Shear Stress Stimulates Phosphorylation of Endothelial Nitric-oxide Synthase at Ser1179 by Akt-independent Mechanisms

Recently, we have shown that shear stress stimulates NO production by the protein kinase B/Akt (Akt)-dependent mechanisms in bovine aortic endothelial cells (BAEC) (Go, Y. M., Boo, Y. C., Park, H., Maland, M. C., Patel, R., Pritchard, K. A., Jr., Fujio, Y., Walsh, K., Darley-Usmar, V., and Jo, H. (2001) J. Appl. Physiol. 91, 1574– 1581). Akt has been believed to regulate shear-dependent production of NO by directly phosphorylating endothelial nitric-oxide synthase (eNOS) at the Ser residue (eNOS-S), but a critical evaluation using specific inhibitors or dominant negative mutants (Akt or Akt) has not been reported. In addition, other kinases, including protein kinase A (PKA) and AMP kinase have also shown to phosphorylate eNOS-S. Here, we show that shear-dependent phosphorylation of eNOSS is mediated by an Akt-independent, but a PKA-dependent, mechanism. Expression of Akt or Akt in BAEC by using recombinant adenoviral constructs inhibited phosphorylation of eNOS-S if cells were stimulated by vascular endothelial growth factor (VEGF), but not by shear stress. As shown before, expression of Akt inhibited shear-dependent NO production, suggesting that Akt is still an important regulator in NO production. Further studies showed that a selective inhibitor of PKA, H89, inhibited shear-dependent phosphorylation of eNOS-S and NO production. In contrast, H89 did not inhibit phosphorylation of eNOS-S induced by expressing a constitutively active Akt mutant (Akt) in BAEC, showing that the inhibitor did not affect the Akt pathway. 8-Bromo-cAMP alone phosphorylated eNOS-S within 5 min without activating Akt, in an H89-sensitive manner. Collectively, these results demonstrate that shear stimulates phosphorylation of eNOS-S in a PKA-dependent, but Aktindependent manner, whereas the NO production is regulated by the mechanisms dependent on both PKA and Akt. A coordinated interaction between Akt and PKA may be an important mechanism by which eNOS activity is regulated in response to physiological stimuli such as shear stress. Endothelial cells are constantly subjected to shear stress, the dragging force generated by blood flow. Shear stress triggers a variety of biochemical and physical changes in cell structure and function. For example, shear stress regulates vascular tone and diameter, inflammatory responses, hemostasis, and vessel wall remodeling (1). Laminar shear stress has been shown to play anti-atherogenic roles by inhibiting some of the key pro-atherogenic events, including apoptosis of endothelial cells and binding of monocytes to endothelium (2–5) Although the exact mechanisms by which laminar shear stress prevents atherosclerosis are not known, NO produced from endothelium seems to play essential roles by mediating many effects of laminar shear stress: vessel relaxation, inhibition of apoptosis, and monocyte adhesion triggered by the pro-atherogenic factors (2, 4–6). Although it is well known that exposure of endothelial cells to shear stress stimulates production of NO from endothelial nitric-oxide synthase (eNOS) both in cultured cells and in intact vessels (7, 8), the molecular mechanisms by which shear stress regulates NO production have not been clearly elucidated. eNOS is known as a Ca calmodulin (CaM)-dependent form of NOS (9). Indeed, most humoral ligands, including bradykinin, acetylcholine, and ATP, stimulate NO production from eNOS by raising the level of intracellular Ca , which forms Ca CaM complex (9). In the basal state, the majority of eNOS appears to be bound to caveolin-1 with its enzyme activity repressed in caveolae (10, 11). This tonic inhibition of eNOS can be released by displacing caveolin-1 with Ca CaM in response to Ca -mobilizing agonists (10). Unlike Ca -mobilizing hormones, however, shear stress stimulates production of NO from eNOS by a mechanism that does not require a maintained intracellular Ca level or CaM (8, 12, 13). Other potential mechanisms that could mediate the acute, shear-dependent activation of eNOS include phosphorylation, acylation, and translocation of the enzyme as well as its interaction with other molecules such as heat shock protein 90 (14–19). Recent evidence suggested that the phosphorylation of eNOS at Ser (based on the bovine eNOS sequence and equivalent

Endothelial cells are constantly subjected to shear stress, the dragging force generated by blood flow. Shear stress triggers a variety of biochemical and physical changes in cell structure and function. For example, shear stress regulates vascular tone and diameter, inflammatory responses, hemostasis, and vessel wall remodeling (1). Laminar shear stress has been shown to play anti-atherogenic roles by inhibiting some of the key pro-atherogenic events, including apoptosis of endothelial cells and binding of monocytes to endothelium (2)(3)(4)(5) Although the exact mechanisms by which laminar shear stress prevents atherosclerosis are not known, NO ⅐ produced from endothelium seems to play essential roles by mediating many effects of laminar shear stress: vessel relaxation, inhibition of apoptosis, and monocyte adhesion triggered by the pro-atherogenic factors (2, 4 -6).
Although it is well known that exposure of endothelial cells to shear stress stimulates production of NO ⅐ from endothelial nitric-oxide synthase (eNOS) 1 both in cultured cells and in intact vessels (7,8), the molecular mechanisms by which shear stress regulates NO ⅐ production have not been clearly elucidated. eNOS is known as a Ca 2ϩ ⅐calmodulin (CaM)-dependent form of NOS (9). Indeed, most humoral ligands, including bradykinin, acetylcholine, and ATP, stimulate NO ⅐ production from eNOS by raising the level of intracellular Ca 2ϩ , which forms Ca 2ϩ ⅐CaM complex (9). In the basal state, the majority of eNOS appears to be bound to caveolin-1 with its enzyme activity repressed in caveolae (10,11). This tonic inhibition of eNOS can be released by displacing caveolin-1 with Ca 2ϩ ⅐ CaM in response to Ca 2ϩ -mobilizing agonists (10). Unlike Ca 2ϩ -mobilizing hormones, however, shear stress stimulates production of NO ⅐ from eNOS by a mechanism that does not require a maintained intracellular Ca 2ϩ level or CaM (8,12,13). Other potential mechanisms that could mediate the acute, shear-dependent activation of eNOS include phosphorylation, acylation, and translocation of the enzyme as well as its interaction with other molecules such as heat shock protein 90 (14 -19).
Recent evidence suggested that the phosphorylation of eNOS at Ser 1179 (based on the bovine eNOS sequence and equivalent to human eNOS-S 1177 ) by a sequential activation of phosphoinositide-3-kinase (PI3K) and protein kinase B/Akt (Akt) pathway is the underlying mechanism by which shear stress stimulates NO ⅐ production in a Ca 2ϩ ⅐ CaM-insensitive manner (14,20). In the cases of vascular endothelial growth factor (VEGF), sphingosine 1-phosphate, and estrogen, there is strong evidence supporting PI3K activation of Akt, which in turn is responsible for regulating the phosphorylation and activation of eNOS (17,(21)(22)(23). However, published reports (14,19,20) have not provided direct evidence that Akt is indeed the protein kinase directly responsible for phosphorylation of eNOS-S 1179 and its subsequent activation in response to shear stress. For example, it has not been reported whether expression of a dominant negative Akt constructs such as Akt AA or Akt AAA can block the shear-dependent phosphorylation of eNOS-S 1179 and NO ⅐ production in endothelial cells (19,20).
In addition, it is becoming increasingly clear that Akt is not the only protein kinase that can phosphorylate eNOS-S 1179 . Other protein kinases, including protein kinase A (PKA), protein kinase G (PKG), and AMP kinase have also been shown to phosphorylate eNOS-S 1179 (24 -27). It may be possible that eNOS-S 1179 can be phosphorylated by different protein kinases depending upon each given stimuli.
Here, we examined whether Akt regulates phosphorylation of eNOS-S 1179 in response to shear stress in bovine aortic endothelial cells (BAEC) by using adenoviral constructs expressing a constitutively active form (Akt Myr ) or dominant negative forms of Akt mutants (Akt AA and Akt AAA ). Because our initial results indicated that shear stimulates phosphorylation of eNOS-S 1179 in an Akt-independent manner, we examined other protein kinases to characterize their role in the phosphorylation. Our results demonstrate that, unlike VEGF, shear stress stimulates phosphorylation of eNOS-S 1179 by an Aktindependent, but PKA-dependent, manner.

EXPERIMENTAL PROCEDURES
Cell Culture-BAEC harvested from descending thoracic aortas were maintained (37°C, 5% CO 2 ) in a growth medium (Dulbecco's minimum Eagle's medium (DMEM) containing 1 g/liter glucose (Invitrogen) and 20% fetal bovine serum (Atlanta Biologicals) without antibiotics) (28). BAEC used in this study were between passages 5 and 10. Unless specified otherwise, 2 million cells were seeded in 100-mm tissue culture dishes (Falcon) and grown to confluency in the growth medium before exposure to shear stress.
Adenoviral Infections-BAEC were infected with recombinant adenovirus at ϳ90% confluency. Cells were infected with adenovirus in serum-free DMEM for 1 h and then incubated for 48 h in a growth medium before the treatment. Three different recombinant adenoviral constructs (Ad-Akt Myr , Ad-Akt AA , and Ad-Akt AAA ) were used to overexpress Akt mutants. Akt Myr is a constitutively active Akt mutant generated by fusing a myristoylation signal to its amino terminus (29). Akt AA is an Akt mutant generated by substituting Ala at two major regulatory phosphorylation sites (Thr 308 and Ser 473 ) (29). Akt AAA is another Akt mutant in which the phosphate transfer residue in catalytic site (Lys 179 ), in addition to Thr 308 and Ser 473 , was replaced by Ala (22). Both Akt AA and Akt AAA have been shown to inhibit Akt specifically in dominant negative manners (22,29). Recombinant adenovirus encoding ␤-galactosidase (Ad-␤-gal) was used as a control. Infection efficiency of BAEC with recombinant adenovirus at 50 multiplicity of infection (m.o.i.) was close to 100% as determined by immunohistochemical staining of ␤-galactosidase as described previously (28).
Shear Stress Studies-A confluent BAEC monolayer grown in a 100-mm dish was exposed to non-pulsatile, laminar shear stress in a shear medium (phenol red-free DMEM containing 0.5% fetal bovine serum and 25 mM Hepes, pH 7.4) by rotating a Teflon cone (0.5°cone angle) as we described previously (30,31).
Immunoblotting-Aliquots of cell lysates (20 g of protein each) were resolved on a 10% SDS-PAGE gel and transferred to a polyvinylidene difluoride membrane (Millipore) (28). The membrane was incubated with a primary antibody overnight at 4°C, and then with a secondary antibody conjugated with alkaline phosphatases (1 h at room temperature), which were detected by a chemiluminescence method (28). The intensities of immunoreactive bands in Western blots were analyzed by using the National Institutes of Health IMAGE program. The following primary antibodies were used: polyclonal antibodies for phosphorylated forms of Akt-Thr 308 (pT-Akt), Akt-Ser 473 (pS-Akt), eNOS-S 1179 (pS-eNOS), and ERK1/2 (pERK1/2) and total ERK1/2 from Cell Signaling Technology; a polyclonal antibody for total Akt from Santa Cruz Biotechnologies; polyclonal and monoclonal antibodies for total eNOS from Transduction Laboratories; and a monoclonal antibody against a hemagglutinin (HA) epitope from Roche Molecular Biochemicals.
NO ⅐ Assay-A confluent BAEC monolayer grown in a 100-mm tissue culture dish was exposed to laminar shear stress by rotating Krebs-Ringer carbonate buffer (25 mM NaHCO 3 , pH 7.4, 118 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl 2 , 1.2 mM MgSO 4 , 1.2 mM KH 2 PO 4 , and 11 mM glucose) with a cone in a 5% CO 2 incubator at 37°C. To measure accumulation of NO ⅐ in the medium, a 1-ml sample was collected, replaced with fresh medium, and kept dark on ice until nitrite assay. After shear exposure, cells were washed with ice-cold phosphate-buffered saline and scraped in lysis buffer to measure the amount of protein and subsequent Western blot analysis. A fluorescence assay using 2,3-diaminonaphthalene was used to measure nitrite, because it accounts for more than 90% of total NO ⅐ metabolite accumulating in the medium in response to shear stress (31)(32)(33).
Statistical Analysis-Statistical analysis was performed by the Student's t test. The significance, p Ͻ 0.05, based on at least three or more independent experiments, was considered to be statistically significant.

Shear Stress Stimulates Phosphorylation of eNOS-S 1179 and
Akt-Thr 308 /Ser 473 -We first confirmed that shear stress stimulates phosphorylation of Akt and eNOS. As shown previously (14), exposure of BAEC to an arterial level of laminar shear stress (15 dyn/cm 2 ) stimulated the phosphorylation of eNOS-S 1179 in a time-dependent manner (Fig. 1A). The phosphorylation of eNOS-S 1179 was apparent as early as 2 min after shear onset and reached a maximum by 30 min. During this time period the amount of eNOS did not change as determined by Western blot using an antibody detecting total eNOS. Shear stress also stimulated phosphorylation of Akt at the two key regulatory sites (Thr 308 and Ser 473 ) as determined by Western blots using antibodies specific for each phosphorylated site (Fig. 1B). The time courses of shear-dependent phosphorylations of Akt at both Thr 308 and Ser 473 residues were essentially identical. In addition, the time course of Akt phosphorylation was very similar to that of eNOS phosphorylation.
PI3K Regulates Phosphorylation of eNOS-S 1179 and Akt-Thr 308 /Ser 473 in Response to Shear Stress-To examine whether the shear-dependent phosphorylation of eNOS and Akt are regulated by PI3K-dependent mechanisms, BAEC were pretreated for 30 min with the PI3K inhibitor wortmannin, and then exposed to shear stress for up to 30 min. Treatment of the cells with wortmannin completely blocked sheardependent phosphorylation of eNOS-S 1179 ( Fig. 2A). As shown previously (14,31), wortmannin also blocked shear-dependent phosphorylation of Akt-Thr 308 and Ser 473 (Fig. 2B). In contrast, wortmannin did not block shear-dependent phosphorylation of extracellular signal-regulated kinase (ERK) 1/2 (Fig. 2C). It was noteworthy that wortmannin consistently abolished even the basal phosphorylation of Akt (Fig. 2B), whereas it did not inhibit the basal phosphorylation of eNOS-S 1179 ( Fig. 2A). The results shown in Figs. 1 and 2 clearly demonstrate that shear stress stimulates phosphorylation of eNOS-S 1179 and Akt-Thr 308 /Ser 473 by a PI3K-dependent mechanism, confirming previous reports (14,31).

Expression of a Constitutively Active Akt Mutant (Akt Myr )
Induces Phosphorylation of eNOS-S 1179 -To address the issue whether Akt regulates phosphorylation of eNOS-S 1179 in response to shear stress, we used three different recombinant adenoviral constructs to transiently express a constitutively active Akt mutant (Akt Myr tagged with HA) or two dominant negative constructs (Akt AAA and Akt AA tagged with HA) in BAEC. First, BAEC were infected with Ad-Akt Myr (10 or 50 m.o.i.) or Ad-␤-gal (50 m.o.i.) as a control. The dose-dependent expression of Akt Myr was confirmed by Western blot with antibodies for HA and total Akt. Expression of Akt Myr also activated Akt as determined by Western blots with antibodies specific to phosphorylated forms of Akt-Ser 473 and Akt-Thr 308 (Fig. 3). As shown in Fig. 3, expression of Akt Myr alone was sufficient to stimulate phosphorylation of eNOS-S 1179 . In con-trast, expression of ␤-galactosidase did not have any effect on phosphorylation of eNOS-S 1179 , showing the specific effect of Akt Myr as the kinase.

Dominant Negative Akt Mutants Inhibit Phosphorylation of eNOS-S 1179 When Stimulated by VEGF but Not by Shear
Stress-Here we directly examined whether Akt regulates sheardependent NO ⅐ production by phosphorylating eNOS-S 1179 by infecting BAEC with dominant negative Akt mutants. First, we checked whether expression of Akt AA inhibits shear-dependent NO ⅐ production in BAEC. As shown previously (31), infection of BAEC with Ad-Akt AA (50 m.o.i./cell) prevented production of NO ⅐ stimulated by shear stress demonstrating the essential role of Akt in eNOS activation (Fig. 4). In contrast, when cells were infected with Ad-␤-gal as a control, shear stress still stimulated NO ⅐ production robustly (Fig. 4). This result dem-

FIG. 1. Shear stress stimulates phosphorylation of eNOS and Akt.
Confluent monolayers of BAEC were exposed to laminar shear stress (15 dyn/ cm 2 ) with a cone-and-plate viscometer for the time periods indicated. Cell lysates (20 g/lane) were analyzed by Western blot with antibodies specific for phosphorylated forms of eNOS-S 1179 (pS-eNOS), Akt-Thr 308 (pT-Akt), and Akt-Ser 473 (pS-Akt). The membranes were re-probed with antibodies detecting total eNOS and Akt to monitor equal loading of samples. Densitometry was performed to quantify phosphorylated bands, and the graphs show mean -fold simulation values Ϯ S.E. (n ϭ 3).

FIG. 2. A PI3K inhibitor blocks phosphorylation of eNOS and
Akt, but not ERK, in response to shear stress. BAEC were pretreated with vehicle (Me 2 SO) or a PI3K inhibitor (100 nM wortmannin) for 30 min before exposure to shear stress (15 dyn/cm 2 ) for the time periods indicated. Phosphorylation of eNOS, Akt, and ERK was determined by Western blot as described in the legend of Fig. 1. As controls for equal loading, the same membranes were re-probed with antibodies that detect the total amount of each protein. The phosphorylated bands were quantified as in Fig. 1, and the graphs show means Ϯ S.E. (n ϭ 3).
onstrates the dominant negative effect of Akt AA on NO ⅐ production in response to shear in a specific manner.
Next, we examined whether overexpression of dominant negative Akt mutants would inhibit phosphorylation of eNOS-S 1179 stimulated by shear stress. In this study, BAEC were infected with Ad-Akt AA or Ad-Akt AAA and cells were subjected to shear stress (15 dyn/cm 2 for 60 min) 1 day after the infection. The whole cell lysates were then analyzed by Western blots to determine the phosphorylation of eNOS-S 1179 . Because VEGF stimulates phosphorylation of eNOS-S 1179 by Akt-dependent mechanisms (17), the growth factor was used as a positive control. To our surprise, overexpression of either Akt AA or Akt AAA did not have any significant effect on the phosphorylation of eNOS-S 1179 stimulated by shear stress (Fig. 5, A and  C). On the other hand, overexpression of either Akt AA or Akt AAA substantially inhibited phosphorylation of eNOS-S 1179 induced by VEGF as expected (Fig. 5, B and D). Overexpression of Akt mutants induced by infection with Ad-Akt AA and Ad-Akt AAA were demonstrated by Western blots using antibodies specific either to total Akt (Fig. 5, A and B) or HA tag (Fig. 5, C and D). As controls, cells were also infected with Ad-␤-gal, and it showed no obvious effect on eNOS-S 1179 phosphorylation induced by shear stress (Fig. 5, C and D). These results clearly demonstrate that the dominant negative Akt mutants prevent eNOS-S 1179 phosphorylation if it is stimulated by VEGF but not by shear stress. It is interesting to find that Akt AA blocks shear-dependent NO ⅐ production without inhibiting eNOS-S 1179 phosphorylation. Taken together, these results show that shear stress stimulates phosphorylation of eNOS-S 1179 in an Akt-independent manner.
Role of PKA and ERK Pathways in Shear-dependent Phosphorylation of eNOS-S 1179 -The above results showing that the shear-dependent phosphorylation of eNOS-S 1179 is not mediated in an Akt-dependent manner prompted us to consider alternative mechanisms. In addition to Akt, several other protein kinases, including PKA, PKG, AMP kinase, and ERK pathways have been implicated in eNOS regulation (24 -27, 34). We chose to use selective inhibitors of PKA and ERK pathways to examine whether they regulate shear-dependent phosphorylation of eNOS-S 1179 . First, BAEC were pretreated with 0 -20 M H89, a highly selective PKA inhibitor for 30 min, and then sheared for 30 min. As shown in Fig. 6A, H89 inhibited shear stress-dependent phosphorylation eNOS-S 1179 in a dose-dependent manner with a maximum inhibitory effect observed at 10 M.
This result could be derived from two different possible mechanisms. One possibility is that PKA is an upstream regulator of Akt, which is then responsible for phosphorylating eNOS-S 1179 directly. The other is that PKA, without involving Akt, is responsible for phosphorylation of eNOS-S 1179 directly or indirectly. If the former were true, H89 would be expected to inhibit Akt phosphorylation. If the latter were true, then H89 would not prevent Akt phosphorylation. As shown in Fig. 6B, the PKA inhibitor did not inhibit phosphorylation of Akt at the two major regulatory sites (Thr 308 and Ser 473 ). This result favors the latter mechanism that PKA is responsible for phos- BAEC were infected with Ad-Akt Myr at the indicated level. As a control, Ad-␤-gal was used. Two days after the infection, cell lysates were prepared and analyzed by Western blot using antibodies specific for phosphorylated forms of eNOS-S 1179 (pS-eNOS), Akt-Thr 308 (pT-Akt), and Akt-Ser 473 (pS-Akt). Antibodies detecting the total amount of each protein were used to re-probe the same membranes. Because Akt Myr was tagged with an HA epitope, an HA antibody was used to show expression of the transfected protein. phorylation of eNOS-S 1179 without involving an Akt pathway.
Interestingly, H89 augmented the phosphorylation of Akt both in basal and sheared cells (Fig. 6B). This result showing that, even when Akt was intensely stimulated by H89, eNOS phosphorylation was not increased (rather, it was completely inhibited) provides further evidence against the role of Akt as the protein kinase directly phosphorylating eNOS-S 1179 .
Next, we tested whether ERK pathway regulates the shear stress-dependent phosphorylation of eNOS-S 1179 . BAEC were pretreated with 0 -40 M of PD98059, a selective inhibitor of MEK1/2 (the upstream protein kinase regulating ERK1/2), and then sheared at 15 dyn/cm 2 for 20 min. Phosphorylations of eNOS-S 1179 and Akt-Ser 473 stimulated by shear stress were not affected by this inhibitor (Fig. 7, A and B), whereas ERK phosphorylation was inhibited in a concentration-dependent manner (Fig. 7C). These results demonstrate that ERK pathway does not play a significant role in shear-dependent phosphorylation of Akt-Ser 473 and eNOS-S 1179 .
Treatment of BAEC with cAMP Induces Phosphorylation of eNOS-S 1179 -To further examine the role of PKA in phosphorylation of eNOS-S 1179 , we stimulated the protein kinase by incubating BAEC with a direct activator, a cell-permeable cAMP analog (8-Br-cAMP). As shown in Fig. 8A, treatment of cells with 8-Br-cAMP alone maximally stimulated phosphorylation of eNOS-S 1179 within 2-5 min, which was sustained for as long as 60 min. In contrast, during the same early incubation period (up to 5 min) Akt phosphorylation was not at all stimulated by 8-Br-cAMP (Fig. 8A). Only after 15-min incubation with 8-Br-cAMP, a modest increase in Akt phosphorylation was observed (Fig. 8A). These results demonstrate that eNOS-S 1179 can be rapidly phosphorylated in the absence of Akt phosphorylation in response to cAMP. These also suggest that eNOS-S 1179 phosphorylation can be stimulated by an Akt-independent, but by a PKA-dependent, mechanism. In additional dose curve studies, it was also found that a concentration of 8-Br-cAMP as low as 250 M significantly stimulated phosphorylation of eNOS-S 1179 (data not shown).
Next, to determine the specificity of PKA pathway, cells were pretreated with or without H89 (5-20 M) for 30 min before incubating with 8-Br-cAMP. H89, as low as 5 M, significantly inhibited phosphorylation of eNOS-S 1179 induced by cAMP (Fig. 8B). On the other hand, the PKA inhibitor did not inhibit Akt phosphorylation induced by 8-Br-cAMP (Fig. 8B). In fact, as shown above in Fig. 6B, H89 augmented phosphorylation of Akt induced by cAMP. These results demonstrate that cAMP stimulated eNOS-S 1179 phosphorylation by the PKA-dependent mechanisms in an Akt-independent manner.

H89 Does Not Inhibit Phosphorylation of eNOS-S 1179
Induced by Akt Myr -Thus far, our results showed that H89 prevents eNOS-S 1179 phosphorylation without inhibiting Akt phosphorylation, which is the key regulatory step in stimulating Akt kinase activity (35). However, we could not rule out a possibility that H89 could have inhibited the Akt kinase activity in a nonspecific manner. To examine this possibility, BAEC were infected with Ad-Akt Myr to overexpress a constitutively active Akt mutant (Akt Myr ). Cells were then treated with or without H89 for 60 min to examine its effect on Akt-dependent phosphorylation of eNOS-S 1179 . As expected, overexpression of Akt Myr stimulated phosphorylation of eNOS-S 1179 (Fig. 9). However, H89 (up to 20 M) did not inhibit the eNOS phosphorylation induced by Akt Myr . If H89 nonspecifically inhibited Akt activity, it should have inhibited the phosphorylation of eNOS-S 1179 induced by Akt Myr . However, that was not the case. Therefore, this result strongly demonstrates the specificity of H89 as a PKA inhibitor without any significant effect on Akt kinase activity.
PKA Mediates NO ⅐ Production in Response to Shear Stress-So far, our results showed that PKA mediates phosphorylation of eNOS-S 1179 in response to shear stress. Does the PKA pathway play a functionally significant role in regulation of eNOS activity (NO ⅐ production) in response to shear stress? To answer this question, BAEC were pretreated with 10 M H89, 100 nM wortmannin, or vehicle control for 30 min. Cells were then exposed to shear stress for 30 min, and nitrite accumulating in the shear medium was determined. As shown in Fig. 10, treatment of BAEC with H89 prevented shear-dependent NO ⅐ production to the same degree as wortmannin. This result provides a strong support for a critical role of PI3Kand PKA-dependent mechanisms in NO ⅐ production in response to shear stress. DISCUSSION The first significant finding of the present study is that overexpression of the dominant negative mutant of Akt (Akt AA or Akt AAA ) did not inhibit phosphorylation of eNOS-S 1179 if it was stimulated by shear stress (Fig. 5, A and C). In contrast, Akt AA and Akt AAA did inhibit eNOS-S 1179 phosphorylation when it was stimulated by VEGF (Fig. 5, B and D). These results indicate that Akt is not the protein kinase phosphorylating eNOS-S 1179 at least in response to shear stress. This result prompted us to search for the alternative mechanisms by which eNOS-S 1179 is phosphorylated in response to shear stress. The second important finding is that we identified an FIG. 6. A PKA inhibitor H89 blocks phosphorylation of eNOS-S 1179 without inhibiting Akt phosphorylation induced by shear stress. BAEC were pretreated with vehicle (Me 2 SO) or H89 (5-20 M) for 30 min before exposure to shear stress (15 dyn/cm 2 ) for 30 min. Cell lysates (20 g/lane) were analyzed by Western blot with antibodies specific for the phosphorylated forms of eNOS-S 1179 (pS-eNOS) (A), Akt-Thr 308 (pT-Akt), and Akt-Ser 473 (pS-Akt) (B). The membranes were re-probed with antibodies detecting total eNOS and Akt to monitor equal loading of samples. The phosphorylated bands were quantified as in Fig. 1, and the bar graphs show means Ϯ S.E. (n ϭ 3). Note that H89 inhibits shear-dependent phosphorylation of eNOS-S 1179 , whereas it augments both basal and stimulated levels of Akt phosphorylation. alternative protein kinase, PKA, that regulates phosphorylation of eNOS-S 1179 and NO ⅐ production.
Recent reports have clearly demonstrated the essential role of PI3K in regulation of eNOS in response to shear stress (13,14). It was further proposed that Akt, which is activated by PI3K-depedent mechanisms, is directly responsible for phosphorylation of eNOS-S 1179 and subsequent NO ⅐ production (13,14,20,31). However, this assumption has not been tested directly. Circumstantial evidence that has been used to support the role of Akt in phosphorylation of eNOS-S 1179 and NO ⅐ production in response to shear stress is following: 1) The PI3K inhibitors inhibit shear-dependent activation of Akt, phosphorylation of eNOS-Ser 1179 , and production of NO ⅐ (13,14). 2) Overexpression of the constitutively active Akt Myr mutant can increase phosphorylation of eNOS-Ser 1179 and NO ⅐ production (14,17). 3) Overexpression of Akt AA inhibits phosphorylation of eNOS on unknown Ser residues (20). These findings clearly demonstrate that the shear-dependent activation of eNOS is regulated by the PI3K-dependent mechanisms. Most of these findings have been reproduced, expanded, and confirmed in the FIG. 7. ERK pathway does not regulate shear stress-dependent phosphorylation of eNOS-S 1179 . Confluent BAEC were pretreated for 30 min with increasing concentrations of PD98059, a MEK1/2 inhibitor, and then exposed to shear stress (15 dyn/cm 2 ) for 20 min. Phosphorylation of eNOS, Akt, and ERK was determined by Western blot as described in Fig. 2. As controls for equal loading, the same membranes were re-probed with antibodies that detect the total amount of each protein. The phosphorylated bands were quantified as in Fig. 1, and the bar graphs show means Ϯ S.E. (n ϭ 3). were analyzed by Western blot with antibodies specific for the phosphorylated form of eNOS-S 1179 (pS-eNOS) and pS-Akt and pT-Akt. The membranes were re-probed with antibodies detecting total eNOS and Akt to monitor equal loading of samples. In A, the phosphorylated bands (pS-eNOS and pS-Akt) were quantified as in Fig. 1, and the bar graphs show means Ϯ S.E. (n ϭ 3). current study (Figs. 1-3). However, these results alone are not sufficient to conclude that Akt mediates phosphorylation of eNOS-S 1179 in response to shear. None of the above studies performed a key experiment whether direct and specific inhibition of Akt (i.e. use of dominant negative Akt mutants) would result in prevention of eNOS-S 1179 phosphorylation in response to shear stress.
We showed recently that overexpression of Akt AA blocks shear-dependent stimulation of NO ⅐ production (31) as confirmed in the current study (Fig. 4). In fact, this result also led us to assume initially that the Akt AA effect would be the direct result of inhibiting eNOS-S 1179 phosphorylation until we examined this issue in this study. If Akt were indeed responsible for the shear-dependent phosphorylation of eNOS-S 1179 , expression of dominant negative Akt mutants in BAEC should have inhibited it. However, the Akt mutants failed to inhibit eNOS-S 1179 phosphorylation when it was stimulated by shear stress (Fig. 5, A and C). It could be argued that the failure of Akt AA or Akt AAA to inhibit the shear-dependent phosphorylation of eNOS-S 1179 is due to their insufficient expressions or ineffectiveness as dominant negative inhibitory mutants. However, this possibility is not likely for the following reasons: 1) The same Akt AA or Akt AAA mutants completely blocked sheardependent NO ⅐ production in the same experiments (Fig. 4). 2) The identical mutants effectively inhibited phosphorylation of eNOS-S 1179 if it was stimulated by VEGF in the same studies (Fig. 5, B and D). Altogether, as well as in previous reports (17,22), these findings clearly demonstrate that Akt AA and Akt AAA constructs used in these studies effectively inhibited Akt activation.
The differential effects of the dominant negative Akt mutants on eNOS phosphorylation (Fig. 5) provide strong evidence supporting a new concept that phosphorylation of eNOS-S 1179 can be regulated by the Akt-independent as well as -dependent mechanisms depending upon each stimulus. In the case of VEGF, we confirm that phosphorylation of Akt is required for phosphorylation of eNOS-S 1179 . In the case of shear stress, however, Akt does not play an essential role in the phosphorylation of eNOS-S 1179 . Then, how does shear stress regulate phosphorylation of eNOS-S 1179 in an Akt-independent manner?
It is important to emphasize that Akt is not the only downstream target of PI3K in endothelial cells. It has been demonstrated that PI3K activates phosphoinositide-dependent kinase-1 (PDK1), which phosphorylates and activates not only Akt, but also many other target kinases, including PKA, PKG, PKC, serum-and glucocorticoid-inducible kinase, and p 70 S6 kinase (36,37). Moreover, it is becoming increasingly clear that eNOS-S 1179 can be phosphorylated by other protein kinases, including PKA, PKG, and AMP kinase in addition to Akt depending on each stimulus (24,26,27). Therefore, we began to screen the role of other protein kinases in shear-dependent phosphorylation and activation of eNOS.
In this study, we examined two protein kinases, PKA and ERK, because they have been implicated in regulation of phosphorylation and activity of eNOS (27,34). First, we showed that the ERK1/2 pathway (by using the MEK1/2 inhibitor) does not play a significant role in shear-dependent phosphorylation of eNOS-S 1179 (Fig. 7). This is consistent with our previous finding that inhibition of ERK1/2 pathway by treating BAEC with pertussis toxin (inhibitor of G i/o -protein family) had no effect on shear-dependent NO ⅐ production (31). In comparison, the ERK1/2 pathway has been shown to stimulate eNOS phosphorylation at sites other than the Ser 1179 residue in response to bradykinin, and that inhibition of ERK1/2 activity stimulated eNOS activity (34). Taken together, these results suggest that the ERK1/2 pathway is not involved in phosphorylation of eNOS-S 1179 .
Next, we found that PKA plays a critical role in shear-dependent phosphorylation of eNOS-S 1179 in BAEC. Treatment of BAEC with H89 blocked shear-dependent phosphorylation of eNOS-S 1179 without inhibiting Akt phosphorylation (Fig. 6, A  and B), suggesting a role for PKA. It was further demonstrated that stimulation of PKA by using a cell-permeable cAMP analog, 8-Br-cAMP, alone maximally stimulated phosphorylation of eNOS-S 1179 within a few minutes (2-5 min) (Fig. 8A). It should be noted that Akt phosphorylation was not at all stimulated by 8-Br-cAMP during the same time period (Fig. 8A). Only after 15 min or longer incubation, 8-Br-cAMP modestly increased phosphorylation of Akt (Fig. 8A). These results clearly illustrate two points that phosphorylation of eNOS-S 1179 1) does not have to require Akt and 2) can be regulated by a PKA-dependent manner in response to shear stress.
In our study we tested whether the effects of H89 on phosphorylation of eNOS-S 1179 were due to its direct inhibition of Akt activity. If this possibility were true, treatment of cells with the inhibitor would have prevented the eNOS-S 1179 phosphorylation induced by the constitutively active Akt. However, this possibility is highly unlikely, because H89 did not have any effect on phosphorylation of eNOS-S 1179 if it was induced by Akt Myr (Fig. 9). Interestingly, we observed that H89 treatment alone strongly increased phosphorylation of Akt (Figs. 6B and 8B). At this time the mechanism underlying the stimulatory effect of H89 on Akt phosphorylation is not known. One of our speculations is that basal PKA activity may be involved in controlling dephosphorylation of Akt and that H89 may inhibit The membranes were re-probed with antibodies detecting total eNOS and Akt to monitor equal loading of samples and expression of Akt Myr , respectively. H89 has virtually no effect on eNOS-S 1179 phosphorylation induced by Akt Myr overexpression. the dephosphorylation pathway. Consistent with this speculation, PKA has been shown to regulate protein phosphatases (27,38). Nevertheless, what is clear at this point is that, even when Akt phosphorylation was maximally stimulated by H89, eNOS phosphorylation was completely inhibited. This argues against the role of Akt as the protein kinase directly phosphorylating eNOS-S 1179 .
Collectively, our results indicate that PKA plays a critical role in phosphorylation of eNOS-S 1179 . In support of this finding, eNOS-S 1179 has shown to be phosphorylated either directly by PKA catalytic subunits in vitro or by treatment with isobutylmethylxanthine (increases cAMP level in cells) in BAEC (25,27). However, the significance of PKA-dependent phosphorylation of eNOS-S 1179 in response to physiological stimuli has not been determined until the current study. Our study establishes for the first time that PKA pathway plays a critical role in eNOS-S 1179 phosphorylation under a physiologically relevant condition, shear stress. Furthermore, we established the functional significance of PKA pathway in regulation of eNOS activity (NO ⅐ production) under shear stress condition. We showed that treatment of H89 prevented shear-dependent NO ⅐ production in BAEC (Fig. 10). In support of our finding, the cAMP and PKA pathway has been shown to induce NO ⅐ production in isolated arteries (39).
One potential mechanism by which shear stress stimulates NO ⅐ production is through activation of PKA, which in turn regulates phosphorylation of eNOS-S 1179 directly or indirectly. However, it is not clear at this time whether phosphorylation of eNOS-S 1179 is directly responsible for NO ⅐ production in response to shear stress. This caution is especially important in light of the previous reports showing that NO ⅐ production from eNOS can be regulated independently of the phosphorylation status of eNOS-S 1179 (26,35). For example, bradykinin has been shown to stimulate phosphorylation of eNOS-S 1179 as well as NO ⅐ production (26,34). However, treatment of the cells with a PI3K inhibitor (wortmannin) did not inhibit NO ⅐ production, whereas it blocked the phosphorylation of eNOS-S 1179 in response to bradykinin (26,34).
Interestingly, we found that expression of Akt AA inhibited shear-dependent stimulation of NO ⅐ production without affecting eNOS-S 1179 phosphorylation (Figs. 4 and 5). These findings demonstrate that, despite the lack of its effect on eNOS-S 1179 phosphorylation, Akt still is a critical mediator of shear-dependent NO ⅐ production. Then, how does Akt regulate eNOS activation in response to shear stress without regulating phosphorylation of eNOS-S 1179 ? One potential mechanism is that there are unidentified amino acid residues in eNOS that can be phosphorylated by Akt. Alternatively, Akt may activate other protein kinases and phosphatases, which then regulate phosphorylation of eNOS on other sites. In support of these ideas, in addition to Ser 1179 , eNOS has been shown to contain several other phosphorylation sites, including Ser 116 , Thr 497 , and Ser 635 (14,17,19,26,27,40) and some unknown sites as well (20,34). Another possibility is that Akt may regulate eNOS activity by regulating other regulatory molecules such as caveolin, CaM, and heat shock protein 90 (10,41). These speculative ideas await further studies.
Based on our results as well as previous findings reported by other investigators, we now propose a following scenario by which shear stress regulates eNOS phosphorylation and NO ⅐ production as depicted in Fig. 11A. Shear stress stimulates activation of PI3K, which in turn activates PDK1/2. PDK1/2 is then proposed to stimulate both Akt and PKA. PDK1 has been shown to phosphorylate and activate PKA (36), although it has not been studied whether shear stress activates PKA by the same mechanisms. Establishing this pathway will require further studies. PKA is then proposed to stimulate phosphorylation of eNOS-S 1179 directly or indirectly. In addition, Akt is proposed to regulate NO ⅐ production by unknown mechanisms, including phosphorylation of eNOS at sites other than the S 1179 residue. As comparison, VEGF-dependent stimulation of eNOS-S 1179 phosphorylation and subsequent NO ⅐ production by the PI3K/Akt-dependent pathway are shown in Fig. 11B.
In conclusion, the current study demonstrates that both PKA and Akt play critical roles in regulating phosphorylation of eNOS-S 1179 and the enzyme activity (NO ⅐ production). However, the mechanisms by which these two protein kinase pathways regulate eNOS phosphorylation and enzyme activity seem to be quite different, depending on the given stimulus. In the case of shear stress, phosphorylation of eNOS-S 1179 is regulated by a PKA-dependent, but Akt-independent, mechanism, whereas the NO ⅐ production is regulated by the mechanisms dependent on both PKA and Akt. The activity of eNOS in cells may be controlled through a coordinated regulation and interaction between the two protein kinase pathways, Akt and PKA. FIG. 11. Proposed roles of PKA and Akt in eNOS regulation by shear stress. A, shear stress is proposed to activate PI3K and PDK1/2 in sequence, which in turn stimulate both Akt and PKA. PKA is proposed to regulate phosphorylation of eNOS-S 1179 directly or indirectly, whereas Akt phosphorylates eNOS at sites (?) other than S 1179 residue. A coordinated regulation of eNOS by both the PKA and Akt-dependent mechanisms is proposed to result in production of NO ⅐ . The link between PDK1/2 and PKA has not been established as described in the text. B, in contrast, VEGF stimulates eNOS-S 1179 phosphorylation by PI3Kand Akt-dependent mechanisms. Specific inhibitors (wortmannin, Akt AA , H89) and stimulators (cAMP and Akt Myr ) are indicated.