gradients.
Methods and Results— In human umbilical vein endothelial cells (HUVECs), immunofluorescence revealed strong eNOS membrane staining at the cell–cell junction colocalizing with platelet/endothelial cell adhesion molecule-1 (PECAM-1). In PECAM-1 –/– mouse aorta, eNOS junctional localization seen in the wild type was absent. Similarly, junctional staining was lost in wild-type aorta near intercostal artery branches. eNOS/PECAM-1 association in HUVECs was confirmed by coimmunoprecipitation. When HUVECs were subjected to a 0.5s impulse of 12 dynes/cm2, a transient disruption of the eNOS/PECAM-1 complex was observed, accompanied by an increase in eNOS activity (cGMP production). Ramped flow did not trigger complex dissociation or an increase in cGMP production. In a cell-free system, a direct inhibition of eNOS activity by PECAM-1 is shown.
Conclusions— These results suggest that eNOS is complexed with PECAM-1 at the cell–cell junction and is likely involved in the modulation of eNOS activity by FSS temporal gradients but not by FSS itself.
We hypothesized that the particular sublocalization of eNOS at the cell–cell junction would render it responsive to activation by FSS temporal gradients. Our results suggest that eNOS is complexed with PECAM-1 at the cell–cell junction and is likely involved in the modulation of eNOS activity by FSS temporal gradients.
Key Words: shear stress ? temporal gradients ? eNOS ? PECAM-1 ? endothelial
Introduction
Nitric oxide (NO) produced by the endothelial isoform of NO synthase (eNOS) is a fundamental component of cardiovascular homeostasis. It regulates systemic blood pressure, vascular remodeling, and angiogenesis.1,2 The most important physiological stimulus for the formation of NO is the fluid shear stress (FSS) continuously generated by the flow of blood on the endothelial cell (EC) layer. The importance of this mechanical regulation is particularly well illustrated by the preferential distribution of atherosclerotic lesions in regions where endothelium is exposed to rapidly changing levels of FSS (high temporal gradients ).3,4
In vitro studies have shown a biphasic production of NO in response to flow. Sudden onset of flow induces a burst of NO production which is both calcium- (Ca2+) and G-protein–dependent. In contrast, steady FSS that follows induces a lower but sustained release of NO and is both Ca2+- and G-protein–independent.5 Moreover, the initial and transient production of NO is directly related to the rate of change in FSS rather than to its absolute magnitude, whereas the sustained release of NO is directly related to the level of FSS.5 Hence, shear-induced NO production appears to be a superposition of two independent mechanochemical pathways.6 However, the initial molecular events enabling eNOS to respond discriminantly to steady and unsteady flow are still not understood.
We hypothesized that the plasma membrane is a site of FSS mechanochemical transduction and that different plasmalemmal molecular mechanisms are involved in the transduction of constant FSS versus TG in FSS. We speculated that rapid changes in FSS will deform the plasma membrane at the interface with rigid domains. Such membrane deformation would lead to membrane protein signaling.7 Interestingly, caveolae, which, in contrast to the major part of the plasma membrane, appear as more rigid liquid-ordered domains, have been described as sites of eNOS activation after FSS.8 The glycocalyx has also been implicated as a mechanosensor structure.9,10 However, the role of glycocalyx components in modifying the shear stress response in ECs is still unclear. The cell–cell junction is also such a rigid domain, characterized by a marked concentration of membrane-associated cytoskeletal elements, as well as numerous intercellular bridges that closely juxtapose the two plasma membranes of neighboring cells (platelet/endothelial cell adhesion molecule-1 junction, adherence junction, tight junction, and gap junction).11 Therefore, we hypothesized that eNOS involved in the mechanotransduction of rapid changes in FSS would be localized at or near the cell–cell junction. In contrast, we expect that steady FSS induces diffuse increases in membrane fluidity on the apical plasma membrane. Hence, eNOS involved in the mechanotransduction of steady FSS would be rather broadly distributed along the plasma membrane.
PECAM-1 is a transmembrane glycoprotein primarily localized at the cell–cell junction of ECs, where it is involved in homophilic interactions with PECAM-1 of neighboring ECs.12 Interestingly, PECAM-1 has been shown to be tyrosine-phosphorylated in response to FSS stimulation.13 Furthermore, it has been suggested that PECAM-1 may function as an inhibitory receptor, interacting with various signaling molecules through its Src homology (SH2) containing protein tyrosine phosphatase (SHP-2) binding domain.14 Therefore, we hypothesized that PECAM-1 might interact with eNOS in primary ECs, hence providing a compartmentalization of eNOS with other signaling proteins, while localizing the enzyme in an area prone to sense any sudden change in membrane tension after a rapid change in FSS
The present study describes a series of experiments which have explored eNOS cell–cell junction localization, eNOS interaction with PECAM-1, the dynamics of this association after different FSS stimulations, and concomitant eNOS activation.
Methods
Cell Culture and FSS
Harvested human umbilical vein endothelial cells (HUVECs) were plated onto slides as described.15 Cells were starved overnight before the experiment in DMEM plus 1% BSA.
Slides were assembled into parallel plate flow chambers.16 Starvation medium was used as perfusion medium. Exposure of the cells to laminar FSS was accomplished using a syringe pump (Harvard Apparatus) driven by a computer, as described.17 At the end of varying periods of FSS, cells were dipped into ice-cold PBS and immediately lysed. Sham control cells were mounted onto the chamber but not subjected to shear.
Animals
All procedures were in accordance with institutional guidelines for animal research. PECAM-1 knockout mice (B6;129S-PECAMtm1Lex) and wild-type mice (B6;129SF2/J) were purchased from the Jackson Laboratory (Bar Harbor, Me). Animals were genotyped at the Jackson Laboratory and any offspring were genotyped in our laboratory.
Immunofluorescence
HUVECs on slides were fixed in 4% PBS-buffered paraformaldehyde (PFA) for 45 minutes at 4°C, treated with 25 mmol/L NH4Cl, and then permeabilized with 0.1% Triton X-100. For double labeling, HUVECs were successively incubated with: (1) 10% normal serum, (2) goat anti–PECAM-1 IgG (Santa Cruz Biotechnology, Santa Cruz, Calif; 1/50) and rabbit anti-eNOS IgG (Santa Cruz Biotechnology, 1/25), and (3) donkey anti-goat IgG (Alexa 488, Molecular Probes, Eugene, Ore) and donkey anti-rabbit (Alexa 568, Molecular Probes). Cells were washed with PBS between incubations. Slides were mounted with "Slow-Fade" reagent (Molecular Probes).
For en face deconvolution microscopy, mice were anesthetized with Ketamine 90 mg/kg/Xylazine 10 mg/kg, IP, and then perfused with ice-cold 4% PFA (15 minutes, 4°C). Aortas were carefully dissected; cross sectional segments were cut and then opened longitudinally. Branches and blood flow direction were recorded. Permeabilization and staining steps were similar to those for HUVECs, except nuclei were stained with DAPI (Molecular Probes). Preparations were mounted in Gelvatol.
Slides were observed under a DeltaVision deconvolution microscope system (Applied Precision Inc.) (For deconvolution methodology, see Falk).18
Please see http://atvb.ahajournals.org for a description of similar studies performed using en face confocal microscopy on rat aorta.
Cellular Lysates
HUVECs were harvested in octyl glucoside (OG) buffer (60 mmol/L OG, 50 mmol/L Tris·HCl, pH 7.4, 125 mmol/L NaCl, 2 mmol/L dithiothreitol, 50 μmol/L EGTA) with protease inhibitors and sonicated. Lysates were centrifuged to remove insoluble material and stored at –80°C.
Protein Assay
Protein in cell lysates was determined by a BCA protein assay (BioRad).
Coimmunoprecipitation
Aliquots of cell lysates (500 μg protein) were precleared for 1.5 hour (4°C) with Protein G-Agarose beads (Santa Cruz Biotechnology). Supernatants were incubated overnight with goat anti–PECAM-1 (Santa Cruz Biotechnology) or rabbit anti-eNOS (Santa Cruz Biotechnology) polyclonal antibodies at a final concentration of 8 and 4 μg/mL, respectively. Antibody titration experiments (not shown) demonstrated that these concentrations led to the quantitative immunoprecipitation of PECAM-1 and eNOS, respectively, from cell lysates. Protein G–Agarose beads were added for an additional 1.5 hour incubation (4°C). Bound immune complexes were washed 3 times with OG buffer and once with 50 mmol/L Tris · HCl (pH 7.4), 150 mmol/L NaCl. Immunoprecipitated proteins were released by boiling in sodium dodecyl sulfate (SDS) sample buffer for 5 minutes.
Immunoblot
Denatured immunoprecipitated proteins were separated on 7.5% SDS-PAGE gels and transferred to polyvinylidene fluoride (PVDF). After blocking with 5% nonfat milk in Tris-buffered saline 0.1% Tween 20 (TBST), membranes were incubated with primary antibodies against eNOS (monoclonal, Transduction Labs) or PECAM-1 (Santa Cruz Biotechnology) for 1 hour in TBST/milk. Bound primary antibodies were detected with horseradish peroxidase–conjugated secondary antibodies: goat–anti-mouse IgG (Pierce, 1/5000) and rabbit anti-goat IgG (Santa Cruz Biotechnology, 1/2500), respectively, followed by chemiluminescent SuperSignal substrate (Pierce). Band intensity was quantified on unsaturated X-ray film by a digital image analyzer (Quantity-One; BioRad). All comparisons were made relative to sham controls.
cGMP Accumulation
NO production in response to FSS was assessed by measuring intracellular accumulation of cGMP. HUVECs were treated with 0.5 mmol/L isobutylmethylxanthine (IBMX) for 30 minutes (to inhibit phosphodiesterases) and 6 mmol/L L-Arginine (substrate of eNOS) for 1 hour including FSS stimulation. Control cells (2 μmol/L ionomycin in DMSO or DMSO only) were used in each experiment. Subsequent steps were performed on ice with 0.5 mmol/L IBMX to prevent cGMP degradation. Cells were scraped in 500 μL 65% EtOH. Resulting extracts were centrifuged to eliminate the insoluble fraction. Supernatants were dried at 60°C in a SpeedVac and stored at –80°C. cGMP levels were measured using a cGMP enzyme immunoassay, following the acetylation procedure as recommended by the supplier (Amersham Pharmacia Biotech). To normalize cGMP values, protein content was measured in each pellet.
eNOS Activity Assay
Please see http://atvb.ahajournals.org for a description of eNOS Activity Assay performed in a cell-free system.
Production of CD31 Cytoplamic Domain GST-Fusion Protein
Please see http://atvb.ahajournals.org for a description of the production of CD31 (PECAM-1) cytoplasmic domain GST-fusion protein, required for the eNOS Activity Assay.
Data Analysis
Data are presented as mean±SEM of the indicated number of observations or as percentage of control. Statistical comparisons between groups were performed using the Student t test with significance level of P<0.05.
Results
eNOS Colocalized With PECAM-1 at the Periphery of ECs In Vitro and In Vivo
To determine whether eNOS and PECAM-1 colocalize, a confocal double-label immunofluorescence study was performed on freshly isolated, confluent, serum-starved HUVECs. A striking eNOS staining was present along the cell periphery in addition to Golgi-like labeling (Figure 1A). This peripheral staining suggests a localization of eNOS at the cell–cell junction. Indeed, immunolocalization of PECAM-1, a major component of the ECs intercellular junction, demonstrates an almost identical peripheral pattern (Figure 1B), and PECAM-1/eNOS double staining demonstrates clear colocalization at the cell periphery (Figure 1C). No fluorescence was detected in negative control slides in which normal serum was substituted for the primary antibody or in cells incubated with secondary antibody alone (data not shown). Similar controls were included within each of the following sections of our immunostaining study with similar results (data not shown).
Figure 1. Major colocalization of PECAM-1 and eNOS in HUVECs in vitro. In serum-starved confluent primary HUVECs, double-immunofluorescence staining revealed peripheral eNOS (A) and PECAM-1 (B) expression. Superposition of eNOS and PECAM-1 staining demonstrated colocalization in yellow-orange (C). Bar=25 μm
As confluence increased, the emergence of a strong eNOS membrane staining was synchronous with the appearance of PECAM-1 staining at the cell–cell junction (data not shown).
To confirm the physiological relevance of the results obtained in our in vitro model, the same staining procedure was performed on rat and mouse aortas. Please see http://atvb.ahajournals.org for results on rat tissue.
eNOS Was Present at the Cell Periphery in PECAM-1 +/+ Mouse Aorta But Was Absent in the Knockout Mouse
Deconvolution microscopy was used to visualize eNOS/PECAM-1 immunolocalization on a straight section of aorta from PECAM-1 knockout and wild-type mice (Figure 2). Antibodies and fixative used were the same as above. In the wild-type mouse aorta, eNOS staining presented the same pattern as the confluent HUVECs: a Golgi labeling and a strong peripheral staining (Figure 2B). PECAM-1 was also detected at the cell periphery (Figure 2A) and a predominant colocalization of the two proteins was observed at the cell–cell junction (Figure 2C). In contrast, in PECAM-1 knockout mouse, the strong eNOS Golgi-like labeling remained but the peripheral staining was absent (Figure 2E and 2F). The absence of PECAM-1 protein in the aorta of PECAM-1 knockout mouse was confirmed (Figure 2D). These results clearly demonstrate that eNOS peripheral staining in mouse aorta endothelium is dependent on the presence of PECAM-1 at the cell–cell junction.
Figure 2. eNOS distribution is altered in PECAM-1 knockout mouse aorta. ECs located in the straight section of the mouse aorta are shown. Staining was performed on wild-type (WT) mouse aorta in A, B, and C, and on PECAM-1 knockout (KO) mouse aorta in D, E, and F. Double staining was performed to localize PECAM-1 in A and D and eNOS in B and E. Superposition of eNOS and PECAM-1 staining demonstrates colocalization in yellow-orange (C and F). Cell nuclei were labeled using DAPI (blue). Bar=15 μm
In Wild-Type Mouse Aorta, Peripheral eNOS Staining Was Absent From the Branch Point of the Intercostal Arteries
In straight sections of the aorta, endothelium is reportedly primarily exposed to unidirectional FSS with a high mean and small TG. In branched arteries, the ECs are believed to experience large TG in FSS with a low mean.3,4 Deconvolution microscopy was used to visualize eNOS/PECAM-1 immunolocalization near an intercostal artery branch in wild-type mouse aorta (Figure 3). Although anti–PECAM-1 stained the cell–cell junction strongly (Figure 3A), eNOS junctional staining was reduced (Figure 3B). However, the strong Golgi-like eNOS labeling was still present. Consequently, eNOS/PECAM-1 colocalization was barely detected at the cell–cell junction (Figure 3C). Such results suggest that chronic TGs in FSS disrupt eNOS/PECAM-1 colocalization.
Figure 3. eNOS distribution is altered in mouse aorta branched areas exposed to high TGs in shear. ECs located near an intercostal artery branch of the aorta are shown. Double staining was performed to localize PECAM-1 in A and eNOS in B. Superposition of eNOS and PECAM-1 staining demonstrates colocalization in yellow-orange (C). Cell nuclei are labeled using DAPI (blue) Bar=15 μm
eNOS Coimmunoprecipitated With PECAM-1 and PECAM-1 Interaction Decreased eNOS Activity In Vitro
Based on the prominent colocalization of eNOS and PECAM-1 within confluent HUVECs, we next investigated whether the two proteins interacted biochemically at the cell–cell junction. For this purpose, detergent-soluble lysates were prepared from confluent serum-starved HUVECs, and both eNOS and PECAM-1 were immunoprecipitated. OG detergent preserves only direct protein–protein interactions. To further enhance the stringency of the immunoprecipitation process, a high ionic strength wash buffer, containing 150 mmol/L NaCl, was used. As seen in Figure 4, immunoprecipitation of eNOS coimmunoprecipitated PECAM-1 from HUVECs extracts (Figure 4A), and conversely, immunoprecipitation of PECAM-1 coimmunoprecipitated eNOS (Figure 4B).
Figure 4. Association of PECAM-1 with eNOS and its effect on eNOS activity in a cell-free system. eNOS or PECAM-1 were immunoprecipitated from HUVECs using polyclonal antibodies. Samples were prepared for Western blot (WB) analysis. A, Under nondenaturing conditions, immunoprecipitation of eNOS coprecipitates PECAM-1 in the presence of 60 mmol/L OG detergent and 150 mmol/L NaCl wash. B, Under identical conditions, immunoprecipitation of PECAM-1 coprecipitates eNOS. C, eNOS activity was measured in the presence of varying concentrations of either GST or GST-PECAM-1 cytoplasmic domain fusion proteins at 37°C for 30 minutes. L-NAME, a competitive inhibitor of eNOS, was also used as control. The experiments were performed in duplicate, and eNOS activity was presented as net radioactivity after subtracting the value for the negative control. The effect of GST-PECAM-1 on eNOS activity was compared with GST at the same concentration using a t test assuming unequal variances. Significance at P<0.05 level, two-tailed, was achieved at 20 μmol/L GST-PECAM-1.
To determine the effects of PECAM-1 interaction with eNOS in vitro, we measured the effect of PECAM-1 on eNOS activity using recombinant proteins. We found that 20 μmol/L GST-PECAM-1 inhibited eNOS activity by 36% compared with 20 μmol/L GST alone (Figure 4C).
FSS Caused eNOS Dissociation From PECAM-1
To determine whether the eNOS/PECAM-1 complex is indeed affected by FSS, highly confluent serum-starved HUVECs were subjected to three different profiles of FSS. TGs in FSS were generated by a sudden onset of FSS (0.5s at 12 dynes/cm2 or 24 dynes/cm2), which we have referred to as an "impulse" in previous studies.6 To apply similar levels of FSS without significant TG, we used a slow ramping in flow over 30s, which reached 24 dynes/cm2 and stayed at this level for 0.5s before ramping down to 0 dynes/cm2 over 30s. Repeated high shear gradients experienced by ECs at branch points were simulated by "pulsatile flow" (repeated impulses; 0.5s every 3 seconds for 30 minutes at 12 dynes/cm2).
A 0.5s impulse of 12 dynes/cm2 resulted in a rapid disruption of the eNOS/PECAM-1 complex (Figure 5A and 5B) within 7s, which was maximal after 15s. After that time, the proteins rapidly reassociated. The reassociation of the complex reached a maximum between 45s and 1 minute. The exact time of the complex reformation varied slightly between these two time points from experiment to experiment, a variation which we attributed to differences in HUVEC batches, but was consistent within duplicate analyses of a single batch. The levels of eNOS immunoprecipitated with PECAM-1 subsequently returned to their prestimulation level. Sham controls (0s in Figure 5A and 5B) were used to normalize immunoprecipitation results. This allowed us to eliminate artifact due to any mechanical stimulation incidentally applied to the cells while mounting the slides on the flow chamber, as well as artifact possibly induced by postshear treatment of the cells. Detergent-soluble lysates were prepared at various times after FSS and PECAM-1 protein was quantitatively immunoprecipitated. As shown in Figure 5C, where eNOS and PECAM-1 binding was assessed 30s after the end of the FSS, the slowly ramping flow (24 dynes/cm2) did not trigger eNOS/PECAM-1 dissociation, although it generated much more shear than the impulse. This clearly demonstrates that eNOS/PECAM-1 dissociation is triggered by a TG of FSS rather than by the FSS itself.
Figure 5. Single FSS impulse and pulsatile FSS stimulations induce a transient dissociation of PECAM-1/eNOS, whereas slowly ramping the FSS does not. Confluent serum-starved HUVECs were subjected to an impulse in FSS of 12 dynes/cm2 for 0.5s. Cell lysates were prepared for PECAM-1 immunoprecipitation and gel electrophoresis. A, A densitometric analysis of five independent experiments reveals the association-dissociation kinetics of PECAM-1/eNOS after the impulse. eNOS densitometry results were normalized for PECAM-1 detected by a second immunostaining performed on the same blot. Results are percentage of sham control. Values are means±SEM (*P<0.05 from sham control). B, A representative Western blot (WB) is shown, demonstrating that eNOS transiently dissociates from PECAM-1 after the impulse in FSS. Dissociation of both molecules was maximal around 15s to 30s. The association was restored 1 minute after FSS. C, A densitometric analysis of four independent experiments reveals the absence of PECAM-1/eNOS dissociation, 30s after the stimulation, when the stimulation was a slow ramp in FSS as opposed to an impulse. D, A densitometric analysis of three independent experiments reveals a partial PECAM-1/eNOS dissociation 30s after pulsatile flow (12 dynes/cm2, 0.5s impulse, 3s intervals for 30 minutes). In C and D, values are means±SEM (*P<0.05 and **P>0.05 from sham control).
Cells subjected to pulsatile flow for 30 minutes showed a partial dissociation of the eNOS/PECAM-1 complex (Figure 5D). This is consistent with the in vivo observation of a reduced junctional eNOS staining near intercostal artery branches, where ECs are exposed to repeated high TGs in FSS. Results observed after a single impulse in FSS were similar to those obtained after a pulsatile FSS. This suggests that data obtained from single impulse studies are good indicators of the in vivo effects of high TGs in FSS. However, the response to the pulsatile FSS was less marked than the response to the single impulse. This could indicate some degree of acclimatization of the cells to a repeated impulse stimulus.
A 2-Fold Augmentation of cGMP Accumulation Was Observed After Impulse
The role of eNOS/PECAM-1 dissociation on eNOS activity was indirectly measured by assessing the cGMP amount accumulated within HUVECs at various times after FSS. A nearly 2-fold augmentation of cGMP accumulation was observed 1 minute after an impulse in FSS compared with sham controls (Figure 6A). No further FSS-induced cGMP production was detected after this time, indicating that cGMP synthesis is very rapidly triggered and downregulated within the first minute after FSS. This time frame is identical to the one of the eNOS/PECAM-1 complex dissociation–reassociation. No augmentation of cGMP production was observed during the first five minutes after slow ramping in flow (24 dynes/cm2), but large quantities of cGMP were produced when cells were exposed to an impulse of the same intensity (Figure 6B).
Figure 6. An impulse in FSS induces a concomitant augmentation of eNOS activity. Slowly ramping the FSS does not trigger such an increase in eNOS activity. NO production from HUVECs in response to FSS was assessed by the intracellular accumulation of cGMP. Confluent serum-starved HUVECs were incubated with 0.5 mmol/L IBMX for 30 minutes and 6 mmol/L L-Arginine for 1 hour. Cells were then subjected to an impulse in FSS of 12 dynes/cm2 for 0.5s in A, and to an impulse in FSS of 24dynes/cm2 for 0.5s or a slow ramping toward a FSS of 24dynes/cm2 in B. At different times after impulse or ramping (0, 30, 60, 120, or 300s), HUVECs were lysed and cellular extracts were generated. As shown in C, in each of these experiments, control cells were stimulated for 15 minutes with 2 μmol/L ionomycin or with the vehicle alone (DMSO). cGMP levels in the extracts were normalized for cellular proteins content. cGMP production values are percentages of sham control. A and C display means (±SEM) from 3 independent experiments (*P<0.05, **P>0.05). B shows one representative experiment of two.
Discussion
This study is the first to demonstrate, by means of coimmunoprecipitation and in situ immunolocalization, that eNOS and PECAM-1 are closely associated in human primary ECs in vitro and mouse and rat aortal tissue in vivo. This suggests that PECAM-1 is responsible for the localization of eNOS at the cell–cell junction. Additional compelling experimental evidence to support this conclusion is that eNOS peripheral staining was absent in the endothelium of PECAM-1 knockout mice in situ, whereas it was present in the wild-type controls. In addition, results showed that eNOS/PECAM-1 association is rapidly and transiently disrupted after a sudden change in FSS but not after steady FSS in ECs. These in vitro data were corroborated by in situ immunostainings demonstrating the absence of eNOS/PECAM-1 colocalization in wild-type mice endothelium in regions exposed to rapidly changing levels of FSS (high TGs). Moreover, the dissociation of the eNOS/PECAM-1 complex was paralleled by an augmentation in eNOS activity in freshly harvested ECs. Finally, in a cell-free system we were able to show a direct inhibition of NOS activity by PECAM-1. The fact that PECAM-1, in the absence of cells, is able to directly affect eNOS activity, suggests that these proteins not only colocalize but also interact and influence protein function. This is further evidence of the effect of PECAM-1 on eNOS protein localization and activity. Therefore, we provide the first evidence to suggest the involvement of an eNOS/PECAM-1 interaction in FSS-induced mechanotransduction.
The eNOS/PECAM-1 colocalization we observed in human primary endothelial cells mimics the staining pattern previously reported in immortalized murine microvascular ECs and native rat cardiac endothelium.19,20 Its physiological relevance is further strengthened here by the demonstration of eNOS/PECAM-1 colocalization in rat and mouse aorta. The conservation of this colocalization through three species suggests that it has an important function. The correlation between cell confluence, PECAM-1 appearance, and eNOS presence at the cell–cell junction confirms published observations using an immortalized murine EC line.19 In addition, the results obtained with PECAM-1 knockout mice are consistent with the notion that PECAM-1 presence at the cell–cell junction is necessary for eNOS peripheral localization. The colocalization of eNOS and PECAM-1 immunostaining is corroborated by the coimmunoprecipitation of these two molecules. This procedure was performed under highly stringent conditions which preserve only direct protein–protein bindings, eliminating any weak interactions. Together with the demonstration that PECAM-1 directly inhibits eNOS activity in the purified protein preparation (Figure 4), this strongly suggests that the two proteins form a direct and tight association.
The close association of eNOS with PECAM-1 is important in light of recent developments in PECAM-1 biology. PECAM-1 has recently been suggested to function as a mechanotransducing protein13 and has been shown to be tyrosine-phosphorylated within 1 minute in ECs exposed to a "step change" in FSS.13 Furthermore, the activation of extracellular signal regulated kinase (ERK), one of the early EC responses to FSS, has been shown to be dependent on PECAM-1 phosphorylation. The recruitment and binding of SHP-2 to phospho-PECAM-1 at the cell periphery eventually triggers activation of ERK.13 Interestingly, we have recently shown that ERK activation is more specifically sensitive to TG in FSS.15 The tyrosine phosphorylation of PECAM-1 occurs in a time frame similar to that of the eNOS/PECAM-1 dissociation observed in our study. A recent article by Jin et al suggests that flow-stimulated tyrosine kinases recruit phosphoinositide 3-kinase and mediate activation of Akt and eNOS.21 Whether, and to what extent, these events are intertwined is presently unknown.
The fact that the eNOS/PECAM-1 was not affected by a slow ramping FSS, even at high shear levels, strongly suggests that this mechanoreceptive system is specific to rapid changes in FSS. The localization of the complex at the cell–cell junction is consistent with this discriminatory ability because it is located in a region of the membrane presumably sheltered from the FSS itself. In fact, one would expect mechanotransductory molecules sensitive to the FSS (and not the TG of shear stress) to be concentrated at the apex of the apical surface of the cell, where the FSS is maximal.22
During recent years, several groups have demonstrated that activation of eNOS, both in conjunction with and independent of intracellular calcium flux, occurs through the allosteric binding of eNOS with neighboring regulatory proteins. In this regard, several eNOS-associated proteins have been identified, including calmodulin, Caveolin-1, Hsp 90, dynamin, NOSIP, and some G protein–coupled receptors (for detailed review see Fleming et al).23 In addition, most of the kinases known to phosphorylate eNOS, notably on serine 1177, are believed to be concentrated near eNOS through a scaffold protein. Our data suggest that PECAM-1 is also involved in an eNOS signaling complex at the cell–cell junction. Currently, we are investigating which other known eNOS-associated proteins might be present in the eNOS/PECAM-1 complex.
In conclusion, we have shown that eNOS is closely associated with PECAM-1 at the cell–cell junction and that this complex is dissociated by TG in FSS but not by FSS itself. Current studies focus on the role of intracellular calcium and G proteins on eNOS/PECAM-1 association as well as on eventual interactions of eNOS/PECAM-1 with previously described eNOS-associated proteins.
Acknowledgments
The authors thank the University of California, San Diego Cancer Center Digital Imaging Shared Resource staff for their assistance with the deconvolution studies. This work was supported by National Institutes of Health grant HL40696.
References
Ignarro LJ, Cirino G, Casini A, Napoli C. Nitric oxide as a signaling molecule in the vascular system: an overview. J Cardiovasc Pharmacol. 1999; 34: 879–886.
Davies PF. Flow-mediated endothelial mechanotransduction. Physiol Rev. 1995; 75: 519–560.
Zarins CK, Giddens DP, Bharadvaj BK, Sottiurai VS, Mabon RF, Glagov S. Carotid bifurcation atherosclerosis. Quantitative correlation of plaque localization with flow velocity profiles and wall shear stress. Circ Res. 1983; 53: 502–514.
Ku DN, Giddens DP, Zarins CK, Glagov S. Pulsatile flow and atherosclerosis in the human carotid bifurcation. Positive correlation between plaque location and low oscillating shear stress. Arteriosclerosis. 1985; 5: 293–302.
Kuchan MJ, Frangos JA. Role of calcium and calmodulin in flow-induced nitric oxide production in endothelial cells. Am J Physiol. 1994; 266: C628–C636.
Frangos JA, Huang TY, Clark CB. Steady shear and step changes in shear stimulate endothelium via independent mechanisms-superposition of transient and sustained nitric oxide production. Biochem Biophys Res Commun. 1996; 224: 660–665.
Gudi S, Nolan JP, Frangos JA. Modulation of GTPase activity of G proteins by fluid shear stress and phospholipid composition. Proc Natl Acad Sci U S A. 1998; 95: 2515–2519.
Rizzo V, McIntosh DP, Oh P, Schnitzer JE. In situ flow activates endothelial nitric oxide synthase in luminal caveolae of endothelium with rapid caveolin dissociation and calmodulin association. J Biol Chem. 1998; 273: 34724–34729.
Pohl U, Herlan K, Huang A, Bassenge E. EDRF-mediated shear-induced dilation opposes myogenic vasoconstriction in small rabbit arteries. Am J Physiol. 1991; 261: H2016–H2023.
Hecker M, Mulsch A, Bassenge E, Busse R. Vasoconstriction and increased flow: two principal mechanisms of shear stress-dependent endothelial autocoid release. Am J Physiol. 1993; 265: H828–H833.
Klymkowsky MW, Parr B. The body language of cells: the intimate connection between cell adhesion and behavior. Cell. 1995; 83: 5–8.
Newton JP, Buckley CD, Jones EY, Simmons DL. Residues on both faces of the first immunoglobulin fold contribute to homophilic binding sites of PECAM-1/CD31. J Biol Chem. 1997; 272: 20555–20563.
Osawa M, Masuda M, Kusano K, Fujiwara K. Evidence for a role of platelet endothelial cell adhesion molecule-1 in endothelial cell mechanosignal transduction: is it a mechanoresponsive molecule? J Cell Biol. 2002; 158: 773–785.
Newman PJ, Newman DK. Signal transduction pathways mediated by PECAM-1: new roles for an old molecule in platelet and vascular cell biology. Arterioscler Thromb Vasc Biol. 2003; 23: 953–964.
Bao X, Lu C, Frangos JA. Mechanism of temporal gradients in shear-induced ERK1/2 activation and proliferation in endothelial cells. Am J Physiol. 2001; 281: H22–H29.
Kuchan MJ, Jo H, Frangos JA. Role of G proteins in shear stress-mediated nitric oxide production by endothelial cells. Am J Physiol. 1994; 267: C753–C758.
Gudi SR, Clark CB, Frangos JA. Fluid flow rapidly activates G proteins in human endothelial cells. Involvement of G proteins in mechanochemical signal transduction. Circ Res. 1996; 79: 834–839.
Falk MM. Connexin-specific distribution within gap junctions revealed in living cells. J Cell Sci. 2000; 113: 4109–4120.
Govers R, Bevers L, de Bree P, Rabelink TJ. Endothelial nitric oxide synthase activity is linked to its presence at cell–cell contacts. Biochem J. 2002; 361: 193–201.
Andries LJ, Brutsaert DL, Sys SU. Nonuniformity of endothelial constitutive nitric oxide synthase distribution in cardiac endothelium. Circ Res. 1998; 82: 195–203.
Jin Z, Ueba H, Tanimoto T, Lungu AO, Frame MD, Berk BC. Ligand-independent activation of vascular endothelial growth factor receptor 2 by fluid shear stress regulates activation of endothelial nitric oxide synthase. Circ Res. 2003; 93: 354–363.
Barbee KA, Mundel T, Lal R, Davies PF. Subcellular distribution of shear stress at the surface of flow-aligned and nonaligned endothelial monolayers. Am J Physiol. 1995; 268: H1765–H1772.
Fleming I, Busse R. Molecular mechanisms involved in the regulation of the endothelial nitric oxide synthase. Am J Physiol. 2003; 284: R1–R12.(Implication for Flow-Indu)